STATEMENT AND

INFERENCE
 Oxford University Press
London Edtnbu,gh Glasgow Copt'1'hagen
New rork 'Toronto Mel.bourne Cape Town
Bombay Calcutta MaJra1 Shanghai
Humphrey Milford Publisher to the UNIVtRSITY
STATEMENT
.. AND
INFERENCE· )

WITH OTHER PHILOSOPHICAL PAPERS

BY

JOHN COOK WILSON

M.A, HoN, LL.D. (ST. Air;R~ws), F.B A.
SOMETIME WYKEHAM PROFESSOR OF LOGIC IN THE UNIVERSITY
or OXFORD; FELLOW OF NEW COLLEGE. HON FELLOW OF ORIEL

EDITED FROM THE M~S., &c.

BY

A. S. L. ~~RQUHARSON)
Fl;LLOW OJ' UNIVERSITY COLLEGE

WITH A PORTRAIT, MEMOIR, AND SELECTED

CORRESPONDENCE

IN TWO VOLUMES

VOL. I '

-
OXFORD
AT THE CLARENDON PRESS
1926
 Pnnld ,,. E11gl&ttd
Al Ille OXFORD UNIVERSITY PRESS
Ily John J olm.,,,.
/'mlU, lo tlle Unwe,isly
 TO
MY MOTHER AND MY FATHER'S MEMORY

'From the memory and good name of my father,

honour and manliness. From my mother, to serve God
and give to man , ID shun not every evil deed only but
every evil thought. Simplicity of heart and home, and
to keep far from the path of riches.'
M. AURELIUS ANTONINUS.
 CONTENTS OF VOLUME I
• PAGE
TO THE READER vii
MEMOIR xii
ELENCHVS OPERVM (List of Published Works and Con-
tnbutions to Reviews) • lxv
TESTIMONIA lxxiii
FAMILIAR LETTERS, 1859-1914 lxxvii

STATEMENT AND INFERENCE

A_nalysis of Contents 3

PART I. INTRODUCTORY
I. The Definition of Logic 24
II. The Relation of Knowmg to Thmkmg . 34
III. Logic and Cognate Studies. General and Special
Logic 48
IV. Logic and Theones of Knowledge and Reahty 60

PART II. STATEMENT AND ITS iRELATION TO

THINKING AND APPREHENSION
I.Apprehension in general 78
II.The use of the term' Judgement' in Modern Logic . 92
III.Opinion, Conviction, Bebef and Cognate States 98
IV. The Distinction of Sub1ect and Predicate in Logic
and in Grammar II4
V. The Analysis of the Statement or Proposition into
Sub1ect and Predicate 127
VI. The Confusion of Predication with Objective Rela-
tions 138
iv CONTENTS
PAGE
VII. Certain Objective Distinctions and Doctrines of Pre-
dication . 149
VIII. The Meaning of Grammatical Forms 170
IX. The Symbolization of Forms of Statement 192
X. The Copula and the Modality of Statements 212
XI Synthetical and Analytical 'Judgements'. The
Relat10n of Propositions 231
XII. Negation or the Quality of Propositions. 247
XIII. Erroneous Attempts to define' Judgement' 274
XIV. Apprehension, Conception, and Statement 295
XV. The Quantity of Proposit10ns and the Universal 330
XVI. Classification 354
XVII. Definition . 377
XVIII Denotation and Connotation 386

CONTENTS OF VOLUME II
PART III. INFERENCE
I. The General Nature of Inference 412

II. The Syllogism 435

III. Principles and Method of the Pure Demonstrative
Sciences . 454
IV. The Universal Character of Inference and its Rela-
tion to Immediate Apprehension 479
V. Simple and Complex Ideas and Problematic Con-
ceptions . 491
VI. Hypothetical Statement and Hypothetical Argu-
ment 525
VII. Direct and Indirect Argument. The Fictions of
Modern Mathematics 553
VIII. The Mathematical Expression of Probabibty and its
Relation to Reality . 569
 CONTENTS V

PART IV. SPECIAL LOGIC

!AGE
I. Induction. The General Form of the Experimental
Methods . 578
II. Defects of the Modern Theory of Induction. The
Formal Cause in Bacon • 596
III. The Axioms of Modem Induction . . 6o6
IV. Failure of Empiricism to explain the Laws of Thought 616
V. The Method of Physics and its Relation to Pure
Mathematics 631
VI. Symbolic Logic . 635

PART V. TENTATIVE INVESTIGATIONS AND

PHILOSOPHICAL CORRESPONDENCE
I. The Nature of a' Thing' 664
II. The Universal and its Differentiations 670
III. Modern Fallacies about the Universal . 677
IV. Classification of Classes 683
V. Predication predicated of itself . 689
VI. Relation and Quality 692
VII. Categories in Aristotle and in Kant . 6g6
VIII. Universals. Last words 707
IX. Trichotomy (Letters to Professor H. Goudy) 71S
X. Correspondence with B. Bosanquet 728
XI. On the Notions of a Class and of Classes . .752
XII. Primary and Secondary Qualities . 764
XIII. Letters to Mr. H. A. Prichard, Fellow of Trinity
College, Oxford Box
XIV. Later correspondence with B. Bosanquet . 818
xv. Divine and Human Consciousness . 830
XVI. Rational grounds of Belief in God • . 835
2773•1 a3
vi CONTENTS
PAGE
POSTSCRIPT . . 870
SOURCES AND DATES OF THE SECTIONS OF THE
LOGIC LECTURES 885
CHRONOLOGICAL TABLE 888
INDEX TO PARTS J-IV • 891

LIST OF ILLUSTRATIONS
John Cook Wilson. From a photograph by Elliot and Fry
(see Letter No. 70, p. clxii) . Frontispiece
J. C. Wilson playing with Mr. Prichard's sons
(see Letter No. 31) • To face p. cxvii
 TO THE READER
THESE volumes embrace what seemed most valuable and
best suited to publication of the late Professor John Cook
Wilson's philosophical lectures and speculations. His reputa-
tion for Greek scholarship stood so high with his contemporaries
that a strong desire was expressed that his best work in this
field also should be colJected under one view. Expense and
the inchoate condition of his unpublished critical studies have
prevented this. Two important sets of lectures were accessible
in pupils' note-books and were put generously at my service,
To fit them however for the learned public would have
demanded laborious recasting, while the student has easy
access to invaluable work in English on both these topics,
Plato's Republic and Aristotle's Theory of Demonstrative
Science.
His collections upon the Greek Scriptores Musici and Tactici
and his careful studies in Plato's later dialogues only Wilson
himself ccmld have arranged.
It seemed therefore best simply to make a list of his published
works, with a note of the places where tbey are to be found
and an indication of their nature. This I have done
His logic lectures, on the other hand, could be presented in
a form which does not indeed do justice to his powers but
will, it is hoped, be of assistance to the philosophical student.
Moreover, he had himself clearly contemplated their publica-
tion. They are supplemented by two special courses, by
selected passages from his memoranda-especially a paper
which he was preparing to read to the British Academy-and
by his philosophic correspondence. The latter has, it will be
viii To THE READER
seen, the imprimatur of no less an authority than Bernard
Bosanquet.1
The sources of this work I have indicated in a Postscript,
where I have also set out the approximate dates of the several
constituent parts. This will explain and perhaps justify in-
equalities and inconsistencies of expression and even, at times,
of doctrine. These could hardly be removed without serious
danger of misrepresentation. Some recompense will be found
in the light thrown on the path which led Wilson to his goal.
I desire here to thank those gentlemen who so freely put their
lecture notes at my disposal. In the end I had an embarras
de choix.
My little Memoir is written dehberately from the point of
view of a single observer. To correct its deficiencies or
partiality I have added a selection from Wilson's familiar
letters. These are meant to give ' some idea of what was in
the man's mind in its original unmitigated form, without
apology or attempt to soften it down '. 2 I might add ' what
was in his heart ' also. The series covers, so far as was possible,
his hfe from childhood to age. I have had no thought by my
choice to idealize the writer or to disguise his amiable weak-
nesses. Wherever I could I have given the letters entire.
Part of Boswell's secret lies in the scrupulous setting-down
even of the trivial and commonplace. In these things we
want a man, not a hero or a lay figure. My only omissions
have been made to avoid giving pain to relatives or to living
writers. I have left out some characteristically intemperate
outbursts provoked by differences of opinion on philosophic
themes, not because I desire to represent a passionless sage
but because they would not be understood. Otherwise the
1 Part V, x, xiv, esp. p. 826.
1 R. L. Nettlesh1p of his own account of Plato.
 To THE READER ix
letters are nearly exactly as they were written, punctuation,
spelling and all. The more scientific letters are in Part V ;
their length and the care given to their composition display
very successfully the immense time and ardour Wilson lavished
in his friends' assistance and in the attempt to win converts
to his way of thinking. To illustrate his singular generosity
m this regard I have been old-fashioned enough to include
a selection of Testimonia.
The work has cost me most of my leisure hours in the past
four years. That the result is very imperfect I do not need
to be told, but I was anxious not to delay the book too long.
For obvious reasons and others which do not concern the
public I could not get to work until the Easter of r92r. It
will be recalled that Green's Remains took a long time pre-
paring and other examples will readily occur. Yet Nettleship
found Green's works either already printed or in their manu-
script form ' in general, contmuous and coherent '. Tins was
not my good fortune. Wilson left a great mass of material.
He himself called hts handwnting ' impressionist '. Thus the
mere ocular labour of reading the manuscnpts through has
been most exacting. Moreover some parts, and those impor-
tant, were in great disorder, and age increased his inveterate
tendency to discursiveness and prolixity. In editing an intri-
cate and abstract discussion it is very hard, I have found, to
be always certain where one's author is repeatmg htmself and
where he is merely making one more parallel in his laborious
advance. To o:nut or to rearrange may involve serious mis-
representation. Mr. J.C. B. Gamlen however, my co--executor,
desired me to use my judgement in all these matters, and I am
in consequence entirely responsible for any changes made in
the order and expression. Certainly the book appears to m~
to give a sufficiently exact view of Wilson's faith.
X To THE READER

I desire to thank those who have been so kind as to lend

me letters and manuscript essays, especially Professor G. F.
Stout, F.B.A., Mr. H. W. B. Joseph, and Mr. H. A. Prichard.
The Editor of Nature has courteously permitted the republica-
tion of part of a letter upon Inverse or a posteriori Probabihty,
and the late Bernard Bosanquet, F.B.A., most readily sent all
Wilson's correspondence with him for the use of an editor.
The late F H. Bradley, O.M , with whom I more than once
discussed the publication of Wilson's criticism of himself, said·
' Why not ? it may help the sale of my own book ; but I
cannot promise to read 1t' Mr. W. H. Fyfe, Head Master of
Christ's Hospital, and Mr. Gamlen were good enough to criticize
the Memoir. I have, m conclusion, to thank the Delegates
of the Uruversity Press for undertakmg, in these hard times,
the publication of a book designed to do honour to a dis-
tinguished occupant of the Wykeham Chair of Logic, and the
Staff of the Press for the care they have given to its pro-
duction. One of the Delegates, Mr. W. D. Ross, Deputy
Professor of Moral Philosophy and Fellow of Oriel College,
has helped me very much by readmg the proofs and making
valuable corrections. I am grateful to Dr. A. C. Ewing,
formerly Exhibitioner of my own college and sometime Senior
Demy of Magdalen College and Bishop Fraser's Scholar of
Oriel College, for making the Index, and to Messrs. Elliot
& Fry and to Mr. Prichard for permission to reproduce
photographs, each in its kind so perfect.
A. S. L. FARQUHARSON.
3 GROVE PLACE,
OXFORD,
Zflth August z9a5
MEMOIR
 MEMOIR
Laboriosus et diuturnus sapientiae Miles.
ALMOST all we have of Epictetus, the Stoic philosopher, are
lecture notes which a Roman governor gave to the learned world
of the second Christian century, with an apology for their want
of form and fintsh. In the ded1cat1on Arrian says that ' when
Epictetus spoke, his purpose was to move lus audience to
virtuous ambition · he had no second thought If then thcs<.'
memoirs do the same, they will do what a philosopher's words
should do ; if not, you must believe that when the Master spoke
he forced the class to feel what he meant them to feel '
Any reader of the lectures here prmtrd, who wai. privileged
to hear them dchvered, will understand what Aman n1eant.
He will miss the living voice Not that Wilson had any rare
graces of gesture and delivery, but that he could communicate
lus own earnestness to his hearers. ' He had a tone in oral
delivery which seemed to convey sen'le to those who were
otherwise imperfect rcc1p1ents ' 1
Wilson was not a moralist but a logician, yet he so touched
what appear to many the dry bones of logical doctrine that he
was as one contending for moral verity. This was clear as
he read his lecture, clearer in his mformal dass, clearest of all
m conversation beyond the school. Thus it was well said by
a colleague, 2 not a professed philosopher, that, whatever the
theme, Wilson was the best talker he had known, because all
he heard and said he took so seriously. He forced his class to
feel what he meant them to feel ; that is surely the sign and
secret of the great teacher.
1 C. Lamb of S. T. Colendge, Edmonton, 21.xi 1834 The Rev L. R
Phelps, Provost of Onel, one of Wilson's first pupils. says 'he made
philosophy absorbingly interesting and, yet rarer gift, made you tlunk 1t
really mattered ' (letter to H W. B J , 19.11 17).
• Mr L L. Pnce, late Fellow and Treasurer of Oriel.
 MEMOIR •,..,h,xiii
...
r "
The breath has not altogether left the body of these disc~
Unfinished and uncouth as they remain, they still exhibit unmis-
takably the height which this vivid and indefatigable spirit
had won in half a century of arduous and perhaps too solitary
devotion to Truth. Behmd the inconsistency of mcompleteness,
amid arid tracts of controversy, we may trace the footsteps of
a persistent mvestigator, the purpose of a livmg mind.
This then, which he reached so hardly and held so tenaciously,
1s what Wllson would have desired to live; beside this the
temporal record of his days must appear tr1v1al, almost irre•
levant. Yet men m the busy world, who once heard him and
who, conung on ,m old note-book, c1.re surprised lo find how
much they hc1vc forgotten,1 m.iy care to learn c1. httle more
,tbout their old m,tster ; Im, many friends m Oxford will piece
out the imperfection of an attempt which piety has dictated
I am to try to depict, ,1s I knew them, not only that eager
absorption m the problems of lus Chair , that boyish confidence
m the general mtercst of lus own pursuits (the zeal which so
easily besets the student} , the fierce mtellectual hatreds, the
Jealousy for truth, the uncomprom1smg exposure of what he
deemed imposture ; but the httlc things also that endeared
him to us ; lus chase of the impossible, Ius volunteering
enthusiasm, his confidence m your sympathy, his everyday
goodness, s1mpl1C1ty and human kmdncss; all the minute
diversities that crowded and dissipated the tireless sequence of
his hours.
Wilson was already a professor and a philosophic force in the
Oxford m which I found myself as a freshman in 1890. Already
he was, mamly by his Volunteer activity, a familiar figure round
whom legend began to form. Some said, for example, that
Mrs Wilson had been a Fraulein Moltke, others that her father
was Hermann Lotze, the philosopher-and this I have seen in
prmt. For the men who read Greats he was fast becoming an
intellectual stimulus which recalls the power of Socrates over
Athenian youth Barely turned forty, he appeared already old,
save m voice and movement. Below medmm height, his small-
• Suggested by words of S1r Herbert J Creedy, KC B, Under-Secretary
of State for War He lent me lus copy of Wilson's Lectures o# Inference,
with some of Wilson's autograph corrections.
xiv MEMOIR
neu was emphasized by his broad forehead and full philosopher's
beard, red once, now fast turning that beautiful white which in
red hair is the compensation of age. His keen yet pensive eyes
looked out from under dark, strongly defined, eyebrows ; his
short firm step and rapid motions were an index to a decided
and impetuous character. He would be seen riding most reek•
lessly through the traffic on a bicycle, a rare vehicle in those
days for gownsmen, young or old, though a few years later it
brought Wallace to his untimely end.1 On Sundays you might
meet him walking in the Parks or Mesopotamia, in a double-
breasted frockcoat and a minister's round felt hat with guard
attached, looking much what one thought a German professor
might be When the University Corps paraded, he led a small
section of cyclists, in blue serge knickerbocker uniform and
white spats, and we came upon him in the vlllages round Oxford
or saw him stand on Cumnor Hirst, the centre of a miniature
staff, directing the manceuvres of the sadly small companies of
enthusiasts who m the 'nineties kept alive the tradition estab•
lished by Warre of Eton.
' Id vas an audumn afdernoons, vay down in '89,
De pully poys of Oxford vas geranked 1n pattle line
All brebared for vight and ploonder, und 'tvas peautiful to see
De phdosopede gontmgent und de footman-cavallrie.' 1
In those days he lectured in a converted racquet-court in
Oriel Street He first discoursed very quickly and fluently but not
very clearly, so far as delivery went, on the day's topic; using
frequent blackboard illustrations from geometry, and usually
overstepping the thirty-five to forty minutes he allowed himself.
He then paused and said : ' I will dictate a paragraph.' This
he read from the note-book 3 of an earlier pupil so rapidly that
it was impossible to get 1t all down and versions had to be
collated after the lecture. He was so much engrossed that he
failed to appreciate our difficulty. In later years I remonstrated
1 Wilham Wallace, White's Professor of Moral Philosophy, was killed

by a fall from his bicycle 1n the spnng of 1897.

• From 'Des (SIC) Kochmannshed ', the 011/ortl M41ruir1e, 4.xi1.89.
Wilson served in the OUR V Corps (1st Vol. Bn. Oxfordshire L.I.) from
..--._1904 Lieut. 1889 ; Capt 1891 Started a Cyclist Section 1889;
P.-s: iTgc,; T 1891; retlred as Hon. MaJor with V.O. in 1904.
• At that date, the notes of Prof. W. A. Cra1gie.
 M:IU,tOilt xv
with him on this but he only replied: ' What they couldn't
hear was probably not worth hearing.' His appearance and
gestures were an odd mixture of the serious and the comic,
recalling Boswell's description of Johnson, only here there was
a tiny cockboat of a man, there an unwieldy galleon. Never•
theless, his force of character constantly prevailed over any
tendency to levity, and his power over the class may be gathered
from a single narrative. He had mtroduced into his lecture
a reference to the mathematical calculus. The clock was on the
stroke of one when he ceased to dictate ; then without warning
and looking fixedly at a man m the first row, he said : ' Do
you all understand the calculus ? ' Under cover of the silence
which followed he walked smartly to the door, locked, or pre-
tended to lock, 1t, and then standing there with his back to it
said with decision · ' No one shall leave this room until you all
grasp the essentials of this simple matter ' And then and there
he returned lo the blackboard and began his demonstration,
until (such was his ardour and so late the hour) all in the room
appeared convmced that the veil was lifted from what had been
and remained to most, I fancy, a region quite unknown. It
was hke the hghtnmg flash and then the outer darkness. Yet
philosophy too has its phenomenon of conversion and such
moments may irradiate the soul.
The mtellectual enthusiasm of this, the prophet's faith in your
powers and his own, the lack of proportion, even the occasional
absurdity, were part and parcel of Wilson's native composition
and are represented m the lectures now prmted. What is absent
1s the wonderful energy diffusmg itself from. teacher to hearer ;
now awakening and dlummatmg, now dazzling or numbing him,
not however so much by the qmck verbal fence of a Socrates
as by the depth and seriousness of his own conviction. He
compelled you, at least at the moment, to assent 1£ not to
comprehend. No one thirty years ago was a better antidote
to a certain spirit of irony, of intellectual suspension, almost of
scepticism, which ruled m Oxford, easily charming the restless
mind of adolescence, so lately released from the discipline of
school. Destructive his dialectic was, too polemical perhaps,
yet positive and bracing by contrast with those cold negative
currents which filled the air; the irony, say, of Arnold or of
 xvi MEMOIR
Nettleship, the higher criticism and the many contemporary
solvents of inherited belief.1
' We have but faith: we cannot know;
For knowledge is of things we see.'
So sang the Laureate from hts intellectual throne; but Wilson's
faith was in knowledge ; the essence of his creed a knowing
which saw its obJect and apprehended what it saw.
In 1892, the philosophy professors m Oxford were William
Wallace, Wilson, and the late President of Corpus. Mr Case
was lecturmg chiefly on Bacon and Aristotle. He had lately
published his Physical Realism and the doctrme was opposed to
the current views Re.tlu,m of any kind w.i.s not then the
fashion m Oxford Wallace lectured in Mor.i.l Philosophy,
prmc1p.1lly on the period from Kant to Schelling, blending
much metaphysic with 111s morals Harsh m voice, rugged and
remote m manner, lus lcarnmg and literary allusions, his rare
grace of style, softened his moral rigour Suggestive, appre·
ciat1ve, seldom dogmatic, unmethodical and discursive, hts
lectures flowed on without begmmng or end Those were the
pleasantest hours of the week. Wilson was a valued complement
to his colleagues. Difficult to hear and to follow, confused in
expression and at times m the sequence of ht& argument, severely
unadorned m dict10n, he yet seemed to develop his subJect from
the beginning. Shtrkmg no crooked questions, he seemed in
the end to make some rough places smooth. His confidence was
comforting to begmners. He was still under the strong mfl.uence
of Kant and Green, hts logic purer and less contaminated with
the theory of knowledge than it later became. To omit from
this account the writings of Mr. F. H. Bradley would be to
ignore the most stimulating of all influences, at that date, to
the novice.
Wilson discoursed on logic for three hours in each week and
during each successive term. He began in the fall of the year
1 The phrase was suggestod by ReJ,gious Cllanges ,n 011/ord dunng tie

last Fifty Yeal's, by R W Macan, D Litt, Oxford, 1917; My remarks an,

the recollections of youthful impressions; a broader and saner outlook will
be found in my old tutor's paper. Dr Macan says little of the lugl\_er
cntica. I viV1dly recall Dr T. K Cheyne's bold and eloquent sermons
delivered as Canon in residence at Rochester, a current of mountain air
fl.owing into a rather confined and narrow atmosphere.
 MEMOIR xvii
and concluded, or rather failed to conclude, on the eve of the June
examination. He gave, besides, informal instruction, two hours
once a week. Few men commenced, fewer still continued this.
He was less known or less regarded then by the college tutors.
Later, old pupils, now teachers themselves, sent their men to
this class and it became an important feature of the philosophical
course. I remember however one year when, after the first
week, the present Bishop of Carlisle 1 and myself were the only
attendants. Mrs. Wilson was ill and away from Oxford and he
did not always keep his appomtment. He took us at Oriel, m
the set we understood had been Cardmal Newman's. He was
informal in the traditional Oxford manner ; we two, each in
a large arm-chair, he standmg before the fire. We were denied
the weakness of a note-book. He stood up or walked about,
thmking aloud. He would invite a problem and then ' develop '
it at leisure, in two or three successive meetmgs. He had a way
of becommg rapt m a brown study, sometimes for five minutes
together. Then he would relax his neck muscles, 11 his head
bowed forward very much and his eye dimmed and lost its
expression. You were present at what seemed a philosophic
travail. Sometimes he gripped the mantelpiece, swung himself
back and looked long and very earnestly at a portrait of Jowett,
staring through 1t rather than at it. 3 Sometimes he walked to
the wmdow and gazed mtently and sadly at what I supposed
was the evening hght on Merton Chapel tower.
' He would take, in his informal instruction, what men thought
simple questions and show how much lay in their solution. His
mind, said a pupil, was like a vice ; so hardly could slovenly
thought or uncriticized phrase escape from the tenacity of his
attack.' 4 Indeed he seemed always to be trying to disentangle
the question on its own merits, and I believe that he was
1
The Right Rev H H Williams, D D., Hastings Exhibitioner of The
Queen's College, Oxford, afterwards Fellow of Hertford and Principal of
St Edmund Hall, a pupil and intimate fnend of Wilson
' This has beon observed m many teachers Philostratus in his Lnies
of the Sophists notes it I have road the same of Schelling.
1
These charactenstics are more racily descnbed by S Ball in The Onel
Record (Sept 1915), vol 11, p 248, and in the Oxfcwd Magannt: of 22 x 15.
• From the Proceedings of the British Academy, vol vu, by Mr. H. W. B.
Joseph He klndly lent me his matenal and gave me leave to use his JUSt
and elegant mem01r (Referred to as H. W B J.)
2 7731 b
xvtli M-EMOIR
scrutinizing his own thought rather than ours. And he varied
very much. Some days he certainly repeated himself and went
round and round like a man skating figures. So have I watched
him cutting very small circles on the ice in the meadow on
Iffl.ey Road. In a letter 1 he himself speaks of a later occasion
when he was ' very tired and really only talked mechamcally,
and without truly attending '. That was sometimes his case
even then; mainly, I think, from domestic anxiety. He had
a way of holding his breath at intervals and then releasing it
in a hmg and loud expiration, and he would wind up one hair
of his beard and break 1t with a vicious snap. He said, on his
portrait being painted, that his friends complained that his
beard was too short. ' They don't understand that my beard
varies. When I am working at a problem something has to
come out. If the problem doesn't, the beard must.' If a subject
interested him peculiarly, especially when he was polemical, he
would reveal signs of physical excitement, signs which grew
more marked with age. The preliminary signal was the snapping
of his fingers in a scrie'f; of rapid pistol shots from behind his
back, and he was often betrayed into unphilosoph1cal explosions.
Thus : ' The time has come for stronger language . what the
devll does the fellow mean ~ ' When discussmg the paralogism
of Achilles and the Tortoise, he would break out, almost as
though Zeno were in the next room : ' What docs the fellow
mean by never ? '
Thus he attempted to impress upon his clasc; the serious
nature of the quest for Truth and revealed by his own practice
the method of pursmt ; he never deluded his pupils with ready-
made result or easy epigram and so, when he reached a con•
clusion, he was most confident of its certainty. Naturally these
methods did not suit all men ; a few, even some with a gift for
philosophy, soon deserted his lectures and many found the
informal instruction, especially as he grew older, too much of
a soliloquy and with too little attention to their own problems.
One of these said : ' If the virus did not bite, it made one
immune,' and laid his own immunity complacently to the pro-
fessor's account. But ordinary youth is tolerant of foibles in
a respected teacher and these idiosyncrasies of Wilson appeared
1
To Mr. H. A Pnchard. (Referred to as H. A. P.)
 MEMOIR xix
to most an inseparable accident of his devotion to their con-
version. One last reminiscence before I turn to his earlier years.
About ten men were present at the first informal class of that
year. He was treating by request the Kantian paradox . ' the
mmd makes nature, the material it does not make.' He paused
in his familiar manner and bending forward looked fixedly in
the face of a Balliol man in a ragged scholar's gown. He,
supposing himself to be interrogated or in a spirit of solemn
miscluef, blurted out : ' But why shouldn't that table be there,
Just where we see it?' Silence attended the result. The pro-
fessor sprang once into the air ; said very fiercely indeed :
• Why shouldn't it?' and then relapsed mto reverie. The
scholar never returned, but I have sometimes wondered whether
the shock set Wilson determmedly to work clearing the path
which after many days led him far from the idealist solution he
then accepted or appeared to accept. I say ' appeared ', because
even then he did not altogether acquiesce, outside the lecture-
room, m the objective ideahst solution which for many years
he continued to dictate, m slightly modified forms, to his
audience.

The only son of the Rev James Wilson,1 a Methodist mmister

of the New Connexion, by his marriage with Hannah Cook,1
daughter of John Cook of Newcastle-under-Lyme, John Cook
Wilson was born at Nottingham, the 6th June 1849. In
January, 18591 he was sent to school with a Mr. and Mrs Morgan,
to Shireland Hall, near Birmingham. A sheaf of his letters from
this academy are preserved but, save for two studied and care-
fully indited compositions ' announcing the approaching vaca-
t10n ', they are not dated except by the days of the week. Like
most boys' home letters, they are largely made up of answers
to mqu1ries about health and recreation, and of requests, often
very firmly pressed. But not for money. The little boy already
1
James Wilson, born 2 x1.14, died r 111 1902
1
Hannah Cook, born s vn.18, died 16.v11.02. A letter to J.C. W, on his
mothn's death says: • She was one of the bnghtest and cleverest old
ladies I have ever known.'
b2
xx MEMOIR
realized his parents' straitened means and (to his mother) is
most regretful about torn clothes, broken garters, and lost shoe•
laces. He shows a marked enthusiasm for study, music and
drawing, for flowers, the capture of butterflies and dragonflies,
and for scientific toys like the kaleidoscope. This last he
promises, on his return, to his httle sister to comfort her in
some trouble. The simple veracity and strong affection of the
grown man are already there. The letters nearly all actually
fill three to four sides of an ordmary folded sheet ; one page
or more to his father, one to his mother, the third or part of
the fourth to the favourite younger sister whom he calls ' Micky
Meg '. 1 Here arc some extracts · ' I am in Delectus in Latin,
I read some sentences such as Cessator esse n.olo. The minister
nearly always preaches about Charity.' 'You tell me I must
not tell tales of the boys but Mr. Morgan said that when one
boy hurts another the boy must tell him nevertheless I will do
what you tell me.' ' Give my Grandmama my love, tell Emma
she is a skunk and give Maggie this letter.' 'Thank you for
the butterfly, I hke shooting with a bow very well so I bought
a little [bow] and 2 half penny arrows but the bow broke money
thrown away but I hope you will forgive me this once. I gave
6d. to Mrs. Morgan's birthday present to which I do not think
you will object.' ' I don't like to enter the cricket club because
we have to pay but I can find plenty of amusement without.'
' I am in octagons and hexagons now. I have not time my dear
papa to send you a specimen of what I am drawing I am sure.
I practice every day and the music master comes on Mondays
and Thursdays their are 2 drawing days Tuesdays and Fridays.
I draw on Fridays we arc lear(n)ing that funny tale of John
Gilpin in Cowper which I like very much.' ' What you say
about the cuckoo brings something to my mind they call me
Cook here to distinguish me from the other boys and William
Hedges changed it into cucko. I assure you Papa that I have
plenty of play.' 'As regards my study we have just been
examined I want a NEPOS bound in boards with copious notes.•
' I am just got into a melody of Beethoven in music but I cannot
play a regular time. I cannot keep time to every note yet.'
He was very homesick and looked forward eagerly to the end
1 Margaret Wilson, born 1856, died 1886.
 MEMOIR xxi
of his first half on the 16th June 1859, the school assembling
again on the Ist August. Just before the ISt June he writes :
• I am not very happy [his first sentences are usually by way
of reply J. I am the most unhappy boy in the school. I can't
help crying for you and Mamma. My chum who always has
been my friend kicked me out of bed the night before last to
amuse a new scholar and called me names, but I got him to be
as warm a friend as ever, but notwithstanding this I am very
unhappy.' ' You ask me to strike the key of pleasure but I have
none to tell you of.' Of chddish piety these are examples :
' I hope the Lord will bless me, dear Papa. There is a boy in ·
our school named Penrose who used to live at Wake.field who
opened his mmd to me and we resolved to pray and read and
try to be good so talk a great deal more religion to me.' ' The
Lord does hear the prayer of the young for I asked the Lord
to grant many conversions to your mmistry.'
Mr. and Mrs. Morgan were evidently good kind people ; how
long he remained w1th them is not certain. In September 1862
he entered the Grammar School of Derby, as a boarder in the
school house, Mr. James Wilson being at the t1me stationed at
Derby. After two or three years, a new master, the Rev. Walter
Clark, B.D., was appomted. He had been Captain of the school
al Shrewsbury, under Kennedy. A scholar of Magdalene, he
was placed fourth m the 2nd Class in the Classical Tripos, being
bracketed with H. Cecil Raikes, afterwards Member. of Parlia-
ment for Cambridge University. 'When he came to Derby he
found everythmg at a low ebb : there were but few boys, there
was little or no public spirit among them ail.cl for some time
the school had had httle or no success at the Universities ' 1
Wilson saw things so much transformed by Mr. Clark's energies,
that 1t was hard to remember what the school really had been.
He chose his masters well, encouraged modern subjects and
natural science, improved the games, and erected a boat-house,
and, in the result, there came from Derby School, Wilson him-
self; a 2nd Wrangler; Mr. E.W. Hobson, a Semor Wrangler;
many scholars at both Universities and one man who rowed in
the Oxford Eight and took a ' first ' in Classics. Mr. Clark was,
Jowett said, ' a good scholar and a man of uncommon energy
1 Letter of J. C. W., 10 111,84.
xxii MEMOIR
and . • . raised the Grammar School from small beginnings to
the position of one of the first Grammar Schools in the king-
dom '. 1 Wilson was Captain of the School from 1865 to 1867,
and of the cadet corps 1866-7. After he went to Balliol his old
head master continued to help him by receiving him into his
house m the vacat10ns, when ' he had much important readmg
to do '. 2
Mr. J. W. Sharpe was some three years his junior at Derby
and a constant friend, correspondmg with him on mathematical
and philosophical topics. 'He was', he says, 'an honest, true,
brave, high-spirited man and was the same when a boy at
school. The head master befriended him and he always spoke
of him with gratitude and affection, and would never pay any
attention to any strictures upon him, which many of us boys
were ready enough to make. In fact, I never knew Wilson to
listen willingly to blame nor to contemptuous speech of any-
body,-nor would he use such speech himself, except for sins
of philosophy or scholarship, and there he would break out and
become vmlcnt, even were the matter but a Greek particle.' •.. 3
'He was 17 years of age m the year 1866, when I, then nearly
14 years of age, became a boarder. . . . He was head of the
school at the time. He befriended me in every way. He was
a kind, high-spirited, cheerful-minded boy, always ready to
suppress any disorder which he happened to consider it desirable
to suppress. No one resented the forcible measures to which
he promptly applied himself, for he enJoyed a smgular reputa-
tion amongst the rest of us boys for smcerity of character and
directness of method. In fact, I have never met anybody who
excelled him in these particulars. . .. Games were not of much
account among us except as an exercise and a pass-time and
Wilson took little trouble about them, though he was a strong
and athletic boy. Most of us were idle, some worked hard and
Wilson worked very hard. He was a very able mathematician
by nature and was good enough for (one of) the first three
wranglers in any year at Cambridge.... His father, ... as is
1 Te-sb.momal letter to Mr Clark, s m 84

• Lett.er of J. C. W , 10.111.84.
• Letter of Mr J W. Sharpe, sometime Fellow of Gonville and Cams,
25.v 16, to H. W B. J.
 MEMOIR xxiu
the custom of the Methodists, was never more than three years
on one station ; and his son once or twice told me that they
were twice reduced to dire straits, i.e. on two stations, because
his father would not model his teaching on the opinions of some
powerful deacon or other, who thereupon took care that the
teachings of poverty should be inflicted upon the recalcitrant
and stiff-necked minister. Wilson's stores of energy, both
physical and mental, were to me always astonishmg, at all
times of my life. Nothing depressed him, and no prospect of
work appalled him, and hardly ever indeed proved too much
for him.' 1
Professor Hobson says : ' Wilson was, when at school, much
interested in the drilling of Volunteers. One day when the Drill
Sergeant was absent, Wilson was set to dnll us small boys. We
imagined we were gomg to have an easy time, but we found we
were quite mistaken, as Wilson turned out to be quite a martinet
who compelled us to do everything with extreme care. I have
a vision of seeing Wilson m the Entrance Hall of the School
busily turning over the leaves of an immense Dictionary (prob-
ably Liddell and Scott), when I was told he was preparing for
the Oxford Semor Local Exammat1ons. I remember the feclmg
of awe with which the sight msp1red me.' 2
It was T H. Green's reputation and m part Green's initiative
that brought lum to Oxford. Desiring to bridge the gulf which
separated the poorer classes and the less privileged schools from
the Universities, Green had earnestly supported a favourite
scheme of Jowelt's by which boys who had done well m examma•
t10ns hke the Oxford Local might, provided they had offered
Latm and Greek, be elected to an exhibition at Balhol. 3 Wilson
profited by this He came up with one of these exhibitions in
January 1868, assisted also, I believe, by his head master's
generosity. 4 The College had instituted a hostel in St. Giles',
1
' J W. S to H W. B. J , 21 v 16
Letter of Mr E W Hobson, Sc D , LLD , F RS , Fellow of Christ's,
Sadleirian Profeb'!or of Pure Mathematics, Cambridge, to H. W B J ,
15 V 16
• For Green's interest in the scheme see Worhs, vol 111, p. cv11 (Bradley's
memoir)
• • In several instances, to my knowledge, he generously helped those
who had difficulty 1n meetmg the expenses of a University education' :
Letter of J C W., 10,111.84.
xxiv MEMOIR
where poorer men might lodge and board at less cost than in
college. Green himself hved there both before and after his
marriage m 1871 to John Addington Symonds's sister. Here
Wilson read for both mathematical and classical honours.1 He
had few friends, lived to himself and was happy enough, studying
very hard, rismg rarly and s1ttmg up late. In November 1869
he was raised to the status of mathematical scholar and resided
continuously until the summer term of 1873, never moving mto
college 2 Thus he missed one of the happiest and best of
influences, the society of young men congregated within the
walls of a college. At Balhol Green's influence was strong upon
him; he followed his courses upon St. Paul's epistles as well as
his ordinary instruction. He attended Chandler, a very learned
Greek scholar, and an early course by Wilham Wallace on the
Politics of Aristotle. He heard Mr. Case too and his hfe-long
friendship with him depended largely on their common friend
Aristotle One thing he would say he learnt from Green, to
divide his own lectures into paragraphs. This has left its mark
on the present lectures. When he began to teach, he imitated
too Green's patient method of wcighmg every detail of an author
he was cnticizmg, spending an mordmate labour on minuter
issues, which the advance or change of ideas has itself anti·
quated Both Wilson and his master thus often seem unable
to discern any good in the author they are handling, m this the
very opposite of Wallace. This tendency has a bad effect on
the bcgmner, rousmg his natural sense of fair play and thereby
producmg the reverse of what is mtended, or else makmg him
run off with the notion that philosophy preserves the errors of
the past in order to refute them, while science quickly leaves
mistaken theories behind.
His mathematical teacher was Henry Smith, 8 a fascinating
and richly endowed character for whom Wtlson had an undying
admiration. Menttonmg Chandler, Green and Smith in his
' ISt Math. Mods 69: 1st Classical Mods 70; ISt Math. Finals 71 ;
1st Lit Hum 72.
1
The late Rev F H. Hall, Fellow and Dean of Oriel, told me that the
night of his election at Oriel was the first Wilson had spent w1thm the
walls of a College
1
H. J. S. Smith, F.R S., Ireland Scholar 48 ; Fellow of Balhol 49 •
SaVJhan Professor of Mathemabci, 6o-83; Fellow of C.C C. 73-83 •
 MEMOIR XXV

inaugural lecture 1 on the 15th October 18891 he says: 'What•

ever a man's own unworthiness may be, he is allowed to praise
his teachers. For me, however, the task is altogether too
difficult. I have had such great teachers.' A reviewer of
Smith's mathematical papers, posthumously published, says that
the lesson they convey 1s that ' only an mvestigator to whom
any slurring over of difficulty or exceptional case 1s absolutely
repulsive, can hope to make a real advance '. 2 Tlus 1s perhaps
the great lesson that he learned from these great men. Those
who were not privileged to work with Wilson, or who have not,
as I have, looked through his manuscripts, can have little
idea of the way m which he took this lesson to heart. ' An
investigation', he wrote to a favourite pupil, 'carried on per•
severmgly for a long time may end in the discovery of a fact of
consciousness which upsets the theory so laboriously worked
out. The utmost gain one has seems to be that one has found
out what will not do. Now this 1s a gam, but one is not at once
prepared for the new effort which 1t suggests.... The trouble
1s that one feels life is so short, ars longa but philosophy seems
very much longer.' 3 This 1s the scholar's last lesson, the clue
perhaps to what 1s sometimes called Oxford irony.
He went away more than once with Jowett, who became
Master m 18701 to Malvern, but not, I thmk, as an under-
graduate. He shared the common affection of all Balliol men
for their great Head, but not the extreme reverence of some hke
Wallace. He was not in the least restrained from critic1zmg
him, even his work on Plato. He perhaps underrated somewhat
a philosophy that was so much compact of common sense ; the
Master's irony too and subtle suspension of Judgement rather
escaped lum. It will not be amiss to add one authentic anecdote
to the multitude that have gathered with years round that
venerable head. Wilson took an essay to the Master and himself
described the event as follows : • I knew the subject was a stdI
one, for I had chosen 1t myself. I had taken uncommon care
over it ; m fact I suspected I had gone into it a little too deeply
' Printed less introductory paragraph at§§ 355-64
1
Ma1or MacMahon m Nature, 27 1x 94
• J C. W. to H. A P, IS xii 06, quoted more fully 1n Mind, N. S ,No III,
P, 300 Mr. Pnchard has kindly allowed me a free hand m usmg his article
and his pnvate correspondence w1th Wilson.
xxvi MEMOIR
for Jowett. So I read it very slowly and pretended to pause
to take breath, to give him time. When I had fimshed, he
looked up and said, "Been running, Mr. Wilson? "' 1 This
story he would tell in great glee, although probably he had not
succeeded so well as he supposed. The remark at least showed
Jowett's acquaintance with his pupil's habits of life. Wilson's
regular exercise, at this time, was to run down to Folly Bridge,
get into a sculling boat at Salter's, and scull as far as where the
Free Ferry now is, never farther ; then scull up agam, and so
at the double once more back to his books. 1 The little red-
bearded figure was famd1ar to the watermen along the Is1s and
no doubt to tutors on the towing path
On the 10th April 1874 he was elected Fellow of Oriel, out
of a strong field. 3 He was Justly proud of his success, in later
years, and used to say : ' they were better men in some ways
than myself but I defeated them by weight of metal ' He had
mdecd gained a double first m both Moderations and the Final
Schools and now offered as a candidate at Oriel, besides the
literae humaniores subjects, mathematics and the De Anima of
Aristotle. He however put down Greek Iambics for his principal
special subject, unnecessarily as this was a normal part of the
examination. He was then, as later, covetous of recognition in
the narrower fields of scholarship.• ' When he came to Oriel,
he was obviously overworked, rather emaciated and very languid.
He used to play the piano a good deal, and, I fancy, got mto
trouble with Provost Hawkins for playing at unsuttable hours. 5
The Provost also obJected to persons smokmg in Wilson's rooms,
which were over his study, as he said the smoke came down
the chimney. . . Cyclmg, hockey and volunteering did a good
deal for lus health and gave lum the robust look he afterwards
1
The words are, I thmk, precise Wilson told the anecdote to my wife.
The sayings and anecdotes are not selected m any partial spint and are,
unless otherwise stated, experiences of my wife or myself
• This and the amusing fact about the Iambics I owe to F H Hall
1
It included Mr. Andrew C Bradley and F. H Peters, one of the
bnlbant Harrovians of that epoch, afterwards Fellow of University.
• viz pure scholarship or criticism In 1873 he won the Chancellor's
Pnze for a Latin Es,ay, 'Quaenam fuent revera Epicureorum ph1losophia'
' By his own admission the chief offender was Mr Thomas Case, after-
wards Waynflete Professor of Moral and Metaphysical Phllosophy, Fellow
of Magdalen and President of C C C He was then a Tutor at Balbol
 MEMOIR xxvii
had. I don't think he took much interest in men's sports,
though he used to come to bump suppers. The only remark
of h1s I remember in this connexion was when he saw some
men practising " heading " a football in the quad and told them
he was glad they had found some use for their heads.' 1
These last notes are by a close friend, born on the same day
in the same year as himself, whom he found already a fellow
when he came to Oriel in 1874. The Common Room was then,
as 1t had long been, a distinguished body, though without the
colour for which 1t was best known in Newman's time. It
included D. B. Monro, W. Stubbs, J. W. Burgon, Poste, Bryce
and C. L. Shadwell. In after years Freeman, Froude, John
Wordsworth and T. K. Cheyne were, as professors of the
University, fellows, and in 1883 a well-known Aristotelian, Sir
Alexander Grant, became an honorary fellow of his old College.
Wilson did not reside long within the walls, soon marrying
and establishing himself m North Oxford. In 1873 and 1874
(or 1874 and 1875) he had v1S1ted Gottingen, 1 no doubt specially
to hear Lotze. 3 To Judge from his lecture notes he did not
attend the logic course. Though he always retained a deep
reverence for Lotze, he was not so much influenced by his
doctrine as would have been natural at that impressionable age.
He was however affected, doubtless, by the general reaction
from the cntlcal 1deal1sm of Kant, and from Hegel, which was
then in the air. His development was rather due to reaction
from the English writers, the reaction from Mill, as well as from
Green, Bosanquet and F. H. Bradley. He adopted Lotze's method
of discussion followed by dictation, but without the ease
which enabled Lotze to dispense with notes in either part. At
1
Letter of F. H H, 30 XI 21
• I have not been able to verify the years He himself stated in wntmg
that he was at Gottmgen 1n 1873 and 1874: be stated however that he was
elected at Oriel m 1873, a year before the actual date, and may have been
wrong here also (cf. Letter 37, p. cxx1v)
• Hermann Lotze succeeded, after an mterval, to Herbart at GOttmgen.
Wilson repeatedly employed Lotze's 1llustrabon of the relation of mdlV1dual
colours to colour, to elucidate 'the universal•. He was also much affected
by Lotze's views as to the moral, aesthetic and religious consciousness.
He held Ills view of pumsbment as retnbutive and the difficult op1mon
that the pleasure we take in it5 beauty or goodness 1s what gives the
beautiful obiect or moral act its value. (He detested but used this word
valut.)
xxviii MEMOIR
Gottingen he made himself nearly as famtliar with German as
with his native tongue ; he was however not a devoted student
of German literature, even of the philosophers. In Hanover he
met the lady who was to become his wife, and it was this
perhaps which led Miss Margaret Wilson to accompany him on
a visit to Germany m 1875. Returning to England, he began
to lecture and instruct at Oriel and Lincoln in the Michaelmas
term, 1875 ; for the two colleges then shared teachers in several
subjects. He took freshmen in ' ordinary ' logic (Mansell's
Aldrich and Mill, with Fowler) and in Dr. Magrath's and
Trendelenburg's selections from the Organon. More advanced
courses were in the History of Greek philosophy, with Ritter
and Preller, the Development of philosophy from Bacon to Kant,
both the Ethics and the Atialytics of Aristotle, and a charac-
teristic and thorough set on the Republic of Plato. At that
time the system which now makes a college lecturer virtually
a University teacher was beginning ; thus men of other colleges
heard him. He also taught from time to time for Balliol,
Pembroke and Hertford. He was for a while Junior Treasurer
of his College and a keen and sometimes unorthodox Librarian. 1
He was Junior Proctor in 1885, and Pubhc Exammcr in Classical
'Greats' m 18871 but 'found the responsibility of decidmg on
men's classes so harassmg that he held his office for only one
year of the three years' term. At all times and m all matters
he was incapable of domg thmgs by halves.' Otherwise he took
little part in University affairs, spending what leisure he had
from teaching in mathematical researches, m symbolic logic and
in Greek and Latin criticism. After fourteen years' diligence as
a lecturer-he seems, to Judge from the University Gazette, to
have been the busiest philosophy lecturer of the day-he had
the good fortune in l 889 to be elected, at the comparatively
early age of forty, to succeed Thomas Fowler in the Chair of
Logic, among his rivals bemg Venn and Mr. Case. From his
election to his death he occupied himself with the development
of the lectures here printed, in his subsidiary studies, in volun-
teering and in the work of a Governor of his old school. In
the last he took much interest, especially in the attempt to
secure his old master preferment and, on his premature death,
1 He was Subdean and L1branan 75--92, Junior Treasurer x 87-x 89
 MEMOIR xxix
in the choice of a fit successor. A strong and convinced Liberal,
in both political and intellectual questions, he happily did not
feel drawn to work of a semi-political semi-educational character,
within or outside Oxford, confining his exuberant energy to his
duties as student, teacher and citizen-soldier. He became fellow
of New College m 1901 1 by virtue of his Chair, and an honorary
fellow of Oriel in 1909. The • little University ' 1 of St. Andrews
honoured itself and him, in 1906, by conferring on him the
degree of Doctor of Laws, honoris causa, and he became a fellow
of the British Academy m 1907, a distmctlon he did not much
relish except as ' showmg the opinion of him entertained by
Caird and Bosanquet '.
This, with the Presidency of the Oxford Philological Society
m 1901, completes his small talc of worldly honours, but when
he died on the 11th August 1915 it was said truly that • for
many years he had been by far the most influential philosophical
teacher at Oxford ' and that ' since Green no one there had held
a place so important in these studies '.11 The explanation hes
not so much in what he taught or wrote as in what he was.

His years as a tutor were devoted, so far as production goes,

mainly to classical criticism. He read his first paper to the
Oxford Ph1lological Soc1ety ' On rearrangements of the F1fth
Book of the Ethics', in 1879, and became a contributor to the
Academy, then at its best, the 'Journal of Philology and the
Classical Review. The Oxford Philological Society, as a glance
at its Proceedings will show, was then m the heyday of its
strength. Scholars hke Monro, Ellis, Bywater, Rhys and James
Murray were in their pr1me and humaner themes were introduced
by Henry Nettleship, Pelham, John Wordsworth, Warde Fowler
and Dr. Macan. In the hst of members are other redoubtable
names, whose silence was as impressive to a novice as the
1
A v•01topu1,u,s of affection. ' Our Judgement 1s that the little Un1ver-
s1ty has done itself great honour m selecting you for a degree; I do not
think there 1s anyone whose clauns appear to me so strong In fact, of
course, your great place and prestige, as well as the solidity of your reputa-
tion, make it really a distmction which you bestow upon us, a distinction
of a type wlnch I strongly covet, that of showing that really good men
WIII accept our honours and that the httle place knows how to value what
1s good.'-B B. to J. C. W , 21 11,06 (' Our ' includes Dr. G. F. Stout )
• H W. B. J , P. of B. Academy, l.c.
XXX MEMOIR
standard set by the more frequent readers. In such society
and with these men as critics Wilson won his spurs, with papers
which here and there show growing philosophical power and
interest, always thoroughness and critical and analytic ability.
He had begun with a study of the De Anima. Torstrik's edition,
no doubt, had turned his attention to what may be called the
higher criticism of the philosopher. Acute minds were just
then applying to the vu]gate of the Stagmte the kind of analysis
which had given such remarkable results in Old Testament
studies. They were finding traces of later recension in the
various treatises contained m the textus receptus and pro-
nouncmg, somewhat prematurely, on their origin and genuine-
ness. Wilson was attracted into this field, partly by the example
of Monro and Grant. In 1879 he publtshed his Aristotelian
Studies J.1 He concluded that the Seventh Book of the Ethics,
chapters i-x, were not written m their present form by the
philosopher for the Nicomachean Ethics, nor by Eudemus, but
contained traces of not only two but sometimes of three versions
of the subject, contaminated by a peripatetic later than Eude-
mus. In 1912 he published a revised edition. He had then
come to the conclusion, resting on maturer reflection and know-
ledge and on his experience in seeing work of his own through
the press, that the variants probably represent different drafts
by Aristotle himself, left side by side in the rolls of his notes,
disordered by oversight or accident and so preserved in their
present shape by some reverent editor.
In I 882 he gained the Conington prize for an essay ' On the
manner in. which the Aristotelian wntmgs have assumed their
present form, &c.' Richard Shute was put second by the
judges Newman, Bywater and Monro. Both candidates kept
the essays by them, in the vam hope of some day bringing them
to completion. Shute's was published posthumously in 1888,2
but Wilson's never saw the light. He continued to work at
philosophic texts and at one time was to have joined Bywater
in editing the Nicomachean Ethics. At least he discussed the
• See Ell"nchus Oj>erum, infra, p lxv1.
• On the history of the process by which tho Anstotel1an wntings arnved
at their present form (with memoir), Oxford, 1888. ' Shute died prematurely
in 1886, a man of extraordmary acuteness and force of intellect '-R. W. M ,
I.e., p. 43. Cf. Bywater 1D Arehir, f. G. der PhJlosoJ,Jue, iu. 4, pp. 654-5.
 MEMOIR xxxi
more difficult places with him and for many years they spent
Attic nights together in these studies, Wilson also taking much
pains over Priscianus Lydus, whom Bywater edited for the
Berlin Academy. At the same time he was a busy tutor and
lecturer and was working, as will be seen, at symbolic logic and
at cognate semi-philosophical semi-mathematical topics.
The effect of all this was a certain diffusion of attention and
a devotion to relatively minor issues, which have left their mark
on hts work. He felt some anxiety to justify himself and writing
later to a philosophic friend says : ' Doubtless my great fault
is to be interested keenly in different and hardly commensurate
subjects, which I work at eagerly in turn ...• I feel no inclina•
t10n to write a general treatise on logic but only the parts to
which I think I can make real additions. I know the public
is more impressed by a "book" on a whole subject, which is
indeed a measure of the public's judgement.' 1 He was weighted
also by the intense effort of a surely mistaken endeavour to find
a mathematical refutation of what, as a philosopher, he believed
to be the fallacy of non-Euclidean geometry. 'There 1s nothing
so exhausting, either, as this feverish and energetic grappling
with proofs of this kmd. I'm m excellent health, but I assure
you that the top of my spine simply aches with the efforts
I have been making lately and when I have this sign I take it
(as) a sign that I must knock off mental work for a bit. I should
have added that my work 1s alc;o impeded by domestic worries.' 1
He had indeed all his hfe anxiety, pecuniary and other, in regard
to his aged parents, his sister and Mrs. Wilson. All this rein-
forced his native tendency to concentrate on single and minute
issues, though he naively attempted to excuse himself by
fostering a contempt for the wntmg of mere books and the
vanity of premature publication. Lackmg self-critlc1sm and
convinced that 'the (printed) letter killeth ', in his anxiety to
keep an open nund he reserved for an audience of youths work
that merited the Judgement of the world. He maintained how•
ever with some justice that the necessity he had lain under of
elucidating texts to novices had been of the greatest service to
his scholarship, and here Bywater was at one with him. ' It
forced me to translate the beggars,' Wilson used to say, and
1 J.C. W. to S C. P., 13.v.01.
xxxii MEMOIR

Bywater often suggested ironically that German editors made

faulty emendations ' because they would not use, or (better still)
make a crib '. So Wilson praises Poste for publishing his classical
editions always with a translation. 1 The desire for exactitude,
the trick of the teacher, the logician's care for method, all con·
spired to lead Wilson to part regretfully, 1£ at all, with his
scaffolding ; he would insist on showing you how he reached
the truth, however tortuous the route he had pursued. This
made his later papers mtolerably tedious, affecting even the
stylt> ; notably when, after an evening spent with the Greek
tact1C1ans in a promised discussion of the Greek word for induc•
tion, he ended close on midmght with the remark : ' I hope
now to have convmccd you all that the tactical uses of the
verb throw absolutely no bght on its meaning for philosophy.' 8
Herc, as in logic, he built his knowledge on a few first-rate
authors, commentaries and grammars, but secondly on compre-
hensive induction, almost of the Bacoman kmd. Rightly too
he laid stress on preserving a fcclmg for natural usage, on the
force of the actual context and on the tendency to inequahty
and iteration of writers absorbed in their subject-matter. In
his lectures he recurred agam and again to the unconscious
evidence of natural speech and would press convincingly the
folly of looking even in Aristotle for dead terminology or isolated
dogma. For him the language even of the Organon was no
petrified lava stream but a flowing river. I remember a dis-
cussion m which the Greek word Ev8vr {straight) was under
dispute. H P. Richards, rather a precis1an among plulologers,
held that the word must mean' directly to a destmat1on '. Inas-
much as the context implied a change of d1rectton, he desired
to make an emendation. Wilson stoutly opposed this. We met
at Wadham College, where you approach the Common Room
by a newel staircase. As we came away Wilson took my arm
and chuckled, in my ear: ' We surely went up straight from
Hall, we didn't go up screwed.' Here is an example of his way
of handling a problem of interpretation : ' Before considering
1 ' Nor did he ever offer a commentary on a clasS1cal text without

a translation '-Ob,, of J Postc by J C W m the Oxford Maganne, 4 v1 oz.

• 0 P S 1901. Those who were present will remember the prohx
introduction on the logic of evidence to the paper on 'The Similes of Homer•,
given to the same society m 1908.
 MEMOIR xxxiii
your present argument, I re-read the chapter carefully, with the
idea of commg fresh to it and letting it tell its own story, without
prejudice, for I think before I may have been rather affected by
what one thinks are the real facts of the assignment of " praise"
and therefore by what one expects Aristotle to mean. . . . ' 1
He then agrees in part with his correspondent's statement
of the difficulty, but thinks he has missed what is probably the
most important point. This point he states and tests by com•
parison with the context Next he gives all the passages from
Bomtz where the crucial words occur, only arranging them in
!us more logical order. He then selects the one which is, he
is sure, dec1s1ve and so returns to show that his own old
interpretation (modified not unnaturally by now) was mdeed
correct and that lus questioner had m fact unearthed a mare's
nest.
The year of his election to the Chair was devoted, in spite
of the necessity of preparing an inaugural lecture and energetic
volunteering, almost wholly to reviews followed by a pamphlet
of 149 pages, large 8vo, 'On a recent edition of the Timaeus'.
Tiu"> contains only a part of what he had adumbrated in the
Classical Remeu,, the treatment of scientific and philosophical
questions being absent Yet, as 1t stands, a competent critic 2
pronounces it ' an astomshmg example of wide and precise
knowledge of his subject and of close reasoning applied to the
thought both of Plato and his editors'. No future editor will
be able to ignore the learning Wilson has lavished on this work,
though its chief value hes perhaps m the reminder to young
scholars of the honesty and rigour required of those who would
be worthy of the scholarship of the past. The embers of the
controversy have long died down, so that the story may be
told, without preJud1ce, as an illustration of Wilson's nature
Archer-Hmd's ed1t1on had been greeted with a chorus of praise
m the reviews. This appeared extravagant to critics in Oxford.
1
Letter to F H H. 22 1x 13, on Anstotle, Eth Nie uo1b 18 seq
The dec1S1ve passage referred to is Cat 5b Is • What I feel about the
whole passage (of the Ethics) is its weakness and formalism He doesn't
get at what 1S really essential m the matter of praise and yet it is not hard
to do 1t • Wilson handled the famous doctnne of Purgation in the Poetics
with f'4Ual uncompro1I11smgness.
1
H. W. B. J, I.e.
~773 l C
xxxiv MEMOIR
Shrewd Cambruige scholars were of the same opinion. J. E. B.
Mayor, that most learned of Lat1n scholars, wrote to Wilson as
follows : ' I told - - that I thought Cambridge was iosmg its
reputation for solid ploddmg scholarship and was running after the
ignis fatuus of brilliancy. Men write in the newspapers and try
to raise a momentary excitement about one another's publica-
tions, but are far more eager to produce something new than
something true. Holden and a few others keep to the old paths,
but as a rule I am sorry to hear of an undergraduate reading
the classics with the new ed1t1ons.' 1 The reference 1s obvious.
The editor of the Classical Review, J. B. Mayor, held much the
same opm1on. Wilson made two stmgmg attacks upon Archer-
Hmd in language provokmgly magisterial m tone and suggesting,
hardly covertly, literary dishonesty and pretence This was
intended A rather contemptuous and unguarded reply in the
Academy Wilson answered m the same number. The pamphlet
appeared m December and was ignored, but Archer-Hmd, poor
man, never agam trespassed m this preserve
There can be little doubt that Wilson's motives were mixed.
His prmcipal and Just desire was to vmd1cate Enghsh scholar•
ship and to defend two d1stmgu1shed foreign scholars agamst
what appeared to him ignorant and slipshod censure. ' It will
not do to allow foreign critics to thmk our standard of an
edition of a classical author so far below theirs, or our notion
of the mterpretatlon of ancient philosophy so anachronistic.'
His choler was further moved by lampoons of which he had
been made the butt both m the Cambridge Review 2 and m the
Anstophamc verses 8 which accompamed the December, 1889,
class-hsts at the sister Umvers1ty. In both cases his second
Chr1st1an name, his mother's maiden name, was played upon.
The Aristophamc verses were sprightly and fair enough con-
sidering the occasion, but the ' tribute ' in the Review accused
him of lying. It was headed by a verse in Greek :
WI; fr71Tvp.wt MArEIPON (Tf )"€ ,ca.\oiiuiv &viJpw1ro,.•
1 J EB M 1.oJ C W,9x1189
• 14 211 89 under 'Poetry•, p 278
• e g ib,,)p 4AAollads Tes 1taTaTo£fv11,u lr/>f/
maECI' Tapa/•111, A.tnliopi,,, ~lf"M,
• ' How Justly do men term you COOK •
 MEIIOIR XXXV

I found among his papers a reply in similar doggerel :

1
MayE&poJI El'lfE IT ol, Kadir, a.irrov i\.a.8~1',
.A,,,,pr&#A'laar 3' -Ei\.a</)ov O'U &,xoppowCJJi 1
• Whene'er you find conceited trash,
Smash, Wilson, SMASH.
When sc10list agam writes book,
Cook, Wilson, COOK
His Goose,
Again in its own Juice '
To the end he was hke an old troop horse, neighing and
prancing at the voice of the trumpet, and once in the fray he
cared only for victory, was dissatisfied unless 1t became a rout.
Moreover, he caught at the chance for wmnmg a reputation for
exact and sohd scholarship. That this last was one of his motives
I judge to be certain, from an experience of my own. Wilson
came m one mornmg and left for my opm1on some printed work
111 a by-path of Greek scholarslup. I sent him that evening
a short note on it and was surprised next day by a second visit.
' I want you to work this up,' he said, ' and publish it ' ; addmg
gleefully, ' it is not a bad thmg for your reputation to attack
pretentious scholarship, wherever you find it, and to demolish
1t.' He followed tlus up by a letter m similar terms Not till
much later did I realize that the man whom I was to attempt
to smash was a tritagomst m the old quarrel. Jowett with
worldly wisdom wrote : ' I hope your controversy 1s progressing.
Everybody speaks of the desirableness of moderation , which
1s also a most provoking thing to your antagomst. I desire to
stir up your Christian Charity by the last remark. I thmk
some genuine and deserved compliments might be paid to the
scholarship of Cambridge.' 2 How little he knew his man I
Wilson steered lus own course resolutely and published the
pamphlet in all its uncompromising fullness. He pricked, as he
used to say, the bladder ; quartered the Hmd iu quite decisive
fashion Jowctt's last words suggest another reflection. In
those days there was m Oxford a strange preJudice against
Cambridge, now happily unknown. From this even the Olym-
pian Bywater was not immune, though m him 1t was largely
' • He d1dn•t:see that Cook was Just the name for you. You made mince-
meat of the Hmd, dec1S1vely '
1
To J. C. W., 25 1v 89
C2
xxxvi MEMOIR
playful. No one admired men like W. H. Thompson and Hugh
Munro more than Wilson and he did. The general opinion,
however, was that the ' Archer ' was not worth the artillery
concentrated upon him and that Wilson's proper remedy was
to edit the book himself.
Wilson had been concerned by what he thought the philo-
sophical shallowness of the ed1t10n and outraged by attacks
upon Aristotle made ' with the acrimony usually reserved for
contemporaries '. His words in regard to interpreting Plato in
the tight of modern philosophy still deserve attention 'We
have no preJud1ce agamst the attempt to understand ancient
thought by the help of modern; we venture to believe that
Plato and Aristotle are ltkely to be best understood by those
who have an interest in modern metaphysics. But there 1s
a stage m such interpretation wluch has brought discredit on
1t a stage which the individual will still often have to go
through, but out of date (we had hoped) m the progress of the
race When a man with a new enthusiasm for some modern
system of metaphysics begms to see, behmd d1fferences of
formula, affimties between 1t and the doctrmes of Plato and
Aristotle, 1t sometimes happens that what to him is a discovery,
disorders the Judgement, so that the reaction agamst the dull
annahsttc treatment results m an opposite extreme, almost as
much to be deprecated Ancient thought 1s crudely treated as
if it were modern and the natural sense of a text is either not
seen at all or passed with contempt. To the student himself
somethmg hke finality seems achieved ; but instead of being
the end, it 1s not much beyond the beginning of critical inter-
pretation · it 1s merely a stage before the development of
an historic sense. It may, nevertheless, deserve respect ; for all
must begin and all must wish charity for their own short-
comings.' 1 This and the subsequent passages upon the inter-
pretation of Plato as though he were a modern idealist show
the maturity of his mmd at this date. His teachmg was mdeed
most excellent where he generalized in regard to error, showing
that a given mistake belonged to a type, which he then charac-
terized decisively. He constantly reminded his pupils of what
he cruelly turns upon his antagonist in this passage of arms,
1
The Classical Re11iew, vol. m, No. 3, p. 119 seq.
 xxxvii
a saying in the Parmenides : ' you are still young : the time will
come when philosophy will have a firmer grasp of you, if I am
not mistaken,' and he made them seek the general fallacy under•
lying a particular dogma or an apparently novel heresy. Thus
m regard to the Megarian problem called the • Liar ', he would
mamtam that the puzzle about the truthfulness of Epimenides
relies for its effect upon the erroneous notion that a Judgement
can make a statement about its own truth or falsity. This he
then grimly developed in order, as he hoped, to destroy the
coherence theory 1 of the nature of Truth.
Wilson only reluctantly learned the lesson that the parties to
a controversy rarely succeed m convincing any one but them-
selves. He did not fear controversy; on the other hand, it
disturbed him ; he did not enjoy it dispassionately as Matthew
Arnold would. Neither did he ever lose his love of victory.
He says frankly in a letter, ' I can talk or discuss better with
one , if there arc more, tbe desire to wm a victory disturbs
me.' 1 This made him appear cross and irascible to some who
did not know hrm, Just as hrs self-confidence provoked, if it
fettled to amuse, more modest mmds He was on the point later
of becoming drawn mto an attack on a much younger man,
James Adam. He resented an improper reference, as 1t seemed,
to Monro m that lamented scholar's essay on Plato's Nuptial
Number. 3 He wished to put hrm publicly m his place. For•
tunately the matter was ended and Adam later consulted lum
on a mathematical question in the Republic " The old warrior
formed a friend~hip with his junior, and Adam, with charac·
tcnst1c modesty, enriched a valuable edition with the frmt of
Wilson's special learning This change of attitude was m keeping
with all we know of Wilson. Once you had his friendship, he
would see all the good in your work and help you lavishly to the
better. But you had to defer to him and his preJud1ces. Only
provoke lum by opposition and he would, perhaps for a season,
perhaps for ever, see you and your work from an angle and,
what is more, discover moral turpitude m mere mtellcctual
1
As in Mr H H Joachim's The Nature of Truth, 1906, Cf Logu,
Lectures, I 240; Phliosoph1cal Letters, Part V, xm
1
Cf J C W. to B B, 7 and 18 vu 03, infra pp 729 and 740
• The Nuptial Numbe,, of Plato, J Adam (London), 1891, p. 9, note 1.
• Vule Testimo,ua, infra, p. lxxiu.
XXXVlii MEMOIR
obscurity. This side of his nature may be traced not unfanciT
fully to his father, also a fierce controversialist, and to an
inherited nonconformity. Alexander Kilham, the founder of
the schism from the Wesleyans to which Mr. James Wilson
belonged, is said to have exhibited m his wntmgs an undue
proportion of invective and to have injured his cause 'by an
occasional virulence of aspersion that was not m harmony with
his general character'. The same is true of Wilson. , !Ie was
the most lovable of friends, but some stram in him, a certain
ignorance of the world, the undue solitude of his early manhood,
an inherited pugnacity, made him a dangerous and sometimes,
I thmk, an unscrupulous antagonist Scholars, whose studies
involve the patient, often unrewarded, search for truth, will
forgive Wilson his zeal m causes he deemed sacred. He is an
extreme example of the schoolman's tendency to see things
too close.
He was the same m all he undertook. A mathematician who
intended philosophy, he was diverted mto the fascmatmg paths
of scholarship. Mathematics, he said himself, are the best
preparation for 1og1c , he found m them a d1sciplme and, in
later years, a recreat10n Yet he appears to have been, as
mathematicians Judge, no more than a competent mathemati-
cian. ' He would have made h1s mark, 1f he had made mathe-
matic-s his speciality,' 1 says the friend, who yet tried to dissuade
him from the costly labour of his one mathematical treatise, his
waiyv,uv y~pw1, On the Traversing of Geometrical Figures. ' With
his many-sided interests he hardly gave sufficient time and thought
to the subJect to make himself really conversant with the modern
aspects of the underlying problems.' 2 This Judgement of
Dr. Hobson is the opinion also of those best able to speak in
Oxford. It would be strange 1f 1t had been otherwise. But he
could not shut his ears to the s1rens' song and thus my table is
strewn with the wreckage of his many ventures m those delusive
waters. Quite a late packet of manuscript ' On contmu1ty and
change' 1s endorsed' This discussion written 1905-6 (apparently)
must be entirely revised I am not sure that what is said of
d1rect1on of a curve of a pomt is sound. J. C W. 191 I.' The
• J W. S to H W B J , 25 x1 16,
1 E.W. H. to H W. B. J., 2,Vl,16.
 MEMOIR xxxix:
latter subject might well occupy a philosopher ; but much of
his work, often endorsed in some such way, belongs to those
by-paths of the human intellect of which Lewis Carroll 1 has
said : ' This field of mathematical research, with all its wealth
of hidden treasure, is all too apt to yield nothing to our research ;
for it is haunted by certain ignes /atui-delusive phantoms that
flo~t before us and seem so fair and are all but m our grasp....
Alas for him who has been turned aside by one of these spectres,
who has found a music in its mocking laughter and who wastes
his life and energy m the desperate chase.' The words were
self-critical, but might well have referred to Wilson. The snare
which besets the man of science is to put aside problems per-
tment to his studies, feeling the time not yet ripe to attack
them ; so he forgets them and grows narrow-minded ; the
philosopher's temptation is to overrate his powers; to take all
time and all existence for his own, and the sands of life are run
out before he has well begun.
Writing what he calls an apologia pro vita mea, 2 Wilson says :
' You know that mathematicians in our day believe m the
poss1b1hty of constructing a new kind of space or spaces. That
1s to my mind the mere illusion of specialists, who can't under-
stand philosophy or metaphysics and are cocksure they do.
I long to destroy the abortion I thmk it possible to do this
m a way wh1ch would convince metaphys1cians. But to k1l1 the
thing one must convince mathematicians m the only way m
which they arc pcrv10us, 1.e. show 1t ends m a mathematical
contradiction.... Remember I don't mean that the refutation
of it depends on such a discovery-it can be refuted without.
But 1t would be convmcmg to mathematicians, 1f one could find
it .... If one could prove that 1t (the original absurdity) led
to the contradictory of 2 sides of a triangle are greater than
a 3rd, one would have done the trick. Now I have contmually
thought that I had discovered the required contrad1ct10n and
1
The late Rev C L Dodgson, Senior Student of Christ Church The
quotation 1s from A new theory of Parallels', by C L. Dodgson, 1890, p xv1.
1
J. C W. to S C P., 13 v.01 Cf (same letter) • I may say without
vamty that I do a quantity of hard thinking about puzzles and you will
easily understand that thlS cannot show bulky results. But 1t 1s a far
greater satisfaction to me to solve puzzles and do sometlung which m this
way will last '
xl MEMOIR
then found myself wrong. It 's noted as a treacherous subject.
But the fascination is tremendous and when it 's on me I work
almost mght and day. I have really, when the thmg seemed
commg out, thought through a whole mght till the break of day
and later I I can only compare myself to Bernard Palissy the
potter, who finally burnt the very floor of his house for the dis-
covery of his famous enamel If I succeed 1t will be admitted(ly)
worth the time spent. And of course I may not. How.ever, the
process has caused me to produce a number of theorems m this
pseudo-geometry-some already known, but mathematicians
tell me my methods arc ongmal and I believe they have the
advantage of s1mphc1ty. So that 1f I must finally give up
the " quest ", I may m undertaking the philosophical refutation
have the advantage of proving myself not an outsider m the
mathematical department, for of course the mathematicians hke
to say the philosophers who disagree with them don't under-
stand the mathematics They won't be able to say that of me.
I have however had a certain important success It seemed to
me one could develop some startlmg paradoxes, if one apphed
i.uch geometry to force and motion I constructed a system of
statics on the basis of this pseudo-space and proved among
other thmgs [that] in such space that two fleas could pull an
elephant with ease provided they were pulhng at two ends of
a bar and he at the middle, if the bar was made long enough,
thus I!-
Eleph.
t
+
Flea (N B no fulcrum.) +
Flea
One of the most d1stmguished of our mathematicians (an F.R.S.)
present at the mectmg of the Math: 1 Society here, in which
I showed tlus, said at the end that I had shown that either this
kmd of space must be abandoned or the current conception of
force must be changed, and he gave it as his opinion that it
was probably not the conception of force which \\Ould be
changed. . . Meanwhile I have had another attack of the
subJect, it seizes me off and on hke a fever and for several
weeks-because I found a new track-it has however ended m
nothing and I am making myself put it aside in order to fimsh
 MEMOIR xii
my i,,-ay0>)'11. On the whole this " quest " 1 it is that has hindered
me most and you will feel that, I think (sic), if it succeeded,
it would be a far more valuable thmg than the otl\er matters
on which I am working : and if it was once done, I could simply,
as I imagine, reel off the other things '
Thus we have httle but the memory of Wilson's obiter dicta
to oppose to the development of symbolic logic, which has gone
so far beyond the pomt at which he was assa1lmg it ; a develop-
ment perhaps the most mterestmg of all modern science to the
philosopher; for 1t 1s, or pretends to be, a branch of logic and
runs quite counter to the philosophy of mathematics which
Wilson taught and even to lus theory of knowledge. For, in
its essence, it appears to be the mmd d1ctatmg terms to reality
and 1t suggests what is indeed a metaphysical world. Wilson
treated Mr. Russell and the hkc almost contemptuously ; he
employed common sense and his dialectical skill to overthrow
their theories and the victory was, as such victories tend to be,
a vam one The two armies moved perpetually round ec1.ch
other's flank Wilson knew where the Vl.eakness but not where
the strength of the enemy lay; he ,vorked at these sub3ects
fitfully, however vehemently, and m the end there 1s httle to
give to the world I have printed two public lectures, 2 dealing
with Boole and Venn and implicitly with some sides of the
modern doctrine They are sufficient to show Wilson's lme of
attack and the brilliant ,..,ay in which he handled the topic.
There comes a stage m every thinker's history when he is no
longer open to conviction by the force of his adversary's objel.•
tions This came very soon to Wilson even m ph1losophy, and
111s doctrine of reality, for example, will be seen to be a statement
of lus own conv1ct10ns rather than a strictly pos1t1ve proof of
what he held.
In 1892-4 he conducted J. long debate by correspondence with
the author of Alice through the Lookmg-glass on tlus &ub3ect, and
mdeed the letters nught almost have been extracted from that
book. Carroll's view and method will be found m his Symbolic
Logic, Part 1, 3 a httle book that can be mastered in a week.
1
Alluding to Plato, Ap 22 A and 23 B • Logic Lectures, §I 371-400
1
Symbolic Logic, Part P, by Lewis Carroll, 1896 The extracts are from
part of a correspondence, dated from I I xi 92 to Christmas Eve 92. Wilson's
xlii MEMOIR
His limitations are easy to see but to resolve his difficulties is
quite a serious task. Wilson was led away into an attempt to
show that fie could do this kind of thmg JUst as well or better
than Carroll, and he found his antagonist too much for hun.
The correspondence flows on with an easy pers1stency and pro•
voking elus1venec;s on the one side ; on the other with a growing,
rather dogmatic 1rr1tab1bty. Carroll, while asking for symbolism,
insists on concrete instances, and many of the letters turq round
a problem concerning the 'walking out of six married couples',
certain conditions hm1ting the combmations. The thing is
purely mathematical, as Wilson saw, but underneath Carroll is
trymg to resolve a difficulty he felt in hypothetical thinking.
He asks, for instance, for the logic of the true statement : ' If
I were to run to London m ten minutes, you would be very
much surprised.' The nature of the letters may be gathered
from the following extracts ·
CARROLL. ' I should be much interested to hear which, 1f
any, of these problems you think soluble by ordinary methods
or by any existing methods '
WILSON. 'Some problems qmte wrong and remainder soluble
by ordmary methods I knew a priori, before looking at your
quest10ns, that the ordmary methods must suffice, 1f the argu-
ments were sound because ... the figures of the syllogism are
complete for the purpose '
CARROLL. 'PECCAVI.' (A week later) 'At last I have spotted
the fallacy . . You made an assumption you have no right to
make: viz., that two contradictories are true at once.' The note
on this letter is 'This 1s a curious mistake of Dodgson's.'
CARROLL. ' I am charmed with your letter Just received and
regard 1t as a real "feather in my cap" that I have caught
the Professor tnppmg So you would really have the courage
to assert that the two Rules. (1) When I go out, I wear my
hat · (1i) When I stay in, I do not wear my hat · arc " contra•
d1ctory ". Yet may I venture to assure you that I own an
unbroken allegiance to both and never disobey either of them.
The note on this is : ' an extraordmary illusion of Dodgson, he
has made an elementary mistake.'
letters were preserved by Dodgson and may be seen 1n the Chnst Church
Library.
 MEKOilt xliii
CARROLL, ' I think you are a little hard on me, in our dis-
cussion. For, whereas I read all you send me, you decline to
examine arguments of mine, on the ground that, being already
convmced that your theories are sound, it is superfluous to
examine any argument which is alleged to disprove them. This
puts me at a disadvantage and rather remmds me (if I may
venture to draw such a parallel) of the JUry who, havmg heard
the evidence against the prisoner, told the Judge they were
convinced of his guilt and that it would be superfluous to hear
any evidence in his favour. Would you mind, Just for once,
reading a few remarks on the two rules which you and your
fnend 1 assert to be contradictory ? '
One outcome of the long duel fought m this way between the
two principals, and mvolvmg more than one second on Wilson's
side, from November 1892, till the summer of 1894, was Dodg-
son's dehc1ous article m Mind called ' A logical paradox '. 2
I might cite other mstances to show how wide and scattered
Ins mtercst and his labours tended to be Thus he lacked the
breadth and comprehension, the sense of proportion, of greater
<;cholars and thinkers. There is a certam narrow mtensity m
lus work and, if 1t is not too fanciful, he may often be compared
to a boy workmg at his sums, absorbed m relatively small
problems. His method of work m philosophy was the same.
He was not an mdustrious reader, he hardly knew modern or
contemporary writings and he used Prantl's and Ueberweg's
histories freely. If he had occasion to treat some thinker he
was of course too good a scholar not to go direct, at least to
the opening of that author's works ' I treated you ', he writes,
' ?.S I do any philosopher whom I happen to be reading. After
reading about a couple of pages of your paper, I saw what your
pomt and what your difficulty was. I then put it aside until
I should have thought out my own account of the matter clearly,
after which I should again take up the paper and read the
1
The late Rev W. Warner, Senior Student of Chn,;t Church, one of
Wilson's earliest private pupil,; Warner read with him for Lit. Hum at
Malvern,
• Mind, N S , vol in, No II The subiect waci resumed by Miss Jones
in Mind, N, S, No. S3 The article by Wilson (signed W) ism No. S4
Wilson did not keep his MS. but the evidence of style is conclusive and
he referred to an article of his m Musa at this date in a letter Seep cxix
xliv MEMOIR
sequel. I think this the best plan anyhow, when one is energetic
enough, though it may sometimes take one a long way beyond
the immediate question one finds actually proposed (though this
hardly happened in the present case) : and 1t suited me best
this time because the pomts you raise have been before me for
some time and I have been fcclmg the necessity of recons1dermg
them and rewriting this part of my lectures.' 1 Thus he read
and conversed rather with a view to his own difficulties than
to yours and rarely gamed a sympathetic mstght mto the author
he was studymg. Moreover, his own mmd was not creative and
his severe method of mquiry took the form of using as a fulcrum
for himself the pos1t10n reached by others. He was emmently
a critic, constructive 1£ at all only by dmt of instmctive previous
negation In fam1har talk you found that your mind was his
whetstone and written work he put to the same use. His treat•
ment of Mr. F. H Bradley 1s most easily explamed m this way.
It began m its present form m talks m the ' Long' of 1901,
~pent with me m North Yorkshire, and Wilson after his manner
earned his criticism too far. Y ct the very mtensity of the onset
shows his respect for the enemy , he treats him after all as he
does Aristotle and Kant, where he supposes them to be wrong,
and he has written in red mk on the manuscript the famous
apology of Anc;totle to his master . ' Where both were f nends,
it was a duty to put Truth first' 2 Bywater, associatmg his
name with his text of the Nicomachean Ethics, has called him
pre-emmently an Anstotehan 3 (a m.1.n most full of Aristotle),
and Monro 1s said to have ' ranked him among the first Greek
scholars of his time ' , ' Suc;cm1hl, from a distance and with
equal justice, taxes lum with undue subtlety. 6 Certamly his
mdustry was at times that of the spider rather than the bee.
Thus he remmds one m all his work of a seventeenth-century
1 Letter to H A P , 18 xu oo
' ,lµqwiv -,rip 6v-rou, <f>//\mv 01110v rrp,mp.iiv Ti]v dA~IJflav Ji.th Nie 109(:,& 16
• 'V1rum d1co s1 qu1-s alms 'Ap1aToT1A,trwTaTov, I. C. Wilson', prefcLCe
to Ar1stolelis Ethiea Nicomachea, Oxford, 1890 An allusion no doubt to
Winstanley's us<' of the same epithet of J Harris m Aristotelis de Poetica
Laber, Oxford, 1780
' S. B. Oriel Record, l c
5 • N1m1a subt1htas • He calls Wilson • v1r acubss1mus' in Anstotelas
Polie,ca• (Teubner), 1882, p xx1, and • m cntlc1s operation1bus v1r vabchs-
mmus ', 1b., p. xxvt, with a spice of irony.
 MEMOIR xlv
scholar, in form and in a certain mercilessness ; and their spleen
and prolixity are only tolerable in the greatest. His sad unfinished
work is in melancholy contrast to his friend Bywater's Erasmian
urbanity, concentration and perfect finish. These characters of
imperfection and incompleteness are impressed on his logical
remains also. There was indeed the excuse of age and fa1lmg
strength, distraction and lack of time, but the evil was more
deep-seated than any remedy could meet and arose in part from
an mtellectual vanity in striking contrast with the simplest and
sweetest of moral natures. I thmk he knew in his heart that
desire for glory alloyed his love of truth ; he faced indeed his
love of victory but he wa'l for a philosopher too much interested
m fame. The passwn grew upon lum to be original m his work.
He would at least satisfy himself that what he discovered he
had found for himself This made him so hard on what he
imagined to be dishonesty m lesser scholars ; not, I thmk, the
pure love of truth. Ile would choose rather ' a flower of his
own getting than much better gathered to his hand '. So he
reqmred to be directed upon problems which were incomplete,
and refused to be governed m his mvest1gattons by results
already secure. Workers m the same vmeyard, except personal
friend-;, he regarded with mstmct1ve, almost childish, rivalry.
He wa'l at no pams to acqu.unt lumself sufficiently with the
work of lus contemporanes, forgot that a ltvmg study progresses
only by the tentative and common development of questions
that at the moment mv1te and engross men's attention. He
never told Mr. Bradley, working m the same city, the errors he
detected tn his work. Thus he neglected that ' contribution of
arts the one to help another, that vanety of particulars for the
correctmg of customary conceits '. These faults, for they are
faults, helped perhaps to make lum the great teacher he un-
doubtedly was. ' Rarely has a professor been such a teaching
power in the U mversity or exercised such an ascendancy in its
philosophical studies. He was not only a teacher of other
teachers' pupils, but a teacher of teachers themselves ' 1 In the
last years of his hfe I have been told that on more than one
occasion, at the close of his last lecture m the summer term,
the whole class stood and gave the venerable little figure an
• S, B., I.e.
xlvi MEMOIR
ovation. This, so far as I am aware, is unique in Oxford, where,
except at public lectures, applause 1s forbidden by convention
and your audience effectively conceals any approval or dis-
approval it may entertain.
His chief recreations in middle age were hockey, volunteering,
and the war game. He was a keen hockey player, being with
the late Master of Balhol one of the dons who shocked Pater's
delicate nerves, when he lighted upon their combat.s _in the
Meadows He had resigned when I Joined the War-game Club,
so that I cannot speak at first hand of his skill and ardour.
Mr. Spenser Wilkinson was the founder, Sir Charles Oman,
H. B. George, and W. Sanday constant devotees. The last
named was reputed a Marat in his handlmg of cavalry, in odd
contrast with his gentle voice and manner and his benevolentia
tkeologzca. He was evidently a born fighter, however. Wilson
once wrote · ' Some years ago I published a series of criticisms
of his fussy emendations of Aristotle I showed up the lot,
I remember to the great JOY of Sanday who II can never have
enough of fighting" and sent me a congratulatory postcard on
having demolished the supposed Cambridge authority on his own
chosen field-" all along the hne ", I thmk he said.' 1 Of Wilson
the story went that he lost a campaign by delay in one of the
enemy's towns When asked by the umpire what he was doing
all those precious hours, he rephed . ' Sacking it, of course, and
putting the women and children to the sword.' This recreation
did not lead him mto any wide reading of m1btary history ; he
worked, as his manner elsewhere was, exhaustively at each
problem and to such good effect that he was the first amateur
in Oxford, to my knowledge, who reahzed the ability of Lord
French's earlier operations m South Africa. His study of the
war game and his volunteer service were sides of his strong and
native love of country. His feeling for England may be traced
here and there in his papers. A logic lecture 1s dated by
a marginal note, in Latin, ' I write this two days after the news
of our lamentable disaster m Africa '. 2
1 J. C W to A S L F, 10 v 12 The articles 1n the Academy, • Recent
emendations, &c ', are directed agamst suggestions made by the late
Professor Henry Jackson, 0 M.
• 'Haec scnbebam tert10 die post nuntiatam misernmam 1n Afnca nostram
cladem ', viz the battle of Isandhlwana, 22.1.79,
 MEMOIR xlvii
His simplicity and his share m the volunteer movement in
Oxford are brilliantly satirized in the ballad published m the
year ot his Timaeus campaign.
• Shtand oop, yoong man,' der Kochmann gried, und blaced him
on his feet,
'By vay of ransom you moost schvear-my ladest vork to readt,
If ve ish daken brisoners, id dakes moosh geld to free us,
I gi/s to you mein l1ddle book, dot treats of de Timaeus.'
The undergraduate world was aware of the controversy and
re;oiced as young men will One day m camp Wdson com-
manded the combined Cambridge and Oxford cyclists and was
ordered to attack the infantry battalions. The verdict of the
umpire was adverse and, on returning to his own Imes, Wilson
explained how the reverse was due to a blunder of his Cambridge
subaltern. ' Docs any one know the name of the / ellow,' he
asked, with somethmg of Mr. Pickwick's stress on the last word
There was a pause and then a scholar of Corpus took two paces
forward, saluted and said ' I think the name was Archer-Hmd,
Sir,' and so fell back smartly to the ranks. Wilson worked hard
at his quahfymg courses and at Musketry, mastered the War
Office manuals and a small but excellent book on Mmor Tact,cs, 1
and practised conscientiously Military Map-drawmg. Once a
v.eek for the whole afternoon he would take his cyclists mstruc-
tlonal rides m Road and Village Reconnaissance, divcrs1fymg
the route both ways by ' developing ' mmor tactical problems,
smted to cychst advanced and rear guards, m our behoof.
Probably we knew the track past Ch1lswcll Farm and all the
scenery of Thyrsis as well as the poet himself, and there are few
vdlages round Oxford that we omitted He prmted one of the
.first, if not the first, manuals of Cychst dnll and another small
pamphlet on Reconnaissance for his men, and parts of the drill
were adopted by the authorities. A brother officer 2 thus appre-
ciates him : ' The points m h1s character which first impressed
1
Wilkinson Shaw's The Elements of Modern Tactics (Ed x1 1900)
• MaJor F A Dixey, D M, FR S, TD., Fellow, Bursar and Sub-Warden
of Wadham, Curator of the Hope Collect1ons. I think our old Volunteer
battalion mess would give the palm to Wilson's work for the OU V. and
its successor the Contingent of the O T C Great nevertheless 1s the debt
to Dr Dixey and his contemporanes and to the many regular officers,
active and retired, who commanded, were Adjutants or Company officers
of, the battahon.
xlviii MEMOIR
me were his single-mindedness, his strict conception of duty and
his clearness of vision, carried out with practical prom;>titude
of action. He was tenacious of purpose, thorough in everything
he undertook, a master of techmcal detail. As I came to know
him better, I learned to appreciate more and more the unselfish
zeal with which he strove for the highest efficiency m himself
and in those under lus command, and his worth as a loyal and
st1mulatmg comrade His v1e\\-s were clear-cut and his,d~cisions
prompt. This quality came out not only m military matters
but also m general conversation, which often took the form of
friendly controversy On these occasions he was not content
with forming, or even with expressing, lus own opimon, but he
would manifest an eager desire to drive 1t home m the mmds
of others. . . It must be confessed that m presl>mg his point
he was apt to lose the sense of proportion and measure ... and
the earnestness with which he would urge the importance of
right views on these and like subjects had for some of his hearers
an effect which was really amusmg . . . But the respect and
affection so widely felt for Wilson were such as can only be
mspired by a nature at once strong, thoroughgomg, simple and
sincere ' There is \Vilson the man and soldier to the hfe On
parade he v. as very strict 111 essentials, as the little boy at Derby
found lum ; he knew lus drill perfectly but had a poor word
of command. In the field, he tended to magmfy a skirmish
mto an action, loved guerrilla methods rather than the ordered
disciplme of regular combat. On a small scale mdeed he was
singularly like the remarkable Confederate leader m whose genius
the professor and the professional soldier so oddly blended. He
would have been as fearless, morally and physically, as Stone•
wall Jackson m actual battle. He had the aggressive and
supremely self-confident nature, great endurance and activity
of mmd and body, and demanded the last ounce from his men
With a noticeable slimness m manreuvre he would cheat when
he could. When remonstrated with for recklessly exposmg him-
self, he broke out: 'Don't be a fool; can't you see, I want the
umpire to see me.' I think he proved that very considerable
military ability may be developed m the unprofessional soldier,
given certam native qualities, the study of principles and con-
stant and precise care in resolving problems set on the ground
 MEMOIR xlix
and controlled by trained officers. At the time of the war in
South Africa, the boys in the street saluted him as 'Cronje' 1
and by this sobriquet he became known to famous soldiers.
That Cronje was also professor of Logic was a puzzle the regular
soldier's mind refused to solve. On an occasion at Aldershot,
Lord French, whom Wilson much admired, visited the camp.
Wilson had seriously hurt his knee by coming off his machme,
in the most reckless descent of a hill I have ever seen a man
of fifty attempt, and could not take the field. Lord French,
after the mspect10n 1 said suddenly ' I must see Cronje ' and
turned with his staff officer down the officers' lines. Wilson was
seated m his tent, not fully dressed, but he came out, hke
Socrates on a famous occasion, and was soon deep m talk with
the general, without a trace of self-consciousness. He frequently
commanded in what used to be called ' sham fights '. This he
much enJoyed, and left no stone unturned to wm, Writmg to
a brother philosopher z he throws m these remarks at the end
of a long letter: 'One thmg I am lookmg forward to as
fac1htatmg my work 1s givmg up volunteermg. 1 shall have
completed 20 years of 1t. I delight in it because I am so fond
of tactics, but it doesn't simply take the place of ordmary
exercise. I find 1t mvolves a good deal of energy, because
I have to tlunk and arrange, being m a responsible position.
Moreover I have so often to command on Field days, and as 1t
1s no good domg 1t unless it 1s done thoroughly I go over the
ground carefully and take the officers, who are to serve under
me, over 1t afterwards-that may mean several afternoons.
Besides 1l excites me greatly, which wouldn't matter so much
1£ 1t wasn't in term. In this way I lost all last week as far as
private study goes I thmk, however, I gave the officers a very
interestmg time. We had to defend agamst the town and
yeomanry, they two maxims and we only one, both sides cyclists
but we no yeomanry I am generally made to attack m our
wars, and havmg to defend I thought I would show an example
of an aggressive defence, giving the enemy no comfort. We
succeeded on the problem given us entirely and captured one
of his guns, while I succeeded m keeping the other from ever
1 Some boys with less understanding shout.ed ' KrooJer •
• B. B : the remainder of the letter 1,;; given in Part V.
11 773•1 d
 MEMOIR
getting into a fair position. I like this in the vatotion especially
at Aldershot, where last camp I was presented by the Colonel
with a cucumber as field marshal's staff in consequence of a suc-
cessful command. You will comprehend the fascination of
tactics anyhow. Last year, or rather longer ago, we defeated
Cambridge here (on our ground however) and by using a wood
I kept back and neutralized about half a battalion of their
infantry with about 25 of my cyclists (as the umpire ,a~terwards
mformed me). I used to think this was mere recreation but
the more responsible the work the more I find it is not so : it
is urgently necessary I think to give up and I have got over
the pang, especially as I leave the cyclists m charge of - -
• • • • •. I can turn out with the corps I daresay sometimes,
but it will make all the difference that I have no responsibility
for training and efficiency. I gave up Kriegspiel long ago, it
was too exciting. Well, now I am starting to" face the music".'
Here, m his own words, are his love of victory, his passion
for a problem, as well as his indefatigabihty There was a kind
of restlessness in his intellect which made his thought not merely
discursive hut sometimes wandering ; he was rarely content as
are most men m their maturity with doing a thmg well, he
coveted the prize of recognition. The phrase ' as the umpire
afterwards informed me' betrays the writer.
Time would fail to tell all that I might of his exploits m this
mimic warfare, of how his cyclists were on the pomt of sur-
prising a line of outposts, when their position was revealed ' by
the old man's beard sticking up out of a drain pipe ' ; of the
occasion when, late m an afternoon, after a turning movement
outside the prescribed bounds of manreuvrc and even off the
Aldershot map, Wilson suddenly opened rapid fire upon the
battalion standing at ease after the day's fighting. Similar
incidents will no doubt occur to all who have been familiar with
operations carried out with the ambition and the spirit of the
amateur tactician. Suffice it to say that he did a great work
for volunteering m Oxford, not only by begging from Dean
Liddell, m the vacation when the governing body was away, the
use of the Meadows for drill, and by collecting subscriptions, but
by his example, his contempt for ridicule, his serious study, and
by twenty years of teaching his subordinates. It was Wilson
 MEMOIR li
who did most to foster the enthusiasm and knowledge which
later made the Officers Training Corps a success in Oxford.1
He was as happy as a lark at Aldershot, very pleasant in the
mess, never felt by his juniors to be censorious and, on an
occasional guest night, he could be a boy with the best. He
had a natural regard for the best type of soldier, did not import
h1s dialectical skill mto h1s conversation with them, never fell
into the easy temptation of suggestmg, as the clever amateur
sometimes will, an over-confident belief in his intellectual
superiority. His military life 1s of a piece with the rest. He
chose a recreation which was not an mtellectuaJ refreshment,
and was never rested except perhaps m body. Only a very
strong physique could have stood all that he asked of himself.
A contemporary at Balhol, 2 who could not be accused of indo-
lence, once replied to a question as to Wilson's undergraduate
days, • I only remember that he worked pamfully hard '. So
1t was all h1s hfe. He would have worn himself out very much
sooner, 1f he had not had so sound a constitution When he
stripped you were surpnsed by the mould of his frame-an
excellent chest and all the body well proportioned. He had
a good appetite and heartily enJoyed the scent and flavour of
simple dishes His heart was very strong , he was very active,
even nimble, and he had great nerve He told me once, when
we were clambermg m the clerestory of Beverley Mmster, that
he used to fear looking down from a height • I determined to
and did overcome it by use, and I have smce been about freely
on the roof of Cologne Cathedral and never felt a qualm.' He
was however a wretched sailor and this ma)- account for his
' The Corps was launched m Oxford by a puhhc meeting which Sir
Herbert Warren, K CV O , Hon D C L , President of Magdalen, the then
Vice-Chancellor, General Sir Ian Hamilton and Viscount Haldane, 0 M,
addressed Lord Haldane told me that on this occasion Wilson said to
h1m, ' Don't let us talk of the new Reahsm, I am thmkmg only of the
organization of armies • The introduction to the Oxford Unw,rstty Roll
0/ Se,vice, p 1x, doe,; unconscious mJUl>bce to tbe • body of eccentncs •.
who preceded the O U O T C The volunteers m Oxford were much more
than this and during the Boer War upward,. of 800 served m camp m two
su1.cess1ve years The numbers m the OTC. were increased m 1913 largely
through the remarkable devotion to duty and the umque peuonal magnet•
Hm of the late Brig -Gen R C Maclachlan, Rifle Bde. He was first
AdJutant and then O C the Contingent He fell m action
1
The late Sir Thomas Raleigh, K.C S I., afterwards Fellow of All Souls.
d2
Iii MEMOIR
original dizziness. He would have been a good athlete had not
his sight been defective and his interests elsewhere. He was
astigmatic and when looking at pictures, if he forgot his glasses,
would correct the failmg by an optical artifice of his own. He
had indeed a variety of devices which diversified life for him
and were allied to his love for physics. I am sure he was more
keen on the mechanical toys he gave to children than they
themselves were. This was especially so when th& gyroscope
became common. He loved puzzles, and above all the joy of
di11covering the answer for himself, and this led him away into
those tempting inqumes which wasted too much of his time and
diverted his energies from his proper tasks. He invented or
recalled from boyish days a curious device producing a stereo-
scopic effect and wanted to get it patented as a Christmas toy.
I wrote to a fncnd, an optician, who replied courteously saymg
he knew the principle and suggesting a practicable hne m the
matter. Wilson was quite angry, said the man had missed
the real point, and soon after appeared to have forgotten the
idea that was to have made his fortune.
What then was the private life of this man, so reckless and
restless in work and recreation ~ He was certainly a most
lovable fnend to his younger contemporaries. Like many great
teachers, he found himself most easily m the company of his
juniors. He succeeded m discharging the balance of years and,
m the country especially, he made you feel as if he and you
were undergraduates together. In fact he only became young
when he began to grow old m years. One of his juniors puts
this very well : ' He was a delightful holiday companion and
a careful, enthusiastic and energetic guide to good scenery and
to other good tlungs as well. At times he would show a most
boyish vigour; '\\alk, chmb and run with the best; at the age
of sixty he bathed, on a sudden impulse, m an ice-cold tarn on
the snow level in Switzerland, and he could be on his legs for
hours with a total disregard for food.' 1 He had a capacity
unusual in clever men and one you would hardly have expected
from so contentious and quarrelsome a writer, the power of
listening quietly to anything you had to say upon subjects in
which he was interested but not proficient. This is true especially
• C. W. B to H. A. P , 1 v 19 Quoted also in Muid, I.e.
 ME:t.101g liii
of music and the fine arts. In our visit to Beverley it was
a delight to see him absorbed in the beauty of the churches,
with all the fresh enthusiasm of a schoolboy. So in regard to
music ; he would constantly come to our house to hear the
pianoforte and would sit buried in the sounds. Staying with
us m the country, just after his father's death, he looked up at
the end of Beethoven's A fiat sonata and with reference to the
funeral march 1 said. 'That 1s the utterance of a man who had
reahzed that death 1s a definite good : it is the expression of
a master's conv1ct1on.' He was quite other where hrs scholar's
vamty was mterested. Unless you happened on something that
chimed with his opinion or suggested a new side to a problem,
he was curiously mtolerant of d1fference On the other hand,
rf you consulted him as a superior, sohc1ted hrs advice, he would
lavish endless pams in enhghtenmg you, regardless of his own
employment, and was seriously troubled 1f you were unconvmced
He was hke a religious man trymg to secure a convert He
admired the classical poetry of which V1rg1l 1s the great exemplar
and would occas1onally (as when I told him we were off to the
ltahan lakes) burst mto lines hke ·
' Anne lacus tantos ? te, Lan max1me, teque,
fluctibus et frem1tu adsurgens, Benace, marmo.'
These he recited ore rotundo and with a gust hke that with
which he would enJOY a salad Drama and romance he would
have formally correct, and expressive, in m.1.tter, of the normal
moral outlook. In the graphic arts he admired the delicacy
and minute prec1s1on of hne engravmg and old mezzotint (he
esteemed Lalanne, for instance, highly), and, except where
nature was concerned, enjoyed the more formal and detailed
schools of pamting. He was a lover of the Flemish and Dutch
painters ever smce a vacation tour in I 892 with C. L Shadwell
m the Low Countries. Bach and Beethoven were his favourite
composers and, as his own thought appeared not seldom to
'find no end m wandering mazes lost', so he delighted to forget
and find himself in the fam1har yet ever fresh, the mexhaustible
yet defined wanderings of music. Whatever may be the verdict
on his central philosophic tenet, he was certamly most objective
1 Op. 26. ' Sulla morte d'un eroe ',
liv MEMOIR
in his enjoyment of beauty, allowing the reabty of artist's vision
or of nature's pageantry to speak direct, never throwing a veil
of sentiment or melancholy emotion over what he saw or heard.
He enjoyed the sounds and smells and tastes and sights of his
world, disliked suggestiveness, occasional writing, exotic flavours,
anything like what is called impressionism. He would come in
to a concert and sit by one and follow the score. He kept his
place by counting the time on his fingers, as he w~s not a
practised reader. Rarely he played a little himself and once
gave us a rendering of the first movement from a Dussek
sonatina. 1 His time, which he had found difficult at ten, was
now perfect, his touch rather staccato ; he was as excited as
a child over it and hke a child thoroughly enJoyed his little
performance Whatever he did he never suffered from that kind
of nerve-exhaustion which brings in its tram depression and
self-depreciation. In connexion "11th music I saw him for the
first and only time m c-ompany with his father. I was sitting
in the gallery of the Town Hall, when my attention was attracted
by a tap on my shoulder. There was W1ls~n behind me and
by his side stood an even smaller man, with a longer and more
silvery beard. The older man's mmd was clearly fazlmg, but
his son had hopes that the Joachim Quartet would give him
pleasure He hardly spoke but sat very quiet and patient
through the music and nothmg could have been tenderer than
his son's bearing towards him. As they went out together, the
younger supporting the elder, J was irres1st1bly reminded of that
scene in the Roman Imperial senate which 1s said to have earned
for Antoninus his cognomen Pius.
This innate reverence, thts pietas which appears even m his
boyhood letters, found satisfaction in a deep love and admira•
tion for his parents, 2 whom he induced to remove from Clanfield
and settled for their declining years at Ishp, and m an engrossing
devotion to his wife She was a Hanoverian lady, Charlotte,
daughter of Amtmann a. d. Schneider, of Gifhorn m the province
of Hannover, and educated m the Limebcrger Heide. Neither
1 Op. 20, No 1.
• Dr. C L Shadwell, late Provost of Onel, wrote to J C. W. of • the
happy contentment wluch seemed to attend the last years of her I.de •,
when Mrs J. Wllson ched 1n 1902.
 MEMOIR Iv
she nor her husband had any illusions as to the ultimate aim
of the Prussian monarchy. One at least of her brothers had
fought in the Hanoverian army in the Six Weeks' War. When
war broke out in 1914, Wilson tried to get his wife's nieces, who
were (except his son in South Africa) his only relatives, to come
and live with him in Oxford. At the same time he was wntmg
to beg me to inform any person of influence I could approach
that he was convinced that the only way to secure decent
treatment for our men imprisoned in Germany was to ' have
out German prisoners of high rank and, after due notice given,
to hang them in cold blood ' unless the outrages ceased.
His married home was first at No. 26 Winchester Road;
later he removed to I 2 Fyfield Road, a simple and peace£ul
house where he worked in a study opening out of the drawing-
room, the two rooms being heated by a German stove. The
study led to the garden and looked east to Elsfield. His books
were few, the room always tidy, the furmture severe and good.
Over the mantelpiece was an original landscape by van de Velde
and on the shelf a portrait of the elegant and handsome Lotze.
In this quiet privacy Mrs Wilson and he entertained their
friends, principally to Sunday midday dinner, and a very
pleasant host and hostess they were. She insisted on his
brushing up for these occasions and in his fine blue double-
breasted frockcoat he looked hke an old officer of the Crimean
d..i.ys, carrying himself erect and enJoying the good things of the
table The conversation was homely and easy, the Wilsons
delighting to brmg together in this simple way acquaintances
of theirs, young and old, some of whom were perhaps not too
much befriended m Oxford. It might have been the table of
a professional man or pastor in a quiet provincial town , so
simple yet generous was the fare, so homely and unpretentious
the appomtments The waiting was assisted by a very small
boy in page's buttons and the control of a succession of such
pages exercised Mrs. Wilson very much. Her friends in Oxford
were amused and sometimes fatigued by her perpetual pre-
occupation with the tenant of this office, whom she would always
refer to as ' ze boy '.
Mrs. Wilson's health was, after the birth of their only child,
gravely affected until her death in January 1914. She was
lvi MEMOIR
indeed almost a chrome invalid. ' This threw upon him a burden
of daily nursing and household duty which he shouldered with
an unvarying fortitude and patience.' 1 She was a simple
creature, with a remarkable fondness for and understanding of
children and animals ; she had plenty of shrewd common sense,
and among her husband's older papers are traces of an attempt
on her part to help him as an amanuensis. But she had not
the kmd of mind to be of assistance to him in his 1Vark ; in
fact, though she admired him ·for his position, she had no
sympathy with his labours. She had a letter of his m her hand
when I called one day and, supposing 1t to concern a philo-
sophical subJect, said : ' I am so glad , John has written to say
that he has fimshed his dull book.' She was heartily amused
when the word ' dull ' was amended to ' drill '. Her illness made
her hysterical and I have known her husband sit with her, during
one of their hohdays, for hours together, on the stairs of the
house, holding her hand to soothe her, on a heavy day when
thunder was m the air. And so 1t was m many of the smaller
troubles, physical and moral, of their hfe. He would be sharp
and even cross with her occasional banalities, his intellect pre-
dominating, but he really suffered with her m all her many
atlments, spent hours in the Bodle1an reading up medical treatises
on her presumed illness and, when she was dymg, he seemed to
suffer an anguish more than mortal. At her funeral, 2 which
started from the chapel at New College, his stony grief was
most pathetic, and his devotion to her m hfc and death must
have made many of his friends feel that this strong passion
justified what had often seemed to them the intolerable waste
of time spent by the professor m running small domestic errands
and m listening to the dally tale of petty d1fficulties and
grievances which most married men are content to leave in
happy ignorance to their wives.
Like many other intellectual men, like his friend Lewis Carroll,
he was very fond of children and played with them on all fours
• H. W B. J, le
• Wilson wrote an account of this sad event for a German newspaper.
The article speaks of • a beauty and stateliness wlu.ch even m beautdul
and stately Oxford have seldom characterized a private funeral . . to the
accompaniment of beautiful and touching music the famed New College
ch01r sang the two lovely hymns "Jesu, Lover of my soul" (Wesley) and
"God moves in a mystenous way", that noble poem by Cowper••
 MEMOIR lvii
like a big dog. He never forgot them at Christmas and would
go the round of Oxford with valentines for his many child
friends Here are two valentines and an Easter poem, of 19()8
and 1909 ·
' But Jane, my dear,
This is leap year,
Why aren't you courting me ?
But p'r'aps 1t means
That little queens
Are served on bended knee,
And court to them and not by them must ever rendered be.'
' If Valentine day 's Sunday,
Whatever can we do ?
Hug Saturday or Monday ~
Well I don't know, do you ?
For Saturday 's too soon
And Monday 's much too late
Says Jane, " Oh how you moon,
Professor Addlepalc '
Both Saturday and Mondc1y
Belong to Valentine ;
I hope, Professor, one dc1y
Your wits will brighter shine" '

' In quamt olcl Germany,

Where fames most abound,
And where the cleverest housewives
In all the world are found,
They tell us that the Easter eggs
Are laid by th' Easter hare ,
And yet they don't know why, Jane:
We do-We have them there
We know the reason why, Jane,
In England, aren't you glad ?
The Easter hare 's the March hare
And the March hare 1s mad.'
This 1s one out of many letters to the same child friend,
written at a time when his wife was dymg
13 Jan. 1914.
{Forgive bad writing. I have so much to write that I have
to write quickly.)
Mv DEAREST JANE,
It was very sweet of you to write me another letter
because you were not sure whether you had written. You did
lviii MEMOIR
wnte, but I am very pleased to get another letter, for the first
showed the influence of the whirl of Xmas gaieties and was
what we grown-up people call ' merely common form '.
You hoped I had a lovely Xmas as no doubt you did and the
people to whom that was possible, but I was consoled by finding
that you hadn't even signed your name-for which we grown-ups
have another long word 'preoccupat10n '.
I knew you really had been overwhelmed by the good things
of Christmas and had been put out of your real self. But your
dear little second letter made everything all right. • I· am so
glad to hear you have had such a; pleasant time.... I am sorry
to say Mrs. Wilson is very very ill.
With much love.
j. COOK WILSON.
He did not talk commonly on so-called religious topics, m fact
I do not think he was in the normal sense of the term a religious
man Perhaps he thought 1t unwise to discuss such subJccts
with eve1 y friend He read a str1kmg paper m Professor Sanday's
lodgmgs upon the Ontological proof for the Existence of God. 1
It was not all read, though the reader stopped close upon mid•
mght. Its purport was that belief m God, apart from revelation,
1s not to be proved rat1onally1 but rests upon the umversal
mstmct of reverence for some bemg higher and better than
ourselves Thus it was akm to Jacobi's saymg that 'the behef
m a God 1s an innate devotion before an unknown God '.
Upon his father's death, he came to stay with us in the
country and we had many walks and talks. One evening we
were standing m the graveyard of Great Bnckh1ll, Beds., which
commands a magnificent prospect. The sun was smkmg with
more than his usual splendour and the thought of Gray's Elegy,
a favourite poem of his, was in both our mmds I ventured
to question him upon immortality. He remamed quite still,
gazing out as I had seen him in his old rooms at Oriel, then
quickly turned the conversation back to the Elegy and so to his
father's memory. Somethmg however of his more intimate
1
Sanday collected the representab.ves of most diverse rehg1ous views to
hear what was advertised as a paper ' On the Ontological proof for the
existence of God' The notes revised by Wilson are pnnted m Part V m
their rough, unstudied form, with a few later addib.ons. Wilson used to
say that the essence of the matter lay 1n St Anselm's ontological proof,
restated in modern terms The paper itself, however, 1s neither a statement
nor a cntlcism of that proof.
 MEMOIR lix
feelings in the matter may perhaps be gathered from the following
letter written four months after Mrs. Wilson's death :
12 Fyfield Road, Oxford.
II May, 1914.
The news of your dear boy's death was a great shock to
me....
My heart goes out to you both in your trouble and I assure
you of my deep sympathy. Many hearts will go out to you,
and that you will find a comfort as I did I have looked often
at the kind words you wrote to me when my own sore trouble
overtook me, and I little dreamed I should be sending a message
of sympathy to you m a similar calamity before I had thanked
you for them. The truth is that I have only been able to begin
my answers to my kmd friends rather late, and for many reasonc;
can only get on with this expression of my thanks rather slowly.
I have found 1t a great solace. May you be rewarded now for
your share m comfortmg me then Perhaps I might be allowed
to add a word beyond my own personal sympathy. I will tell
you quite frankly that 1t seems to me one's philosophy, 1£ one
faces out the whole bitter problem, leads mcv1tably to a con•
vtchon of personal immortality You know that I am quite an
uncomprom1smg thmker, and so you may be glad to know my
deliberate conv1ct1on I daresay that some well-meaning philo-
sophers-espec1ally when not m the reality of a bereavement-
nught be afraid to profess the ordinary and, as 1t seems, uncritical
rehg10us behef not only m 1mmortahty but m our reunion with
our beloved ones who have done with the mortal and changeable
body. I am glad to thmk that an absolutely severe philosoph1c
consideration seems to me to lead to this conviction of the
ordmary religious consciousness Love is the victor over death.

With kmdest regards,

Yours truly,
J COOK WILSON.

My dear wife took such sympathetic interest m your dear

boy's illness and her heart was full of pity for you. She did
so keenly feel the sorrow of others. 1
At this period he became once more as he was m his cluldhood,
overflowing with affection and most dependent upon his friends.
He might well have said, in his own choice phrase of childhood,
1 J C. W to F J W.
 MEMOIR
'You ask me to strike the key of pleasure, but I have none to
tell you of '.
Yet our memory of him must not be saddened by the deeper
shadows which fell upon the evenmg of his days. He had a very
happy span of years, ' passing through common hfe with the
natural emotions of common men '. He suffered m the deepest
part of that common nature during his sister Margaret's insidious
illness. He had to face the mevitable loss of two beloved parents
and last of all that of his own wife , yet even after Mrs. Wilson's
death his natural elastic gaiety began to reassert itself. The
summer before he had bought a motor-car and this gave him
needed rest and recreation. ' I got something this summer
which gave me much pleasure-a little motor-car I do so
enJoy taking my friends m it In general I can only go m the
mornings I have on the doctor's advice been motormg again
lately. I wasn't well and thought I'd better not: but he said
I must and it has done me good Lately I took Mrs. Ross to
Church Eynstone (for a servant's character 1), 32 miles and back.
Also I took out the Rector of Exeter and Professor Butler,
formerly of New College, and Mr. Jenkinson .. I know you
will be glad to hear this because I have not been able to get
away this Xmas [1913] at all' 1 Earlier than this he used to
draw Mrs Wilson about m a trailer, that earlier type of the
side-car, and would come m, exuberant with health and spirits
and the pleasure of givmg pleasure to her. One day he came
back from a big nde alone and, turning to his wife, quoted
with allusion to the case with which he had compassed so many
miles:
'Away went Gllpm, and away
Went postboy at his heels I -
The postboy's horse right glad to miss
The lumb'rmg of the wheels '
I am sorry that I never saw lum m full career in his motor-car
nor ever drove with lum, but those who did have assured me
that he bated no whit the recklessness which had been his in
the old days of the bicycle.
And so his life passed m Oxford, a tutor of the old school ;
to work or to lecture in the morning ; to doze for an hour after
1
J C W. to J C R. F (aged ui), 13 1 14, quoted also above, p. lvu.
 MEMOIR lxi
lunch (unless he went afield for exercise), to get up refreshed and
work again till dinner and after dinner up to a reasonable hour
once more ; this, with interruptions to run errands for his wife,
to deliver notes of his own or hers by hand to save the postage,
to make excursions to the Bodleian or other library ; once
a week to discuss philosophic topics with a few intimate and
younger friends over afternoon tea ; this was the tenor of his
hfe in term. In vacations, especially in the summer, he went
away from Oxford taking work with him ; to Germany, to
France, to the Lake country, to the moors or sea, wherever he
hoped his wife might benefit and be happy. 1 When away he
worked hard or, 1f he took a day off, he took it without hesita-
tion and drank the sweet pleasures of the countryside or studied
the architecture and antiquities of the neighbourhood to his
heart's content On these expeditions he had sometimes the
company of a congemal younger friend or married couple He
might have said 'when you have seen one of my days, you
have seen a whole year of my life; they go round and round
hke the blind horse in the mill ', only he would have said 1t
without the repining of the poet Gray and he certainly believed
that the horse made some progress and was m no sense blind.
On the contrary, regular though his round was, his life was
ceaselessly diversified ; hke a young hound he caught at any
and every intellectual trail and only the rigorous performance
of his academic duties kept him from mfi.mte digression. He
drew a little, taught himself to read Italian, and once began
a novel, written m the Miss Mitford or Mrs. Gaskell vein.
Mr Joseph, after enumerating the tale of his comparatively few
printed works, adds with a spice of humour : ' And among his
remains are papers on the existence of God, on the conception
of Life, on Homer, on Greek musical modes, on Greek tactics,
on the undergirdmg of Greek ships and the beaks of triremes, on
umversals and on the good will ' 8 To justify the tenmty of his
published work we might perhaps apply to him Goethe's saying,
' the man who has hfe m him feels lumself to be here for his
own sake, not for the public '.
1
He left Oxford practica:lly every vacation Some of his letters, which
are pnnted below, are a commentary on the old text coelum non animum
'""'""' • H. W. B. J ., 1 c.
 MEMOIR
His energy of mind contmued even after the pernicious
anaemia, which at length carried him off, clearly manifested
•
itself. ' Lying mostly m bed, he still worked privately with
a few pupils, and even in the spring of 1915 came down to
college and delivered half a dozen lectures.' Keenly interested
in the progress of the war and opt1mist1c as to our eventual
victory, he wrote occas1onally to the papers, notably upon the
formation of an independent cyclist force and upon ,tbe pre•
servat1on of the teeth of soldiers. He wrote to me more than
once. especially in regard to mechanical devices for protection
against, and the destruction of, submarines. These I was able
to submit m the right quarter and hke so many similar letters
they received careful attention from the experts m the subject.
But the ideas were naturally ' too mathematically conceived ' 1
from want of contact with reality. He was, for example,
strangely unaware how a shell fired from a ship will ricochet
away along the surface of the water. His cyclist idea was
adopted and he was gratified by a leading article m the Man-
chester Guardian which described how, m the stress of war,
cychsts had protected themselves agamst ~avalry by making
a zareba 2 with their machines This was a practical not10n of
his own which had been very much ridiculed m Oxford common
rooms years before
In a sense the moment of his death was timely ; he did not
hve to endure that early summer of 1916, when each day brought
its new tale of young scholars killed or wounded, nor the long
months of war which followed. He passed away after a few
days of unconsciousness on the IIth August 1915 1 and we were
summoned to Oxford to attend his funeral three days later.
As I reached New College the closmg words from St. Paul
were bemg read · · ' steadfast, unmovable, always aboundmg m
the work of the Lord, forasmuch as ye know that your labour
1s not in vam m the Lord', and the August sun, not long past
his meridian, was kindling the glass m Wykeham's ante-chapel
and stainmg the pavement with its broken rays. From Oxford
1 Said of Wtll.iam Chdhngworth, a ma.n • of a stature little superior to

Mr Hales ', by Lord Clarendon, Life, ed 17 S9, p 30 (Quoted m Characters

of the Seventeenlh Century, D. Nichol Smith, Oxford, 1920, p 177)
1 Called by htm a lager , cf M1l1tary Cwling, 1890, note to p 9 (pp

17-18)
 MEMOIR lxiii
we drove to Islip, where he was laid to rest beside his wife and
parents and sister in the churchyard which commands the
valleys of the Ray and Cherwell. A month later I was standing
on the bttle eminence, called Mont Rouge, watching our English
shrapnel bursting over the grim lines of the enemy front, and
only with difficulty could I free myself of the notion that my
old company commander in the 1st Volunteer battalion of the
Oxfordshire Light Infantry was at my side So strong was his
vitality that, for the first time in the presence of actual warfare,
I turned involuntarily as if to hear his appreciation of the
features of the ground. That closing scene was appropriately
set amid the beauty of the College which endows the chair he
held ; the broken sunlight seemed to speak of the intense
pleasure he took in the glory of the physical Universe and m
the power of mmds hke Newton's; an English churchyard 1 is
a good resting-place for the son of a minister of the Methodist
New Connexion, who with a full measure of the polemical temper
of that sect loved well the order and dignity of the Nat10nal
Church; finally, the stern struggle m Flanders seemed no
mappropnate background to the close of a hfe dedicated m so
much of its rare le1o;ure to training a succe . . sion of scholars m
the rudiments of military science.
Remarkable though these fragments of Wilson's higher thmk-
mg may appear and highly characteristic of his fearless and
uncompromising mmd, they do not exlubit, save dimly, the
power of his confident and passionate assertion of a hardly won
philosophic creed Except m their literary form they cannot
present the unaffected simphc1ty of their auth'>r, nor the almost
excessive devotion to detail that was his characteristic in small
thmgs as well as great. No words of mme will convey a true
notion of the exuberant and boyish happiness which triumphed
for so long over domestic anxiety and sorrow, the entire absence
of a scholar's self-consciousness and timidity, the affection for
friends and the love for parents and little children, winch m his
nature ran so strong and deep
1
In his account in German of Mrs Wilson's funeral he says of what was
to be his own resting-place ' The grave itself, adorned with a celtic cross
of white marble, occupies the highest point of the lovely churchyard, with
an extensive view of a pretty landscape.'
 MEMOIR
We shall always regret that the necessity of a long absence
prevented most of us from waiting upon Wilson's declming days.
His loss makes a greater gap than we expected in the circle that
enjoyed his friendship and profited by his words. When we
returned, Oxford, swollen by the unexpected volume of maturer
students, was only gradually resuming her normal channels,
had hardly subsided within her banks. She was recruiting her
heavy losses, restoring and remaking her studies, discussing how
best to commemorate her dead, ambitious of a future worthy
of her past Oxford the same no doubt in essence, yet how
changed for the men of my generation ! In the tyranny of war,
manhood had slipped into middle age ; our own ranks were
thinned, and we marked for the first time that the majority of
those who taught us were no more Without the familiar faces
Oxford 1s strangely unfamiliar; but of all that is gone, more than
my gay company of the Trammg Corps , 1 the happy careless
friends, my pupils , more tl1an the cheerful and witty Charles
Fisher, 2 or Foster Cunhffc's 3 curious melancholy, even more
perhaps than the gentle and cxqu1s1te irony of Reginald T1ddy,'
I miss the little figure, the hght quick step, the good grey head,
the secure affection, the profound plulosoph1c ardour of Wdson
himself. A. S. L. F.
March 1921.
1
E Company, 0 U O T.C, was handed over by me tn 1913-14 to J L
Johnston, formerly Junior Demy and afterwards Fellow of Magdalen and
then Fellow of New College He se1ved with the 52nd m France and fell
at Festubert 12 v 15 He would certamly bave r1ben 1n Church or State
I wish specially to mention two of my Colour-Sergeants, both of New
College H T Culhs of Alleyne•~ College of God's gift at Dulw1ch, Scholar
of New College m 1899, wa& on leave from India when war began He
fell at Armentillres 10 xii 1 'i, servmg m the 12th (Service) Bn Rifle Bde
The other 1s Geoffrey W Polson, lulled at the Aisne, dunng the first allied
counter-stroke, 15 ~x 14 He wai, a subaltern in the zst Bn Royal High-
landers There were no better NCO 'sin the old Volunteer Bn. and the
new Tr0.1mng Corps Of the cadets who attended my lectures m nuhtary
sub1ects 111 May and June 1914, sixteen out of forty-four fell in the war
• Charles D Fisher, Scholar of New College, Semor Student and Censor
of Cbnst Church, lost m the foundering of HMS In111ne1ble 1n the action
off Jutland, p v 16
• Sir Foster C H Cunliffe, Hart, of New College, Fellow of All Souls,
formerly Colour-Sergeant of E Company, 0 UR V Bn., lolled serv1ng with
the 13th (Service} Bn Rifle Bde. on 10 v1116 at the battle of the Somme.
' Reginald /· '.E T1ddy, f,cholar and Fellow of University, .Fellow of
Tnwty, one o Wilson's cychsts Killed 1n the trenches when commanding
the men of his village in Oxfordshire m the 2nd Terntonal Bn Oxf. &
Bucks. LI. on 11.vm.16.
ELENCHVS OPERVM
TESTIMONIA
 LIST OF PUBLISHED WORKS AND
CONTRIBUTIONS TO REVIEWS
[The articles m German penod1cals are wntten in that languagt' The
following abbreviations are used ·
0. P. S. =- Transactions (or Proceedings} of the Oxford PbilologJcal Society.
J of P = The Journal of Philology
Cl R = The Classical Rev1l'W
Cl Q = The Classical Quarterly
The remark.c; m square brackets are to indicate the character of the arbclec;
and 'IOml't1mt's to c;how add1tional content~ ]

1873.
Chancellor\, Latin Es~ay, Quaenam fuent rei•era Epicureorum
philosophia, recited m the Shcldonian Theatre, Oxford, 18 June,
:\-IDCCCLXXIII, by John fook Wilson, B.A , Mathematical Scholar
of Balliol. (Oxford)
1879
On rearrangements of the Fifth Book of the Ethics
0 P S , No 83, 14 March.
Aristotelian Studies, I On the structure of the Seventh Book
of the Nicomachcan Ethics, ch 1-x. (Oxford )
0 P S, ~o 8,+, 9 ~fay
1880
J Cook Wilson, An~toklic1.n Stud1cc., I, \'om Verfasst-r
Gotting1sche gel Anzc1gen, St 15, 14 April
[An account of his book, with d1scm.,ion of interpolations m
Eth E11d, Cat and An Pr]
Prof Suscm1hl's Ed1t1on of thl" l\'1comachea1i Eth1cs
The Academy, 26 Jun<'
1881.
Aristote, Morale a Nicomaque (8me hvre) Par L Levy.
Phil Rundschau, i 39.
(Review of the book, with suggestions as to the text-structure
and the existence of parallel version,; here and elsewhere.]
Notes on some passages in the Politics
J of P. x, pp So-6. ..
[Stmcture of text, Bks IV. 3; III. 10-n; V 1-3; VII. 1-3
and 12-13 (13-15) and 1260a 22, 1298a I, 1342b 23,l
 ELENCHVS 0PERV1\I Jxvii
1882.
(i) Zcllcr's interpretation of Xen. Mein. iv. 2, 34 and iii. 2, 4.
[Socrates' eudaemonism.]
(ii) Plato, Phil. 3I A.
(iii) Aristotle's criticism of the definition of pleasure in the Philebus.
(iv) Theophrastus, De Sensu, § 90. [emendns. of Plato, Tim.
66 D ; Arist. 443a 2I, 359b 32,]
0. P. S. 1881-2, pp. IO-I3, IO Feb.
Conjectural emendations in the text of Aristotle and Theophrastus.
J. of P. xi, pp II9-24.
[Aristotle, 443a 6, 444 8 16, 11 2, 453a 27, 457" 31, II34a. 1, 056•
29, and Thcoph De Smsu, § 90 j
Stud1en zu Arh,totcles Politik von H Bucl1-.rnschutz
Phil Rund~chau, ii 39
[Attempts to prove that ordu of books in tlw Aristotelian
vulgate 1s as old as Anus D1dymt1'>' t-pitom(' ]
The interpretation of certain passages of the De .lmm,i in the
editions of Trcnddcnburg and Ton,tnk
0 P. S. 1882-3, pp 5-13, 17 NoY.
'ApttTToT,.>..71s Tl'Ep2 iftvxijs by E. Wallacr, :\I A.
Phil Rundschau, ii. 47.
1884
The gemunencss of An~totl(', Rheforir, ii 25-6 [also text of
14020. 29 and b 19l
The poss1btl1ty of a conc<'pt10n of the Enthymcmc earher than
that found m the H.lwtonc and th<' Pnor Analytics
0 P S 1883-4, pp. 4-6, 22 FPb. and 7 :March.
1885.
De Amma, 431" 24-6
Meteorologica, 1v, ch. 8-9 [on TEyi.ro, = soaking].
Nie. Eth. 1097b 8 [cf 1887].
0. P. S. 1884-5, pp. n-13, 6 March
A recent emendation of Aristotle [Metaph. 1035a 14].
The Academy, 2 May
1886.
Nie. Eth. n33a 14-16
Platc4IR, 330 E.
Nie. Eth. iii. t. 17.
0 P. S. 1886-7, pp. 2-4, 5 Nov.
e2
Jxvib ELENCHVS 0PERVl\-f
1887.
The sphere of Corrective Justice in Nie. Eth. v.
Nie. Eth. 1143b 5,
On Trendelenburg, Eletnenla Log. Anst., § z.
[Doctrine of truth and falsehood in de bit. and de An.
Simple notions.]
· 0. P. S. 1887-8, pp. 2-6, 4 Nov.
Recent emendations of the Aristotelian text.
The Academy, 3 Dec.
f\'inchcates, against H Jackson, MS. readings of Etlt. Nie.
I. vii 7-8 ; VII xm 2 , II vu 14 ; \'I v 4 and 6 ; JI.
\'JI I ' \' Vil l ' I \'l, I ]

18t(8.
Xie Eth 1097h 8, 1145h 23-4
0 J> S 1887-8, pp. 20-1, 3 Feb
~ome recent t'lll<'ndabons of Aristotle and Plato
fVmdicat<'s MS rt'admg-. of Eth l\"1c u45a 27, b 30, 1147b 31,
n77a12,n7<J'1 1fi,IX x §3,andPlatoR .n8E,andillnstrate.::
A 'i, tendency to dhphcal langua!?c ]
A recent emendation of SophocJt>1,
!Vind1catei. MS reading of AJax,646-9,against van Leeuwen.]
The Academy, 18 and 25 Feb.
Lange's theory of thl' cond1t10nal sentence m Greek
[E: originally a relative and (probably) a temporal pronoun.]
0 P S 1887-8, p 22 (for brief account see ib 1889-
90, pp 54-5], 8 June.
1889
Some recent t'mcndat10ns m the text of Plato
[Vindicates the MS. readings in R 537 c and 402 A.]
The Academy, 23 Feb
The Timaeus of Plato, cd R Archer-Hind
[Rev1ew1, m] 1 The Clas-.1cal Review, iii, pp. 114-23, 183-4.
11 The Oxford Magazine, 13 March.
Mr. Archer-Hind's 'Last Word'.
The Academy, 8 June
On an Evolutionist theory of Axioms, an Inaugural lectun-.
Oxford.
On the interpretation of Plato's Timaeus. Critical st$es with
reference to a recent edition. London (D. Nutt).
Manual of Cyclist Drill for the use of the Cyclist Section of the
O.U.R.V.f Oxford.
 ELENCHVS OPEK\rM

1890.
Military Cycling, or Amenities of Controversy. Oxford
On some apparent anomalies in the use of p:,j.
0. P S. 1889-90, pp. 23-56, 13 June.

1892-3
A point of Infantry drill.
A. and N. Gazette, 16 Jan 1892
The pseudo-Aristotelian treatise<;, De JI,,/ elisso, Xc1topl1anc, <,,wgia
Cl. R. ,·ols. v1 and vii
[Six articles on Apelt'i. edition, containi. incidentally many
suggested emendations ]

1894
Ani.totlc, Sopli. El. 166b 32, 171b 37, 172a 25 and Top. 1621! I.
0. P S. 1894, 12 May

1895
lcsllmonia for the tc:xt of Arbtotlc's Etli. Sic, Mctap!i. and
An. Po
[From Ptolemy, Theon, &c J
CI R. ix, pp. 1-4

1890.
Am,totle'i. ClaSblfication of the arts of ,1cqui&1tion.
rscc 1902 J
Cl R. :x, pp 184-9, l\lay

1897 and 1898

Zu Anstoteles' Pohhk, 1258b 27-JI.
LA sequel to the article in Cl. Rc-v of 1896.]
Archiv f. Gesch. d Phil x1 2, 246-62 and xh. 1, 50-4.

1900
Suggestions for Cyclist Road Practice in Skirmi:.hing and
Patrolling. Oxford.
Inverse or" posterior, Probability.
[A proof of the principle and an explanation of the philo-
sophical meaning of probability.)
Nature, 13 Dec.
lxx ELENCBVS 0PERVM

1901.

On Aristotle's Poetics, 1451• 22.

Cl. R. xv, pp. r,.S-9
On the meaning of bra.ywy'l in Aristotle and other Greek writers.
O. J>. S, 7 June; Cl R. xv, p 430
Probability, Jame!> Bemoutlli's theorem
[A i.1mpler proof of the theorrm]
Nature, 14 llarch

1902.
F..dward Postc
Obit. notice in The Oxford l\lagazinr, 4 June.
Eth. Nie. VII. xiv. 2 and xii 2.
[Correction of articles dated 1896, 7 and 8 ]
Alrya.\mrpi71'E&a and Mrya.'A.Dlfvxla. in Aristotle.
On the u-rr~1~p.a.Ta of Greek ships. [Cf J. Adam, 1'he Republic
of Plato, h, p 44-5 ]
Plato, R. 616 E Lagainst Kron•~ ed. of Proclus in Rep.; astronomy
of the myth of Er, the Armenian]
Cl. R xv1, pp. 23-8, 203,234 and 29.?-3, 0. P. S., 31 Jan.
The astronomical concephom, m Plato, R , Bk X
[V1dc J Adam, I c , 11, App VI, pp. 470-9 J

1903
l\1emo1r of the Rev T. W Fowle, 1\1.A , 1.tte Rector of fabp,
Oxfordshire Oxford
On the g<•om<>trical prohl<'m 111 Plato'& l\Jeuo, 86 :E.
Note on LAnstotk] de lw 111,ec CJ70 8 5
J of P , , ol ,._ '\ vm, pp. 222-40
Eliz Ni, u35b 19
Cl R XVll, pp 384-5

1904.
Pseudo-Euclid, I ntroductio Harmonica [text on Concord!, emended].
On tbe Platonist doctrine of the d.crvp.PA.,,,.o, d.pi8p.o{.
The problem of the Greek modes [criticism of D. B. Monro].
Musici Scrit,tores Graeci, Emendations and discussions.
Cl. R. xviii, pp 150-1, 247-60, 278; 0. P. S., 12 Feb.,
and 387-gr
 ELENCHVS 0PERVM lxxi
1905.
On the Traversing of Geometrical Figures. Oxford.
Addendum to the above. Oxford.
Homer, Od. xxiv. 336 seq [vindication of the MSS.]
The idea of tcd.Oa.po-,,;; in Aristotle's definition of Tragedy.
Cl. R. xix, pp 144-7 and 321-2 (0 P. S , 2 June).
Lewis Carroll's Logical Parc1.dox [signed W.]
Mind, N S , No 54, pp. 292-3

19o6.
On a supposed solution of the '!•our-Colour Problem ' [against
Archbishop Temple]
The l\lath<'matlcal Gc1.zctk, in, No. 58, pp 338-40.

1907
(1) :\Iemonal Notice of D. B l\Ionro
Burs1an's Jahresbc•richt, xxxu B, pp 30-40.
(n) David Binnmg 'Monro, a short :\lcmoir Oxford
[Translated, with alterations, from (1) ]
(m) David Bmmng :\lonro, 1836-1905.
Proceedmgc; of th<' B A. 1907 [ (ii) abbreviated].
Plato, R. 442 B
Et/1 Nie 114811. 23 [emrndation]
Cl R X'\.1, p 106

1908
Clement Ale:<.. Stromateis, 1 158 [emendation].
Cl Q. ii, p 293

1909
On the use of d.U' .;; m Aristotle
Plato, Pliil , 31 c.
Clement Alex , Stroni iv. 23 [emendation].
Cl Q ih, pp. 121-6, 216-17

1910.
Natural anomalies in original compo&ition.
[Refers to a paper to 0. P S , 1909, on The Similes of Homer.)
Eth. Nie. IV. iu. 15
Cl. R. x.uv, pp. u8 and 144-5.
lxxii ELENCHVS 0PERV:M
19n.
Eth. Nie. n23b 31.
Cl. R. XXV, pp. 132-5.
1912.
Aristotelian Studies, I {1879), Reissue of, with additions. Oxford.
[A Postscript on the authorship of the parallel versions and
supplem. index.]
Inaugural lecture {1889), Reissue with mtrod. sect. omitted.
Oxford.
Syllogism of the Abbe and the penitent.
The Athenacum, 10 Aug.
1913
A metapby&1cal problem written in Greek in mutation o! the
&tyk of Aristotle
The o~f. Mag xxxi 16, 6 ::\larch.
Plato, Soph., 244 c LCnticism of H. Jackson J
Cl. Q vb, pp 52-3.
De Motu A nim 69Ba 16-24
Rh. 1403b 21 Sl'q , Po 1449a 2J !:-Cq
Phy. 231b 21.
Eth. Nie. n22b n-18.
Plato, Tim. 37 c.
Catullus, lxili. 31
Metaph. 1048• 30 seq.
J. of P. xxxu, pp. u3-4, 137-65, 166, 167-g, 300-1.
Po. 1451• 22, 1447b 13-16.
On the meaning of M.,..os m certain pasi.agcs m Anstotlc'& Nico-
HUJClsetan Ethics.
[.\o-yos to be tram,lated by' reason'.]
Cl R xxvii, pp 7-9, n3-17.
 TESTIMONIA
'To Professor J Cook Wdson, of New College, Oxford, I owe
a special debt of gratitude for undertaking in response to my appeal
an exhaustive discussion of the astronomical difficulties m Book X
and unreservedJy placing at my disposal the full results of his
investigations '-J Adam, The Republ,c of Plato (1902), vol. i,
pp. ix, x (Cambndge)
' It is only through the kmd co-operation of Professor Cook
Wilson that I have at last been able to form a definite view as to
the meaning and '>0lution of this e,.tremely complicated problem.
From 616 B to the middle- of 617 B, my commentary is mainly
based on the exham,t1vc criticisms and mvest1gatlons which he has
sent tom<' '-lb, vol ii, p 441
'Thii. Appendix 1s the result of prolonged d1scusr;ions with Pro-
fessor Cook Wil'ion and frc·ely r<'produccs nearly all his most
important argument1, '-lb, vol. 11, App. VI to Book X, p. 470.
' [The Editor] has to thank Mr J. C. Wdo,on ... for the most
cordial and ample as'iistance in dealing with the numerous passage&
in which mathematical knowledge was required. It is believed that
the translation of these passages will, owing mainly to his help, be
found on the whole correct and mtelligible '-Lotze, Logic• (Eng.
Trans.), ed. by B Bosanquet (Oxford, 1888), vol. 1, p. vi.
' The Editor has to thank Mr. J C. Wilson . for ample and
ready assistance when consulted on passagrs involving the technical
language of Mathematics or Physics, 1f the Author'& meaning in
such places has been intelligibly conV<'yed, this result is wholly due
to Mr. Wilson'i. hclp.'-Lotze, Metaphysic (Eng Trans.), ed. by
B. Bosanquet (Oxford, 1884), p. v1
'Line of treatment [of Plato's Parmemdesj suggested by Pro-
fessor Cook Wtlson.'-B. Bosanquet, Plato's Theory of Forms
(Oxford, 1903), p. 14
' Sed est cui prae ommbus grates agam atquc habeam, VIrum
dico si quis alius 'Apt.cr,w~TOv, I. C. Wilson: quem ut
socium mihi in hac editione paranda futurum speraveram, ita post•
quatn instantiora eum negotia alio averterant, alienum opus adiuvare
ct, quantum potuit, amico suppetiari non recusavit ; haud exiguam
Jxxiv TESTIHONJA
enim libri partem una relegimus, crebrisque colloqaiis collato studio
diflicillima quaeque excussimus.'-AristoteUs Eflffca N-ico,nacllea,
I. Bywater (Oxford, 1890), p. vii.
' The essential symmetry of the inverse and the direct methods
... is shown by an elegant proof which Professor Cook Wilson has
given for the received rules of inverse probability.'-F. Y. E.,
Encycl. Brit., x1, p. 378b, note Io
'To Professor Cook Wilson in particular, who read over the
whole of the proof pnnts of the work and made a number of acute
criticisms, I am much indebted '-Trichotomy in Roman Law,
H. Goudy (Oxford, 1910), preface. (German translation Dreiteilig-
keit im 1innische1i Recht, E Ehrlich, 1914.)
' ... the debt wluch I owe, in common with so many of his older
or younger pup1b, to Prof J Cook Wilson, whose death occurred
while these sheets were pas::.mg through the pres!> Various foot-
notes will show the~ that I haw made of his unpublished teach-
ing; but hi:. tllness prevented me from subm1tt1ng to him what
I have written, and h1:. authority mu!>t be made responsible for no
errors that I have made. His few and scattered publications can do
little to convey to strangers the power and stimulus of his personal
teaching. And there are subject!> on which, by his combination of
scholarly and mathematical with philosophic insight, he was quahfted
as few have been to produce new work of real value.'-Anlntro-
duction to Logic 2, H. W. B. Joseph (Oxford, 1916), pp. vi and vii.
' My obligations arc many and great . . . to Professor Cook
Wilson, to have been whose pupil I count the greatest of philo-
sophical good fortunes Some years ago 1t was my privilege to be
a member of a class with which Profeso;or Cook Wilson read a por-
tion of Kant's Critrquc of Pure Reason and subsequently I have
had the advantage of discussing with him several of the more
unportant passages. I am especially indebted to lum m my dis-
cussion of the following topics . the d1Stinction between the Sensi-
bility and the Understanding, the term "form of perception", the
Maaphysical Exposition of Space, Inner Sense, the Maaphysical
Deduction of the Categories, Kant's account of "the reference of
representations to an object ", an implication of perspective, the
impossibility of a " theory " of knowledge and the points con-
sidered, pp 200 med.-202 med., 214 med.-215 med. and 218. The
views expressed in the pages referred to originated from Professor
Cook Wilson.'-Kant's Theory of KMWledge, H A. Prichard
(O.xford, 1909), pp. ui-iv.
 T.ESTIMONIA bav
' Throughout this Essay I am deeply indebted to the critic:isms
and suggestions of Professor Cook Wilson. In particular, I h&ve
substantially adopted his account of the distinction between abstract
terms and adjectives, in place of a less satisfactory view of my
own.'-Personalldealism, ed H. Sturt (1902). Note I to Professor
G. F. Stout's essay on E,-,,o,,,
' I owe to conversation with Prof Cook Wilson the first suggestion
that this view is one for serious consideration, but for nothing in
my working out of it can I claim his authority.'-Problems in lhe
relations of God and Man, C C J Webb (London, Nisbet, 1911).
' Dedicated in affectionate gratitude to the memory of a great
thinker and a great teacher, John Cook Wilson, sometime Wykeham
Professor of Logic in the University of Oxford.'-Dedication of God
and Personality, C. C. J Webb, Fellow of St. Mary Magdalen
College, Oxford. (London, Allen & Unwin, 1918 )
' ... as my lamented and honoured teacher, the late Professor
Cook Wilson, did in a paper of marked originality, which made
a great impression on those who heard it read at Oxford, and which
I hope may hereafter be made pubhc, when the return of peace
shall have set his literary executors free to carry out the pious task
of giving to the world what he has left behind him '-lb., p. n9,
foot-note 6.
Referring to Martineau's works on Ethics and Religion, Professor
C. C. J. Webb says: 'In later years I have re-read them with
greatly mcreased admiration and have seen how well this writer
deserved the commendation which I recollect my lamented teacher,
Professor Cook Wilson, long ago bestowing on him for his bold
faithfulness to the facts of our common moral expcrience.'-Divbre
Personality and Huttian Life (London, Allen & Unwin, 1920).
'So ist das Verhiiltniss schon nchtig \-On Herrn John Cook
Wilson, dem Ari!.totelesfon,cher, beurthellt, der auf gutige Ver-
wendung des Herrn Prof. 0. Francke . die drei Hss , wie ich
sagen darf, mit ausserordentlicher Genauigkeit collationirte.'-
Awaoi M-yoi, Ernst Weber, Phil -Hist. Beitr. (p 35).
' In the attempt to interpret this difficult passage I owe much
to the late Professor Cook Wilson, who discussed it with me.'-
Aristotle's Metaphysics, W. D. Ross (Oxford, 1924), vol. il, p. 268.
'I am indebted to my friend, Mr. J. Cook Wilson, Fellow and
Tutor of Oriel College, for many valuable hint-. in connection with
the Introduction.'-A.ristotle's Psychology, Edwin Wallace (Cam-
bridge, 1882), p. vii.
lxxvi TESTIMONIA
See the following notices ·
S. Ball in The Oriel Record, Sept. 1915, pp 246-8, Oxford Mag•-
zint, :z:z Oct. 1915.
The late Rev. F. H. Hall in The Onel Record, Sept. 1915, p 245,
' Peace and fresh service to that imperious spirit '.
Mr. H W. B Joseph, Proceedings of the B. Academy, vol. vii,
pp. I-II.
Mr. H. A. Pnchard, Mtnd, N S., ;,.;,.nu, No. III, pp. 297-318.
FAMILIAR LETTERS
1859-1914
(Numbered 1-72 indusive)
 The following letters were addressed :
To Mrs. W1lc:;on senior ~0<1 1-6, and 8-12.
To I. Bywater. ~o 7.
To Mr. J W "harp N'o 13 (from Wilson's own copy).
To Rev. S C Parm1ter. ~o. q.
To B. Bosanquet Nos. 17, 24-6, and 59.
To a Schoolmaster. No 27 (from Wilson's own copy}.
To Mr. H. A. Prichard. Noc:;. 28, 30, 31, 34, 39, 6o, and 61.
To Sir W. Henry Hadow. No. 58.
To Mr. Colin Gilray. Nos. 62 1 63 1 and 70.
To C. Cannan. Nos. 64 and 67-9.
To Lieut -General H. D. Farquharson. No. 71.
The remainder were to myself or my wife. They are often
undated but I have indicated "hat I believe to be the dates.
I have to thank those ladies and gentlemen who have per-
mitted me to use this material
 1
[Shireland Hall, Birmingham.]
Wednesday aftnoon
[? 10th August 1 1859.]
MY VERY DEAR MAMMA,
I received your dear letter this morning, the reason that
I wrote my letter so bad was I couldnt keep from crying. I hope
that all are well give my love to dear little Maggie, many kisses
for you. I have written this very badly as the other and for
the same reason. Miss Simmonds does not know the reason my
collars were iron moulded, she says I shall not want the 2 collars,
she sends her very kind regards to you, I hope that Grand•
mamma is better give her my love and lots of kisses. Your
letters make me cry so, your ever
affectionate son
JOHNNY,

Mv VERY DFAR PAPA

I have to tdl you a trouble that makes me very miserable
you know you told me to tell J. Wright and J. Mackenzie that
as they were honourable young gentlemen they would pay me
for my umbrella but I could never get a favourable opportunity
to tell them for 1f the boys heard me tell them about honourable
they would shout after me I cannot exactly explain why and
that would make me more miserable. Jim Wright says that
J. Mackenzie says that he will not pay for it yet because he
does not want to break mto his pocket money yet. I had a nice
journey when I left you 1t was about l past 6 "hen "e got
here. I don't know how many boys 2 we have Albert Ellis is
my chum. I am going mto a book in Latin called Nepos and
1 This was wntten apparentlr 011 the second Wednesday of his second Half

The first page 1s to h1, mother, the letter to his Fathe-r takes page z and
part of page 3.
1
The next letter gives 49 as the number The preceding letter says, ' When
we eot near Bl1'Illlngham John McKenzie took mv umbcrclla and had a ,;cuffle
with James Wright and the:1- broke the handle of'
lxxx FAMILIAR LETTERS
I shall try to get into Astronomy. The drawing master thought
that I should win the prize in this half. Dear Papa I cannot
live away from home. If you want me to die send me to this
place or away from home. I do not say anything of this to
anybody. Lots of kisses. Your affectionate son
JOHNNY.
P.S. I will try to write well.

2
Shireland Hall,
Birmingham,
December 1st, 1859.
}fy DEAR PARE!l<TS,
It is with great pleasure that I inform you that the
vacation will commence on the 16th inst. The school will be
reopened on the 1st of January. Please to send the money for
my travelling expenses as soon as possible and give me all
needful instruction~ for my journey. Please to give my love to
all at home. llopmg that you are quite well,
I remain,
Your affectionate son,
JonN COOK WILSON,

3
The Rectory,
Grasmere.
July 19 [? 1872].
Mv DEAR MAMMA,
J am much obliged to you for enclosing Mr. Case's letter.
I am much amused at your apologising for opening it seeing
that you appear to have read it. I am glad to hear the touching
account of Puss : you don't say whether he gets his lights very
regularly. You needn't envy my staying here. It rains with
few intermissions in bucketfuls. The most useful of all garments
is a mackintosh : moreover I find it an awful grind to keep
these men going: but I don't give them more than 3 hours
a day among them. Consequently I would much prefer to be
like you, at home : than hke me, here. . • . Today was the
festival or ceremony of ' rush-bearing ' at Grasmere Church,
 F AMILi AR LETTERS Ixxxi
All round the churchyard wall inside and I think on the top
stood children with wooden frames covered with flowers and
devices made of flowers and rushes. I don't know what it all
means.•.•
Papa need not put B A on my letters any more :-once was
enough (July 20).

4
Gottingcn June 16
f? 1875].
~l y DEAR MAM MA

I am afraid Papa wont be allowed to retire, but you

mustn't bother much about the arrangements of the house you
may be sent to as you wm so soon be out of it again-by the
middle of September at latest, I hope. I have just written to
Papa. I told him how mistaken his answer to the President
was and how he might repair it I am glad too that you found
yourselves lodging in a pretty country-that is a great comfort
in rase Papa retires. Anyhow I trust that tlus year you will
begin to have a quiet hfe for the first time. We ought to be
very comfortable, 1f \\e arc careful and when the furmture is
paid for we shall be pos1ttvely well oft I hope to kec·p my own
expenses even v. hen at Oxford withm the limits of my Fellow•
slup We have good "eather here, not too hot I and Meggy
have had some good walks. She is a great comfort to me in
very many ways I am very thankful to have such a ~ister.
I don't wonder at the people here being fond of her. She 1s
decidedly very much better m health. I don't trouble myself
now much about my Enghsh acquaintances here. Some of them
I don't care very much about and on the "hole I feel com·
fortabler at home or walking about with Meg than '"ith them.
I forget whether I told you that I had a very kind letter from
Butler, the new Dean of Oriel. I am sure he will be a very
good friend to me hereafter, your aff echonate son
J. C. WILSO~.

2773•1 f
lxxxii FAMILIAR LETTERS

5
Sept. 13 [1882].
Mv DEAR MAMMA,
Y csterday I went with Shadwell to see our Littleworth
estate. I walked about a mile and a half to Bessclsleigh where
Shadwell, dnvmg from Oxford, picked me up. The road, the
one I hope to traverse when I bicycle to Clanfield, is a beautiful
one We had a splendid day and I looked down upon you with
the opera glass from the lull, Clanfield church being visible
among the trees Of this attention you "ere of course uncon•
scious, We found everything going on well The estate is
a beautiful one and extensive, comprising a fine manor house
with a small park (for a park) at Wadley.
(The rest is an account of the fascinating ways of his baby
son, including ' the way the sly dog hints what he wants is
amusing. "Dat Father's cake, poo Father wants it", as though
his own anxiety "as that I should cat my own cake This
transparent hypocrisy 1s a" ful at such tender years.'] Best love
to all, your affectionate son
According to this evening's Echo, the great stronghold Tel·
el-Kebir was taken this morning by our troops

6
August 7th, I 888.
Hotel de la Marine
Arromanches par Bayeux.
Mv DEAR MoTUER,
I was glad to get your letter. I tried today to get
a photograph for you with a peasant woman m the cap of the
country but was unsuccessful. They wear a u•hite cap with
a white starched band bound over it something like this only
the effect, as you will believe, is much prettier [pen-and-ink
drawing]. I have been twice to see the beautiful cathedral of
Bayeux-architecture nth (one of the towers), 12th, 13th and
15th century. In the nave are fine Norman arches, with the
genuine Norman Byzantine decorations, especially the zigzag
round the arches [sketch) but the greater part-of the architecture
is much later especially in the chancel or choir, which looks like
 FAMILIAR LETTERS lxxxiii
our Early English decorated. The central tower of the catbedral
1s I think of 15th century (time of wars of Roses) and is very
beautiful, the tracery of the windows in it (open without glass
like those in tower of St. Mary's, Oxford) is very light and
elegant. Ralph and I went up into tbe lantern (300 feet and
over high). I also saw all the famous Bayeux tapestry, attributed
to Matilda, wife of Wm. Conqueror, with the history of expedi-
tion to England ending in Battle of Hastmgs. Bayeux is
a picturesque medieval town. There arc shops "1th old China
-French mainly &c-1 tried one of them but find it dear and
only bought a few trifles. . . I took Ralph to Bayeux on the
tncyclc The people thought it so charming to see him sitting
behind. Today we made a short expedition and Ralph had
the- pleaE-ure of examining a ship wrecked m February last on
the sands.
7
Hotel de la Marine
Arromanchcs
To I BYWATER, 93 Onslow Sq London. par Bayeux.
Aug. I 6, I 888.
As you may suppose I don't get much time for work here.
I have however got far enough m A.-H. to be able to form an
opinion on his capacity for the non-scholarship part of the
T1m(aeus). It took a ternble long time to make sure about the
mathu~ and music, with his authont1es. As to math. I find he
has but little knowledge of Gk mathuc and writes authonta·
t1vely, as usual, making all kmds of inaccurate statements.
Nearly evcrythmg 1s straight out of Martin and Stallbaum and
tlungs thdt ought to be noted are not because he knows nothing
beyond what is ms1de these sources The music is straight out
of Mart. and Stallb: and no acknowledge' beyond his preface,
even where quite necessary. He doesn't know the subject well
enough to reproduce these authorities properly or to distinguish
between what ought to be given in full and \\ hat can be abbre•
viated. (Often like the schoolboy who, havmg cribbed the
answer, is unable to shew the working or doesn't know it matters.)
As to the philosophy he doesn't understand the elements of his
business. I draw your attention to the pretentious note on
P• I06, which is utter nonsense, as a te11t passage.
f2
fxxxiv FAMILIAR LETTERS

8 Arromanches
Bayeux.
Aug. 24 (Friday) 1888.
Mv DEAR MOTHER,
The above little map shews the places which I visited
lately-Caen, st Lo, Coutances, Avranches. Caen contains fine
specimens of the early Norman Churches, one built it is said by
wm Conqueror and another by his wife. But perhaps one of
the most beautiful was at Coutances. Town set high on hHl,
abt 6 miles from sea. The Cathedral spire serves as a sea mark.
I was at a part of the Sunday morning mass there and the
architectural effect was much enhanced by the imposing cere·
monial. I think of all the cathedrals and churches I have seen
that of Bayeux ha,:; the most beautiful exterior on the whole,
I shall probably go there agam tomorrow on my tricycle and
take Ralph who enJoys 1t all very much .
. . . We went along the coa,;t the other day to sec an isolated
rock standing in the sea, 1 a<'ccssiblc at low tide. A striking
object. I made a sketch of it which I will shew you. Such
things rare on tlus coast. On the Cornwall coast they are
common enough.
9
Dec. 4. I 889.
26 Winchester Rd. Oxford.
Mv DEAR MoTnER,
•.. My article promises to be successful. It appeared
yesterday and the same evening I had a note from a Professor
here congratulatmg me on having completely smashed my
Cambridge opponent. I have had great luck in Kriegspiel (the
Wargame) and lately completely defeated our best player, or at
least one of the best. He had however the great disadvantage
of having (through an accident) httle time to prepare his plans,
whereas I had a long time. We had a game last week in which
a real military man (a retired Colonel) commanded the opposite
side to that in which I served as a subordinate (this time). He
suffered a tremendous defeat, lost a whole battery and a third
of his infantry.... Your affectionate son, J. C. W.
• {La demo1selle de Fontenable<1 1
 FAMILIAR LETTERS lxxxv

'S Gravenhage.
Wednesday, June 29, 189:?.
Mv DEAR MoTHER,
Tomorrow (30th) is dear Meggy's birthday as you re•
minded me in your Jetter. There wdl be fresh flowers on her
grave and fresh remembrances of her in all our hearts. You
would be glad to hear that I had such a good passage. We do
not go to Amsterdam after all until tomorrow (30th) [after all]:
i,,o perhaps I may find a letter from you there. We shall take
Haarlem-the tulip place-on the way, principally to look at
some pictures. Of course the tulips have Jong been over. I shall
be thinking of you and dear Meggy. Mr. Shadwell is as pleasant
a travelling companion as can be imagmed. The hotel is a very
nice one indeed-a sort of picturesquenesc; and · magnificence
about 1t. It is lighted throughout, bedrooms and all, with
electric light. The I Iague is a beautiful place with its canals
J.lld parks We have been to Delft from here, where the famous
pottery used to be made. Yery httle made there now. We have
also been to the famous Dutch watering place called Scheven•
ingen, near here. To it and in many other directions there are
trams, steam, electric and horse power. Everything in Holland
seems as neat and clean and polished as possibly can be. With
bei,,t love to both, yr. affectionate son,
]. C. W.
11
Hawthorn Cottag~
West Malvern.
August 29, 1892.
Mv DEAR MoTHJ:R, •
I wonder "hether you have seen the sad news from
~w1tzerland in the paper. I first saw a paper m which it was
reported that Professor H. NettJeship had died from cold and
exposure on the Alps, he and his guides being overtaken by bad
weather as he was going up Mont Blanc and losmg their way
1n the snow storms. It would have been a terrible thing as he
was just going away from Oxford for a term to be with his son
who is to commence a musical education in Berlin in the autumn,
lxxxvi F AMILi AR LETTERS
However the Pall Mall Gasette said that it was R. L. Nettleship,
the unmarried brother at Balhol. This I expect is the true
version for he usually went to Switzerland in the summer.
I have also a letter from Bodmgton th1s morning who speaks
of R. L. Nettleship and says he met him last year in Switzerland
with his t,vo guides and that he seemed to be undertaking then
expeditions which were rather beyond his strength. You will
remember that he hved with h1s mother an invalid almost
entirely confined to her bed, to wlu>m he showed every attention.
No doubt when he "ent away some of the family would come
and stay with her, probably the wife and children of her painter
son, so she would not be alone. It \\lll be a most terrible blow
to the poor old lady.... The loss "dl be much felt at Balbol.
He was Green's successor and their prmcipal philosophic tutor
there. . . . The men "ere very much attached to him for he
had a certam nobleness of character. One of his most intimate
friends was Mr. Warde Fowler of Lmcoln College, the man who
is such a good musician and slightly deaf. He will be much
missed m the University too He was certamly one of our very
best men. He was so gifted and could write and speak so well.
He was a brilliant scholar and I thmk one of our ablest n1en
m philosophy.

12
Oriel College
Wednesday.
p 18921
Mv or:AR l\loTHl!.R,
... I am to stay m and take care of myself I thought
however I would send you a letter by tram as I can't come
myself. I haven't eaten or drunk anything apparently to upset
me. Perhaps I have been thinkmg too much I am nearly all
right again. I wanted to tell you that I asked the doctor to
visit you a httle often er, because I thought it prevented you
from getting too nervous m the intervals. He wrote in his reply
an encouraging account of you. Although you have felt so
depressed the action of the heart is steadier and also the other
matter which teases you seems to present no disquieting symp·
toms at all. I am so glad and so grateful. I expect you are
 F AHILIAR LETTERS Jxxxvii
depressed by this uncommon bad weather like everybody but
more in proportion as you are so poorly. I have explained to
the Cook about the brawn. He quite understands what you
mean and will be glad to prepare it. Shall I let him do some
for Saturday? To dinner tomorrow we have Mr. Heberden of
Brasenose and sister, Mr. Warde Fowler of Lincoln College,
Mr. and Mrs. Madan, Bodleian Library, Mr. and Mrs. Warner
for whom the party 1s given, MISS Wordsworth, Principal of
Lady Margaret Hall and two others whose names I forget.
With best love, yr: affectionate son,
J.C. W.
The Clarendon Press present to me has come at last. 3 folio
volumes, handsomely bound in half-morocco. You perhaps
remember they gave it me for crit1ci&mg a book for them.

13
[posterior to Nov. 1900.]
When you say you occupy the • common mathematical posi•
tion about the relation of the geometrical diagram to geometrical
thought ' I have a remark to offer which I thmk may interest
you and make you thmk twice about the meaning and value of
the common mathematical position . . . when we mean real
mathematical thought I should say it was pretty fatal to disagree
with the common mathematical position But the 'position of
a mathematician ' on a given question 1s not necessarily • a
mathematical position ' ; he may not be Judging qua mathemU,
nor usmg his mathematical faculty at all, though mathematical
matter m.r1y be mvolvcd. For mstance, he may conduct a process
quite nghtly m mathematics; but there ts a certain reflection
on the process, philosophical, logical, whatever you like, which
1s certainly not mathematical, but the exercise of a totally
different faculty. And it is a great mistake of a man to suppose
that because he is (humanly speaking) mfalltble in the one
process (math0 ' ) he is equally successful m the other. Thus,
e.g., there are people with a perfect knack of doing certain
physical things, but either utterly unable to say how they do
them, or, if trusting to their perfect practice they imagine they
must have the right theory of them, give [sic] a most incorrect
account.
lxxxviii FAMILIAR LETTE~S
We graduate our vocal apparatus with marvellous accuracy
when we smg a melody, but this power does not help us in the
least to say how we do it. That belongs to the physiologist and
psychologist, so far as anything is known of it,
Suppose a gemus among vocalists-a Patti or a Reeves-
presumed to dictate on the question of how it was done, the
scientific investigator of the larynx &c. wouldn't trouble himself
much about that. He would admit that a musical ear was
necessary to conduct the investigation and concede readily that
the great singer or mus1c1an had a better musical ear than
himself, but he couldn't affect, even for courtesy's sake, to
pretend to care about their op1mon on his subject. The' common
position' of musicians would be authoritative to him upon
a. really musical question, but he would be foolish to defer to
their common pos1t10n on the phys1ological question, if they
had one. One can imagmc how Plato might have worked out
the point in a dialogue
Now there is the same kmd of difference between conducting
a math1 process and nfl,ecting (in certain ways) upon the method
and presuppos1t1ons of the math1 process and faculty. A man
may conduct other processes of reasoning too, not mathematical,
and fail utterly m the analysis of what he is doing.
For centuries mathemat1c1ans were content with the view
that geometry had axioms, def111 &c, for its principles and given
premisses, and got the rest by sytlog1sm, a view originated, as
far as records go, by Ar1stotlc In which they shewed they
didn't understand processes which they conducted so rightly
that 1t seems a paradox to say they didn't understand them.
Of course they understood them in one way but not in another.
The common mathematical position here was wrong and it was
reserved for Kant, both a real mathematician and a philosopher,
to free our minds from this illusion.
Now emphatically the question of the use of the figure in
relation to geometrical thought belongs not at all to the sphere
of mathematicians as such. One must know a certam amount
of Geometry to be able to handle the question, just as a man
must have some musical ear to study the larynx &c in relation
to sound, but that is not enough nor even to be a very accom•
plished geometrician. The common mathematical position in
 FAMILIAR LETTERS lxxxix
t/zis-quption must be carefully distinguished from [the] common
mathematical position, when mathematical is used in it~ true
sense : it is really a position or attitude of mathematicians, but
it is not a mathematical position : and therefore, while I abso-
lutely defer to the (common) mathematical positmn, I haven't
the slightest respect for the common position of mathematicians
10 this question. As they were wrong 10 a body about the
syllogism, so the maJority (I suspect) are wrong in the modern
question and indeed the error 1s a consequence of the older one.
So if all the mathematicians took off their coats to me (to parody
Plato) I shouldn't run away.
There is however a natural reason wJ1y mathematicians should
have been misled so far as to thmk this a part of their own
subject and therefore one on which they arc particularly com•
petent to pronounce The speculations, which produced hyper•
bolical and elliptical geometry and the theory of (3 +n)•
dimensional space, seem math 1 and purely mathematical, but
they are not. They mvolve mathematics together with phdo•
sophical or metaphysical reflection on mathematical processes :
for they arc only possible by a theory (and a false one) of the
use of the figure in ordinary Geometry. The mathematician is
thus unconsciously conducting processes \\ luch belong to that
reflection which I have characterised as not an exercise of the
mathematical faculty at all and he, supposing that be is but
acting as a mathematician, 1s proportionately confident and of ten
contemptuous of phdosopluc questioning and doubt. The mathe·
mat1cal investigations here can only be called ' geometrical ', or
relating to some thinkable 'space', through a nustaken theory
of the position of the figure in ordmary geometrical (or Euchdean)
reasoning: a theory wlurh is not a mistake in mathematics but
a mistake in the phJlosophy of mathematics ; but, through the
unfortunate confusio11 of the two processes and faculties m the
same subject, it is presented as a necessary development of
mathematics and therefore 1s bound up mistakenly with the
honour of mathematics as such. We therefore who study
philosophy and presume to form an opinion on what is a phdo·
sophic and not a mathematical problem are put in a false
position .•.• Fortunately I can claim to be what perhaps no
living mathematician can claim, one of the inventors of hyper•
XC FAMILIAR LETTERS
bohc geometry. For I discovered the main features and theorems
of the subject for myself years ago before I knew that mathe-
maticians had done it already and I cannot therefore be treated
exactly as an outsider.
As I have said the people who invented the new Geometry
were mixmg geometrical and philosophic thinking, invading
a sphere of which they knew next to nothmg-unconsciously-
and they she" themselves extraordmanly incompetent in it.
I refer especially to the Pohsh geometricians who started it, who
talked the greatest nonsense when (necessarily) touching on the
metaphysical question which seemed to them mathematical.
These good people take an authoritative tone to such as myself
as if we "ere mterfermg "1th their proper province ; whereas
it is they who have made mroads mto ours and while they
thmk that \\e raw and untrained ones are presuming to judge
in their subject : 1t is they who raw and untrained are presummg
to judge m ours I recall an amusing instance. A man {now
dead) considered one of the greatest hvmg authont1es on hyper-
bolic geometry, with a great assumption of superiority, brought
agamst me m a debate (London Mathematical Society) as
a crushing obJection that to every theorem in the supersensible
geometry corresponded one in the Euclidean ( I knew the
commonplace argument) and gravely argued (as 1s customary)
that this was a strong proof of the validity of hyperbohc geo•
metry 1 The pnnc1plc involved m the argument ts such <1. ridi-
culous fallacy and so easdy seen to be so by anybody accustomed
to ordinary logic that I hardly knew how to preserve the outward
appearance of good behaviour And to have this infanttle non·
sense gravely and authoritatively put before one as an important
piece of instruction. There 1s nothmg so irritating as when
a man who 1s really a great authority m his own subject pro·
nounces on another of which he hardly knows the elements-
not indeed from conceit but simply because, through a confusion,
he thinks it within his own.
1
Infr.i., I 3.20
 FAMILIAR LETTERS xci
14
Fyfi.eld Road
I2
Oxford.
22 Sept. 1901.
Since IV.Tote to you I have been to Glasgow as representative
of the Oxford Philological Society-of whJch I am this year
President-and I saw some very distinguished people, among
them Lord Kelvin. I was mterested to notice that his face
confirmed my Jmpress1on of him as a man devoted in the most
smglehearted way to his subject, mathematical physics, and
utterly unable to see beyond it. It is amusmg to see how the
matha• bow down to him and attend seriously to the extra-
ordinarily naive utterances he dehvers on the philosophical side
of his subject. He seems to me a mere child m such matters.
Have you noticed the worship of the ' mighty atom ' in the
British Assoc•. Rucker (I suppose 1t ts the same man) delivered
once a lecture m Oxford on ' Action at a distance '. It was
mdescr1bably funny-such elementary ignorance about the
nature of the quec;t10n he was dealmg with One could have
reduced him to the most artless contradictions (w1thm his own
lmuts) and I was strongly moved to do so but hadn't quite the
courage. It was some years ago. If it happened agam I think
I skould have the courage now I was glad also to see and
speak to Professor J. B. Mayor of Cambridge, who is one of the
most learned and accurate men of his generation and I think
one of the most honest.
So from the end of June till now I have worked at the Greek
tacticians hke-the best bishop-you know what I mean. . . .
I wanted to get my work into <the) shape of a treatise before
the vac• is out I daresay I shan't do that especially as the
continuous application JS begmnmg to take it out of me. But
I hope certainly to publish my results rec1.sonably soon. The
thing was really too tempting to neglect. I hope my friends
won't despair of me when they hear that the promised i,ray(l)y,j
is not ready for I think scholars will think the work I turned
aside for worth domg.... I haven't published my address 1 on
the nature of one's conviction of the existence of God: and
I have generally thought it best not to lend the MS. because
1 Infra., H 505-8:z,
xcii F AMILi AR LETTERS
it is not adequate to what I actually said. It 's too much of
the nature of notes for an address. If you are very anxious to
see it however perhaps I had better let you have it-that is
when you have really time to read it-you would be more
sympathetic than a stranger and perhaps know what I was
driving at even when only indicated. But then I am afraid it
would be necessary for me to make additions to the last part
or even rewrite 1t to make 1t clear, for being pressed much for
time I couldn't write out at all properly the very end of it but
that didn't matter for my address, for it "'as the most important
thing I had to say and I was m no danger of forgetting it. But
I haven't seen it for a long time and I don't know how much
I should have to do at it. If it was fairly much I could with
the present preoccupation of my ' pint of brams ' (would it were
at least a quart I) hardly put myself back in the right position
just now at least. As a matter of fact I should bke you to see
the argument (though I should mfimtely prefer that you had
heard it-' the letter kdleth ') and so later on I shall sec if
r couldn't make the l.ittcr part fairly intclhgible to a reader.
I trust everything 1s prospermg with you. I am very glad to
say that the place I was at in Yorkslure (Ravenscar) suited my
wife better than most places we've tried. For which I am very
thankful. •.. We were counting up to-day the number of
eminent specialists she has been to and found them nine at least.
But I am beginnmg to hope that she really will get much better.
She can tricycle a httlc, rt suits her better than walking.
With kind regards, yours truly.

15
si Edward'~
Islip, Oxon.
t May 1902.
It 1s perfectly splendid. I can't tell you how glad I am.
Your cup of happiness is getting pretty full and I congratulate
you with all my heart. I wish I had a daughter myself, I should
have adored her : and 1t 's a blessing to a boy to have a sister.
My own sister whom I lost when she was 30 was the light and
blessing of our household. She was sweet and adorable beyond
words.
 FAMILIAR LETTERS xciii
My mother says it is so nice to have a little girl first and
I am to tell you that. But I expect these things are relative.,
that most mothers prefer a son, if the idea of preference is
allowable, and most fathers (apart from the artificial prejudice
in favour of a son and heir to maintain the family succession)
prefer a daughter.
I always thought it so touching that Victor Hugo should
present his bttle grand daughter with the words 'c'est ma petite
grand'fille que j'adore '. I can so enter mto his feelings ...•
How very good of you to write to me at once about your
good fortune. I am afraid, by the by, that you will have been
pursued by a letter from me about defaulting Cyc-hsts. I didn't
know you were off. To-day (Friday) I am in lsLIP and sleep
also, as my mother's condition has begun to give cause for
anxiety. With very kind regards to you both and hearty good
wishes.
Yours truly.
I'll tell you about my interview with old Poste when I sec
you. He was calm and cheerful. Did I tell you I tried diameter
myself at first as ao8Eia-a ypaµ.µ.~. Thanks for date of Butcher's
solution Bywater with whom I spenl yesterday evening is
rather epns of my solution and wants it published 1

16
Recluse Lodge
Freshwater, I.W.
8 Aug. 1902.
We are truly sorry to hear that you have lost that noble dog 2
It h, well that he did not have a longer 11lncss-poor fellow....
I do trust the httle daughter is getting on better and that you
have no anxiety about her. I didn't deduct anything from the
account you sent for I do not know what the ' At Home ' cost.
We can settle that next term. We are resting here, only my
friend is very fond of discussions and, as he is an invalid and
doesn't often get such a chance, I have to humour him. But
it is tiring sometimes and I shall try to keep the thing within
1 Plato, .lleno, S6 F, f, of Plulolog,•, XX\'ui, pp 222--40 (1903)
• llafe Bernard, one of the noble<it of a. noble breed. Buried, by kind
pernw1s1on of His Grace the Duke of Bedford, on his estate at Aspley Guise.
xctv FAMILIAR LETTERS
limits. You will miss Rafe sadly, but it is well that the little
daughter had arrived and with her and with music and gardening
(which I think delightful) you and Mrs. - - should have
a pleasant summer. I have brought here my Vergil and my
In Memoriam with Bradley's commentary, which I shall greatly
enjoy. I had already read some of it. The weather has been
deplorable but our friends are charming. My friend's daughter
is a good musician and plays the 'cello wdl. I may confess
I never felt so old, but I trust the feeling will pass--it is
depressing. With very kind regards from us to you both,
yours truly.

17
To BERNARD BosANQUET.
22 June 1903 .
. . . The reason I did not publish '\\ as that tho' Boole is the
fountam there- has been a good deal of development of the thing
m Germany and I thought one must attack the thing as a whole.
Consequently I got Schroder's book- a frightful tlung to tackle
and there I stuck because I was obliged at the time to turn to
something else and I have never had the time since to read
Schroder. I contented myself with giving a public lecture 1 in
which I endeavoured to show. (1) that merely as a calculus it
went on a wrong principle, not recognizing that a symbolic
calculus must be developed from the particular matter it relates
to i (whereas 2) they are forcing the matter into algebraic
symbolism, as 1f the latter were the only possible ; (2) I shewed
that as a cakulus it was involved m a contradiction which
mathematicians as such must admit to be such-violating the
principle of algebraic calculus as such ; (3) I shewed this latter
not merely verbal or formal, but that by it the identity of the
extension of any class could be proved with that of any other,
e.g. if A and B are class symbols, as they are, then the equation
A=B can be established K•hatei,er A and B represent; (4) I
pointed out some minor fallacies ; (5) taking the calculus on 1ts
own merits (as (intended) to solve certain kinds of problems)
I can verify my first point, about fitting the form and the
matter, by producing a calculus of my own of exceedingly simple
1
See ff 371-400 1 ' when ', original
 FAMILIAR LETTERS XCV

character, which is suggested by the nature of the problem to

be solved. It solves 1 with ease and simplicity all the problems
attacked by the symbolic logicians whose calculus is in com•
parison very cumbrous ; e.g. Boole's methods require, for such
complicated problems as Dodgson used to devise, an extra•
ordinary amount of working and quantities of equations, whereas
my method does the business quite easily These complicated
problems do not appear m Boole or his followers at all. Elimina-
tion from 12 or 20 (e g) premisses of a number of terms.
Dodgson used to make them and I have many in the unpublished
2nd part of his book, of which he sent me proofs Dodgson had
a method resembling Boole's in some essentials. but of course
he had (the) advantage of making the problem. Dodgson was
astonished, as he freely said, at the way m which I solved all
such problems and was able to shew him that he had in them
superfluous premisses, or premisses one or other of which might
be alternatively superfluous-things he had not found out him•
self. I never shewed him the method, for he died Just as he
was busied in the 2nd part of the book and I hadn't had a good
opportunity. This 5th part I was unable to develop in my
lectures (two only). Edgeworth, who is a disciple of Boole and
all such, came to me after (the) lectures and said I had ' put
pegs into Boole', but said also an important use had been made
of Boole's calculus in application to probability .•• and I had
better consider that carefully r ..•
2
] I hadn't at the time
done this. A Cambridge friend of mine, v.ho knew Boole's
calculus well, told me he had never read this part and so, as he
1s a very thorough man, I supposed it of no particular importance.
However I read it and soon found that the main theorems were
actually false-false that is from a mathematical point of view.
I suppose as mathematicians as such don't mterest themselves
generally in symbolic logic, these mistakes which are demon•
strable beyond a doubt and vitiate all Boole's account of
probability have escaped notice. There's absolutely no doubt
of the error, but as a confirmation I may say that I sent one
of Boole's solutions to a distinguished mathematician (F. R. S.)
and asked him how a certain result could be got. I didn't send
1 The onginal has ' they solve •.
• A. digresS1on on the sad drowmng of two Oxford men.
xcvi FAMILIAR LETTERS
the working but the problem and its Boolian solution. He
replied that he didn't see how it could be got-that is to say
a certain element in the solution. Now I knew it couldn't be
got for the problem presented no data for determining the
clement in question, which therefore must remain indeterminate.
Part of Boole's error, which is a most ingenuous one, consists
in JUSt giving a determinate form to what in the theory of
mathematical probability couldn't have (precise determination).
The curiouc; thing is that his theory sometimes brings a true
result and I had therefore to discover the reason of this and
determine when it could happen. It took me some time but
I did eventually find the reason and have got to the bottom of
the whole matter.
This agatn I didn't publi--11 because I thought 1t ought to go
with the cr1t1cicim of the Boohan calculus in general ; which,
for the reason givrn, had to wait and other things came pouring
m upon me I tlunk now, perhapci, 1l would have been well to
have published the rnttc1c;m hy 1tsdf, 1t might have prevented
such mistakc<i as R Ruc;c;('}l's and ccrt.unly I c,hould have
destroyed any 1<lca of .Boole's infalhbihty by bhe\\mg the mis•
takes m probab1hty "luch all mathcmallc1ans would admit.
The more so as I hear from Edgeworth that recently considerable
use has been made by some people of Boole's probabihty
methods. There is another tremendous fallacy in the subject,
the belief that the forms dealt with m ' symbolic logic ' are
general forms of all wference \\ hatever. They arc only forms
of syllogistic inference and one c.in shew that mathematical
inference as such is not syllogistic and that the supposed reduc•
tion of (e.g) F.uchdean proof to c;yllog1stic form 1~ a fallacy.
I have postponed reading Dedekmd and Cantor's ll1e11genlehre
for the same reason for which I postponed Schroder (as also
Peano). In general the time I am able to give to quasi•
mathematical speculation has been engrossed for the most part
lately by a renewed effort as you know to find the kind of
contradiction in morc-than-3-dimensional space which would
make me master of the situation, because it would convince the
rank and file of mathematicians and thus they would at least
not suppose the philosophic criticism, by which I intended any•
how to attack, somehow wrong.
 FAMILIAR LETTERS xcvii
I will take this opportunity of telling you that during the
last three terms I have simply thought mcessantly on the sub-
Ject, so much so that I felt I was in danger of • breaking • my
mind for a time in 1t. It seemed constantly as though I had
succeeded this time and of course the result of success would
be so very important that it seemed worth while to strain every
effort. The argument I sent you attracted some very con-
siderable mathematicians, who studied 1t with great care, as
well as some other arguments of a different kind which I did
not send you. Later on some very good men disagreed with the
others in at least not feeling secure about some of my results.
The net result seemed to be that my mathematical friends had
never analysed certam notions about movement m curves and
so disagreed in their view of them and, when I pomted out
certain consequences of their view, changed it. Now, for my
purpose, a hne of mathematical argument is of httle or no use
unless 1t 1s qmte convincmg to all, even ordmary, mathe-
mat1c1ans, and so far therefore I consider my effort after all
a failure. For 1t won't do to risk anythmg upon a doubtful
mathematical demonstration. Even 1f one's phtlosoph1cal crib•
rism accompanying 1t was sound 1t would be suspected because
of the suspicion of the mathematics This 1s what really hap·
pened to Lotze, who of course gave a great deal of time and
thought to these things And his example 1s a warning to
smaller people.
Of course I am sufficiently disappointed, as the attraction my
view had for some very critical and clever mathematicians made
me thmk I had done it at last and could begm now to put my
views mto prmt, more especially as I have almost worn myself
out these 3 terms by mcessant thmking, because it always
seemed that I could get the thmg into a shape which would
convince all and (I) contmually found again it came back to
some point of the same disputable character. I have resolved
therefore now to make myself no more anxious about this, but
develop a criticism of the (more-than)-3-d1mens1on (geometry)
directly on the books in which the theory 1s set forth I shall
111ev1tably from time to time return to the cherished project,
but I shall not devote myself to 1t cxclus1vely, but rather think
of 1t occas1onally when there's nothing else to do--m the train
2 773 I g
xcviii FAMILIAR LETTERS
or in the waiting-room at a railway station-and take my chance
of a lucky thought or inspiration. However this prudent resolve
is not all I have to shew, (1) I have made some new develop•
ments in' queer' geometry, which will shew that I am qualified
as an original thinker in the mathematical theory itself and so
I shall get more heed to my criticism of it all ; (2) I think
I have got light at last on the relation of motion on the curve
to motion on the tangent-the fundamental pomt on which my
mathematicians waver and I think my result explains their
different views and why they should waver. This I might
communicate to you sometime when I am not writing a volu-
minous letter. I am absolutely under contract, so to say, to
write on certain questions relating to Plato and Aristotle this
vacation; but I have been asked by the philosophical coterie
you wot of, to deliver a public lecture at least as soon as may
be on these modern vagaries connected with symbolic logic
{including Dedekind and such hke). I may perhaps therefore
take Dedek10d & Co. to the Schwarzwald m the vacation, but
as I must make the Greek a first charge I doubt whether I shall
be ready with my public lecture by next term
I am send10g you now herewith my criticism of the ' class of
classes ' fallacy As to Royce's theory, the map in a map, this
again I haven't read but have had told me by such competent
people as J. A Smith. Prima facie it seems a mere form of
stating the familiar truth that, if we take any two magmtudes,
no matter how much greater one is than the other, we can always
divide them m the same proportion. The position of a point
on the one map can always be represented by a proportion of
certain magmtudes (coordinates and their relation to height and
breadth of map) and the 'corresponding' point in the other
map by the same proportion. And this really reduces to the
simple truth that any magmtude can be divided 10 the same
proportion as any other. One might just as well take a line
and another which 1s a part of it, and consider a set of pomts
determined 10 the larger and their correspondents on the smaller
-Just as well as take a surface (the outer map) and a surface
within it I should be glad therefore to see your own criticism,
dealing with the thing in detail. I daresay you could without
difficulty send me the essentials of it. I would in return send
 FAMILIAR LETTERS xcix
you a criticism of the fallacy reproduced in B. Russell's article
concerning
I 2 3 4 S 6 . . ad inf.
2 4 6 8 IO 12 • ad mf.
[Here follows an unsympathetic criticism of Prof. Moore and
of Mr. B. Russell ]
I see I forgot to say that I found myself a certain difficulty
m the argument sent to you which none of my mathematical
critics discovered. I regret that I have unaccountably mislaid
your kmd letter, no doubt 1t wdl turn up, and so am answermg
from memory, yours truly.

18
Fyfield Road,
I2
Oxford
30 June 1903.
I do hope you are better of that touch of sun. For your very
kmd letter very many thanks. I was so glad to hear the field
day on Monday was such a success. By a happy chance the
problem was the same that the Colonel himself set upon Satur-
day : so our men would have the advantage of that day's
experience on the same ground and with same obJect. I con·
gratulate you also on managmg so well with the faithful few.
Kindly remember me to all who care for the message and you
may go down as low as beutenants for the sake of B--.
Mrs. C. W. is engaged m the arduous toil of getting me into
the right clothes to see the Major 1 married this morning. Wir
smd beide sonst etwas medergedruckt.

19
Kurhotel-Schbnwald.
Schonwald be1 Triberg, Baden.
29 July 1903.
I cannot remember whether I wrote to you or not, but give
you the benefit of the doubt. This place 1s really capital and
I could recommend it if I were sure the rainy weather we have
been having was exceptional and not normal. The hotel is
quiet, comfortable and quite moderate in charges. Actual
1 Now the Right Rev the Lord Bishop of St Albans
g2
 C FAMILIAR LETTERS
locality not striking, but good centre and fine views within
about a mile. Between two centres of the clock industry,
Furtwangen and Tr1berg. Tr1berg is very fine in its scenery.
It doesn't suit my wife as well as I could wish. She misses the
tricycle very much, as I feared she would. I hired her a bath
chair from Fre1burg, which has extended her radius a good deal.
Yesterday I pushed her up a road with frightful gradient (45°
as Dodgson might say) in a pelting storm of rain {which over-
took us) for a frightful way. Thought I should be stiff all over
this morning but am not stiff at all.

20
Hannover.
Sat. 29 Aug. [1903].
. . . I do hope you are better Perhaps what you want is
a fortnight {say) of mere physical exercise and enjoyment of
open air, carefully avoiding any ph1losoph1cal or Academic work.
I believe this absolute laying aside of work is often imperative
m such a condition as you seem to be (m). It 's difficult, I dare-
say, but 1t pays We wmd up our holtday to-night by going
to Fidelio and expect a great treat We start home to-morrow.
We wonder what the surprise 1s and hope it may be that you
and Mrs. - - are coming to hve m or near Oxford. Kmd
regards from us to both.

21
Oxford.
7 Sept. 1903.
We have just been having a hohtary tea m our little garden
which looks perfectly lovely. Oxford 1s now m the throes of
St. Giles' Fair My wife was noticing the other day how
many suicides there are m the papers. For myself I am suffering
from such acute nervous depression that I might think it
alarming if I had not had it before now. We spent our last
fortnight m Hannover : it 's a place which always has a most
dismal effect on me (tho' such a beautiful town) .... I suspect
the strange weather has helped, it was simply stifling at first
jn Oxford I have got what I suppose to be the male equivalent
for hystena-somet1mes badly. It makes me understand the
 FAMILIAR LETTERS ci
queer things that sane people sometimes surprise one with, for
I expect they get into this extraordinary nervous phase. I may
however myself count on being preserved from them, because
I am conscious of the situation. It helps me also to understand
hysteria in general. I am doubtful sometimes whether there is
anything properly insane about it ; whether 1t 1s not that one
feels and realises things clearly and acutely, for instance, one's
utter powerlessness and insignificance, also the loss and absence
of friends and relatives,-too clearly for one's proper balance
which is cond1t1oned by a certain callousness But when it's
pretty bad I do thmk it 1s something hke a conscious msamty.
The feeling is, I thmk, the mental analogue of the misery {even
agony) of sea-sickness. It's worse than any mere physical
illness and the effect of 1t I dread is mabihty to do any real
thmkmg or reading. . . . I even seem to dread tlus mechanical
work coming to an end, when I must face real work. As
a matter of fact I have now-being the next thing promised-
to wnte out my article on the points you wot of m the trireme,
not a great mental task, the material all ready in the shape of
notes, but I shrmk from it. Also I have to keep up my study
of Dedekind and Cantor, which I broke Wl'll during my stay m
the Schwarzwald (that was splendid) but scarcely was able to
touch m Hannover ...
My wife has been also very depressed m Oxford (of course
I can't confide my real cond1t1on to her ...) ...

22
Oxford.
17 Sept. 1903 .
. . . I am glad to say I am distinctly better. I feel the worst
is over and I am able to do some work.... One feels so much
coming back to Oxford and not being able to go and recount
one's experiences of travel to one's parents It 's my tirst
summer without one of them and till last Easter 1 I had both
of them My wife feels our loneliness in this respect very much,
tho' of course 1t doesn't concern her anything like so nearly....
1 [vu:. Easter IQO:Z]
cii FAMILIAR LETTERS

23
[June 1905]
I wish you had been at my paper on Friday or that I had
been at your Inspection.... Fortunately I. B. was there him-
self. He was much pleased with my paper and said he (had)
never seen the point put properly before. We made a night of
it together after in the Randolph.... It wouldn't do for them
of Ascalon to know the rifts m the philosophic lute and don't
repeat to the philosophers either

24
12 Fyfield Road,
Oxford.
I I May 1904.
l trust that you have had a refreshing Easter holiday, and
are now able to devote yourself to your studies agam I want
to report progress on some of the subjects on which we have
corresponded. I am always moved by what Locke calls ' the
most pressing uneasiness ' and have had to deviate more than
once from my study of Dedekind. However, one of my devia-
tions has brought me to it again m, I hope, a useful way.
The ' Greek M us1c ' earned me off first, for the tlung seemed
to work out so well and to be confirmed by further examination
of ancient classical and post-classical authorities. I read a paper
on it here which was attended by the Provost of Oriel and the
Principal of Brasenose · and was greatly confirmed by the
result. I told the Provost beforehand the new passage in
Aristotle. I understand he came thinking he could dispose of
it, though how he could I can't imagine, as it certainly is obviously
fatal to his own view. However, I revised all the classical
evidence, without the Aristotle, and I think satisfied my audience
that the modes differed both in interval and pitch. I then
showed that the Aristotelian passage confirmed this, and from
it alone one could get the kmd of theory I advocated. Monro
hardly made a crit1C1sm. Heberden thinks he realized the
Aristotelian passage couldn't be got over, and indeed I had led
up to 1t by showing that there were other important passages
which couldn't be got over and had been misinterpreted by the
Provost and others. It was a feature of my theory that one
 FAMILIAR LETTERS ciii
recovered a quite simple and natural interpretation of such
places. I was told afterwards, that the Provost had been
greatly impressed. Well I started to get this ready for the
press, when I read an article upon Eth. I. vi, in connection
with the Idea Number theory by one R. G. Bury in the February
Classical Review, which turned out a still more pressing uneasi-
ness for it gave me a great opportunity of saying something on
the d.rruµ.PA.f/T'O& ,lp,6µ.ol, upon the Idea-Number theory, and on
Plato's philosophy of Mathematics-things (some of them) which
I have given in lectures I had by me for years, but had no
occasion to publish. The article in question showed that there
was really need to say something. So after patenting the
'Music' by a short abstract of my paper in the Classical Review
(it will appear either in May or June) I gave myself to the new
work This took up a good part of my energies in the Easter
Vacation. I have written and indeed printed, for the proofs
are finally corrected, a long article entitled auvp.f3>..f/T'O& a.p,8µ.of.
for the Class. Review. This will probably, or a first instalment
of it, appear m the June Class. Review. I used the opportunity
to give two ' Pubhc Lectures ' m Oxford on the subject, and
am glad to know that I gamed the adherence to my views on
the Platomc and Aristotchan questions involved (mamly in
Aristotle, Metaphysics, Plato, Republic; lhavoia, et id genus omne)
of the kmd of people here one wishes to convince. The reason
why I may suppose this to interest you at all is that d.rrvµ.~. ap.
led me to give a deliverance on Dedekind which I have put
in a foot-note. It relates not to the book 'Was sind und
was sollen die Zahlen ' but to ' Stetigkeit und die irrationale
Zahlen '. I am clear now that the attempt to introduce con-
tinuity into ' number ' itself comes from a fairly ghastly mistake
as to the meaning of ' the number two ', ' the number three '
etc. etc. that I am sure about, but I think that I have made
it very probable that it was from a very accurate understanding
of what these expressions do mean, that Plato or a Platonist
got the conception of d.rr11µ.fiA.11T'ot ap,6µ.ol, This I have put
forward in my article, and have chaffed the modern metaphysico-
mathematician for a mare's nest constructed from fallacies the
Platonist saw through long ago. I have through the comparison
of ' parallel versions ' in the Metaphysics shown, I think, that
civ FAMILIAR LETTERS
one writer in the Metaphysics particularly wished to dissociate
Plato's name from the later' Idea-Number' theory. This same
writer does associate Plato with the o.rrvfJ,~"'1/To, aµ,8µ.ol, and it'
is one of my points that it arises out of the earlier Platonism
and has nothing to do with the later Idea-Number theory as
such. I have also given an explanation of the Ta ,,.ua'6 for
Mathematics, which I dare bet you will approve, and I think
I have shown 1t a great mistake of interpretation to think any
such doctrine m the Republic. This I thmk inter alia disposes
of some of the great nonsense H. Jackson talks about Plato,
but I had no occasion to name him, as the essentials of the view
I traverse are a mistake at least as old as Ueberweg.
I will send you a 'deprmt' of the whole thmg when I get
some. I might have got on to my • music ' m the vacation but
was interrupted by two claims on me. Lady Welby insisted on
my g1vmg a set opinion on her book, and I had to read it nearly
through I knew already enough of it to sec 1t wouldn't do,
despite her cleverness, and would gladly have declined the task
but she wouldn't let me off. I broke the result gently to her
and had a very charming reply. Agam a friend bringing out
a rather good analysis of the Ethics for pass men {the pass
part) submitted 1t to me and I spent a good shce of my time
in writing and argufymg about that. So you see I haven't been
able to help myself (' bin wirkhch sehr unschuldig ', what you
once kmdly quoted) One absolutely can't avoid these thmgs,
and 1t ts only now that I am gomg to begin to get the Greek
Modes ready for press Fortunately now it is simply a matter
of straightforward work and can be got through almost mechani-
cally. I don't thmk I can mterlard Dedekmd with 1t (much
as I wish) because I have got with help of notes of pupils to
rewrite my commentary on the Posterior Analytics (delivered
about one and a half years ago) as some of the tutors are good
enough to desire very much that this should be done, and I must
use these notes while I can keep them. Meanwhile I itch to go
for the other Dedekind books, I really believe I can do for 1t,
but I know I must get the ' music ' out of the way first. I saw
with mterest your review of Moore's Ethics m ' Mind '. As
I haven't read Moore and am httle hkely to, I have no right
to opmion but you seemed to me to say Just the nght thing,
 FAMILIAR LETTERS CV
and what is perhaps a good deal harder to do, just in the right
way. I asked a very capable Oxford tutor if he had seen it
and what he thought of it, and I found he was very pleased
with it and thought it remarkable that you could manage it at
all, considering the amount of your lecturing work.
I have not been afraid of writing about my own work, not
only because I know you are so sympathetic, but because I want
you to understand I have not let go the investigation which
you think [it] rather a duty for me, though I haven't been able
to give it first attention, and also by explaining what has
hindered me, to show that I was really justified in leaving it
for a time.
You yourself have no need I expect for any such ' apologia',
you manage to go straight forward and finish off important
pieces of work.
I do trust you are all right and vigorous. For my own part
I seldom had such a good Easter vacation.
I was at a lovely Sussex village standing on the Downs four
miles from Arundel and within easy ride of a very dear friend,
an old Oriel pupil, a most interesting young artist with such
splendid ideals. He lives on other side of Downs near Amberley
at a place called Bury-if you happen to know the neighbour-
hood. My friend's name 1s Wethered. Ile has now a picture
m the New Gallery-if you go tell me what you think of 1t,
I haven't seen 1t. Some of his work 1s very successful (I think)
and certamly 1s all very imaginative--some of 1t I don't like,
but he's uncompromising and sticks to his prmc1ples. He 's gone
on landscape but can draw figures wonderfully. He made
a pastel of me, which I think 1s very fine vigorous drawing,
and the critics here agree. There 's a good show of portraits
from the Colleges here now. You may see an account of it m
the Burlington Magazine by Blakaston.
cvi FAMnIAR LETTERS

25 22 May 1904.
12 Fyfield Road,
Oxford.
(1} o.rruµ.{JJ1.11ror means 'not add1ble' as is abundantly proved
by Aristotle's d1scuss1on in the Metaphysics. I suspect you
were thinking of o.rrvp.p.Erpor.
(ii) You say on the universal of number ' It is clear ... that
twoness cannot be added to twoness and make fourness ; but
twoness must involve some sort of structural relation to fourness,
I should have thought '
I don't follow the ' but'. What it introduces 1s no objection
whatever to what precedes it. Don't you rather mean that you
are accustomed (or perhaps we are all when not reflecting) to
think of the structural relation of two to four as the construc-
tion of Four by the addition of two Twos, and that as this
clearly wont do (for you agree I have shown that) since by Four
the number Four 1s meant, a new question arises What is the
structural relation of Two to Four, if any ? How ought their
relation to be represented ? What 1s there m the relation in
virtue of which three comes m the serial order between two and
four ? If I have got the real meamng of your difficulty clear,
I think the answer 1s also clear.
I have pomted out that the proposition ' two and two make
four ' means that any particular two added to any particular
two makes a particular four, or two particulars added to two
make four particulars. Two means always two of, twoness 1s
twoness of (cf. Aristotle's defimtion of o.pLfJfJ,or as 1rAijfJor
µEµup71µl1 1ov-which 1s very suggestive, as all his say10gs about
number and magnitude are).
Now that any particular two added to another makes a par-
ticular four hes m the nature of Twoness and Fourness. This
seems to me the simple answer to the question about structural
relation I imagine proposed. You may say that if one group
of elements has twoness and the other also twoness, they con-
stitute a group which has fourness, but I think the more natural
language the best. So agam as to the serial order :-The serial
order 2, 31 4, corresponds simply to the fact that the nature of
Twoness and Threeness and Fourness 1s such that a particular
 FAMILIAR LETTERS cvii
three consists of a particular two with a particular one added
to it, and that to get a particular four another particular one
must be added to the particular three.
Thus it takes two steps of the same kind to produce a four
by addition to a two, only one such step to produce a three by
addition to a two and only one such to produce a four by addition
to a three. These productions accordingly form a series : and
in the serial order the production of the particular three comes
'between '-in the sense explained-the production of the
particular four and the two (or the product10n of the two).
This is the simple meaning of the order of ' the series of
natural numbers '. The order 2, 3, 4 does not mean that the
numbers 3 and 4 arc formed successively (one before the other)
by the addition of the number one (=oneness) to the number
2( = twoness), and of the number one again to the number 3,
which is mere nonsense, but that a particular three and a par-
ticular four are produced successively by the addition of a
particular one to a particular two and the addition to them
again of a particular one.
May one not compress the principle which explains all these
things into this statement?-' Just as squareness has not got
four equal sides etc. but is the having of four equal sides, so
twoness is not made up of two umts, but means the being made
up of two umts.' I think the key to all the fallacies and con-
fusion can be got from this.
By the way, you would, I suppose, agree with my criticism
of Dedekind, for his theory of Continuity depends precisely on
the fallacy that the number four 1s a magniJude differing from
the number three as another magnitude by one unit.
(ui) As to BLavoLa-1 mtended to write but think it best to
take one thmg at a time, so will put off till I hear from you
about the above. I will only say that I feel pretty sure that
when you have reconsidered what I have said with Plato's text
before you at leisure you will agree with me I can't pretend
that I think there can be really two opinions. The objection
you make seems so directly dealt with in what I submitted to
you that I think it might be merely enough to ask you to look
at my argument agam when I can send you a deprint.
I am sorry you didn't agree at once (the criticism I should
cviii FAMILIAR LETTERS
have rather expected was that it was not at all new to you)
but am not exactly surprised : for my· experience has shown
me that it 1s often more difficult to convmce acute people about
something very simple than on a difficult pomt, especially when
they have got biased by some current view. You quote our
friend Stout as apparently agreeing with you in doubting my
interpretation. I might also quote my Oxford audience 1£ we
are to count heads (as' heads'), Some of them, of co\:lrse, have
given very special attention to the question, and I felt sure they
were (some of them) in favour of some such interpretation as
Ueberweg's. There was a moment when I saw I had captured
them-one notices these things m an attentive audience-and
it was when I challenged any one to say whether they could
make anything of aui,-..ETp05' avT~ and TfTpO:ywvov aw~assigned
m the most defimte way by Plato as the object of lluivo10-
except the ll1'a1 of these geometrical elements.
And I saw also the pomt fully taken and appreciated that
the origin of the fallacious mterprctatlon was the confusion of
lluivo,a '\\Ith the faculty of mathematical mvestlgat10n-whereas
p.a871p.a.Tuc~ hke a,aAo:nK~ wac:; a process 'outside' the four
farultu.•s I found out after people were convinced.
I will only ask you for the moment to try a method which
I find it useful to fall back upon when I can't make headway
with what I beheve the simple, obvious, and necessary meaning
of a passage in Plato or Aristotle.
I ask those who, I believe, are mamtammg an unnatural and
artificial view to try the experiment of puttmg what they believe
the meanmg of the text into the Greek wluch they think would
best and most clearly convey it. The comparison of the result
with the text tells its own story , and 1t becomes evident that
Plato, humanly speaking, couldn't possibly express himself as
he has done if he had mtended what was attributed to him-
especially when as m the present case the expression he actually
uses would convey naturally the very opposite of the supposed
meaning.
I pointed it out in my lecture that when we satisfied ourselves
of the absolutely necessary meaning of Plato's making the objects
of a,&vo,a the ' originals of which the phenomena arc copies ' and
the awcl TtTpdywvo11 etc., that point of mterpretat1on was irre•
fragably fixed and we must not mind what difficulties PJato got
 F A'MILIAR LETTERS cix
himself into afterwards by his distinction e.g. of a,,h,o,a and
voiis, It would be a great mistake to alter the necessary
meanmg of the Greek because of any such considerations-the
confusion if any would be Plato's look-out. But I contended,
and satisfied people, there was no such confusion and Plato said
something perfectly intelligible
The fact is the distinction Plato makes between the• obJect
of a,&vota and voii~ 1s one which in essence appears within
mathematics, and no doubt was suggested to him by mathe-
matics 1t is simple and not difficult and nobody really following
the text and undcrstandmg the thing 10 question in mathe-
matics, and not confused by the Aristotelian passage, could have
two minds about 1t.
I will explain this later and put the answer to your objection
again and more at large than in the printed article when I next
write, unless you prefer merely to wait and take another look
at the article.
Joachim has been givmg me an amusmg account of a debate
with Stout at St Andrews. From what he said I have a sus-
p1c1on I should have been m essentials with Stout Some things
Joachim said seemed right, but I can't help thmkmg that 1£ he
had put them properly they would be found m agreement with
Stout's view. But I may be qmte out of it.
Yours truly,
J. CooK WILSON.
PS. I hear Stout 1s not brmgmg his little boy when he comes
to Oxford, that seems rather mean.

26
13 November 1904.
Oxford
' An anxious mqu1rer ' this Sabbath mornmg who, though
a young student (a BA) and not an 'expert' in mathematics,
though he knows a decent modicum, 1s convmced of the non•
sensical character of the construction of space out of an aggregate
of points and has been attacking this effut1t1on as represented
by B. Russell (m a College Society), has brought back to my
mind a matter which I ought to have wntten to you about
long ago
You will remember my writmg to you about that (tn my
ex FAMILIAR LETTERS
opinion) rather foolish fallacy of B. Russell's about a class being
a member of itself. I showed by an artifice how the game
could be played with any class, and emphasised the fact that
it was a mere fallacy of language. You weren't altogether
happy, because you didn't, I think, altogether see what I was
driving at. But later I put the theory on what I must think
a sound basis 10 writ10g to Stout. There remained, however,
one difficulty 10 Stout's m10d 1 which roughly came to his
accepting a very fallacious verbal mistake of Russell's, in fact
that a class as many could be dist10guished from a class as one
and (ultimately) represented as a member of itself.
These things provoke me for, though they depend on fallacies
of mere language which I am obliged to think very childish,
they are insidious if people aren't accustomed to analyse forms
of expression-and so many, even logicians, are not I therefore
sat down to a determined effort to dnve the hobgoblin out of
Stout's mind, and took great pa10s to get all clear. Stout had
fought determinedly but this final effort conv10ced him abso-
lutely (that was long vacation 1903, I thmk) I should bke you
to know this in case you had thought there was anything in
the thmg · and 1t drives me wild that these mere muddles of
language which annoyed Plato and Anstotlc so much in the
petty and trivially clever debaters of their day, should reappear
at (all), after all philosophy has done, d.Od actually be mistaken
for somethmg the least worth having.
Russell's fallacy may be refuted m a hne or two, thus·-
A class 1s a unified manifold of elements A member of
a class is one of the elements so unified To say, then, that a
class is a member of itself is to say that a unified manifold of
elements is an element in the same unified manifold of elements.
(The peculiar way in which the members (elements) of a class
are unified doesn't matter, one only need use the part of the
conception which 1s sufficient to show the fallacy ) This being
so the fallacy can only be due to some verbal error and the
remainder of one's work 1s to show what the verbal error is,
which has made it seem plausible. I think I did that, but for
mere refutation('s) sake the above 1s enough.
Yours truly,
J. COOK WILSON.
 FAMILIAR LETTERS cxi

'Zl
rTo a schoobnaster 1n reply to the words ' 1-rners generally find a difficulty
1n Tngonometry from the fact that a revolvtng hne only baa a BlgD 1n 4 definite
pos1tions and 1s regarded as having no sign 1n all other positions There
seems no way (1n elementary Tngonometry) out of this ddliculty except
co1111ent1011 ' ]
(? 10 Feb. 1905.J
(Such an answer would never do, for the following reasons :-)
Nothing in the way of truth can ever be settled by' convention',
though all the elementary books speak as if the use of + and
- for direction was, as you said, JUSt a convention. All con•
vention can do is to settle what symbols sha)) be used for a given
idea, we can mvent the symbol and ' agree ' (convention) to use
1t, we cannot mvent the idea. For instance we may have the
convention that + shall represent division That is absolutely
an convention can do. This as a general principle 1s quite
self-evident, but we can bring it to the reductio ad absurdum
test, in a particular mstance, easily. For instance, if we tell
the boy the use of + and - for two opposite directions is
arbitrary and a convention, he might reasonably say: 'Well,
1f + doesn't itself mean direction, what right have you to use
1t for d1rect1on ? It 1s a sign of a certam meaning or idea,
viz. addition, how can you make 1t the sign of another meaning
or idea without confusion ? Either you are usmg 1t as having
a different meaning from add1t1on or you are not. In the first
case, why on earth don't you use a different sign as the meaning
1s different? In the second case (i e. 1f you don't use it as
having a different meaning) you are obviously makmg a mis-
take.' And he might involve you in the reauctio ad absurdum
by saying ' Good I If 1t 1s mere convention, we might just as
well use the sign of multiplication for one direction and the sign
of d1vis1on for the opposite direction ( x and +) ' There could
be no answer to this upon the convention principle and you will
find at once, if you try and in the ordinary co-ordmate system
substitute x for + as a sign of direction and + for - , that
it won't work the least bit, but you will get quite ridiculous
results. Try it : I have. I am not saying what the true answer
is but only pointing out the futility of the convention answer,
which we all ahke have been taught.
cxii FAMILIAR LETTERS
May I add to the support of my imaginary schoolboy that,
after he says + must mean one thing or another, he might go
on to remark that he actually finds in trigonometrical examples
before htm + used in the same example both for sign of direc•
tion and sign of addition. Heaven save you for the present
from the imaginary boy I
Note agam that m co-ordmate geometry you may see no
difficulty in an equation hke y = mx + c. You will substitute
for y, as representing a co-ordinate, -b, say, and calculate the
-b-c
value - - for x as a co-ordinate. But what happens when
m
you have a product xy,. To begm with, this represents no
co-ordinate with direction at all, but a rectangle. How can it
hJ.ve a sign at all ? And suppose we give x the value -a, and
y the value -b · then xy = ( -a) ( -b). How can we say that
the two minus signs of direction can produce a plus sign. How
could they be multiplied " And are we to say that the resulting
+ sign (1f we take the algebraic rule) 1s a sign of direction?
(To consider now, so-called imaginary quantities). The mis-
take m people like De Morgan ts to suppose they have to look
for an •interpretation' of ✓-1, where it does not (or could not)
represent an 1mpossib1hty, misled by the analogy (not rightly
understood) of those problems which give - I (or some number
with - sign) simply m the answer and thus shew the question
of the problem has to be answered m the negative, while the
same - quantity m another reference would give a positive
answer and express a possibility. For instance, 1£ you ask what
will be a man's gain under certain condrttons and working the
equation find 1t == - 3£1 the simple meaning is that he gains
nothing at all and further that his failure to reach gain is
measured by 3£. If now you ask what he loses, you get the
answer 3£. The first answer shows it 1s impossible he should
gain and how far he falls : the second shows that he actually
loses and how much.
Now we want ✓ -=--i" explained, so far as it represents an
1mposs1bihty. Any interpretation.· of it as representing some-
thing possible and founded on a really different calculus and
symbolism 1s quite irrelevant and a false use of the loss analogy.
The true parallel is that the ✓ -=-i answer, hke the - 1 answer,
 FAMILIAR LETTERS cxiii
gives a definite measure of the impossibility. In the profit
question, -3 shews, so to say, how far the transaction fails of
being possibly gain. The ✓--:::j_ answer shews also a measure
of how far the thing fails of possibihty. For in an instance (of)
it, a difference like a• -b1, which ought to be positive to give
possibility, and the magnitude of its negative value, 1.e. of the
value under the root m ✓a• - b1 = ✓ -c•
shews how far it is
off possibibty. In the case of a circle for instance 1t shews how
far a circle comes off reaching a given hne which 1t does not
reach, as can be easily seen from a figure. If r is the radius
and d the distance of the line from the centre, then the y
,a
co-ordinate of (the) point of intersection = ✓ -d2• And r• -d,1
gives the exact measure of the 1mposs1b1lity, when it is negative,
and accordingly gives a distmct1on between one circle and
another (e g. of same centre)-the very (definite) measure of the
1mposs1b1hty, varymg from one circle to another, actually seems
to define and detcrmme each circle. It 1s quite simple d bemg
constant, the value m question of the 'degree' (shall I say) of
the impossibility determmes the correspondmg value of r and
so the circle 1
28
[Oxford]
20 June, 1905.
You asked me what Bywater thought of my views on
Kd8apcr,,. 2 I have been to see him to-day with a result both
amusing and unexpected. As l anticipated he does not think
of giving up the ' purgation ' interpretation of the passage and
apparently does not dispute my analysis of the metaphor of
purgation so applied I should have thoiight that nothmg
remained but to admit Aristotle was wrong, and badly. But
Bywater seriously holds that Anstotle 1s right because this 1s
all that can be said to excuse tragic representation on the stage.
Plato was right, he says, it 1s wrong to have tragic representation
at all. From the point of view of the moralist and the politician
this titillation of the emotions ought not to be permitted (these
are Bywater's exact words), Plato's view bemg right, the only
possible excuse for performance at all is that given by Aristotle,
says Bywater, to let off, purge off these undesirable feelings.
• Cf t 116, p 268 • Seep lxx1, O P S, under 1905.
2 773 l h
cxiv FAMILIAR LETTERS
From further conversation, after this remarkable deliverance,
I gathered Bywater thought that tragedy only gratified (from
the ideal point of view) a morbid liking for seeing and hearing
horrors. ldeaJly it shouldn't be gratified at all (so Plato), but,
if at all, only by way of harmless outlet and purgation (Aristotle}.
I feel the wirked mirthfulness of the slave in Plautus (or is
it Terence?) who' mixed things'. I have driven some into bad
translations or fantastic creations quite away from anything in
the Poetics, or into that confusion of metaphors I prophesied,
and Bywater I have driven into the arms of Plato. He certainly
developed his adhesion to Plato's view of the drama after my
criticism for, when I first told him my difficulties, such a solution
had not occurred to him. I thmk I may leave the people who
disagree with me to deal with one another. I have got views
written down for me by various people and J. A. S. has ex·
pounded viva voce the positive side of his view (he agrees, you
know, with me on the negative side) to me lucidly and ener•
getlcally. I have not had time to study thoroughly the written
papers, I have been too unavoidably busy with other things:
but everything I have yet seen or heard, whether by itself, or
setting opposites against one another, confirms me in the view
I advocated m my paper and m my impression that there was
a deplorable confusion of ideas on the subject You said you
had heard some things which seemed to shake my position.
I wanted to ask you what they were but haven't had the chance.
Don't write, I'll wait till I see you.
N.B. Miss Gwyer, for whom I wrote the essay out of which
I developed my paper, is very properly delighted at the rumpus.
Yours
29
5 Granvdle Place,
Portman Sq.
w.
3 Jan 1906.
The Morals of Marcus makes a good play, 1£ well acted and
it was well acted. Carlotta was played wonderfully well by
Miss Carlisle. . . . I had read the novel but had forgotten even
that I had read it. Of course it came back to me as the piece
developed. If I had known, I hardly think I should have gone.
 F AMILi AR LETTERS CXV

The book produced a painful impression on me. The writer has

committed what is to me the ' unpardonable sin ' against human
nature. Mrs. H. Ward has done the same thing in the Marriage
of William Ashe. . • • The novelist first creates a lovable
character perhaps with some eccentricities (in order to palliate
the offence he, the writer, intends to commit) and then makes
the character do something horrid (something simply disgustingly
wicked when looked at honestly) in order to make a sort of
affecting tragedy. What is done would be quite impossible for
any real person with the character the novelist has depicted up
to the false and criminal step. It would be impossible for any
person however stupid, 1f they had any feeling, any ordinary
sense of duty, whereas the heroine (I observe this class of writer
doesn't take this sort of liberty with a man's character) has
been represented as having a lovable nature-very sensitive to
obligation and at least with tender and refined feeling I On the
other hand it would be impossible for any person without any
decent feeling if they had a httle ordmary common sense. And
yet Kitty m Mrs. Ward's novel ts very clever and Carlotta isn't
a mere fool. In short to make the tragedy at all possible the
person would have to be an idiot both morally and intellectually.
The trick of shockmg the reader is mere vile vulgarity and
always makes me furious. Human nature 1s poor enough but
1t is a great deal bf'tter than that. It 1s therefore to my mind
profanity as well as vulgarity In the case of the present novel
or play the writer has not only been vulgar but stupid . his
own representation has made what happens particularly impos-
sible. One very marked feature of Carlotta's 1s the innocence
and naivete with which she tells everything to ' Sir Marcus ', her
flirtation with the grocer's boy for instance. Obviously the first
thing she would have done would have been to tell Sir Marcus
how kind, how very kind Lady Mamwarmg and Pasquale had
been, that the former had explained to her she couldn't marry
Sir Marcus (and so he couldn't save her from the Effendi that
way) and P. had been so very kmd as to promise to take her
safely away out of reach of Hamed. The author has mdeed
written himself down an ass as well as a vulgar profaner of
human nature. You will never find anything approaching this
in any great master, it 's modern decadence and love of paradox.
h2
~·
Georp Eliot, who harrows one's feelinp trofitably, i'nlild K'nl'
have made aucb a blunder. Just think of her WOll1eD. 'hle
v~dict of the common public on such 'tlings is that they aire
• unnatural ' and the verdict is right, only too merciful. The
cleverness of the writer in making us first love his character ia
no excuse, because he simply prostitutes himself (I mean it) to
produce an effect. . . .
30
[Oxford]
Feb. 9 [1900].
- tells me you were maintaining that there were duties
which couldn't be claimed as rights by others and that R.
opposed tais. You may like to know that this is my view,
communicated often to pupils when I was a tutor a propos of
the question of the difference between acts of justice and
generosity. The view is the direct development of the theory
of the right of the individual mamtained in my Republic lectures.
Duties I distinguish into those whose performance can be claimed
as a right by others and those where it cannot. The former
constitute Justice. Their performance is not according to [the]
use of language called ' mentor10us ', their violation e contra is
a crime and Justifies compulsion and punishment (see Rep.
lectures). The performance of the other is accounted meritorious
and is the sphere of generosity as opposed to justice. It 's
a demerit not to perform them but not a crime. (Not a good
augury for R.'s book) Kind regards to you and Mrs. --.
My conduct has not been good enough for me to appear before
my ' dual control '.
31
New College.
19 Feb. 1906.
I am touched by the Joy of my dual control at my • decora•
tion ' 1 and I hope 1t will propitiate them for a time at least.
I hoped that hidden away in the University Intelligence of The
Timts it would not be noticed, for no people doing their duty
in term would be likely to see 1t; and I think few, if any,
philosophers have seen it except yourselves. But when people
1 Hon. LL.D. at St. Andrews.
J ( \\ 11'.on J>laymg "1th l\lr Prichard•., -.on-.
 1' AMILIAR LETTERS cxvii
take in •The Times as their domestic paper and then lie abed
unemployed one 1s not safe, for
Satan finds some m1sch1ef still
For idle hands to do.
I wouldn't be so affected as to ask you to hush it up, but I don't
want attention drawn to it unnecessarily. It really has not the
value such an honour ought properly to have and mean [sic],
that one's achievements are appreciated m the outer world It
only means that Bosanquet and Stout have a high opm1on of
me , that I knew already, and I value 1t far more than the
degree. The only thmg of real va!ue from the outside of oneself
is JUSt the estimate of such people and of one's colleagues here.
As to the degree 1ti,elf, 1t 1s, as I have !>aid to Bosanquet, the
first instance, as far as I know, where the affection of my friends
has Jobbed me mto a d1gmty and as such it delights me im-
mensely I was sounded about 1t something hke a year ago
and I said I thought d1stmctly such tlungs were only appropriate
for people who had pubh!>hed books of a certain (considerable)
thickness and were not m my \\ay at all Seriously I do thmk
that people who brave publicity are those to whom such honours
are really appropriate Otherwise you encourage the merely
local reputation which 1s the danger of the greater Umvers1t1es.
(Have I not already offered 1t as a sop to my momtors ?)
The children were delightful the other day You don't get
to know what children really are till you play with them and
see them play with one another. Some of my bttle friends do
not stand the test as well as I should hke. But I can say
without flattery that yours do. I did not know Charley very
well before but I thmk him a delightful child after seeing as
much of him as I did yesterday
With very kmd regards to both of you, yrs truly.

PS. I simply can't come to teas this term-at least not yet.
I am bothered about my lectures next term bacause 1t is exactly
the time when I ought to letture on a number of difficulties left
over in the Poster1.or Analytics and I am not ready. I have
continually yielded to the ' more pressing uneasiness ' and so
put it off. I can't risk advertismg 1t because 1t 1s quite possible
cxviii FAMILIAR LETTERS
I may not find time m the Easter vacation As I have to go
to Scotland for the degree on April 3d I may take my wife to
see Edinburgh and stay a short time with a friend in Glasgow,
who has sent a pressing invitation, and this doesn't look hke
time for woz'k. I am very unhappy about it, because the
Summer term is JUSt the one in which I ought to give this
lerture
32
[Oxford, ? June 06.]
. . . I am grateful both to you on whom far more. work
devolved than upon the men and to the men themselves who
certainly worked under great difficulties and drawbacks. 1 Well,
give them a jolly good dmner on my part on Monday and say
how much I wished I could be among them. They are stunning
chaps even if they haven't won....
He 1s perfectly and exceptionally capable of doing the work
but apparently has some sort of kmk in ]us mmd and just
' bucks ' as some of these clever people will. What a lot of
good it would do such creatures (God's creatures after all) to
come into camp-just a httle ragging, for which we could acquit
everybody all nght, would be most salutary. Figure me gomg
to-morrow and trying to coax this fractious ass of a gemus with
p.uA.1xto,r i1r,fcraw when I should hke to - -....
Really I am a person of little readiness of mmd, the Col.
presented me with a cucumber (or lone) as field-marshal's staff
at luncheon. What an ass I was not to put 1t in my pocket I
He would, in common honour, have had to pay and my wife
would have been most pleased with me, as economical Hausfrau,
for cucumbers are dearish this year.
33
Redfern House, Bosham.
[19 Apnl ? 1905.]
Utmam me una tandem re nunquam im1tatus esses, dico
,ca,coypa</>(av. Litterae tuae nov1ss1mae v1x legi possunt. You
might have kept (the) Bradley letter till term. 1 There is much
• Extract from a long letter about a competit,on (Wolseley Cup) at Bisley.
1
A letter from me on Mr A C Bradley's cntiC1Sm of Shakespeare Wilson
thought at this time that he had improved on Mr Bradley's theory of the
l'ragedies Letter to A C.F., 27 m 05 (notprmted) Tlusletterisoutof order
 FAMILIAR LETTERS cxix
force in what you object but there are reasons why I did not take
the obvious and simple interpretation : and as I am sworn foe
to interpretations which are too subtle and ingenious you may
suppose I have special reasons. These are best not detailed in
a letter, we will talk them over. You perhaps do not remember
enough that Hamlet is an inhabitant of the same kind of world
as Ophelia, when you talk of a 'sterner' age than ours. Valde
dolemus quod uxor tua, nobis con1unct1ssima, Justo tardius con-
valescit. My article on Odyssey XXIV has appeared m Clam.cal
Review. 1 Prmters made awful hash of my article in Mind. 1
Stout is penitent-not my fault.

34
Bad Kreuth,
Bavaria.
13 July [19()6) .
. . . My wife is disgusted with Kreuth. We began with sultry
thundery weather which is for her the most trymg thmg of all.
Then suddenly mist and ram and cold November weather.
To-day we are Just sh1vermg about. . . . Fortunately for me
there 's a mce young fellow here named Franz who was about
IO years ago learning English m poor Abbott's house m Oxford.
He went an expedition with me the other day mvolving a climb
to the saddle of the Plauberg (otherwise Blauberg). The climb
turned out extremely steep and m one place where we had to
scramble over a scarp of naked rock the track practically dis-
appeared. There was a glorious crop of gentians at the top.
I read m the guide book (afterwards) the path was called ' uber
den Fels ' and was ' nur fur schwindelfreie ' Fortunately I am
' schwindelfrei ' m a most desirable degree. There can't be
much the matter with me for my young companion was much
more pumped by the ascent and begged me to stop and give
him breathing space. He remarked that I was evidently ' very
well trained ' (he meant ' m good training '} and expressed his
surprise, when we got back, to my sister-in-law on my walking
and climbing powers. I attach importance to this incident
1
The Classical Review, xix, pp 144-7.
• On the strength of tins and of internal evidence I have included the
artlcle 1n Mind, N. S,, No 54, Signed W., 1n Wilson's works
CXX F AKILIAR LETTERS
because through my other ' Beschwerden ' I had begun to fear
that I had entered a period of permanently diminished vigour,
which considering the work I have before me had been causing
me some anxiety .1
Today I have got the proofs of my Monro memoir from
Berlin.
Kindest regards from us both to you both.

35
New College.
28 Feb. 1907.
I understand the difficulty. It looks as though one ought to
try to do some recrmting in some of the Colleges. We can talk
about that when we meet. We might have a short pow-wow
in certain selected Colleges. It's a pity the Cyclists are not
better supported and it's rather hard upon you because in the
last two years the work of the Cyclists has been on a much
higher level than it ever has reached before . . If you like
I could draw up with you a short statement of the nature of
the work of the Cyclists which could be seht to captains of
compames or other persons mfluential m the matter of volun-
teering. It 1s clear to me that the difficulties of recruiting have
increased a good deal from the causes you mention smce I left
the Corps (except that cause which 1s among what you mention,
courteously, but really 1t isn't that I). Of course I was not
finding fault with you, I am not so blackly ungrateful.

36
Monday.
[20 May, 1907.]
The letter . . . proved to be of the d1stressmg kind I feared.
. . . I should like to come and talk to you about 1t, for though
no very decisive step need be taken at the moment I feel as
though I should break down if I couldn't confide the main
things to somebody....
If you happened to come here and there was some one else
in the room you could say you wanted to talk to me about
something and when we were alone you could say first some-
' Part of this letter 2s pnnted 1n Part V.
 FAMILIAR LETTERS cxm
thing about the Cyclists or any other subject. If I don't see
you before I should like to see you Wednesday, I suppose my
best chance to see you (or Tuesday) would be after 4, but it
may be very difficult for me to get away. My wife will want
me to tea and be afraid of my being out too late. Would there
be any chance of finding you in about 2.30 ?
Bywater's pubhc lecture is 2.15 {I think) to-day. 1 I am afraid
no chance of my going. By irony of fate, another of my letters
was to say I was recommended for election to the British
Academy : but I am in such trouble I can't appreciate it.
Please do not say a word about the Brit. Acad. to an;yone
(except of course Mrs.--).
Many, many thanks.

37
Villa Taunus, Ems.
4 June {and after I) 1907.
. At Cologne where we spent the night. . . I thought
it best to go and have a restful time in the Cathedral and was
glad I did. I found it had lost none of the charm it had for
me when I was younger. It is wonderful what an effect such
architecture can have upon one, it is so calming. Such lofty
vaultmg as that of the aisle and choir at Cologne, with such
light and graceful forms, is subhme without being oppressive,
it is the subhm1ty which encourages and lifts one up in a manner
sympathetically to itself. A service was just concludmg without
music, but the responses of the congregation had a very beautiful
effect : the vast echoing spaces transmuted the sound till it was
no longer like human voices but bke the breakmg of waves
heard at a distance. One was far enough from the congregation
(Defense de circuler pendant l'office I) to get this effect. I wanted
to renew my acquamtance with the choir chapels, partly to see
some beautifully carved woodwork figures in high relief on (the)
altar-piece and partly, as I thought, to see some Flemish pictures
by Memling and Kranach. But my memory played me false in
both respects. I suspect I was thmking of some other church
seen since-possibly at Bruges, for though I found a magnificent
carved-wood altar-piece, 1t was not in the choir but on the east
• Tiu Er1Um111" Pf'otJWtJ&1f1110Jt of Greelt ""d ii, prec-ursors (pubhahed 1908),
cxxii FAMILIAR LETTERS
wall of the S. transept and the figures were not what I had in
memory, really far better. [Here follows praise of the altar-piece
in Radley School Chapel.] ... I doubt whether you would find
anythmg better in any continental church and it will transport
you m imagination to Flemish and German medieval churches.
. . . In the ' Three Holy Kmgs ' Chapel there were some very
beautiful figures of the 14th century (the above is dated 1520)
but they were ordmary detached images (the three' kmgs' and
I thmk two other images) and apparently they had been put
into the altar-piece m modern times-the altar-piece itself bemg
by a modern Dutchman (and excellent). The aforesaid very
large altar-piece-sort of big triptych or tetraptych-in the tran•
sept came out of a destroyed church.
As for the pa10tings there was no Memling or Kranach but
to compensate, the famous picture of the Adoration of the Magi
(not 10 their chapel however) by Stephan Lochner, 15th century.
It is triptychal, the colourmg is very beautiful and the faces
and figures are quite out of medieval stiffness and almost
modern. It is in wonderful preservation and 1s said to be the
picture mentioned by Albrecht Durer 10 his journal of his visit
to the Netherlands. How thankful one must be for these great
painters and one is so glad to thmk the names of the quite early
people are not forgotten (Lochner was from Switzerland but had
settled m Cologne) and also that the names of the first architect
of Cologne cathedral and of his successors are also preserved.
It is delightful too that the orig10al sketch of the towers and
W. fa1;ade has been recovered (they say part of it was in Darm-
stadt and the rest in Paris) and is to be seen in the Cathedral
June 9th. I mtended to send this off days ago but my time
is taken up by my medical exercises and restmg after them and
writing necessary letters (I wrote 7 postcards yesterday and
I write every day to Mrs. Wilson, she is often so lonely especially
as her knee prevents her from walking). 1 I must also count the
music which is my principal resource. I listen to that from
4 to 5.30 every day and generally take a good walk between
that and supper. After supper there is another concert, generally
the best in the day.... I chose a train which gave me a little
more than two hours in Cassel, for I wanted to renew my
• He had left Mrs Wilson at GOtbngen.
 FAMILIAR LETTERS cxxiii
acquaintance with the picture gallery there. 1 It contains some
fine Rembrandts. One of these is the celebrated picture of his
wife Saskia. . . . The picture is fine enough but what he saw in
Saskia I don't know. Even he has not succeeded in making her
look interesting.... Much more interesting are several portraits
of himself, one very fine. The most celebrated is, I believe, a large
picture 'Jacob blessing the sons of Joseph. It 1s a fine thing but
rather monotonous m colour for Rembrandt. There are also
some splendid portraits of various people by him. I was happy
to find quite a nice landscape, The shore of Scheveningen by
Adrian van der Velde. That's the man of whom I have an
original landscape drawing over the mantelpiece in my study.
His metier was to put the cattle in other people's pictures,
e.g. Hobbema's, and so one doesn't often see a whole picture by
him. There is also a splendid and well-known Van Dyck, The
painter Snyders and his wife . . . also some gems by Metso and
Ter Borch.... Cassel itself is worth a visit ; it is ideally situated
and has a delightful old-world look, Just an ideal of the old·
fashioned capital of a German principality. There is at least
one fine Titian and I was much struck by a small cattle piece
by Paul Potter, it seems a sort of preparation for the famous
Bull. 11 There 1s the same wondedul rendering of the glossy hides
of the cows and of their horns but still 1t is far behind the
famous bull, the cows are more hke single studies beautifully
detailed just put on the same canvas, whereas the Bull is a fine
composition. By the way there 1s a curious picture of Rem•
brandt's, celebrated and I daresay you know it but I had
forgotten it. It is called The Woodcutter's family and 1s supposed
to be a Holy Family. It seems an absolutely Dutch interior,
very delightful are the mother and child ; the father is dimly
seen a little distance off, perhaps through a door into a shed,
chopping wood. . . . I came by chance on quite a beautiful
picture by Antony More : perhaps you remember that one of
his was about the finest in one (the first I thmk) of the exhibi•
tions of College pictures m the schools. It would delight the
heart of Woods, and doubtless has delighted it for he is sure to
• He had VlSlted it in 187 s m company with Miss Margaret Wilson and
probably later.
• He had adnured tlus when v1SJtJng Holland with Dr. C. L. Shadwell in 189:z.
CX.XJV F AHILIAR LETTERS
know the Cassel gallery. The amiable custodian, who was with
me, told me there were other Mores but I hadn't time to look
them up. The line hither passes Marburg . . when I first
went to Gottmgen 1874 I went that way-ils situation is most
romantic, the bmldmgs delightfully quamt and medieval. The
town seems to pour down the semicircular side of a high hill
into the Lahn. The configuration of the hill occasions the most
picturesque situation for the buildmgs. I always think it one
of the sights of the world and compare it to Innsbruck, which
also is built high and low on a hill-but it is rather at the foot
of a mountam. I fancy Marburg must be an excellent place to
philosophize m. Leyden I always thmk another, Leyden so
wonderfully still and remote from the world. You can imagine
people there wrapped up m classical scholarship or philosophy,
as mdeed its great men were, a place from which physical
science would be bamshed as mere ' stinks ' Gottmgen was,
but Gottmgen, I feel, fuit. It has grown tremendously. The
suburb where my wife 1s m a nursmg home was utterly unrecog-
msable, fine handsome houses grossstaedhsch [sic] m fact and
the idylhc countnfiedness of Gottmgen seems gone for ever.
It seems only appropriate that the prmc1pal philosopher should
be a psychophysic1st. Nevertheless the old town w1thm its
'Wall '(the latter, which used to surround 1t and was a favourite
walk of mme, 1s levelled now m many places but the site of it
remains as a sort of boulevard}, not so very much altered, is
now a kmd of kernel of old-fashioned med1eval-hke bu1ldmgs,
surrounded by large suburbs with modern bmldings. It makes
a sad 1mpress1on on one, the simple charm of the old place is
gone, the time when 1t seemed all Umversity. By the way on
one house outside the 'Wall' I saw a tablet with the names of
Brahms and Joachim. Ask Joachim if that was h1s uncle.
I certamly don't remember to have seen 1t there ' m my time '.
This place (Ems) 1s delightful. ... The position, on the Lahn,
is just wonder£ul, steep rocky hills go sheer up from the river
which has delightful wmdmgs. These hills are well wooded and
there are delightful walks to the tops of them .... One of my
greatest delights 1s the fine band. It plays to us from 7 to 8.30
wlule we are drinkmg our waters, but I don't get much out of
it then. There's a fine concert in the afternoon 4 to 5.30 when
 FAMILIAR LETTERS CXXV

we have our coffec. They do play well. One feels it is so

accurate. I always think it greatly contributes to the pleasure
of music to see the quaint or beautiful shape of the instruments.
A fine trumpet polished like gold, the quaint bass viols and the
'cellos with their delightful golden brown colour. The Dutch
painters felt this you can see, for they delighted in beautiful
paintmg of the lute and other instruments (Ter Borch, Metsu,
Van der Meer). Apropos there is a Lady playing a lute by Ter
Borch in the Cassel gallery, and a Metsu where, besides a lute
played by a lady, there is a fine coloured 'cello lymg on a table,
which the painter evidently delighted to pamt. 1 That is the
weak side of the modern piano. The Grand is simply an ugly
instrument, unless you happen to see right on to the key-board,
which is its redeemmg feature. Do you know I thmk the earlier
people felt that, and that is why the spmet, clavichord &c. were
so often highly decorated, mamly with pictures, even inside the
hd so that when it was raised to give more sound (as m the modern
grand) there was a delight to the eye. I believe they felt the
artistic difficulty and that was their way of getting over it.
I distinctly remember this brought out in Dutch pictures, where
the painters represent the spinet or harpsichord, decorated in
this manner. I remember one quite clearly (but not the name
of the painter) in which the open hd of the spmet or harpsichord
shews the picture on the inside of it 2 The music here is a great
solace It prevents me from feeling too lonely. We had a
Wagner Abend the other day, it was splendid. 0£ all the Wagner
I know I hke the Tannhauscr music the best and especially the
overture with the wonderful Pilgrim melody Don't you think
that it is characteristic of Wagner's style to have the melody
principally m the tenor or else baritone. I believe much of his
peculiar charm is due to this. But I must stop now so that
I may stop somewhere. Kmdest regards to both of you.
J CooK WILSON.
PS. Once more a thousand thanks for all your kmdness to
me and care of me
1 Cf. Metsu, The music lesson, Nat Gall , No 838

• Perhaps the wnter was thmkmg of Vermeer's A lady at Iha 111rginals,

Nat. Gall, No 1383
cxxvi FAMILIAR LETTERS
38
18 June 1907.
Villa Taunus, Ems.
Your letter was a Wohlthat (which means, for otherwise you
are sure to misconstrue, 'a refreshment'). Yesterday I wanted
you badly, remembering our visit to Beverley. As I came here
I saw two Romanesque churches, from the railway, in most
picturesque positions and I visited them both yesterday after•
noon. . . . One is the cathedral of Limburg on the Lahn. The
description of it ought to make your mouth water. It is built
on the top of a rocky eminence descending in sheer walls of rock
to the river, which curves beautifully at the foot of it. The
town itself climbs up the other side of the hill to the cathedral.
The town is itself most picturesque, all sorts of crooked and
narrow Gassen, such quaint houses and buildings. The whole
reminded me of Albrecht Durer's beautiful landscape back-
grounds, the quaint buildmgs he loved, and JUSt this ptle of
architecture-church or castle on the top of a sharply defined
rock. It is an extraordmary emotion to find what one 1s only
accustomed to in pictures and of a long past age, and so belongs
to the dream side of one's hfe, actually still existing m all its
quaintness and beauty. One wonders if the people about belong
to the past too m their ways and dialect. And it was with
a sort of satisfaction that I learned that the people in the village
similarly dominated by the second Romanesque church were
almost exclusively Catholics How wicked the artistic sense can
make one I must tell you a story of old Bodmgton of Lincoln
a propos of this. He came once to stay with us when we were
at a place on the coast of Picardy We were in the cathedral
in Boulogne where there is a famous statue of the Virgin, which
came down from heaven on to a boat (I think). There happened
to be a pilgrimage from some of the neighbouring villages to
see it · mainly women of course, on some sort of festival. It
was a pretty and a touching sight. Old ' Bod ' (bilious Bod
was his Oxford title} said. 'and we have lost all this by our
beastly Protestantism.' I won't guarantee ' beastly ', but it was
either beastly or ' words to that effect '. He said it quite
seriously without a wink or a smile but looking quite cross.
You see he wasn't master of his artistic feeling. Yet he wasn't
 FAMILIAR LETTERS cxxvii
even a ritualist at home. I didn't dare to chaff him and I might
have died of the inward bleeding of suppressed chuckles. Of
course it meant a want of the sense of humour (aren't you
profoundly thankful you have a sense of humour? It ought
to be accounted a virtue, for it saves the virtuous from being
prigs}.
The Limburg cathedral, date about 1218-42, begun just before
Cologne, is architecturally most fascinating ; it is built in
a transition style from Romanesque to Gothic. The main out•
Imes of the buildmg are Romanesque, the usual pair of high
W. towers, two small towers at the end of each transept with
the usual Romanesque gables and diamond-shaped roofs to them,
and a central conical tower [illustration], which I guessed to be
an imperfect restoration and I learned afterwards that this
tower had fallen in ;-seven towers m all. It was curious to
see these seven towers {besides their Romanesque wmdows}
decorated with pomted Gothic arcading. The effect at a distance
was fine, the Romanesque quite dommatmg; but near at hand
it had the effect of degradmg the severity of the Romanesque
into mere quamtness and the Gothic shewed itself the truly
beautiful as opposed to the quaint. Inside, the church was all
thoroughly transitional, capitals sometimes purely Byzantine,
sometimes of the type of the Early English decorated capital
[illustration] and the same shape as opposed to the squatter
Romanesque. Here and there were striking features of Early
Enghsh such as this [1llustrabon]. Then there was a triforium,
a thing unknown I suppose to Romanesque proper, rather
a shallow one with Gothic arcadmg . instead of the fiat expanse
of painted wall below the clerestory of a Romanesque church.
The main pillars showed the same transition to Gothic as you
may observe in the four pillars at the Junction of transepts,
nave and choir in Christ Church. 1 Finally, everything was
painted in the Romanesque fashion. This one saw was the
mistake, it did not smt the graceful Gothic forms introduced,
even if the colour had been better, but made them look tawdry,
at least deprived them of force. It was well one had Christ
Church in mind (more hke this than Beverley) for one saw how
1 VIZ the Cathedral Church in Oxford. The wnter as not quite correct about

the style and date

cxxv:iii FAMILIAR LETTERS
really beautiful the combination would have been, if left in pure
uncoloured stone. The whole thing was full of lessons and you
will realise I could hardly get out of the place. One could see
the spirit of architecture struggling into the beauty of 13th-
century Gothic, and of course in the light of later achievement
it seems strange (tho' it really is not) that once the improvement
began it hadn't got the Gothic outline everywhere as at Beverley.
By the way, in the arcading of the triforium the shafts of the
pillars appeared to be black marble bke Beverley and the nave
of GJoucester, so that feature appears. Under the triforium what
I may call a gallery for want of remembering the technical name
(Emporen is the German) runs all round the church as in Notre
Dame and the shafts of the middle pillars of the corresponding
arches were black, painted the custodian said and I verified
[11lustration]. In Chnst Church there 1s a beautiful combmation
of Romanesque (note especially the very beautiful and severe
capitals m the choir) and Early English but it appears there
to be a combination proper, due to different times of building,
in different styles, but m Limburg 1t is transition proper, the
elements being combined from the first. One lesson it teaches
is the great beauty of the fully developed tendency in Beverley
and the great beauty of the analogous combmation in Christ
Church, for mteresting as Limburg 1s, it is far inferior in real
beauty to these other two. The choir had an apse and a deambu-
latory-the latter very bke Malvern Abbey-church. It was
severely beautiful from the outside, with a pretty sort of gallery
near the roof (arcaded) and two strong, simple, unadorned flymg
buttresses I tore myself away from this to go to Dietkirchen,
a village about It miles off direct and 2 along river bank (Lahn).
This was a similar picturesque situation A rocky eminence
with village clustering about it-one side sheer precipitous walls
of rock and the summit crowned by a pure Romanesque church
-no transepts and only two towers, the usual two western,
those very high with a severe and noble effect, dominating the
landscape in a striking way because Just here the valley widens
into a plam, so that the isolated rock with the high buddmg
upon it 1s thrown into extraordinanly bold relief. You will
remember this sort of thing in Durer's (and others') pictures,
where it looks exaggerated. The church, supposed to date from
 FAMILIAR LETTERS cxxix
the 11th century (and it looks the part}, is very solidly built.
The aisles are vaulted so that the ' galleries ' above them have
stone floors. Painted all over inside, qmte right, but if the
colours had only been better chosen it would have looked
infimtely better (this is the place where the people are all
Catholics}. The two towers are united near the top, high up, by
a covered bridge, at quite a fearsome height. The diamonds of
the roof of one of them were sharper than that of the other-
it's not merely my bad drawmg [illustration]. The towers and
indeed the whole bmldmg were of a beautiful bluish grey. How
wise the Church was to build in this imposing way. It was the
necessary set off to the castles of the savages we call the kmghts
of the middle ages. It must have powerfully impressed the
imagination m those rude times I should have said that the
ecclesiastical foundation of Limburg dates from about 800.
There is a beautiful effigy of the founder, of the 13th century,
about contemporary I suppose with the cathedral, the features
in style just hke those of the New College statues (over gate
on both sides and founder's tower).
What a pleasmg picture you draw of your father . . . . Yes,
Sewell was what I should have said They know nothing at
Radley about the carving except th.i.t they thmk 1t came from
Cologne. I am sorry to have to tell you that Mrs. Wilson seems
rather worse than better. . . . I have lied.rd of a capital place
m the Harz-a Kurhaus m the forest and I hope to cheer her
up there. I shall be able doubtless to get a chair to wheel her
about in as I did m the Black Forest and I hope she will forget
her troubles 1 ••• Kant says somewhere, I thmk, that' imitation
has no place m morals' This 15 qmte true as meant, but it has
always struck me that as a telling practical precept the very
opposite is true. When I am cons1dcnng how I ought to behave
in certain circumstances (I really do consider sometimes), I find
it far better than any maxim or precept to think how so and
so would have behaved in similar circumstances. One realises
the thing then properly, and it is really the best corrective of
one's natural impulse and want of self-criticism. Thus if I am
' He goes on to speak of his election to the Bntl!.h Academy (on which
I had ventured to congratulate him) and of Demard Bosanquet's chanmng
letter about 1t, 'hardly to be beaten for tact and courtesy.'
2773·1 i
CXXX FAMILIAR LETTERS
getting, I feel, out of temper in a debate, I think how Sanday
would behave in similar circumstances-that is, of course, when
I do stop to think about it. And when I feel my defect in the
ordinary courtesies of life, I pull myself up by thinking how,
e.g. Bosanquet or Strachan Davidson would behave. The ' voice
of reason ' is as a nothing to be compared with the memory of
such incarnations of 1t in others and I am sure it is one of the
best safeguards in any man's life. At first one is too proud to
do it, but I have got over that. As to studies of peasant life
read Auerbach's Barflissele (Little barefoot) and tell me what
you think of that. It 's not long and it is most restful and
I think beautiful.... I should love to hear from camp, but you
mustn't write if you can't get it in, I know how it is there.
N.B. I am anxious you should give up either your infantry
cy. or the Cyclists-one at least if not both-you can't stand
1t. If Lattey takes Cyclists, I would help a bit. I hope to come
back much better....

39
Sunday July 7 [1907].
Johannser Kurhaus bei
Zellerfeld (Harz).
Though I owe J - - a letter a long time I must wnte to you
first, for I am not comfortable about your work It seems
obvious you should not have begun on the Kant so soon after
term. I know the importance of sticking to a piece of work
hke that, if 1t is to be got through-and the work of a tutor
1s so entirely engrossing during term, that in the vacation one
must stick steadily to a thing even agamst the grain sometimes,
remembering that the ' term cometh when no man can work '.
But 1t was clearly imprudent, considering what you did at
Easter, to start on such a hard subject as Kant when term was
hardly over. As I said I suppose it means you must in con-
sequence of other vacation arrangements take the time for work.
Of course 1t 's better to arrange to have two or three clear weeks
of rest and recreation after term. But if you had to take your
main holiday later, then nevertheless you ought to have had at
least a clear week after term and then only done routine work,
 FAMILIAR LETTERS cxxxi
such as College business or lecture work or reading. You really
must believe me about this. You will do your work all the
better if you wait till really fresh for it and certainly get more
done too, if you have the self-restraint to wait a fairly good time
after term. If you're at the Kant still, I should venture to take
the risk of advising you to stop it.
As to your experience now, undoubtedly if you take the kind
of view we seem agreed upon (at least m essentials) of the
original source of confusion in Kant, the more his theory is
carried into details the more confusion it will naturally shew,-
that is only to be expected. As to style, I trust you will take
to heart what I said on my postcard. In order to be qmte
clear m your thmkmg, your expression at first must be just what
1s natural to you, or you won't really say what you mean. Of
course what you don't hke is the labour of writing the thing
twice over. Now I will confess to you. I generally write first
a rough draft, then I begm to write out as for prmt . but
presently I begm to think or rethink and sometimes seriously
alter, and then there has to be a second rough draft. Then
I try to make certain that the third (is final) and 1t often is not.
Lookmg back on the various articles &c. I have prmted, I came
to the conclusion that I generally write a thmg four times by
or before the last version ,-a thmg wluch would frighten me
1£ I realised before it was to be so. Each time I try of course
lo have fewer rev1c;;als but generally 1t comes to the same thing.
For you (as for anybody) I am sure the first one or two \.ers10ns
&hould be in the words and forms of sentence most natural to
you. Don't be afraid of repetition. IF you mean the same
thmg, say the same thing or you will really get altering your
argument. In the final form you can use literary devices lo-
overcome the verbal repetition. I am fairly sure Aristotle wrote
m this kmd of way-more especially that 1f a word was the
right and natural word in the present context, he kept 1t though
he had used (it) just before m a different sense appropriate to
the previous context. Of course a mathematician never thinks
of varying the phrase, because he wishes to be absolutely
accurate. I repeat that you should certainly not look at the
MS on Universals if the thing is at present off your mind.
I only left it with you because at one time it was worrying you
l 2
cxxxii FAMILIAR LETTERS

I am sure now I don't want to see it myself till I come back

to Oxford. The simplest thing is-if you will be so kind-to
leave it in my college room in any drawer of my kneehole table
which happens to be open. I am glad you have done corre•
sponding with - - . I know what weariness that sort of thing
can be. [There follows an account of the failure of Mrs. Wilson's
clinical treatment at Gottingen. He fetched her from Han-
over] ....
This is a charming place (we came here last Tuesday) and
I hope it will do her good. But the bother is that her knee is
not well and 1t prevents her walkmg except the tiniest distances;
so she is kept to the vicinity of the hotel-which however is
very charmmg and I suppose about 2000 feet above sea-level.
I can recommend this place. The hotel is charming, countrified
with spacious verandahs for coffee &c., plenty of public rooms,
excellent and abundant food ... lovely scenery m the midst of
firwoods and forests, remote from human habitation, delightfully
quiet, such a sense of rest about it-nobody rushing about to
see things. The society quiet, domestic, middle-class Germans
of a superior sort. . . . Also Professor Fick 1s here, the dis-
tinguished comparative philologist (retired) ; as he was at
Gottmgen we have common pomts of interest to talk about.
Hts wife and my wife too were girls together in her native place
in the Lunebcrger Haide. As for me, though otherwise very
much better, my cough is not gone but the doctor thought rt
certainly would go later. My great worry 1s that my wife
doesn't get anythmg hkc the enJoyment she ought because her
knee confines her to a narrow area Previously I have always
w,d a ' Bathchair' (light) and wheeled her long distances, but
• this time she won't hear of it, though really 1t is excellent
exercise for me. I do hope J. A. is relieved of anxiety about
his sister; there 's a man here who reminds us a good deal of
him .••.
With kindest regards from us both to both of you and love
to the children, yours truly.
 FAMILIAR LETTERS cxxxiii

40
Hannover. Ellernstr. 2.
I Aug. 1907.
It was so kind of you to send me a card here. We leave
Monday Aug. 5 and hope to be in Oxford on Aug. 7. My wife
picked up a good deal in the last ten days in the Harz in general
health and her knee at last gets distinctly better. The gout in
her fingers got very bad indeed and she is here trying a new
electric treatment which has certainly had remarkable successes.
I take her twice a day for the treatment, which may be described
as getting certain gout remedies directly into the fingers from
the outside (instead of through the stomach) by means of
electricity. The inventor was confident he could cure my wife's
fingers in IO sittings (for which we JUSt have 5 days), but so far
his ant1C1pations have not been realised as to rate of progress
and a good deal of pain is caused. The treatment is fairly
expensive and I confess to some anxiety now about its success.
Weather bad again after one or two fine days, on one of which
we left Gaslar [sic]. I renewed my acquaintance with Gaslar on
the way. The last Harz week I made (the) acquamtance of (a)
delightful German officer on the General Staff who had spent
the last month in England with Sir John French, &c. I had
rousing good walks with him. I must tell you about him later.
In the Harz the birds suffer from the multitudinous squirrels.
As to cows, distribution of their breeds follows d1stribullon of
races (Fick told me), e g. m Tyrol, the Slaves, the Germans,
Italians &c have their original breeds of cows. You will enJOY
the famous beauty of the Italian lakes
te, Lan maxumc, teque,
fluctibus et fremitu surgens, Benace, marino.
Let me have a card again soon ; affectionate remembrances
to both.
41 Hannover.
Aug. 3 [1907].
I was very glad to get another card from you here-I needed
it. The gout treatment seems a failure and therefore also
a serious waste of money. I am so sorry, for I had hoped my
wife had at last found something to give her relief. Yesterday
cxxxiv FAMILIAR LETTERS
came the very sad news that her only surviving brother, the
pamter, who had been for years in America was dead. It was
a shock to her.... I forget the Wages of Sin, tho' it interested
me much at the time. But I do trust you don't read English
books when abroad I L. M. always struck me as an effective
writer.
42
Hannover, Ellernstr. 2.
7 August [1907].
My wife has had to go into Hospital w1th diphtheria and we
may be delayed a week or ten days. She is gomg on all right.
I have written to ask Mrs. - - if she would look m at our
house as the house-parlourmaid (good hearted but headless} is
there with a new cook and a new boy I I She might kindly help
the headless one, 1£ m difficulty .... If you are not at Oxford
it does not matter.
43
Hannover. Ellernstr. 2.
7 August [1907].
There 1s no end to our troubles. My wife's expensive 'gout
cure' seems absolutely to have failed. Moreover we were to
have left yesterday (Tuesday), delaying a day on account of the
cure, but my wife got an attack which the doctor pronounced
diphtheria on Tuesday and the same night she had to be removed
to the hospital. I was in an agony of fear because the doctor
said it was rather advanced or, at all events, the white ' skm '
in the throat was extensive and he was clearly anxious. We
ought to have sent to him sooner but it seemed a mere infiam•
mation of the uvula at first. However she is doing well, the
serum was inJected last night and this morning she was, rather
to the surprise of the doctor, without fever and apparently the
wlute membrane has disappeared (two doctors testified to its
existence last night). I was the more frightened because she
always takes the worst possible view of herself-this of course
impedes recovery and is serious in such a dangerous matter.
She was quite overcome when she went away and said she knew
she should never come back. But that is all different to-day.
'Joy cometh in the morning.' I have been to see her twice and
the second time took joyfully a scolding for having bought her
 FAMILIAR LETTERS CXXXV

a book to read which she considered a waste of money-especially

as she supposed all infectious patients' books were burned. She
was also very afraid of the hospital after her experience of
Gottingen, but she is very well cared for and most fortunately
the head-nurse m her department is a great friend of my wife's
mcces here. It is troublesome that the hospital is fairly ex-
pensive. All expenses arc covered by about 17 /· a day, which
no doubt is not really much compared with England, but our
expenses have rather mounted up this vacation. [There follows
much about going to see the house-parlourmaid, with the new
cook and new boy.] ... You can imagine what trouble we have
had in the house about disinfecting, especially as my sister-in•
law with whom we arc staying has seven young girls m her
house who go to schools, and German regulations are strict ....

44
Ellernstrasse2. Hannover.
22Aug. [1907].
Your kind letter has cheered us. I regret to say that there
is no immediate chance of our travelling. My wife came out
of hospital on the 13th, but soon developed after-effects (appa-
rently) of the moculation and has taken to her bed. She had
such pam in all her limbs. Now it seems to have settled m the
knee which was bad originally. She suffers such pam when she
stands that, after frmtless efforts to be up and use the limb,
she has had to go back to bed and wait to see what complete
rest will do. Of course she is in the doctor's hands. We are
m a pitiable plight. I had hoped to settle to steady work m
Oxford by the second week m August and I really can't do
here the philosophical work which I am upon, One is so dread-
fully unsettled. I long too to be back to our house and garden,
1t is Just the time of year for it. We are much m the way here,
for my sister-in-law was preparing to go to America and she
sailed on Tuesday I On the Ist of September a young lady
returns who occupies one of the rooms we have and a deplorable
hugger-mugger must result if we arc still here, for if my wife
had to go to a Chnik again, I think she'd go out of her mind.
They are touchingly kind to us m the house and do all they
can for us. Another trouble 1s that the Kai~er comes next week
cxxxvi FAMILIAR LETTERS
and so, if my wife 1s not well enough to travel before, we should
be delayed, for the trams will be too crowded for us. Please
ask Spenser to let me know when the Hertford fellowship is, as
I have promised to examine. Hannover disagrees with us both,
so our enforced stay 1s the more tantalising It was so kind of
you to go to our house so soon. Kindest regards.

45
Ellernstrasse 2. Hannover.
28 August [1907] .
. . . Many thanks for your very kind letter. I hoped to start
on Friday or Saturday, but my wife is now suffering with
a floating kidney trouble wluch comes and goes intermittently
and she 1s afraid about starting We all here have no doubt
she ought to make up her mind to start: 1t 's a good deal
a matter of courage and nerves but we don't like to tell her.
It 1s just about the last straw to me and I don't know what'll
happen, if we don't get off tlus week It's all I can do to bear
up at all, and I have begun to suffer from (a) very disquieting
feelmg of pressure on the brain, hkc what I had after strychnine
p01soning. Perhaps I'll write you a letter, but I want to get
off this because of the question about (the) Hertford fellowship.

46 [Oxford]
13 Nov. 1907.
We will consider the return of my MS. after the turmoil of
term is over. Perhaps you won't forget about 1t even if I do .
. . . I was sorry to nuss you but sec no prospect for the present
of havmg time for a talk. I am examining for the John Locke
and for B.Sc. I do hope Mrs Cook Wilson will let you examine
for me in the former next year. Many thanks for the kind
message from your wife. One of the compensations of age 1s
that one's lady friends send their love to one. Some of mine
are quite reckless m the way they transmit such messages. I will
tell you an amusmg example when I see you.
I have thoroughly sifted --'s case, interviewing all kinds
of people . . I will tell you all about it when we meet. It
was an interesting study of evidence. Everything proved accord-
mg to my view, before I got the unexpected ear-witness....
 FAMILIAR LETTERS cxxxvi1

47
[5 Granville Place, Portman Sq. W.]
26 Dec. 07.
It was one of your kind thoughts to send me a letter of good-
will this Xmas. These things are a great help ' m dieses Lebens
Jammerthal '. Your letter suggests a delightful picture of
domestic happiness which you and your wife so well deserve.
But I miss one thing-you haven't told us about little Jane's
Xmas pleasures, and what she said and did.
In the doubts and difficulties about good and evil m our lives,
there is one thmg which is most comforting and encouragmg to
me and that 1s that m so many young cluldren tender and
affectionate and even unselfish feeling ts so soon developed. It
does not come as the slow result of a long experience m which
by repeated and painful effort a ' natural ' selfishness 1s gradually
ovcrcome-somethmg which needs much reasoning and reflec-
tion : but it is there almost at once-like a revelation. The
fcelmg of gratitude for services rendered seems to us, by habit,
so obvious as to need no explanation and it docs not surprise
us even in the animals. But how much reason it implies-the
realisation of the existence of another personality, the attribu-
tion to it of agency, rcahsmg that some one else has done this
for us and has caused 1t as willing it-and that this person also
1s well disposed to us. All this seems to me implied m gratitude
for benefits received and yet qu1te tmy chlldren are grateful.
Besides, at a lower level, they arc not merely pleased with what
they get but they are quite obviously pleased (even 1f the feclmg
1s not developed enough for gratitude) with those who have
given them a pleasure.
It is a most encouraging thmg for our view of humamty that
quite the highest and best feelings and the amount of intelligence
that they imply arc found quite at the bcgmnmg and not as
the result of careful thmkmg about experience-as a slowly
learned lesson.
Of course the familiar rubbish about ' instinct ' is no help and
1s a sort of irreverence to the human spmt. So far as there is
any truth m it, it is an imperfect recognition of the fact that
affection for others is primitive in the sense that 1t is underived
cxxxvih F A:MILIAR LETTERS
from anything else. It is indeed a rational emotion, reason's
own immediate activity and undesired possession. You might
just as well call the highest thinking possible 'instinctive', and
it is ' instmctive ' in the same sense. So you see the immense
importance of httle Jane-that mighty atom. We saw Marie
Tempest m - - , a poor play but she was as excellent as ever.
We dmed with the C-s on Xmas day. Mrs. C - and her
two daughters arc daughters of Anak. ... Friday and Saturday
we arc hearing Wyndham and James Welch. Next week
Charley's Aunt. . . . I hope little Jane got her doll all right (from
Mrs Wilson) as well as the mechamcal toy, which does she hke
best? I have no poetry for Jane this time (J01ccy p) calls it
' rhymes ' I) for I have exhausted my vein in writing German
verses to a delightful httle girl in the Harz to whom I'm Onkel
John. {I have brought Caird's addresses with me, you and your
wife would greatly enjoy them) A pamtcr friend of a Jewish
lady friend of mme, who has long wanted me to sit for my
portrait, has collared me at last and I am to begin sitting
to-morrow (Friday) I The painter is first rate. My wife much
plagued with gout. Both of us join m kmdest wishes to you
both, yours affectionately.

48
S Granville Place, Portman Sq. W.
31 Dec. 1907.
It was so kmd of you to write to both of us but one envelope
and one stamp would have done-youth is so reckless. We arc
so sorry to hear that little Jane was so unwell, but it often
happens after parties to children otherwise healthy enough.
Please give her our love. We both send our hearty good wishes
to you all for the New Year. Please tell Jane I was delighted
with her pretty little letter and wish her a Happy New Year.
She doesn't say what the castle was budt of m which I was
imagined to be.
The drier air in London is doing my wife good. She is dis-
tinctly better than when we came. I take care to wrap up well
after the theatre-this is in reply to your husband's kind moni-
tion. This morning I took Miss C - and Miss G-- to look
 FAMILIAR LETTERS cxxxix
at the Rembrandts and other Dutch pictures in the National
Gallery. The arrangement has been much changed in the last
year {' hew management ') and I think for the better. E.g. three
pictures of De Hoogh's, which I always used to compare w1th one
another-two hanging apart in (the) same room and another in
a different room, are all hung together. Artistically they are
best in 3 different rooms and I think the artist would have
preferred that : but for purposes of study one wants them side
by side. But perhaps after all this is wrong as a principle of
arrangement in a gallery, m fact on 2nd thoughts it seems
possibly horribly wrong. However pictures of Ver Meer (or
Van der Meer), Maes, Metsu, Ter Borch &c. are now brought into
the same room, which seems both convenient and right. I sup•
pose Jane would say it was no use putting good wishes at the
end of the letter as they have been already expressed in the
letter I Yours truly.

49
[Oxford] 29 Jan. 1908.
µ711l/ µ' lpv,u µ&x11s t/n'll.Jwv 'ITEp"ovlil JJ,E 'ITfl(JUS, 1

50
[Oxford] 8 Feb. 1908.
Confidential
The teichomachy 8 has an amusing sequel. You r<"member
Murray abandoned the Jones, Brown, Robinson passage, though
it is the one he quoted for (the) 1st year of War in his book
and the only one. He told me he really meant to quote .S 32,
not M 1-32. I had considered the passage myself and never
imagined tt contained any reference to the first year of the war,
but I did not feel competent to deal with rt decisively last night
on such short notice. I have looked it up in Monro's notes
{xili-xxiv) Monro's explanation (wluch excludes (the) idea of
• Iliad, xvm. 126, with </>l'-lo11aa ,rrp, of course Achilles to his mother,
'Hold me not back from battle, though you love me I will not obey.'
' The ' fight at the wall • was a debate on the 7 Feb at the Oxford Plulo-
logical Society upon part of a paper called On the Suniles of Homer read by
Wilson to refute Professor G. Murray's ' advanced• views m Homenc cnticism.
It was a field-day with he.,,vy guns ; Sir Arthur Evans, among others, being
present at lea9t for part of the ti.me.
cxl FAMILIAR LETTERS
1st year of war though he doesn't say so) is obviously right and
entirely convmcing. It explains also the point of the passage,
which the mterpretabon of Murray's friends does no~ in the
least. In fact the 'advanced' ones have misconstrued and
badly. Just look at Monro's note. Godley agrees, he tells me,
entirely with me about the Wall. He thmks, as I do, the sup-
posed d1fficult1es about its occasional absence purely imaginary.
The whole wall argument, I think, mcludmg these precious
construes, rs a heartless hoax, due to a lying dream sent by
Zeus to the advanced ones.
From a remark of Murray's to me after, I gather he was
under some illusion about your remark to him about the way
Mill's contradiction could be explained. I suppose in such
circumstances a man naturally thmks hastily a new pomt hasn't
been noticed by his opponent and therefore is m his own favour.
Of course the contradiction can be explained-I thought I had
given the explanation myself-and I did, only I didn't dwell on
1t. I thought of workmg it out quite clearly last night but
didn't want to waste time, as 1t didn't matter.
Obviously all these contradictions have an explanation. 1 The
first thing 1s to establish their existence cmpmcally : one then
explains them, and not by the rational workmgs of consciousness
as I abundantly indicated. So m the 'first over the wall',
I offered an explanation as to how the poet could get confused
over it. In fact this is the 'fightmg m the dark ' which the
advanced school do not understand You were doing this when
you offered an explanation of Mill's contradiction to Murray,
but he didn't realise that.
The valentines haven't appeared in the shops yet.
Yours truly.
PS. Godley said he thought I ' had ' Murray generally,
because he (and even more the school he follows) depends upon
contrad1ctions, despite what he said-indeed he shifted ground
(gracefully), for the view specially criticized depended wholly on
contrad1ct1ons.
1
See 'Natural anomalies in original composition', The Classacril Revi,w,
vol. xxiv, p u8.
 FAMILIAR LETTERS cxli

51
Oxford.
7 April 1908.
It was very kind of you to write, for I felt when you left that
two guardian angels had gone off duty. I saw Jane in the
street the day you left but haven't yet paid her a call ...
Thank you for the view of Wells. I have never been, and
always hope to go there and to Glastonbury. It looks very
attractive in the picture. I go in to sec Caird sometimes : he
came here this afternoon-a lovely day 1t has been and the
buildings looked wonderful in Oxford. My head is distinctly
better. Yesterday I had headache until about tea-time but
to-day there was none and the oppression was very much less.
I could have done some work I am sure this morning, only I had
a good deal of business out of doors which I naturally put into
the morning. Indeed I have idled all day, for I mowed the
lawn this afternoon (1t was very nice to do m the sun, I delight
m the smell of the cut grass too, it 's refreshing) and after the
Cauds were gone, I had to go out with my wife on business
(to look at kitchen ranges, as we have succeeded m getting our
landlord to contemplate a new one for us). Joseph 1s here for
a few days. In the latter half of the day, 1 e after tea, I have
been able to work fairly , but it 's frightfully hard stuff I am
thinking out. I have been ' stuck up ' nearly a week on one
point, but I have at last got hght, I think. ... My wife also is
gettmg dreadfully nervous about the trouble of going away the
last week to Woburn Sands. If however she gives it up, I must
go away a few days myself-probably to my friend in Sussex.
I do hope you are thoroughly resting, both of you, and enjoying
yourselves. I am fully content here when able to work. It 1s
really very nice here. Mrs. Case has lent me Sir Sterndale
Bennett's 1 Life. It is so mce-you ought to look at it. I am
solving a ' logical ' conundrum sent me by a man in Kentucky,
US A. I have to do these things sometimes. It's an [sic]
awful piffle and he ought to have done it himself or got help
easily at home, but then his institution 1s ' The Potter Bible
College'. Very kmd regards to both from both, yours truly.
1 Mrs T Case's father.
cxlii FAMILIAR LETTERS

52
Villa Taunus, Ems. June I, 1908.
Herc 's a geometrical problem for you suggested by Roman•
esque architecture. Suppose a tower 1 has a square section and
each wall terminates in a gable as at A and D. Let AB be the
height of the gable. Then if the planes of all such triangles as
ACD arc continued so as to cut in such lines as AE, DE, four
4-sided figures such as ACDE will be formed. Prove that, what-
ever the length of AB may be, these figures are true rhombuses :
so that AE, ED, DC, AC are equal. The same is true if the
section of the tower is a rhombus and not a square. The proof
is delightfully simple.
I had a dehghtful journey here The Surrey landscape was
golden with a splendour of broom and gorse. (Read a man's
dissertation for B.Sc. before leaving and this is the effect, for
he couldn't say anything simply. Lud I I let him have it.)
Dover, where (via Reading) I had five hours, was en fete for
the French President. The promenade pier was outlined in
electric lamps. So were the promenades on sea front and
another pier (not the Admiralty), and the scene lovely beyond
description at night. There were about 30 Torpedo boats in
harbour and from 9 to 9.30 they gave a display with their search
hghts-prmcipally by dluminatmg the chalk chffs. I had a
capital passage and a most comfortable coupe to myself all
night. Arriving at Cologne at midday, I was able to start by
the one o'clock boat to Coblenz. It was a perfect day for the
Rhme. It has rained for me on most previous occasions. The
view of Bonn and the Siebengebirge was very fine as we came
up the river The S1ebengebirge was very imposing and looked
the ideal home of Saga. I had forgotten how very wide the
Rhme is about there-the width of the river adds much to the
effect. Had a most amusing experience of the dependence of
' apprehension ' of objects on physical conditions. As we got
abreast of the Siebengebirge near Konigswinter and it was
opening out fully, something got in my eye, I believe an eyelash,
and you want another Johnny to get out an eyelash for you.
1 This JS Jllustrated by a drawmg of a typical German pyramidal roof,

like Somptmg tower in Sussex. A, D are vertices of gables, C a tower angle

 FAMILIAR LETTERS cxlth
The pain, increased by futile efforts to get rid of the thing, was
Just maddening and I couldn't manage the view with one eye.
Now if it happened to a lumpish creature only intent on his
Rhine wine and the tempting food they give you on the steamer,
there would be something to be said for the ' fitness of things '.
However by heroic efforts I at last got rid of the nuisance.
Thereafter were the most splendid views of the Siebengebirge
10 the rear as we ascended to Rolandseck. The space between
the Siebengebirge and Rolandseck I think is one of the finest
t.hmgs on the Rhine or anywhere and one saw it to perfection.
Yet one of my ra10y times on the Rh10e I saw 1t 10 a more
poetical mood, for it was near sunset and the rain had cleared
and storm-clouds ht up weirdly by the fine sunset were hanging
about the Drachenfels. It looked like the land of the Opera,
Wagnenan, not hke the world at all. There were lovely views
all t.he way up to Coblcnz. Bywater, years ago, had stayed at
the 'Giant' Hotel and so I piously went there, tho' somewhat
alarmed to find 'Hotel zum R1cse-Furstenhof' painted outside
it. However It was one of the few starred in Baedeker. I can
recommend it, Its situation close to the nver is charming and it
is not at all extortionate. I asked for an inexpensive room and
got one, on the first floor too Room and breakfast (plentiful
coffee and rolls) 4 35 marks. The next day I went up Ehrcn:
breitstem and succeeded 10 gett10g a nse out of the stohd and
trusty-looking sentmcl whose face melted mto a smile from one
ear to t.he other. By the way I took another sort of rise out of
an officious railway guard, who when we got over the German
frontier at Hcrbcsthal told me I must put my luggage off the
seats up on to the nets. I had indeed occupied the whole
coupe with it, and it 's fair to say people might now be expected
to get mto the tram 10 greater numbers, it being morning.
I did it and then some people came and secured their places
by putting luggage upon them. The guard came back and
remonstrated with me for not having put up my luggage (the
other people had gone out again). Without jaw J said it wasn't
mine. He pointed indignantly at a piece which really was hke
one of mme. My reply was bnef and German, not forcible-
the facts were that, and he hastily departed with as complete
a silence as I ever remember to have observed (the German for
cxhv FAMILIAR LETTERS
him is wie ein begossener Pudel). I spent some time trying, as
I have before, to discover where Turner painted Ehrenbreitstein
from and how he got the elements of his picture together. I got
unexpected help from an intelligent ' unteroffi.zier ', whom I
asked about a curious-shaped monument conspicuous in the
foreground of Turner's picture. He told me there was one like
it on the Karthause, the high ground on the other side of
Coblenz, opposite Ehrenbre1tstem, so that the whole town hes
between the Rhme and the Karthause. The consequence is that
the point A represents the pomt from which Turner would get
the view of Ehrenbreitstem that he represents, the scenery he
puts m his immediate foreground 1s at pomt B on the other
side of the Moselle and at some considerable distance. This is
what the ' Master ' does (Do you know, a man rev1ewmg my
memoir of Monro and wishing to pay me a compliment said it
was done in a masterful manner?) Can't you imagine the
idiotic delight of a •pragmatist' (excuse the low slang word)
at discovermg that Turner was a p. t and identified truth
with whatever was convenient to him. I continued my journey
by Rhme as far ac; the mouth of the Lahn in perfect weather
and very fine views astern of Ehrenbreitstem, which detaches
itself more and more from the adJacent highground from this
pomt of view. The mundung of the Lahn mto the Rlune is
very beautiful, on one side the rums Burglahneck and on the
other the Schloss Stolzenfcls (modern rcstorat10n) m picturesque
lights, the former on the Lahn, the latter on the Rhme. Then
a short ndc by tram to Ems. I was much welcomed by my
landlady and her husband (the latter 1s playmg to me at the
moment, for he 1s one of the principal bassists in the band, which
I am listening to as I write m the Kurgarten, where I have
been having my coffee). The husband was just engaged in an
improvement m my room ; which had a homelike feeling for
me (antecedent to wluch ?). I missed a most convenient high
desk out of it and was informed that a ' candidat ' (young man
in orders and with a view to a charge) had been allowed it in
my absence, being strictly warned he must surrender 1t when
I came. I told them it did not matter, but they insisted and
the poor candidate had to disgorge I am really most com•
fortable m my room, it's rather large and plenty of accommoda•
 FAMILIAR LETTERS cxlv
tion for one's clothes and all one's etceteras. It is quiet and
retired being at the back with nice view on to the hills-tho'
I own to buildings of a ' back ' kind between. By the way ask
Mrs. - - for me (you're sure not to know yourself) what the
exact meaning and motive of Sullivan's Lost Chord is. They
play it here occasionally and translate Der V erklungene Ton,
which seems wrong. I can't guess from the music itself what
it 's about. Aubrey Moore did have a Jest about Noah's arc
and the lost chord a propos of my immortal inaugural lecture.
It's very beautiful here but I want some one with me to share
in enjoyment, but by dint of writing this to you in very beautiful
surroundings I have ' raised ' you in a kind of way (Samiel I
Samiel I Erschein I) and don't feel so lonely. Send me a line
when convenient. Tell dear little Jane I don't forget her at all.
She's ineffably sweet (no danger you'll tell her that). Look up
Mrs. Wilson when you've time and when you haven't. I have
quite cheerful cards from her. I can't say much yet about my
kur. I have to sit in a thick fog (artificial) l an hour a day
fas though a philosopher hadn't enough of that anyhow). A new
nose treatment seems a mistake (new for me this time) and it
produces the symptoms 1t ought to destroy : I must consider
this with my doctor.
Kindest regards to both of you.
Went to English church on Sunday on chance of finding there
some one to walk with, but the few men there seemed no class:
so I was punished, but I only gave sixpence to the collection.
Haus Taunus, Ems.
IO June, 1908.
I enjoyed your interesting letter (in wh: you ignored my
question) but you left out the Prince of Denmark. There was
nothing about J. C. R. F., the topic of greatest interest: the
dissertation therefore did not ' attain the highest degree of
merit '. I am sending her a little model of the ordinary German
wheelbarrow, which I hope will appeal to her refined taste. Tell
her to go and give my love to Mrs. Cook Wilson (that will amuse
J.C. R. F.). Glad you met the P. of 0. 1 and do not doubt the
acquaintance will ripen. Gratulor Garsington 3 titulis accedere
nostris .•..
1 Dr Shadwell, late Provost of One!.
• The cyclists had sketched and reported on this village.
:1773•1 k
cxlvi FAMILIAR LETTERS

53
12 Fyfield Road. II July, 1908.
Louis Dyer had, it was supposed, very painful neuritis m (the)
left arm. He went to Matlock for a fortnight-suffered agonies
and they idiotically dosed him with opmm, which the doctors
here had to try to get out of his system. The symptoms how•
ever soon made them suspect grave m1sch1ef, malignant disease
of the bone. It was thought that 1t might be enough to remove
the top part of the arm near the 'ihoulder, just as in Strachan-
D,1V1dson's case, the bone gradually forming a new surface for
the JOmt. But 1t was feared the arm might have to be ampu•
tated. This alas proved necessary: 1t was mahgnant, but the
tlung had spread to shoulder and (the) shoulder had to be taken
off too (done m London). He seemed makmg a good recovery
when a clot of blood got on (the) bram and caused paralysis
from which he died very soon. J saw the poor fellow a day or
two before he went to London. He was so brave and even
cheerful. I was filled with admiration of him.
The doctors said his case was anyhow hopeless. If he recovered
for a bit, he would have got bad agam for his system was per-
meated with cancer. I didn't know he was such a friend of
yours. I have known him for more than thirty years. Poor
W--, I wrote to him. They sent round a prmted acknow•
ledgement to their numerous sympathisers .... It's capital m
Oxford now. I take a header at our boathouse at 7.30 unless
(the) mornmg happens to be too v.mdy (tho' entre nous I thmk
that's all fudge as we have the boathouse to dress m) I can't
help thinkmg they made a mistake about Grenstcd. If you write
to him convey my sympathy. I have been greatly interrupted
in my work since we are back-can't be helped; but we have
JUSt secured Mrs. --'s cook, who seems to be a treasure. She
will come after we return (Sept. I 5, circa) and I trust hfc will
become easier. You can't conceive how my wife worries over
housekeeping and she has indeed had baddish luck since we
returned. However we have got what seems a capital new boy.
We seem as if we should have some peace at last. Of course
I have had so much bother with the house and getting the new
cook (advts. and letters) and my wife also bemg so poorly, that
 FAMILIAR LETTERS cxlvii
the work has (been) very sadly interrupted. At Llandrindod
I shall have very much fewer duties and (I trust) fewer worries,
and I hope, tho' in a boarding house, really to get some work
done. I have, I confess, spent some time with Bywater-that's
natural. I do feel so sad sometimes about his going. (N.B. clrco.A•
aa"Tor is, on account of privat:, proparoxytone.) How sweet
of dear little Jane remembering us m her prayers Little Alan
Stout does the same for Prof. Bosanquet and for me. I must
write and tell Mrs Stout I hope he keeps 1t up, for we both
need it. (You remember my addition to the usual health at
mess, ' The King, God bless him I ') How you must have been
delighting m httle Jane Yes, we just miss We depart Tues·
day. Bywater has given me some more books, the photograv.
of his picture and a photog of the Ba.le picture of Erasmus.
J - - was here a day and gave me Nicholson's 'etching' of
Oriel. Much love to Jane. Our very kmd regards to you both.
Yrs truly.

54
Montrose, Aspley Heath,
Woburn Sands
[last week of July 19081
It is so kind of you to renew your offer to take me into
your party m Cornwall. I.can conceive nothmg more delightful
than to stay quietly m your happy family circle, with little
Jane as the centre thereof. It is so kmd of you to be willing
to extend the radius of 1t to mclude me (Postulate Let 1t
be granted that the family circle may be described with Jane
as centre and of such radms as to mclude C. W.) But 1t isn't
really possible. I doubt whether it would be advisable to leave
Mrs. Wilson alone in Oxford till the 18th when we go to Lian·
drindod. She has been hysterical lately and has no friend to look
after her there Moreover she is nervous about her fingers, which
are very bad, and wants to have them examined by the Rontgen
rays. I think it might have been a good thing if she could go
to Llandrindod alone or be there alone, if she could have been
with a friend there (Mrs. D-) but that plan has broken
down ; Mrs. D-- will be there but can't be in the same
boarding-house. I could then have gone somewhere myself and
177,1•1 k2
cxlviii FAMILIAR LETTERS

got some work done. It wd. have been nice to come to you
then but I shd have had to work a good deal and it would be
too late for you (latter half of August). Meanwhile don't worry
about me. I am much better, have done some satisfactory
work, if no great quantity One of my speculat10ns seemed to
go so easily into Aristotelian Greek that I wrote it 10 that
tongue 1 Tl seemed really clearer and more adequate to the
subject than English. I tried it on Gilbert Murray (with whom
I had previously discoursed on the adequacy of Gk as a philo-
soph1cal language), I sent him no English as the thing won't do,
unless it tells its own story He was delighted and said he
thought it couldn't be so clearly said m English Love to Jane.
Kindest regards to Mrs - - and yourself With much gratitude
for yr kmd help and kmd mvitatlon.
Yours truly.

55
Oxford 7 Aug 1908
We returned here July 31st The last week .tt Woburn Sands
was very hot and my wife was murh tried It was difficult to
get a cool place When the sun went down, however hot the
day, it was usually too damp (at once) to sit out. [There follows
a long and distressing account of the radiography of Mrs Wilson's
hands and of the defcct1on of the new boy He was followed
by a 'motherless' boy, a great relief] ..
Bywater 1s here from time to time I am to dine with him
on Sunday. I was so sorry to see Grenstcd 10 the 2nd class
Perhaps I made him too clear and all he said was mtelhg1ble,
or perhaps he failed m history . . We go to Llandrmdod on
the I 8th and stay 4 weeks I am qmte hopeful that I shall
get some work done at Llandrmdod. What I have done has
been pretty satisfactory and I am not troubled by that dis-
tressing oppression in the head . . I hope you have a piano
I am looking forward with horror to 'Kurkapelle' at LI It's
awful. ... Our garden and verandah are JUSt delightful now-
bkfast on verandah difficult to beat anywhere 10 Europe More
kmd regards, yrs truly
1 Published 1n the O!11/ortl Macaz1ne, vol. xxXJ., No. 16, 1n 1913.
 F AIIILIAR LETTERS cxlix

56
Oxford.
Sunday 16 Aug. [1go8].
I have left the keys of our boat (' Witch '} m College lodge
for you. . . . I went first to your house but nobody there to
take the keys. I cut off the dead flowers (a quantity} from the
httle standard which flowered so freely, so that you may have
a chance of seeing it do 1t again, for which I claim a place m
your ' spiritual exercises '. Been seemg a good deal of Bywater,
last glimpses of him m Oxford. He 1s so pleased with Raleigh's
lecture on Johnson and wished me to tell him so. Look up
Bywater yourself, he's here the weekends till about end of
August. Much love to Jane W1r haben schon eme Kochm,
cm Schatz I
57
5 Granville Place W
3 I Dec [19081.
This is to wish you a Happy New Yt'ar all of you, and to
thank you for another kmd letter describing your Xmas festi-
vities which mterested me I suspect little Jane didn't get the
mechamcal toy (' galloping maJor ') which was to have been
sent on 23d If not, wd. you mmd applymg at the shop which
rails itself • Sports Club', or some such name, opposite the
Town Hall m St Aldate's and wd. you mmd at same time
askmg 1f they sent a toy from me to the Carntts (mcchamc.i.l
swan) in Holywell. Yrs
58
[To 1llustrat«- his theory of the genesis of the Anstoteban text, Wilson had
~ent a correspondent the shp proofs of the b'lt of contents of hu, book On
llie Traversing of Geometncal Figures. These showed a very considerable change
and growth 1n the book after the first proofs had been pnnted and m the
process of proof correction This letter was m consequence l

[Oxford] 2nd March, 1909.

As you are good enough to take the matter seriously, let me
ask you to consider another point. What would happen if,
instead of having to deal with the easily altered shps of printer'&
type, I had had to give my draft to the scribe who wrote the
roll or, more probably, had begun to dictate it to him myself ?
k3
cl F A'MILIAR LETTERS
These alterations and additions would all be excogitated while
the professional scribe's calligraphic copy was being made, and
I suspect more than one was dictated at once. The list I s111t
you does not betray all the interpolations I had to make in the
sections whose title 1s unaltered and alterations too to prepare
for the new. I 1magme that, for a writer contmually recom•
posing while he was d1ctatmg the final book, the roll might have
to be cut and strips mserted, perhaps sometimes a new strip or
piece was simply pasted over the place for which a new passage
was to be substituted contammg the substitution. Both pro·
cesses were hable to accidents. The pasted piece might become
detached and leave the old text, mcons1stent with some new
matter further on. Mistakes might happen JO the cutting and
insertions; the begmnmg e g of a passage might be at the
bottom of a column and, when the next column was cut straight
<
down, the piece at the bottom of the former column) might
remam and thus you might have a duplicate head or tail of
a chapter. And there are other thinkable accidents. More
especially short sentences, which poJOted forward or backward
to a context removed or altered, might remam This would
account for some such phenomena JO the Politics (of Aristotle).
My mam idea, you see, 1s that the prmc1pal disturbances of the
text were not due to later scribes or later d1slocat1ons of the
MS. (I profoundly d1sbeheve cspecially m such dislocations) but
that they probably happened JO the writmg out of what was
the final form of the book, as for pubhcatlon, owJOg to an
1d1osyncrasy wh1rh I fortunately exemplify m my own (im)proper
person. [This view had grown on Wilson through a study of
the Politics] Ah,o there arc certam stylistic characteristics,
which I thmk were due to the mere fact of dictation as I have
occasionally expounded 1n lectures

59
Wed. 7th April, 1909.
12Fyfield Road, Oxford.
We got back on Monday. The time was rather unprofitable
because the whole fortmght had scarcely a warm or sunny day.
This made a great difference to my wife, who couldn't get out
much, and she returned both unwell and dispirited. Unfor•
 F AKILIAR LETTERS cli
tunately a change of domestics has turned out badly, or partly
so, for though we have a very nice woman as cook, a young
German house-parlourmaid very specially recommended to my
wife turns out to be a minx and we must be rid of her as soon
as may be. This sort of thing affects my wife very much and
has caused a return of hysteria, her condition having been lately
very neurotic. These things affect my stupid head also. Besides,
I have had to be constantly employed as errand hoy during the
uncomfortable transition period. and have besides to go with
my wife sometimes on her necessary errands, because it isn't
quite safe for her to be alone. It is hard on her for she suffers
otherwise enough from painful ai thntis 10 the finger Joints, for
which we have tned much in the way of remedies and in vain.
So I haven't been able to get to our suh1ect till this afternoon,
and as my head isn't very well I will confine myself to the first
point. If I don't do more it is quite good for me : indeed it
is a kind of welcome anodyne. 1

60
The Boar's Hill Hydro, near Oxford.
16 June Thursday 1910.
Your question was as usual qmte relevant but when developing
a contmuous argument, as opposed to discussing, I find it difficult
to switch my mind off to another question however germane.
And so I forgot that at the beginning of my paper I said not
that we thought of hving as = growing but that growing was
the fundamental conception we had about the hvmg thing, and
I gave as an mstance that we thought of a seed like a bean (not
as something growing but) as something which could grow,
whereas a pebble could not. This 1s an mstance of somethmg
we contemplate as belongmg to livmg matter as against dead,
though we don't think of it as actually growing, and our idea
of it as such, i. e. as living and not dead matter like the stone,
depends on the idea of growth. Later on in the sequel (unread)
of my paper I have pomted out that m the more scientific con•
cep~ion of the hvmg thmg as feedmg, growing, reproducing (and
moving otherwise) one sees that the bottom of the thought is
that life, as the hving quality of the hvmg thing, is the principle
1 The remainder of this lettel' ,s 1n Part V, xiv.
clii FAMILIAR LETTERS
which makes 1t have the actav1t1es of growth, movement, &c.
though it is only thought of defimtely through these observable
activities, in which (according to a familiar usage"of speech in
such cases) we should say 1t 1s 'mamfested '. These two parts
of my paper (if I had only thought of it) give the answer to
your question and objection so far as it implies 'alive'=
' growing', in the sense of giving what I think about it, at any
rate, and had expressed I had begun to think about ' decay '
but, partly m (the) hurry to get it done and partly through
much preoccupation, qmte forgot to develop it in the written
paper As to the plant at the stage when it perhaps 1s growing
no more, we, I thmk, clearly regard 1t as ahve, as long as the
sap rises and is absorbed, 1. e. as long as it is ' feeding ' : which
really involves (1f we reflect so far) that, if then not growing
absolutely, 1t 1s losing matter (wasting) which is replaced by the
feeding, and thm, 'growing' relatively is going on. The dying
plant I suspect we thmk of as having parts of it dead, parts
which no longer feed, m which the sap doesn't rise, and are just
dead matter as lifeless as what never hved, as the w1thermg bp
of a blade of grass or the husk of wheat: ' dying ' meaning then
havmg more and more parts dead, and also probably as having
less and less sap nsmg m the parts we still think ahve. They
feed less and less until unable to feed at all and cease not only
to grow but to be able to grow
The more difficult case however seems to me to be the one
which I mentioned at the begmnmg of my paper, viz. that of
the seed. For, whether 1t be really growmg or feeding m any
sense or not, we don't th111k of 1t (normally) as domg either.
But ~e do think of it as something which can feed and can
grow : whereas the dead parts of the plant which have ceased
to feed cannot feed or grow again. I suspect that m practice,
while we certamly don't think of the seed as dead, hke the
husk, we don't naturally think of 1t as (actively) ahve or strictly
speaking (as) · ahve' at all, simply because of its quiescence.
While we don't think of the :;eed as dead (wh· N.B. ts obviously
not the same thmg as thinkmg that 1t is not dead) I suspect
we only come to reahse that 1t is in some sense ' alive ', when
we find certam seeds won't grow any more than a pebble--then
we naturally come to think of those which do germinate as
 FAMILIAR LETTERS cliii
having been ' alive ' tn some sense when not germinating.
Growth is therefore exactly the test used to decide even here,
in what is not actually growing, whether it 's ' ahve ' or not.
Not only, then, that which 1s growing but also that which has
the potentiality of growmg is considered by us (as) alive. Our
thoughts are so filled by the idea of growth, the actuality by
which we test the potentiality, that we don't reflect on what
the seed actually is as opposed to its potentiality-but N.B. that
tf the question occurred to us at all we shouldn't answer ' 1t is
a dead thmg but one with the potentiality of hfe '. Probably
we thmk of it m a way which corresponds to the formula
'suspended animation', even if we don't use this phrase. For
I suppose this formula 1s itself the result of the observation
that somethmg which doesn't manifest the ' activity of life '
nevertheless can't have been dead, because it afterwards shows
such activity. It 1s Just the kmd of uncritical formula which
belongs to imperfect reflection-or reflection m 1ts first stage.
We are probably helped by the analogy of sleep m the eonsc1ous
Irving thing, and mdeed the word ' dormant ' 1s what we are
likely to use
Yours truly.
61
The Boar's Hill Hydro, near Oxford.
21 June 1910.
Your letter gave me great pleasure. I am glad to thmk that
I have been of some use to you , such generous acknowledge•
ments as yol.V's are a great encouragement. And it was a no
small encouragement to have a few men hke yourself who never,
or ' hardly ever ', missed the weekly two hours' talk for the
whole year and who showed such unwearied attention as you
did. I always greatly enJOY the informal class myself and it is
very mce to find from ttme to time that my pupils enjoy it
also. Not long ago a young American professor writing to me
said our conversations in the informal instruction class were
such a pleasant memory to him, and your own kmd expressions
brought back his words to me
I trust that if I can be of any use to you m the future you
will not scruple to write to me
Owing to domestic troubles, which quite prostrated my invalid
cliv FAMILIAR LETTERS
wife, we fled up here before the end of term and my last informal
class was held here. With kind regards and best wishes, yrs.
truly.
PS. I should have rephed sooner to your letter but didn't
get 1t until to-day when I went down to College.

62
[Shap, (? July), 1910.]
My remarks on the teaching of mathematics concerned the
teacher's own method rather than the text-books . . . The
defect of method m the qmte elementary teaching of Geometry
1s that boys are usually given Eucbd or some other geometrical
treatise to learn the demonstration of propositions. This, I
thmk, should not be done. The teacher should propose each
propos1tton to the pupils as a problem to solve. Let them try
first For instance, ask them how they would describe an
equilateral triangle, 1f they were c1llowed a pair of compasses
and a ruler. If the boys :find a solution of their own, it may
be right, even 1f not the orthodox and, 1f 1t really assumes too
much, this can be pointed out, and the pupil should try to prove
the assumptions he makes, which will often be quite right. If
they cannot find the solution, they will often become far more
able to appreciate the proof when 1t 1s given to them by the
teacher. And sometimes 1t will be possible to lead them on
from the pomt they have go(t) to. If possible, I should do this
from the very begmmng. E. g , tell a boy what you mean by
an isosceles tnangle and then ask him to find out how to make
one. Then you can lead lum on to an eqmlateral. But in a boys'
class this may not be always thought possible. I shouldn't agree
with that view-or else the boys may have so bttle idea of the
subject that you can't quite begm this way Then I should
teach them by gomg through the steps of the discovery myself :
getting them, when they understand it, to help by finding or
suggesting the next step after you have given them the start.
The boys will get much sooner a real interest m the subject m
this way. Moreover every proposition will be to them a 'rider',
and a class taught in this way will find ' riders ' far easier than
those taught in the ordmary way Boys who can follow and
reproduce rightly and intelbgently demonstrations in a given
 FAMILIAR LETTERS cJv
text-book are often very bad at riders and are afraid of the
'problems' in the Geometry paper, because their minds have
never been accustomed to originate thinking, but only to under·
stand the given argument ready worked-out. In the case of
those theorems which require a construction, it is the construe•
tion which is the important thing and the text-book, such as
Euclid, n8Ve1' shews how the construction is found.
It is absolutely necessary that the teacher should try to show
how the construction may be arrived at. The whole difficulty
in many riders depends on finding the right construction. No
wonder then the boy as ordinarily taught finds difficulty, because
he has never learned from the text-book how the construction
1s found in the various theorems. He is accustomed to have
the hard part done for him, that 1s to say to have the construc•
tion given to him, whereas he ought to see how every step in
1t 1s suggested by a consideration of the problem.
The real difficulty, as regards time, 1s that it means so much
extra work for the teacher. You must yourself study the given
demonstration beforehand and try to see how the discovery of
the various steps m it is suggested. This is not always easy :
but I can make a suggestion which may be of use. A theorem
may be quite evident m some special symmetrical case and this
may suggest what to do, when the symmetry is disturbed. For
instance, I suspect that the theorem of Euclid 1 47 (Pythagoras'
theorem of the square on the hypotenuse) was first suggested
by the case where the right-angled triangle 1s isosceles. This is
the case worked out m the Meno (of Plato).
As a matter of fact a theorem and its construction are not
infrequently discovered without looking for them, m the course
of some other investigation. So that, m such a case, it is not
true that the construction was discovered by considering the
needs of the problem, because the problem itself had not occurred.
Some other problem was before us and, in making our con•
structions and seeing what followed from them, we observe
something that part of our construction necessitates and then
this can be turned into a separate problem, the construction for
it being what we have accidentally discovered. I daresay if you
find yourself sometimes hard up to see how a construction or
demonstration was discovered, I could help you.
clvi FAMILIAR LETTERS
As you reahse, the work of the teacher is at first increased,
but the class will make far greater progress, though they may
have, say, only done half a book of Euclid while another has
done two books. But your class wtll know and wdl not easily
forget, whereas the other may have to go over the ground again.
One very important and usually neglected thmg is to have
the figure drawn on the board as correctly as possible. Let the
straight lmc be ruled, and the circle described by a pair of large
wooden compasses, \\-luch you can easily get. When I was
a boy, I first got to understand Geometry through a teacher
who used wooden compasses. I remember I first realised what
an angle meant and what equal angles were, when he drew an
angle by ruhng along ms1dc the legs of the compass-the com·
pass bemg put flat on the (black)board-and then carrymg the
compass to another part of the board and makmg an equal angle
by rulmg ms1de the legs of the compass agam, the mchnatlon
of the legs not havmg been varied When I went to a higher
school and there wasn't such apparatus, I made a pair of wooden
compasses and used 1t in class with the approval and apprec1a•
t1on of my class master
In teaclung Algebra and Arithmetic you should treat the
' book work ' m a s1m1lar manner, letting the boys thmk 1t out
as far as possible themselves and only helping them when
they fad.
A second thing I am very anxious about 1s that the teaching
of elementary mechanics, mcludmg hydrostatics, should begm
as soon as possible In many schools the boys are kept working
at Geometry, Algebra and Trigonometry long before they begm
Mechanics As soon as a boy knows the elements of Tngo·
nometry he 1s able to do Mechanics and should begm at once.
The more abstract studies arc so wearying, and the apphcat1on
of Geometry and Trigonometry to Mechanics gives them quite
a new interest-the sooner therefore Mechanics are introduced
the better. The ordinary practice, which I have described, tends
to kill interest in the subject It 1s as 1£ you kept a boy grmdmg
always at Grammar and exercises m Latin, until he had got
through all the exercise books, before allowing him to translate
a Latin author and giving him that mterest.
There is another very important point and that is the teaching
 F AKILIAR LETTERS clvii
as regards the use of signs of direction, i. e. plus ( +) and (minus)
(- ), used as signs of direction in Trigonometry and afterwards
m algebraical Geometry, or analytical Geometry as it is called.
But now I admit this is a very difficult matter, because mathe-
maticians, even the best books, are in the dark about it. The
usual thing to say 1s that it is a ' convention ' to represent
direction by the signs plus ( +) and minus (- ). This 1s sheer
nonsense. You could never get the mathematical results (and
they are marvellous) out of a mere convention. If it were mere
convention one might ask why should not the signs of multi-
pbcabon ( x ) and division (+) be used as signs of direction ?
And to this question the ' conventioners' have no answer. It
is possible however that you mayn't have to teach Trigonometry
and we may let the sleeping dog he for the present.
In Algebra and Arithmetic you should take every trouble to
make boys understand proportion and mcommensurability.
Nixon's 'Euchd revised' has a capital treatment of the incom·
mensurabihty of the side and diagonal of the square-Euclid's
own, if I remember right, but I haven't the book before me--
which makes the mcommensurability clear without the help of
algebra or square roots. The only books I have ever seen
myself which treat the elements in the right kmd of way are
those by De Morgan There is an Anthmetic and, 1f I remember
rightly, an Algebra. As regards Algebra, a title I am sure of
is De Morgan's Trigonometry and Double Algebra, but that is too
advanced. 1 think there is an elementary Algebra I do not
always agree with him, especially m the Double Algebra (which
wouldn't concern you), but he goes the right way about it and
is suggestive and sbmulatmg [1t]. For instance, his remarks on
Concrete and Abstract number, upon the meamng of the minus
sign m the answer to a problem (even though this latter may
not be altogether right).
Another thmg you will have to be careful about 1s the meanmg
of an answer with ,/ =- I, m an algebraical problem : and here
I am afraid you would find no help whatever from mathematical
books. Suppose the problem is what is the value of x when it
satisfies certain conditions, the boy must be taught the difference
between the answers x =o and x = ,/--::::X. You see I have sent
you a fairly lengthy reply instead of ' merely mentioning
clv1ii FAMILIAR LETTERS
literature', but the reason 1s that I know of no satisfactory
literature on the subJect.... With kind regards, yrs. truly.
[The Elemetds of Ardhmetie •, A. De Morgan, 1835, 5th ed. 1848 (see
especially the Prefa~. London (Taylor & Walton)
T/,e Elements of Algebra, A De Morgan, 1835.
Tngonomet,,y and Double Alg,bra, A De Morgan, 1849
The Conntnon of Number and Magndude, A De Morgan, 1836, 18 probably
the book to wluch the letter refers for the d1stmct1on between abstraet number
and tlungs counted or repeated
De Morgan's Differential and Intelf'al Calculus, 1842, 1s the forerunner of
certa.Jn modem books which endeavour to make the Calculus 1ntelbgible to
a reasomng mmd, not gifted mathematically It 1s still worth perusal by
plulosoplucal studentci ]
63
South View House, Shap,
Westmoreland. 18 Aug. 1910
My friend 1 recommends as about the best guide to modern
method in teaching mathematics A school course of Mathematics,
by David Mair. Clarendon Press
He says that Mair has done more to improve mathematical
teachmg than anyone else in England He (Mair) is the senior
exammer of the C1v1l Service Commission From what he says
I thmk Mair's book would give you very valuable suggestions.
My friend thmks the books by Hall and Knight are probably
the best among ordmary text-books. These authors sometimes
wrote Jomtly and sometimes with other authors. But he says
he won't venture to recommend them as sound m principles,
though they have much merit as simple guides I should thmk
probably you would have to have Mair for your own guidance
and possibly have to give the boys Hall and Knight's books.
By the WclY, when you do teach geometry and are trying to
interest the boys m the ' thmkmg ' method, you would probably
gain their confidence 1f you let them have the usual proofs of
the Pons Asinorum and then gave them, after they had studied
these, the simple demonstration without construction which
I think I gave to you. Of course I would Jet them try them-
selves first before shewmg them the usual proofs, some boy
might well hit upon the proof I gave you. Yours truly.
PS. I regret I have mislaid your letter, so this must go
round by Umversity College.
1
[The late Mr. J. W. Russell, lecturer of Balhol College ]
 FAMILIAR LETTERS clix

64
7th May, 1912.
You formerly spoke feelingly of my aspersions upon the intro-
duction to Post. Analytics, 1 1. I, and, as these were only made
in lecture, I concluded you had become acquainted with them
either by seeing a copy of my notes or hearmg of 1t from a pupil.
My reasons were stated in lecture thus : ' The passage profes_ses
to sketch the subject but it 1s a mere analysis of the first chapter
though it has one line to the effect that the u,c/1/m is 'llEp'i.
411'oouf,v and 1s of ,m<TT~JJ.71 a,roou,cTmj, and there is the
farther difficulty that J. d.. is properly the subject of the Posterior
Analytics' I added that the Post. An. began with a general
statement, hke the Ethics, which wasn't a sketch of the subject
and that that was the Aristotelian manner. It has occurred to
me since that this kmd of thing is not likely to have been done
by a Peripatetic philosopher. He wd know that i,r. clvoo, was
the proper subject of the Post. An. No reason why he should
go out of his way to make that kmd of mistake. On the other
hand when wntmg Pr. An mtrod0 Aristotle may well not as yet
have conceived the special log. treatment of science as in Post. An
In fact m wntmg on syllog demonstration a:. opposed to mere
dialectic he thought he was writmg on Ell'. a.ll'oo. Afterwards
he felt the Pr. An. was too general and somethmg more special
required on the sciences and so he wrote the Post An -and
as to the analysis bemg of the 1st ch: 1t 's hkely enough Arist.
might have written so, mtendmg, after the part to which the
analysis referred, to go on with another introduction to the next
part, but as a matter of fact he went straight on without doing
1t. This 1s the kmd of mistake an author might make, but not
at all so likely a mistake for an mterpolator who wanted to
write an introduction. He'd be likely to give something general
covering the whole treatise. The origin of my note I on Tim. 37 c
was this :-Somebody, who had before him Archer-Hind's
foolish note on the place, prepared a juggling emendation in
which i.olwv was somehow got out of ,holc,w, really supposing
the interpr0 of the kind A. H. had in view. The editors of the
periodical in which the article with the emendation was offered
1 A abp for Pnor Analytics, ste the author's words, p. 31 • See p. lxxai.
clx F AM:ILIAR LETTERS
sought advice. J. A. S. told me of it and I wrote this paper
which the editors communicated (typewritten) to the would-be
emender. Afterwards I sent my paper to the Journal of Philo•
logy. I enclose my article from Cl Journ ... Yours truly
j. COOK WILSON.
In same no. of Cl. Journal (April) <J of P. )xiii, p. I 36)
Henry Jackson has come a cropper in trymg to emend punctua-
tion of Soph. 244 c.
65
New College. 31st May 1912 .
• . . By the way as I was gomg at a fair pace to-day, bemg late
for lecture, on my bicycle I passed M. I column and I heard
one of the men r1dmg at the head call out ' There 's that old
man agam I' How smoothly the waters close over us!
Yours truly.
66
12 Fyfield Road 2nd June 1912 .
. . • the matt<-r of the accouut of the pleasures of knowledge
m the Republic. Plato describes these as mcident to the filling
of the soul, leavmg one m the difficulty that they then seem to
be mixed pleasures at least and perhaps wholly pleasures of
relief of pam The difficulty 1s directly met m the Philebus, for
Plato there expressly lays down that the pleasures of learning
are not the removal of (the) pam of want, because the absence
of knowledge 1s not felt as a pam otherwise, as Plato says
acutely, we should not forget (or we shd be conscious of the
pam of the losing of knowledge, etc ) ... [he contmues about the
Sophist and the Parmenides] for 1t consists m what I confess 1t
has taken me some time to arrive at, trymg to appreciate the
quite plam meaning of the dialogues apart from the pre-
suppos1t1ons wJ11ch encumber us so much m our study of
him....
67
15 Dec 1912. Oxford.
I regret that owing to the great demands on my time through
the disquieting and distressmg change in my wife'b condition
{she has been getting worse smce beginnmg of August con•
 FAMILIAR LETTERS clxi
tinuously) that I haven't time to write to you as I should like
about the Arist8 question. But I think I can in a few words
perhaps comfort you. (I) The IT'tp&v ,., 811 business is very
fully dealt with m my ordinary Aristotle lectures. (2) It is not
really relevant to the particular question before me-the con·
fusion about the use of the word predicate in both ancient and
modern times. Whereas the remarks in Post. An. (are) relevant.
You would realrsc this if you could see the long discussion I have
on the distinction of subject and predicate, which perhaps I may
some day put before you. Indeed I mtend to do so. (3) The
kmd of diff. Aristotle gets mto (as I suppose) on the occasion
when he had ttme (?) to reflect on the nature of predication was
important to me, ~nd he's not the least excused by the fact
that he wanted to meet a special difficulty. It's a bad look-out
1f he had to make a lot of mistakes to do 1t (4) There may
be a 'higher synthesis' which JUsttfies a circular defimtion, but
this is JUSt the place at any rate where 1t won't do (5) Don't
misunderstand these ' extracts of Logic ' I They are merely
printed because, owmg to a dra~t1c rewntmg of my lectures
under great difficulties, I can't give the men anything hke
a complete course without prmtmg pieces for them to read
which I shan't have ttme to lecture on I send them to you
because some of them are likely (I imagine) to interest you.
I thmk, e g , what I have written on' Modahty' and on Bradley's
' Ultimate Reahty ' might be of the kmd I intend this vacn.
lo prmt a good btt more (e g. on subj and pred. in the smtence
of the form A 1s B) and on the negative Judgement .... Yrs truly.

68
I2Fyfield Road, Oxford.
22 May 1913.
The enclosed contam extracts from my lectures which I was
obliged to omit 111 delivery. I make a pomt of letting my
hearers have a fairly full course, covering the usual main topics
and latterly I have helped myse\i out in. this way. Don't ti:oub\e
to study it. I shall never ask you even 1{ you nave rcaa' any
of it. If you have any leisure and interest you might look at
the §§ on the Categorical and Hypothetical propos1ttons as these
clxii FAMILIAR LETTERS
contain a covert polemic against Bradley. If you have any
further curiosity you might look at the §§ on the Negative
conceptian. Yours truly.
69
New College. 27 Dec. 1913.
I send you of my Logic Extracts Pt. II §§ 26-60 (with excep•
tion of § 57). I thmk I sent you §§ 61 seqq. before (didn't I)
but a better print. Part one 1s m the press and will follow.
Yrs. truly.
70
12Fyficld Road. Oxford.
Oct. 22, 1913.
I was greatly pleased to get your letter. I have a very vivid
and pleasant recollection of you. I am much mterested m your
present candidature for a lectureship in Otago University.
I resolved to write you a testimonial though you did not ask
for it. But soon after I had a request for one through Mr. Wylie
and have sent it through him. Your degree m Greats is quite
good enough because you were considerably handicapped at the
start, and I daresay with the same advantages as people here
you would have secured a' First' comfortably
It was very considerate of you not to call m January but the
fact is that m that month at the begmning my wife had a sudden
and decisive change for the better. I thmk you rather overdid
it m courteous consideration, for when a man is going so far one
can always spare a little time for him. I regret much I have
no photograph I could send you. Elliott & Fry photographed
me some few years ago, but I was far from well and the result
1s most melancholly [szc] I shouldn't hke anybody to have it.
But I must think of having some made and will remember you.
However I don't think it will be yet awhile, for unfortunately
either in consequence of over-exercise in the autumn-grubbing
up trees, the hardest work I ever did-, or of an unrecognised
influenza attack, the doctor doesn't know which, I have got,
the doctor says, a weakness of the heart (I suppose so-called
'athlete's heart') which he takes so seriously as to prescribe as
much rest as possible for the present. He allows my morning
lectures but I have not been permitted informal instruction this
 FAMILIAR LETTERS clxili
week. The last years I have had 2 delightful classes, one of
men and one of women. The women were good, and one .got
a ut and another a 2nd. This year also I have a good women's
class and the new men's class seems promising.
This Summer was the best holiday we have had for years-
in Westmoreland near the scene of the Aisgill railway disaster.
The air did my wife an extraordinary amount of good. The
• squires ' m the neighbourhood were extraordinarily hospitable
to us. One of them housed my motor-car m a lordly stable.
I have acquired a car and the art of dnvmg it and am glad that
my wife was able to go about with me, though the country was
too hilly for her. She never quite hked coming down a steep
hill though I always went slowly. The car of course contributed
much to the pleasure of our holiday. I have a very good friend
in N.Z., Bevan Brown, Headmaster of the Christchurch School,
and Professor Macmillan Brown. With kindest regards from
Mrs. Wilson and myself, Yours truly.

71
I2 Fyfield Rd. Oxford.
November 9, 1914.
To L1EuT.-Cot. H. D. FARQUHARSON.
DEAR SIR,
The enclosed letter from my friend your brother Spenser
must serve as my mtroduct1on and commendation to your kind
services.
In view of the difficulties caused by submarines and floating
mines I have some suggestions which I should hke to put before
the naval experts of the Admiralty. I want to secure that they
shall be attended to merely, and if they seem of no use, I don't
want the Office to trouble to do more than send me a line to
say that this is so. I had intended to send them now, but in
the meantime an urgent matter has come to my knowledge
which I must write about at once.
From an absolutely reliable Danish source I hear it is believed
in Denmark that the German fleet is only being held back until
certain new guns are ready which will out-range our guns by
a matter of two miles. If this is so, it accounts for Admiral
Tirpitz's boast that he would surprise the British Admiralty
about this very date.
clxiv FAMILIAR LETTERS
I daresay the Admiralty know all about this but it is welt
to make sure. Yours truly,
J. CooK WILSON.
PS. My suggestions 1 will follow in another letter.

72
MILITARY CYCLING
To the Editor of 'The Times'.*
SIR,-The letter of mme on military cycling which you were
good enough to publish has not only evoked a very kind notice
from the Manchester Guardian but has brought me a com•
munication from a d1stingu1shed cyclist officer which I trust the
War Office may thmk worthy of very serious attention. It
seems that we have no cyclist battalions with our Army in
Belgium, although the country in which it is operating is specially
suited to the operations of cyclist forces. Now we have at home
quite a considerable number of cyclist battalions, smart, well
trained, and enthusiastic, and each with its machine-gun detach•
ment. Their ranks, I beheve, are full, and their reserves, in
some cases at least, of equal number. It seems that these are
being kept in this country to perform a service which could be
rendered by much less valuable troops, to their exceedmg regret.
If the commanding officers of these battalions are anything like
as good soldiers as the one who wrote to me, we might expect
quite remarkable results 1£ they were given their chance in the
present war.
Such men naturally burn to be at the front, where I have no
doubt they would give a brilliant demonstration of the special
value of cyclists as fighting troops. I may repeat that I hope
the War Office may give this matter their earnest consideration.
I have the honour to be yours, &c ,
J. COOK WILSON.
1 They were sent on the 10th Nov They concerned (1) the destruction of
submannes, (11) the closmg to submarines of a passage through a maned area,
and (in) the destruction of contact mine&
• Pubbshed II Nov 1914. The article in 1.hf' Manchester Guard,an wa.'I
under date the 30th Oct 1914, and referred to a letter from Wilson in The
Times of the 29th October The Manchtslef' Guardian said, 'All honour to
those who have the capacity to thank for themselves an a science which of
all others most needs onganabty and independence of thought, the courage
to give their View'! to the world and the pat1'11ce to wait for thE'lr acceptance.•
 STATEMENT AND INFERENCE

21731 B
 ANALYSIS OF CONTENTS
PART I. INTRODUCTORY
I. THE DEFINITION OF LOGIC PAG'B

§ I. Logic, hke the sciences, begins with special problems, not with
a general definition of its nature and province 24
§ 2, One scientific problem suggesting another, cognate questions get
grouped together Reflection on such organic groups is a new
land of thmkmg 25
§ 3. This generahzmg tendency is due neither to logiccu thought nor
to the impulse towards definition Definition is the search for
a universal to cover a whole subJect and thus to permit its
differentiation . , 26
§ 4 Coherent problems do not necessarily belong to the same science.
Ancillary sciences 27
§ 5. The true nature of abstraction, a gradual realization of a common
universal . 27
§ 6 The ideal of definition. Definition and classification cannot even
m a priori ,;c1ences precede their development Logic 1s no
exception 28
§ 7 No defirutton of logic nor explicit recognition of it as a special
department m Anstotle's works 30
§ 8 Modem attempts to define logic Prov1Siona.lly it may be called
some kmd of study of thought m d1stmctton from a study of
things 32

II. THE RELATION OF KNOWING TO THINKING

§ g, Thmkmg embraces : (1) knowmg, which is neither perception nor
feeling, e g. the apprehension of universals and of relations ;
(11) the formation of opinions, wondering and deliberating,
which are, none of them, activities of knowledge 34
§ 10 Thinking, mall its formci, is based on knowing It 1s understood
through itself and knowing. To ask for a definition of either
is to ask an unreal question. The genus consciousness and its
sp,ci,s knowing arc sui generis • 37
f 11. Nominal and real definition. Definitions of terms must begm
from their appllcat1on in usage • 40
I 12. Socrates' attempt to define moral notions. Ha audience must
have known moral ideas m one sense. but not m another 42
5 13 Appbcat1on to the d1stmcbon between perceiving and thinking.
Pecubar nature of the activity of comparison in sensation and
perception 4S
B2
4 TABLE OF CONTENTS
III. I.OGIC AND ITS COGNATE STUDIES PAGE
Sr 4.Ordinary usage distingu1Shes logic from metaphysic.,; and gram-
mar. The d1stmct1on bes within the province of thinking 1tseli ,48
f 15 The method of logic 1s determined by its subJect-matter. Logical
analysis makes explicit the universal forms implicit m all
thmkmg . 48
§ 16. The sense in which the method and form of logic are a pnon 49
§ 17 Grammar 1s not of words and logic of thought. Both study the
forms of thought so
§ 18. Twofold ac;pect of all thmkmg as (1) event, (11) meanmg Psycho-
logy embraces the whole of consc1ousnec;s and considers
especially the event aspect and time-order of thmkmg, logic
considers the meanmg within its own narrower field Physio-
logical psychology SI
f 19 Time and time-order m logic The question how m fact a thought
occurs is psychological 54
§ 20 The sciencec; study the ob1cctJve, logic the sub1ect1ve side of
thmkmg Metaphysics studies both and their relation, in a way
of its own. Theories of reality are metaphysical, not logical • 54
§ 2r Logic includes in its Vtew subjects bke eptstemology as ancillary
to itself. The bnk between 1t and cognate subJects 1s the
desire to mvestlgate truth • 56
§ 22 The false distinction between pure and applied logic arose from
the notion that syllog1shc was a general form of pure reasonmg 57
§ 2 3 General and special logic may be distinguished, provided we do
not suppose that general logic can be developed m abstraction
from the reasoning actuabzed in the sciences Logic ebc1ts
the forms of the apprehe11S1on of obJects, not conceptions
which are of the nature of the obJect (e.g. Cause) • 58

IV. LOGIC AND THEORIES OF KNOWLEDGE AND REALITY

§ 24 Can the definition or logic as a study of thought be mamtamed
m the face of varieties of 1deal1sm ~ 6o
§ 25 By accepting the vtcw of ordmary realtsm we do not escape these
difficulties 61
§ 26 What IS meant by saying that obJccts are u•itlnn consciousness ? 62
§ 27. The obJect as content of thought. lmphcation of everyday speech
about thmkmg and thought. The 1dentificat1on of what we
thmk with thought • • • 63
§ zS. The d1stmction between the form and the matter of thought is
ultunately ideal1sbc • • • 64
§ 29 What 1s apprehended cannot be abstracted from the apprehension
of it. The analogy of graspmg an obJect • . . 66
I 30. Ordmary usage llllphes an 1denttfication of what we thmk with
thought and mtends to include our subJect1ve act of appre-
hension . • 68
f 3 r. Substances and relations. Modern metaphysics tends to destroy
the distmct1on Necessity of a d1stmctton of thought and what
lS thought from interrelated ob1ects Bodies to be related must
be d1sttngu1shable from the relations between them 10
 TABLE OF CONTENTS 5
l'AG:S
f 3:z. Objects apprehended need not be part of the nature of the
apprehension of them, nor part of the apprehendmg conscious-
ness. The dec1S10n a.s to whether they a.re or a.re not rests upon
an exa.mma.tion of the objects themselves 74
f 33. Attempts to explain apprehension by theories of knowledge and
reality are futile 7S

PART II. STATEMENT AND ITS RELATION

TO THINKING AND APPREHENSION
I. APPREHENSION IN GENERAL
§ 34 Appreheru.1on 1!> the &tartmg-pom\. of logic The di&tmctton
between what lb apprehended and its apprehem,10n lb indepen-
dent of conflictmg theories of reality 78
§ 35. Essential difference between bc1cntific .i.nd plulosopluc thmkmg
Dawn of the logical con'ictou">neb'> 79
§ 36 The development of the logical com,ciou1me1:,1, m the tracl!ttonal
logic , the study of inference 81
§ 37 Ancient logu.. start1:, from the proposition , the 1,yllog1sm is not
the outcome of grammatical analysi1:, but anse!> from argument
1n deb.i.te. 84
§ 38. Modem log1cal tre.i.tt1:,e1:, 1:,tart from Judgement. Confns1on1:, m the
use of this term. Judgeme11t 11, dec1S1on reached from doubt.
Judgements of perception arc falsely bO termed 84
§ 39. Ongm of the d1stmctton between the theories of Judgement and
of mference Modem logic mcludes under Judgement both
what 1s and what 1s not the re&ult of mference 85
§ 40 True Judgement mcludcs mfernng The d1stmctton of an activity
called Judgement from mferencc lb fictitious 86
§ 41. It anses from a fal&c abstraction Knowledge, op1111011 and belief
a.re erroneously grouped under the title of Judgement 87
§ 42 The modem logic of Judgement 11:, c1. logic of statement It con-
fuses true logical forms with grammatical and mct..i.phys1cal
A m1Staken way of d1stmg111shmg Judgement from mfcrence b8
§ \3, Our study of apprehension mcludeb what arc trad1tionally called
the theones of Judgement and of Inference. Grammatical and
metaphysical mqu1ncs are also involved 90

II THE TERM ' JUDGEMENT' IN MODERN LOGIC

§ 44. Judgement does not properly mclude every mental attitude
which may be expressed by a simple statement 92
545. Is th1nkmg to be included m or excluded from percept.ion ? 94
§ 46. The relation of opinion to Judgement Mearung of the form A U1
probably B • 95
§ 47. The umty of the activities of thmklllg lb not to be found m
a common umvcrsa.l of which they are bpec1es 97
6 TABLE OF CONTENTS
III. OPINION, CONVICTION, BELIEF AND COGNATE
STATES
PAGB
I t8. The identity of the form of statement is no guide to the attitudes
of mind of which 1t 1s the verbal correspondent. They must
be themselves exammed 98
§ 49. Knowledge and Opinion 99
§ 50. Opinion and Behef 100
§ 51. Practical decision and the feehng of confidence which accom-
panies opinion and behef 102
§ 52. D1fficulbec; involved in the idea that Judgements may be false . 104
§ 53 Alternative solutions of the problem of false Judgement, both
untenable 108
§ 54. States of consciousness which simulate Judgement or opinion
Meaning of the phrase ' I was under the 11npression that ' 109

IV. SUBJECT AND PREDICATE IN LOGIC AND IN

GRAMMAR
§ 55 Ambiguity of the traditional defimbon'l of sub1ect and predicate
m logic 114
f 56 Unambiguou'l btdkmcnt of tht• prmc1plc undcrlymg the trddi-
bonal d<'fimhons I l 'l
§ 57. Enghsh distmguishes the predicate by the •predicative' stress
accent 119
§ 58. Logical analy&is of the statement mto sub1ect and predicate IS not
verbal but depends upon meanmg The grammatical tenns
should be • sub1cct words ' (nommatlves) and • predicative
words' 122
§ 59 The relation of the distmct1011 of sub1ect and predicate to gram-
matical forms 124

V. ANALYSIS OF THE PROPOSITION INTO

SUBJECT AND PREDICATE
f 6o. The trdditional form S 1s P and the difficulties 1t mvolves The
ordmary theory contradicts its own ddimtion 121
§ 61 The form A is B ongmated m Anstotle'& analy&is of the syllogi&m
and m his treatment of the premiss as the standard form of
proposition 129
§ 62. Ar1Stotle's study of inference began from the Socratic search for
definitions and m the mvesbgabon of philosophical and rhe-
torical debate , not 1n mathematical or grammatical analyslb 131
163 Probable origm m the syllogism of the i.ymbohc andly&1s of the
standard form of sentence, with its principal verb the verb
of bemg • 1 34

VICONFUSION OF PREDICATION WITH OBJECTIVE

RELATIONS
f64 The identification of the sub1ect and predicate with S and P in
the symbohc analyi11s Criticism of this modem symbobsm
of the proposition 138
 TABLE OF CONTENTS "I
PAGB
f 65. Objective and 1ubJeetlve elements in the chatinction of 1ubJect
and predicate • 139
f 66. The theory of the syllogism confuses predication with the state-
ment of obJective relations. The words sub1ect and predicate
are otiose m syllogistic theory 141
§ 67 Confusion of predication with the relation between the universal
and the particular 145
§ 68, The incorrect use of S 1S P as the clue to the analysis of proposi-
tions 1'1 due to Anstotle The resultant error 1s to express an
ob1ect1ve relation m phraseology applicable only to a subJect1ve
distinction Ordinary language has no des1gnat1on for the
ob1ective relation m question 147

VII OBJECTIVE DISTINCTIONS AND DOCTRINES OF

PREDICATION
§69 Necessity of investigating grammatical form., Relation of
grammatical to so-called metaphysical forms 149
170. Attributes d1stmgu1shcd from (a) subJect and (b) subbtance m
regard to part1culan, 151
§ 71. 'lhe 1:,ubJect and attribute relation 1s an m-.lan<.e of that umty
of d1vers1ty i.ecn in tlungs and the elements of things Things
.i.nd their relations 155
§7.2 Aristotle's doctrine of predication He neglects the d1stint--
tion between obJective and subJective with resultant confusion 159
§73 A modem doctrine as to the true subJect of every existential
Judgement being the ultimate Reahty • 166

VIII THE MEANING OF GRAMMATICAL FORMS

§ 74 The obJecbve meaning of the forms of the sentence Expression
of attnbutes and relations . I 70
§ 75 Nouns denoting attributes m abstraction from any subJect 173
§ 76. Double function of the verb a.,; indicating (a) assertion, (b) part
of the natu1e of what 1s asserted lt'l temporal function 175
§ 77 • To be • 1S no mere copula but expresses attributive bcin~ Use
of its present tense in abstractmn from time reference Identi-
fying statements 181
§ 78. Verbal expre1,1:,1on of umven.al statements C.rammat1lal forms
of existing language express particulars of universals 187

IX THE SYMBOLIZATION OF FORMS OF

STATEMENT
f 79 Grammatical sub1ect and predicate and subJect of attribution
Early language restncts the latter to substances . 192
I So Can a statement be without a logical subJect ? 196
f 81. Improved symbolism to represent the form of statement w1tb
• to be • for its prmc1pal verb Symbols for the analys19 into
sub)eCt of attr1but10n and attnbutivc , 197
 TABLE OF CONTENTS
l'AGB
f Sa. Symbolism to express sentences whose nominative 18 an abstract
noun denoting a universal. Ordllla.ry lmguistic usage faithfully
mirrors a true metaphysic of universals , 2o6
I 83. General proof that every affirmative statement can be resolved
mto a form of the type all (some, &c.) A 1s B, or S 11, P , 209

X THE COPULA AND MODALITY

584. The meanmg of the ho-called copula Aristotle's doctnne.
Medieval ongm of the notion that the copula 1s a sign of
predication 212
§ 85 Statementb wluch seem to deny every lond of bemg to a subject 214
§ 86 An ancient view that the copula expresses identity Identity
and difference m the statement 217
§ 87 The possible identity of the so-called sub1ect and predicate The
true concC'pbon of a mathematical equation 219
§ 88. Common terms Anstotehan gc11era and species A d11,bncbon
m the Categories between• of a subJect' aud '111 a sub1ect' 2.io
I 89. Traditional c!ai,1:,ificat1011 of forms of :,tatemC'nt The modality
of i.t.itemenb 223
§ go. Modd.lity .ind the ob1e1..hve meamng of ' to be ' 224
§91. Mod.ii d1.,t111ctions are bub1echve z25
§ 9.i. The affed1onb of the copula involve aftmn.itlons about our own
state of mind 2z S
§ 93 Modahty m respect oi time Bcmg m general 11, wider tban any
temporal di~tmctlon~ Special idiom of the verb • to be • 226
§ 94 Abbreviated forms of expression (impersonal verb<;) and t.hc
relation of logical analysis to commands, wishe& and mterroga-
tions 228

XI SY:!IITHETICAL A11.D J\}.ALYTICAL STATEMEN1S.

THE REL.\11O:N OF STATEMEN1S
§ 95 An apparent JUbtlh1...iilon of the analytical &tatC'ment Erroneous
theory that every synthetieal statement can become analytic • 231
S96 Another apparent Justification of the analytical statement. The
use of formulas m mathematical demonstration . • 233
§ 97 General i.olutlon An identical statement is impossible and so,
stnctly, is an analytical 234
§ 98. Ca.tegoncal and hypothetical statements dlbtlngu1shed Can the
true predicate be not absolutely affirmed of the true sub1ect ?
l:>upposed reduction of the categorical to the hypothetical form
of statement 235
§ 99 The uwversal categorical btatement cannot be reduced to the
form of c1.n bypothebcal :,t.atement and retam its full meawng • 236
 TABLE OF CONTENTS 9
PA.GB
t roo. Absurd attempt to reduce the categorical singular statement to
the hypothetical form 238
§ 101 The hypothetical statement depends on the categoncal form and
1s not a d1Stinct form co-ordinate with 1t • 239
§ 102. The hypothetical statement does not affirm a relation Proper
meaning of •if•. The protasis refers to 1eal clements but
whether they stand in a certain relat10n 1s uncertain. In what
sense the consequent 1::. the result of the antecedent. R1tl11ctao
cul absurdum proofs The hypothetical statement 1S an infer-
ence ma non-hypothetical argument • 241
§ 103. All statement 1s categorical Its hypothetical form states abso-
lutely a conncx.ion between two problems about reality. Its
non-hypothetical form states a~olutely a connexion between
realities Strictly the hypothetical form 1S not denvable from
the non-hypothetical . 243
§ 104 An ambiguity m the form m which a condition 11, ex.pressed.
Thi::. ambiguity lb present m the c,i:act demonstrations of
mathematical ::.c1ence 243
§ 105, 'lhe hypothetical statement cannot formally be c..onvcrted 111mply
bec..au::.e of the ambiguoub ei.presb1on of the condition Accurate
::.cient1f11., hypothetic..al ::.tatemcnt::. arc ::.miply convertible 245

XII NEGATION OR THE QUALITY OF STATEMENTS

§ m6. In vrct'k philosophy the d1fficulhe::. as to negation were meta-
physical Modern logical difficuliieb relate to the form a::,
bUCh 247
§ 107. An apparent reducbon of the negative biatement lo dll affirmative
form 249
§ 108. Can we entertain a conception of not-A-nes::. and does any true
uwverbal correbpond to 1t? Error m making not-A-ness a uni-
versal. Rei.ulting mfimte regress 251
§ 109 Not-A-ne&& regarded from the side of apprchch&1011 InverH1on ol
the relation between defimilon and d1St1nction 255
§ 110 The negative i,tatement ab ::.uch cannot convey a positive deter-
mination of 1t~ nomtnabve 258
§ 1r r. Negation and afunnat1on arc mexphcable 1 he connexion of the
negative statement with (1) advance 111 knowledge and (..?) the
affirmative form The general form of statement embraung
both positive and negative forms , 26o
§ 112. The d1Sttnct1on of sub1ect and predicate m the general form of.
statement and 1n the negative statement • ..?b..?
§ 113 The two forms involve one another but are not co-ordinate 264
§ 114 The ob1ective meamng of the negative statement 265
§ 115 Statements that seem to assert complete non-e'tlStence , 266
§ 116 Imaginary quantities m mathematics Their true nature 268
I n7. The natural expression of negative thmlang The apprehension
of negation is an apprehenSion of two pos1t1ve reahties as
cWfenng • 271
IO TABLE OF CONTENTS

XIII. ERRONEOUS ATTEMPTS TO DEFINE ' JUDGEMENT'

PAGB
I 118. The statement 1s not a combmat10n of ideas 274
I 119. Combination of ideas m regard to truth and falsehood The
mental picture. Ideas wlnch are combined regarded as con•
ceptions • 276
f 120. Thought and its obJects. Apprehensions without anythmg appre-
hended are empty 277
1121 We cannot combine our conceptions or apprehensions, we can
apprehend a combmation . 278
I 1.2.2. A modem theory of the ideal element in' Judgement'. Erroneous
analysis of sign, symbol and meaning Refemng=Judgmg
The ideal content 1s m no sense ideal • 279
§ 123 The theory rests on the same pnnc1ple as the copymg idea theory
but with add1tlonal comphcabons and confusions • 285
I 1.24 Signs, symbols and meanmg in thJS modem theory of 'Judge-
ment' .288
§ 1.25 Signs or symbols are c-onvent1onal and arbitrary True uses of
symbols : (1) words for thoughts, (11) (a) algebraic, &c symbols,
(b) thmkmg by l'.Ords Mental images are not symbols of
umvcrsals, we 1magme a parbcul:.r 291

XIV APPREHENSION, CONCEPTION, AND STATEMENT

§ 126. Ancient and modern views of statement as a combination of
conceptions or ideas 295
§ 127 Two mam uses of the word conception m English 299
§ 128. Justification of the ordinary d1stmct1on of conception from
op1mon, Judgement, &c 301
§ 129. Conception does not precede Judgement. Conception as used m
mductive processes and m the assoc1atlon of ideas 303
I 130. Second mam use. The conception of duty, of force The' con-
cepts ' of science 304
§ 131 The mental 'wnteut • of concept10ns a figment due to the
copymg idea theory Second sense of conception rests upon
a confus10n 306
§ 132. Neither ordmary linguistic usage nor sc1entlhc usage 1s con-
ceptual1stic. 'Concept' m modern logic as a substitute for
'term' • ~
I 133 The elements of statement and of the corresponding thought
Are these elements pnor to the wholes to which they belong? 310
§ 134. The d1stmction m the statement of knowledge 1s not between
conception and Judgement but between incomplete and com-
plete apprehension Simple sensation (apprehension) 15 neces-
sarily an element ma wider apprehension 312
f 135 The elements of apprehension per se have no reality except as
elements m the obJect. No real thought corresponds to the
verbal expreSS1on of the sunple elements The nominative m
a statement. Confusions as to the' formation' of conceptions 314
 TABLE OF CONTENTS II
PAGB
t r 36. Conception in statements of opinion and belief. The problematic
conception. A relation may be perfectly simple • , 318
I 137. Can a simple conception be an imagination? Imagination and
expenence. In any statement the s1D1ple conception must
be true 319
I J 38. Differences between imagination and expenence. The unagina-
tion of a past sensation 18 not identifiable with a present
sensation of the same kind 320
f 139. Further confirmatory considerations, Psychological theory of
memory Association 324
f 140 The simple conception is the apprehension of the blmple obJect
Psychological figment once more 326
§ 141. No active apprehension of the entirely simple is possible Anger
is not a rudunentary emotion but argues developed mind.
The dawn of the apprehending consciousness 327
XV. THE QUANTITY OF STATEMENTS AND THE
UNIVERSAL
§ 142. The distinction of uruversal, particular and singular statements • 330
§ 143 The nature of a universal is not affected by the number of its
particulars 333
§ 144. The universal, called Ubually, but. ma.i.curately, the univen.al
conception 335
§ 145. The universal 1s present m all statement 336
§ 146. Thmkmg and perceivmg 337
S 147 The quality, charactenst1c or mtnns1c being of the universal 340
§ 148 The unity of the universal The universal cannot be explained
1n terms of something other than itself 344
§ 149 Umversal-ness incorrectly descnbed as itself a universal 348
§ 150. Fallac1es about the u111vcrsals of number m modem mathematical
pseudo-metaphysics • 352
XVI CLASSIFICATION
§ 151. ProbleIDS connected with the universal Conceptualism, re.i.bsm
and nommahsm 354
§ 152, Error of enunciating logical rules for d1v1b1011 or describmg its
fallacies • 354
§ 153. Classification results naturally from experience. Its study appears
very early 1n the h1Story of logic 3 55
§ 154. Genus, species and d1/fn'enl1a Summa gene,a and 1njimfllJ specaes 356
§ 155 Extension and mtension of the universal. Extension said to
vary inversely with mteDSion 356
§ 156. Inadequacy and arb1trarmess of the d1Sbnct1on of genus and
diffe,ent,a, If based upon the above view • • . , 351
I I 57. Generic and differential elements m figures and colours They have
two chara.ctenstics (1) the d1ffwndia 15 not added to the
genus ; (u) the relation of genus and different1a is not reciprocal 3 58
§ 158. True notion of genus in relation to sp8c18s ; the proper meaI11ng
of different.a • • • • 359
§ 159. Extension of Arutotle's formula of potential and actual m regard
to genus and species to the relation of species and mdlVIdual , 360
 TABLE OF CONTENTS
PAGS
I 16o. Two cliHerent lands of class universals • 361
I 161. Origin of the logical theory of div1S1on. TIU'ee cases which have
to be considered when dividing a class 361
I 16.2. Distinction of spe"es The fumlamenlum divmonss and the ideas
of co-ordmab"n and subordlllat1on 363
§ 163. Meaning of oppo1,ibon and contradiction of spe,ies Some species
arc differentiated in a fixed ordff 364
5 164. Order is nece~i,ary to brmg about either absolute or relative
oppos1t10n of spe&ies Not every class adD11ts of these oppo51•
t1om, 366
I 165 Dichotomy 367
I 166 True division 1s by co-ordmate positive dsfferm#iae Logical
abstraction D1V1Sion 1s achieved by the sciences, not by logic.
Division m the empirical sciences is never exhaustive 367
§ 167 Kmd and quality d1st111guished from quantity and degree 36g
§ 168, Can a difference of degree amount to a difference m kind ?
Changes of curvature and of temperature 369
§ 109 Natural lands m Anstotle Modem ideas of a na#ural classlficab.on 371
f 170 Meaning of a natural principle of division. In the empirical
sciences 1t cannot be more than tentative 372
§ 171 Mill's doctrine of real kmds His confusion 373
§ 17J. Two kmds of umve~als (1) defimte, (11) problemabl, Retur11
to Aristotle 37 S

XVII DE:FlNlTION
§ 173 Dehmt1on 8b (a) bearch for the esi,ence, (b) proceedmg byget1u~
and dijferent,a • 377
§ 174 The object of defimt1on. Artifi.ciality'IOf Anstotle'b acwu11t 378
§ 175. Definition applied to moral notions 379
§ 176. Definition 8b of esi,ence m opposition to property Anstotle's
account of essence ,• 380
§ 177 Two grave difficultie& mvolved in the idea of defi.mbon of essence 381
§ 178. Div1S1on and defi.mtlon The relation of definition by essence to
defimbon by genus ,1.nd dsffiwen#ia 382
5 179 By definition, bC1ence gaim1 both m cleames~ ,1.nd 111 extent.
True scienbhc propositions are convertible lhe search for
analogous elements arises from the di!:ttmcbon of genus and
d1Jferent,a 'lhe progress of science 383
I 18o The reason why science prefers one kmd of clas!:t1fi.ca.t1on to
another 384

XVIII. DENOTATION AND CONNOTATION

•
I 181 Method to be pursued in tlus kmd of mqu1ry . . 386
I 18.2 Is the distinction betMlall &onnote and denote leg1tunate ? 381
I 183 Necessity of first mvestigating the facts of lmguistic 11Sage 388
I 184. The meanmg of calling nouns and adjectives• names of' 388
§ 185 Concrete general nouns hke man 389
f 186. Abstract nouns like we•g/rl, vsrlue • • • 389
I 187. In what sense are adJecb.ves termed 'names' ? 390
 TABLE OF CONTENTS I3
PA.OB
I 188. On• signify• and• mean• 1n relation to the words in a sentence 390
f 189. Direct and indirect meaning • 392
I 190. False distinction of pnmary and secondary meaning 392
I 191. Implication and llleanmg 393
f 19.2. C8,J1 an abstract word hke heavmess mean the substance which
has the attribute ? 394
I 193. The term• denote' is adequate to express the relation descnbed
by an abstract word . 396
5 194, The meaning of adJectives m the sentence 3g6
§ 195 What 1s meant by meaning ? • • 391
§ 196 Grammatical function of the ad1ecbve 398
§ 197 Grammatical function of the abstract noun 398
§ 198 Concrete general names (nounr,) Their function m the sentence 399
§ 199 The 111vest1gation shown to be of grammatical forms and not
truly logical 400
§ 200. Denotation supposed to denote 1.he members oI a class and
connotation to imply their common characteristics as such
members • 401
§ 201 Confusions of the traditional logic Concepts and terms . 401
§ 202. Mill's treatment of • to denote • and • to connote • , 402
1203 The normal meaning of del)ote is abused m these theories • 403
f 204. Futility of the inquiry about the denotation and connotation of
proper names
I 205. The difficulty which is concealed under this whole question is the
senous one of the relation of the apprehension of the umversal
to that of the partJcular
S2o6 Mistoncal sketch of the use of the word connotare m Ockham and
other scholastic wnters •
§ 207. Emergence of the modern difficulty m Thomas of Strassburg
The d1ffic11Jty 1s implicit in Ockham

VOLUME II
PART III. INFERENCE
I. THE GENERAL NATURE OF INFERENCE
I 208 Inference regarded as the study ot certain ways of reaching
op1mons and Judgements Concl•n and proof are words of
ordinary speech, not technical logic:tJ terms • 412
f 209. Inference d1Stmgu1Shed from e:r.perl.ence. Ongmatlve and recep-
tive activity. Denvation of one fact-from other facts. MntJ
inference and direct apprehens10n • . • 413
I 210. Dlssatzsfact1on m regard to inferred laiowledge, Intr1ns1c and
e:xtnns1c evidence There may be different lands of inference 414
§ 21·1. Provisional account of inference Difficulties in the Vle'WII of
formal logicians Pure thought, imagination and experience • 41 s
I 212. The d1Stinction between prem1'ls and conclu'11on Novelty of the
conclusion • 416
14 TABLE OF CONTENTS
PAGB
1213. Mediate and immecba.te inference • • 417
1214 All inference is strictly speaking 11nmedia.te • 418
1215. Limitation of the idea of immediate inference in the ordmary
syllogistic logic. Form and matter of inference , 419
I a 16 Im.mediate mfereru:e Relation of the conclusion to the premiss 42 J
f 217. The mconceivab1hty of the opposite as a cntenon of truth • 424
f 218. Reality of the advance in knowledge made in an immediate
inference 42 S
f 219 Certain proces,es which seem hke a repetition of the premJ.SSeS
but which simulate the verbal form of inference 42 S
I 220. Distinctions prerequisite to a theory of inference 426
f 22 J. Solution of the prccecbng d1fliculbes 427
1222. Inference in general is the apprehension of one fact as necessitated
by a different fact or facts • 429
§ 223 Immediate inference and the syllogism The conclus10n of an
inference, whether syllogi'!bC or not, is always immediate • 430
§ 224 The relation of apptthen,1on to proof 431
I 225 An instance of apparent dtfforonce between conclusion and
premisse'I . , 43 I
f 226 Sub1ectiv1ty of the idea of inference a.,; a connexion of 'Judge-
ments• 431
f 227. Prov1S1onal recogrut1on of the universal character of inference • 433

II SYLLOGISM
§ 228. Problem of the ~yllogisbc logic 435
S229. Limitation m the form of the solution proposed by the syllog1Stic
logic Two particular premisses can give a conclusion. Un-
d1str1buted middle 435
§ 230. The method of syllog1sbc logic 1s a priori and not analytical. It
1S not a branch of logic but a science hke mathematics • , 436
§ 211, The syllog1Sbc method does not proceed by' pure' thinking alone 437
§ 23z Formal logic 1s not an cwganon. The syllog1sbc 'rules• are
nugatory , • 438
t 233. A further bm1tabon of the syllogistic logic It 1s unsuitable m
form for quite simple arguments • . 438
§ 234 The syllogistic form is not the general form of demonstrative
reasoning. • •
I 23.5 The syllog1stic form 1s a species co-ordinate with other forms of
inference . • 441
1236. The middle term Syllogistic reasoning does not succeed in
abstracting all matter from its argument • • 441
f 237. The fallacious reduction of arguments to a syllog1'!bC form The
rule (ax1om) of syllogism fallaciously made a premiss . • 443
f 238. Relation of the conclusion of an mference from two premisses to
either premiss taken smgly 446
f 239 Apphcat1on of the preceding cons1derabons to the syllogism • 441
I 240. Relation of the conclusion to both premisses taken together.
The cohefttlce theory of truth • •, 449
124t. The reduction of one form of argument to another. Anstotehan
reduct10n. • 452
 TABLE OF CONTENTS :r:5
III. PRINCIPLES AND METHOD OF PURE DEMON-
STRATIVE SCIENCE
PAGB
5242. Attempted use of the syllogism m demonstrative science. Geo-
metry is not syllogistic m method 454
5 243. Actual procedure m geometncal reasoning 455
§ 244 The imperfection of the geometncal figure. A new d1stmction in
our apprehendmg faculty • 456
§ 245. The apparent place of the syllogism in geometrical reasomng 457
§ 246 Geometrical advance 11l1Srepresented as the subsumpbon of a
particular case under a universal The true process is the
d1Scovery of the umversal m the particular instance 458
§ 247. The syllogism of proof 1s in fact only recognized by reason of the
analytical process of discovery. Memory of by-gone proof 459
f 248, Application of the preceding investigation to a.II demonstrative
syllogisms. Extension of Kant's principle m regard to mathe-
matical inference • 461
I 249. The particular <1yllogism regarded as a genera.l1zat1on of some-
thing apprehended m the particular instance 462
S2 50. The application of the universal to the particular 464
f 251, The fir,t principles of geometry, The postulates are theoretic
constructions 464
§ 252, The axioms are incorrectly d1stmguished from demonstrations.
Both are based upon mtu1bon 465
I 253 Nommal and non-nominal definition m Euclid's geometry 466
I 254 M1Sconceptlon-; m regard to the nature of defin1t1on m science • 467
f 255. False d1Stmcttons between conceptions and Judgements that such
conceptions arc vahd for their obJects • • 468
f .l56 The difference between defimbon m mathematic~ and 1n the
empmcal sciences 46g
f 257 Essence and property Aristotle's d1stmction between them • 470
f 258 Essence m mathematical science 471
f 259. Rec1proc1ty of e~sence and property Condition and condit10ned.
Direct and md1rect proof in geometry A thmg may be both
self-evident and demonstrable 47 3
f 26o. Algebra and the science of pure quantity depend on perception
or 1magmabon for their reasoning 476
§ 261. The three uses of symbols or perceptual units. The processes
symbolized are not merely analytical • 477

IV. THE UNIVERSAL CHARACTER OF INFERENCE AND

ITS RELATION TO IMMEDIATE APPREHENSION
I 262. The umversal is necessary to every inference • 479
I 263. Inference as dependent and indirect knowledge contrasted with
knowledge which 1s not inferred 481
S26-4, Expenence as direct apprehension further contrasted with infer-
ence. Circumstantlal eVJdence Inference in the emp1ncal
sciences • 482
16 TABLE OF CONTENTS
PAGJ!:
I 26.5.Inference 1n the demonstrative sciences 482
f 266. Illustrations of geometrical mference, both mediate and im-
mediate . 483
I 267, Further consideration of the nature of med1at1on 488
I 268. Ax10matic and md1rect reasonmg Self-evidence, direct and
axiomatic character, of the conclusion of an inference Some
mathematical axioms are rules of inference expressed differ-
ently Direct knowledge not snpcnor to med1ated, but CO•
ordinate with 1t 48q

V Sll\lPLE AND COMPLEX IDE\', AND PROBLEM.\.TIC

CONCEPTIONS
§ 269 Vanches of hng111st1c form to expn•qs categorical, hypothetical
and cond1bonal stat<-ment'I 491
§ 270 The relation between the hypothetical and catc>goncal form.,
Two mistakes about the relation Suggested classification 492
§ 271. D1fferenct's of meanmg between the cond1bonal and causal forms
The dt'IJuncttve statement 493
§ 272, Hypothetical form'! and hypothetu,al meamng A cond1ttonal
statement may or may not be hypothetical m actual mcamng
or mteutiou 4g4
§ 273. Logical reflection results m artificial and fallac1ouc; forms of
expression 495
§ 274. The element of hypothesis m human thmkmg 1c; neither fict1bous
nor wholly uncertain The element of fiction m statements • 497
§ 275 Complex • idea,; ' may be a.'I impossible to create as &1mple ideas 497
§ 276 Simple ideas or conceptions Locke's distinction of ideas into
simple and complex Relations Modes 498
§ 277. The meaning of simpl1C1ty and complexity when applied to ideas.
Equality and umty 500
§ 278 Manifoldness ch~tmguished from complexity Genus, species,
uruversal, particular 502
§ 279 The attribute of complexity must be d1stmgu1shed from the
complex
§ 28o Idea and obJt>Ct Ob1ective nature of the d1stmcbon of three
. . 504
from threene'ls 506
1281 The attnbute of complexity itself may or may not be complex • 506
§ 282 A thing is not complex becaurie it necessitates a plurality other
than itself A+tnbutes are not necessanly complex 507
I 283. D1stmction between three kinds of complex conception, (a) full
and proper, (b) indefinite and (c) problematic 508
§ 284. The doctrine that simple ideas cannot be fictions Correspon-
dence of the mental image with its obJect m experience 511
§ 285. Some ideas (e g substance) are certainly not given by • experi-
ence•. Failure of empmcal doctnnes, like the association of
ideas, to explain such ideas • 512
I 286 Other ideas wluch are, in a sense, not given in experience.
Equality, straightness, &c Plato's doctrine of remmiscence • 513
 TABLE OF CONTENTS r7
PAGB
§ 287, Distinction between ideas like •cause• and ideas hke •equality•.
Neither cause nor time 1s empirically given. Subjects
(=substances) are not given Origin of empincISm and natural
meaning of expenence, Authonty of our apprehension m
regard to the obJects of expenence sr S
§ 288. Can sunple ideas be made ? Ob1ect 11Msus imagination. In what
sense a fiction of the imag111ation is unreal. No statement
contains :fictitious ideas s18
§ 289 Possibility of fiction m complex ideas Geometrical construc,
bons Problematic conceptions have the verbal form of con-
ception only. They are neither made nor fictitious 521
§ 290 A complex conception 1s the apprehension of a complex ob1ect
umtmg simple elements, not a complex of the apprehensions
of such clements. • Question-conceptions • m the science, 523

VI HYPOTHETICAL STATEMENT AND HYPOTHETICAL

ARGUMENT
§ 291 Problematic conception is one condition of the hypothetical
element Belief and probab1bty Mistaken belief does not
create a conception 525
§ 292 The accurate form of expresq1on of the • 1£' clauc;e Imphc1t
reference to the certain and the real . 526
§ 293. Accurate expression of the problematic statement, (1) where the
problematic conception is, and (11) where it 1~ not, believed to
be realized ~26
§ 294 The difficulty of the '1f' clause m hypothetical statement 527
§ 295 Supposition and assumption. \Ve cannot avoid the meamng
of' 1£' 530
§ 296. The hypothetical statement m an mferentral process Modus
ponens and tollens The words ' supposed •, • assumed •,
'imagined•. 'pretend' and 'ficbhou'I' 532
§ 297 The hypothetical statement and its correspondent mental attitude
cannot be replaced by the non-hypothetical statement 536
~ 298 The hypothetical statement 1s an mference 536
§ 299 The real character of the thmkmg which corresponds to hypo-
thetical expression. The explanation o{ the Jallacia const1-
quentis 537
§ 300. Aspect of the hypothetical attitude which accounts for the verbal
fonn of the hypothetical statement 539
§ 3or Two forms of statement, not two statements, are contemplated
as altemabve forms of the answer to a problem • 541
§ 302. Imagmabon m problematic mqu1ry Ch1maeras 543
§ 303. Nature of the hypothetical argument, modus tollens 544
§ 304. An • 1f • clause refernng to something which 1s not m fact uncertain 54 s
§ 305. D1s1unctive statements are (a) non-problematic and (b) pro•
blematic • 546
§ 306. DlSJunctive statement 1s categoncal. It states a relation between
two forms of statement or between two problems 547
§ 307. Hypothetical statements are 1mphcit m the problematic d1s-
1unct1ve statement 548
111773•1 C
18 TABLE OF CONTENTS
P.AGB
I 308. Non-problema.tic conclusions arising from the combina.tion of
a. dlsJunctive statement with a.ssertonc prem18Se9 • , 548
I 309, The so-called reduction of hypothetical a.nd disjunctive argument
to non-hypothetical and non-dLBJunctive forms Anstotle'a
nght m,;tmct in the matter 549
5 310 How best to represent the complete reasoning m hypothetical
argument It J'I a case of relative, not syllogistic argument • 550
§ 311 Conclu'lion'I m regard to hypothetical reasoning The true
character of the assertions made • SSI

VII DIRECT AND INDIRECT ARGUMENT MODERN

MATHEMATICO-PHILOSOPHICAL FICTIONS
I 312. Linguistic u'lage 'As'lume' and ''luppose' belong to the
practical, not the theoretical, sphere • 553
§ 313 Correct form of a retluctio ad absurdum proof 555
§ 314 The hypothetical form of every such proof can be re-placed by
a non-hypothetical proof Illu,;trabons from propos1t10ns m
Euclid (1. 7 and 19) The principle of contradiction as a prellllSII 555
f 315 General proof of the contention as to the replacement of indirect
by dire-ct proof . 557
§ 316 True function and importance of the ad absurdum fonn In non-
hypothetical proofi. the reasons for the constructions do not
appear The hypothetical process combines both method and
demonstration Analys1,; and synthesis 559
§ 317 Origm and character of systems of non-Euchdean space • 561
§ 318. ConfuSJons of thought which underhe the theory of these systems.
The idea that a problematic conception has become real 563
§ 3 r9. This pseudo-geometry uses no conception of space other than
Euchd's . 565
§ 320. Fallacy committed m the attempt to Justify these systeIILS by
analogy or correspondence • 567
I 321. Fallacies of one-d1mcns1on and two-d1mellS1on spaces in abst,acto 567

VIII. PROBABILITY AND ITS MATHEMATICAL

EXPRESSION
§ 322 No degrees in knowledge or reality Bebe£ bes between Judge-
ment proper and ignorance Possible and probable 569
§ 323. Analysis of belief Enu-ralio simple1t. Number of 1DStances
affects human vac1llat1on . • 570
§ 324 Pos,;1ble and probable m reference to a smgle fact or event 571
I 325 Number of alte-mabves hm1ted Mathematical definition of
probability The meaning of • equally hkely ' 1s purely sub-
jective . . 57 3
f 326 Mathematical analysis provide'! a standard for the more and less
m evidence. The modulus of eVJdence • • 575
f 327 L1m1ts of the application of the modulus. Testimony. Statistics
Fallacious tendency to refer to reality the formula of our own
ignorance. 575
 TABLE OF CONTENTS

PART IV. SPECIAL LOGIC

I. INDUCTION. THE GENERAL FORM OF THE
EXPERIMENTAL METHODS
PAGE
§ 328. Origin of induction. Induction has been regarded (1) as a move-
ment from particulars to a general statement or (u) as argument
from the conditioned to its cond1t1on • 578
§ 329 Different londs of induction. Anstotehan • mducbon ' and the
axioms of science and of philosophy Bacon's attack on the
old mducbon 58o
I 330. Inductio ,Pn mumnationem simplicem 583
§ 33r, Presuppositions of en simp argument 584
§ 332, The value of a numbn of instances. ~86
§ 333, Further defect m the ordmary analysis of en simp argument • 587
§ 334 The expenmental methods Difference 588
§ 335. The isolation of mstances 589
§ 336, Agreement • S90
§ n1, Concomitant variations Bacon•~ Tabula graduu,n 59r
§ 338 The general form of which the expenmental method'! are species ~93

II FORMAL CAUSE IN BACON. DEFECTS OF THE

MODERN THEORY OF INDUCTION
§ 339 The ehmmabvl.' methods discover the fact and not the reason.
Hes1tatmg use of the teleological idea in B10logy . 5()6
§ 340 Importance of causation m the modems The problem limited
unnecessanly S97
§ 341, Bacon's conception of the formal cause umque 597
§ 142. Is the form of heat merely a definition of heat? 599
§ 343 Forma and simplex natura 6oo
§ 344. Bacon's 1rremed1able confu~1ons m his a< count of the Formal
cause 6o2
§ 145 The naturii seems to be human &en&abon m1'1takE"nly tran.~terred
to its ob1ect (,03

III. FAILURE OF INDUCTION IN THE PRESENCE

OF ITS OWN AXIOMS
§ 3-46 The old induction retains a place in the new 6o6
§ 347 The axioms of the new induction repose upon en simp proof&
Nature of these proofs 607
§ 348. The exact form of the en simp proof of the prmc1ple of causat10n
or of umfomuty. Its effect on the modem theory (>09
§ 3-49 The relation between the conclusions drawn from the old un-
sc1E"nt1fic- method and those based on the new SCienbfic induc-
tion 610
I 35<> Reciprocity of condition and conclittoned The plurahty of causes 611
C2
20 TABLE OF CONTENTS
PAGB
I 3.51. The conditioning and the conditioned are necessanly reciprocal 611
5 3.52. The appearance of a plurality of conditions is illusory 613
f 353. Smular illusion in the appearance of a plurality of causes • 614

JV FAILURE OF EMPIRICISM TO EXPLAIN

THE LAWS OF THOUGHT
§ 354 The association oI idea.c; Herbert Spencer•~ b10log1cal e~planation 616
§ 355 Snppoc;ed reconc1habo1i between empmc1sm and transcen-
dentahc;m 616
§ 156 F.volubomst doctrine of .i.x1u111~ Ax1oml:I a pr,011 to the indi-
vidual, a posteriori to the race 617
§ 357 Relation of the theory to human desire,; 619
§ 158 The theory propoc;e,; to account not only for the ongrns but for
thl' vahd1ty of ax1om:1.br ideas 621
§ 359 The criterion 1s negative and the re,;ult of mtm1tely long experience 622
§ ,oo The theory is sclf-contrad1ctory from a b10logical standpoint
It is consistent with its own contradictory and endc; m uncon-
scious scepticism 62 3
§ 161 The Circulus in probando on wlue,h the do<-trme rests 625
§ 162 Thought is made to tcc;t the vahdity of itc; own prcsuppoc;itiom, 626
§ 363 The cntenon 1c; prepostt>rous A general rule ec;tabhshed by its
employment m a special science 627
§ 364 Evolution m ,cience d1stmguished from evolut1omst phllO'lophy 627
§ 365 The doctnne of mnate ideas Locke's 1gnorat10 elent:lu Leib-
mz' reply Cartec;1an d1stmcbo11 of mmd from maHrr Kant's
solution 628
§ 366 The subJect1Hty of Kant's answer 1 rue and obJective solution 630

V THE METHOD OF PHYSICS

§ 167 The sc1encec; fal~ely railed apphed mathematics 631
§ 368 The non-mathematical element m the physical science& 632
§ 36g A mistaken view of the function of deduction m mathematical
physic, • 6ri:
§ 370 A twofold d1&t111cbo11 m the field oI mathem.i.tie,.i.l physics.
Statics and Dynamics contrao;ted with the sciences of Heat,
Light and Sound Physical and psychical correspondence 633

VI SYMBOLIC LOGIC
§ 371 Symbolic logic conducts formal inferences by the aid of an
algebraic calculus • 635
§ 372 Though tlus logic pretends to be mathematical 1t must be studied
by the logician • • 636
§ -~73 The method of logic cannot be mathematical, because mathe-
matics 1s a special science and the methods of the sciences are
a part of logic's object matter 636
 TABLE or CONTENTS
l'AG:S
I 374, TJus logic liOlves partu:ular scientdic problems and ignores the
real problems al logic 637
§ 375 Its general problem JS bllllphc1ty 1Wllf, 1ts 1,pecral problems may
be 111m11tely 111tr1cate b37
§ J76 Ongin of the &tudy Analogy of Iormd.l log1c to the science of
pure quantity Unw1,e choice of algebnuc symbols by thu,
logic C139
§ 377. The unllkehhood of the symbol!, oI one bClt.nce bemg &uited to
another. The danger that th1~ logic may confound terms w1th
units of quantity 640
§ 378 A and B representing classe&, AB &tand~ for what lb both A .and
B. MeanmgofA+BandA-B (4o
§ 379 Multlphcation of the ,;ymbols Boole's fundamental rule of
thought~=~ • 641
§ 38o. Symbolb of operation 1n mc1.illematiu, and m logic Un1ty
unfortunately chosen to -,ymbohze Rf'al1ty. Any empty
clasb =o • <>42
~ 381. If x 1 =not x, x+.!-'= 1 lb the fnndament"-1 cquct.bon Equations
of dichotomy 043
§ 38.1 A= B to be mterpreted oI the l.lasse-, m exten'i1on, not. m mtcru,1on 043
§ 383 '!he vital quest.ion for the calculu& 1s to expre&s all A 11, B m an
equation. The form A=vB, (where v=' -,ome ') 11, .i. bogu'i
equation 644
§ 384 Boolt.'s equation X\: '=o depends upon conceivmg all X 11, Y d.b
hypothetical, viz as conveymg no defirutc statement ~ to the
existence of X lhls fundamental equation i'i an 1l1U&1on 644
§ 385 Elimination 1n the calculu& Particular propo&1tion1, are sc1.1d to
unply the ex1Stence of their &ubJeCtb, raibe equivalence of
All X 1s Y to No XY 1 e'l:1-,tq lhe true contradictory of All
~~y 045
§ 386 D~proo{ by red ad abs of the above equ1vcllence <>47
§ 387 lhe true mferenl.e from AU X 1'1 Y 1'> all XY 1 1., YY 1, but lhlb
g1veb no equation 648
I 388. 1he attempted mathemat.u,al mtcrpretatmn of XY•=o Equa-
tion of 1mpos:,1bleb to zero • 649
I 389 x+ a-1 = I fal5ely mc1.kes reahty ml.lude unrcal1tlcs (1) All 1mpo&•
s1b1htles are equated to the same and therefore to one c1.nother,
(11) each unposs1b1hty 1s equated to zero 649
I 390. (1) and (11) above would be fallacies 1ll algebra lmagmary
quantities 1n modem mathematic'! The equation of 1mpos-
s1bles or 1mc1.gU1c1nes to one another 111volveb the equation of
c1.ll real quantities to one another 650
§ 391 Illu&tration from geometry. The equation of 1mpos1>1bleb mvolves
the equation of all realities to one another . 651
§ 392. Reasons why tlus paradonccl.l equating 111 adopted • 652
I 393 By the rules of the calculus we can show by red ad abs that
the equation of all unpob&lbles mvolves the equation of all real1, 653
It
§ 394. Meanmg m the calcalus of the mverse operation y . Proof by
the calculus that a real class JS part of unreality , • 654
I 395. All clilbliCs morcove1 can be.. proved to be pred1c..c1.ble oI one d.llothcr 656
 TABLE OF CONTENTS
PAGS
f 396, True c:oncluaions may be drawn from fal.'le prenussea , 656
I 397, Why the trr0neous representation of Barbara JD the calcul\18 GOii
not influence the conclu&1on • • • 657
I 398. The equation of dichotomy. The cond1tlonal statement which
the equatJou includes lb employed and the erroneous add1t1on
which 1s necessary to make an equation 1S ignored 658
1399 The equational element 1S m fact never used m this logic, The
truth of its conclusions depends wholly upon the element of
predication 659
f 400. Distinction between categorical and hypothetical btatement.
Imposs1bilibes arc but the reverse of Necesi.ities That every
frue statement would involve every other must not be con-
sidered a fatal obJection to the calculus 659
Notes to Chapter 6 • 6oo

PART V. TENTATIVE INVESTIGATIONS AND

PHILOSOPHICAL CORRESPONDENCE
I. ff 401-.105 The nature of a • thing ' 664
II. ff 406--414 The umversal and its ddlerenbabons The md1-
v1dual 670
III H 415-421. Modern fallacies about the umversal 677
IV. §§ 422-429 Classdicatio11 of classes 683
V H 430-432 Predication predicated of itself 68g
VI ff 433-437 Relation and quality • 6g.1
Vil. H 438-450. Categories in Aristotle and Kant Note 6g6
VIII H 451-459 Umversali. List wordi. 707
IX §§460-476 Trichotomy (corresponnence with Henry Goudy) 715
X H 477-500 Correspondence with Bernard Bosanquet 7.18
XI II 501-518 On the notions of a class and of cla&Ses (c.orrei.pon-
dence with Professor G. F Stout) 752
XII. II 519-540 On primary and secondary qualities 764
XIII. ff 541-552 Correspondence with Mr. Harold A Pnchard Sot
XIV. ff 553-563 Later correspondence between J C. Wilson and
B Bosanquet 818
XV. I 564 D1vmc and human colll!e1ousness • 83o
XVI ff 565-58.2 Rational groundl> ol Belief in God, with a letter to
l'rofeb&Or C C J Webb • • 835
 PART I
INTRODUCTORY
 I
THE DEFINITION OF LOGIC•

§ I. A c.LNER<\L dcfi111t10n of the &cope and nature of any

subJect ,~ u&uc1lly expected to precede the cxpos1t1on of the
subJect itself So Logic, to be systematic, might naturally be
required to open with a clear defimhon of itself, from which its
province would be determined and the order of the subjects
within that provmcc Y ct there is considerable disagreement
about the investigations to be included under the term logic,
and writers who agree m the mam about the sub3ccts to be
studied will still differ m their general conception of the whole,
that 1s as to what the proper definition of logic may be.
It 1s a stnkmg fact that there should be any doubt at all and
with the consider.i.t1on of tlus fact we must begm First, we
observe that the sciences m general arc provided with accepted
dcfimt1ons 1 especially the exact sciences In the case of a quite
new science or study, mdccd, we should not be surpn&ed 1f there
were some doubt as to its defimt1on ; but logic 1!> not a new
!>Ludy, 1t 1s probably nearly as old as sc1ent1fic geometry.
Secondly, it might perh,tp& &eem obvious that the logician could
not even begm lns subJcct unless he knew its dcfimt1011.
These d1fficultic~ arc really founded on a misunderstanding
clllcl w.i.nl of obsen al10n of the actual state of thmgs 111 the
sciences thcmsclws, whether provided w1th an accepted defini-
tion or not In none of them did 1:,ncntifiL rcflcct1on bcgm from
a quite gcm'1,\l dcfi111t10n of the ob3cct of the given science, but
with particular problems b wluch the needs 01 interests of hfe
and experience 111 one way or another suggested For example,
some one had occasion for prachcal purposes to find the distance
[ 11 The rclat1011 of clasi,11ic~t1011 to dcfimt1on is discussed m Part II,
1,,h'>, XVI-XVll
b Probably i.uggcblcd m the firbt mi.tancc by Manscl's Aldrich', p lxxvi,
note g, and the quotations there from Plato, R 527a, and Comte We
1nay compare 'Artem 1nvcmend1 cum mventu, adolescere posse, statuere
<lebcmus' Bacon, t,., 0 1 130 J
 The Definition of Logic
of a vessel from a point of observation on the coast. This
purpose led in the end to the discovery of the group of pure
theorems and problems connected together under the title ' con•
gruent triangles ' • and to the study of these questions and those
of similarity of triangles as we find them in Euclid's Elements.1
Thus a given question in a science bke geometry does not
originate m any general conception, for example of space, nor
m any conscious desire to study what this 1mphes, for the
investigator may not yet have entertained either. His solution
depends upon the particular character of the problem within its
own limits. It is possible then to know what we want in one
way b but not to know 1t m another, and 1t 1s the first way which
is important in a given science, where the mvestigator knows
quite well what he wants, but may never have reflected whether
the question belongs to the science, say, of space or to the science
of pure quantity.
So far then, logic 1s not m an mfcnor condition as compared
with the sciences. It starts as they do with special problems
(the analys1s, for mstance, of the forms of argument) and, if it
has not yet, hke some sc1enccs, got a clear definition, in one
important sem,c at least 1t docs not need such a dcfimtion. For
neither arc the sciences themselves helped to the discovery of
their truths by a general defimtwn of the kmd of tllinkmg to
wluch their special problems belong
§ 2. Wlulc, howe,cr, investigation starts with particulJ.r
• Books I, II and VI
[• Wllson at lai,t came to giving Euclid, Element~, 1. 1, as his illustration
here I have substituted Elem 1 4, 26, &c , as historically more 1,a.tisfactory,
d Proclus 1n Eucl 1 352, and Sir T Heath, Thirteen Books of Euclid, &c.,
i 304-5. At one 1,tage, Wilson used one of the l>ythogorean problems con•
nected with transformation and apphca.t1on of areas, a subject which lS an
even better example of the pomt he is making but more difficult for the
non-mathematical reader My illustration 1s connected trad1tiona.lly with
Thales' determination of the distance from the shore of a ship 1n the offing.
b Marked m the MS ' to be expanded ' I have endeavoured to explain
by an addition what 1s meant The phrd.be recurs with a different beanng
m I u, of persons who know particulars but have not yet attended to their
uwversa.l impllcatlon The phrase 1s used by Ar1Stotle, m the sense mverae
0£ that m I u, m A,i Po 7rb 5 (cf. Eth Nie 1147• 6). Wllson's use here
lli, I think, an unconscious remuuscence of Tylor, Anlh,opology (1881), p. 318,
where, in connexion with the dawn of geometry, the author says ' the
Egypila.ns then knew and did not know about geometry '.J
 INTRODUCTORY
problems, thought of severally and not conceived as parts of a
whole, the solution of one problem will yet lead to another; some·
times because it needs that other for its own solution, sometimes
as suggesting that other. In this way there grows up something
like a systematic body of knowledge whose parts have organic
connexion. After a ttme this connexion of problems and
theorems with one another suggests the question whether there
is any one general conception which covers them all. Now that
1s a question which does not condition or ortgmate the activity
of the science and, accurately speakmg, does not belong to that
activity at all. On the contrary it presupposes the procedure
of the sc1eru:e as already ex1stmg and arises from a new kind of
thmkmg, i.e. not the thinking which constitutes the method
of the science but reflection on that method itself.
§ 3. Now there is a progress always towards unity ma science
itself-that is to say towc1.rds a unification of truths already
ascertained, which, though akm to the attempt to find a defim-
tton of the given science or department of thinking, is not quite
the same thmg. For instance, m geometry we may have a proof
of the properties of the circle based upon its definition, and an
accurate proof. Yet we feel that we understand better when
we find analogous properties m other curves and arrive at
.a. general conception of come sections, of which the circle is only
one. Then after recogmzmg this unity, inasmuch as all these
curves have the same kmd of property, we are led to a new
demonstration not confined to the circle alone, but valid for all
conic sections. Here we have not made the demonstration of
the given propcrlles of the circle any more certain, but we have
improved it. We have a greater insight now into the fact, when
we sec it as an mstancc of a more general prmciplc. And more
than that, we can even say more accurately why the circle has
this property. It has it not qua circle but because of the pro-
perties which it shares with other come sections. Here the
desire for unity actually improves the demonstrative process
within the science itself, but 1t is not due to a mere philosophic
reflection outside the science, nor does 1t aim at a definition of
the science. The impulse to find a definition is not directed to
finding more general theorems which will unify under them•
selves a number of special theorems ; it is rather the attempt
 The Defl"'tion of Logic
to find a universal which will cover the whole and to differentiate
it so as to be able to map out the various departments of the
science. Thus we say, for instance, geometry is about space,
algebra about pure quantity, and so, if the question is once asked
what the definition of logic 1s, we might suppose that all we
have to do is Just to observe the common characteristics which
cohere in one whole, in the subject of logic as it has grown up
historically.
§ 4. A system of theorems, however, which is coherent does
not necessarily fall under one and the same science. The
problems of one science may lead to those which belong properly
to another. Hence the fact that a given group of problems
coheres docs not guarantee that they all belong to the same
science, and therefore a generalization from their coherence might
be wrong. In his Elements it 1s necessary to Euclid for the
theory of proportion between Imes, areas and volumes, i.e.
between geometrical quantities, that there should be a discussion
of the general theory of proportion. This accordingly looks like
part of geometry and forms Book V. But though necessary to
geometry it 1s not a part of it, belonging really to the general
theory of quantity. And there are other books of Euclid's
Elements which also belong to this general theory, though their
subject 1s studied with direct reference to geometry.
The coherence, then, of a set of problems does not necessardy
show that they belong to the same science. Similarly in logic,
though a given problem which we may decide to be logical,
e.g. the validity of thought m relation to reality, is connected
necessarily with certain other problems, 1t may nevertheless be
true that those other problems belong to metaphysics, psycho-
logy or even to grammar. But that will not excuse us from
considering them, and we must not yield to the temptation of
avoiding a difficulty mcident to our mvestigation on the plea
that it is ' extra-logical'. The solution of a problem in one
science is often necessary to the solution of a problem in another.
Algebra and geometry are different sc1ences, but algebra is
useful to the geometrician.1
§ 5. There is a further difficulty which affects some subjects
more than others. Even when we have the right group to
1 Cf.§:n.
a8 INTllODUCTOllY
generalize from, i.e. a group really unified by one universal, tho
discovery of the common element in it is not always easy.
The traditional idea of abstraction • is here most misleading.
We get to think of abstraction as if 1t were merely leaving out
the elements in which a number of complexes dtff er, and retaining
the common clement. This presumes clear knowledge of the
elements m each complex and so the procedure is the simplest
possible. But the actual procedure 1s usually very different,
The elements are not as a rule before us m such clear analysis
and we often feel, as we say, that there 1s an affinity, but find
it hard to discover m what it actually consists. Take, for
instance, the Socratic search for defimt1ons. 1 Here the attempt
is made to find a umvcrsal to umfy a class of particulars,
especially the umversals of the various virtues. People d1s•
cussing the subject agree as to whether a given act 1s just or
unjust, and we don't find in this Socratic mvestigatton that the
real difficulty is about particulars. And yet the mvestigators
are liable to disagree as to what 1t 1s m Just acts which makes
them just and the discovery of the common element often
involves a considerable amount of argument and investigation.
Abstraction mdeed 1s not so much the picking out of one element
already recognized from a number of others already recognized,
but 1s usually a process m which the abstracted element is for the
first time commg into clear consciousness. This process is often
slow and the recogmtton of the true universal grows clearer and
clearer as our experience itself grows, or as the science which
is concerned itself progresses. The act of abstraction then, even
when we have the right matter to abstract from, may be difficult,
and it 1s difficult not only m the case of logic but m other
departments also of what is called philosophy.
§ 6. The definition of a science or of any study then does not
help to the solution of 1ts particular problems and, in the case
of logic, it 1s well to remember at the outset that the value of
our solution of a given question does not depend on our being
right or wrong in the general defimtion of the subject. Yet
thought aims at a certam completeness in the grasp of its
1 Cf. If 11, 12, 35, 166, and 175.

[• Derived from Ueberweg, Log,,;, I 51, and Lotze, Log,,;, t123, 121 and 1.57,
but the pomt goes back to Kant (Logu;, 1. 1, I 6 ct al.).]
 Ths Dlfinitlrw,. of Logic
objects which is not attainable without definition, We want to
know not only the solution of problems, or how one necessitates
another : we also seek for their general connexion under some
unifying universal or characteristic. Our impulse is not satisfied
till we get a conception, the articulation of which we suppose
will cover the whole of the science or department of study.
The ideal of the definition of an inquiry is that while it covers
the whole subJect it should be such as to indicate within itself,
by its own differentiations, the total divisions of the subject.
It may then, by its very generality, indicate special fields which
ought to be, but have not at present been, mvestigated. For,
when we have the universal, we may discover species implicit
in it which may not as yet have been approached in the course
of the actual solution of the problems of the science in question.
This ideal definition cannot be obtained in every science because
we cannot always differentiate a priori the genus given by the
definition, and so determine a priori the species and the depart•
men ts thereof. For these latter may depend upon experience
to reveal their subject-matter. Yet the more the empirical
sciences advance, the more they progress to the unification of
their theorems on the one hand and to their ideal definition on
the other. They endeavour to systematize the knowledge which
they have, though clearly such systematization must always be
provisional.
In the exact sciences, where the method is a priori (what this
precisely means we shall determine later), 1 it might seem at first
sight that, after definition of the whole, the determination of
the subordmate definitions ef the parts mdicated by the general
definition might be possible by simple successive differentiation
a priori. For m such sciences we ran differentiate universals
into their species a priori and know that our classification is
exhaustive. Thus we may divide geometry into plane and solid
geometry. Then the study of plane figures may be divided into
that of rectilinear figures, curvilinear figures, and the combina•
tion of the two, and so on.
Yet this kind of differentiation is but of a limited character.
The actual working of a science, in the solution of its particular
problems, leads to departments of investigation and so to
I II 16, 243.
30 INTttODUCTORY
principles of classification, which we can arrive at in no other
way. Hence, these important differentiations of the province
of the science cannot possibly precede its actual development,
and the field cannot be mapped out at the beginning, a priori,
u a mere consequence of the general definition of the science.
Thus it turns out that not only in the empirical but also in the
exact or a priori sciences, we have to wait for the progress of
the science, in the solution of its problems as they occur, in
order to effect the successive differentiations of the general
notion of the science and to articulate its departments. To
take a simple instance, the division of triangles mto right-angled,
acute-angled and obtuse-angled depends upon the proof of
Euclid, Elements, i. 32. One example among many in a higher
department is the discovery of the Cartesian method of co-
ordinates, which brought with it classifications of geometrical
subject-matter unsuspected before. The reason of this will
appear hereafter when we have seen what the method of the
mathematical sciences really is.
It is important therefore to observe that the differentiation
of the definition of an mquiry depends upon the particular
subject-matter and can therefore only be effected by scientific
thinking proper, not by philosophic or logical reflection. In
logic, too, we must expect that the differentiation of the depart•
ments to be investigated will depend mainly upon the develop•
ment and discussion of particular problems, and that 1t would
be as futile here as m science to try and formulate a complete
scheme of its subjects beforehand.
§ 7. Logic, as the history of Greek philosophy shows, is no
exception to the rule which has governed the development of
the sciences. It began without any definition of itself. Even
after logical inquiry had become more or less independent and
methodical, it remained without any such defi.mtion. In Plato's
case this is not surprising, for, though he made some contribu-
tions to logic proper, these are not advanced so far as to make
it obvious that they form a separate provmce of philosophical
investigation. Aristotle however has become conscious of the
separate nature of the subject, for he thinks of it as a whole
and has devoted special treatises to it. Yet he neither gives
a definition of it nor has he anywhere a discussion of the sphere
 The D,ji,MM of Logic
peculiar to logic, Of the inquiriet grouped together at the head
of his works and commonly styled the Organon, the Categoriu
has not even the pretence of an introduction. The Prior
Analyacs opens with a few lines which profess to sketch the
subject but, with the exception of one line, we find merely
a summary of the first chapter. The subject is said to be
• proof ',1 properly the subject of the Posterior Analytics, This
book (one of the author's most mature works) begins in truly
Aristotelian fashion with a general proposition, ' all teaching
and learning of an intellectual kind arises from previously
existent knowledge '. 11 This is not a definition of the subject
he is about to mvestigate ; it suggests however what is more
clearly expressed in the opening of the Prior Analytics, that his
main intention is to discuss inference. If so, the discussion of
propositions, for example, might have come in as subordinate
to the general purpose. The De lnterpretatione has a more
systematic opening than any other of these treatises. It dis•
tmguishes thought from things and language from thought.
But, though enumerating particular topics to be discussed, it
has no general account of the subject as a whole. Moreover, its
style is scholastic and it is a bare analysis of the early part only
of the De lnterpretatione itself. Nor is there in the body of the
Organon any attempt at a definition, and the only thing like
a general name for logic m Aristotle is Analytics ; 8 even that
is not intended by him to designate all the treatises in the
Organon.
We should naturally look for some definition of logic in
a passage of the De AnimtJ 4 where he classifies the various
forms of knowledge. But, though he is there speaking of philo•
sophy as well as of the sciences, he does not mention logic even
indirectly. We may conjecture a reason for this. He is thinking
of the processes of the attainment of truth and may very well
have thought of logic not as a process of this kind; that is,
not as a study by which we add to our positive knowledge but
rather as an analysis of the mental forms and processes which
1 D•,2 drilr,£11,, 1111E ,.,~IA'I' dwoh11rT1djr. See however my A nstotelian Sluds,s,

I•, p. 89, note 1, [Cf. Letter 64.]

I Diio'a 11.laO'IIIIAla #Cll ,rii,,11 palrµ,11 ll&1&•01JT&q l,r 1rpovr,apxa60',,r .,,...,,., , .__
Cf. the opening of E#II. Nie.
• Trl 11.vaAuTurd. • De Anuna, 1, ch, t.
3Z INTRODUCTORY
are common to various departments of knowledge and opinion.
Thus logic may not have come before his mind when he was
classifying the sciences, though it does not follow that he con•
sciously excluded it. This suggestion is in agreement with the
vague term Analytics, for that seems to imply not the discovery
of new truth but rather the analysis and examination of what
we have already. Beyond this the word tells us hardly anything.
The application of the term ' logic ' to the subject is attributed
to the Stoic,;.• The adjective from which the term is derived
has in Aristotle a very different sense. The absence of a defini-
tion in Aristotle did not escape the notice of his scholastic
interpreters. 1
§ 8. Aristotle then, who founded the study, offers no definition
of logic and, at the outset of our inquiry, it is safest to adopt
a broad description of the subject and not to attempt a more
precise or detailed definition Some writers, for example J. S.
Mill, make inference the main subject of logic. Yet in Mill's
work and in the ordinary text-books there are topics other than
inference and those not treated merely in relation to inference.
Hence we :find wider definitions. Whately, for instance, calls
logic ' the science and the art of reasoning ', and Mill, at one
place, ' the science of the operations of the understanding which
are subservient to the estimation of evidence '.
Whatever their differences, however, all accounts seem to
imply that thought as such is the special object of logical inquiry
and that logic owes its existence and its difference from the
sciences to some sort of distinction between thought and things.
Our view then of the nature of logic may depend upon our view
of the nature of this distinction, as well possibly as of another
distinction within thought itself.
Thought in general may be distinguished from other aspects
of human consciousness ; sensation, for example, or our practical
and emotional faculties, such as wishing, desiring, willing and
1 Arlis Logieae Rutlimenla, ftom the ten of Altlnch •, H L Mansel. Oxford,

1862, p xhii [The reference 1s m fact to Ramus not to the Schoolmen.]

[• Rather the earher Peripatetics. The Stoics included under ,-c) ~

,UflOI, (a) Dialectic, (b) Rhet.onc, Logic in our sense falling in (a). The mistake
(which is common) probably arises from a hasty remembrance or reading of
Ammcmius tt1 CAI., ch. 1 ]
 The Definmo,i of Logic 33
liking. Thought in fact means only that activity of conscious-
ness by which we know or try to know.
It is, at all events, better to leave the definition of logic in
a very general form at the beginning, as some kind of study of
thought. We shall thus be less likely to be unconsciously com-
mitted at the very outset to some preconceived theory which
might prejudice the development of the subject ; 1f, for example,
we were to say that ' logic studies the laws or forms of thought
in general '. The word law would in fact introduce associations
which we shall see later have promoted a mistaken view of the
office of the logic of inference.•
[• The descnpt10n of Logic is developed from Mansel, 1 c, Im and lxv,
probably m the hght of Hegel, Logic (Encyc ). § 19 I have put mto the
text the definitions which the author used as illustrations to the lecture ]

D
 II
THE RELATION OF KNOWING TO THINKING
§ 9. THE descnptton of logic as bemg some kmd of investiga•
tion of thought m d1stmctton from things may seem a ~e
enough generahzatton, yet we shall see that it involves certain
difficulties. Before cons1dermg the d1sttnctton we may naturally
ask for a definition of the word thought itself. The thought with
which logic has to do seems obviously connected with knowledge,
and, 1£ we examine the normal usage of the word in English,
we shall be led to the view that this connexion of th10king with
know10g 1s a universal charactensttc of the word. We shall also
see that we can give no defi01t1on, 10 the ordmary sense of that
term, either of thmkmg or of knowmg or 10deed of certam words
cognate with these.
In exam10mg the meanmg of a word such as ' thought ' in
philosophy, we must remember that it 1s a term of ordinary
language. It may have acquired, rightly or wrongly, some
different mean10g 10 philosophical wnt10gs and we must be
careful not to confuse the two. Thus 111 the philosophy of
Berkeley thought comes to be used in the general sense of
consciousness. Tlus is ahen to the normal use of language ; but
the philosophy of Berkeley and the alhed systems have come
so murh into vogue, that we are m danger of forgetting this
and falling into confusions. 1
Let us then first endeavour to follow actual linguistic usage
and take for granted the application of the name thinking to
certa10 kinds of consciousness and its exclusion from certain
other kinds. We shall then ask what is common to the things
1 So also Locke's use of the word ' idea • • was aben and still remains ahen
to the normal usage of ordinary English And the vogue of the phdosophy
of Locke has popularized a confusion
[• 'Some immediate obJect of the mind which 1t perceives and has before
1t' Epistle to Reader in An Essay, &c , cf Letter I to Stlllm.gfleet. The
word 11 used by Hobbes, but witltout emphas1S, m this aense, e.g. Lo11e, i 5,
§18and9]
 The Relation of Knowing to Thinking 35
to which the name is applied. Thinking then, in its normal
use, always has something to do with knowing.
There is some knowing, viz. the process of reasoning, to which
the name ' thinking ' is applied without any doubt. It is true
that, according to an idiom of our language, when we prove by
reasoning that the angles at the base of an isosceles triangle are
equal we should not be said to think that the angles are equal,
but to know that they are. We might therefore vaguely suppose
that perhaps the process of reasoning is to be called thinking
and that knowing 1s the result of such a process. This would be
a mistake ; for the process of reasoning is precisely the activity
of knowing, as will become obvious when we discuss inference,
and here, therefore, the ordinary idiom encourages a fallacy.
Thus then there is a certain kind of knowmg, which must be
called thmking 1f anything is.
But, on the other hand, there are activities, often, at least,
called knowing, which would not be called thmkmg. If every
apprehension of the nature of an object is taken to be knowledge,
then perception (or at least some perception) and the appre-
hension 1 of a feelmg would be knowledge ; yet, according to
the natural usage of language, they would not be called thinkmg.
This is probably because we regard thmkmg as an originative
activity of our own (not that we have clear ideas either of activity
or of origination), whereas we regard neither our perceptions nor
our apprehension of feelings as originated by ourselves. On the
other hand, the comparison of feelings, or of perceptions in
general, we do regard as thinking, because we seem to ongmate
these processes of comparison ourselves. Thus the apprehension
of relations, either of some or of all of them, is accounted
thinking, and probably the apprehension of universals 1s similarly
always taken to be thinking, as depending on acts of comparison.
If then some perception I is knowing, there 1s some knowing,
i.e. the experience of perceiving and feeling, which is not
accounted thinking.
1
Having a mental state, for instance a feeling, whether 1nvolvmg our
apprehe11S1on of 1t or not, 1s not, as such, to be 1dentmed with the appre-
he11S10n of the state For the mental state of desmng, e g , 1S not an apprehen•
sion of desinng, nor 1s 1t any kind of apprehension
• On perception .&nd thmking, cf. H 13, 4S, 146, 209, 263
D2
 INTllODUCTORY
Again, as not all that might be called knowing is thinking
so also not all thinking is knowing.
The formation of opinion is undoubtedly called thinking ; but,
though based on knowledge, it 1s not knowing. Indeed, according
to the English idiom already referred to, if we say we ' think '
A is B, it is understood that we are not prepared to say we
• know ' A 1s B. We are accustomed to say ' I don't know but
think so '. Opm1on, in fact, is a dec1s1on that something is
probable and, though not a decision of knowledge, is based upon
~ ur knowledge of the evidence available.
Further, there is something still more remote than opinion
from knowing which would be called thinking, viz. questioning
or wondermg. •
When we have not got to the truth which we happen to be
seeking, nor formed an opinion about it, but are wondering what
is true and putting questions to ourselves about it, we should
be said to be thinking. This certainly is the ordinary view, and
it seems natural enough when we reflect that this wonder is the
force which brings mto play that thinking which is the invest1ga•
tion of a given problem.
In an inquiry, first comes this questioning activity when we
set a problem to ourselves This implies that we know some-
thing of a given subject but are ignorant of some aspect of it
which interests us. We put to ourselves questions. our attitude
1is obviously not that of knowing, nor even of having an opinion,

,but an attitude in which we wonder what the truth is. We

.may find the answer by experience or some other direct appre-
hension : or else we may see that the facts known to us at the
start necessitate certain other facts and thus reach the goal by
reasoning, a form of thinking which is knowing. If we arrive
in either way at the knowledge which we seek, our undecided
and interrogative attitude ceases. If our data are not enough
we may either remain undecided, or we may form an opinion.1
1
H 44, 46, So, 5 1
[• Cf the fuller dlsc:WISlOD in Part Ill, ch vi, I 295, &c His atatemeut in
regard to its neglect seems enggen.ted. Sandel'IOll, Logaca, llfflS eontj,lfltlt••
(1615), Appendix II, ch r, para 11, ' Suspic10 Slve Dubitatlo est habitus, &c. •;
Hobbes, L"""'-, ch 7, Locke, I c , iv. 2. 5 (' precedent doubt '); 'Et ii
ya du doute avant la d6momtn.tion •. Leibniz, Nouv,..,. E....,, &c., iv,
I 5 : Kant, Lope, Introd , x et al ]
 The Relatioff of KftO'llling to Tmffking 3'I
That thinking has something to do with knowing may be
seen also by considering the thinking which is not directed to
scientific knowledge, for instance deliberation in regard to action
or artistic production.
When a man is planning something he would certainly be said
to be thinking. He is partly wondering and inquiring, partly
learning and knowing, and partly forming opinions as to what
would suit his purpose. Under this head would come the process
of literary composition, the production, say, of a poem or a play.
The imagination which belongs to such processes seems, as mere
imagination, rightly enough not accounted thinking, though
there is no thinkmg without imagination.• Remembering again
can only be called thinking because 1t is more than mere
imagination and involves apprehension.
We have found, then, included in thinking, activities of
knowing of the kind which 1s ppt expenencing, viz. ~sonin_g, 1
apprehension of umversals, and apprehension of relations between
things experienced; activities also which are not knowledge,
viz. inquiring, forming ~imons, w_gnderin~ and deliberating.
§ IO. We must now ask what 1s there common to these
activities m virtue of which they are all called thinking.
Consider what is common to wondering and knowing. Won-
dering involves knowing. We know something of a given subject
and we know that we do not know something else of it, or there
would be nothmg to wonder about. But wondering or question•
mg JS not 1dentical with this latter knowledge. Wondering
presupposes the idea of knowledge and is unintelligible without
it, because wondering is wondering what is true, although itself
not identical with any form of knowledge nor with the desire
for knowledge. Further m explanation we cannot go, for the
inquiring attitude is unique, cannot be expressed in terms of
anything else, is its own explanation. Similarly each of the
other forms of thinking owes its existence to knowing and can
only be understood through itself and through knowing. Both
thinking which is not knowing and knowing which is thinking
seem rightly called activities of consciousness 1 and of conscious-
• There II a possible acceptation of the term conscsOTUflus which would
c• NNUI ON lnw .,,.., . .ra<1pa-ro,, Anstotle, M ... 449b 31. Hobbel
continually insists upon this.]
 INTRODUCTORY
ness which is not the experience called perception or feeling.
This, then, is what is common to them : but it is a universal
which is not confined to them ; for willing and desiring, which
are not thinking, are also activities of consciousness. But beyond
the common universal of activity of consciousness, these forms
of thinking have no further differentiation of it to unify them.
What does unify them is the fact that the one, thinking which
is not knowing, entirely depends on the other, knowing, and is
only intelligible through it. This brings us to the general answer.
The unity of the activities of consciousness, called forms of
thmkmg,. is not a universal which, as a specific form of the
genus activity of consciousness, would cover the whole nature
of each of them, a species of which thinking would be the name
and of which they would be sub-species, but lies in the relation
of the forms of thmkmg which are not knowing to the form
which 1s knowing. Those which are not knowmg arise from the
desire to know or from some other relation to knowmg and are
unified with knowing by a special relation, dependmg m each
case upon its peculiar nature and sui generis, mtellig1ble and
only intelhg1ble by a consideration of the particular case. This
therefore is a case where the ordinary idea of definition 1s not
applicable Ordmary definition is a statement of the general
kmd (genus) to which the thing to be defined belongs and of
the characteristics of the particular sort (species), that is the
differentiation of the kind (genus), to which the thing to be
defined belongs. Thus all species of rectilinear triangles are
called by the same name m consequence of a common quality,
the havmg three straight sides. This is what is usually called
an explanation of the nature of the thing. In a given case we
may ask agam for a definition of the assigned genus and we may
go on to similar questions about this new definition In this
process we must obviously come to something which cannot be
defined, in the given sense of defimtton, or the process would
never end. Definition, in fact, itself presupposes the ending of
the process in elements which cannot be themselves defined, in
make it inaccurate to call all thinking an activity of conscioumess. If con-
11C1Dusness be hunted to consciousness of some obj1cl, wondering would not
be coDSC10usness It mvolves the consciousness of ob]ecta but is not itself
the consciousness of an obJect But then neither would desire be CODSC10ua-
ness, for desire is not consciousness of the object desired nor of its absence.
 The Relation of Knl1fftng to Thinking 39
so-called ultimate distinctions explicable from themselves alone.
This does not leave our notions indefinite, because the nature
of such undefinable universals is perfectly definite and is appre•
hended by us in the particular instances of them. In the case
of thinking, several kinds of thing are called by the same name,
not because of a quality common to them but because of the
manner in which they are associated in reality through the
peculiar relation of one of them to the rest and the nature of
their dependence upon it. The same 1s obviously true of far
less abstract universals such as colour and sound and even of
infimae species such as blue and red.
But now, since the other activities to which the name thinking
is applied depend upon knowing and to understand them we
must have the idea of knowing, it might seem that, though
there cannot be a defimtion of thinking (as definition is ordinarily
understood), we must ask for a definition of knowledge. But
the genus consciousness and its species knowmg are universals
of the kind just characterized ; no account can be given of
them in terms of anythmg but themselves. The attempt in
such cases to give any explanatory account can only result in
identical statements, for we should use m our explanation the
very notion we professed to explam, disguised perhaps by
a change of name or by the mvention of some new term, say
cogmtion or some similar imposture. We have in fact an
instance of the fallacy of asking an unreal question,• a question
which is such m verbal form only and to which no real questioning
in thought can correspond. For there are some things which
cannot be made matter of question. Indeed we cannot demand
an answer to any question without presupposing that we can
form an estimate of the value of the answer, that is that we
are capable of knowing and that we understand what knowing
means , otherwise our demand would be ridiculous. Our
experience of knowing then being the presupposition of any
inquiry we can undertake, we cannot make knowmg itself
a subject of mquiry m the sense of asking what knowing is.
We can make knowing a subject of inquiry but not of that kind
[• A favounte point, denved partly from 'tuch a vtew 111 •nadnuuible and
auch a question therefore has no point or at any rate leads to an answer
drllerent from that wluch 1t expects ', Lotze, 1 c , I 9 ]
40 INTRODUCTORY
of inquiry. We can, for instance, inquire how we come to know
in general, or to know in any department of knowledge.
§ 1 I. In the preceding investigation we have followed actual
linguistic usage. We have not disputed the application of the
name thinking and it might seem that we could do nothing else
but acquiesce m it, provided we are examming the meaning of
a name. The meamng would be a fact, m the sense at least
that the application of the word to certam thmgs was a fact,
which we &imply recognize as existmg. We have, for example,
distmgu1shed thinking from perception ; a d1stmct1on which, m
ordinary usage, has become a commonplace. It appears, for
mstance, m Browning's Sordello .
'Thought may take perception's place
But hardly coexist m any case.' •
Suppose now that, with some modern philosophers, we con-
tended that perception m fact involves thought and that the
ordinary distinction is incorrect. If we criticized actual usage
in this way, whether rightly or not, 1t would seem that we were
appealing to other data than usage and so not really mquiring
mto the meaning of a name, smce that must be determined by
actual usage. We would appear to correct the application of
a name because of our knowledge of the nature of the thing or
quality signified, a knowledge which enables us to see that that
nature is not present in some object to which usage attaches it.
Accordingly, such an inquiry would seem to be not an inquiry
into the meaning of a name but mto the nature of a thmg.
On the other hand, in our supposed knowledge of the nature
of the thmg meant by the name, m virtue of which we cr1t1C1ze
the application of the name, it would seem that we must have
started anyhow from the datum that the word is in fact applied
to the given thmg or quality.
[• The quotation from Browning was denved through H H Statham,
.drc/tdeaure "' tl&e Poels, ch I •
• Because perceptions whole, bke that he sought
To clothe, reJect so pure a work of thought
As language . thought may take perception's place
But hardly coexist in any case,
Being its mere presentment-of the whole
By parts, the simultaneous and the sole
By the successive and the many '-Worl&s (1888), vol. 1, p. 114]
 The Rlltmcm of KflOUling lo Thinking 41
As a matter of fact, in such investigations we are sometimes
examining the nature of a name and sometimes the nature of
reality without any very clear idea of the relation of these two
inquiries. In philosophy there is no denying that, however it
may come about, we are interested in inquiring what we and
other people really mean by certain terms, e.g. by cause, force
or thinking; nor is this interest confined to philosophy. General
biological theory, for instance, suffers very much because an
inquiry is not undertaken into what the word life already means
in ordinary usage and what guide there is to its meaning in the
things to which people apply it.
Yet there seems to be something odd in the idea of such
investigation at all. If we know English we do not inquire into
the meaning of •chair', •blue', 'loud', or •circle'. We are
supposed to know the meaning from our use of the words. Nor
should we seriously try to find out the meaning of the definition
of a circle by examining various circles and asking what they
had in common. Again, when we do appear actually to use
such a method, as when Socrates asked • What is Justice? ',
whether we thmk we are investigatmg the nature of the thing
or finding out what people think it is {that 1s, the meaning of
the name), we depend obviously upon knowmg that the acts we
are examining arc all really • just '. It would be no use abstract•
ing from acts that are not Just.
How then do we know that we have included no acts that
are not Just ? If it is through a conception of the nature of
justice, a conception by which we test particular actions, then
we have already what we profess to be looking for.
It seems to remam that we can only take as data the actual
application of the names : if so, we are at the mercy of usage.
We cannot criticize it : even if we found in it anything apparently
contradi~tory we should be helpless to decide. Again, what
security have we that we make the right generalization and find
out what people really mean ? If we took a look at various
ellipses, we should not be likely to find the abstraction which
mathematics gives as their common definition, and in a look we
might see no difference to speak of between a parabola and
a hyperbola. Would it not be the safest way, as it is also the
easiest, to ask people who use the word what they mean ? This
4Z INTRODUCTORY
would certainly be the right way in the case of the ellipses or
different sections of the cone
§ 12. The solution of these difficulties, as might be anticipated,
is to be found in what was said about abstraction, 1 namely, that
there is a certain feeling of affinity between particular cases, the
nature of which we do not clearly understand and cannot
formulate. This explains the paradox that we are able to
criticize the data on which we seem wholly to depend. The
application here of the word ' feeling ' 1s due to a proper instmct
in language, m so far as it 1s realized that we have not here
clear apprehension (or clear thinking) and therefore any such
definite word as knowing is avoided. But really, feeling is not
the right name nor has ordmary speech got a name for it. There
are in fact certam conditions of our consciousness which are akm
to thinking, akm to apprehension in general, but are neither.
We shall have occasion later 2 to recognize their existence other•
wise : we shall find, for example, a condition of consciousness
which simulates judgement and opimon, but 1s distinctly neither.
In general much difficulty is caused in logic by the attempt to
express everythmg m terms of a clear thmkmg consciousness •
These other conditions of consciousness arc not recognized and
so the phenomena which belong to them get misinterpreted.
The log1C1an here has naturally been affected by language.
There is a want of terms for these conditions of consciousness ;
their existence is imperfectly mdicated by the use of such a word
as feeling and the impulse of the philosopher is at first simply
to criticize this, because he realizes there is some confusion of
thought in the employment of the word.
It is difficult to describe such conditions Just because there
is no proper language for it, but we can mdicate their character
by describing the corresponding facts of consciousness. There
are certain prmc1ples which exist implicitly in our mmds and
actuate us in particular thoughts and actions, as 1s shown by
their operation in our attitude to particular cases. But we
realize them at first only in particular cases ; not as definite
1
I 5. I I 54•

[• • It is not necessary that everything should become reasoned. knowledge

Much 1B quite clear and plain wluch yet shrinks from defuution,' Lotze,
GntfFGI Pltysiology, p 163]
 TM Relation of Knowing to Thinking 43
general or universal rules, of which we are clearly conscious and
by which we estimate the particular cases. On the contrary,
there is no such formulation to precede the particular cases :
the principle lives only in the particulars.• This can be under-
stood by means of examples. Take, for instance, the logical
abstraction of the syllogism. People argue quite correctly in
particular syllogisms : they see the necessity of the conclusion
from the premisses in a particular case ; they are entirely
unconscious of the general rule. Thus the abstract form of it,
when first presented in logic, comes as somethmg new, while their
acquiescence in the form or prmciple depends on an appeal to their
own consciousness m which they have been implicitly using it.
A more important example, both in itself and historically, is
to be found in moral rules and definitions. It seems absurd to
say that a person who 1s distinguished for the justice of his
conduct does not know what is just, and he might be rightly
mdignant 1f you denied that he knew the meaning of Justice ;
yet he nught easily be puzzled if asked to define 1t. Now owmg
to the unity of such a principle implicit in our minds there
must be an affinity in the cases where we do use the term justice,
We know the JUSt man has a principle, and always treat him
as if he had ; and yet, as we see, the prmciple lives for him only
in its application in particular cases Indeed, the term applica-
tion itself is somewhat misleading, becaui;e it rather implies that
we first have the rule consciously and then apply it, which is
not the case.
This affinity finds its first expression and recognition m the
appearance of a common name. Often, as in the case of justice
at present before us, this name has no clear, decided, definite
meaning. It corresponds only to a general consciousness of
affinity which has not yet arrived at a clear understanding of
itself. When we begin to feel the want of a clear notion to
correspond to the name and, what is perhaps more important,
when we become conscious of the need of a definite rule in
action, something, in fact, to make our Judgements in regard
to what is just more reliable, how have we to go to work ?
[• Anstotle,D1M11n 45o&1al, Locke,Essay,iv 7,111: 'Nowtheammal,
fN animal, does not exist : 1t 11 merely the universal nature of the individual
arumals ' Hegel, Log1c (Encycl ), I 24 ( 1) )
ff INTRODUCTORY
Obviously we must start from the facts of the use of a name,
and shall be guided at first certainly by the name : and so far
we may appear to be exammmg the meaning of a name. Next
we have to think about the individual instances, to see what
they have in common, what it is in fact that has actuated us.
This seems by contrast to be the examination of a thmg or
reality as opposed to a name. At this stage we must take first
what seems to us common in certain definite cases before us :
next te&t what we have got by considering other instances of our
own application of the name, other instances (more accurately)
in which the principle has been workmg m us. Now, when thus
thinking of these other instances, we may see that they do not
come under the formula that we have generalized. If we feel
satisfied (and it is only by thinkmg about the particular cases
that we decide whether we are or not) that these really belong
to the rule, are in fact 1ust, we require an enlargement of our
formula. The defimtlon was too narrow.
Again, arguing from the formula itself, we see how it neces-
sitates that certain cases should fall under it : but when we
consider such cases we find that we do not in ordinary life apply
the notion to them and, if m our moral consciousness we are
confident that they do not accord with our prmciple, we have
correspondmgly to correct our previous generahzatlon. This
time the definition is judged to be too wide.
Observe that in every such step we rely upon the rightness
of our use of the prmc1ple in particular cases ; this does not
mean that we are sure of ourselves m every case, but that there
""' cases at all events about which we are sure. This explains
what in the Socratic attempt to find definitions would otherwise
be paradoxical and mexphcable. The definition depends for its
correctness on the assumption that the people who wish to find
the definition know what is just already and know it in the
most important way, from a practical point of view. Yet all
the while they are supposed to be trying to find out what justice
is; and so, with the ordinary analysis, the whole procedure seems
irrational. We understand now that the people who are to be
instructed by the Socratic method do know in one way,• and
everything depends on their knowing in that way, but there is
[• Cf. note 11, p 25]
 T/18 Relano,- of KflOflling lo Tht,-Jnng 45
another way in which they do not know and this it is which
gives the investigation a rationale and meaning.
There is a further stage when we have, or think we have,
discovered the nature of the principle which h,:U really actuated
us. We may now correct some of our applications of the name
because we see that some instances do not really possess the
quality which corresponds to what we now understand the
principle to be. This explains how it should be possible to
criticize the facts out of which we have been drawing our data.
§ I 3. Let us apply these general considerations to the case before
us, the distinction in the normal use of language between thought
and perception. When the exclusion of perception from thought
is called in question we have the paradox of a challenge of data.
It being understood, though not necessarily after a clear inves•
tigation, that the apprehension of universals is thinking, we find
it contended that this apprehension is found in perception • but
that this fact has been overlooked. Perception m consequence
has been erroneously distmguished from thmking.1
But there is a simpler ground for recognizing thought in
perception. We are sure that reasoning is thmking, that com•
paring is thinking, processes which involve both the appre•
hendmg and mquiring attitudes. That being so, these are our
certain data and m these activities we recognize consciousness
concerned with knowing. We recognize also that we think of
such processes as originated and conducted by ourselves-
originative activities we may call them. We see this operating
in our view of feeling ; for, though it is consciousness, we
distinguish it from thinking ; and again we see the distinction
in our ordmary dismclination to call perception thinking because
1 We shall see reason later to coIISlder whether t1us 11 really Justified and
we shall challenge the view that m ordinary perception the UD1versal lS really
apprehended as such. (I 147)

c• Ultimately from Anatotle, ' Mal id, afdarrN& ,..1. "" .,,.. lunar
t r .r.,,,.., mllllldAov lnw, ol'w M,--, ', An. Po. 1ooa 16 Wilson is
doubtleu ftfernng to the current Oldord ideahsm. • The sunplest act of
pm:eption is a judgement• was almost a commonplace then, cf. T. H. Green,
Wora, u, pp. 170-2; and • In all human perceptaon thought is pre&ent, 10
too thought 1s the 111Uversal in all acts of conception, recollection, &c.', and
again • Man therefore is always thinking, even in lus perceptions ; if he
obeervea anything, he always obeerves 1t as a 111Uversal ', Hegel, I. c., f 24 (r).J
46 INTRODUCTORY
of the element in it which we seem in no way to originate ; an
element also which seems to be what is mainly important in the
matter.
Consider a sensation and our knowledge of it. The mere
having a sensation, though 1t is consciousness, is not knowledge
and must be d1stmguished from apprehension. To know what
a sensation is I must recognize 10 it a definite character which
dist10guishes it, e g, from other sensations. I recognize, let us
say, thc1,t 1t is a pam, and then agam a burn10g, or a prfcking,
pain, as the CJ.se may be. But this implies comparison of pain
with other sensations and other pams; and thus by the activity
of comparing we go beyond the mere passive state of being
pained, and this activity we are sure, ex hypothesi, is thinking.
Thus though the sensation is not orig10ated by us we require
an originative act of consciousness to apprehend it.
The same is true of other obJects of consciousness 10 percep-
tion, which we do not ordinarily suppose to be sensations,
e g. objects seen as extended m space. Whatever passive
element there 1s (and we certainly do not suppose ourselves to
origmate the shape and colourmg of thmgs) the apprehension
of the characteristics of what we perceive mvolves a comparison ;
and comparison we take to be thinking. If this ts so, the knowing
part of perception would after all be thmking and the distinction
whereby the knowmg 10 perceptmn was excluded from thinking,
would only be a popular inaccuracy.
Yet here we must be careful to avoid an overstatement. It
is not fair to condemn the ordmary view wholly, nor 1s it safe :
for, if we do, we may lose sight of something important behind it.
Distinctions current m language can never be safely neglected.
In what we ordinarily recogmze as comparison we have before
us two obJects at least and apprehend each of them distinctly.
As we should say, we are thinking of the nature of both. But,
in the apprehension of the definite quality of a given sensation,
we are as a rule not consciously comparing it with the quality
of another sensation which we distinctly remember and so have
before us. We are not concerned primarily with the qualities
of other things, but only with the quality of the object before
us : our interest is 10 it and not in them and the fact seems to
be that we have a consciousness of 1t as havmg a quality differing
 The Reltition of Knowing to Thinki,ig 47
from that of other objects in general, but not a consciousness of
other objects in detail. The particular qualities then of other
things being in abeyance m this way and our interest being in
the distinctive quality recognized in the object, we can under-
stand how the fact that there is a comparison comes to be
overlooked and how we seem to be merely appreciatmg the
quality of the object by itself In short, we are really comparing
but do not recognize that we are. This then shows that the
comparison in this case (though obviously necessary to recog-
nizing the quality of the obJect as something distmct m itself
and not just to be confused with anything at all) is different
from ordmary comparison, and requires special recognition.1
• The sub1ect of the d1&tmct10n of thought from perception will be resumed
later m the discussion of universals, § 146.
 LOGIC A~D COGNATE STUDIES GE.SERAL
A~D SPECIAL LOGIC

§ 14 BEFORE c.om,1dermg the d1stmct1on between logic as

a study of thmkmg and the sciences as special studies of thmgs,
we may make a prov1s10nal d1v1c;ion within the provmce of
thinking itself In ordinary speech we dtstmgutsh logical mqu1ry
from the mqmry mto questtonc; which we call grammatical,
metaphysical, ,rnd so forth We feel that there is a certain
affimty between logic and other studies which may be c;a1d to
deal with thought a'> their obJe(,t, yet \11,C are av.arc .ilc;o of
a difference betV1ren 1t ,rnd them ac; well af, ht't\\cen them them·
selves. These d11lercncec; \\C shall now prov1,;;1onally consider,
m order if possible to <letermme more clearly what it 1s that
we mean "hen we spc,1k of a question as bemg a question of
logic
We shall also ask prov1s1onally ' In what sense 1s logic an
a priori subJect, as pure mathematics appears to be~ ' and
agam ' What is the ground for the famthar d1v1S1on of logic
mto pure and applied logic, a d1stmctton, tf not c;uggcsted by,
at least analogous to a popular d1vis1on of mathematics mto
pure and apphcd, or pure and mixed, mathematics~ '
§ I 5 The nature of what logic mvesttgates 15 what prescribes
its method It 1s always the special nature of the object
mvestlgated which delermmes the method of mvestigatton. We
cannot settle the nght method of any science, without attending
to the special character of itb objects Logic bcmg an activity
of thmkmg must presuppose the forms of thmkmg, because 1t
must use them, and therefore 1ts study of these forms cannot
be m the way of either doubtmg their vahd1ty or estabhshmg
1t For m such an examination logic would have to presuppose
the validity of what 1t would be cnt1C1zmg. All it can do 1s
[• This chapter wai. 111eant to 'be revised, but l"tillt left alone except II 22-3 J
 Logic and its Cognate Studies 49
to disentangle the universal from the particulars in whu:h it is
manifested ; in other words, in this logical activity, thought
itself is but recognizing its own universal forms. This is why
some parts of logic are so simple : for at all events such forms
as are the presupposition of any knowledge whatsoever must be
simple and obvious when once pointed out There is thus
a certain appropriateness m Aristotle's term Analytics, and it
is significant that m grammar a procedure hke that of logic is
called the analysis of sentences. A quahficat10n, however, must
be made Analysis 1s often understood to imply a whole of
which the parts are explicitly known before the analysis , 1 but
logical elements are for our ordinary consc1ouc,ness only 1mphc1t :
we use them without reflecting on them, JUSt as we use gram-
matical d1stinct10ns long before we have any knowledge of
grammar. 2 Logic docs not merely an,tlyse 1t makes exphc1t
what was 1mphc1t
§ 16 There 1s a <:ensc m wluch both the log1cal forms and the
method of logic may be called a priori 3 It 1s true that they
are not actualized as matters of consc1ouc,ness till we have
experience of particular thought'> ; but their validity is not
derived from <:mgle mstances That 1s to say, they are not
learned a posteriori from instances regarded as their evidence
and therefore h,tble to corrf.'ct10n by further expcnence of
thmkmg (On the contrary they belong to the subject as sub-
Ject and c,o must rond1t1on f.'Very part of the thmkmg experience.
The object m experience which we thmk about cannot con-
tradict them, for the forms of thought are properly forms of
thinking and not of 'what we thmk' or 'what 1s thought'
(even from an 1deahstic pomt of view), and so either the forms
of what we thmk, that 1s the obJect, are irrelevant to the forms
of the thinking, so that there can be no contrad1ct1on ; or, 1£
there could be any contrad1ct1on, 1t could only arise from this,
l §5 • Cf Le1bmz, Nou1•oaux Essazs a • §§ 6 and ..l4l,

[• No reference given Perhaps he was thinking of• II est vra1 qu'd ne faut
point s'1magmer qu'on pu1s,e bre dans l'Ame ces etemelle& 101s de la ralSOn
a\ bvre ouvert ma1s c•e~t assez qu'on le~ pu1S&e tlecouvnr en nous a\ force
d'attention a quo1 le~ occasions sont foum1es par IC'! sens', Introd, and the
passage about the Chmese language ending, • c•c~t ams1 qu'on pos~de b1en
des choses sans le savo1r • Book I,§ 21 Cf Lotze, Logic,§ 358)
3 773 I E
50 INTRODUCTORY
that the forms of thmkmg might necessitate something m the
object to be thought about. Now, 1f that were really possible,
we could not intelhg1bly represent any object as contradicting
them, for to do so we must thmk abo1:1t 1t and, tf 1t possessed
attributes contradictory to what the forms of thmkmg demand
m 1t, 1t simply could not be the obJect of thought. S1mdar
cons1derat1ons apply to the forms of apprchens10n m general.)
Not only arc the formc, a priori m this sense, but so also 1s the
method of their mvest1gat1on For, though first recogmzed in
particular inc,tanccs of their use, they are at once seen to be
mdependent of them, smce otherwise we could not go beyond
the mstr1nces analysed, except by way of conJecture Now we
have a ccrtamty of their umvcr,;ahty wluch 1s mcompat1ble with
.m a postertort ongm Thought can recog01.:e 1ts own laws by
rcflectmg upon itself
~ 17 Grammar cannot be simply d1-;tmgmshcd from logic on
the ground thc1t 1t dral., \\1th lr1nguage wlule logic deals with
thought Grammar deal~ with lang•uge only as the symbol of
thought, a,; enahlmg us to understand thought when expressed
in words Thus 1t '>l'ck-; for general formc;; of cxpress10n which
have the same kmd of rnd11Terencc .is log1rnl form!> to the c;pcc1fic
cont mt rxpres!><:d, and c,omettml's thc&e general forms comc1de
with the log1ral forms Their prmc1pal difference 1s that logic
deals with thought qmlc generally and m absturt10n from any
particular hngu1c;t1c mode of expression (though 1t 1s true that
log1c1anc, may be greatly 111fluenced by the form'> of expression
111 their own languagt·) , gramm,tr treats of forms of thought so
far as they h.i,·e bernmc recognized 111 hngu1st1c formc; Forms
of thought get direct rcrog01tlon m grammar ""hen in the
language studied there happen,;; to be a general word-form
correspondmg to them or some kind of general rule For
example, noun~ have not all the ~.1me termmat1on m a given
language, yet tlwy all ~tand under the same set of rules m
relation to verb-; Forms of thought get also an indirect recogm-
t1on \\ hen the different species mto which the form subd1v1des
have special word-form'l corresponding, though the general form
itself which comprises them all has not Clearly then logic and
grammar so far agree 111 that both involve a study of forms of
thought applicable to all kinds of obJects. Grammar however
 Logic and its Cognate Studies 51
is the more hmited in scope, because it studies them primarily
only so far as they have received expression in the general
word-forms of a given language.
§ 18. If psychology meant the study of mmd in general, logic
would only be a part of 1t, but then the same would be true
of ethics, politics, aesthetic, the philosophy of history and the
theory of literature In practice psychology has not this general
scope; 1t has become a special study of a scientific character,
and in some ways 1s experimental and empmcal like the empirical
sciences Psychology 10 this narrower sense of the \\<Ord differs
from logic by mcludmg subjects which logic exdudes, for example
the practical and emotional aspects of consciou,;ness and ques·
tlons about the physical cond1t1ons of consc1ou,;ncss The real
difficulty m d1o;tJngmshmg the two comes m those departments
10 which they have m some sense a common ohJert, for psycho·
logy generally mrludes some study of the mental processes
connected with knowmg and behevmg, processe,; such as havmg
apprehensions or forming Judgements and opm10ns. It must be
confessed that the attempts often made to define the province
of psychology 10 d1,;t10ct1on from logic and other wgnate sub·
Jects arc vr1.guc and unsat1-,factory, for the truth 1,;1 success
depend,; upon ,;omc knowledge of logic and metaphysics, 10
wluch !.UbJcrts p,;ycholog1sts are not ..t.lway'> at home
In every ,Ht of thought we must recognize a twofold aspect.
In the first pl.tee, it appe,us a,; an event, when regarded as the
activity of J. particular thinker bccau,;c 1t happen,; at .• parti-
cul.ir time, and m a particular time-order with reference to other
events Agam 1t 1s connclted 10 th1,; time-order with all those
occurrences, physical or mental, which may be ,;aid to con·
tribute m any way to the fact that the per,;on thmkmg thinks
a particular thought at that particular time, for example the
thmkmg suggested by the sight of some memorandum we have
made. But there 1s another and a totally d1stmct a&pect 10
every thought , <tins aspect 1s <;imply the tl11nk1pg a:. thmkmg
about somethmg, an apprehension of somethmg, whether accom•
pamed or not by a conJccturc or a question about 1t.) This
aspect 1s the Sdme whenever the thought occur<; and is wholly
unaffected by that other aspect of the thought as an event. It
rs somethmg complete m itself, wholly mdepcndent of the time•
E 2
52 INTRODUCTORY
order as such and of anything which conditions that time-order
as a mere time-order.
It is quite true that (this aspect of the thought as apprehension
of a fact connects 1t with other thoughts as also being appre•
hens1ons, and through the nature of the facts apprehended ; it
is true also that) this very connexion (which we may convemently
describe as a connexion of the contents of the thought) may
cond1t1on the occurrence of a given thought as a subjective
act1v1ty at a given time. Nevertheless this connexion of the
thoughts (m respect of what 1s apprehended in them) 1s not
such temporal connexion at all. It may 10fluence but it cannot
be mfluenccd by the temporal order To take an example, we
may have the thought of a triangle. Now the content of that
thought necessitates a property of the triangle and we, appre-
hending the first content, may apprehend afterwards that 1t
neress1tates the &erond The reason of that hes m the objective
relation 1b,df Tlus, though appn,hended m time, 1s not a tem•
poral relation, and m it the two clements <l1stmgmshed as
corre.,pondmg to the two acts of thought arr not related as
before and after Tins d1stmct1on appears m every process of
learmng or mventmg a proof ; the various &teps are apprehended
by a particular lhmker m a certain order m time , 1t • takes
tum time' as we say to discover the argument, the amount of
time bemg conditioned by his mental capacity and his physical
state Here then the acts of thought seem to present them-
selves as events bearmg an essential relation to time, and even
camally connected with other events, (they may possibly even
havl' the appearance (though that 1s a m1smterpretation) of
bemg connected "11th one another m the temporal way of cause
and effect) Nov.. though this may be so, yet what the thinker
m this process understands, the meamng of the argument, does
not as such enter mto anv of these temporal relations That
meanmg obviously hes m the apprehension of the premisses and
of their connexion with the conrlus1on. The apprehension of
a premiss is something complete m itself and, whatever the
temporal occasion of the apprehension, what 1s apprehended is,
e:xcept m a sense to be presently explamed, entirely free from
temporal connexion and its temporal character may therefore
be ignored. S1mtlarly the connexion of the premisses with one
 Logic and its Cognate Studies 53
another and with the conclusion is not temporal in the sense
that it has anything to do with the time taken to apprehend
it The relation of premisses and conclusion (as thoughts about
something,) is not that of cause and effect, nor even a relation
of succession in time The premiss is not an event preceding
the conclusion, for then it would be over and gone when we
reach the conclusion, whereas 1t 1s the presence of the premiss
which 1s the cond1t1on of our having the conclusion at all.
Premisses and conclusion are, m their essence, m no time relation
and, 1£ we tried to represent them as m time at all, we should
have to represent the various apprehensions as completely
synchronous in an mdiv1S1ble time
Now logic 1s concerned with our thoughts not as events but
with that side of them which 1s not event , 1 1t 1s concerned
with (their character as apprehensions of ob1ects, or as con-
Jectures and op1mons added to those appreheno;1ons; that is,
1t is concerned with) the truth or falsehood of what we thmk
Logic never considers the way in which we come to thmk a given
thought at a given time, except so far as the process is entirely
withm (the activity of apprehension as such,) and dependent
therefore on relations, which are entirely non-temporal, (between
the thoughts, apprehens10ns and so forth concerning a given
object matter), and not upon anything belonging to their
character as events It is true that we may trace such con-
nexions between the contents of our thought man order of time
m our own mmds, but m our apprehension of them the idea of
(this sub1ect1ve) time and everything temporal (wluch does not
belong to the nature of the obJect thought about) 1s abolished.
This gives us a distinction between logic and psychology in
so far as both deal with thought II Psychology treats thoughts
as events and, qua psychology, is mainly concerned with their
time-order , often indeed 1t has to regard the merely temporal
conditions under which the content of a given thought becomes
a matter of consciousness to a given thmker at a given time.
1
Mr Bradley gets mto great confusion about this charac.ter of thought.
Cf f 122
• The word psychological 1s too often used m a vague and confusing manner
even 111 logical treatises See, for example, S1gwart's Log1k, vol 1, § 32 (p 203,
bne g of German text 1873) [' No Judgement JS uttered without a psycho•
logical ground for its certamty ']
54 INTRODUCTORY
Moreover, (besides certam physiological conditions which are
absolutely nothing to logic,) psychology 1s occupied with (the
temporal conditions of) such phenomena as memory, association
of ideas, 1magmat1on m general, (treating a side of them which·
doe!> not belong to logic) and with processes necessary to per-
cept1on, (which logic as such does not consider. Nor, finally, 1s
logic concerned with what 1s paradoxically called the uncon-
scious mechanism of consciousness )
§ 19 It mu!>t not be suppo-,ed however that all consideration
of the subJective time-order 1s nece!>sanly excluded from logic.
Logic ongmated in the reflection of the mmd on its own sub-
jective arhv1ty, and 1s essentially concerned for example with
inference as a subjective act As such, mfcrence repre!>ents an
advance m knowledge 1mplymg a contrast with a prev&0us state of
mmd, before that realization of thought wluch 1s the result of the
process In the statement, too, with 1t., <l1stmctton of subJert and
predicate, a tnnr-order 1s, we !>hall see, mvolveci Tins cons1dera-
tion of the ttmc-ordt•r 1s ueces!>ary for log1c, \\ hatrver p!>ychology
may from 1t., pomt of view ha.ve to say, and nothmg whatever
1s here borrowed from p!>ychology a!> a science. But the d1s-
tmction of before and ,liter 1s treated quite differently m logic
and 111 psyd10logy. (In the d1-,tmct1011, for mstancc, of subject
and predicate, logic 1s concerned with the mere fact that the
apprehension correspondmg to the one precedes that which
corresponds to the other but not with anythmg else rel,Ltlng to
the tune; not with .my reason for the fact and ccrtamly not
with any temporal rea!>on for 1t nor, 111 general, with anything
cond1t10111ng the temporai sencs, qua temporal) In the case of
mference logic 1<; concl·rncd only with !>uch reasons for the time•
order a!> he 111 (the nature of the apprehension itself) If we
ask whether the concept1011 of a figure whose mtenor angles are
together equal to two nght angles could precede the conception
of a three-sided rect1hncar figure, the answer depends entirely
on the matter of the conceptions themselves, 111 other words on
the relations of what 1s apprehended
§ 20 (Though metaphysics has no generally accepted defim-
tion, the word 1s m frequent use and this 1s so far Justified
masmuch as 1t corresponds to a certam affinity, felt rather than
clearly understood, m the subjects to which 1t 1s applied Thus
 Logic and its Cognate Stu.dies 5S
we feel that certam problems are rightly called metaphysical,
though we may be at a loss to define metaphysics ; Just as we
are sure that a certain act ts generous, though we might find
it difficult to define generosity so as to draw the lme between
it and justice. This 1s why philosophers are more likely to agree
as to what questions should be called metaphysical than upon
::my defimt1on of metaphysics. The vague popular conception
of it seems to be that it H, not science but philosophy ; and
further, not any philosophy but some lughcr kmd of plulosophy.
Logic, m a strict acceptation, and etlucs and politics may all
seem not to be metaphysics, yet we find questions occurrmg
m logic and also m ethics wluch we should naturally call meta·
physical. Examples of such questions arc, the reality of uni-
versals, the nature of causation, the problem of freedom and
necessity. The word metaphysic origmated m the name given,
by some unknown arranger of Aristotle's works, to the treatise
which Aristotle himself spoke of as concerned w1th First Philo·
sophy 1 and the name only means that the treat 1sc m the order
of his system was c011S1dere<l to come '.1fter The Physics',
although Aristotle's Physics would itself accordmg to our modern
usage be said to be mamly metaphysical
The popular idea of metaphys1cc; correi,ponds so far with the
contents of the Aristotehan treallse, masmuch as he thought
tlus first philosophy 1 to uc a higher km<l of study lo wluch his
Analytics was a prl'hmmary, and as moreover he does not mclude
etluci, w1th111 1t He also d1st111gu1sl,cd it from the sciences, as
treating of Bemg m general while they each consider some
special part of Bemg. Thuc; Aristotle's metaphysics mvest1gates
the elementary or fundamental prmciplcs, 2 and God as the
supreme and most perfect Bemg ) If we now try to formulate
what 1s essential m the affinity we feel between metaphysical
problems, perhaps it 1s truest to say that metaphysics ha!> for
its ultimate object and ideal a complete understanding of reality,
and that not as opposed to the thmkmg subJect but as mcludmg
the subject. It seeks at all events a completer understanding
than 1s contamed m the sciences and so 1t is bound to let no
assumption or presupposition pass unexamined. The sciences
are also mqumes mto the nature of reality, for they assume
I dpxal.
 INTRODUCTORY
conceptions and statements which they use and develop, but
which, as sciences, they neither examme nor cnt1c1ze. Geometry,
for instance, assumes space, but, as geometry, does not criticize
it. Generally, scie11ce assumes the reality of ob1ects of a knowmg
or perce1vmg subJect and accepts a certain oppos1tion between
the two ; these presuppositions metaphysics cxammes. Logic,
such as Aristotle origmated, studies thought and brings to hght
its presuppositions, but still it makes assumptions which, as
logic, 1t does not investigate. Exammmg thought as the sub-
Jective element m apprehension and so assummg the difference
of subject and object, it assumes that m experience the subJeCt
can know (that the object is there and also) something about it,
it assumes m short the workings of thought as data and arranges
them. The criticism of these assumptions, (whether explicitly
faced m logical treatises or not,) 1s metaphysics Metaphysics
1s bound to raise the whole question of the nature of the relation
of thought to reality and therefore an ideahsllc theory of reality
such as Berkeley's belongs to metaphysics and not to logic
Similarly with subJects usually included under the title of theory
of knowledge We may l>ay then shortly · <science studies the
obJect1ve side of thought, logic the subJcct1Vl'), metaphysics
studies both and the relation between them But tlus 1s not
enough; metaphysics btudies them m a manner d1ffcrcnt from
that m wluch they are studied by logic and the sciences Meta-
physics (does not propose to add to the sciences w1thm their
own hm1ts ; for example 1t docs not study geometry m order
to develop new geometrical the,orems it tries to complete the
sciences m another \\ay, a way m ,,luch they cannot help them-
selves), by understanding both their prel>uppos1t1ons and the
orgamc connexion of the <.hffcrent parts of rcah t y "Inch are
severally studied by them
§ 21 'We can no\\ see how the 1,tudy of thmkmg, of the bemg
of the apprehending thought, may go beyond a strictly logical
activity and comprise subjects usually contamed under the term
theory of knowledge, for example, the vahd1ty of thought in
relation to reality {a problem wluch involves metaphysical
questions proper) or the reclhty of the umversal, and the pos•
sibihty of gcttmg knowledge from percept10n 1 two problems
which have also some relation to psychology. The association
 General and Special Logic 57
of ideas seems to belong peculiarly to psychology, yet it comes
naturally into any discussion of knowledge. All such subjects
are united by the desire to mvest1gate thought as true,-this is
what relates them to logic , and though not parts of logic
proper, they are so connected with it that the logician is obliged
to consider them, JUst as the geometrician must study the theory
of proportion, though it belongs to the province of pure quantity.
On this account we must never· put aside a question on the
mere ground that it 1s metaphysical. It may be as necessary
to logic as the theorems of one science sometimes are to those
of another 1
§ 22 • A d1stmct1on 1s sometimes made between ' pure ' and
'applied' logic, which seems to be partly a legacy from an old
mistake about the theory of the syllogism and partly due to an
maccuratc d1stmct1on made m the sciences The terminology
1s borrowed from the d1v1s1on of mathematical science mto pure
and applied mathematics This agam 1s grounded on the fact
that the theorems of geometry and of the calculus of pure
quantity arc applied to the geometrical and quantitative rela·
twm, of bodies m movement or m eqmhbrmm 2
Now the old view of syllog1sttc inference was that 1t formed
the general method of all demonstrative mfercnce whatever, so
that the methods of the demonstrative sciences would be merely
an apphcat1on of the rules of the syllog1st1c theory ; they were
accordmgly called deductive sciences. Hence, perhaps, may have
arisen the idea that the logic of a particular deductlvt: science,
1 e the logical theory of it!> method, would be the application
of the general theory of demonstrative mference to its special
case. Similarly, when the theory of mduct10n began to have
I (,f § 4

[• §§ 22-23 A digre~sion pnmanly directed against Lotze Wolf divided

Logic into Theoretica and Praclica, a distinction restmg on the old Logica
Docens et Utens Kant used the term!. General and Special Logic, probably
denving them from the distinction between General Law or the Law of Nature
and Special Law m Junsprudence Hegel uses applied logic for Physics and
Psychology, Works, v1, p 49
In the 1909 course Wilson added ' In the general treatment of the sub1ect
a senous error has been caused in the theory of inference, IDasmuch as an
attempt has been made to find a quite general proce~s of inference based
upon mere abstract considerations about the general form of propositions '
He was refemng to symbolic logic (infra, p 59) j
58 INTRODUCTORY
more importance in logic, it was perhaps thought that there was
an apphed mducttve logic of a special inductive science.
But even if this view of the syllogism and of the deductive
character of the demonstrative sciences were right, as it 1s not,
the d1stmctton of pure and applied 1s maccurate, confused and
quite worthless even m the sciences from which 1t 1s borrowed.
For in,;tance, 111 the application of geometry to mechanics the
geometrical theorem'> remam exactly what they were m pure
geometry There 1s no applied geometry wh1rh 1s a special geo•
metry of mechamc" , the only geometry m the .i.pphcd geometry
1s pure geometry If anytlung could be called applied geometry,
applied algebra, &c., at all, 1t would have to be for mstance
the science of mechamcs 1t'>clf as a whole
But this, agam, 1s quite maccurate The apphcation of the
gcncral theorems of a <;Ctcnce to particular mstanccs never pro•
duces a new science. The so-called applied m.i.themattcal sciences
are new sciences a!> compared \\1th pure mathematics, and they
are so because they add to pure mathemat1rs Rr1ent1fic principles
of their own which .i.re not m any !>l'nse derived from pure
mathematic!> Tht> term' applied mat hemat1<''> ', used to de<;rrtbe
the whole nature of these sc1ences, mvolves a very grave con-
fus1on Smularly, 1£ \\ e allow the untrue assumption that the
demonstrative sricnrcs arc 'deductive', m the technical sense
of the word, the only logic of them would be the logic of the
syllogism, the so•c.tlled ' pure logic ', JUSt as the only geometry
111 'applied mathematics' remams pure geometry. There could
be nothmg which could be d1stmgui1:ihed as an applied logic from
the pure logic
But it 1s quite untrue that the syllog1st1c logic presents the
general form and rules of all demonstrative reasonmg, as we
shall see hereafter
Agam, even 1£ the syllog1st1c theory of a general form and
general rules of all mference whatever \\ ere replaced by the
theory of some other gt·neral form and general system of rules
of which every particular mference was a mere application, the
first objection would hold and the idea of an applied logic would
be altogether mcorrect.
§ 23 The distinction of general logic and special logic seems
much more promising. It looks as 1f there might well be, for
mstance, logical considerations applying to reasoning in general
 General and Special Logic 59
and also logical study of special forms of reasoning peculiar to
different departments and to special sciences. Nevertheless we
must abandon altogether the idea of a general logk which could
be developed a priori without a study of the processes of
reasoning m the sciences. We shall see m the sequel that we
can only find out the true nature of mference, and even the
true meaning and pos1t1on of the syllogism itself, by cons1dermg
the actual methods of the mathematical scICnces. Anything
which could be called a general logic of inference can only arise
m this way, and m the general treatment of the subject a serious
error has been caused by the attempts wluch have actually been
made to find a qmte gem·ral process of mference uased upon
mere abstract c-ons1dcrat1ons about the general form of proposi-
tions.
On the other hand, the sc1encc'l, for 1m,tance, wluch arc
demonstrative, while they have certam common characterisucs
of method, have special methods of their own, wlnch they them·
selves discover These depend upon the ,;pec1al .!>UbJect-matter
of the sciences and therefore could not possibly be anticipated
by the log1c1an, or derived from any general characteristics of
inference. These admit of logical study, which may be rightly
called special logic m contra,;t with the study of the common
characteristics of inferencl', which may be called general logic,
or a part of general logic But, m the nature of the case, such
special logic would not be an ,3.pplicat10n of the general logic of
inference, nor derived m any way from 1t.
Further, a particular science or set of '>Clenccs may employ
some special conception which 1s not found m other sciences.
Such conceptions may be presuppos1t1ons of the science or
sciences, but are m no sense mvestigatcd by the sciences them·
selves. Their exammat1on, then, would belong not to scientific
thmkmg, but to what may be called, in general, reflective
thinkmg or philosophy, though not to that form of reflective
thinking which is logic. For these conceptions, of which cause
JS an example, as they belong to the nature of the object, are not
logical conceptions (which belong to the apprehension of the
object} and, accordingly, they are not the subject of logic but
of the kind of philosophy called metaphysic<; Logic, however,
has to attend to these conceptions m order to see whether they
may determine anything m the form of the apprehension.
 IV
LOG IC AND THEORIES OF KNOWLF.DGE
AND REALITY

§ 24. To return to our provisional definition of logic as some

kmd of study of thought. There are other studies besides logic
which may be said to deal with thought as their object or at least
to mclude a study of thought; such studies for example as Gram-
mar and Psychology, of which we spoke m the last chapter. There
is however a serious and perhaps unexpected difficulty which
threatens to confuse the provinces of logic and all science whatever.
This difficulty arises from the distinction often implied when logic
is defined as the study or a i,tudy of thought, the distinction of
thought from thmgi, or from reality. It probably seems that this
distinction enablei. us to separate off logic from the mathematical
and empmc.tl sc1ence1> For wlule the sciences think about
things, logic tlnnks about thought. Yet tlus d1stmction apparently
presupposes that thmgs are somehow in their essence independent
of conscwusnei.s. It rests doubtless upon the ordinary and
popular attitude to the world But a doctrine has appeared in
philosophy which makes it impossible to acquiesce in this
apparently obvious distinction without very careful considera-
t10n. That doctrine is Idealism. SubJective ideali,;m, like that
of Berkeley and Hume and their more recent followers, may
mdeed be said to identify the reality and the thought (as thought
is understood m this school), for it makes the perceived thmg
entirely a part of our consc10usness and existing so long only
as some one 1s conscious of it. Absolute idealism, hke that
wluch appears m the writings of Green, professes not to make
reality subJective 111 this sense and yet holds that reality is
somehow identical with or constituted by thought If this were
so, the distmction between logic and the sciences, which seems
so clear, could not stand m the form in which 1t has been stated,
for science as studying thmgs would necessarily be studying
thoughts.
 Theo,ies of Knowledge and Reality 6I
§ 25.• It may well be supposed that such difficulties exist only
for idealism and are no concern of those who do not accept it
m one or other of its forms. Yet we certainly cannot escape
them by adopting the ordinary realism, which is usually opposed
to 1t and gets phdosophic expression m the works of such a writer
as Locke. This reahsm, though it holds that things are entirely
different from our thought and independent of 1t, 1s associated
with a theory of how we come to know the existence of thmgs
and their properties, which ends in the adm1ss1on of what it set
out to deny. The mmd, it would be said, gets knowledge of
things different from itself, or any state of itself, and the thing
1s certainly not a state of consciousness How, then, do we
know the existence of such thmgs ~ Knowledge arises from
perception. Perception again 11, described as an effect 10 our
consciousness of the action of things on our bodily organs. This
effect is what Locke calls an ' idea ', and it 1s sometimes vaguely
called an ' 1mpress1on ' Why is this knowledge ? Because this
effect (or 'idea') representc; the object, or (better) JS a copy 10
consciousness of the object We naturally ask how we know
that it 1s a copy ; for 1t 1s not enough that the ide.t should be
a copy, we must know that 1t J'i a copy or 1t would be of no
use for the purpose We can only know that a copy 1s a copy
by compar10g Jt with the ongmal, and lo do th1,; we must
apprehend the onginal itself But, ex ltypothest, 1t JS the copy
only which i,; before us m consc1ousntss, and further, if the
original were before u,; and apprehended, the copy would be
superfluous Again, the theory cannot guarantee that the sup•
posed effect m our mmds, the ' idea ' or ' 1mpress10n ', is at all
hke the thmg which 1s said to cause it ; and, worse than this,
1t cannot guarantee that there 11, any original, and, worse again
than this, 1t cannot even account for our havmg the idea that
there 1s an original Thus ordmary or empirical realism ends
m the subjective 1deahsm 1t was mtended to avoid. Indeed, 1t
was by reflectmg on Locke's theory of knowledge ,ind dcvelop10g
its consequences that Berkeley was led to hie; own doctrine.

[• Wllson gave much thought to thts apparently simple section He said

' I have tned to put what seems the true ground of Bcrkeleiantsm, though
Berkeley himself never, so far a!. I know, got h1~ own position with such
clearness • (Letter to F H Hall, 7 v1 14) ]
6.z INTRODUCTORY
For our purpose these difficulties are not so interesting as the
fact that this theory after all unconsciously involves the con•
sequence, which subJective idealism expressly states,-viz. that
the only ohJect the mmd can apprehend, so far from being
something independent of consciousness and ' outside ' it, is
somethmg which cxic;ts only within consciousness as a state of it,
as much indeed a mere state of it as is pleasure or pain. This
result leads to a more general prmciple with which it is not
1denttral-thal the real obJcct is not directly present 1 to conscious-
ness, or that we c1re not conscious of the • real ' The empmcal
theory of reality mdeed, affirming that the existence of things is
wholly mdepcndent of thought, necessarily leads either to sub-
Ject1ve idealism or to sccptlClsm of the obJect These things are
not obJects of consciousness , we never can be conscious of them.
§ 26 Ordmary reah'>m then re<;ults, through Locke's philo-
sophy, m .i theory winch necessarily mvolves the assumption
that c1,ll ohJc-clc; ,,c- seem lo be conscious of arc merely mental
existences If \\e c,,1y th.it what I'> dtrt:ctly present to conscious-
ness must be ' within ' consc1ouc;ness, there lh a danger that
this should he taken to mean th,1t what th directly present to
consc1ou'>ncss 1s c1 state or a pc1rt of consciou<;ncs~. So Berkeley
takes 1t for granted that what we arc directly conscious of (or,
as he would say, ' perceive ') ts merely mental The only
question bet\\-cen lum and Ins opponents 1s whether there 1s
somethmg else besides this merely mental ' somethmg '. And
as Locke, whom Berkeley has m view, docs take 1t for granted
that what 1s immediately apprehended 1s merely mental, Berkeley
1s Justified m getting from this assumption the general conclusion
he ,unvcs c1t, though he c1rnvcs .i.t it, 111 fact, by an unsound
argument.• The \\ ord ' withm' 1s a metaphor from space, and
1 The phrase • dlrl.'l tly present ' 1~ given for purpobeb of polemic, as that

which would probably be used In 1lbelf 1t 1s liable to an obvious cr1bc1bm

[• Referrmg to ' But, t.ay you, though the 1deat. themselves do not exist
without the mmd, yet there may be thmgs hke them whereof they are copies
or resemblances, which thmgs ciust without the nnnd, m an unthinkmg sub-
stance I answt•r, an 1ded can be hke nothmg but an idea, a colour and
figure can be bke nothmg but another colour or figure If we look but ever
so little into our thoughts, we shall find 1t 1mposs1ble for us to conceive
a likeness except only between our ideas ' Berkeley, Pr,nuples, I VIII. The
al'((ument depend~ upon the assumption that two thmgs hke one 8.llother
must h.i.ve the same J.md of cxu,tence]
 Theories of Knowledge and Reality
the statement that what is directly present to consciousness
must be within consciousness is simply a defimtlon of what the
word ' within ' means m this reference to consc10usness. The
word ' within ' adds nothing to the idea of ' direct presence to
consciousness', and cannot therefore imply in addition that this
means state of consciousness, unless that is already necessitated
by the idea of direct presence to consciousness. Sub1ect1ve
1deahsm then holds, whether this is expressed or not, that what
1s directly present to consciousness-what we are ' 1mrned1ately
conscious of '-is something mental and mdecd a state of the
i.ubJect's consc10usncss, and this is treated as somcthmg self•
evident and merely to be taken for granted Absolute idealism,
on the other hand, though not makmg the obJcct identical with
subJect1ve thought, appears to mJ.kc 1t essentially a realization
of thought, of thought conc;1dcred as somehow obJectlve. And
thus 1t c;cems common with Absolute 1deahsts to speak of the
obJect as ' content ' of thought •
§ 27. Something of the same sort appears where we should
little expect it,-10 our ordinary way of sprakmg about thought;
that is m language not tcchmcal nor peculiar to plulosophcrs,
but m.thc ordmary and normal u.,e It 1s not due to the con•
sc1ous adoption of or c-vcn to the mflucncc of any form of 1dcahsm
We d1stmgmsh the act1v1ty of thmkmg from 'what we tl11nk' ;
the latter we regard ai:. a part of the whole fact of thought, and
mdeed the most important part of 1l By our •thought' we
certamly mean 'what we thmk ', as well as the thinki11g of it,
and indeed mamly the former At the same time, while we say
we thmk about thmgs, we usually d1stmgmsh the thoughts which
we say we think from the things we thmk about. Certainly, m
the ordmary unphilosoph1cal thmkmg to which this ordinary
f• e g B BO!,anquet, Essentials of I.og,c, rh 3, § 2, where the matter .i.nd the
content of knowledge are identified by contra:.t with the form • We are chiefly
mterei,ted to know what a thmg 11>, viz its content, which 11, no more objective
than 1t 1s sub1ecbve,' Hegel, Logic, § 42 (3) 'Content' lb a tran~lation-~oni
(lnlialt) 1n Enghsh and 1~ therefore bkely to be found 1n the later English
1deahsts In German • Inhalt • m this sense 1s not confined to 1deah~t philo-
sophers, e g • Das Abb!ld selbst als das Resultat der Erkt'nntmssthabgke1t
1st der Inhalt der Erkenntmss • Uebcrweg, Logik, § 2 "The word 1~ now
commonly used m Engh~h to indicate what any one 1s thmkmg, without
further statement of 1b, elements and without begging the question of the
relation of the mmd to its obJet..t)
 INTRODUCTORY
and normal language belongs, people do thus distinguish ' what
they think' (or the so-called 'content' 1 of their thought) from
the thmgs they think about. But what are the consequences
of this language, c;trictly taken? It would ordinarily be said
that the thoughts which we think are not the things-we do
not think the thmg, but thmk about 1t. But what do we think
about the thing ? What 1s this somethmg which is not the
object or thmg, but 'of ' or 'about' the obJect? If it 1s no
part of the nature of the obJect, 1t would seem ac; 1f to thmk
it ' of ' or ' dbout ' the ob1ect would be a delusion (Nor can
we avoid this by making our thought-1 e • what we thmk '
about the thing, a mere mental copy of the thing's nature, for
what we think of the thmg 1s not such a copy-e g what
I think of the red obJect ts its own redness, not some mental
copy of redness m my mmd I regard tt as havmg real redness
and not as havmg my copy of redness ) If v. e ask in any
instance what 1t is we tl11nk of a given obJcrt of knowledge, we
find 1t always conceived as the natUie or part of the nature of
the thing known What we tlnnk of a thing which we know
is what we apprehend in it and must be part of its nature. If
it were not, we should not be knowing. It follows, then, that
1£ we continue to hold that ' what we tlunk' about a tlung 1s
a part, and a main part, of the whole activity of thinking, we
must also hold that this rnam p.art of our thought or thinkmg
1s exactly the nature or part of the nature of the thmg thought
about. Now, whether this leads to any form of 1deahsm or not,
tt certamly did not ongmatc m 1deahstlc philosophy-on the
contrary, 1t comes from the very way of spe,1king which is
natural and habitual with those who do not believe m any form
of idealism It is independent of any theory we may hold as
to what constitutes the reahty of the obJect-whether e.g. it
is merely existent m consc10usness or not It 1s simply the
result of callmg • what we thmk' (or the defimte somethmg
which \\re thmk) 'thought', or 'content of thought '-1 e. of
makmg 1t an mtegral part of the \\hole activity of thinkmg
§ 28. • Now, if we have been led by such cons1derat1ons as the
1 Observe that ' content of thmkmg • 1s metaphoncal It adds nothmg
and may mislead

(• Ongmally a cnt1c1sm of Mansel, Prolegomena Logica, ch vu ]

 Theories of Knowledge and Reality 65
above to include the object somehow in thought, we should
have in some way to recover the distinction of subJecttve and
obJective, which is really the main inspiration of that empirical
theory of reality which otherwise seems so inadequate In the
nature of the case there would be no object which 1s mere
object, for under the present hypothesis the obJect for us would
be, as an aspect of or clement in actualized thought, the obJect
of a subJect, the one bemg correlative to the other. The two
sides, then, would not be d1stmgu1shcd any longer as thought
from not thought but taken as const1tutmg together m their
inseparable unity the actual particular thought. The real
thought would be not one side of this relation, but both sides
together ; the clements, though mscparable, being nevertheless
distmct. 1 It would follow that we cannot found the d1stmct1011
of logic from science on the d1st111ctlon between thought and
something which 1s not thought, but on some d1stmct1on sub•
s1c;ting within thought itself Now there 1s a distinction on
which the definition of logic 1s sometimes founded, which seems
(whether intentionally or not) to be of this kind. Logic 1s some-
times said to deal with the ' form ' of thought as opposed to
its 'matter'. These terms are metaphors and cannot be a final
explanation of anything, but the d1stmctlon has the charac-
ter1st1c that it appears to be made w1th111 consciousness and
w1tlun thought. It probably ongmated from an mstmctive
tendency to put ' what we think' w1tlun thought or thmkmg
itself, but also, probably, without any clear consc1ousnus that
this might result m puttmg the object itself w1thm thought.
The total reality, then, of thought would be this mseparable
umty of subject and object, and thought would be not one side
of this rclat10n, nor the mere attitude of the subject to the obJcct,
but both sides in their umty. On this view, the d1stmct10n
between logic and scicnce would be that science studies the
obJcctivc side of the whole reality of thought and logic the
1 Such elements m a whole must not be confounded with parts, i e parts
of an aggregate, for that generally implies that such parts can exist without
one another Anstotle formulated this lond of umty of elements which are
yet d1Stmct m the well-known phrase----lo-T, µiv TalrrJ, Trl 3' ,lvai ain-oir
fnpo11, 1 e mseparable m existence, but d1stmct 1n essence [Quoted loosely,
the nearest expression being EO'T& ,..~,, ol'iv To.6r6v, Trl r,• rT11a1 f71po11, De
An,ma, 4248 25, cf E. N I ?Joi' u, al]
~HI F
66 INTRODUCTORY
subjective side (the content of the individual thought) in abstrac-
tion from the objective side. And here we see perhaps some
room for the application of such metaphors as form and matter,
for neither element is real in abstraction from the other. These
terms however must be used only as convcmcnt abbreviations,
not as supposed explanations. It is interesting to notice that
they would aiise quite naturally from that trndency to identify
\\hat we tlunk with thought, a tendency which seemed so inno-
rcnt anrl whose conscquencer; we did not ant1c1pate. They are
however, strictly ,;peaking, tcrmc; proper to ::in idcahstic view
only and should therefore be av01<led unlcs<, we hold such a view.
It makes uo difference v.hcthcr they were or were not directly
intended to be the result of that view.
§ 29. The <l1stinct10n thus arrived at between logic and the
sciences not only follo\\ s from the implications of that ordinary
view of the content of thought wluch we had under cons1dcra-
tion, but it 1s the one v. h1ch suits any form of 1<lcahsm.
Moreover on.hoary language, wh1cl1 no one tlunks of challeng-
ing, appcarc; to involve as its consequence 5omr sort of identifica-
tion of the obJcct \\1th the thought, an 1dentificat1on certainly
not suspcctecl by those ,, ho use tlus language, and so ,ve naturally
turn to reconsider tlus language itself
The consequences arc really drawn from language by help of
the view that \\ hat we, in knowing, tlunk of the object 1s what
we apprehend m the obJect and must belong to the object
itself This view can hardly be quest10ned and we have seen
the contrad1ct10ns involved in denymg 1t. We have then to ask
whether ' v. hat we thmk' should he called ' thought ' in the
ordmary way-1 c not as :,omcthmg merely on \\ luch the
activity of thought 1,;; C"-crr1scd but of the nature of thmkmg
itself, and wholly compnsrd w1thm that nature. That 1s the
only proper meamng of the metaphor ' content of thought '.
That the ordmary ,1c\\ 1 though so fam1ltar, involves remark-
able conscqucncrs, may be seen by an analogy a When we grasp
an object m our hands, though graspmg cannot exist without

[• Wilson frlt the difficulty of the metaphor of• grasping• Fmdmg a remark
on 1t m a pupil's note-book, he wrote a Ion~ note on the question wluch I do
not reproduce. An~totle used 9,-,Ei" (to touch), Metaph w51b 24. Cf Reid's
Inq1my (ed Hamilton), 2, § 3 note, and note to § 34 infra]
 Theories of Knowledge and Reality 67
an object to grasp, we not only distinguish the object from the
grasping of it, but we never think of calling 1t a part of the
whole fact or activity of grasping. Nor do we represent 1ts
being as wholly comprised in that of grasping. Is then ' what
we think ' to be called ' thought ' in the ordinary sense, which
means ' content of thought · and 1s 1dent1cal with actualized
thought? In favour of this 1t may be said that 1t seems hkcly
to be right, because 1t is the natural and universal mode of
expression m ordmary untechmcal language, ancient and modern.
Tlus seems evidence that the distinction is not artificial, as the
mclusion of the grasped obJcct m the grasping ccrtamly would
seem to be.
!wen m the case of graspmg 1t might be said that though the
body grasped 1s not a part of the whole actuality of grasping,
yet its resistance to the pressure 1s inseparable from the
grasping and only comes mto existence when the object is
grasped
But the strongest argument m favour of what appear to be
the impltcat10ns of the ordmary use of language seems to be this.
If we say that thought proper (m the case of knowledge) is
nothmg but the apprehension of the obJcct, 1 e. the apprehension
of • what 1s thought', and that • what 1s thought' (i.e. the
nature of the obJcct) is not of the nature of thought or apprc•
hens10n itself, tlus abstraction of what 1s apprehended from the
apprehension of it, of what 1s thought from the thmking of 1t,
seems to make the act of thmkmg or apprehension en1pty and
meaningless. This argument is perhaps more convincing 111 the
case of a umvcrc;al proposthon, where we are not so much
influenced by the customary oppos1t10n between the mihvidual
tlung perceived and our perception or thought of tt. We apprc•
bend that the product of two odd numbers 1s odd. If now we
abstract what we apprehend here, the essence of the act of
apprehending seems to be gone. Hence the nature of \\-hat we
thmk seems m this instance to belong essentially to the nature
of thinkmg. There 1s a kmd of parallel to this in feeling ; there
is no feelmg apart from the defimte quality of what 1s felt-say
heat or cold-the idea of feeling seems altogether empty if we
abstract what is felt, the quality of what 1s felt ; and here, at
any rate, our ordmary attitude, whether nght or not, is to make
F2
68 INTRODUCTORY
the quality a part of the whole reality of feeling, and without
any hesitation.
Of these three considerations, the second appears to be of no
value, because this property of the object, viz. its resistance,
only gets reahzcd in the fact of its being grasped. The corre-
sponding feature in apprehension, if the argument is to be
relevant at all, would have to be that the properties of the
ob1ect known should only come mto ex1:;tcnce m the sub1cct1ve
act of apprchenclmg them That however would be m plain
contrad1ct1on not only to the ordinary view, with the language
of \\ luch wc arc at present concerned, but to the nature of
sc1cnt1fic knowledge. We have, therefore, to turn our attention
to the first and tlurd cons1dcrat1ons
§ 30. To hegm with the firr;t cons1dcrat1on, derived from the
natural moJc of lingu1st1c cxprcss10n It "111 prove in the end
that the phrase ',,hat we thmk about a tlung' is ambiguous.
There 1s a sense m which ' v. hat WL tlunk ' must be a part of
the activity of thinking and entirely '"1thin the being of thmking.
Let us see what is charartcn-;tic of the c-.1sc where it is so
natural to call' "hat \\e tlunk ', thought That seems to come
out mol:,t dearly Ill the kmd of thmkmg wluch 1s called opinion.
Suppose v. c arc of op1111011 that A is B, and it turns out we are
wrong. It would be said, m the natur..il use of words, that what
we thought was untrue · or, agam, that what we thought about
A was untrue Herc ' what we think ' cannot be the nature of
A, for the nature of A cannot be wrong or untrue. So far then
we have a fairly ckar distinction to JUsllfy the contrast between
'thinking about the tlung' and • thinking the thmg ', and
between' what we thmk about the thmg' and 'the tlung '.
We should say that '"hat v.-c thought' was untrue 1f the
opinion was wrong But then, if tins 1,; so, there must after all
be included m ' \\ hat we think ' the thmkmg, because 1t 1s only
the thinking that can be wrong ; and the fact 1s that what 1s
here termed 'what we thought', and said to be untrue, is a
particular act of thmkmg, in ,vh1ch the thinking is inseparably
included. It differs from thmkmg if thinking stands for a uni-
versal : for it differs from thinking 111 general as being a particular
realization of thmking. In the case then of opinion, it is easy
to understand how ' what is thought ' 1s included in the activity
 Theories of Knowledge and Reality 69
of thinking in a way in which what is grasped is not included
m the activity of grasping.
The case where thmkmg means knowledge is obviously the
most important, for, as we have already seen, whatever is
mcluded under thinking depends entirely upon the notion of
knowmg. In this case also when we say that 'what we think'
is • thought ', or ' what we tlunk about anythmg ' 1s ' thought ',
we mtend, of course, not merely that the thinking of it 1s
thought, but that ' what we thmk ' is ' thought' m the sense
of that which makes the thought a real indiviclual thought as
opposed to the empty form of thmkmg m general. We should
also say that tlus thought ( :s what we thmk) is true masmuch
as it is knowledge, and consequently, as before, whether the
thmkmg 1s knowmg or opm1on, it follows that it necessarily
includes our subJectlve act.
It makes ~o difference, then, '" hether the thinkmg 1s knowing
or opm1on: the whole point 1s that we do attach these epithets
' true ' or ' false ' to ' what we thmk '.
Now, whether we know that A 1s B or form the opinion that
A is B, what Is the accurate answer to the quest10n ' What 1s
It we thmk about A? ' The accurate answer 1s, 'That A is
B '. With regard to the expression (observe) we can rightly
say ' that A 1s B is false', or ' that A is D is true'. This shows
that the expression ' that A is B ' 1s not equivalent to ' A's being
B' : for A's bemg l3 represents a fact which cannot be true or
false, so that when the tlunkmg 1s knowledge, and therefore
thmkmg A 1s B is true and is knowmg that A is B, the expres-
sion ' that A is B ' is still not equivalent to ' A's bcmg B ',
or 'the B-nec;s of A'. What we thmk about A 1s, m fact,
properly understood, always ' that A 1s B '. This, then, seems
to be the accurate expression for th.:it upon which so much
depends-' what we thmk about A'.
Our previous difficulty arose from makmg what we thmk of
A equivalent to the obJcct1ve B-ness or the ol>Jectlve ' A's
bemg B '. Can we say m any sense that m thmkmg A 1s B
(whether the thmkmg 1s opmion or knowledge) B-ness or A's
bemg B 1s what we tlunk of A ? The answer seems to be that
. we can and do, and yet that m truth the expression 1s not
accurate but abbreviated and idiomatic. It has mdeed, in the
 INTRODUCTORY
last resort, to be explained as just meaning that we think that
A has B-ness or that A 1s B ; and consequently, when accurately
expressed, this explanation 1s precisely the same as the one
already given.
So far, m examining the identification of ' what we think '
with ' thought ', ' thinking ' has been used m the widest sense.
Dut our invest1gat1on of the various uses of the word thinking
has shown the importance of asking m every given case what
kind of tl11nki 11g 1s meant, when any such quest10n as the one
before us 1s raised about thmking In the case of op1111on, we
got our answer by considering precisely the kmd of thmkmg
meant-viz. opinion. In the case of knowing that A is B, our
thinking is the apprchcnc;1on of A's bcmg 13, or of the B-ness
10 A; and' what we thmk' of a tlung wluch we know 1s what
we apprehend m it and muc;t be part of its nature 1 If what
we thmk were accounted to be thought and mrludcd in thinking,
this would have to mean that what wc apprehend, the fact
of A's bcmg D, or the D-ncss of A, 1s mcludcd in the apprehension
as a part of the activity or reality of apprehcndmg. Obv10usly
this bst statement woul<l be quite ag,unst the usage of language :
and 1t is the usage of language that we arc cxammmg.
This shows that the use of language when properly understood
implies an 1dcnttficat1on of ' what we thmk ' with ' thought '
only when 'wh.i.t we thmk' 1s taken in the general sense m
wluch 1t mcludcs opmion, and 1s not the nature of the object
thought. Accurately, then, the ' what we tlunk' 1dcnttficd by
implication with thought m ordinary language 1s not a 'what
we think' abc;tr.t( tL·d from the tlunkmg of 1t, but mcludcs
mscparably our su1i1cctivc act of tlunkmg- 1t
§ 31. \\'e come now tu the tlurd conc;1<lcrat1on, for the pre-
ceding has by no mean,;; d1o;;poscd of that.
In kno\\ mg the n.iture of A as B, wc arc apprehending an
obJcct, ,md ' 'i.£ 1/iat we apprehend ' 1s an obJcct, and the apprc-
hcns1011 seems empty and no apprehension at all, 1£ we abstract
from 1t \\ hat 1s apprehended. Take that a\\ ay and what 1s left ?
Tins naturally mclmcs us to tlunk of what is apprehended as of
the bemg of the apprchendmg This time, undoubtedly, what
is apprehended 1s the nature of the thmg.
' § 27
 Theories of Knowledge and Reality 71
Nor is this a mere theoretical development, for, as a matter
of fact, people do vaguely thmk of the • content ' of thought or
apprehension as being the main part of the reality of the apprc•
hcnsion itself: and persistently, m consequence, distinguish this
from the thing. To solve this problem then we shall have to
appeal to cons1deratlons of a more general kind, which are not
confined to the relation of thmkmg and its obJect, but apply
m general to things which arc related.
In popular thmkmg, at all times, there 1s a tendency to treat
individual existences as independent rcaht1cs and, though the
fact that they enter mto relations is recogni1,ed clearly enough,
to regard their rclat10ns as something external to the nature of
the things rcttterl, not as bclongmg to their own bemg. Tlus
which 1s 1mphc1t m the ordmary consc10usncss becomes explicit
m the bcgmnmg of philosophy, and it 1s what we find m
Aristotle a Rclallons arc d1stmguishcd from tlungs and are
excluded from the bcmg or essence of the individual thmg. In
many ways Locke in modern times remmds us of the position
of Aristotle m relation to his own day. For Locke may be
described as a plulosopher who makes explicit the tendencies
and 1mphcat10ns of ordmary popular thmkmg. It seems fair to
say of him that he treats the md1vidual substance as thus remote
from relations There 1s, m fact, a somewhat sharp distinction
between substances and relat10ns both m Locke and m Aristotle.
Modern metaphysics u has revised tlus view as 1t has become
clearer about the connex10n of tlus supposed md1fferent relation
\\ith what 1s called the essence of the thmg, and has come to the
conclusion that the severance of the two 1s artificial But, in
so domg, 1t has tended to advance to the extreme opposite of
the view it crit1c1zecl and more or less consc1ously to break down
the <l1stmct1on bet,-.ccn tlung'> related altogether.
It seems quite wrong lo exclude from the being of a tlung
anythmg wluch 1s necessitated by what we have t..ikcn to be
its bcmg, anythmg from which that is in">eparable For the
latter 1s a ncccsc:;ary clement 111 the complete ..iccount of the
[• In earlier lectures ' Plato and Aristotle '
b 'The general metaphysical theory • • overstatement•, p 72, refers to
Green, e g Wllf'ks, 11 170-2 Hts reaction against this view was fundamental
but he d1d not emphasize it in reference to Green. Cf. Part V, I 564.J
 INTRODUCTORY
thing's being, though it may be different from that aspect of
the thing's being which we first had before us. Suppose A and
B are in a relation R. R is nothing without the special nature
of A and B, and th11s, 1f we follow the tendency above described,
we should say that 111 the being of R, in its completeness, must
be included that of A and B, and similarly B would be included
in the bemg of A TIHS would have to be the case even if R
were a temporary relation Suppose two bodies come into
collision ; the colhs1on 1s nothing "1thout the bodies of which
1t 11:, the colhs1on, and its special n.iture as th1s collision depends
on the special nature of these bodies On the other hand,
R itself similarly belongs to the special nature of A and B, so
that the nature of A or B in tlus way would be made to include
that of R
The apphcat10n of th15 lo the relation of the apprehending
and apprehended 1,; obvious We should be obliged to say that
the bemg of the tlung ".l!> included in the bemg of the relation,
the bemg of the t hmg apprehended somehow mcluded in the
being of the ,tpprchcns1on. We shall find that the general
mclaphys1cal theory wluch would lead to such a result 1s an
over:.ta.tement
If A stands 111 a relation R to B, even 1f R be temporary,
the bcmg of A 1s not mdependent of the bemg of B That 1s
true and accurate. Again the complete account of the being
of A must include the relation to B. That again 1s true and
accurate. But must we, therefore, include the being of B m
A, as seems the tendency m some modern metaphys1cians, and
would the use of ' include' here be accurate ~
The key to the ans\\cr 1s the fact that A and B, however
closely rcl.itcd, arc different and not to be entirely identified
with one another. The dependence of the one on the other,
however absolute, cannot destroy this difference. If now 111clu-
s1on 1s taken in a strict sense, as 1t ought to be, and the being
of B 1s to be mclude<l 111 that of A, for the same reason we
!.hould have to include the being of A in that of B This would
result in a contrad1ctlon, which would only be avoided 1f A and
D were ind1stingu1shably the same
Again, 1f "c take the ' bemg of A ' 111 the Jense of what A is,
then what is part of the being of A must be a part of what
 Theories of Knowledge and Reality 73
A is. So we say 'heavy' is a part of what lead is, since lead
is heavy. Now, if this 1s a particular quality or kmd of thing,
X-ness, we must be able to say that A is X or A is an X, where
X is an adjective or a common term. If, however, X is a parti-
cular hke A itself, we must be able to say simply A 1s X, e g. that
flower is the one I bought yesterday. But if A is related to B,
we cannot therefore say that A 1s B, or A is /3 where /3 1s an
adjective or common term corrcspondmg to B. It 1s not accurate,
then, to say the bcmg of B 1s mcludccl m that of A, or the bemg
of A m that of B. On the contrary, m so far as we say A 1s
m the relation R to D, and m so far as it is the nature of A to
stand in this relation, A's bemg m the relation to B may be
said to be a part of the bemg of A. But A may have a being
other than this relation, a bemg not mcludmg it ; c g. if we
say A is near to B, this implies that A and 13 arc something,
have some nature, wlnch 1s other than this rclat10n, for mstance,
that A 1s a tree and B a house. Moreover, 1t 1s this nature,
other than the relation and not mcludmg it, which is what we
mean by A or B, and 1f we were asked what A and D were, our
answer would be a descnpt1on of JUSt tins nature In short, the
relation only obtams between terms winch arc different from one
another and as havmg natures wluch do not mcludc the relation.
Consider now the bcmg of R, the relation between A and B.
Tlns is relevant to our 1mmcd1atc prol,lcm, smce the appre•
hendmg of the obJcct may be said to be a relation between the
thmkmg subject and the object.
Our difficulty was that 1f we abstract what 1s apprehended
from the apprehens10n of 1t, apprehension seems empty and
meamnglcss. We saw, for instance, that 1f we abstracted a uni-
versal proposition apprehended from the apprehension of 1t, the
essence of the apprehcndmg itself seemed gone Thus the nature
of what we thmk or apprehend might seem to belong to the
nature of thmkmg or apprehendmg. Now, 1t 1s true of a relation
m general that 1t is inseparable from the terms rcl..i.tcd and that
1t seems empty and meaningless 1£ we abstract these terms from
it. Equality is essentially the equality of two tlnngs and we
cannot leave out the things. But docs 1t follow that we should
therefore regard terms related as havmg their l>emg mcluded m
that of the relation ?
74 INTRODUCTORY
Let us consider the example of a collision. This is nothing
apart from the bodies which come into collision : it is inseparable
from them. Abstract the bodies and the collision is gone also.
However, the very nature of the collision between two bodies,
A and B, ncccss1tates itself that A and B should be different
from one another It also necessitates that A should have
a being other than bcmg in colhs1on \\ 1th B, and 1t 1s only as
havmg c;uch bemg that 1t can enter mto the given relation with
B. Agam, while the bcmg of A is not mclu<lcd m and is not
part of the rolhs10n, on the other hand (m that wider sense in
which the bcmg of A is made lo include everything which
happens to A) the colhs10n would be part of A's bemg. But
1t 1s no part of that bcmg of A which 1s 1<lent1cal m all that
happenc; to 1t, and wluch would be called what A 1s ' m itself' .1
And the latter bcmg, wl11ch excludes the given temporary rela-
t10n, is the being already spoken of a,; necessary for A's entermg
mto the relation
We have, then, here a case \\<here n relation, though empty
and mcanmglcsg 1£ we abstract from 1t the terms related, 1s so
far from neccss1tatmg their mclus1on in itself that 1t necessitates
the contrary ; for 1t ncccss1tatcs that these terms must have
a bemg of their own which 1s not included m the bcmg of the
relation.
§ 32. This 1llustration seems enough to show that the insepar-
ableness of the apprchcm,10n from what 1s apprehended docs not
warrant the co11clus1on v. luch 1t seemed to suggest The truth
is, that Just as the coll1,;1011 with I3 1s only possible through
a bemg of 13 other than 1t!., rommg mto colh-.1on, and 1t is with
B as h,wmg sue h hl'mg that the colhs1on t..1.kcc; place, so also
the apprchcn<,1011 of an oh Jct t ,.., only poc:;c:;1blc through a bcmg
of the obJcct other I han its hcmg ,tpprchcn<lccl, and 1t 1s tl11s
bcmg, no part 1t 5clf of the apprchcn<lrng thought, winch 1s what
is ,tpprchcndcd.
Thus, 1f an obJcct 1!. apprehended, 1t does not follow that
merely because 1t 1s apprehended 1t must be a part of the nature
of the apprehension, that 1s part of the apprehcndmg conscious-
ness If that ,,ere so 1t would be entirely mental or, m general,
a state of consc10usncss. \\•hat 1s apprehended, or the obJect
I H 70, 71.
 Theories of Knowledge and Reality 75
apprehended, may be a state of consciousness, yet even then-
it would not be a part of the apprehending consciousness ; or
agam it may not be a state of consciousness. VVhich 1t 1s can
only be determined by an examination of the nature of the
object in question itself and certamly not from the mere con-
s1dcrat1on that 1t 1s apprehended in general, or is an obJccl of
thought. Y ct it 1s this consideration wluch 1s the sole basis of
such 1dcahsm as Berkeley's, and, one may ask, 1s there as yet
any system of idealism of wluch this 1s not true?
§ 33. We have seen m what sense the habit of making ' what
we tlunk ' belong to the nature of thought itself may be JUStlficd.
On the other hand, when' ,,hat we Lhink' means what 1s appre-
hcncled, what 1s thus thought or apprehended is an obJect wluch
1s not to be reduced to J. part of the bemg of tlic apprchendmg
thought Yet tlus 1s actually donl' by 1mplicallon when the
idcnt1ficabon with thought of 'what we tlunk' (or the so-called
content of thought) 1s applied to the apprehension of the nature
of the obJect. For the only thmg that c-an be found as' content'
of the apprehcndmg thought 1s the nature of the obJcct appre-
hended. The mistake however is not noticed because thinking
is still supposed to be tlunking something about the tlung and
not thinking the tlung, whereas, if the act1v1ty of consc10usness
1s to be called 111 tlm, case tlunkmg at all, it must be tlunkmg
the thmg
But now, m discussing the d1ffirulty about the belong"ing of
the ob3cct to the being of the apprehending thought, we have
been led to sec th.it conversely, m the wider sense of the being
of the obJcct, there belongc, to tlus bcmg the fact of its being
apprehended, and therefore the apprehension. As the way m
which the obJect bdong'l to the being of the apprchenswn docs
not rc<lurc the obJcct to terms of the apprehending thought, so
also th1'l bclongmg of the apprehension to the bcmg of the obJcct
does not warrant our reducing the bemg of the apprehenswn to
terms of the ob1cct nor make 1t of the same kmd Neverthcle'>s,
this latter is Just what 1s 1mphc1tly done m the old theory of
knowledge of thmgs through idcas-wl11d1 really still mfluenccs
people who would unhcs1tatmgly rcJect it when put to them
explicitly-for the idea in quest10n 1s notlung but a sort of
mental replica or reproduction of the obJect, sometimes called
 INTRODUCTORY
a copy, and the apprehending of the object is reduced really
to having such an idea; that is to say, is reduced to the existence
of such an idea in the mmd.
• Theories of kr.owledge and reality, in the futile attempt to
explain apprehension (1 e. to explam the absolute presupposition
of any explanation), have been much affected by these two con·
trary and one-sided tendencies, earh of them an overstatement
of the mterconnexion of the bemg of the obJrct and the being
of the apprehending thought ; the tendency to reduce the obJeCt
apprehended to terms of the apprehension of 1t, and the tendency
to reduce the apprehension to terms of the object. Now neither
can be reduced to the other ; neither expressed or explained
in terms of the other.
[• Cf Part V, §§ 54r-52 ]
 PART II

STATEMENT AND ITS RELATION TO

THINKING AND APPREHENSION
 I
APPREHENSION IN GENERAL
§ 34. WE must now endeavour to begin a systematic investiga-
tion w1th i;omc- unifying concept10n, remembering that we could
not ~cl c;uch a conception unksc; the material which 1t 1s to
unify were, m part at lrast, already before us We must not
expect to dctcrmme exhaust1vdy all the departments of our
inqmry beforehand, and we must also take into account the
begmnmgs of logic, because logic-, hke other studies, has developed
hH,toncally only from the c;olution of particular problems.
If we hold that the thing 1s not- to be identified at all with
thought, we can chstingmsh logic as a study of thinking from
science as a study of tlungs. Tlus, however, may be put in the
more general form that srtcncc- studies the object apprehended,
and logic the apprchem,10n,a a study wluch should include the
various act1v1t1cs called thinking . science, in fact, being a way
of apprehending ob3ect-,, an<l log1c a reflection upon apprehension
m general
Dcgmning w1th the idea of logic as a study of tlunking, we
at first looked naturally for a umvcrsal "h1ch might embrace
all tlunkmg. But the various departments of thinking have not
[ 8 Thi~ important krm 1~ nowhere definl'd by \Vtl~on, who appears to use

t a~ equwaknt to .\u~totle', 1•ciqa,r For long he used ' recogmtmn ' to

rxpre.,., the 11unmh.1tl' <'ogruzanc,• of lh<' ohJcct and eonv1tbon of 1u, being,
The dlfhlulty l,1tent Ill the won! 1., 1t, concealed mt'taphor and the fact that
smle 1t became an 1-.ngh.,h word 1t has t<>ndcd morr and more to mean belief
(sub1cct1ve) m \\h,1t may or m.1y not be real, e g apprchrns1on of death nnd
danger In N E D I l 5 1t 1~ dcfinNI a'! ' the ac.tmn of laymg hold of with
lhc 1,cnsr'I' and 1b 7 'thr artmn of gra-,pmg w1th the mtcl\cct' Sec also
s v m Mr Omons' A Shakespeare Glossary Price u~es 1t as a tcchmcal term
for ' the ~oul's Po\\cr of surveymg and cxammmg all thing~, 1n order to Judge
them, that IS, a power conven,ant about umversals and actively d1scemmg •
(Brdtih Moralists, Selby B1ggc, 11, § 593) Price sr.-rms to suppose that Plato
used ur'1Af/rf,ir m this sense, but 1t belongs to later Greek philosophy, especially
the Stoic The Latm word seems used by the Schoolmen as a general term
for the receptivity of the senses (cf notes to I§ 44 and 147, p 341) Lotze
u~es 1t for a sen~1t1vcn<'ss which precedes perception (Logic, 020) ]
 A pp,ehension in General 79
the kind of unity which this search implies ; they have not
a unity in the sense of a common universal of which they are
the different species. Nevertheless those of them which are not
knowledge depend upon knowledge ; they exist only through
the impulse to know and arc understood only through know-
ledge. Their unity with knowledge or apprehension and their
umty also with one another both depend upon their relat10ns
to apprehension. We arc thus led to the consideration of apprc•
hcns10n m general, both that wh1ch 1s perceptual and that which
1s not, as the primary subJcct of mvcst1gat1on, 1f we start from
logic as a study of thmkmg. We shall arnvc at the same result,
1£ ,ve start from the idea of logic as mamly concerned, as its
l11story shows it to have been, with inference
If then we make apprehension in general the starlmg-point,
tlus includes knowledge and 1s the key to the actlv1t1cs called
thinking. In this general form also we have a d1stmcl1on wluch
1s adequate to characterize the difference between logic and
scJCnce ;nd 1s not affected by any theory as to what constitutes
the reality of the obJcct, for it 1s based simply on the d1stmctlon
of apprehension and that which is apprehended ; and 1t docs
not obhgc us to decide between the confhctmg theories of reality
of \\h1ch we have spoken
§ 35 a This d1-,tmclton of logic from science is of great impor-
tance. Sc1cnt1fic thmking is csc,cnt1ally d1ficrent from any kind
of philosoplucal tlunkmg and the common hab1t of callmg logic
a science, wluch results from defining SCJ(,nce as method1cal
study m general, 1s to he deplored as obscuring one of the most
vital d1stmct10ns m the field of knowledge
In our ordmary expericnc<' and in the sciences, the tlunkcr
or observer loses himself 111 a manner m the parllcular obJect
he is pcrcc1vmg or the truth he 1s proving That 1s what he is
thmkmg about, and not about himself, and, though knowledge
and perception imply both the d1stinctton of the thmker from
the obJCCt and the active workmg of that distmction, we must
not confuse this with the statement that the thmkmg subject,
m actuahzmg this d1stmctJon, thmks exphc1tly about himself,
and his own activity, as distinct from the obJect.
The process may be desc~1bcd as one in which the thinking
[• Lotze, Logic, I 332)
8o STATEMENT, THINKING, AND APPREHENSION
subject, already realized m some activity of thinking, passes to
a further realization of this activity-this 1s the process from the
point of view of apprehension-, or we may describe it as a pro-
cess in which th" already partly apprehended object becomes
further apprehended or has some further opinion formed about
it. The sub1cct1ve clement in this unanalysed unity of appre-
hending and apprehended becomes afterwards itself an object
of consciousness 'flus is a new kind of thinking, which we may
call retlectivc, as dt,tingmshcd from sc1ent1fic tlunking and our
ord111ary thinking about objects, and comes into existence in the
conscious attempt to attain knowledge. For here the subject
distinguishes Ju-. own 1nc,omplctc state from the completer state
wluch he desires ; his attent10n is directed to himself and his
thmkmg act1v1ty, and he is able to ' abstract' lumsclf, as we
say, and tlus thinkmg activity in grneral, from the various
stages in which 1t is mamfcstcd. Tlus advance leads to the
abstraction of thmkmg as such, as our subjective activity, and
later comes the recogmt10n of specific forms which belong to this
activity and arc, in a sense, mdependcnt of any particular ob1cct.
It 1s the discovery of these abstractions wluch constitutes the
beginning of logic. Yct we mu&t not suppose that they are at
first made \\1th a comciousness of how they differ from the
abstract10ns which we make m our ordinary experience and
thmking. Thus, even when attention comes to be directed to
them, it may be found difficult to determine what constitutes
their peculiar character. It is understood that they arc some-
how specially abstract as compared with our more familiar
notions, but that 1s not enough. No process of abstraction,
however far 1t 1s earned, will get the properly logical notions
out of our <.onceptions of objects as "c have them m experience
and m the sciences. For, 1f carried to its extreme pomt, such
abstraction would end, say, in the mere abstract conception of
bemg in general, but would not take us mto the region of those
conceptions wluch properly belong to logic. This itself 1s decisive
evidence of the difference between the sciences and logic. The
reason 1s, simply, that such abstracllon proceeds entirely from
conceptions of the obJect known and cannot therefore bring us
to conceptions \\ h1ch arise from a cons1dcratlon not of the
object, but of the kno,ung of lt.
 A pp,ehension in General 81
§ 36. Now in the activity by which the subjective side naturally
first gets recognized-i.e. becomes an object of the reflective
consciousness-one kmd of apprehension of objects, viz. experi-
ence, docs not so naturally suggest the reflection on our own
subjective activity. The explanation is that the subJect here
seems mainly passive ; we seem to be acted on by the object ;
something comes to us without our making or seeking. We do
not then naturally inquire at first what arc the forms or laws
of experiencing, much less ask the qucstwn whether there are
rules for domg tlus properly, because it seems sometlung out of
our power. But there are processes of apprehension which
depend upon our own desire for knowledge and arc not expcn-
encmg (m the normal sense of the word), processes which we
originate and which we conduct, as distinct from the action of
the object upon us. Here 1t 1s that the rccogmtJon of our own
activity naturally begins, and it is to such processes, including
the mquirmg activity associatC'd with them, that the word
thinking as meaning an activity of our own is m ordinary usage
restricted. The rccogmt10n of such processes brings with 1t the
question. 'Can we find general forms wluch belong to 1t ~•
together with the further question ' Can we lay down rules
for 1ls safe guidance?' In this VI ay ongmatcs the study of
mfcrcnce.
It 1s thus that logic has, as a matter of fact, been ronccrncd
with the forms and rules of thmkmg m tlm rc5tncted sense and
not with those of every kmd of apprehension or of a~quirmg
knowledge, and we find that the study of inference has been
from the first the mam, if not the only, problem of logic •
§ 37. If we look to the actual starting-point of the part prc-
hmmary to inference m the Aristotelian logic and the anuent
logic founded upon it, we do not find 1t to be some consideration
of the general character of apprehension. The bcgmnmg is made
from an examination of the verbal statement or enunciation 1
of a fact ; and, smce the d1stmct10ns arnve<l at are those upon
which the theory of syllog1st1c mfercncc 1s based, we may suppose
1
rl•J<t>avrm, m Arn,totle e g De Int 17a 25

[• The last paragraph 1s exaggeratedly put ' rulu, • m tins sense Logic 1s
called an Organon, see§ 232)
2773•1 G
82 STATEMENT, THINKING, AND APPREHENSION
that the kind of questions to be asked about verbal statement
and the nature of its analysis were suggested by the interest in
discovering the general forms of inference. Inference, m its
verbal cxprc,;sion, bcmg made up of statements, the search for
a general form of inference leads to a determmatlon of the
general form of 1,t,1temcnt, with its cl1v1S1on mto species. And
the clement!> wluch arc d1'>tmgu1shccl w1tlun the form arc of
a kmcl suggtsted by the contemplat1on of a syllog1st1c argumcnt. 1
Nc,w the <l1stmct1on w1thm the verbal form of statement
arrived at with such an mterest m mm<l-a d1stmct1on sub-
servient to the ana]y1,1s of the syllog1sm-1s one which belongs
to the being of the object and not to our apprehension of it,
and so 1s not logical Tbat a distmctlon wluch belongs to tlungs
should be discovered through the medmm of the statement is
not surpnsmg, for the statement dec;rnhes the nature of the
thmg, and to the ohJcehvc d1stmrt10n m question corresponds
a verbal d1,;t mrtion ,, 1thm the gener.tl form of the statement.
It 1s quite nght that the d1..,tmc.t1on arrived at should be, not
a logical one, but \\1lhm the nature of the ohJcct, because it 1s
such a d1c;t111ct10n \\ Juch is nccc•c;c;ary for the purposes of syllo-
g1st1c mfcrcncc, a,; "ill .tppC'ar on e::,.,.an11nat1011 Y ct, though
not a log1cal but a mct..iphys1ral conception, 1t has a true place
111 logic because' 1t 1c:; the result of a log1r.il mqu1ry and 1s mtro-
duccd as ncc.cc;sary to the solut10n of a logical problem
But, though 1t 1s an ob1cct1w <l1stmct1on which 1s made and
used, the ordinary logic docs not realize thnt 1t 1s so and actuaJly
confuses 1t in tcrmmology with a truly log1cal d11,ttnction, that
of subJcct and prcchcah' Y ct the t\, o arc so chfTerrnt that
\\ luk the d1stmct1on m quc,;t1011 1s m <'l''-'mry to syllog1o;t1c theory,
the du,tmc t10n of suhJeC.t ,111<1 prc1h<'ate 1s ah<:olutdy useless to
th,\t theory. l'ar,ulox1ully t·nc,ugh t hL· tc-rm111ology of suh1cct
and pn•chratl' 1s ah, a) s used m the prc..,l·nta t10n of the syllog1stic
theory, "1thoul ho\\cvcr producmg conf u!>1on The reason is
that the erroneous termc:; ,\re Ile.\ er usC'd 111 that theory m their
proper senc;;c, but arc confined to the ohJcct1ve <l1stmct1on, and
what 1s really me.mt 1s made rk.1r C'nough for practical working
by the symbolism adopted. ~ot uncommonly those who apply
a pnnc1plc rightly 111 particular cases arc unable to give a correct
• Thi~ pomt 1i, rC'i,umcd m §§ 61-2
 A pprekension in General 83
account of its general character. The confusion in this case
comes from a mistake about the meaning of verbal forms m
relation to 'predication', and the exact nature of the objective
d1stmct1on of wluch we have been speaking will have to be
reserved for a special 111,·cst1gat1on, where \\ e shall show how
the confus10n makes itself felt m a noteworthy utterance of
Aristotle's upon predication 1
It 1s mdeed possible that the idea of a general form of state-
ment and of its analysis mto such ekments may have been
growmg up already through the interest taken by Greek thmkers
m the metaphys1c..il questions of the umty of a thing m its
attributes and the umty of the universal m its particulars. To
the former correc:;pond:, Ill statement, as Plato says m the Sophist,
the calling of one thmg by many names; to the latter corre-
spondc; the callmg of many tlungs by the s.1me name. Never-
theless the analysis of the syllogism probably brought with 1t
the precise formulation and actual symbolism, for v,e do not
find this before tl,e '"ork of Aristotle, and it is connected with
the characteristics of the syllogism m a i:.trikingly direct manner.
How far the analyi:.1s arrived at m logic was promoted by tlw
contemporary progress of grammar v,e cannot say, for we know
so little about the l.ttkr Grammar doec:; presuppose that the
.1b~tract10n of a general form of statement- the sentence-has
been arrived at, and that would facilitate and perhaps influence
such a treatment of the propoi;1t1on as \\ e find m the logic of
Aristotle On the other hand, the analysis which he reached 1s
not a grammatical one; 1t docs not coincide with the distinct10n
grammar proper makes w1thm the sentence, either of clauses
(for these arc altogether 1gnore<l) or of parts of speech. The
,tnalys1s however, "Im h the syllog1~m would n.tturally i-.uggei:.t,
hes to hand, the '-'Cry name mdcc<l under winch state-
ment 1,; invest1g,LI c<l 111 this loe;1c 1s not a gr.Lmmatiral one, but
derived from argument m debate For propo1>1t10n 2 me.1nt
1 § ;z

[a Propob1bon, 11p:Ta<m The or1gm of the word m An,totlc's log1<. lb not

certain TnndcJcnburg thought it was dcrivccl from gr,Lmmar and intended
to express 'the <.ond1t1onmg •. \\'1lhon follow'> Ammomuh, (Au-you,) .:,,
'lrpo'rfl11oplvou. {,,rJ .,;;,,, t111A.\o-yiua<18ai Tl /j01J.\opl110,11 .,.,,, 1tr11vOJVois .,.,,, .\u-yo,v ,rpoTa<111r
lrnJ .,;;,., m.\acoiv tvopa(oplro11, (in De Inl 2• , cf 4 11 (ilu&&c)]

G2
84 STATEMENT, THINKING, AND APPREHENSION
originally a premiss put forward in discussion, to be accepted
and argued from, or to be questioned and argued against.
§ 38. Logic, m some· rc<'ent books, has followed the ancient
trad1t1011 111 Legu,mng "1th "liat 1s m fact the !i>tatcmcnt and
its analysis Tl11c; 1c;; trul ,1l'>o of the more .idvanccd and ph1lo-
supl11c modern log1r, though 1t profec;c;cs to ch.,cuc;s a mental act
called JU<lguncnt an<l not the verbal cxprcsc;1on of that act a
Thus, m llm more advanrl·<l modern lo~1c, we find the term
JU<lgl·ment .,nb!'itltuted for the propoc:;1t1on an~l the enunciat10n
of the olcl logw Tim, no doubt ,trosc be.cause the older tcrmmo-
logy sterned to the more plulosoplll( pornt of view madequate
to log1t rcgar<lc<l as the !>Ludy of thmkmg 1 he propos1t10n
wac; -.o 11,1mul ongm,illv on ,1crou11t of 11!> rclatwn to debate,
and cnunn,tt1on too refn., to the verbal cxprc.s..,1oll A!> 1t, \\as
hd<l tli,tt I he mtentwn of logic\\ ,ts prmwnly to cxamme thmkrng,
and verbal t''("prc~c..wn only c..o f.1r a-. c..u!J<,1dury to th1::, ob3ect,
the word Judgement c;tcmul more arrurate ,rnd w,ic; ac.cordmgly
c;ubc;t 1tut ct! for the tl rm pir,poqt1011 Tl11s, ho,\ cvrr, \\ ac; done
without a full .ipprct 1,1!1011 of thl ronseqm llCL'> ,llld the result
h,1-. been a Ll'Ttam ,111101111! of c ontuc:;1011 'J he11· 1~ " furthc.r
eonfu,1011 111 the cli,-,t m, l 1011, uc..u.1! 111 tlrn, ach .imed logic, bet,, ccn
.l tl1eory of mkrcntl' .ind a thl·ory of J11dgcnwnt, and about the
true n,1tun· of tJu.., ~croncl rnqu1ry and its relation to mference.
I lo,, then doc., the 1h~t1nct 1011 bl t \\ cen the~ci\\ o mqumcs naturally
ansc and \\lut I'> the confusion that rec;ults fromfalc;eabstrac.t10n ~
§ 39 The f.tn11li.1r concc.pt1011 of a theory of JUdgemeut as
cl1stmgu1sl,ed from a theory of rnfrrencc sremc; lo ongmatc thus
The proc.css of rnfr1 encc 1s seen to prc-suppo..,e knowledge already
ga1tll'd or ornmon., ,1lrL,Hly formed The process itself 1s con-
Cl'l\l'd, m l'ilect, ac; the apprd1lntl111~ of \\hat thu,c prc\.lOUS
apprl'hcnswn., or op1111onc; rn·ress1tate m the \\ ay of other kn<)\\ -
ktlgl' or opm1on I knee 1l Leconws C\ Hknt that there must
be apprchens!\>nc; not got by mfcrcnrt or n·a,omng Tlus appcars
m the f,u111lt,1r statunent tl1,1t till re mmt be undemonr,trated
prcnu::,ses or there \\ ould l>c .t Ill'\ Lr•emlmg process I b
J § ,~<I
(• Ifrfrrrmg to Ho~anquct':, I ,,i:1c a11,I F H Ilra<llcy•~ The I'rrnc1ples of
I o.~u (181:q)
1• Inf,·rcnLc us{'(] to bC' ba1d to l><'gm ' rx f>rnecogmhs ct f'rarronce5s1s • The
\'IC\\ 1~ 111 An~totlc, .4.11 l'o 7~• 7 an<l b 18, Metaph 1006• 8]
 Apprehension in General 85
Certain apprehensions, then, are recognized as not bcmg
inferences, and also as being material of inference. These arc
called propositions or Judgements (not that these terms arc con•
fined to them alone), the name 'judgement' being preferred to
proposition rn modern logic, because the word ' propos1t1on ' 1s
associated with the verbal statement rather than with the mental
al.t1v1ty. The study of such apprehension<; would be necessary
to the 'itudy of mfcrcncc, as it 1s here conce1vcd, smce they arc
1tc; material, but d1stmct from the study of mfcrcnce smce they
arc not mfcrences.
Thus, 1f the namec, ' propos1t10n' and 'Judgement' were
confined to apprehcns10ns winch "ere not mfcrence<; and 1£ the
theory of the propos1hon or Judgement meant the c;tudy of them
a<; such, the d1v1s10n mto the theory of Judgement and the
theory of inference \\ ould be JUSt1fiable and the nomenclature
c,o far correct But the study 1s not so concc1vcd nor 1<; the
termmology thu<; restnrted , for the clcc;1gn.ttlon 'Judgement '
1c; not confmcd to \\ hat 1s not mfcrrcd, but mrludes what 1s
mkrrccl, whether knO\\ lcclge or op11110n
§ 40. Tim, secmc; to come <1bout ,1.., follo\\;; The knowledge
,, l' get l,y 1nfc1 encc Ill the sc1cnccs 1s stated m a verbal form
,,. l11ch sr~rnfic<; the ll,ttun of the tlung known and that only,
i•ot the nature of our apprehcns10n of 1t , and the i,tatcmeut
of the fart, onuttmg, as it uftt11 docs, the grounds of it discovered
Ill the mfrrencc, supprcs&c<; d!l trace& of the existence of the
pro<.css. For lll<;tance, ' the square on the hypotenuse of a right-
<tngled triangle 1s cqudl to the sum of the squ.ircs on the other
t\\o 1,1des' ii, a dcscnpt10n of the obJcl.t1ve f..ict, without any
rcfcrcn<.e to our suLJcct1ve state Such a statement preceded
by the \\ ord ' therefore ' H, the last statement lll the verbal
expression of an mfcrencc But without th1<, word, wluch con·
ncl.ls it \\ 1th the proce&&, 1t 1s given as the so-called concluci10n
of the mfcrence 1 he form of the conc.lu;,10n, not cont..immg
the grounds on \\ h1ch \\ c based it, promote~ the h,tl lit of rcprc•
c;ent111g the mental activrty v.lu<.h corresponds to rt (the appre-
hension of the fact) as a rei,ull d1c;t111ct £10111 the reasoning
protcs& by v. luch it 1s got, and the fall.icy of regarding it as
something in itself apart from the process and po&sessed some·
hov. alone v.1thout the process; whereas the possession of 1t ii:.
36 STATEMENT, THINKING, AND APPREHENSION
JUSt the inferring procesb itself The mental activity corre-
sponding to 1t, the having or posses!>mg it, wluch 'having' can
only be the apprehl'ns1on of the fact expressed by the statement,
or the belief m 1t, tl1Uc; , irtually becomes a result of the infernng
process and 110 somethmg chffcrent from 111fcrnng or reasoning.
Like the apprehrm,10ns \\ h1ch arc not mfcrcnccs, 1t 1s called
Judgement. ltc, wrhnl fmm 1s tal,cn to be the statement of
a Judgement as d1'itmgu1shcd from the statement of an inference
that 1s to c;ay, the statement of sometlung Judged as opposed
to the c,tatemcnt of c;onu thing mfcrred. Tlus carries with it
111ev1tably the <l1stmet1on of Judging from mfrrrmg and thus the
fiction nf a kmd of mental act1v1ty called 3uclgcment or Judgmg,
as somethmg d1stmct from mference, comes mto existence. The
truth however 1s that 1£ \\ e allow ' Judgement ' the meaning
necessary for the view under cons1derat1011, viz the havmg any
knowledge, behcf, or op1mon, the Judgement called the con-
clus1on 1s not anytlung apart from the process of inference by
which 1t 1s att:.11nc-cl, and the 1IH ntal ,u ti\ 1ty of J1tdg111g tlus
judgement-the ' havmg' of 1t spoken of aboYe- 10:, prcc1.:;ely
the process calkcl 'att.11mng' 1t, y1z the mkrrmg of 1t, and
mfcrnng 1s merely a part 1cular wny of 1uclgrn~.
Dut more than tlus, 1f "e take 3udgmg m its moo;t correct
and natural i;ensc, that ,., as dec1s1011 on evidence after dehbera-
twn, then mfcrnng 1s Just one of those forms of apprehending
lo "lurh the "orcls 3udpng and Judgement most properly
apply. 1
§ 41. The fam1h.1r d1strnct10n then of Judgement from inference
m\.olvcs a confus10n of thou({ht. Kno,, ledge, op1111on and behef
arc regarde<l as fe>rms of the s,imc sort of mental activity, termed
;udgcmeut, and tins actIY1ty, called Judgement, 1s made the
subJcrt of ,i Sl'pnratc 111q111ry Just bccam,e 1t 1s regarded as
• Thl' acti, c "urkmg of tlw fal,t- cli-t1m hon between a JnrlgC'mC'nt m gcnl'ral
and the procc-',< b} "ha h 1t 1, ,lttatnC'd is su•n rn the trad1t1011al doctrme uf
rrduct1on (of tlw l mfcrmr IH{llrl'' to thL Fir&t) Tlu, rC'ductton 1s crroneou~.
though appan•ntl} qmh. um.h,1llt•ngul a, and it depends on the fallacy c-xposcd
above Cf §; 11

l" The &t.ilL111c11t u, too gC'neral, d (l' g ) 'The first three figure, are the
l'lt)pcs of threl' real and c-sst-11t1.1lh d1tfcrl'nt operations of thought•. Schopen•
hancr, I he ll'orld as H 1/l a11d ldca, Bk I, t-h ,.. • Tu bomc people this
,1 e n•duct1011 to the 1st Figure) has seemed &upcdluous,' Lotze, Logic, § 91 J
 Apprehension in General
somethmg different from inference. This fictitious severance of
Judgement from mfcrence results from ,L false abstraction. The
rn11clus10n of an mfcrence 1s really the verbal st.ttcmcnt of
a fact, the eAi,,tence of wluch is inferred. Tins statement,
,lbstracted from the mfcrnng process, 1s treated as 1f it could
be conceived ,, 1thout the mfcrrmg activity. The supposed
activity of 'havmg' or apprehcndmg 1s called JUclgmg or Judge•
ment and the statement itself 1s called a Ju<lgcment, or more
accur..1.tcly the express10n of a judgement.
Tlus act1v1ty, then, called Judgement, as ch:,tmgu1shed from
mfcrcncc, 1s a fiction, there ts no such tlung Judgement m
tins fictitious sense, besides mcluc1mg non-mfcrrcd knowledge
and knowlcdgr said to be the result of mfcrencr, mcludes also
op1111on and behd, for the s1111ple rcac;on that what 1s known,
pcrcr1vccl, suppoc,rd, or believed may Le expressed 111 an identical
verbal form. The man who knows that A 1s n, whether as
perce1v111g tlus or not, whether as mfcrnng 1t or not, and the
man \\ ho holds the opm1011 or be!tef that A 1s n, may equally
use the form of statement that A 1s B, and under ordmary
un-umstances do &o use iL
Now, 1f this common form ,,ere an cxprcss10n of the mental
,1ll1tude of the pc1s011 us111g 1t, 1t woul<l be reasonable to expect
to fm<l a common and Ct>'>l,lltial c..lcmcnt 111 the mental attitude
corresponding to the verb.ti form But the form merely slates
thP uaturc of \\hat \\C know to be, or tlunk to be, existent,
\\1th complete ab<;tract1011 of the fact that 1t 1s for m, matter
of knowledge, conJccturc, or behcf. So far from bcmg an expres•
!:>Ion of our mental attitude, 1t says notlung about 1t whatever.
A is B means th,tt a cerlam ov3ect has a c<-rtam nature or
quality, 1t Jocsn't matter \\hethcr the statement 1s true or
not, that 1s what 1t me.am,
§ 42. The trad1t1onal d1v1sion of logic, therefore, rnlo the
theory of Judgement and the theory of mkrcncc rcc,t<; upon an
erroneous pn11c1ple Strictly !:>peakmg, the llung- called Juclgc•
mcnt, wluch shoulcl be the sub1ect of wh,tt 1!:> CJlh J the theory
of Judgement, 1s fict1llous but the confus1011 1s concealed by
the fact that the verbal cxpress10n common to matter of know•
ledge, both 111fcrrcd and not mferrecl, ancl maller of opmion or
belief, and mistakenly supposed to be the expression of a mental
88 STATEXJ'tNT, THINKING, AND APP RB BENSIOlf
activity called judging, does duty for this activity in this part
of logic and forms the real object of study. Indeed, the logic
which in modern phrase is to be a logic of judgement is, quite
unconsciously, a.logic of statement.
If this is so, what should we expect? We should expect the
inquiry to be directed sometimes to what the verbal form
signifies and sometimes to the verbal form itself. If the inquiry
is into what the given verbal form signifies, since that is the
nature of the object only, with no reference to our thought
about it, we should expect the result to be abstractions which
belong to the objective reality and not to the apprehension of
it, nor to our thought about it m general ; objective forms,
that is, not forms of the subjective.
And this is what has actually happened. We do find in this
part of logic abstractions which are of what belongs to the
nature of the object (objective forms of the kind called meta•
physical, not true logical forms at all) and, further, metaphysical
forms may be confused with logical, as we shall see m the case
of the familiar d1stmction of subject and predicate, where
a logical and a metaphysical d1stmction are unconsciously
combined m the same designation. No wonder that in some
modern ph1losoph1cs logic is md1stinguishable from meta-
physic.
On the other hand, if the inquiry is really directed to the
verbal form, we should expect to find abstractions which belong
to grammar and to bngu1stic form in general, associated with
the logical and metaphysical abstractions. And this, again, has
actually happened. Many of the mqumes in medieval logic are
of this kmd. In modern logic an mstance of it 1s the theory
of the connotation and denotation of terms-" hich, indeed, has
a medieval source 1 This instance 1s mterestmg because the
subject has proved so confused and puzzling. One must venture
to thmk the secret of the confusion to be that the d1stmcbons
attempted concern the grammatical functions of certam word
forms, a fact which has not been realizcd. 2 Another instance
is the theory that all universal propositions arc hypothetical,
a fallacy which has arisen because the logicians who hold the
view do not realize that they have before them the question of
• Part II, ch. 18.
 8g
the meaning of certain forms of speech, a purely linguistic
question. 1
In general, when the logical, grammatical and metaphysical
notions are not confused with one another, there is a tendency
to pass from one to the other without a clear consciousness of
the transition and to associate them as if they were of the same
kind.
But though the general character of this part of logic may
have been misconceived, that need not prevent the presence of
true logical mquiries in the traditional theory of judgement and,
indeed, of grammatical and metaphysical mqu1ries which belong
to the subJect because they subserve the logical. If there are
such logical mquiries and such justifiable metaphysical inquiries
in this traditional part of logic (and there seem to be), we have
to ask what the rationale of their grouping as a special part of
logic is. If they differ, and they do seem to differ, from what
belongs to the theory of inference, how should the part of logic
to which they belong be characterized ? From the point of
view of the wrong distinction of inference from something called
judgement, the difference between the two parts of logic would
have to be characterized thus :-the theory of inference studies
the subjective side of thought in the reasoning process by which
we attain a thought consisting m knowledge or opinion ; the
theory of Judgement studies the subjective side of thought in
the form of knowledge or opinion, considered in itself and apart
from any process by which it may be attained.
After what has been said, it will be evident that the formula
betrays its own mcoherence.• The subjective side in question,
when the thought attamed is knowledge, is the apprehension of
the fact and, m the case of what would be called the Judgement
attained by inference, the apprehension is the inferential process
itself. Thus, according to the above formula, we should, in the
theory of Judgement, be studying {m the case of a judgement
which was the conclusion of an inference) the apprehension of
something, considered entirely apart from its apprehension.
1
H 98, 103, and 312 [Ueberweg, Logia•, § 94]
[• Tlus, as so often in Wilson, 1s self-polemic. He states his own old attempt
to 110lve the problem and condemns it, as he would have said, unc;ompro-
m1smgly]
go STATEMENT, THINKING, AND APPREHENSION
And, in general, we have found the whole basis of the view
which would result in such a formula untenable.
§ 43. What then have we to substitute? For we shall not
necessarily reject the whole of an inquiry because it has been
conducted under a m1sconccpt1on. We have seen that the idea
of logic as a study of thinking in its various kinds led to the
consideration of apprehension m general as the primary subject
of investigation in logic Starting now from the form of appre-
hension \\h1ch is reasoning, that is from our interest in the
subJective side of thought as 1t appears in reasoning, we observe
that inference or reasoning depends upon apprehensions which
arc not inferring. We arc then naturally led to the idea of
some study of apprehension m general as apprehension, whether
inferential or not. This would be a preliminary to the study
of inference and so far accord with a feature of a traditional
part of general logic, namely that part wluch, though sometimes
entitled the theory of conception, 1s nowadays usually embraced
under the title, the theory of judgement. Apprehension being
properly restricted to knowledge and opinion bemg formed in
the effort to get knowledge, we might further mquire into what
is common to the attamment of knowledge and the formation
or opinion, more especially as what would be called the state-
ment of an opinion and the statement of knowledge are so often
{indeed commonly) the same in form.
Such an investigation, agam, we should expect to lead naturally
to an examination of the verbal form of statement, not merely
because of this formal sameness in the verbal expression of
opinion and knowledge, but m order to see what light the form
of expression might throw upon problems about the mental
state. But then we should clearly recognize that it is the verbal
form which we are exammmg and how such examination 1s
relevant to our purpose. This, again, would correspond to
a feature of the theory of judgement, for example the classifica-
tion of the various forms of propos1t1ons, though the inquiry
would be conducted in a different way and with a better chance
of avoiding confusions.
Again, metaphysical conceptions such as substance and attri•
bute might have to be recognized and considered in so far as
they, m turn, may assist m the understanding of subJects con•
 A pp,ehension in Gene,al 9:r
nected with apprehension or thinking in general. But here,
again, we should avoid confusion by recognizing that they are
of what belongs to the object and not to the apprehension of 1t
and so should not confuse them with logical forms. In this way
we might expect to mclude those parts of the traditional theory
of Judgement which can be v111dicated as having a place in logic
and to understand their relation to one another and to the
general scope of the mquiry. This part of logic then we may
describe as concerned with statement and its relation to thmking
and apprehension.
 II
THE USE OF THE TERM 'JUDGEMENT' IN
MODERN LOGIC

§ 44. WE have next to consider certain difficulties which result

from assuming with the advanced modern logic that there corre-
sponds to the simple statements, whose common characteristic
is that they claim to be true, a common subJective activity
called 'Judgement•,• on the ground of which the statements
themselves may be called judgements. Difficulties arise at once
from the natural meaning of the words ' judge I and ' judge-
ment', and from Lhe ordinary v1ev. of thinking Judgement is
a word taken from ordinary usage :ind ought to re tam what
is essential m its meaning there. Certainly it is adopted as if
this were so, for no proviso is made. A Judgement is a dcc1S1on.b
_____
To Judge is to decide. It implies previous indecision;
,. ____ -- ------ - a previous -
[• Cf 'every act of consciousness 1s a Judgement and therefore a belief 111
the presence of its obJect' Mansel, Prolegomena Logica•, p 320. Smee
Mansel, the notion has been very common in England.
b Cf ' The last opinion in search of the truth of past and future 1s called
the " Judgement ", or " resolute " and " fin.ii sentence " of him that " dlS-
courseth" ', Hobbes, Ltviathan, ch 7 Among the Schoolmen, the usage of
Ockham 1s very close to that of Wilson, ' Inter actus intellectus sunt duo
actus, quorum unus e~t appreben~1vm, . , abus actui. potest dic1 1ud1cativus,
quo mtelle<.tus non tautum apprehend1t ob1e1,tum, 1,ed et1am !lb assentit vel
dissentit nulh ai.,,eutm1us per mlelleLtum, n1si quod verum eX1Stimamus,
nee abcw d1ssent1mus, ms1 falsum reputamui.,' Ockham, Sent Prolog. qu 1. O
(quoted m Prantl, Geschicllle, &.c , m, p 333, note 753) ' Sed quando 1ud1cat
rem ita se habere, s1cut est forma quam de re apprehend1t, tune pnmo
cognosc1t et d1c1t verum,' St Thomas Aq Summa Th I qu 16, a 2. Judge-
ment 1s usually held to involve also compan~on or' d1scnmmat1on ·, e g "they
that observe their difference and diss1m1htudcs , which 1s called ' d1&tm-
gu1shmg • and ' d1scemmg ' and ' Judging ' between thmg and tlung . are
said to have a' good Judgf'ment' ", Hobbes. Ler•ialhan, ch 8 Cf infra f 108
Certainly m ordinary usage the word mcludes opm1on, e g ' not conceal their
alteration of Judgement ', 1 e. as the context shows ' an op1t11ot1 formed of the
truth or falsehood of a doctnne ', 'Walton, Lafe of R Santlwscm, pp. h 3 and
h z , and ' The last opuuon m search of the truth of past and future IS
called the " Judgement " ' (Hobbes, quoted above) )
 'Jflllgement' in Modem Logic 93
thinking process, in which we are doubting. Those verbal state•
ments, therefore, which result from a state of mind not preceded
by such doubt, statements which are not decisions, are not
judgements, though they may have the same verbal form as
Judgements.
But now many of our so-called judgements in perception are
of this kind. If I see black letters on white, I can say that
the letters are black and the paper white, but these statements
do not, in their ordinary use, represent results consciously
arrived at after a period of indecision. Ordmartly I apprehend,
it would be said, the white and black directly, without any
previous questioning or doubtmg. These are, in the true sense,
apprehensions, although not the kind of apprehension which is
Judgement, m the strict use of that word. Further, there are
in logic and metaphysics certain artificial statements, called
judgements, which yet not only seem without title to the name,
but do not even in ordinary life get expression in the verbal
form of statement supposed to be Judgement. According to the
doctrine omnis determinatio est negatio, the perception of white,
for example, would be said to mvolvc necessarily 1ts d1stmction
from other colours, the apprehension of this distinction, and
therefore negative Judgements expressing the distinction. Now
these would have to be of the form, ' white is not black ',
' white 1s not red ', &c. Yet no one ever, in natural thinking
or speaking, either makes or expresses these so-called judge•
ments. I will return to this later in the discussion of ne~ation. 1
Again, there are statements of opinions, and these in this logic
rank equally as Judgements. But now, if I consciously form
the opm1on merely that all A is B, I am not sure that all A 1s
B ; nor do I decide that A is B. The verbal formula then, all
A is B, if regarded as the expression of my op1mon, is not the
expression of what I decide. That statement therefore is not
the expression of a decision and therefore not the expression of
a judgement. There is here a thmkmg process which contains
decisions, the decision bemg not that A 1s B, but that there is
such and such evidence for it ; and m opinion, the dec1s1on
about the evidence is what alone is entitled to be called a judge-
ment.
1 Part II, ch. 12.
94 STATEMENT, THINKING, AND APPREHENSION
So far then we find,• putting aside that fictitious notion of
judgement of which we have spoken, that in its normal use
judgement does not include every mental attitude which corre-
sponds to the statement A is B ; that perceptions in general
are distinguished from Judgements ; and that (excluding certain
apprehensions and perceptions which do not seem, in any natural
expression, to get the form of statement) perceptions which do
find expression in verbal statement, and opimons, as well as true
judgements, whether they have any common element in their
nature or not, may still have corresponding to them the same
form of verbal expression, i.e the statement.
§ 45. Now logic, when 1t professed to deal with propositions
in general, had no difficulty in including these varieties of state-
ment No distinction was made and log1c1ans did not even
think of making one. Modern logic 1s in a different position; for,
if it sub~titutes Judgement as an improvement upon proposition,
it should realize the consequences c,f choosing such a sigmficant
term and should exclude such statements as these, or at least
explain how they can possibly be regarded as judgements. But
in general, owmg to the misleading identity of the verbal form,
these statements are not excluded and no reason is given for
retaining them.
Consider next, from the standpomt of the ordinary view of
thinking, the mental attitude which corresponds to these various
kinds of statement, which are thus identical in form. If logic
starts, as it usually or often does, from a definition which
commits it to the view that it is a study of thinking and if it
really confines itself to thinking, then it might be said that
there \\ ould be no difficulty about statements representing per-
ceptions, for these arc not the results of thmkmg proper and so
would not enter into the field.
If, again, logic starts with Judgement proper, to be consistent
it would have to exclude perception and statements of percep-
tion ; that 1s experience and the statement of experience ;
simply because perceivmg is not, m the ordmary view, either
thmkmg or judging. But the logic m question does not exclude
[• The ongmal MS has a margmal nott.> · ' add to this from preV1ous
version '. I could not trace this, but my redraft expresses WllsoD's mature
views, I believe ]
 'Judgement ' in M oele,n Logic 95
them and, if we are not to exclude them, what then is the
justification for retaining a discussion of them in this part of
logic ? The true explanation would appear to be that logic is
obliged to study them if for no other reason than because they
have the same verbal form as true Judgements. For certain
decisions again, after the process of thinking, perception 1s
required ; for instance, I may see that the question I am
thmking about partly depends upon an experience; if I am,
say, comparing the length of two obJccts, I may have recourse
to putting the two tlungs side by side. Now the apprehension
of one extended beyond the other would be a perception, never
a thinking process. Fmally, even if we are clear that many
perceptions are not, as such, decisions after a state of doubt,
the question will arise whether thinking is to be excluded after
all from perception, and mdeed some philosophers have made
a great point of asserting that 1t docs enter into perception.
Clearly such questions can only be decided by getting some
definite agreement as to what we mean by thinking ; in fact
Ly some such d1scuss1on as that conducted in an earlier chapter. 1
§ 46. On the other hand, while statements of experience would
thus be excluded on the ordinary view of tlunking, opinion
would certainly be held to be the result of a thinking process.
Thus, if logic 1s confined to thinking, 1t might be asked whether
opin10n would be excluded by making Judgement our starting•
pomt.
If it be said that • Judgement is not merely a starting-point,
but covers everything examined m this part of logic, 1£ Judge•
ments moreover arc to be decisions, opinion would be excluded
for the reasons already given.
But we may modify this account of the starting-pomt in
judgement and understand 1t to mean that we start from judge-
ment m its true and proper sense, making 1t not the general
form of the mental activities called thinking but the key to the
understanding of them. Then opinion would be naturally
included m this part of Jog1c, smcc 1t is a process which we
understand through Judgement; both because 1t 1s in the effort
• See Part I, ch 2

[• Cf, § 42, Self-polemic against a d1Scarded solution in earller drafts.]

g6 STATEMENT, THINKING, AND APPREHENSION
to form a judgement proper, a decision on evidence, that we
form an opinion ; and because the estimate of evidence which
is necessary in the formation of opinion involves judgement
proper.
So far, then, we should be Justified in including opinion under
the theory of Judgement i but still there will be something else
besides judgement to be recognized m the formation of opinion,
that 1s to say knowledge, as manifested in such activities as
occur in ordmary perception ; activities, in other words, which
arc not properly speakmg decisions. We may suppose that some
difficulties may be got over by a better formulation in language.
The man who forms an opimon is not entitled to say that all
A is B, much less 1s that an expression of his attitude of mind.
It 1s not adequate to the thought it may be supposed to repre•
sent. A more adequate expression would be 'A is perhaps B ',
or 'A is probably B ', ass1gmng at the same time the reason;
the reason bemg a statement of evidence Now, if we assume
that the word' probably' simply represents an obJecttve charac•
teristic of the evidence, namely, that there is more evidence of
a certain sort m favour of A's being B than against it, then
' A i!I probably B ' Vl,Ould be the statement of a judgement in
the true sense of the word But m this way we should have
gone too far, for we should have failed to provide in our state•
ment for opinion as opinion at all, because we should be merely
expressmg that dec1s1on which we have shown cannot be opimon.
We must therefore either say that the expression' A is probably
B ' docs not represent the opmion as opinion, or else we must
interpret the word ' probably ' in a different manner. The latter
seems to be the most reasonable course The word ' probably '
docs imply the attitude of opm10n • In fact it does add some-
thing to that estimate of the evidence which in the process of
forming an opuuon can be called Judgement. It refers to
a unique mental phenomenon, a certain effect of the evidence
on our consciousness "'hich has no parallel m the knowmg
activity. The expression ' A 1s probably B ' then remains the
[• Cf 'The entertainment the mmd g1\•es this sort of propos1tlons 1s called
belief, assent, or opinion' Locke, Essay 1v 15, I J Wilson was readmg
Locke agam carefully when he wrote all these sections about apprehellSlons
which arc not true Judgements J
 ' Judgement ' in Mourn Logic 91
expression of an opinion and not of a judgement, is indeed in
the strict sense not a judgement. We cannot therefore, by
a mere change of phrase, eliminate the difficulty with which we
started, so as to make the only subject of logic, in this part
of it, judgement in a true and accurate sense.
§ 47. Judgement then being different both from perception
and opinion, the question now arises whether there is any general
form of which these are species. Could we, for example, say
that this form is apprehension in general ? Apprehension seems
to be a term proper only to those Judgements and experiences
which are knowledge , for apprehension must be true because
it is apprehension of the nature of the obJect and this is just
what truth is. Opinion, accordingly, cannot be apprehension,
for even if the opinion that A 1s B should be true, that would
not constitute 1t apprehension for the person forming the
opinion.
Inasmuch then as some judgements at least are knowledge,
the answer to our present question must follow from the answer
to be given to the question whether knowledge and opmion are
species of one common form, and to the further question whether
perception is, as some suppose, erroneously distinguished from
thmkmg. ~ thmgs at least seem clear from the foregoing
discussion one, that 3udgement is not the name of a common
element m the various mental attitudes wluch find their verbal
correspondent m the form of statement, A 1s B ; the other,
that, in accordance with our previous discussion of the supposed
common element in thmkmg, apprehension, a term wider than
true Judgement, is that through which we have to understand
the activities of thinking and 1s also that which gives them their
unity Nevertheless their umty (even 1f we exclude from con•
s1deratlon the mterrogative or wondering attitude) 1s not to be
found m a common umversc:tl of which they are the different
species.

H
 Ill
OPINION, CONVICTION, BELIEF AND
COGNATE STATES

§ 48. WE pass now to some important considerations about

actlvit1es which are naturally described as thinking, at first
sight, and are d1stingu1shed from knowledge, whether perceptual
or not. These activities arc such as opm1on (which has already
been compared with Judgement), convict10n, belief and their
congeners. We shall be led also lo mvesttgate the poss1btl1ty
of false Judgement, to con51der certain erroneous states of mind
and even a certam form of consc10usness which actually simulates
opinion and behef, although 1t 1s, properly speakmg, neither.
This investigation illustrates the necessity which logic is under
of exammmg, m its own mtcrests, act1v1t1es and ideas whirh
may turn out to belong themselves properly to another study
or science. We may say m advance that we shall be obliged
to recognize m consciousness somethmg which 1s not clear
thinkmg and which eludes us Just because we tend to try to
express its character m terms which belong to that thmkmg.
This caution must be kept in mmd in all our succeeding
investigations.
In what has gone before, \\e have cr1t1c1zed a view which is
obscurely bound up with the fam1har d1stlnct1on of Judgement
from inference, the view, namely, that knowledge, opinion,
behef, as well as perception and experience, are forms of one
and the same sort of mental c!.Ctlvity, called judgement. Thts
supposed common activity has been mvestigated with special
reference to its name ' Judgement ' , but the difficulty does not
be merely in this designation, 1t hes m the assumption that there
is one and the same kind of act1v1ty at all, however named,
manifested m each of these mental attitudes ; something more
specific than, say, mere activity of consciousness m some relatlon
 Of,iKion, Conviction, Belief 99
to the object,1-something so definite that a special verbal form,
the proposition, corresponds to it.
Now we have suggested that this view has not arisen in the
only way in which it should properly arise, from an analysis
of these mental states or activities themselves ;-mdeed no
analysis could discover it,-but from the existence of the
common verbal form, the statement, which is indifferently
applicable to all of them.
Yet the form of statement in question is so far from being
an expression of the mental attitude of the person usmg it that
it says absolutely nothing about our condition 10 uttering it ;
whether that condition be one of knowled~fstrong belief, or
of weak opmion. It is therefore a fallacy t assume from this
identity in the form of statement about the JCt't an identity
m our mental attitude towards the object. We must examine
these attitudes or activ1t1cs themselves, if we are to find a
common element m them
§ 49 Consider knowledge and opm1on. We easily see that
opm1on mvolves knowledge ; but we also see that the opinion
itself must not be confounded with that knowledge. It ts
characteristic of the ca.ecs where we form an opimon that we
notice a certam quality m the evidence, 10 virtue of which we
say the evidence known to us rs stronger for one alternative
than for the other. W,: know, that rs, that certam facts are
m favour of A's bemg B, but either that they do not prove 1t
or that there are facts against, though not decisively against,
A's being B But this estimate is not the opinion We are
affected by it so as to form the opm1on, yet the opinion is neither
the knowing which constitutes the estimate nor any kind of
knowledge. It is a peculiar thmg-the result of the estimate-
and we caU it by a peculiar name, opinion. For it, taken in
its strict and proper sense, we can use no term that belongs to
knowing. For the opimon that A is B 1s founded on evidence
we know to be insufficient, whereas 1t 1s of the very nature of
knowledge not to make its statements at all on grounds .recog·
1 We cannot make th.ls vague idea more specific by definmg the relation

to the obJect as, for example, • being concerned with the apprehension of the
obJect' For the apprehemaon of the ob1ect 15 knowledge , thus we should
be merely definmg knowledge as an activity concerned with knowledge.
H2
100 STATEMENT, THINKING, AND APPREHENSION
nized to be insufficient, nor to come to any decision except that
the grounds are insufficient ; for it is here that in the"'tnowing
activity we stop. In knowing, we can have nothing to do with
the so-called ' greater strength ' of the evidence on which the
opinion is grounded ; simply because we know that this ' greater
strength ' of evidence of A's being B is compatible with A's not
being B after all. Beyond then the bare abstraction of conscious
activity, there 1s no general -fharacter or quality-of which the
essential natures of both knowledge and opinion are d1fferentia-
tions, or of which we could say in ordinary language that each
was a kind. One need hardly add that there 1s no verbal form
corresponding to any such fiction as a mental activity mani•
festeq in a common mental attitude to the obJect about which
we l&ow or about which we have an opimon. Moreover it is
vam to seek such a common quabty in belief, on the ground
that the man who knows that A 1s B and the man who has
that opinion both believe that A 1s B. Bebe£ is not knowledge
and the man who knows does not bf'lieve at all what he knows i
he knows 1t. We might as well say at once that knowledge is
a kind of op1mon as that it 1s a kind of belief.
§ 50. We have spoken of op1mon such as is consciously formed
and recogmzed as an opm1on and not .as knowledge by the
person who forms 1t; and, however determined we may be in
such a case to act as if the op1mon were knowledge, our expres-
sion of it shows that we do not confound it with knowledge,
and hence follows what has been said about the distinction of
judgement which is decision from opinion. There is here a cer•
tam simulation of an act of Judgmg which, however, ought not
to mislead us For, though I am not sure that A 1s B {and
therefore, though inclined to 1t as probable, I have not decided),
I may decide to act as if A were B I may have to make up
my mind between two alternative courses of action and, knowing
neither, I may choose all A is B as the more probable and the
one therefore that I shall act upon (although probability is not
the sole ground of such decision). There is then a mental
decision, which may be said to be in favour of all A is B,
a practical decision, getting it is true greater definiteness by the
fact that we act upon 1t, but not the Judgement (or decision)
1/wA is B.
 Of,imon, Conviction, Belief , 101

However, we find in ordinary language another term, belief,

akin to opinion, yet distinguished from it, so that sometimes
a man would actually prefer the term belief to opinion. When
this is done it would be felt that the uncertainty which seems
to be associated with opinion has caused us·to avoid-tiha• ward.
This again implies some (say) 'superior' certainty about belief.
This fact we cannot afford to neglect. Such distinctions in
language are never unimportant In the first place we observe
that the tendency wouJd be to use the term ' superior ' certainty
and to avoid ' absolute ' certainty ; and then we at once reflect
that in certainty there are no degrees and that certainty there-
fore is not the right word. Yet we feel that there is something,
when we compare belief and opm1on, which does somehow vary
m degree and we naturally ask what this is. In the case of
a given belief, for example, that A 1s B, as long as we hesitate
to call 1t knowledge, or at least betray ourselves by usmg the
word belief and avo1dmg the word knowledge, we cannot really
have decided that A is B unless, on reflection, we can say we
know A 1s B. That is why judgement does not seem a proper
designation for this attitude of ours. There is a clear decision
m our resolution to " act on our belief ", i c. to behave as 1£
A were B, takmg some practical resolut10n m consequence. But
this dec1s1on m the case of belief 1s not the whole matter, for
such a decision may also be made m the case of an opinion.
Again, m a practical decision there are no degrees. We decide
to act on a belief or we do not. Nevertheless, comparmg the
practical decisions with one another, we observe a difference.
In the case of a given opinion, we should risk treating 1t as 1£
true for certain purposes ; for others we should not, because the
consequences would be too serious for us if we were wrong. In
general, on a belief we risk more than on an opinion. Thus,
though there are no degrees in decision, we observe a difference
of degree in the importance of the decisions ; but what is behmd
that ? Now it may truly be said that the more evidence there
is in favour of A's being B, the more we incline to riiik on 1t.
But this is not yet the complete account. In a given case (not
in all cases) of belief, we should have a real judgement, a true
intellectual decision, that there was so much evidence for A's
being B; also, a real decision, but a practical one, to behave
 .
IOZ STATEMENT, THINKING, AND APPREHENSION
· as if it were true in a certain practical issue ; yet we refrain
from certain other practical decisions which we should certainly
make 1f we knew that A was B.
Now, this obviously cannot be explained by saying that the
evidence is enough to justify us in the one case in taking A is
B as true, but not in another case and in other circumstances.
The evidence is the same both times and cannot change its
strength m consequence of the difference between the practical
issues. As t>v1dence, if it justifies taking the thing to be true
once, 1t JUSt1fies 1t always. We cannot really account for the
facts unless we take note of that subjective feeling, already
indicated by the word mchnation, m what has been said of
opinion. Corresponding to these different degrees of practical
importance m our decisions in the case of different opinions
and beliefs, there is a varymg degree of feeling of confi-
dencE'. This is sui generis; and we have really been getting
at the recognition of its true positi,·e nature by d1stinguishmg
it from that with which it might be confounded. Such con-
fidence is not an attitude that we take towards knowledge. It
1s a matter of knowledge that the angles at the base of an
isosceles triangle are equal. We constantly apply it m practice,
yet we should never say that one who did so acted with ' great
confidence m the truth ' of this propos1t1on To a high degree
of such confidence, where 1t naturally exists, is attached the
word belief, and language here, as not mfrequently, 1s true to
distinctions which have value m our consc10usness. It 1s not
opm1on, 1t 1s not knowledge, 1t 1s not properly even judge-
ment.
To sum up then, we have a true Judgement or intellectual
dec1S1on that there 1s evidence m favour of A being B. We
have further a certain degree of the feeling of confidence (an
ultimate and 1rreduc1ble feeling) about A's being B, depending
upon our estimate of the evidence and frequently influenced by
ou, wishes or fears. In consequence of this, we risk a decision,
not intellectual, but practical, by resolving to act in a certain
case as if A were B
§ 51. Two remarks may here be made, one on the nature of
this practical decision, the other on the mcbnation, or feeling
of confidence, which accompanies opimon and belief
 Opinion, Conviction, Belief
The decision in question is always practical,• a decision of ·
the will i but it by no means follows that it has no theoretical
application. In the case of a scientific investigation it frequently
happens that the investigator has to choose between two opinions,
to choose which of them he shall treat as if 1t were true, so as
to embody 1t in his theory and make deductions from it, there
being sufficient evidence for neither. The decision then, though
practical, is made in the interests of a theory and the con-
sequences of 1t may be purely theoretical.
The feeling of confidence which accompanies opinion and
belief depends partly, at least, on what we call ' the strength
of the evidence ', and Js stronger Jf the evidence seems stronger.
This idea of strength involves an illusion. It 1s only of evidence
which 1s not sufficient that we use the word ' strong ' at all.
In knowing, we can have nothing to do with the so-called greaur
strength of the evidence m question. The reason 1s we know
that this strongest evidence m favour of an hypothesis, provided
we can call 1t only strongest, 1s compatible with the falsity of
the hypothesis, and so our confidence may be futile However
strong evidence may be, it is not anythmg which can mfluence
reahty ; yet, m this feeling of increased confidence with increased
strength of evidence, we are unconsciously treating it as if it
could. The strength of evidence 1s merely somethmg for us ;
mdeed we never speak of the strength of evidence except where
we suppose that Jt doesn't prove what 1~ stated, that JS when
the evidence 1s not sufficient. The increase m 1t 1s only an
increase in our knowledge, yet we tend to contuse this advance
of ours, our greater hold on the facts as we may call 1t, with
some objective force gainmg a greater hold on the objective
facts. We know, that 1s, that the existence of the facts which
constitute the evidence is not something physically stronger
which overpowers the set of facts const1tutmg the weaker
evidence on the other side, and so necess1tatmg A's being B ;
yet, in opmJon and belief, we at least behave as if this were so
and that, although the strongest cases of circumstantial evidence

[• The meaning 1s that a scientist may be led to a particular senes of experi-

ments (or the geometer to a lme of inquiry) which he would not otherwise
have attempted Cf • practical needs may force us to make choice of one
(alternative) • Lotze, Logic, t 281 , and I 28o end]
m.t, STATEMENT, THINKING, AND APPREHENSION

get refuted by the facts. The illusion is almost irresistible and

is the rule, not the exception, with the student of physical
science and m any department where probable reasoning is
found. This fallacy is often illustrated in the treatment of
probability by its mathematical measure, and in argument from
statistics.
The illusion is reflected m language and subserved thereby.
Thus we say A 1s probably B, where the adverb which refers
solely to our subjective mclinatJon is made to qualify gram•
matically the verb of obJective existence. In any case language
does not clearly and transparently show the real truth of the
matter or the fallacy would not be so umversal 1
§ 52. By contrast with opm1on and belief, both of which
contam an clement which 1s not clear thinking, we have hitherto
confined the word Judgement to true Judgements
This however is not usual, and the fact that false judgements
are supposed to be possible leadci to a further consideration
about erroneou& states of consciousness, the ex1&tence of which
has to some extent been 1mphed already.
The judgement that all A 1s B 1s a dec1s1on on evidence that
A is B, and we must return to the nature of tlus decision If
we know that the evidence 1s insufficient, we cannot possibly
decide or Judge that A 1s B. What we really decide and Judge
about is the character of the evidence Though we cannot
decide on insufficient evidence, we may form an opinion, and
this remams true so long as we have any doubt whatever about
the sufficiency of the ev1dc:mce. Whether, therefore, or not every
judgement 1s a decismn on sufficient evidence- and no evidence
1s sufficient which dot's not abc;olutely prove, whether 10 other
words every Judgement 1s neces<i::mly true or not , this, at least,
is clear that, m judgement proper, 1f a man 'Judges' A is B,
he 1s himself sure that the evidence 1s sufficient
But now, this being so, is 1t possible that 10 Judgement proper
the man who is sure that the evidence 1s conclusive should be
mistaken ? If so, we should have two k10ds of Judgement, the
common element being that, in both, the person who Judges
A is B is sure that the evidence proves that A is B. In the
case, then, of one sort of judgement, the man 1s nght and his
• Cf II go-2 and Part III, ch 8.
 Ojnnion, Con1Jictio11, Belief I05
Judgement would be an apprehension ; in the other case he
would be wrong and his judgement would not be an apprehen-
sion. This distinction is natural and accords with ordinary
usage, and what we have said of opin10n goes so far to confirm
it. For, though we may say the opinion that A is B is an untrue
opinion when A is not B, it is not accurate to say that the man
who forms this untrue opinion is deceived or mistaken, since by
hypothesis, in consciously forming an opinion, he does not Judge
that A is B. He is quite aware that the evidence does not prove
A 1s B, however much 1t may incline him to suppose that A is
B and to act on that assumption. It follows that what we
suppose to be deception, and mistake proper, has not been pro-
vided for under opinion So now 1f we assume, in the case of
Judgement, that, should the evidence be insufficient, a man must
know it 1s so, and therefore decides only on sufficient evidence,
there would be no place in Judgement either for what seems to
us mistake proper. There would then be no place for error
proper, unless Judgement could be false.
Nevertheless, this distinction of Judgement into true and false
1s not so easy as 1t looks, and 1s indeed fraught with difficulties.
It is essential to suppose in false Judgement, as above con-
ceived, that the man should be sure that the evidence in favour
of A 1s B 1s conclusive, for, 1£ he were not sure, he would be
aware that he was only forming an opinion. Suppose 1t possible
that there should be such a mistake about evidence ; what
would be the ordinary way of describing the man's frame of
mind ? Probably, if people were thrown back on the ordinary
categories, as they couldn't say 1t was knowledge they would
call it opinion But this 1s an entirely incorrect use of the word
opinion, for it is certainly not opinion to the man, who does
not regard himself as merely forming an opinion. We may say
it has no more value than opimon, but 1t is wrong to say that
1t is an opinion of the man's. For opinion implies consciousness
of the insufficiency of the evidence, whereas 1t 1s JUSt the charac-
teristic of the case before us that the man 1s sure the evidence
is sufficient. It is only a form of this to say the man does not
know but thinks that he knows. But agam, in any distinction
that can be made between thinking and knowing, this is not
true of the man's own frame of mmd. He certainly does not
ro6 STATEMENT, THINKING, AND APPREHENSION
say to himself ' I think I know ', for that must mean he knows
he does not know.
Smee however the man m question doesn't know and is
unaware that he doesn't know and further behaves as 1£ he did
know, we doubtless mchne to say of him that he tlunks or
supposes or believes that he knows A 1s B. But now we have
seen that the man's attitude of mmd to his own process of
decision 1s not thinking m any sense m v. h1ch v. e oppose thmkmg
to knowmg The fact 1s that, 1£ we look first at what 1s decided,
the man doesn't know that A 1'! B, for ex hypothesi A 1s not B,
nor does he ' thmk ' that A 1s J3 Ill the sense of formmg an
opuuon, nor docs he suppose that A 1s B, nor does he believe,
~lnrtly, th.it A 1s B. The man v. ho decides that evidence proves
that A 1s B doesn't lnmself ' suppose ' A 1s B-the word 1s qmte
.LhL·n to the altitude - .u1d Just the same 1s true of belief
Scrondly, 1£ \\ e look at lus attitude to his o,, n mental process,
not only docs he not ' thmk ', ' su 1>pose ', or ' believe' that A
1s B, but also ht: docs nut thmk, s1,ppost:, or beheve that he
knows A 1s B , for these phra..,e<; art: self-(·ontradirtory To lus
consc1ousnc,,s. then, from neither pomt of \ 1ew, can \\ e apply
the,;e word'> ' apprchcnsum ', • op1mon ', ' suppos1t1011 ', or
• Ld1cf '.
N'o\\ this should l'.t11'>e us hcs1tat1on ; though, perhaps, we
rn.1y mdme to think that the result 1s only that these categones
-opm10n, suppm,1t1on, belief, and apprehens10n-do not apply
to tht· man''> consuousncss and that 1t t'l both nght and sufficient
to say that hL· not only Judges or decide!:> on mcomplete evidence,
but dre1drs that thl' mrompktl' cv1dcncc ,.., proof, m fart to say
th.it th,.., 1!:> ... unplv the n.iturc of false Judgement
\\'t• tmd our!>dvc!:>, ho\\l'ver, fared hy a new and senous
difficulty Tilt• man \\ ho 111,1kt,; the supposed km<l of mistake
dl·c1des that the LT1dcnrc proves , and so he 1s m exactly the
same fr,rnw of m111d .1s v.hcn he decides th,1t the l'Vtdence proves
and 1t really does prove Hts conv1ction \\ ould be of JUSt the
same kmd to hunsl'lf m both rasc", both when he doesn't know
and whcrl' ht:. conclus10n J!:> ti ue (\\ h1ch we n11ght at first be
mchncd to rail kno\\ ledge) If t 111,; "ere not so, the man would
di once become J.\\-are of the <l1flercnce, see that he had not
found the c, 1dence conclus1ve, and would then either form an
 Opinion, Conviction, Belief
opinion or look again at the evidence. His conviction then being
of the same kmd to himself when his conclus1on was right and
when 1t was wrong, he would be unable to d1stmgmsh the one
mental state from the other ; the confident state when nght
and the equally confident state when wrong. We might put
this by saymg that he hac; no criterion of truth or knowmg
That however would be m1slcad111g because 1t would imply the
fallacy that there could be a general criterion of knowledge•
by which we should know what was k110,,mg and what was
not
Alternatively we might ,;ay that the man doesn't know whether
he 1s kno\\-mg or not However, the phrase 'know that \H'
kno\\ ' may agam mislead, because 1t rather tends to imply that
we could conduct a pro< e'>s, for 111stancc prov mg that A 1s B,
.1nd then decide othcnnsc that 1t was a knowing }Jroces"i. But
the decision 1t self would be a knowmg proce.,., and ~o v. c should
get ml o an uncndmg series of knowmgs Moreover 1f we could
so decide, m a new attitude of thought, ..ibout the given process,
\\C should not dl•c1dc that A 1s B until tins second process and
tlus second dcrmon, namely the de< 1s10n that the process of
arnvmg at 'A 1<, B' \\cl"i ,l knowmg- proresc;, had taken place
But m the first prorcs.,, 1ust because 1t 1c; a knowmg process
(by hypothcc;1<;), "e Juve ,dready decided that A 1s il , indeed,
1t 1i; !Jy Uw, proc c•i,s .1lonc that we can so dcc1d1· ,md not by
any dcc1s1on about the prot ec;s 1tc;elf The rom,nou"incss that
the knowrng process 1s a knowmg proce!,S must Le contained
\\ 1thm the know mg process itself
A correct way to put the ca&e before us seems to be that the
two processes, the two states of mmd 111 winch the man con-
ducts his arguments, the correct and th(• erroneou"i one, arc
quite 111d1stmgu1&hablc to the man himself But 1f this 1s so,
as the man does not know m the erroneouc; <;t,1tC' of mmd, neither
can he know m the other state
The conclusion of the process 1s true 111 the lat tLr case, but
that does not make the suppoi,ed <lec1c:;1011 or Judgement mto
knowledge Thus, 1£ the given <'asr of faJ.,c Judgement were
l• A favourite po'>lbon denvC'd from K.i.nt • \ ,utflncnt and at the same
time general mark of truth cannot po•sihly be founrl.' J\11t1k d r V, E-d 1,
P 59, cf 1b p 151]
n>8 STATEMENT, THINKING, AND APPREHENSION
really possible, we should never be sure that any demonstration
was true and therefore there would be no such thmg as demon-
strative knowledge
§ 53 The result then appears to be this Judgement being
dec1s1on on evidence after c:lehberation, there seem to be two
alternatives First, either 1t 1s seen that the evidence 1s insuffi-
cient and then 1t 1s not dended or Judged that A 1s B, and so
at most there results the opm1on that A 1s B, not the Judgement
that A 1s B, and the man 1s not deceived even 1£ the opinion
is untrue, ma:,;much as he kno\\ s that 1t 1s not a certainty. Or
else, 1t 15 nut seen that the evidence 1s insufficient and the man
decide-, on evidence which does not prove In tlus case he. is
entirely mistaken, and this would be false Judgement But this
assumption of false Judgement leads to the 1mposs1b1hty of any
Jemonstr,1tl\-e rcasonmg, that 1s, of any reasoning v.luch would
be knowlt'dge many propt..r -,ense
Hence we art· ll'd tu a sclOnd al!crnahve, that in VI-hat con-
cerns reasomng the only kmd of m1,;;take 1s th,. mistaken op101on,
winch 1s not nm,taken rc<t.,m1111g, ,rnd that 10 rt:a.,onmg itself
there can he no nustakc Tim 1., the same as saying that there
1s no sue h tlung ,1<; f.tJ.,t' Ju<lgt'mcnt But this seems to leave
no pl..icc for the most important kind of error for 1t seems
obvious that a man may be wholly mistaken and, the ev1dt'nce
bemg msuflic1ent, may say, if asked, that he knows A 1s ll, and
has not merely an opm10n that A 1s B Thus he at least appears
to JUd~c and dl'cult- tlut A 15 B on m.,ufiiuent e\-1dencc
The obJt·ct1ons to both altt-rnat1,•es o;cem vahd On the
hypot lws1s that there can lw fahe JUd!!cment we could never
be sure tlut .rny • demonstr.1t10n' \\as knov.kdge Yet we
are surl' I hert• 1!> sur h, .111d so c;omeho\\ false Judgement 1s not
possible On the other hand, there arc cases of deception, 111
the full sense, and error 1s not confined to false opm10n (m the
proper Sl'nsc of opuuon) We conclude, therefore, that the
analysis 1s imperfect, that there 1s real error other than false
opin1011, and lh,LI 1t docc; not hl· 111 false Judgement, taking
Judgement m its stnct and proper sense
The rnmplet c ,111swer must ht· cl den l'cl to the discussion of
mfcrenc-c, but the d1stmct1on on which the solut10n turns "111
be m,i<le dear by a d1scuss1on of stalt':. of consnousness which
 Opinion, Conviction, Belief I09

simulate Judgement and which cannot be described in terms

either of Judgement, behef, or opinion.
§ 54. There 1s another fact of consciousness which is difficult
to describe m terms of judgement, belief, or opinion Suppose
that m our experience a thmg A1, or a kind of thing A-ness,
has always had the characteristic X-ncss. We may easily recog-
nize that we do not know that X-ncss necessitates A1 or A-ness,
but m our experience only A 1 or A-ness has had the charac-
teristic X-ncss There 1s nothmg then m our previous experience
to put us on the alert and make us remember that we do not
know X-nes!> to be ncn·ssanly a sign of A1 or of A-ness.
Suppose a case X 1 of the rharacteristic X-ness, wluch 1s not
A, to occur for the first time 1 111 our experience Sometimes it
hd.ppcns that v. c treat X 1 as tf 1t were- A 1 or A, or our attitude
toward X 1 1s as 1£ 1t were A We take some practical step, m
cunsequcm e, of such importance that when we reflect upon 1t
\\C must sLiy that only knowledge that X 1 WclS certamly A 1 or
A could justify our <1.ct10n, 1f we thought clearly of what we
,,ere about We could not have acted upon knowledge that X 1
was A, for X 1 w.is not A Did \H' act on opm1on or on belief
(m the strict sense) ~ If so, we should have been conscious of
riskmg somethmg · hut 111 the given use we acted unhes1tatmgly
and were quite unconsc10us of nskmg anythmg.
For example, we c:;ce at a little d1!>tancc a pcrson whom ' we
mistake for an acqucuntancc' and without hei,1tat10n perform
somt" act winch 1t would he a liberty to take with any <•lle but
an c1cquamtance, do somethmg m fact which we nghtly say we
should not hJ.vc done tf we had ever suspected he was not an
acquamtancc We did not act on an opmion that 1t was our
fnend ; for, 111 formmg .1.11 opmton, we arc aware that the
evidence 1s msuflic1ent and, 1f we had thought that, we should
never have done the act. It seems more like behef; but, 1f
we had consc10usly made 1t a matter of belief, we should have
d1stmgu1shcd 1t from knowledge, and then agam, ex hypothesi,
we should not have done the act Probably one answer offered
would be that, though we didn't know, we thought we knew.
But this will not suffice Apart from the cnttc1sm we have
already passed on this phrase itself, 1f we really thought we
1 It need not be I.he fi.n.t time, but this case 1s taken as the clearest
 '\"'"
.,IIO STATEMENT, THINKING, AND APPREHENSION
knew, we must have reflected and must have thought the
evidence conclusive, whereas, ex kypothest, any reflection shows
it could not be conclusive
For the kind of cons< wusness we are contemplatmg, a moment•
ous practical dcc1s10n 1s not necessary, but we take the case
where there 1-. surh, helause it serves to brmg out the es&ent1al
feature, the ,Lb5cncc of any sense of uncertamty or doubt, the
action bemg one wl11d1 \\ould not be done 1f we felt the slightest
uncert,11nty.
Pcrceptwn, then, ,md JUdgl·mcnt, apprchl'O!>ton, opinion and
ht hd, seem dll alike exduded It 1& true thcLt, 1f asked, we
might say ' I thought 1t was my friend '-' I believed it was
my friend ' - ' I was sun• 1t \\ as my fnend ', but these expres-
wms .ire ,di 111,Lccuratc The truth 1s, ac; "111 be admitted, that
m the gt\ en 1 ,l'>C, "hcn I pcrLe1vc the fan11har charc1ctenst1cs
of my fr1l'nd, 1t m·ver ' cnt<:rr, mt o my hc,ul ' th.i.t they could
belong to any one else. l don't th1rk about that at all, and so
the pr()( <'"'>t'!, of Judgl·tnent, bl hd, awl opm1on arc 1mposs1ble
The tlunkmg process 1s not fully awakl·ned, and tins fact of
consc-muc;ne~.., thro\\s light upon that general ch,trc1ctenst1c of
all thmkmg, "luch at firr,t m,l) seem to us so empty, viz that
It IS ,10 act I\ 1ty of <011SCIOU~lll'% I
I ,tel u,ill~ pcrCl'IVC X 1 and I trc,tt 11 \\ 1thout reflect10n as 1£
1t "ere A 'l hat 1s, my fr,ime of mmd I'- a5 1£ I \\ ere perce1vmg
c;omt>tlung ,, J11cl1 \\ ,l'> A, and I ,Lill 111 no sen&c conscrnus that
I am pcrcc1\ mg somethmg \\ lud1 might not he A , the existence
of sue h ,Ul ,ttllt ude or fr ,unc of nund 11l'mg proved by the fact
that I further treat X 1 as I\, m tht· pr.1ct1c.tl sense of actmg as
1f 1t were. lt would he truly !>,nd th.it, ,, ht•n I perceive X 1,
I 1mc1g111c ,\ 1 111 t0nscquencc of m)' previous experience But
that 1c; by 110 means enough I do much more, I treat X 1 as
1£ 1t \\ere .1.ctu.1lly A, in regard to my mcnt..Ll .i.ttltudc v.}uch
simulates Judgement and c,mnot po&s1bly be reduced to terms
of 1magmat1on Nor is 1t .i. m1:-takmg of the imagmatJOn of A 1
for the perception of 1t It 1s not that, perce1vmg X 1 and

• Doubtlt•s., v.c do not take ~l"r10u~tv Plato'" account of thought ac; the
dialogue of the ~oul wtth 1t!.elf for a true dt hmhon of thought . nevertheless.
v.hen 1t 1s a q11est10n whether we v.ere thmkmg or not, it 1s often useful to
ask,' Did I say &0 and so to myself~' [R 437c, Tht 189e, Sp/, .163e J
 Opinion, Conviction, Belief HI

imagining A-ness in it, I think I am perce1vmg A1. In the

supposed case, seemg the back of a person exactly hke the back
of my friend, I 1magme his face. I do not at all suppose I see
1t, but, without supposing I see it, I treat the person whose
back I perceive as though he had the face of my fnend, both
as regards my mental attitude and the action consequent
upon 1t.
Nor can we explain what happens by memory, m its proper
sense of a conscious activity. I do not, 111 perceiving X 1,
remember that all previous X's were A, or that 1n all prev10us
cxpcnenrcs of X I experienced A-ness, for, m such a case as
the one before uo;, we do not tlunk of the past at all. And, 1f
I did, remcmbcrmg that m the past I experienced A along with
X, I 5hould notice th£' difference of the presl'nt experience as
of X-ne:.s and not of A-ncss, and thus consciousness would be
awa.kent>d to tlunkmg about the relat10n of A-ness and X 1 ,
wluch 1s Just v.hat doe'>n't happl'n. Thus v.c come to recognize
an ultimate fact of 11Ius10n m our ronsc10usncs<; We perceive
an X and, whereas \\-e have always known X as A, wlule our
experience has never m any way <;uggcstcd that an X could be
without A-ness, 1t does not o<,cur to us to thmk of X 1 as not
A This 1s not formmg the pos1t1vc Judgement that X must
be A Y ct m th1c; consc-1ousness 1 which 1s m one d1rect10n
unawakened, on tht> occas10n of the perccpt1011 of X 1, the
1magmat10n of A, together with the ahsencc of any thought that
an X might not be A, has the :.amc effect th.Lt the thought
that X must be A would have m the fully awakened consc1ous-
nes-. We treat X 1, m the practical scnc;c, as if 1t were A The
difference m the two 1s thc1.t, m the awakened consc1ousncss,
c;uch practical 'treatment' would be preceded by the Judgement
that, though A-ness 1s not perceived, 1t must be there In the
unawakened consc10usness there 1s no such Judgement The
peculiar attitude of consc10usncss wluch simulates Judgement
and which \\ e have designated as treatmg an X as 1f 1t were
A, without the Judgement that X 1s A, eludes all attempts to
express 1t m terms of the awakened consc1ou5ness or m any
other termc; but c;;uc-h as we have been ~1'-'Ing pcf'uhar to itself.
For mstance, 1t 1s erroneous, but 1t 1s not erroneous Judgement,
behef, or opm1on.
:U2 STATEMENT, THINKING, AND APPREHENSION
It is futile, agam, • to try to explam 1t as the inseparableness
of the ideas of X-ncss and A-ncss. Such a formula is no explana-
tion. We should find at once that the conception of 'idea'
itself is confuc;ed, tcndmg among other things to mvolve the
fallacy of trcatmg the ' idea ' as though it could be something
;ipart from our consciousness uf 1t, and that it 1s otherwise quite
unus.i.ble. (' I,lca' 1s one of the most ambiguous words m
philosophy) We should find too that mseparableness 1s an
unma'1agc-able phys1ral metaphor In the end the formula, 1f
intclltg1ble at all, \\ould have to be itself explamed by help of
a description of thl· facts wh1rh 1t 1s to expl,un.
Suppose 1t 1s sa1<l that the attitude of trcc1.tmg an X as A,
which simulates Judgement, simply means we ' cannot help
tlunkmg' of an X as A. There 1s a certain helplessness which
causes our error, but what could be the 'thmkmg'? The term
ts perlups l hw,l·n bcc,msl' the st.i.t e of consctousness 1t refers to
1:. not perce1nng, and the term ?tself corresponds to a natural
uc;e of the wolll 't lunh.111g' m phr.isC's like ' I did so and so,
t hmkmg the man \\ as my f ncnd ' If however it means
'1m,1gm111g ', \\l' Juve sct·n th<1t that 1s not enough Agam the
formul,1 c.i.nnot llll'J.11 thc1t \\ c thmk that an X cannot be con-
ceived \\tthout A-ncs<:, that 1s except as bemg with A, for
a moment's rdlect10n v.ould shov. us this v.as untrue Nor can
1t mran that I could not get X-ness and A-ness apart m thought
(1 e thmk th,it an X need not be A) 1f I tned for I certainly
could The truth 1s that it does not occur to me to try and this
1s tht· real groun<l of my helplessness, so that we have to
cxplam our' c.innot help thmkmg' of an X as A, by the absence
of thmkmg ,tbout the rel.1t10n of X-ness and A-ness in the
ordul.lry scnsl' uf the word tlunl"mg
Thus 1t srt·ms that \H' do not make ourselves clearer by the
use of thl' \\Ord tlunkmg m the formula 'we c.i.nnot help
thmkmg of X .ts A', "h1ch we offer as an explanation For
we h.i.ve to C"-plam thc1t v.e do not mean thmkmg as ordmarily
[• Refernng probably to Lolkt h !( 'ideas that, m themse-lve<1, are not
all of km, come tu he ~o muted Ill ~umt mt'n', mmd'I, that 1t 1s very hard
to <1eparatl' them , they alwa}, ket•p 111 company and the one no sooner oLt
any time comt~ mto the under..tandmg, but 1t~ dSM>l.late appears with 1t ,
and if they arc more than two \\lulh arc thm, muted, the whole gang, &c.'
Essay II 33,§5, cl 1b §17]
 Opinion, Conviction, Belief J:IJ
understood and then to explain what we do mean by describing
that attitude toward X 1, as if it were A, which simulates the
Judgement that X is A
We return then to what seems the accurate account of what
happens, viz that, perceiving X-ness to have always before been
rxpenenced with A-ness, and 1magmmg A-ness m consequence
as m X 1, 1t doesn't occur to us to think of X as without A and
we behave as 1f A were there.
Th1s agrees with an explanation which would often be given,
"h1ch 1s, not that we thought X 1 v. as necessarily A, or even
that we could not help thmkmg of an X as A, but that 1t never
occurred to us to doubt 1t; m truth we never thought about
1t at all It agrees also with a special and d1stmct usage of
ordmctry speech, 'We were under the 1mprrss1on that X 1 was
A ' This 1s perhaps the one of the ordinary unph1losoph1c
answers v.h1ch 1s most adequate, for 1t seems chosen from
a feeling that the ordinary act1v1ty of thought was not there.
the met,1phoncal \\,Ord ' 1rnpress10n' bemg used to suggest
a certam pass1v1ty and helplessness
We have said that the reason for the peculiar state of con-
sc10usness we are here trymg to recogmze 1s that we have never
l"Xpenenced an X without A-nt>ss, that 1s m the case we h..1.ve
taken It 1s now hardly ncccss..1.ry to guard agamst a m1smterpre-
tat1on of this 1 he reason m such a c,ase 1s not a ronscious reason
for us, m the sense of a premiss from which we mfcr. We do
not argue that X 1 must be A, because 111 our experience all X
has been A. It 1s merely the ground of ou. 1magmmg A-ness
when we perceive X 1, and of its not occurnng to us that X could
be without A-ness The kmd of consc1ousncas • here considered
is of constant occurrence, but 1t 1s usually m mstances wluch
do not attract attention because not of the crucial tharacter of
that here selected. Besides there are numerous mstancec:; where
our attitude towards the obJect docs not attract attention
because, though unwarranted, it docs not lead to a mistake.
That 1s, m many cases we unthmkmgly treat X 1 as A and 1t
also happens that 1t is A
L• Kantsa1d that the ground of error 1<, • the occult mfluc.ncc of thr ~ens1b1bty
on the understandmg or, to speak more preu~cly, on the Judgement' Logic,
lntroductmn, vu ]
2773•1
 IV
THE DISTINCTION OF SUBJECT AND PREDICATE
IN LOGIC AND GRAMMAR

§ 55 Tm~ ' propos1t10n ' m the older, ' Judgement ' m the
modern tcnmnology, 1., said to be analy<;ed mto subject and
pred1c..ik, or, agam, mto subject, predicate and copula If this
&ccond analysis be .tdopted, propos1t10ns are supposed to be
resolved, for logac1I purposes, mto the form symbohzcd by S 1s
P , where S 1s said to be the subJect and P the predicate, while
the verh ' 1c, ' 1c. called the copula, ,1c; connectmg the predicate
v.1th the bUhJcct
Now, l.lc<..ordmg to the trad1t1on.1I defimt1onc; • of subject and
pred1catt-, thl' &ubJcLt 1s what m tht:: propo::.1tion or statement
we .ire speakmg about, .rnd the pn·chr.-tle ts what we say of
the subJell or. m a f,umh.ir forni, thr i,ubJect 15 th,tt about
v.h1ch ~omethmg ,., asc;ertl'd, .ind the predtcJte 1c; thctt which
1
1s asserted .1bout 11 Tim, agrt'c::. v.1th the derivation of the
v.ord · predic..tte' irom the Latin term prrduatum, a translation
of Anstotll",; Knn,ro1>01ip.wo1• An.,totlc d1stmgu1shc::, v.1thm the
propos1t1on, r,'i rnr,.,yopovµEZ•or .ind ru ,:;a(J' ov ,:;ar11yopE'irai, what
1s said of somethmg and th.It about \\-htd1 it JI'> ::,aid He never
defint•s the d1::.t111rt10n, anc.l gets mto d1ffirult1es m consequence.
The earhest kno" n formul..1.t1on b of the c.l1stmction, amount mg
to a dcfimt1on of 1t, i.l e1nc; to be that given by Boethms He
1 H \\' B Jo~eph, ll1trod11ct1on to Log,c, pp vu, 145
[• ( f '"""!(t 111.111 d,,., ~11h1u t 1-,t d,,.,, wovon t twa, ,1.11,g<•,agt uud das
Pr.id1k,1t 1,t d,1-, \ufi;<•,.11,:h, .,., 1-,t dtl''- ct\\a.o Tnv1dlt'o (u,mmonplac.e) und
llldll erl,,hrt tl,1durd1 Ill< ht, N.,Ju n, ub, 1 den l·11terotJ111 <l d1cscr be1dc11 '
Hegel, H "~"'"/1<1/I drr I oi:rk,, ,,,., (lr,r/.1, 11>40, vu, p 130)
b \\• l,lll, I thmJ.. tr,LLC llu:, l,uhc-r, 01Taii8a -,a,, ll c 'l!o,1<paT11s >r<puraT<i) Tr)
,.,,., ~. i rrmt1i/JfJ10S Vpui A.i11. ra,, 7~ , ; 11Ep,11aT,; 1tar-rryopo'Up.12 os, &Ur, ,,,. ff'tlYTl
1 1

llO.Tf/'Yap&ll9' >.o-y91 'fC) µii' lart ..,pl uli cl >.t-yo1, T<l a• • ,,,i ••• ,, ov >.1-yuµEIIOI', /tl1i TO ,.,.,
'lllpl o1i ,I Au-yOi, i11ro,n/µ1vov At}fTGI 0/5 11,xi.µoor, Ta.s 1to.T' o.vroii nnrroplo.s, TO
a•• ,pl 0.UTOii >.11uµ1ro11 • KO.T'/")Oput;µ,,,,,,, 0/S 1<0.7' (KEi11011 d-yopn,oµFlfOfl ml Af-yv-
,.,.,011, Ammomu~ 111 .I" Dr l11terprrtatr011r, pp 711-8• The copula 1, called
"poaicGT'l'Yopoi,.., w \ <on temporary c>f Hocthm~. he profes5cs to give the
doctrme of l'rodu, (.po 85 A D ) I llf' d(•hmttou or1gmated probably m the
gr:i.mmanans, e g ' alt(•rum (~c verbum) l'~t quod loqmmur, alterum (bC
nomen) de quo loqu1mur' Vumhha.n. Insl 1 4)
 Subject and P,edicate II5
says·-' The parts of simple enunciation are subject and pre•
dicate : the subject is what supports the predicate locution, ..
the predicate is what is said concerning the sub1ect.' 1 Herc
we at once observe a certain awkwardness in usmg the concep-
tion of the predicate m the definition of the subject, and the
conception of the subject in the definition of the predicate.
This 1s clearly the forerunner of the modern formulae that we
arc cons1dermg These formulae are both inaccurate and
ambiguous For, in the first place, 1£ we consider the part of
the sentence to which the term ' predicate ' is applied, we shall
find that the above definition docs not accurately dct.cnbc 1t.
It 1s inaccurate to say that P 1s what 1s ac;serted or c;aid of S.
In th£' normal and non-teduucal use of language, 1f we asked
what wac; asserted of S m this sentence, the answer would be
not simply P but that S 1c; P If some onr, for m,;tancc, said
Jones was a good <1rt1<;t, ,ind some one dsc, hearmg 1mperfcctly,
,1:,ked ' what w.is said of Jones ) ', the answer would not he
'a good artist', nor would any one reply "' a good artl<;t' was
&aid of Jones", but ' that he 1s a good artist'. The difficulty
makes itself very dearly felt if we attempt to ,tpply the given
dl'fi111t1on of predicate to the negative propos1t1on In ' S I!>
not P ', what 1s ,tated about the 50-c.illed suh1ec-t S 15 'tJul
1t 1s not P ' we t ,mnot !>ay thctt P 1s wh.1t 1c; .1ssertcd or st.tied
about S. Consistency with the procedure of r,tllmg P the prc-
d1cate, m the .iffirmativr c;tcttement S 1s P, can 011ly be preserved
either by confining the given definition of suhJect ,md pre 11cate
to the affirmative statement, or cl,e by the e1roneou, device of
reducmg the negative to the c1.ffirmatJve form through the '>Ub-
st1tut1on of ' S 1s a not-P' for ' S 1s not P ' The m1!>t.tke
which would be comm1tte<l 111 such a reduction will be cons1clcrcd
later, when "e come to treat of the negative <,l.11 c-ment But
the truth seems to be that the trad1llonal clcfi111t10n 1s usu.illy
given with the affirmative st..i.tement alone m vJCw, the negative
being forgotten a Till'; 1s but in consonance with the unc.ntlcal
1 S1mphc1um nem enunc1abonum parlt,b ~unt ~ubudum .. tquc prcd1catum,

sub!ectum e,;t, quod predu.ab o;usc1p1t cl1<-tmnem prechc.atum ucro r~t.

quad chc1tur de ,;ub1ecto Boeth1m,, lntroducho ad Gutcgor1cos syllog1s1no~,
ed Basie, 1570, p 562 (Prantl, Geschtc hie, 1 p 696)
L• But cf • A propos1tton, auordmg to the. <.ommon ::,1mplc definition,
I 2
n6 STATEMENT, THINKING, AND APPREHENSION
character of the development of this part of the ordinary
logic
Another inaccuracy in the ordmary way of speakmg about
predication may be noted here It 1s usual to say that in the
statement, steel 1s hard, what 1s predicated of steel 1s hard-ness,
and yet, in the analysis of the propos1t10n, 1t would be said that
' hard ' 1s the predicate So that ' what 1s predicated ' would
actually not be the &amc as 'the predicate'
To return to the mam pomt, we may say shortly that m the
ordinary uc;e of language 1t 1c; the whole statement ' S is P '
which 1s 'what 1s said about S' So, after all, the phrases
'what 1s said of' and 'what 1s asserted about' are not an
ordinary, but a peculiar, use of language; m fact are really
techmcal phrases found m logical treat1sec; Thus, 1t being
required to explam the techmcal term 'predicate', the explana-
tion 1s (unconsciously) given by another techmcal phrase, which
Il!,elf rcqmres cxpl.-rnatwn
The accurate description then o' P 111 ' ~ 1s P ' 1s not that
1t 1s what 1s a'i'icrted of thr suh3ect, Lut simply thdt 1t 1s what
the subJcct 1s asserted to be I But now, 1f we take th1'i as
a corrected defimt1011 of p1ed1cdte, what would Le the predicdte
m the &tatrment 'A ,,alks fast', a form of expression m which
It 1s not said that A is anythmg ~ Statement<; of this latter
form have to be co11!,1dered, because, the given definition of
subject and predicate bcmg expressed without hmitation, 1t 1s
naturally implied that any form of statement or d.c;sert1on has
a sub3ect and predicate. Now, 1f the 1mphcat10ns of th1c; defim-
t1011 arr accurately followt'd (wlurh, observe, involves dtscardmg
the copula from the general formula of the analysis of a pro•
pos1t1on or statement), tlten 'Jonec; "alks fast 'has a pred1c.ate
If tlus pred1cJ.tc 1s ' \\ alks fast ', then, as before, It would not
be accurate to s,1y that the predicate 1s what 1s asc;erted of the
subJect, for "hat H, cu,serted of Jones 1s ' that Jones walks
fast'. As m the propos1t1011 '\V1l!tams 1s a good runner',
what 1s called the predicate 1s, accurately speaking, not what
• ttaTf/"ropovµ,vov could mean 'that winch the ~ubJt•ct 1q stated to be•
is d1&course 1n "h1ch something 1s affinned or dcmed of qomethmg The
predicate 1s the name denoting that which 1s affirmed or denied, &c • Mdl,
A Sysltm of Logic, Bk I, ch I, § 2 Cf however infra, p 249]
 Subject and Predicate II7
is asserted of the subject but what the subject is asserted to
be, so in ' Jones walks fast ' the supposed predicate would be
what the subject 1s asserted to do We should thus have
a different account of the predicate for each of these two pro-
pos1t1ons, whereas we require for 1t one and the same definition.
Tins difficulty 1s not avoided by reduction of all st.itements
to the type 'A 1s B ', on the ground that in every statement
m which the verb 'to be' (that 1c; the copula) does not occur,
there is nevertheless something about which an assertion 1s
made and something asserted of 1t, and therefore, accordmg to
the given definition, both subject and pred1cak The fact is
that, whereas the trad1t1011al analysis of the proposition 1s mto
sub1ect 1 copula, and predicate, 111 the trad1t10nal definition of
subject and predicate, the copula has been overlooked
But, besides these maccurac1es 1 the formula of the definition
is ambiguous, and if it were our only guide we could not tell
what was subject and what predicate m a given !,tatement.
How could we determine the sub1ect and predicate, for mstance,
111 ' Wheatley 1s farther than Headmgton from Oxford ' ? It
\\ould probably he said that Wheatley 1s the !,Ubject, for 1t 1s
wlut we are tlunkmg .ind c;peakmg about, but we arc also
tlunking and speakmg about Headmgton an<l O~ford, for
example that Ilca<lmgton H, nearer to Oxfur<l ch.111 Whe.itley 1s.
~ 56 But no,\ a d1stmct10n may exist, an<l be recognized m
particular example,;, and yet not be clearly formulated or under-
stood in the abstract. By considering such example& we may
perhaps discover what the ordmary imperfect and ambiguous
defimt10n is a1mmg at
Suppose a stranger m Oxford stops before the Bodleian
Library, and asks lus gmde • What bmldmg 1s that ? ' The
reply 1s, • That building 1s the Bodlc1an.' If we arc- to find an
apphcat1on for the ordmary definition of subject and predicate
to the reply, 1t would be natural to say that what the guide was
c;peaking about was the buildmg pomted out by the str,mger, and
thus ' that building' would be the subject. What he says about
the subject, agam, might naturally be taken to be that 1t 1s the
Bodleian So that, 1f the predicate 1s 'what 1s said of the subject',
the predicate would be ' that that bmldmg is the Bodle1an '.
Again, suppose instruction was bemg given m the properties
II8 STATEMENT, THINKING, AND APPREHENSION
of glass, and the instructor said ' glass 1s elastic ', 1t would be
natural to say that what was bemg talked about and thought
about was ' glass', and that ,,.hat was said of 1t was that 1t
was elastic Thus glass would be the subJeCt and that it is
elastic would be the predicate. The same would hold 1f we
arrived at the st..itemcnt ourselves by an mvestigat1on of the
properties of gla%
We oLserve that m the first example there 1s a peculiar stress
accent upon the word ' Bodle1an ', which 1s absent from ' that
bu1ldmg ', and, 111 the second mstance, that there 1s this same
stress upon the word ' elastic ', and that 1t 1s wantmg from
' glass '. We observe, further, that m each case the stress accent
1s upon the part of the sentence which conveys the new mforma•
lion, and that 111 each case the words from '\\hich 1t Js absent
arc thm,c '\\luch correspond to what v.ould seem to be the
subJect.
From these examples v.c may l.il'm,c an unambiguous dJstmc·
t10n, v.h1rh ,;ecms to be the rationale of thr tradJtJonal account.
Thl' subJert of a st,ttcml'nt mc1y be ddined as '\\hat wc were
tl11nk111~ ..ibout ,ls wc thought 1t 1 or conceived 1t, before we
arnvctl ,1t the statenwnt, or before v.e h,¼d the statement com•
mumc,1tetl to us, "lulc the predicate 1s the new fact wluch we
statt· about 1t or haH' communicated to us Or, to put 1t
otherwise, the !:.Ubject 1s v.hat we v.ere thmkmg of, as we thought
1t or conceived 1t, before forming the Judgement, opm10n, or
beltcf v. luch the statement expresses, or to '\\ h1ch, at least, 1t
corresponds, "luk the prcd1c.1tc 1s the new fact, or what we
supposl' to he the llLV. fart, v.]ll(h \\l' come to kno\\ about 1t;
10 other \\ords the new thmg v.h1ch we state about it. S1m1larly
m the r,1se of .1 !e-lJ.tcmcnt not strictly arnved at by ourselves,
but commumcJ.tcd to us by another, the predicate for us would
be what W,\s statl'd by the spcJ.kt'r to be the new fact about
th.it of winch he was speakmg. or ,, hat was mtcndcd to be the
new mformation given us by lum
We may arrive quite !>1mply at the same prmc1ple by reflecting
upon the 1mphcations of the ordmary defimtion. The predicate
1s defined as what JS asserted of the subJect Now an assertion
1s either for the mformat1on of others, or merely the verbal
expression of knowledge or opm1on we have arrived at for our-
 Subject and Predicate II9
selves. Clearly the purpose of such assertion 1s, in the one case,
not to inform others of what they already know of a given
subject, but to tell them something wluch will be new lo them ;
and, m the other case, the expression, as a statement of know•
ledge or opimon v.luch we have arrived at for ourselves, embodies
something new. It 1s true that there may be occasions when
we repeat to others what they know already, or to ourselves
what we have already conceived, but that cannot be the original
reason for makmg an assert10n. Thus ' what 1s asserted ' should
properly be somethmg 111 the subJect considered or talked about
wluch 1s new as compared with ""hat we ourschcs or our hearers
ortgmally conceived 1t lo be a
§ 57. If we now examine the actual practice of grammatical
treatises, we shall find that the analysis of c;cntcnces mto subject
and predicate does not agree with the prmc1ple ""hich we have
found to be properly implied m the ordmary defin1t1ons of thci.e
terms Consider first the analysis of scntencec; mto subject and
predicate. Suppose our c;tranger now wishes to find the Bodleian
,md askc; 'Which bmld111g 1s the Bodle1an ~ ' The reply 1s 'That
bu1lclmg 1s the Bodle1an '-m the s,tmc wordc; ,ls before, but with
a. different strebs•au:ent It would now be natural to say that
what was thought and c;poken about was the Bodlc1an and
that this therefort: \\ac; tlw i.ubJect: that what was said of 1t
wac; that the butldmg pomtcd out was it So agam, m the
statement ' glass 1s elastic ', 1f the matter of mqu1ry was clast1c1ty
and the quest10n wab wh.i.t substances poc;scs.,ed the proi:erty of
elasticity, glass, m accordance with the prme1ple uf the defini-
tion, would no longer be sub1ect, and the kmrl of stress wluch
fell upon •elastic' when glass was the sub3cct, would now be
transferred to • glass ' In both cases the stress falls upon the
[• This whole doctrmc of the !>trc&& on the prc1hcatr ongmatecl, I 'IU'lpcct,
1n Whately,' Drtft of a Proposition', Element~ of Logic•, II 1v, § 1, c'lp • And
ob<.erve, that when a propos1t10n 1s 1.ontra&tcd with one wht( h ha'> a different
predicate, the Pred11.atc 1s the emphatic \\Ord, a'I" th1., m.111 •., .i murderer",'
and so on I have 1.orrccted 'accent' to ''!tre'IS' .ill through, \Vilson u .. mg
the two words indifferently Whately draw'! attrnt,on abo, properly, to
verbal 1.ollo1.abon and mtonat1on, the last bemg ignored 1.,y Wilson See
Paul, Princ1p1en der Sprachgesch1chte•, § 173 (Enghsh Tr.in'!, p 114) The
English grammarian ha'I to weigh (a) word-order, (I,) '>trrs'I, (r) intonation,
(d) the use of form-word,;, and (e) mftectlon S\\CCt, Sew 1:nglish Grammat'
(1900), I 8I et seq On p 229, mfra, .iccent apparentl}-=-tone]
120 STATEMENT, THINKING, AND APPREHENSION'
word or words in the sentence which correspond to the nc-"'
infonnation, and 1s absent from the words which correspond 1,,
what was sui-,posed to be in the object before the 10format1011
conveyed in the statement was acquired. Thus the same form
of words should be analysed differently accordmg as the word.,
are the answer to one question or another.
But the ordinary analysis rnakes no d1stmction whatever
between the cases where the same form of &entence 'glass 1s
ela,;t1r ' 1s the answer to two d1ffcrent questions. In both cases
abke the nommatlve case to the verb-glass-is treated as the
logical subject and ' ela!>t1c' (or elasttc.ity) a!> the predicate.
The differrnce of stress 1s entirely ignored, a serious error, for
the stress-accent 1s an important part of the whole express10n
This mistake mvahd,1tcs the ordmary method, and shows its
complete failure to realize the prmc1ple which underbes its own
general definitions Nor 1s tins surpnsmg, smce this pnnc1ple has
not come either to clear consc10u:,nes:, or to precise formulat10n.
Next consider the traditional atr0unt of ..t '>entence like ' glass
is elastic', "'hl'rc glass 1s Loth the grammatical and the true
logical subJcct.
Take first, m, is usual m grammar, the \\Ord elastic to be the
predicate Can a jUstificat10n be found for the ordmary restric-
tion of the predicate to a part only of what H, asserted of the
subject? In the sentence 'glass 1s elastic' the word 'elastic'
doesn't by itself convry the new mformat1on, for that requires
the whole sentence-for its expression and thus the restncllon
of the predicate to the word ' elastic' 1s not really consistent with
the defimt1on of the predicate a:, \\hat 1s asserted of the subject,
for this dors not properly mean any words but the fact which
they expre::.s, that 1s, m tlus 1llustr..1t10n, the elast1c1ty of gldss
Elastic 1s however the word m the sentence \\luch relates to the
supposed new fact about glass, tl1e other words only relatmg to
what 1s already kno" n (or supposed to be kno" n) of the object.
We might then give as a rat10nale of the ordmary account,
where the word ' elastic ' 1s called the predicate, that the pre-
dicate, as '"hat 1s asserted of the subJect ', 1s confined to the
word m the sentence referring to the ne"' property of the
sub1ect
But, though this agrees \\1th the ordinary usage where the
 Subject and P,edicate '*'
nominative case to the principal verb is the true su~ject: t~e
.i.pphcation of the principle to the case where th~ nommat~ve IS
not the true logical subject leads to a resultfq u1te at ~ar1an~e
111 th the ordinary analysis. When the stress a1ls upon g1ass ,
m • glass 1s elastic ', there is no word in the sentence which
denotes the actual subject elasticity ; the word ' elastic ' refers
to what 1s already known of the subject, and glass, which has
th(• c;tress, 1s the only word which refers to the supposed new
fact 111 the nature of elasticity, that 1t 1s found m glass Thus,
accordmg to the proposerJ formula, 'glass' would l1ave to be
the predicate In 'that building 1s the Bodlcmn ' (with the
stress on 'that'), the predicate \\ould sunilarly be 'that' or
' that building •.
But the ordinary analysis would never admit that 'glass'
was the predicate 111 the given sentence and clast1c1ty the sub·
Ject. It would probably be said that 1£ anything could be
predicated of elast1<-1ty 1t would not be 'glass', but 'being m
gl.1s& ' , 1t would not be acknowledged that elast1c1ty was the
sub1cct at all or that anythmg v.as predicated of elai;tic1ty m
'glass i,; elastic'. It u., therefore, 1mposi.1ule to find a rationale
for the ordmary procedure on the method proposed, 1f subJect
and pred1catl' are to be words m tlll' -;entenre Besides the
m&uperablc difficulty m the I rc..tlment of the predicate, m tl1e
sentence 'glass 1& elastic', there 1s the further difficulty that,
clast1c1ty bemg the true subject, there i~ no word 111 the sentence
to denote 1t
To meet tlus ou1ect10n, let us suppo&e that the r.ubJect and
predicate are not necessarily words m the 1:,entcn"e, nor even
something denoted by words m the sentence, although one cannot
say that this modified 1mphcat1on is at all rccogmzcd m gram-
matical practice We \\1ll restrict the predicate m a different
way, and make 'clast1c1ty' the predicate 111 the sentence 'glass
1s elastic'
We have then to ask what could be the rationale of an inter-
pretation which makes elast1c1ty the predicate m the given
sentence
When it 1s said that what is predicated of glass 1.s elasticity,
the fact of elast1C1ty 1s '\\hat 1s meant What, no", 1s the exact
meaning of ' elasticity ' when 1t 1s said to be what 1s predicated
122 STATEMENT, THINKING, AND APPREHENSION
of glass in the sentence 'glass 1s elastic'? It could not be the
special clast1C1ty of glass, for that "ould be elasticity as ex1stmg
m glass, which ts the same fact as glass's bemg elastic, or the
fact that glasc; 1s elastic. Thus to say that elasticity ts pre-
dicated of glass would amount to saymg that 1t is predicated
of glass that it 1s elastic, We &hould thus once more have the
whole ,;entence as expressing ' what ts asserted of the subject ',
whereas, by hypothec;1s, a more restricted view of the predicate
is mtended
It remamc; that 'l'lac;tw1ty' should be the quality or property
of elast1nty Ill general, .md ..,o a universal Tlus agrees with
the use ot the word •elastic', through which the so-called
predicate clast1c1ty 1s 1mphed1 because tlus adJect1ve docs not
refer to any particular cla,;;t1c1ty, but to clast1c1ty m general. It
agrees ,tlso \\-1th the me,-mrng of 'bcmg predicated of' m the
usage according to v. luch cl..1.<,t1ut y \\ ould be said to be ' pre-
dicated of glac,c:; ', for that expression 1s taken to be the equivalent
of bemg as!:tcrted to belong to gla.;s So, 1f clast1c1ty, as the
predicate, mrant the special ela5t1c1ty of glass, we &hould have
the tautologou" statement that the da,;tic1ty of glass belonged
to gl.iss The elastir1ty, then, \\ h1ch ts said to be the predicate,
1s a kind of hcmg possc'>!:ted by the subject glasc;, and the new
inform..it1011 11:, that gl.11:,1:, has tlus kmd of bcmg
§ 58. The general prmc1ple under wluch this would fall 1s (1f
we assume the rat10nalc g1"en for the di&tmction of subject and
predicate) th,1t the subJect 1s an obJect as known or conceived
before the mformat1011 given about 1t 111 tht! statement, wlule
the pred1c.itc 1s a kmd of bcmg assrrtrd m the given statement
to belong to tht.' ol,ject, but not comprn,ecl m what was before
conceived to belong to the object
This would apply also to an instance where the word following
•is' ('arc', &c) v.as not an ad1crttve but a common noun.
If, howc\·er, instead of the adJect1vc there "ere a noun which
denotes an mdiv1dual, as m ' that bu1ldmg is the Bodle1an ',
accordmg to the analogy of ' elast1c1ty •, or • bemg elastic·, m
' glass 1s elastic ', the predicate should be ' bemg the Bodle1an '
The predicate here 1s not a kmd of bemg m general, but that
particular bemg "h1rh is bemg the Bodleian The same happens
1f the md1cat1ve of the verb 'to be' is follo\\ed by an adjective
 Subject and Predicate 123

or ad1ectival phrase so qualified that it can only rt'fer to some

mdiv1dual or particular bemg , for instance, ' this blossom is
further out to-day than it was yesterday'.
We may generalize from these examples thus -the subJect
being (accordmg to the previous defimbon) the object as orig1-
nally conceived by us, the new information conveyed in the
statement 1s that the object has a certain reality or being other
than that which we first thought 1t to have. This bemg 1s the
predicate and 1t may, accordmg to the form of statement, be
either universal, that 1s a kmd of bemg belonging to the subJCCt,
or particular, that 11> a part of the md1v1dual bemg of an ind1•
v1dual sub3cct or somehow 1dent1cal with the whole of 1t. The
foregoing would also hold for sentences m wh·ch the principal
verb was not the verh 'to be ', C' g 'Jones w,tlk,; fa,;t ' In
these the predicate 1s a part1cul.tr or a umvcrsal form of bemg,
according as the verb is particularized or not
If then we assign such a meamng to the d1stmchon of subject
and predicate, the words m the sentence which relate to the
predicate have that peculiar stress-accent given lo lhc new part
of the mformation, which we may call the prechcative stre,;s.
These words themselves we may call predicative words, but they
\\tll not always denote the predicate. When, for example,
' elastic ' 1s the predicative \\ ord m ' glass 1s elastic ', the pre·
dicate 1s not denoted by the adJect1vc ' elastic ' but by the noun
' clastic1ty ', a word wluc.h does not appear 111 the given state·
ment In this case, however, the sub3ect ' glasi,' 1!> denJted by
a word m the sentence.
When m the same sentence 'clast1c1ty' 1s the subject, the
predicate 1s ' bemg m glass ' or ' property of glac;s m general ',
and 1s umversal The • predicative word ' 1s now ' glai,s 1 1 and
this has the stress, but 1t does not denote the predicate. The
subJect 'elast1uty' 1s not denoted by any word m the
sentence: all that corresponds to it 1s the a<lJect1ve 'elastic',
and on the analogy of ' pred1cat1vc word ', we may call ' elastic '
the 'subJect word'. In the case, then, where 'gl..tss' has the
pred1cat1ve stress we have ' ~ub3cct words ' and • pred1cat1ve
words', but neither subject nor predicate 1,; denoted by words
m the sentence.
In ' that building is the Bodleian ', the subject 1s denoted by
124 STATEMENT, THINKING, AND APPREHENSION
the words in the sentence ' that building ' ; the predicate also,
which is ' being the Boclleian ', is represented in the sentence
by the words ' the Bodle1an '. Similarly, 1£ the stress 1s upon
' that building ', the subjt.ct and predicate are represented by
words in the sentence.
The above analysis would make the distmcbon of subject and
predicate, one not of words but of what 1s meant by the verbal
expression. We may call this the strict logical analysis, and the
distmcbon of the words of the sentence into ' subJect words '
and ' pred1cat1ve words ' may be called the grammatical analysis.
This, of course, differs essentially from the ordinary grammatical
analysis, which, however, cannot be Justified by any reasonable
mterpretabon of the ordinary defimbon of subject and predicate.
It rema1m, to say that the choice of one or other method of
formulating the d1stinct1on of subject and predicate, in accor-
dance with what seems to be the only rationale of the traditional
defimt1on, 1s a matter of no great moment, for the distinction
is of no importance in logic proper, and mdeed of no use what·
ever for the solution of the usual problems of logic It 1s true
that the termmology 1s in constant use in Logic m such problems,
but, as we shall find, this 1s due to a mere 1llus1on, the distinction
being confused with something entirely different The mam use
of it seems to be to explain a certain form of stress-accent and
its pos1t1on m the sentence •
§ 59. It follows that the distinction between the logical subject
and predicate would not be expressed by the grammatical form
as such. For instance, m the sentence 'that bu1ldmg is the
Bodle1an ', the nominative of the verb is the logical subJect, 1f
the sentence 1s the answer to one question, and 1t 1s the pre-
dicate, 1£ the answer to another question. It 1s the stress and
not the grammatical form which here marks the predicate and,
in general, 1t 1s the stress only which marks the words which
belong to the predicate-what wc have called the predicative
words. It is true that a special grammatical device, such as
' It 1s Jones who is rowing badly ', may help to mark the
predicate or the predicative words with more emphasis, but it
I• Cf Ueberweg,Log,c 1,§68,p 16o 'Thelog1calmeanmgofthegram1nat1-
cal relations m the sentence has seldom been nghtly appreciated •, and the
foot-note there with reference to Trendelenburg, Log. Unt •, u 253,)
 Subject and Predicate I25
cannot do this apart from the stress, as we shall see, for the
stress must fall on the word ' Jones '. With this stress, the
sentence answers the question ' Who is it that 1s rowing badly ? '
Then 'Jones' is the predicate according to our definition. But
the stress may fall on the verb ' 1s ' and then the sentence
answers the question ' Is 1t Jones who 1s rowing badly ? ' or
'ls it Jones who 1s rowing badly? ' The statement, having
a different stress and answermg a different question, requires
different logical analysis and the predicate is no longer' Jones'.
What the predicate 1s may be seen better m an easier case of
the same kind. If we ask ' what sort of a skater 1s Smith ? '
and the answer is ' Smith 1s a good skater ', the stress will
necessarily fall on 'good' and not on 'skater '-to answer with
the stress on' skater' would be absurd,-and the true predicate
1s 'Smith's goodness of skatmg'; accordmgly the corre-
spondmg word m the sentence, namely 'good', has the pre-
d1cat1ve stress The reason 1s that, before the mformatlon was
given, Smith was conceived as a skater and so ' skater ' belongs
to the subject as we have defined subject; the new knowledge
sought 1s what kmd of skater he 1s Agam, ' Smith is a good
skater ' may answer to the quec,t1on ' what can Smith do
well ' ; this time the stress will be on ' skater ' and not on
' good ' The subject 1& Smith conceived as bemg good at some-
thmg or other, so that ' good ' now belongs to the conception
of the subject, which 1s the reason why 1t has no stress; and
' what Smith 1s good at ' bemg the addition to this, m the new
mformation, ' skater ' has the stress and 1s by our defimtlon
the predicate.
Now, 1f we ask whether 1t 1s Jones• who is rowmg badly or
whether 1t is Jones who 1s rowing badly, we are thinking both
of Jones and of the man who 1s rowmg badly and Jones 1s con-
ceived as bemg possibly the man who is rowing badly, or we
should not ask the question Thus there are not two subjects
to the statement,-' Jones' and 'the man who 1s rowing badly'
-(though we have been thmkmg of both of them before),
because information 1s not required about each separately, but
only about them m a relation to one another As related to
[• In the answers m fact both ss and Jones are stressed, and the mtonat1on
1s different.]
r:z6 STATEMENT, THINKING, AND APPREHENSION
one another and so conceived by us, they are not separate, but
form for us a complex subject of which they are the elements,
and what we want to know about this complex subject is
whether its clements stand 111 a certain relation, whether they
are related m the way that Jones is the bad rower The answer
is given in the sentence ' It is Jones who is rowmg badly' or
• Jones is the man who is rowing badly '. In these sentences
Jones and the man who is rowing badly belong to the subject,
as elements 111 what we were th111kmg about and about which
we required 111formation The word 111 either sentence, the
addition of which to the words belonging to the subject brings
the new mformatlon, is the verb ' is ' , consequently, accordmg
to the present theory, 1t is the prcd1Cc1tive word and ought to
be the only word in the sentence that has the stress. Now it
is actually the only word with llus stress in the given case,
which 1s a confirmation of the theory.
In general then we see that, in a sentence of the form • A is
B ', the c;trrss may fall upon either A, or B, or on the verb ' 1s ',
accordmg as the statement answers the question ' what is it
that 1s B? ', or 'what is A? ', or ' Is A really B? ', and the
analysis mto subJect and predicate is different m each case ;
while there ts no exception to the rule that it 1s the predicate
or predicative v. ord that has the strei.s. It results m general
that, smce the grammatical form does not d1stmgmsh the logical
sub1ect and predicate, we cannot determme these m a given
sentence unless \\C have 1ts context, or know where the stress
falls. The quc,;tion what 1s the subject and what the predicate
in a given scntcnc,c 1s futile and admits of no answer, 1f we have
no help either from context, or from stress, or from special
grammatical form
 V
THE ANALYSIS OF THE STATEMENT OR PROPOSITION
INTO SUBJECT AND PREDICATE

§ 60. IN the traditional theory of the analysis of propos1t1ons

it is always assumed that, m order to discover its subJect and
predicate, a given propos1t1011 must be rrduced to the form A is
B, where A 1s the subject, B the predicate Strictly speaking,
the form 1s supposed to be All X 1s B, or some X is B, or this
(these) X is B,, but these may be all comprehended in the one
form A 1s B, or S 1s P. This docs not agree with the view we
have taken of the rat10nalc of the ordinary defimtion. For
accordmg to this, the subject and predicate would not be fixed
m this manner, S m~ght sometimes be predicate, and P, or
a corresponding noun, subJect. But, apart from the view on
which this cntlClsm depends, the ordinary theory is mcons1stent
with itself on this pomt. For, wlule 1t prov1drs apparently for
only one analysis mto subJect and predicate, 1t is virtually com•
m1tted to a vanat10n m the terms which may be so designated.
When a statement has to be reduced to the form A ts B, there
is more than one way of domg 1t. For mstance, ' Hertford
College faces the Bodle1an ' may be reduced to ' Hertford
College 1s the bu1ldmg m front of the Bodleian ' or ' the Bod-
leian is faced by Hertford College ' ' A exceeds B ' may be
reduced to ' A 1s greater than B ', or ' B 1s less than A '. The
ordmary theory has no means of dec1dmg that only one of the
reductions 1s right, and would have to admit both, though the
definition 1s expressed m language which 1mphes that there 1s
only one predicate and one subject m any given proposition.
But the assumption that every statement must be reduced
to A is B leads to a more serious difficulty. If the reduction is
necessary, the inevitable consequence 1s that forms of statement
that are not of this kind have not a subject and predicate. In
that case we could not say that every proposition had a subject
128 STATEMENT, THINKING, AND APPREHENSION
and predicate, but at most only that the meaning of every
proposition can be expressed in a certain verbal form which has
a subject and predicate, while m the other verbal form this is
not the case The d1stmct1on then will be a verbal one, and
peculiar to a special verbal form. If ' Williams runs well ' must
be reduced to the form 'Williams 1s a good runner' before we
can find a subject and a predicate, then 'Wilhams runs well'
cannot be analysed into subject and predicate This obvious
result is in plain contradiction to the ordmary definition ; for
it would follow that statements hke ' there is a God ', ' 1t 1s
raining', not m the form S 1s P, would not contain the d1s-
tmction of something about which an assertion 1s made and
what 1s asserted about 1t Thus strictly we should have the
absurdity that nothing 1s asserted about anything m an assertion
hke 'W1lhams runs well'. If then reduct10n to the form S 1s
P is held to be necessary, to avoid this absurdity we should
have to abandon the defimt10n of subJect and predicate. Analysis
of pr,)pos1tions would then be merely a sort of grammatical
exercise, applying only to statements not of the verbal form
S 1s P, and cons1stmg in the reduction of them to this form,
1.e. in finding a sentence of this grammatical form expressing
the meanmg of the statement to be analysed The nommative
to the verb would be what 1s called the 'subject', the verb
' 1s ' the copula, and the rest of the sentence (what the subject
is said to be) the predicate Why the aforesaid parts of the
sentence should receive Just these designations of subject and
predicate would be unexpl<1.med The terms might be retained
as a technical nomenclature for parts of the sentence of the
given form , but, on the one hand, 1t would be better then to
use proper grammatical terms and call the subJect simply the
nominative to the verb, and, on the other hand, this is not what
1s meant in the ordmary theory, for the defimtions of subject
and predicate were not mtended to define merely grammatical
distinctions. However, the ordinary and so-called logical analysis
1s nothing but tlus kmd of verbal reduction, whatever else it
may be called. Moreover, 1f we were content to take this view
of 1t and treat it as such, there would remam another question.
It would not be enough to assume that the meanmg of every
propos1t1on admitted of this kind of grammatical expression,
 Subject and P,edicate 129
though it always turns out to be true in any particular case
tried : there ought to be a demonstration that the reduction is
always possible, and some reason given why it should be made.
What is the attitude of the traditional logic to these questions ?
So far from discussing them, it is not even aware of them. The
contrad1ct1on of the ordinary theory to the ordmary definition
is not felt and the universal possibility of the reduction is
assumed without any consciousness that 1t needs proof or
justification.
§ 61. The reason for this ov1::rs1ght is that tlus part of logic
has grown up 111 an entirely uncritical way; 1t dates ma manner
from Aristotle, but 1t would be a great mistake to suppose that
it ongmatcd m any carefully thought-out theory of the nature
of the propos1t10n and the mcanmg 1t expresses. If 1t had been
so, we must have found m Anstotle or in the logic subsequently
developed an important d1scuss10n of the form of assertion,
showing that all propositions might be expressed in the form
A is B, why 1t is important to effect such reduction, and why
the parts of such a form should be called predicate and subject.
But, as will be at once admitted, there is nothing of the kind
either m Aristotle or m those successors of lus who developed
the traditional doctrine This 1s enough to show that the
doctrine did not grow up critically. The fact 1s that 1t originated
m a kmd of accident; 1t so happened that the propositions to
which attention was mamly directed in the begmnings of logic,
and to which the d1stmct10n of predicate, and 'that of which
the predicate is stated', 1 was applied, were of the form A is B.
This was natural, because the mam doctrine of this logic was
the theory of the syllogism, and the given form is the form of
the premisses and conclusion of the syllogism ; even though in
Aristotle's own logic the alternative forms 'B 1s m A' (heaviness
is in lead), or 'B belongs to A',• often occur as its equivalent.
The association thus begun between the distinction of subject
and predicate and propositions of the form A is B was continued
without due reflection in the subsequent development of logic.
All propositions, when logic was mainly the theory of syllogism,
I ""''"l"fOPOUJAWIW, -.,· oli #tlff/'(OpeiTai,

[• lnnyxti Tf A Tel B. ' in ', i .. no,,e,,-1..., ]

2773•1 K
130 STATEMENT, THINKING, AND APPREHENSION
had to be reduced for logical purposes to the form A is B, or
its equivalent, and so probably this got to be looked on as the
natural form of the proposition for logical analysis, while subject
and predicate came to be identified with the grammatical
members with which they coincided in the application of this
logical d1stmction to the form The apphcation itself was in
accordance with what has been supposed above to be the dis-
tinction of subject and predicate · for m the syllogism B 1s C,
A is n, therefore A 1s C, the term A usually represents the
subject about wh1l h informat10n 1s wanted, wlnle C 1s the new
clement m the roncc-ption of A wluch 1s got by the syllogism.
Thus, in the conclus1on, A would rightly be called the subJect
and C the predicate. In tlus particular case, then, of a pro-
position of the type A 1s C, sub3ect and prcd1catP, happen to
be A and C, so that the logical d1stmction happens to coincide
with that of the members of this grammatical form; and, this
bemg the form mainly before the early log1c1ans, the mistake
was made of assuming without reflection that sub3ect and pre•
dicate were just the members of 1t. It was a mistake because
it could not be 3ust1fied by the natural meanmg of the undefined
predicate, 1 nor even by the subsequent ambiguous definition, and
certainly not by what we have supposed to be the rationale of
that definition.
We have now to consider. in more detail first how the analysis
of the statement m which the principal verb is the verb ' to
be ' into the clements d1stmgu1shed by the symbols m the form
A 1s B, whether rnllcd subject and predicate or not, did ongmate,
when through the study of the syllogism this form of proposition
had become the form mainly considered m plulosophy. We
have secondly 2 to consider how Aristotle came to have his
attention so engrosc;ed m the syllogism that he treated the rela-
tion of the terms A and B, \\ h1ch is the form of the syllogistic
premiss and conclusion a relation unfortunate} called in later
!Qg1c the relation of predication), as if 1t were the only one wit
which demonstrative reasonmg had to do, thus practically
making the form of the syllogistic premiss the standard form
of the statement in logic, and contributing to the subsequent
tendency to regard this form as the one to which every state-
l ,w.rq-,opoVl'fVCJf, I § 63,
 Subject and Predicate 131

ment has to be reduced for logical purposes. Our inquiry about

Aristotle has to take this direction, because it 1s not the analysis
and study of syllogistic argument, as such and taken by itself,
which could bring about the result described, but merely the
particular position the theory of the syllogism assumed in
Aristotle's investigation of inference.
§ 62 The syllogism is a species of the general form of argu•
ment m which a relation between two terms 1s inferred from
thelf relation to a thtrd. The kind of relat10n with which the
syllogism is concerned, misnamed predication, is the one, accord•
mg to the Anstotehan theory, with which vahd argument or
demonstrative rcasonmg has to do If Aristotle had begun his
study of valid argument by an exammahon of scientific reason•
mg, such as existed m lus time m geometry (the most natural
and the best course), he might have been led to observe two
essential features of 1t. The first is that while in geometry we
constantly find that two terms A and C arc bemg related to one
another by the discovery of their relation to a third term B,
that relation 1s constantly not that of A's being C or, as 1t 1s
said, C's being predicated of A, but, for instance, such relat10ns
as A's equality to C, or A's bisecting C. That this constantly
happens is obv10us, and afterwards we shall inqmre if it 1s not
what always happens. He might then have seen that the rela•
tion important in geometry is not the so-called one of pre-
d1cat10n ; or at least that this relation 1s not the one with which
geometrical reasoning is always or even principally concerned.
That should have been enough to have prevented Anstotle from
giving the false importance he did to the form A is B, and
from making 1t, as he did m effect, the form in which every
statement ought to be expressed for accurate representation of
the reasoning. The second feature he might have noticed is
that propositions continually occur in geometrical theorems
which arc not naturally expressed m the form A 1s B, and can
only be reduced to it with difficulty, together with the important
fact that the rcasonmg is simpler and plainer without the reduc•
tion. Now the truth seems to be, and this 1s confirmed by other;
features of the Aristotelian logic, that he did not base his theory
of inference on a study of scientific methods, but on the charac•
teristics of the arguments used in the philosophic and rhetorical
K2
132 STATEMENT, THINKING, AND APPREHENSION
debating so much in vogue at the time (dialectic, as it was
called) ; and to understand his procedure we must go back to
Socrates.
The inquiries of Socrates, which had such a powerfu1ly
stimulating effect, were not properly logical or metaphysical
but led to the development of logic and metaphysic by his
pupils and their followers. Ile sought for clear and accurate
thmkmg, mainly in the special sphere of moral action, but the
ngorousncs::. of his method aroused interest 111 clear thmkmg
and sound argument in general, and thus helped to awaken
logical study. Ile wanted to find dcfimtc, ecrtam rules for
action, aspiring to make conduct scientific, and tlus led directly
to the syllogistic form and to the predominance of statements
with the verb 'to be' for their prmc,pal verb. For instance,
he supposed that, in order to determine how to act Justly, we
have to discover what justice is, and to express it m a definition.
When the characteristics of Justice were known we should decide
that a given act is just or not JUSt, according as it has these
characteristics or not. The reasoning then would be : such and
such a kmd of act is just : this act is of that kind · therefore
1t is just. This is an inference m the first figure, of the kind
Aristotle aftcnvards called the scientific syllogJsm. 1 And Aris-
totle himself observes that it was to get such an argument that
Socrates conducted lus search for dcfimtions, ' He was searching
for definitions and he was right to do so, for he was searching
for inference and the defimtion 1s the principle (or ground) of
inferences '. 2 Socratic questions then took the form ',lfhat i!
1t that so µ_!ld.2.Q is.J ' and thus the answers were of the form
A is B. He taught that the main thing to aim at was ' know-
ledge of what tlungs are '. 3 Thus tlus form A is B was in the
discourses of Socrates the mam and normal form, and the
knowledge he mamly sought, the proper understanding of
a particular action, was syllogistic, being the subsumption of
a particular case under a universal defimtlon.
The same kmd of search for definitions is prominent in the
dialogues of Plato, and the activity of debate connected wit~
l nMoyit1pot l,11tr"7J,10PlltOf,
■ ••Ali-p,s ,,"q?fl d
Tl In,,,.
T( lt1r,.
Anstotle, Mel 1078b .23,
ITIIUll"(f\<tflcu -rdP ' w••· dpx,) Iii TDII' """-Ao-y1t1,,Gn, .,.a
• virawrw . .• Tl 1 - ,r, nw iinwr. Xen M,m. iv, 6, 1,
 Subject anrl Prsdlcats 133
the teaching of tile Sophists, whether this received its main
impulse from the influence of Socrates or not, seems to have
been generally concerned with the attempt to define current
moral and political notions. It will be understood then how
prominent and important, in the philosophical debating of the
period, statements with the verb ' to be ' for principal verb
became. It seems probable that Aristotle was affected by this,
and that, when he began to determine the general form of
demonstrative inference, he was influenced by the Socratic ideal
of complete knowledge, which was necessarily syllogistic in form.
And as for the ordinary debating, half philosophical, half rhe-
torical, in which also the syllogism would be the form of
demonstration, when a demonstration was attempted, his strong
interest in tlus favourite intellectual exercise of the time is
proved by the fact that he thought it worth while to write
a laborious and exhaustive treatise, the Topics, on the principles
of discussion, and on the various forms of fallacious arguments
and rhetorical devices that were practised. The existence of
this treatise and of the Rhetoric enables us to understand the
bent of Aristotle's mmd at the beginning of his studies, and
thus to explain what might otherwise be a puzzle, the circum-
stance that he should not have begun his investigation of
inference by examinmg the method of that science which by this
time had been so accurately developed, the science of geometry.
We are able to understand how his thoughts received a mis-
leading direction at the outset, which made him miss the
important lessons which m all probability he would have learned
had he been influenced by mathematics (in which he was well
versed) instead of by plulosophical and rhetorical debate. When
he did come later to consider the method of the mathematical
sciences, it was after he had written his general theory of the
syllogism m the Prior Analytics and was already too committed
to his view of the syllogism. He still treated the relation with
which it was concerned, namely that of A and B m the proposi-
tion A is B, as the fundamental one with which reasoning had
to do and, in the Posterior Analytics, fitted forcibly on to mathe-
matical science, as the instrument of its advance, the form
proper to the Socratic subsumption of the particular under the
universal. Later m his philosophical progress, a particular
134 STATEMENT, THINKING, AND APPREHENSION
investigation in the Metaphysics• led him incidentally to a
remark in which the true character of geometrical procedure is
appreciated. But this led to nothing further and perhaps he
did not realize how subversive it was of the Posterior Analytics.
This will be explained when we come to discuss the method of
geomctry. 1
§63. Our second question is: when attention came to be directed
to the form of the syllog1st1c premiss, how came 1t to be analysed
into three mam elements ; as 1t is when a general symbolizat10n,
such as all A 1s B, is given for 1t ? How did such a sentence
as ' All wrought-iron 1s malleable at a certain temperature '
come to be reduced to three parts, and all its complexity reprc•
sented by two letters and one word or at most two words ? Of
the three parts, one consists m all the words that follow ' 1s ',
grouped together as a umt and represented by one symbol, say
B; another consists of the words prccl'dmg '1s' (except the
word ' all '), smularly grouped as a umt and represented by
one symbol, say A, so that the symbolization of the whole
proposition 1s All A 1s D. We are so used to tlus that we may
not have reflected that 1t reqmres to be accounted for. We arc
not asking whether these symbolic terms should have a certain
technical designation, such as subJect and predicate, but why
JUSt these elements should be d1stingmshed and marked by
symLols. The motive of 1t was not grammatical analysis, which
would not be confined to the form with ' to be ', and would
not lead to symbolic representation of the sentence at all, but
to d1stmrt1ons like 'noun', 'verb', 'ad1cchvc ', always with
so-called concrete, and not symbolic, 1llm,trat1ons. Thus, m
Aristotle lumsclf, in the De Intcrpretatwne, we find what is
properly a logical treatise preceded by a grammatical analysis,
which leads to such d1stmct1ons as noun and verb 2 but not to
the symbolic representation and analysis of the proposition, nor
to the distinction of predicate and subJect. The analysts of the
given form of propos1t1on mto the symbolic type ' All A is B '
1 Part III, ch 3 • iivopa, flijpa.

[• &a.powrn -,.ip d,pluitOlltllll • , , iral llui TOUTO ll'O<OVl'TU "fl"(l'&lU11011rm1, Metaph.

1051• 21-33 In the passage m question Aristotle 1s spealang of the process
of discovery by analysis, and not of proof He would not, bke Wll&on, identify
the two See, however, Ross, Anslotle's Metaphysics, 11, p 268)
 Subject and Predicate 135
would be the immediate and natural result of the first logical
reflection upon the syllogism and probably did not appear
before it.
Syllogisms are of course given at first, not in a symbolic form,
but in ordinary language expressing particular arguments. When
logical reflection began to be exercised for the first time on the
syllogism, 1t would become evident that there was a common
element in the two premisses and that this common element
connected them so as to make a conclusion from them possible.
Again, however complex the common element may be and how•
ever many words there may be corresponding thereto, it is not
the several members of it apart from one another which serve
to connect the prcm1sc;cs and brmg the conclusion, but the
common clement considered as one whole. Thus the common
element would be recogmzcd ac; an inseparable umty for the
purpose of the argument. When accordingly the idea of repre-
senting lhc general form m symbols occurred, the set of words
relating to the common clement would get one symbol. Thus,
m ' all wrought-iron 1s brittle · this is wrought-iron : there•
fore this 1s lmttle ', we should have one symbol, say B, for
all that follows the word ' is ' m the minor premiss, and the same
symbol for all that precedes ' is ' m the maJor premiss, except
' all '. This came to be called the middle term, with some
confusion of word and thing.
Agam, the part of the minor premiss that precedes ' is '
(except ' all ', ' this ', &c.), however complex, wa~ recognized as
a whole and a umt (symbol A), not merely because 1t 1s what
this premiss has in common with the conclusion (again excepting
such words as 'all' and 'this'), though this draws attention to
its unity, but for the more important reason that it is only in
its umty and as an inseparable umty that 1t is connected with
the middle term and, so also, only as an inseparable whole is
1t related to the maJor term m the conclusion. This 1s true, even
when the affirmative syllogism 1s not of the scicnt1fic type and
the minor has the form 'some A 1s B '. Thus, for the minor
premiss, the symbolization 'All (this, some) A 1s B' would be
arrived at : what is referred to by A or B, however manifold,
only coming mto the argument as an inseparable unit. This
would be enough to suggest the same general form for the major
136 STATEMENT, THINKING, AND APPREHENSION
premiss, in which the term B already occurs, and so give it the
form ' All B is C '. This would reduce the conclusion to the
form 'All (this, some) A is C '. Moreover, a comparison of
different syllogisms would easily show that the part following
' is ' in the maJor premiss and conclusion might often be
a single word, so that, according to the principles of symboliza-
tion, the corrccl general form of symbol to include each case
would be a single symbol, such as C.
Whether the steps were Just in this order or not, the essentials
may be put as follows. The first logical reflection on particular
syllogisms would show that, in a syllogism, two somethings
were related to one another (each as a umty and a whole, what-
ever the number of its parts) through their relation to a third
something, also as one whole and only entcrmg into the argu•
rncnt as an inseparable whole, whatever the number of its parts.
The convenience of symbobcally representing a general form
once understood, perhaps through the mfluence of mathematics,
it would follow that, each of these somethmgs bemg thus recog-
nized as a unit, the words corresponding to 1t would be repre-
sented by one symbol or 'term '. This woulcl be natural for
all terms, and m any case necessary, from the point of view of
a symbolic analysis of the argument, for the minor term A and
the middle term B. If these symbols are A, B and C, only the
verb 'is' and the words 'all', 'this', &c., remain in the
premisses and conclus1on. So there would result the general
symbohc form, ' All or this B is C, all or this or some
A 1s B : all or this or some A 1s C '. In this way then might
arise the analysis of the statement ,, 1th • 1s ' as principal verb
into three prmc1pal parts, the verb itself, and the two other
terms as above described.
The question will naturally occur whether this type of sentence,
so common m Socratic d1scuss1on, would not of itself and without
tderence to its place m the syllogism have suggested its own
analysis One cannot say this 1s 1mposs1ble, but it seems unlikely.
In the first place, in a propos1t1on of the form • all ABC is
DEF ' (where we must suppose the symbols to represent definite
words), it 1s true that ABC is treated as an inseparable unity,
for 1t is only that which possesses the property ABC-ness
und1v1ded wluch 1s said to be DEF. There would be a motive,
 Subject and Predicate I37
then, for representing ABC by one symbol, say X, and we might
write • All X is DEF '. But it is not true that it is only
DEF•ncss undivided that always constitutes the quality which
all X ( =all ABC) 1s said to have. It 1s possible that X may
be D, and not be thought of as D only in so far as D 1s a member
of a complex whole DEF (for instance, • all English frogs are
amphibious-quadrupeds-that can swim ',-you can say all
English frogs can swun), and this might be true even 1f DEF
were a definition. Thus, though the quality of bemg DEF 1s
possessed as a unity by each X, theIC is not the same motive
for representing it by one symbol as there is 10 the case of ABC.
This in itself seems by no means dcc1S1ve, especially since, if the
idea of symbolization had already occurred at all, a comparison
with statements havmg only one word after ' ic;' or ' are '
would have suggested one symbol for the complex DEF as the
proper general symbol to cover every particular case. But,
secondly, there is a really strong reason for tlunking the pro·
minence of the given type of propos1t1on m the Socratic method
and m ordmary philosophic debate not at all likely to lead in
itself to the familiar symbolism ' All A is B '. The need for
such symbolism did not exist either in the Socratic search for
definitions, or m the ordmary philosoplucal d1sr.uss1on of notions ;
and 1t could not naturally ongmate m either, because it was not
necessary or even useful for the aims of either. The need would
be first felt when systematic logic began m the recognition and
study of the syllogism, and that this really originated the
symbolic analysis of statements, with ' to be ' for principal
verb, is strongly supported by the fact that there seems to be
no trace of &uch symbolism before the syllog1st1c logic of Aris-
totle. The symbolic analysis of the propos1t1011 does not appear
until the syllogism appears, and that this analys1s 1s not found
m Plato 1s important for the issue before us. One thmg seems
certain: the logical formulation of the &yllog1sm was enough of
itself to necessitate the appearance of tlus symbolism, because
it is Just this symbolic analysis of the given form of sentence
that is necessary to represent the general form of the syllogism
in abstraction from particular syllogisms.
 VI
THE CONFUSION OF PRED~ATION WITH
OBJECTIVE RELATIONS
§ 64. IN the foregoing investigation we have considered only
the origin of the general symbolic form fixed upon for state-
ments with ' is ' and ' are ', and not any designation such
as subject and predicate given to the symbols, or rather given
to the parts of the sentence which the symbols represent. It
will have become evident that the symbolic analysis is quite
independent of the distinction of subject and predicate and does
not arise out of any consideration of what is called predication.
That symbolism would have come mto existence through the
analysis of the syllogism and as its necessary instrument, even
if such a technical d1stmct1on as that of subJect and predicate,
or subject of assertion and what is asserted about it, had never
suggested itself. Nor does the proof of this mdependence rest
merely on what we have said about the syllogism. The result
follows simply from an exammation of what the symbolic terms
really do mean in a proposition.
The symbolic terms, however, came to be designated and
d1stmgu1shed as subject and predicate, and by no other names.
So much so that in the modern form of tlus part of logic the
symbols themselves are given as S and P, which mean subject
and predicate. Now, even if in the form A is B {remembermg
this stands for All, this, some X is B) it happened that A and
B comc1ded always with the subJert and predicate, which 1s
quite untrue, it would be a great nustakc to distinguish them
only by these names, unless the relation between them was not
only coincident but also 1dentical -with the relation of subject
and predicate. But they are so far from bemg identical that
the one relation has nothmg whatever to do with the other.
Modern logic has endorsed the mistake by the above-mentioned
substitution of S and P for such symbols as A and B, and thus
the modern practice, instead of improving the old symbolism,
 P,etlication and Objecti7Je Relations 139
has introduced a most serious error into it. The new form,
being only appropriate if the relation of the given terms is solely
that ef predication, naturally suggests that that is actually the
fact. But there is more than a suggestion, for the identification
is actually left, in an implicit form at least, in the ordinary
doctrine of the syllogism. We have already shown the mistake
of supposing what the traditional logic really assumes, namely,
that the distinction of subject and predicate belongs only to the
form A is B and that it coincides with the distinction of A and
B, but we have not shown the mistake in the identification of
the two d1stmctions We shall consider the latter point and at
the same time return to the former.
§ 65. If we assume the account of subject and predicate which
seemed the most adequate to the reasonable features of the
trad1t1onal defimt10n and procedure, the subject and predicate
are both of them ob1ects of apprehension or opinion, but we
apprehend them or conceive them as bemg what they are in
themselves quite apart from the order of our apprehension of
them, and even apart from their being apµrehended at all.
Subject and predicate then strictly mean objects, containing
nothmg whatever that belongs to our subjective apprehension
or opmion They do not mean, for instance, our ideas or con-
ceptions of the objects; for whatever these vague terms, idea
and conception, may mean, they do represent sometlung sub-
jective and distmct from the objects of which they are said to
be ideas or conceptions. We may put 1t thus. Subject and
predicate mean not the idea or conception of an object, but the
object which is said to be an obJect of the idea or conception.
But, while the things called sub3ect and predicate are objects'
without anythmg that belongs to our apprehension of them or
our mode of conceiving them, the d1stmct1on of them as subject
and predicate 1s entirely founded on our subjective apprehension
of them, or our opinion about them, and on nothmg m their own
nature as apart from the fact that they are apprehended or
conceived. It may be said to be a d1stmction not in them, but
in their relation to our knowledge or op1mon of them, and so
not a relation between what they arc m themselves apart from
their being sometimes apprehended. This is the distinction of
the one (as being apprehended earlier) from the other, as having
r40 STATEMENT, THINKING, AND APPREHENSION
its objective relation to the former apprehended later. If we
define object and objective, simply in reference to an apprehen•
sion, as ' what is apprehended ' apart from the given appre,.
hension of it (so that it would include what, from another point
of view, might be called subJective, such as a feeling; because
a feeling may be an ob1ect of opinion or knowledge), then we
might say shortly that the subJect and predicate are objects;
but the distinction of them as subject and predicate is nothing
in their obJedivc nature and hes only in their relation to our
subjective attitude of apprehension or opmion.
This difference of relallon to our subJect1ve attitude, in virtue
of which we call the one object ' subject ' and the other
'predicate', 1s not expressed m the verbal statement of our
knowledge or opinion about them. There is nothing in the
grammatical form, or verbal form in general, to indicate it. ll
1s only the stress of the spoken wordi, or the context that
ind1catesi~- - -- -- - ------
--fi~t--;;ot only 1s it true thal the dist mcbon of subject and
predicate is not expressed or conveyed by any verbal form, but
the d1stinct1on also docs not coincide w1lh any verbal distinc-
tion. Thus in the typical and standard form of the propos1t1on,
as it appears m the syllog1stic logic, 1t is not true that the
subject must be A and the predicate B. It 1s not even true
that B is necessarily the predicative word or words, or A the
word or words winch belong to and indicate the subJect. Thus
the usual procedure of finding the sub3ect and predicate in the
llt.Ltter of the statement by reducing 1t to the form ' A 1s D ' is
a fall.\cy.
If the d1stinrtion doec; not nece<:!>:mly coincide with the rela-
tion between A and D m the form ' A 1s B ', a fortiori it cannot
be identified with this rcl.1t1on. The same follows from the fact
that there is nothmg in the \ crl>al stc1tcmcnl that indicates the
d1stmct1on , for, 1£ the true relation of A and B were that of
subject and predicate, the "ords in the sentence wlueh represent
them, and their connexion m the sentence, must have shown
this relation.
But the most important reason against the identification is
found in what the statement really means. For it 1s this which
is the ground of the other two reasons ; 1t 1s because the sentence
 P,edication and Objective Relations 141

means what it does that it neither indicates the distinction of

subject and predicate, nor has a form which coincides with it,
nor describes a relation which is identicai with the distinction.
To avoid ambiguity, we must be careful not to confuse what
a statement or sentence means 1 with what, in a restricted and
understood sense of the word ' express ', it may be said to1
express. The statement ' glass is elastic ' would often be said
to be the expression of knowledge or opinion m the person
pronouncing 1t ; but the statement itself docs not mean any•
thing about anybody's knowledge or opm10n and contains no
reference to it. It profes~es to describe an objective fact, that
glass has a ccrtam property, and that is its sole meaning. In j
general, we may say, the statement always purpo1ts to describe
somethmg obJective (m the sense explained), so that the d1stmc•
tion of subJcct and predicate, not bemg an obJcctive fact but
one depending on the merely subjective order of our 1houghts,
cannot possibly be identical with any relation wluch the sentence
describes or means. Thus, m the particular form of statement
before us, 'A 1s B ', A means a certain object and B refers to
a kmd of bcmg or a particular being which A has, and the
relation beh\ccn these is obv10usly JUSt the objective relation
of A's havmg this being to which D refers, or wluch B (m some
cases) denotes Tlus has notlung to do with the subJcctivc
order of our apprehension and so nothing to do with any proper
distinct10n of suhJcct and predicate. The 1dt:ntification, there-
fore, of the relation of A and D with the predic-at10nal relation
seems such an obvious fallacy that we might thmk it could
never really occur. But it docs ; m two forms-m the treatment
of the syllogism and sometimes in the discussion of universals.
§ 66. In the theory of the syllogism it is usual to designate
and d1stmgmsh the terms called major, mmor, and middle by
the relation of subJcct and predicate. We arc told, for instance,
that m the first figure of the syllogism the middle term has to
be the predicate in one of the premisses and the subJect m the
other, and the distinction of the figures of the syllogism is made
to turn on the differcnt positions the middle term has in the
premisses, whether, that is, it is subject or predicate. It may
1 [' means •. Here there 1s a MS 1nd1cation ahowing t.hat a paragraph on
' meaning and expression ' was intended to be inserted ]
142 STATEMENT, THINKING, AND APPREHENSION
be, it is said, subject in both, predicate in both, or subject in
one and predicate in the other, a distinction which is to give
the three figures.• The major term is defined as the predicate
of the conclusion, the m10or term as 1ts subject. This could
only be justified 1f the relation of precbcallon was identical with
that of the terms ; for in the logical analysis of the argument
the terms must be distinguished by what 1s essential to them.
The false 1denttf1cat1on, then, 1s 1mpbc1t m this procedure of
logic. But 1t gets direct and cxphc1t expression when the
syllogism 1s said to deal with the relation of predication; for
this cannot merely mean that the relation 1t deals with happens
to coincide with that of predication. If it were contended that
the description of the terms by help of their supposed predica-
t1onal pos1t1on was merely a convemencc, because they happen
to coincide with subJccts and predicates m the manner alleged,
the answer 1s plam that, 1£ they did, tlus description by the
non-essential is qmte absurd m logic But we must add that
it cannot even serve any purpose of convenience, because 1t
assumes what 1s utterly false, namely, that there is a fixed
position in the sentence for subJect and predicate, and 1t cannot
be convenient to make an utterly false assumption m order to
elucidate the essential character of an argument.
The truth, however, 1s that the ordinary procedure cannot be
defended anyhow as a mere convenient use of one relation to
imhcate another. It does really involve the 1dent1ficat1on of the
relations. What the true rationale of the description by means
of the notions of subject and predicate may be we shall inqmre
presently. The mistake made 1s evident from what has been
said about the meaning of the statement m general and about
the true relation of the terms A and B m the statement of the
form A is B, for this 1s the relation with which the syllogism
has to do. But 1t 1s also directly evident from the obvious
characteristics of the syllogism. The syllogism always deals
with the connexion of the objective facts, in the sense we have
given to 'obJect1ve '.\It states that a fact represented by the
statement called the conclusion is necessitated by two other
facts represented by the premisses. This has nothing to do with
[• The reference 1s to Fowler, Elements of Deducl111e Logic, III, ch. hi,
I 2, p. 86 J
 P,etlication and Objective Relations 143
any order in our subjective apprehension of the elements in the
premisses, or with our subjective attitude to the facts at all, or
with anything grounded upon it;) How then can the ordinary
mode of representing the syllogism and its terms be of real use ?
The reply is, that the description is really given by aid of the
(.hstincbon between verbal forms within the sentence, as it is
presented m the symbolic analysis of the proposition, of which
the principal verb is ' to be ' j and provided it is understood
what clements in the form' arc meant it doesn't matter what
they are called. The words subject and predicate arc only
justified m the theory as technical names for A and B in the
form A is B. The names are utterly wrong, but that does not
matter to the purpose in hand, Just as I may call Smith and
Taylor, Jones and Robinson. That will do for many purposes,
as long as I consistently call Smith Jones, and Taylor Robinson.
This is the simple explanation of the procedure and the reason
· why it seems to distinguish the terms of the syllogism. It all
depends on the above distinction of the verbal or grammatical
clements , the terms subject and predicate arc quite unneces•
sary, and it would be both enough and more accurate to say
that, any proposition bcmg represented by the symbolic analysis
' all (tl11s, some) X 1s Y ', m the first figure of the syllogism
the middle term was the X of one premiss and the Y of the
other · in the second figure it was the Y in both premisses, and
in the third figure the X in both premisses.
It might be suggested that the d1stmcbon coul<l be made both
without misuse of the words subJect and predicate, and without
symbols, by givmg the grammatical description of the members
of the syllogistic form of premiss. But there arc two difficulties.
The grammatical term nearest to the so-called subject is
' nommative case to the verb ' ; that, however, would not be
adequate, because somethmg more explicit than the term
' nommative case ' is required. For the purposes of the syl-
logism, we must have the resolution of a nominative case into
' All, this, or some X '. Again, for the word or words corre•
sponding to Y, in the form 'All (&c.), X 1s Y ', there is no
convenient grammatical term to cover all cases. Y may be
either adjectival, or of noun form, whether an individual name
or a general noun preceded m English by the indefinite article.
l44 STATEMENT, THINKING, AND APPRBHEMSION
The fact is, that in grammar the false logical term ' predicate 1
has usurped the place of a proper grammatical designation of Y.
The distinction of the parts of the sentence made by the
symbolic analysis not only affords the rationale of the ordinary
description of the terms of the syllogism in the phraseology of
subject and predicate, but is what in the end saves it from
confusion. It 1s an error to say, for instance, that in the first
figure of the <,yllogism the middle term 1s subject m one premiss
and predicate m the other In All B io; C, All A is B . therefore
All A 1s C, th<" middle term B is the so-called predicate m the
minor premiss, Lut 1t 1s not the suhJect of the major premiss,
for, according to tins logic, that subJcet 1s not B but all B.
This 1s corrected by the above symbolic representation of the
syllogism, which follows on the marcurate defimtion. For in
that symbolic representation it 1s said that B is the middle
term. It would, therefore, be more accurate to define the terms
solely by use of the symbolism and without any names such as
subject and predicate for the symbols, m the manner already
suggested. If the names subJCCt and predicate are retained, it
would he necessary to say, at least, that whereas the subject
has the form' All (this, some) X ', the middle term 1s the X of
the subject of the major premiss in the first figure. We should
also require a mod1ficat1on in the use of the word ' predicate' ;
for it is not always true that the middle term is the predicate
of the mmor premiss (m the first figure): m 'all mammals are
warm-blooded, the whale is a mammal . therefore the whale is
warm-blooded', the predicate, accordmg to this logic, of the
minor premiss must be a mammal, but a mammal is not the
middle term. We should have now to say that, whereas the pre-
dicate has the form Y, or a Y, the middle term is the Y of
the predicate. Herc, therefore-, we find another inaccuracy
in the ordinary dcscrtptlon, the use of the word predicate in the
attempt to dciinc the distinction of the terms without the use
of the symbols. If the symbols really were to be dispensed
with, the logic we arc considering would be put to it to describe
what we have called the X of the subJect and the Y of the
predicate ; for it has no convenient terminology for this at all.
The fact that such a necessary piece of terminology has not
even been considered shows the looseness of the analysis. The
 P,,lletdion antl Objsctiw Rslations %4~
symbolic analysis being the real clue to the ordinary distinction
of the syllogistic terms and being what is really behind it, it
would have been better to use it directly, in the manner described,
for that would have prevented the confusion we have pointed
out. But in any case it is evident that the mere description
by word-forms in a sentence, however accurate it might be, is
inadmissible from the point of view of logic. ( For the syllogism
being concerned with the relations of objective facts, the dis•
tinction should not be merely by tlie verbal forms in which the
facts arc expressed, but by a description of the nature of the
facts themselves.;
The faultiness of the description by help of the verbal forms
is seen w1thm the ordinary syllogistic theory itself. For in this
theory it 1s often said that m the syllogism two terms arc related
to one another by their relation to a third, which 1s called the
middle term. But now, in the syllog1sm, ' all that is elastic is
resilient, steel 1s elastic, therefore steel is resilient ', steel and
resilience arc the two terms related to one another by means of
their relation to elasticity as the middle term. But according
to the desrnption of the terms as subject and predicate, 'elastic'
would have to be the middle term and ' res1ltent ' the maJor
term. For in the given description 1t is not said, in the case of
'All D is C, all A is B, therefore all A i-, C ', that B-ness is the
middle and C-ness the maJor, but that B 1s the one and C the
other. The truth 1s, that the words subJect and predicate ought
never to appear in the theory of the syllogism ; they are never
wanted, because the syllogism has no concern with them what•
ever. This cnt1cism suggests two questions .-What is the
proper way to characterize the d1stmction between the forms
of language corrcspondmg to X and Y m the form All (some,
this) X 1s Y? and, What is the objective relation of those facts
to which X and Y refer? The second question ha'> already
been answered m general terms, but depends for its fuller treat-
ment upon a discussion of the office of grammatical forms in
the sentence m general, a discussion to which the first question
belongs.
§ 67. The other false 1denbficatJon of the relation of subject
and predicate with an objective relation is one found sometimes
in the language used about the relation of universal and par·
2773•1 L
146 STATEMENT, THINKING, AND APPREHENSION
ticular. Thus it will be said that Plato's problem in his theory
of Ideas was to account for the predication of the universal
(which is one) of many individuals. This is a very unguarded
way of putting the fact,;;. The problem of the umty of the
universal in the plurality of its particulars which Plato had
before him 1s a qucc;t10n about obJcctivc facts, and has nothing
to do ,,1th predication, which 1s groundC'd on something sub-
ject1ve It 1s true that Plato may put the problem in the form,
' why do we call many ind1v1duals by the same name ? ' ; just
as he puts another problem, about the umty of the thing and
the plurality of 1ts attributes, in the form, ' why do we call the
same thing by many names? ' But the 'why' precisely means
' what obJcctive ground 1s there for such language ', otherwise
the question would be trivial and absolutely devoid of interest.
The question 1s naturally put with this reference to language,
because 1t implies that our ordinary language of itself betrays
the difficulty. We arc constantly speaking m a way wh1ch
presupposes a very puzzling relation m the facts. It is no
difficulty m language or assertion, as language or assertion, but
a difficulty about the fact wluch the language implies. Yet 1t
is important to philosophy to note the language, because it 1s
valuable testimony to the reality of what puzzles us that it
1s just our ordmary language which docs imply 1t, and not
merely the language of any special metaphysical investigation,
which might naturally be suspected of artificiality.
Thus, 1t 1s an entire m1sundcrstandmg of the situation to
represent what 1s a problem of fact by terminology proper
to a problem of expression The difficulty mvolvcd 1s not one
of expression or of linguistic form at all , 1t 1s a difficulty that
only exist:. because of wha.t the language means.
The futility of rcprcsentmg the problem of the relation of
umvcrsal and particular as one of prcd1cabon, or as the problem
of predication, 1s revealed at once 1f we demand, as we reasonably
may, that the technical term 'predication' be replaced by what
it means. If predication means simply assertion, 1t would be
absurd to describe the metaphysical difficulty of the mhcrcnce
of the umvcrsal m the particular as a problem of assertion
Secondly, if predication refers to the special relation of certain
members of a certain form of assertion (and we have seen that
 Predication and Objective Relations 147
it ha.s come to this in the kind of logic before us), the verbal
relation is not relevant unless it is the verbal expression of the
relation in question. As a matter of fact, the so-called relation
of predication between A and B in ' A is B ' does not express
the obJectivc relation unless Bis adJectival. So that a reference
to the grammatical form would have to be mtroduced, a con-
sideration which helps to show the futility of the identification
in question But, if the relation of predication, or some species
of it, were the verbal expression of the relation of universal and
particular, it would still be absurd to describe the problem of
the relation of certam objective facts as the problem of their
hnguistic expression. Obviously, the only excuse for describing
Plato's problem as a problem of predication would be that we
are thus giving the essential character of the relation itself.
Thus the description 1s convicted of futility unless it is seriously
meant that the relation of umversal and particular is a relation,
or the relation, of so-called predication
§ 68. It will follow that the designation of the terms A and
B by the names subJect and predicate, and their distmction by
these names only in the symbolic form of statement, all (some,
this) A is B, is erroneous on any reasonable mterpretation of
the word ' predicate '. The standmg principle of the traditional
logic, that the subJect and predicate of a given proposition are
to be found by resolving it mto the grammatical form A is B,
would also be an illusion, and the resolution itself little better
than a grammatical exercise The traditional logi" has neither
attempted to give a proof that every statement docs admit of
such a resolution, nor inqmred mto the sigmficance of this fact.
The traditional procedure is thus not the outcome of any theory,
but has grown up uncritically out of an unguarded practice of
Aristotle's, and 1t 1s a mistake to take it seriously as if 1t repre-
sented some profound doctrme. It 1s connected with the sym•
bohc representation of the proposition, which agam grew out of
the needs of the syllogism. This mistaken dcsignat10n of the
symbolic terms and the ordinary distmct1on of subJect and
predicate withm the form of statement (A 1s B), of which ' to
be ' is the prmc1pal verb, dates indeed from Aristotle, pre-
sumably the inventor of the symbolic analysis itself. In his
investigations of the syllogism this would arise naturally because
L2
148 STATEMENT, THINKING, AND APPREHENSION
in the form A is B, when a proposition in a syllogism (and, as
we have seen, it was to represent the general form of syllogistic
proposition that the symbolism was invented), it would usually
happen that A was the true logical subJect of the reasoning,
while B indicated the true logical predicate. This was extended
by Aristotle in practice to any proposition of the form, whether
in a syllogism or not. But further, the association of the logical
subject with the symbol A in' A 1s B ', i e. with the nommative
case to the verb, while natural in the syllog1sm 1 was also assisted
probably by the fact that when information 1s given about any
object, so that the object as previously conceived is the true
logical subject, the tendency of language is to make the word
which denotes the object, as thus conceived, the nommative
case to the principal verb in the statement. This general but
by no means invariable tendency would as such exercise an
important influence on the first assignment of the terms subject
and predicate to parts of the sentence. However this may be,
the mistake in nomenclature and 1ts origm in the symbolic
analysis seem to date from Aristotle, and the general use of the
form as the mstrument of the discovery of the logical subject
and predicate in any statement, a use which became explicit in
the subsequent development of logic, seems to have originated
in his practice.
To sum up our results. In grammar, we cannot retain the
analysis of the sentence into subject and predicate, for it 1s
inconsistent with any either usual or legitimate interpretation
of subject and predicate.
In logic, the use of the form as the clue to the analysis of
statements is erroneous and has led, in the theory of the pro•
position, to the confusion of expressing an obJect1ve distinction
in a phraseology applicable only to the subJective distinction of
subject and predicate. In the syllog1sm, the terms arc in con·
sequence incorrectly described in the language of subject and
predicate, whereas their relation 1s an objective one to which
these terms do not apply and for wluch the ordinary logic has
no recognized dcs1gnauon.
 VII
CERTAIN OBJECTIVE DISTINCTIONS AND
DOCTRINES OF PREDICATION

§ 69. FROM a consideration of the form All (some, this) A is

B the following questions arise for investigation :-the ground
of the possibility of resolving any statement mto the form A is
B ; the true theory of the so-called ' copula ', or verb of being,
in this form ; and the meaning of the equation, in the mathe•
maticaJ sense, which itself belongs to the general question of the
meaning of the so-called copula.
Out of the general discussion of the distinction of subject and
predicate arise the following questions :-the relation of the
ordinary doctrine of conversion to what has been given as
the true prmciple of distinction of subject and predicate; the
analysis m respect of sub3ect and predicate of forms of expres•
sion other than statement, which involves a criticism of a certain
modern <loclrme of the nature of the interjection ; and the
criticism of another recent doctrine, to the effect that the subject
of every existential judgement is the ultimate reahty,
Some of these subjects require for their treatment an examina·
tion of the meaning of grammatical forms. Before these, agam,
we require to consider certain distinctions of the kind called
metaphysical. The two mvestigations are necessarily connected
with one another ; for since the sentence or statement describes
the nature of objects and not any attitude of ours to the objects
described, in the way of apprehension or opinion, its meaning
1s wholly objective, in the sense we have already given to
objective. That 1s, 1t is about somethmg apprehended, m
the case of knowledge for instance, and not about our appre•
hension of it. The general forms, then, m the language of the
sentence can only mean forms of the objects apprehended, or
the objects about which we thmk; they are forms of being, not
of our thought about being, and so far it is vain to examine the
forms of speech in order to find forms of thought. For even
150 STATEMENT, THINKING, AND APPREHENSION
if it should be contended by any chance that they are really
forms of thought, though mistaken for forms of things, or objects
of thought, we must reply that that doesn't in the least alter
the meaning of these verbal forms ; what these forms are
intended to do is to express characteristics of realities or objects,
not distinctions of thought. Even for the extremest idealistic
view there is an object, whether called thought or not, to be
distinguished always from the apprehension of 1t. And it is to
the forms of this object as such, and not to the forms of our
subjective apprehension of 1t, that the grammatical forms corre-
spond. The traditional logic, being based on an examination
of the form of statement or enunc1at1on, comes upon ' cate-
gories ' or ' conceptions ' which are of the kmd called meta-
physical. If then logic, m general terms, is some study of the
nature of our thinking, as opposed to a study of the nature of
the objects thought about (which seems quite essential to the
conception of logic, whatever differences there may be m its
development), the question ought to arise why these conceptions
should appear in logic at all. Now, whether their appearance
in logic can be Justified or not, the traditional analysis has
proceeded with so little consciousness of the true character of
what it is about that the issue does not even get raised. Yet,
with the sole exception perhaps of the distinction of subject
and predicate, the distinctions arrived at are of the obJectivc
kind and are what arc usually called metaphysical ; and neces-
sarily so. Further, as to this very distinction of subject and
predicate, the attempt to find 1t w1thm the forms of the sentence,
which do not provide for 1t, ha!. produced a confusion of it with
an ob1cctive relation.
This leads to another reason for the discussion of the objective
distinctions. We require to elucidate the true nature of the
relation of subject and predicate and to keep it apart from the
d1stinct10ns \\<1th wluch it tends to be confounded. Modern
logic, which substitutes 'Judgement' (erroneously enough) for
'proposition' and 'concept' for 'term', 1s in fact founded
upon the statement or proposition, because it really depends
upon what the precedmg logic had already got out of the
analysis of the verbal statement. The idea of such ' judgement '
with ' concepts ' as its clements would lead no further, 1f 1t
 P,edication and Objective Relations 151

were not for a study of the linguistic forms in which the sup-
posed Judgement is expressed. But we shall consider later the
so-called Judgement and concepts. We shall begin with some
metaphysical or obJcctive distmctions and follow this discussion
by an account of the doctrine of predication in Aristotle, which
it helps to elucidate. We shall then discuss the general relations
which the grammatical forms of the sentence express or imply.•
§ 70.b There are implied then in our thought certain distinc·
tions between clements in the object of a statement as elements
in the object, which arc therefore not distinctions between the
apprehensions of these ob1ects, and so not subjective m meaning.
These must not be confused with the subjective distinction of
logical subject and predicate and yet tend to be so confused.
Such distinctions arc the distinction of substance and attribute
and a more general one includmg this, which may perhaps best
be formulated as the distinction of sub1ect and attribute.
(In such mqumes as the present we have to keep apart two
different questions. The first, what the real nature of the facts
is to which a given word or notion refers, and the second, what
we exactly mean ourselves, whether our notion is adequate to
the facts or not. If we do not keep them apart, we may get
to doubt the existence of a notion, whereas what we ought to
be doubting 1s its adequacy to the facts, or we may be led to
confuse considerations which belong to the facts with those
which belong to the notion. The want of this precaution is one
of the chief causes of perplexity m such modern questions as,
What 1s hfe ? What is force ? and the hke. There will be
incurable confusion if we do not first ask what it is that we
ourselves exactly mean by the word ' hfo ' which we are using
in our problem. If we do ask the question, we are the more
likely to understand what it 1s we really want, and sometimes
our problem may take a new shape or disappear altogether.
The assumption of the d1stmct1ons before us as objective is
rooted m human thought and, in the first instance, we have
merely to try and find out what they are m our thought ; to
recognize them, not to vindicate them. After that we may
(• Cf. Locke, Essay II, xxxm, § 19
b I have marked later additions to tlus section by <) Sec analysis of
dates of fragments.)
I52 STATEMENT, THINKING, AND APPREHENSION
inquire into their rationale and their relation to the facts with
which they have to do.)
Ordinary and popular thinking, like the philosophies for
instance of Aristotle and Locke, which give expression to it, is
accustomed to regard certain existences or reahties as complete
and independent, others as dependent and existing only in
dependence upon the independent realities.
Thus we thmk of a body as an independent existence, whereas
its movement cannot exi5t independently but only as movement
of the body ; and so for its weight and surface, &c. Such
a supposed independent reality, which we call a thmg, m a special
sense, is m fact a unity of real clements (volume, weight,
shape, &c.) which cannot exist except m that unity, so that
the being of one clement seems to enter, in a way, into the
being of another. Moreover such 'thmgs' stand in relation to
one another, and the relations of a thmg to other things belong
to its bemg.
In philosophic thinkmg then we come to sec that things have
not an mdcpendent bemg but that the bemg of one enters
somehow mto the bemg of another Such reflections may cause
us difficulty as to the meaning and justification of a familiar
phrase, viz. '\\hat a tlung 1s m itself', which implies usually
an mdcpmdent bemg m the thmg, expressly contrasted with
its relations to others. But now every element of reality, how-
ever dependent, whether thing, or element m the being of
a thing, or a relation between such realities, must, smce it is
definite and different from other reahtics, have in some sense
a bcmg of its own, or it could not be distinct from other reahties.
It must be &omethmg which they arc not.
We thus get a 1ust1ficatlon for the phrase ' in itself '. We may
say that what a somethmg 1s, and other somethings are not, is
' what 1t 1s m itself '. The ordmary use of the phrase ' in itself '
is somewhat narrower. Suppose a thing T1 stands in a relation
R to a thmg T2 : standfog m this particular relation belongs
to the being of T 1 and to the bcmg of nothmg else, for nothing
else can stand m this relation accurately understood. Yet we
don't think of this as belonging to what T1 is m itself, or as
a part of what T1 ' is in itself '. In fact, 1f a given something
has a part of its nature not constituted, or supposed not to be
 Predication and Objective Relations 153
constituted, by relation to anything else, 'what it is in itself,
is usually restricted to this, and what it is in relation to any·
thing else is excluded from it. Thus we should say that an
orange was in itself yellow and round, but its being on the table
we should not call ' what it is in itself '. This kind of limitation
cannot be made in the case of a reality whose being is entirely
constituted by relation to something else, e.g. the movement
of a body, for clearly such relations cannot be excluded from
what the given reality (e.g the movement) is m itself. Such
a movement stands in a relation to the movement of another
body, or to another movement of the same body, but 1s not con•
stituted by this relation. Accordingly this relation would be
excluded from what the first movement would be said to be ' in
itself'.
(While things or substances are treated as absolutely inde•
pendent realities, the dependent existences may, in their turn,
have other existences depending on them. The absolute or
relatively mdependent reality, considered in relation to what
depends upon 1t, 1s what 1s called in philosophic thinking ' sub-
ject of attributes '. The main idea which determines the use
of the word 'attribute' seems to be that of something which
' belongs ' to somcthmg else, and is thus dependent on that
somethmg. But the ordmary use of the word attribute seems
to, cover two different things, the distmction between which is
not provided for 1nph1losoph1c language. The one is the depen•
dent existence itself, e g. the pomt of a needle, the other is the
possession of it by its subject, e g. the pointedness of a. needle.
As will be seen 1 when the distmcbon is more fully discussed,
it is the latter which, accordmg to the usage of language, should
be called attribute. The former may be called clement, but for
clearness we might call it ' attribute-element ', since (awkward
as this expression is) it will prevent any doubt as to what is
intended. Such an account of the proper meanmg of attribute
involves difficulties m view of expressions current m plulosophic
language. But a reminder may again be given that we have
to recognize and describe actual facts m thought and language,
not to vindicate them. Moreover it is no reason against the
alleged existence of such mental facts that their implications,
I I 81,
154 STATEMENT, THINKING, AND APPREHENSION
if fully thought out, are fraught with difficulties. Ordinary, as
distinguished from philosophic speech, attests the distinction of
attribute and attribute-element m our thought, and provides for
it without confusion. For the latter it uses the verb ' to have',
e.g. ' a needle has a pomt ', and for the former the verb ' to
be', with the adjective corresponding to the name of the attri•
bute, e.g. 'a needle is pointed'. It is quite artificial, as we
know, to say that a needle has pointedness, and the natural
usage of speech docs not permit such an expression at all.)
The subject as thus d1stmgmshed from the attributes, m either
sense, might be supposed to be 'what the thmg is in itself', or
at least some special part of it, as opposed to an ' attribute•
element ' which 1t has, wh1£h also 1s not something which it is,
e.g. we can say a body has surface but we cannot say the body
is the surface. Yet we fincl, 1f we try to describe what this
subject is, that we can only do 1t m terms of some of the things
which we have been callmg attributes in one sense or the other.
For instance in the case of the spherical shape of a body we do
not, it is true, say that th9 body is its roundness or its round
surface, but we do say .it is round.• (Though then the word
'is' gets used with reference to every attribute, the tendency
is to restrict the bemg of the subject proper to what seems to
be permanent and identical m its temporal existence. Here the
subject agam is undoubtedly conceived as something which
possesses those ' attribute-elements' which are themselves
apparently permanent.
Yet, when reflective thought is turned on such 'subject', we
seem to find that nothing can be said of 1t except that it is
what has these attribute-clements. In consequence the concep•
tion of subject, whether substance or not (though it 1s substance
which is mainly in v1ew) 1 comes to be doubted, challenged as
a metaphysical mystery, and finally treated as an dlusion. Here
two mistakes are made. In the first place, we find that subject
(or substance) is merely omitted and the attributes retained,
sometimes with such a defimtc statement as that the subject
1s only the sum of its attributes. This leaves the difficulty just
where it was, because ' attribute ' is necessarily ' attribute of'
somethmg, and presupposes that of which it ts attribute. It is
[• This, llke the whole section, only holds within a certain group of languages.]
 P,etlication and Objective Relations 155
easily seen in any example that this could not be the sum of
the attributes of which the given attribute was an item, and the
discarded conception of subject (or substance) inevitably re-
appears. Moreover, there 1s the familiar criticism that some-
thing is required besides the attributes (in whichever sense) to
hold them together in a whole. But the form of the criticism
seems unguarded and may lead to a umty external to what has
to be unified.
Secondly, it is a common fault in critical philosophy • to
think that to pronounce a notion an illusion 1s enough to
settle an issue, whereas it is imperative to go on to ask what is
meant by illusion in the given case and how the illusion could
come about. If this 1s done, it sometimes happens that the
true character of what has been rejected may be revealed for
the first time and vindicated, in essentials at least.)
§ 71. There 1s, however, somethmg to correspond to the dis-
tinction of 'subJect' and 'attribute', though imperfectly for-
mulated and understood, as indeed we might expect from the
fact that ordinary language has a special form for it. The
rationale of it seems to be not that the subJect is its attribute-
clements (language avoids any form which would mean that),
nor that 1t 1s the sum of such elements. A reality, whether
a thing or not, may be a umty which unites in itself different
aspects or clements : not something over and above them, which
has them, but their unified existence. They cannot exist except
in this umty with one another, and they, in their unified existence,
constitute the one thmg or one element of reality in general
whether thmg or not.
(The difficulty we raise about the notion of' subject' is really
a difficulty about this umty and we are puzzled merely because
we think of the unity in the abstract.) How a diversity can
form a unity, or how a unity must be the unity of diverse
elements in one whole, depends on the particular mstance and
we understand it in the particular instance. Thus we see that
a volume must have a surface and that a surface can only exist
[• Not, of course, referring to Kant. ' To trace any error to its source
will often throw more bght on the subJect m hand than can be obtained if
we rest satisfied with merely detectmg and refuting 1t '-Whately, Log,c, iv.
4, I I, • -,a, lwr,pov •ilropla Ail0'1r TOIII ffp/,npov d.'6opovp.4nn, lO'-rt, M11r, r ot. IO'TW
il-yr,ooiivTas T~v IEul'O"• Ar, Melaph., 995• 28.]
156 STATEMENT, THINKING, AND APPREHENSION
as the surface of a volume ; we seem also to see exactly what
the nature of their unity is, and that no mysterious something
outside the elements themselves is required to unify them. Such
unity of diversity 1s not merely found in what we call things
but also in what are elements of the existence of thmgs. Thus
a movement unites in itself direction and velocity, and we
understand again exactly how it does unite them.
The umty of 3. mind m its attributes, in whichever sense, and
whether conceived as temporal or not, 1s quite different from that
of a material body, and this again we see by considering particular
mstanccs. 1
Now, what have been called ' attribute-elements ' are the
clements m such a whole, and the ' subJect' of them, as it is
called, 1s this whole itself, as a umty. In this wide sense 'attri•
bute '; as ' attribute-element ', would include relations, which
belong to the thing or other given reality. But this would not
agree with the commoner usage of language. Attribute 1s often
confined to the l'lements of 'what the tlung, &c., is in itself',
taken m the restricted sense explained above, so that relations
would be cxduded from them.
If the subJcct of an attnbute 1s a thing, it 1s called a substance
and the d1shnct1on becomes that of substance and attribute.
(As far as language goes, a mind 1s treated m no way as 1f 1t
were different from a thing. For 1t, and what belongs to it,
there is the same grammatical apparatus as for a substance and
attribute, when the substance 1s a thing. We .find this coming
into recognition in Locke when the mind ts spoken of as thinking
substance.•
The question now arises inasmuch as some attribute-elements
1 A notable example of loose tlunlong about umty in diversity 1s the
modem rcprese11tat1on of the 1nd1vidual as a universal because it 1s a u111ty
Jll the d1vcr,,1ty of its quahbcs, &c Thu, doctnnc, which 1S taken as advanced
metaphysics, is nothmg but a d<.plorablt• confusion, due to a. mere verbal
analogy helped out by the metaphy!,tCian's mchnation to paradox, and
absurdest results may be developed from it The umty of the universal 111
its particulars 1s totally different from the umty of the 1nd1vidual substance
as a. umty of its a.ttnbutes (or attnbute-elcmenti.) The pa.rt1culaIS of a. um-
versa.l are not elements m its umty Whether the universal can be the umty
of any elements at all is touched on later (§§ 78 and 8.2).
[• Essay, 11. 23, 122 ; 1v. 3, § 6 Letters to Sbllmgfieet, p. 293 (Worh,
vul 111, 1824).]
 Predication and Objective Relations :r57
at least are themselves unities of elements, as in the instance of
movement given above, how such a umty of diversity differs
from a thing which is also a unity of elements. The answer
implied in our ordinary conception 1s quite clear. The attribute-
element (it would be said) is always a dependent reahty, existing
only as an element in a thing and not conceivable otherwise :
a thing or substance is a reahty not dependent on any other as
element in the existence of that other , or, as Aristotle would
say, the substance 1s a 'subJect' 1 of which itself there is no
other ' subject '. But then it may seem at first that this
answer cannot be vindicated. For we come to see that things
are not absolutely independent. They enter into relations with
one another, and thereby become clements in a wider reality
which comprehends them. In ordmary hfe we freely recognize
that thmgs or substances are related to one another ; we do
not realize how this affects their independence and we still think
of them as the independent realities on which the relations
themselves depend. It is philosophic reflection which seems to
make this idea untenable.
Yet it is not a mere fallacy. We feel that there is a very
real difference between a thing and what we have called attribute-
clements, and that, somehow or another, the thmg or substance
has a higher degree of independence than the element. This
feeling is justifiable, and precise expression can be found for it.
The attribute-element has no nature of its own apart from its
existence m the umty of the thing to whieh it belongs, and we
can form no idea of it without taking account of the nature of
the thing. On the other hand, the thmg or substance, though
as related to other thmgs it may rightly be held an element in
a wider reality which would be the 'one absolutely independent',
has, nevertheless, a nature of 1ts own, not at all constituted by
1ts standing in relations to other substances, and so not con-
stituted by 1ts bemg an clement in the larger umty to which
these telations conduct. We recover thus the true independence
of the thmg as against the overstatement of its dependence.
Language is faithful to the d1stmct1on. Whereas an attribute
is always an attribute of, a thing, in respect of this independent
nature, has a name which cannot be followed by the preposition
1 Sub1ect (substance) iJKcmdp,-,,011,
1~8 STATEMENT, THINKING, AND APPREHENSION
' of ', and, in respect of its dependence in the way of relations,
has another name, a relative noun which can be followed by
' of ', as ' captain of '. This is in strict accordance with the
nature of a relation between two somethings m the proper sense
of this expression. Each of the somethings has, in the cases to
which this expression applies, a nature not constituted by their.
relation at all, and therefore not at all constituted by being
a member of the system which the given relation necessitates.)
The distinction before us, whether m the general form of
sub1ect and attribute, or in the special one of substance and
attribute, 1s a d1stmctlon in the obJects of our apprehension or
thinkmg and not in the order of our apprehensions or thoughts
about them ; 1t 1s, therefore, wholly different from the distmc-
t1on of logical subject and predicate, with which it sometimes gets
confounded. The subject in this objective sense we may con-
veniently term the ' metaphysical' subject, but 1t 1s best defined
as the correlate of 'attribute' (in either sense of that word}.
The d1stmction of subject from attribute is not only not that
of logical subject from predicate, but a subject of attributes in
a statement 1s not necessarily the logical subject. Thus m 'that
buildmg 1s the Bodleian ', 'that buildmg' 1s a subject of attri-
butes and also appears as such m the statement , but it 1s not
the logical subject of the statement. ( A statement can have
only one logical subject, but 1t may contain several subjects of
attributes) Indeed, every element of reality to which the state-
ment refers may be a subject of attributes, and may sometimes,
though not always, be represented in the statement, at least
imphc1tly, as such Thus in ' A walks fast ', A's walking, as
well as A, is a subject of attributes, and one of these is stated,
viz. its havmg a certain speed. In 'the earth moves', on the
other hand, the earth's movement, though a subject of attributes,
1s not represented as such m the statement.
The relat10n of subject and predicate is, as we have seen, in
a certain sense reciprocal, for what is subject of a predicate in
a statement in one context may be predicate of that predicate,
as subJect, in another, and what is predicate may become sub-
ject. But the relation of a subject to its attributes is not
reciprocal. The subJcct cannot be an attribute of one of its
own attributes.
 P,etlication antl Objective Relations 159
The distinction we have here discussed of subject and attri•
butes and relation concerns only particular existence-particular
thing and particular attribute and particular relation. It is of
this usage that we have been trying to find the rationale. How
far a similar distinction may apply to universals we need not
consider at present.•
§ 72. It is instructive now to turn to a doctrine of Aristotle's
which illustrates the confusion resulting from neglect of this
essential difference, the distinction of subJect and attribute
which is objective, and that of subject and predicate which, in
any reasonable account of the meaning of predication, depends
upon a subjective principle. We find m his writings the dis•
tinction of 'what is predicated' 1 and the {subject) 'of which
this is predicated ' 11 established and used for some time before
reflection upon the distinction appears.
In the Posterior Analytics he has at last occasion to stop to
consider 1t and we shall see the difficulties he gets mto ; diffi-
culties which seem to arise from the fact that he has long been
using the d1stmction, without having attempted to define exactly
what it should mean.
With him substance is subject m the sense m which subJect
as the correlate of attribute is 'substrate '. 8 'Accident','
which includes what he calls property as well as accident, comes
nearest the word attribute, m English, but in the Posterior
Analytics he appears, in one place at least, to limit accident to
that which is not in the ' essence '. 6 The elem1mts of the
essence arc 'the (attributes) predicated, or stated, in the what
a thing is '. 6 In the Metaphysics he defines the 'substrate'
as what cannot itself be predicated, whtle everything else is
predicated of it Thus : ' now the substrate is that of which
the rest are affirmed, while itself is said of no other.' 7 ••• 'not
affirmed of a substrate but the rest affirmed of 1t '. 8 So too in
the Categories.-' further first substances, because they are sub-
1 T3 l«JTf/'YOpo!JµEl'OII, I Ta 1ta6' oli traTff"(OpEITOl, • Tel lnr01tdt,1flfOI'.
1
• Tu 1TVµ/31/Jf/trol. ofJula•
• Ta'" T4i Tl lur, 1ta"7-yopovp1r,a An Po 83& 30.
7 T.l 3' brro1t1lpEvl,,, lt1T1 1t11,6' oli rd llAll.a 11.l-yerm, l1tEi110 ll' aim) Jjf/itln itar' 4ll.ll.ov,
Muaph. 1028b 36.
• µ~ d ' lnroitflµlvou clAAd PB' oi TG 4ll.ll.a, ib. 102g& 8.
(• Yule Ockham, m Prantl, u1, p. 368, note 851.J
I6o STATEMENT, THINKING, AND APPREHENSION
strate to all else and all else is affirmed of them, or inheres in
them, are therefore termed substances par excellence.' 1 • • •
'for of first substance (primary essence) there is no predicating;
for it is said of no substrate '. 2 In the Posterior Analytics 8 he
maintains that the substrate cannot properly stand m the pre•
d1cate, but must be the subject of which predicates are affirmed.
Comparmg 'that white obJeCt 1s a piece of wood' 4 with
• the piece of wood is white', 6 he says that the latter is pre•
dicat1on proper, 8 where white 1s predicate and the substrate 7
is subject, whereas the former, m which the piece of wood, the
substrate, appears as predicate, is either not to be called predica-
tion at all, or else is ' improper ' and ' accidental ' predication. 8
What he says then here, m accordance with the other passages,
is that substrate must properly, i e. in proper predication, be
subject: he doesn't say explicitly that the (logical) subJect of
predication must always be a substrate. Nevertheless, that
would be the natural implication and, if this is not intended,
the language is unguarded. But a passage which follows, if
strictly interpreted, would seem to necessitate such an 1mplica-
tion. 8 He says, apparently, that an accident must be predicated
of a substrate (and cannot itself be a substrate) and cannot be
predicated of an accident. From this 1t would seem that an
accident could not be a subject for, if it were, it must have as
predicate either a substrate, which would be impossible, or an
accident, which, by the passage, is also impossible.
But whether we suppose him to mean that the subject of
predication proper is always a substrate, or only that the sub-
strate must always be a subject and never a predicate, the
doctrine 1s full of confusion and quite untenable.
Whatever definition may be given of predicate and subject-
and Aristotle never gave any clefimtion of them, probably
1 IT1 al wpiimJ, ovala1 llui Tl. Tots ci>.Ao,s 11.waau, lnro«<ia8a1 im2 1ravTa Td 4AAa mTd
ToilTow KaT'lf'tOpEia9111 I) ,,, TaVTair rlva1, &d TOVTo l'aA1aTa ova(111 A4'Yol'T111, Cat 2b 15.
1 ...,,a µ~v -ydp Tijs ,rpjJr71s uvaias oilllfµla laTI ltGTf/YOpla, ltQT ol,llu,os -,dp btro1tfl1'4"ou
Al-,«Ta&, Cat 3" 36 • An Po 83" 1.
' Tel AcU#OI' l1tci'vcl lll'TI {vAor,, • Ti) {vAol' AfU/tul' IO'Tlv
1 &wA&is 1taT,r,opeii,, • Til b1ro1tdµrvoi,, 1b 83• 17
1 d ll,) ll,i lfopo817itra,, lno, Til o/$ro, Al-y11" ltllTf/'Y"Pfi", rl, ll' l1e,(vo,s #fTn, µqllapGn

KaT'lf'tOpEil' I) ltGTf/"fOpEIII l'i" I',) c\rrA&is, mTd au1'fJ•fJ71«ds ll~ mT,r,op,iv, ib 83" 14
• bn1telalo, Bi} Tel ltaT'lf'topovµww 1taT,r,opriuBa, ,M, o; 1t1J"7'Yop<iTa11 c\wA&is dAAd
irat'il n/Afl•fl'11t6s. oflT111-ydp Al ,hrolld(11s dwoll1&1tvvovu,,,, ib, 83" 18,
I'•
 Predication and ObjectiTJe Relations 161

thinking the words tell their own story-' predicating ' is the
-subjective act of statement, and so the distinction of subject
and predicate ts a subjective one, a distinction in statement as
statement, and not a distinction in objects. We cannot, there•
fore, suppose that Aristotle would consciously identify the
objective relation of substance and attribute with this relation
of the parts of a statement to one another, though his language
may tend to such confusion. In fairness he may be interpreted
to mean that while the subjective distinction of subject and
predicate is different from that of subject and attribute (an
objective distinction) the substrate shou]d a]ways be made the
subject of a statement and never the predicate.
But then it is a serious defect to define the objective relation
as he repeatedly does through the subjective one. It is a greater
defect that he mvolves himself m a circular definition. He
de.fines the substrate {'subject') through the notion of predica-
tion on the one hand, and on the other hand he defines predication
by the notion of the • subject '. For in order to distinguish
proper from improper predication, he either defines proper pre•
dication as that m which a substrate 1s the subject, or at least
defines improper predication as that which has a substrate as
its predicate.
li we put the best, or most favourable, interpretation on
Aristotle's doctrine, 1t would come, accurately put, to this :-in
predication proper that which 1s the predicate, or the object
s1gmfied by the predicate words, must stand m a c-ertam objective
relation to that which 1s made the subject of the statement.
~hat is to say, the prcd1cat1onal form itself, in the verbal state-
ment, properly implies that the object which the predicate words
signify must stand m the given objective relation to the object
which is signified by the subject words. ~
To maintam such a doctrme 1t would 'be absolutely necessary
to define subject and predicate, and subject and {say) attribute
apart, and then to show from their defimt10ns that a substrate-
a substance-could only be a subject Aristotle-a sign of his
confusion here-makes no attempt to do tlus.
If he supposed that • prcd1catmg' (or 'statmg ') had an
obvious meaning and required no explanation, and 1£ we take
him to imply this definition-' that about which a statement is
2773•1 M
I62 STATEMENT, THINKING, AND APPREHENSION
made 1s the subject and what is stated about it is the predicate',
-then the doctrine is quite obviously false. For anything may
have a statement made about it, and anything may appear as
what is stated or as an element in what is stated. In default
again of a definition we may take the form of verbal expression
and consider what Aristotle suppose& to be subJect and predicate
in that. In the form all A is B, he takes A for subJect and B for
predicate. But, in this form, a substrate, 1.e. an mdividual sub-
stance, may clearly t.ike the place of either A or B, and that
form of statement would be perfectly normal and correct.
If the predicational form agam as such implies that what
stands m the subject and what stands m the predicate must be
in a certain obJective relation, improper predication would be
that in which the objects denoted by the subject and predicate
did not stand in the relation.
It follows that improper predication \\<ould have to be called
either false, or unmtelhg1ble nonsense, as the verbal form would,
ex hypothesi, contradict the matter. But then neither Aristotle's
formula nor his examples arc right. In the formula. ' not to be
called predicating at all, or else improper predkatmn ', ' not at
all ' might correspond to ' nonsensical ', as the nonsensical might
fairly be said to be no predication at all. But the alternative
'improper' (lit. by accident) is incorrect, for this does not
correspond to ' false '. As an analogy to the manner in which
the contradiction of a verbal form produces falsity or nonsense
we may take an mstance m which the implications of a more
concrete formula are contradicted ; m ' A weighs twice as much
as B ', the words necessitate that m the given verbal form A
and B can only apply to bodies. Hence, 1f we put for A some-
thing not a body, e g. 'the rule of three weighs twice as much
as that chair ', we get a statement which must either be called
false or nonsense; but 'that white object is a piece of wood'
is not nonsense, and is not necessarily false. Aristotle himself
says it may be true-' we may say truly ... that big (thing)
is a piece of wood (or wood) '. 1
Again, according to Anstotle, predication would have to imply
in its very form that the predicate was attribute, in the wider
sense, of the subject-as that on which it depended. In improper
1 lorr1 -,dp fl••i• clAPJl&r •• TC) pl-y,,. lir1wo £•Ao,, eT11111. 1b. 83• 1.
 P1edicatio,i and Objective Relations 163
predication, therefore, where the substrate is made predicate,
the mistake or defect should be that the substrate was then
represented as ' accident ', or attribute in general, of something
else as the subject. But this is not so in the instance which
Aristotle gives of improper predication,' the white thing 1s wood•.
The substance called wood 1s not here represented as treated as
the attribute of anything. What, according to Aristotle, would
be the logical subJect 1s the wlutc (thing), which is not white-
ness 1 but the white object, and the proposition does not mean
that wood is an attribute of this On the contrary, the wood
1s identified in the statement with the substrate which has the
attribute of wluteness It 1s therefore not treated as an attribute
of anytlung else, but as a substrate. This appears even in
Aristotle's own account of it : ' when I say the white thing is
wood, I say that what has the attribute " being white " is
wood '. 2 Thus, on lus own showmg, the statement has not that
kind of fault which alone seems to make !11s d1stmction of proper
and improper predication mtellig1ble. Possibly he intends that,
though m saying the white thmg 1s wood we mean that the
object which 1s wlute 1s also a piece of wood, not that the
white object as such, 1 e. as white, 1s the subJect of which wood 1s
attribute, 3 the form of the expression nevertheless 1s only appro-
priate to, and only naturally has the latter mea01ng.' But, 1£ the
form of express10n meant naturally somethmg so contrary to
what we really mean, we should not use it A!'. a fact it does
naturally mean what he says we mean by 1t, so much so that
sometimes it is the only correct form of expression, and the form
he approves would be entirely incorrect. If the question 1s
' what is that white thmg yonder ? ', a correct form of answer
and the natural one would be, ' That white thing yonder 1s
a piece of wood '. ' That piece of wood is white ' would be
absurdly wrong in form and no one would speak thus.
If Aristotle had asked himself how the form he considers not
true predication could arise, he would doubtless have seen that
so far from being a form which was either not to be called
1 ' the white ' 1n Greek may mean white in the abstract, 1 e whiteness.
,.i., "'(dp TO Afu..a., flPa, ,t,G, £uA,w, T6Tf At'r"' 11-r, ~ ,,..,./JI/J.,.., An,..,ji .1....,
• /l.,..,,
EilAw lt1Tlv.
I cwx cl,s .,.,) inratrflJ,lfl'OP .... £ull.q, Til AfUKOP IITTUI.
' After n,.S,s.,..,,,, then m 82• 8, we should supply tra.iTo& 13., ... a 1-tx•i·
'"1/MUftlP'rOVl'O,
11 2
164 STATEMENT, THINKING, AND APPREHENSION
predication at all or else not ' proper ' but only ' accidental '
predication, it was in a certain context the correct one, while
the form he called ' proper ' was quite inadmissible.
We must now ask what considerations led him to the mistake,
and naturally look for an answer to what he says of the difference
between the two forms of expression. The characteristic which
he finds m the statement 'the white (thmg) is wood' seems to
be this. Wht:rea& the subject of the statement is designated
and defined by a certain ' accident ', it is not as having this
' accident ' that it gets the predicate attached to it. But the
predicating a predicate of a subJect only designated by one of
its accidents, and not as havmg that accident, is not on that
account accidental pred1cat1on. To make 1t so would be down-
right confusion of an objective quality of the object matter with
a quality belonging to our subjective statement. There is
another sense which might possibly be given to the accidental-
ness. It might be said that m proper predication the attribute
or attributes by wluch the subject is designated is what makes
the subject the substrate which 1t is 1, and the substrate to which
the predicate attaches In the 'improper' (it might be said)
this is not so, the ' white ' which designates the subject is not
that aspect of 1t to which the predicate directly attaches , the
predicate (it might be said) is only connected with the attribute
which designates the sub1ect by the fact that they both belong
to the same substrate. This, accordmg to Aristotle's own usage,
might be called an ' accidental ' connexion But this interpreta•
tton can hardly be allowed, because Aristotle would not say that
the only ground of connexion between the two different thmgs
was inherence m the same subJect, but that white attached
directly to wood as to its substrate. They would not be treated
in Aristotle as merely both in the same substrate, but wood
would actually be the substrate of white. And, even if this
interpretation were allowed, we must repeat the previous objec•
tion, that a predication which relates to an accidental connexion
is not on that account itself 'accidental ', for, 1£ so, ' the wood
is white ' ought to be accidental predication, whereas, according
to Aristotle, it is predication proper. And, if we consider the
1 lr,p i,al 1-rl•tTo. Tills seems to be the right way to translate thul
difficult clause
 Predication and Objecti'IJe Relations 165
predication itself, we see that it is in no sense accidental, for
what he calls, the predicate is not attached to the white as white
but to the substrate which has the white, i.e. 1t is attached to
its proper substrate, m the sense of being identified with it.
Thus according to his own view it would have to be ' properly '
predicated. 1
Aristotle's statement that 'the white (thmg) is wood' ought
either not to be called predication at all, or else improper and
accidental predication, could be justified only if the phrase in
question was asserted to be predicating wood of the attribute
white (i.e. of whiteness, really) in the object, that is to say of
the object's whiteness For to this (1 e. to the view that wood
was predicated of white) 1t would rightly be replied either that
the relation of the two in the statement is not to be called
predication at all, which would be a correct alternative ; or
that, if it was to be called pred1cat1on, it would not be predica-
tion in the proper sense, for 1t would only mean that the one,
the predicate (wood), was predicated of a subject of which the
other was predicated, and, as we shall see, this 1s so far incorrect
that white is not predicated at all in the sentence. But then
no one does take such a view (1 e. that wood is predicated of
the attribute whiteness), so that Aristotle's supposed cr1t1cism
would not be true of the form itself, but merely of a wrong
representation of it.
The characteristic, however, which Aristotle appears to find
in the form he considers improper predication and the kind of
difficulty he founds upon 1t would not lead naturally to the
exclusion of the substrate from the predicate position ; it would
only necessitate that the subject should be designated by that
which made it the subject to which the predicate belonged,
i.e. by that aspect of 1t which necessitates the predicate. Indeed,
he has quite overlooked that here he is contradicting his own
principles. Socrates is animal I he would regard as quite a proper
form of predication. But in this sentence an ammal means
a substrate (substance), and a substrate as a substrate.
In his doctrine, m this and the other passage, that substrate
cannot properly be a predicate, he seems obviously affected by
a form of language which we shall have presently to discuss.
166 STATEMENT, THINKING, AND APPREHENSION
It is the rule in the earlier stage of language both m the race
and in the individual, the tendency also in the simpler speech
at all times, to make the thing the noun and the grammatical
subject of the sentence (the nominative to the principal verb),
and to express its attributes m their relation to 1t by verbs, or
adjectives in combination with verbs. And it is not natural to
represent the attributes by noun forms. Thus the realities
which are considered as dependent are associated with these
non-nominal forms, as the ones proper to them. Even in the
advanced stage, when attributes have nouns to denote them,
the adjectival forms, &c., remain peculiar to them, for ' things '
cannot be put into such forms
Now Aristotle treats the grammatical subJect (meaning by
this the nommat1ve to the verb} as the subject of statement
and the rest of the sentence as what is stated of it, 1.e as the
predicate. In this way dependence upon some thmg as sub-
strate (subject) of which 1t was attribute, might be associated
in his mmd with the nature of a predicate Thus in chapter 2
of the Categories we find him confusing ' that which 1s said '
with the attribute or attributive.
Fmally, we may say that when Aristotle defines the substrate
or metaphysical subject as that which 1s subject and never predi-
cate he 1s confusing an objective with a subJect1ve distinction.
§ 73. The confusions m the Organon which we have been
investigating appear in a new dress m a modern doctrine to
the effect that the true sub1ect of every ' existential ' Judgement
is the ' ultimate reality ' 1 We may here examine this view in
order to elucidate further the position we have been led to adopt.
What we have said 1mphes so far that all statements or
propositions are about reahty-1 e. some sort of being. Now
the doctrine before us is not that all propositions are about
reality, but that ' ex1stent1al ' propositions II are about reahty
and, further, have the ' ultimate reality' for their subject. The
'existential' propositions are, in general, singular propositions
as distinguished from universal and hypothetical propositions;
the universal categoricals m this theory bemg erroneously
reduced to hypotheticals.
• F H Bradley's Log,c•, Bk. I, ch 11, f 42 (p So).
• The word actually used 1s • Judgements ', an inaccuracy already cnticJ.Zed.
 Predication and Objective Rel«tions rlY'J
The question as to whether the subject of such propositions
is the ultimate reality cannot be profitably discussed unless we
clearly define to ourselves what we mean by the subject of
a judgement, and what by ' the ultimate reality '. When we
have done this, we shall find that the doctrine and the importance
attached to it depend on a mere confusion of analysis, and that,
instead of being a bit of new metaphysics, it is only an ancient
fallacy in a modern dress.
First as to the meaning of ' ultimate reality'. There seems
here to be a special use of the word 'real'. It means the self-
existent and that, agam, apparently means the complete reality
of the world, as bemg the only self-existent, m distinction from
its partial manifestations. The theory must be tested by possible
meanings of the word 'subject' m relation to 'Judgement' or
to statement.
It need hardly be said that in ' existential Judgements ' the
ultimate reality is not what is sometimes called the grammatical
subject. A second meaning of subject which we have distin-
guished is any element m the statement considered as what we
called ' subject of statement (representing knowledge or opinion)
in general '. This we have seen to be entirely relative, every
element being subJect in turn in this sense. Since this kind of
subject necessarily means some element of reality as related to
other elements, 1t cannot be the ultimate reality, as here under•
stood.
But when, m logic-, we speak of 'the subject of a judgement'
simpliciter, we ought to mean what we have called the' logical'
✓

subject, of which the correlative is ' predicate '. We have seen

that what seems really 1mphed in the traditional definition of
subject and predicate is that some particular aspect of a given
object known to us before the statement is arrived at, or, more
accurately, the object conceived as having that kmd of being,
is the subject, and the predicate is what, in the mental activity
which results in the statement, we come to know or suppose to
be some other aspect of the reality or being of the same object.
The subject then is the conception of the ob3ect with which we
start, the predicate a further determination of it which we arrive
at. Now, the conception of the ultimate reality 1s not the con-
ception we thus start with as logical subject, seeking for some
168 STATEMENT, THINKING, AND APPREHENSION
further determination of it. Even if it could be the starting•
point of a philosophic judgement, that is quite exceptional and
does not belong to ordinary life ; and knowledge or opinion
could not possibly have begun in that way. No one, not even
a philosopher, can start his activity of 'judging' with the
general conception of an ultimate reality. For we cannot get
such a conception till we have thought about particulars and,
as we must start with particular realities, there must be many
• judgements ' in which the logical subject 1s not the ultimate
reality, mdeed the vast majority of our 'judgements'. Nor
would 1t be seriously contended that, nevertheless, the ' ultimate
reality ' ought to be the logical subject. For no one would say
that the only subject about which we ought to seek information
is the ' ultimate reality '.
Again, we d1stmgu1shed what might be called a metaphysical
use of ' subJect '. Any reality (whether considered as indepen·
dent and self-existent or not), as umtmg elements within itself,
is looked on as comparatively independent, m contrast with
those elements considered as dependent upon it. The term here
properly opposed to ' subject ' 1s not ' predicate ' but ' attribute'.
Thus a movement may be said to have swiftness as attribute;
and, though we do not take the movement for an independent
reality, we thmk of the sw1f tness of the movement as less
independent, depending as 1t does upon the movement. Or,
again, we think of a reality not as relatively independent, but
as absolutely so ; and here agam the elements which can be
distinguished m it are considered to be dependent and attributes
of it as their subject : and by such independent realities we
mean what we ordmanly call individual thmgs. 1
Now it is such reahttes which are with Aristotle ultimate
realities, 1.e. self-existent, and his doctrine 8 1s simply that the
'ultimate reahty ', as he conceives 1t, 1s the proper subject for
predication.
But, in modern philosophy, when we ask whether these are
independent ultimate realities, we easily find that they are not,
because they are m relation to one another as members of the
same whole of reality. This commonplace of modern speculation
conducts us to the idea of a total reality, of which the individual
1
For tlus and what follows compare f 70, and p. 200. • f 72.
 Predica#ion and Objecti'IJe Relations 16g
things are themselves constituents, as the only independent and
self-existent reality. This, then, would be the ultimate reality,
since ultimate is equivalent to self-existent. It may be rightly
said that the ultimate reality in this sense, as the total of reality
in its unity, is the true subject, if ' subject ' simply means the
independent and self-existent. But this is a metaphysical dis-
tinction which has nothmg whatever to do with our subjective
act of judgement as such, and to say that this is the true or
ultimate subject of ' Judgement ' 1s a mere verbal confusion of
subject in this metaphysical sense with sub1ect in some logical
sense. The criticism we pass upon it is exactly the same as
that which we passed upon Aristotle's view that, m predication
proper, the subject must bl' the substrate, in the proper sense
of that term. For, obviously, the doctrine before us is in essence
the same as that, the only change bemg due to the advance m
modern metaphysics whereby individual substances have ceased
to be thought of as ultimate independent reahties, since they
are to be conceived as in relation to one another.
 VIII
THE MEANING OF GRAMMATICAL FORMS
§ 74. THE difficulties in the Aristotelian doctrine and its
modern congener ilh1strate the necessity in logic of some con•
saderation of the meanmg of grammatical forms. We shall find
it necessary also to the examination of that analysis of the
proposition on which the syllogistic theory rests, the doctrine
that all statements can be reduced to the form all (some, &c.)
A is B, or S is P. We shall see that the objective distinctions
of subject, attribute and relation are reflected m the grammatical
forms and explain them.
In developed language,• the name of anything whatever is
a noun ; or we may say that a word which denotes or signifies
anything whatever, any element of reality whatever, whether
particular subject or substance, or particular attrabute or rela·
tion, or the universals of these, is a noun.
But the converse is not true. Not every noun is a name.
' Bird ', for instance, is a noun ; but bird does not denote any
given particular bird, though • this bird' does. Nor does it
denote the umversal of bird, for that is denoted by ' b1rdness '.
Such nouns are not names, but the general forms of names ;
a true name being produced by the addition to them of a par•
ticularizing word or phrase, as ' that bird yonder '. The names
of universals, as ' humility ', are true names.
It may be added that a noun as the name of anything denotes
it in general as something 'in itself', and different from other
things. Grammarians sometimes divide nouns b into concrete,
which signify particular things, and abstract, which signify
their attributes. This is very inaccurate and confusing.
Abstract is constantly used of universals, and abstractions
[• Lotze, Logic, J J, f 31.
b The reference was specmlly to Kbhner's Greek Grammar (old edition).
Much of this chapter was, curiously enough, moddied, I think, m the llght
of the early sketch of Grammar m Anstotle, Poetics, ch 20, as elucidated by
Bywater)
 The Meaning of Grammatical Forn,s 17r
we know often stand for ' universals '. But the attributes
of particulars are as particular as the subject to which they
belong, and the distinction of umversal and particular a,ru1lies
as much to attributes as to subjects. This verbal confusion (in
grammar) may have contributed to the error of those who
actually believe the solution of the difficulty about the relation
of the universal and particular is that the universal is an
attribute of the particular.
The other parts of speech are not the names of anything :
they do not, in the proper sense of such terms, ' denote ' or
'signify' any element of reality, the adjective 'heavy', for
instance, does not denote, and is not the name of, the attnbute
of weight or heaviness, nor does 1t denote the subJect of heavi-
ness, for that is denoted by such a noun phrase as • heavy
thing'.
The failure to make this distinction causes confusion in the
traditional treatment of terms in ordinary logic, especially in
the account of connotation. We find adJectives regarded as
names, and the confusion caused is irretrievable. 1
Let m, now consider how language represents the being of
any element of reality with its attributes 'lnd relations, and take
first individual tlungs as substances. We will confine ourselves,
in the first instance, to simple sentences , the more complex
with relative clauses and subordinate verbs can be treated on
the same principle.
The given thmg is represented by its name, a noun in the
nominative case. The fact of its having a given ' attribute•
element ' is expressed either by a verb form : e.g ' this star
twinkles', or by an adjective, corresponding to the attribute
(e.g. heavy to heavmess), together with the verb to be: e.g
'this stone is heavy'.
The adjective here does not denote or signify the attribute.
The proper account of it seems to be that it is the verbal or
grammatical form we use, when we wish to represent the attach-
ment of the ' attribute-element ' as ' attribute-element ' to its
subject.
If the connexion of attribute and subject is to be stated, if
we wish, that is, to state that a certain subject has a certain
• On connotation and denotation .,ee Part II, ch. 18.
I7i STATEMENT, THINKING, AND APPREHENSION
' attribute-elemeot •, the verb is required, but if the connexion
is presupposed and not made a matter of statement, the adjective
is attached to the noun without the verb. In such a sentence,
an attribute of the attribute itself (whether the connexion of
the attribute and subject is expressed by a verb alone or by an
adjective and verb) 1s signified by employing an adverb, which
similarly does not denote, and is not the name of, the attribute
of the attribute.
For the expression of relations there are prepositions, particles,
inflexions, cert.lin forms of noun and adJect1ve, and certain kinds
of verbs. The terms to be related are represented by words
denoting them, which are nouns or noun-phrases,-names in the
strict sense. The relations between such terms are expressed
by means of
(a) adJect1ves, with or without the verb • to be' whether
special adJectives of relation, as • equal', in combina•
hon with inflexions, prepositions, or particles ; or by
ord1nary adJectives (which may be called attributive
m contrast with the relative adjectives) subjected to
a k111d of inflexion, that 1s to say, in the comparative
and superlative degrees, with particles (or in some
languages, mfl.ex1ons of a related term) e.g. 'harder
than', foedior omni crimine. This is where the relation
of the two terms is in respect of some common
attribute indicated by the adjective which is in the
comparative or superlative degree.
(b) nouns of relation such as ' gmde ', • brother', ' friend '.
(r) verbs. These are either• active' verbs; expressing a rela•
tion between two terms, one denoted by the nominative
to the verb, and the other, usually, by the accusative ;
or passive verbs , implying relation to a cause expressed
or understood · e.g. : ' he watereth the hills from
above : the earth 1s filled with the fruit of thy works.'
Both active and passive verbs are accompanied by
inflexions and, sometimes, by prepositions. These two
kinds of verbs might be called relational verbs, in dis•
tmction from those which express the attribute (as
distinguished from a relation to a subject of attributes)
and so might be called attributive verbs. Examples of
 The Meaning of Grammatical Forms x73
various kinds of the latter (attributi~ verbs) are: he
falls down, he reddens, we are quarrelling, he walks
about, he stops short.
(d) Prepositions, with inflexions and with, or without, the
verb 'to be'; with which may be associated those
particles which serve to relate clauses to one another,
vjz, the various conJunctions.
We notice, therefore, that the d1stmction of attributive and
relational applies to nouns, adJectives, and verbs. A relation
between A and B belongs to both, but to each in a different
way (in general), umtmg in itself these two aspects of itself
inseparably. If the relation 1s to be stated in the aspect of it
as belonging to A, A 1s m the nominative case to the verb.
B also appears as a noun and may have an mflex1on of some
case other than the nommative. 1
(We need not here discuss the more complicated sentences of
relation where B appears as the nominative case itself to another
verb.) Just as m the case of the adJecllve, 1£ the relation 1s
presupposed and not stated, the verb 1s not used. ' The brother
of Jones has arrived.' And, as m the case of the attribute, the
name of the relation-its noun-does not appear.•
§ 75. This then is the proper function of nouns, adjectives
(when used), verbs and non-nominal forms in general, in all
stages of language. But expression is not confined to this use
of nouns, ad1ect1ves and other parts of speech.
The attachment of the attribute to the subject may be repre·
sented, for instance, by usmg a noun, the name of the attribute,
for the attribute as well as for the subject, and then expressing
their connexion as a relation between them, according to the
grammatical formula for the relation of things. We say, for
example, her face has great beauty, or his expression is one
of undeniable charm, or to hsten to him ts torture. But this
belongs to advanced thinking, it is not found in the simpler
stages of the language. Moreover, even in the advanced stage,
• Nominative case=noun (or name) case, and seems associated with the
idea of the noun, or name, as representmg the thmg named, in at least relative
independence
[• The examples have m many cases been added by me m this and the
following sections of tb1S chapter.]
174 STATEMENT, THINKING, AND APPREHENSION
it is felt to be artificial, at least in many cases ; the simpler
form is recognized as normal and natural. The usage, then,
with which we started, seems to be the tendency of the earlier
stages of language, and of the simpler speech at all times. In
this usage, if information is wanted either as to what substance
has a given attribute or what attribute belongs to a given sub-
stance, the sentence 1s the same in form ; and though the
attribute may be the logical subject, the substance to which
the attribute belongs is still nominative to the verb, the stress,
however, indicating the logical predicate. This also remains
natural in the later stages. So far there is no need for a noun
as the name of an attribute
Information, however, may be wanted about the attribute' in
itself', m abstraction from its subject, not about the question
to what subject the attribute belongs. For some cases the
primitive form may be used For others a variant of it is
more convenient, where the nommattve to the verb is not the
noun which is the name of the subject of the attribute simply,
but that noun together with the adJecttve corresponding to the
attribute, so that the information about the attribute ts given in
the form of a statement about the subject of the attribute as
having that attribute. Thus 1f we want mformation about an
attribute 'magnamm1ty ', the answer may be in the form 'a
magnanimous action 1s so and so '. But, as the attribute gets
to be considered and studied more and more in abstraction from
its subject, such modes of expression become less convenient and
adequate, especially for the needs of science, and the use of
the name of the attribute as apart from its subject, that is of
nouns denotmg attributes, is naturally developed and we get
substantives like' magnanimity', 'stabibty '.
The same grammatical forms are then used for the attribute-
element (represented by a noun) and its attributes and relations.
The same thing again happens for attributes of attributes.
Similarly the names of relations, e.g. fraternity, monogamy, &c.,
come to be used and enter as nouns into the same grammatical
system. Thus the use of the noun form, as nominative to the
verb, for substances is extended to any subJect of attributes or
relations whatever ; but, throughout the development, it remains
normal to represent the attachment of attributes to a subject,
 Tiu Mefffling of Grammatical Forms r75
whatever the subject may be (whether substance or attribute)
by verbal and adjectival forms in which the name of the attribute
attached, a noun, does not appear ; and the same rule holds
for relations. Thus, if an attribute is to be represented as
a subject of attributes or relations, in developed speech we
employ a noun as nominative to the verb ; but, if that same
attribute 1s to be represented as belonging to its own subject,
it is normal to use the corresponding adjective and not the noun
which is the name of the attribute, and the same holds again
for relations. Thus we say ' monogamy is normal in the Aryan
races ', but ' the Aryan races are normally monogamous ' ;
'portability is a characteristic of good currency', but 'gold
coins are easily portable, in small quantities '.
§ 76. A study of the grammatical forms of the sentence would
show us at once that the verb seems to have a special office
in the way of conveying an assertion. A noun like ' John ' or
' humility ' denotes something, means something, but does not
convey an assertion, nor does an adjective like 'heavy'. An
extreme doctrine of the relation of the verb to assertion is the
familiar modern view that, m the form A is B, the verb ' 1s '
is a sign of predication and merely a sign of predication. This
extreme view ought, indeed, to have been corrected by a con-
sideration of other verbs, for their presence is enough to make
a form of words an assertion and yet no one would say that
that was their only function They obviously do not merely
assert or, more accurately, help to produce an assertion, but
show that some special kind of bemg, some quabty, is c1.sserted
to belong to the subject. If we had to empty the verb ' to be '
as copula of all meaning, except that of being a sign of predica-
tion, we should have to leave this same office alone to the more
' concrete ' verbs, which would obviously be absurd.
We are thus conducted to a double aspect in the use and
meaning of the verb. It is quite right to suppose the verb has
a special function with regard to assertion which other words
have not, but it would be wrong to base this (as perhaps is
sometimes done) on the fact that there is no statement without
a verb, for it is equally true that there can be no statement
without a noun expressed or implied. The truth is rather that
the verb is the only one of the general word-forms which has
176 STATEMENT, TkINKING, AND APPREHENSION
an inflexion such that, when it is combined with_ ot~er word·
formsJ an assertion is expressed. This itself implies that it is
not the verb in general whJch serves this purpose, but only
a particular inflexion of it, viz. the indicative mood. Its special
function, then, does not lie in this, that there is no statement
without it, but m the fact that a certain inflexion of it is
necessary to indicate a statement, and that there is no other
word-form wluch ha& a special inflexion for this purpose.
We are thus led naturally to make certain distmcbons in the
word-form& of the sentence. Instead of the utterly confusing
distinction of denotation from connotation, it ought rather to
be observed that the words m a sentence are divided mto those
which are denotative and those which are not Nouns, which are
really names, belong to the denotative. {The general forms 1 of
names hke ' bird ' are not strictly denotative, but are the general
forms of denotative words.) The non-denotative can be elements
in a denotative expression hke 'heavy' in' heavy guns'. This
applies to adverbs, adJectives and verbs. Thus ' the rain which
fell heavily yesterday', as a whole, is a denotative expression
(with non-denotative words as elements) and a noun-phrase. In
general, denotative words, whtle denoting or meaning something,
do not assert anything. There must be word-forms m the
sentence which either themselves assert, or convey an assertion,
or else do this in combination with denotative words. In either
case we may call them assertive word-forms. It is the second
which is the true alternative, and the assertive word-form which,
in combination with other word-forms whJch are not assertive,
indicates that there 1s an assertion, is the verb in the indicative
mood. We have, then, the important distinction of' assertive'
from ' denotative' word-forms, as well as the wider distinction
of ' denotative ' from ' non-denotative '. 1
To resume then the double aspect of the verb. The verb in
the indicative mood is the assertive word-form in the sense
explained But it 1s not merely assertive. That is to say it
doesn't merely signify that there 1s an assertion, the nature of
what 1s asserted berng conveyed only by words other than the
verb {denotative words, or denotative words in combination
with non-denotative), but 1t serves to indicate a part of the
I I 74 • Cf Part II, ch xvu.i..
 T'l,e Mean,ng of G,ammldical Fo,ms 171
nature of what is asserted, the other part being supplied by
the other word-forms denotative and non-denotative. Thus the
verb • to run', in all its forms, involves the idea of the attribute
of running and, in the indicative mood, when combined with
the nominative case is not merely a sign of assertion but signifies
also that the attribute of runnmg is asserted of what is denoted
by the nominative case.
Consider then the aspect of the verb as expressing part of the
nature of what is asserted. Let us first exclude the verb ' to
be ' and consider the others. What are the characteristics of
the verb ? The most obvious is that it not only indicates
a certain part of the bemg of an object, such as running, but
(by means of its tense inflexions) the time also at which the
object, which is subject of this attribute, possesses it. So
Aristotle in the De lnterpretatione distinguishes the verb from
the noun as ' indicatmg, besides, time '. 1 Indeed the two charac•
teristics of the verb which Aristotle detects are that it implies
time and that it is a sign of that which is predicated of some-
thing else. Not a sign, observe, that it 1s predicatt:d, for Aristotle
does not say th1s1 but ' a sign of thmgs which are of the kind
predicated about the substrate', i e. of what is attributed to
the substrate but is not itself substrate. Aristotle does not
recognize clearly• the assertive symbolism m the verb
But now there are other non-denotative words, v1z. adjectives
like ' round ', which imply a particular kind of attribute (round-
ness) and the attachment of an attribute-element (c uved surface)
to a subject, without, strictly speakmg, denotmg either, and
which not having themselves any symbol of assertion serve, in
combmation with other words, to assert the attachment of an
attribute to a sub1ect In general these words imply a given
attribute-element, but without any indication of 1ts connexion
with time, the t1me1 that is, at which 1t may belong to a subject.
1 'Pijµa 114 ,.,..,., .,.,} "f'Oll'"'l,.,."'o" xp/,•011, oli ,,,pas ol,3~11 .,..,,.,.i,,.. )(..,pt,, 1<111 ,.,..,.,,, dtl
1'0111 tta8' I..Jpov llf"fO/AfroN .,..,/AfUW, • • olu11 m lftt.8' l,wo1mµlro11 I) iv b,ro-p.~
D, Int 1611 6-11
--------
(• He appears to do so exphc1tly. Awd ,u11 oil11 m9' lt1.11Td Af-,6p.tva. rd ~fp.a'l'fl
6116p.a1'4 in1 ml O'l/p.aJ,,,11< .,., ••• dM' ti 1,,,.,,, ,) µi/ 06""' "'1Jllll1111• olllli "/G/' .,.;; e11111t I)
1'0 "" .r- '"' "i'6" ,.,. ,., 1'0U W'fl4"f/J41'0I, • • awo ,,i,, .,a,, oilal11 , ....,, ..""".,.,,,,.,,., &~
tni,,f,.,.,., .,...,,. ,,,, &..11 m "'"I'"''""""' """ 1.,..,., 11oijll'1U De Int. 16b 19,)
N
178 STATEMENT, THINKING, AND APPREHENSION
Unlike the verbs we are considering they do not, when com·
bined with a nominative case, express any assertion but
become only a part of a denotative phrase (say, the heavy gun
will· follow the first-tine transport), that is, a phrase which
denotes a subject as possessing the attribute-element. Thus we
say' heavy gun' instead of asserting that the gun is heavy.
Do the attributes implied by these words differ m kind from
those implkd by verbs, or is the difference merely that in the
one case in the corresponding words there is an indication of
the time of lhc existence of the attribute and in the other not ;
so that, conceivably, a given attribute might have corresponding
to it both linguistic forms ? The answer seems to be that there
is a difference in kind. The verb (excluding ex hypothesi the
verb ' to be ') in its objective reference, as distinguished from
the subjective act of assertion characteristic of it, has also the
characteristic of corresponding to an important objective distinc•
tion, to what would usually be called a metaphysical category.
The thing or substance, besides remaining identical with itself
through time, has also a changing existence m time. There is
no such thing as mere change or absolute change. Events or
time processes are only changes, as changes of something per•
manent, for instance, of a moving body, which has them. As
a mere series or succession (if such a thing were possible) they
would not be changes, or phases of a change ; they can only
be that as a series belonging to something which is not the
series or a part of it ; something which has the series, some•
thing to which the members of the series belong, and by reference
to which alone they are elements of a change or changing
process.
Now it is this something, subject or substance, which has
them, which 1s properly said to change. The subject of the verb
• to change ' cannot be the change itself, nor any phase of it.
Thus it is actually true that only the identical can change and
that there is no change unless there is an identity. Now, that
which is identical in its changes, the changing substance, is not
conceived as a mere identity but as having a definite character
consisting of definite attributes or qualities, which remain per·
manent or identical in the change of the substance.
Besides this, there are attributes which are relatively per·
 The Metimng of Gt-aHfffff.dical Forms 179
manent, states of the substance which remain identical through
a certain time and don't change with the time (or at least are
so thought of), but which are not absolutely permanent attri-
butes of the being of the substance, as they may within its
existence have a beginning and an end. Further, there are
attributes which come into existence in the process of change
(as, if a thing is stretched and gets a different shape, or if it
gets a different colour or is in another place). Nevertheless
these are not themselves proresses of change. Their existence
may even be momentary only, but still they are themselves not
processes of <'hange m the substance, processes which the subject
is going through ; they differ m fact from what we commonly
regard as its activities or passivities (though we arc not con-
cerned here with the meaning or justification of these epithets).
All these attributes, the first kind, those conceived as absolutely
permanent, the second kmd, those conceived as relatively per-
manent, and the third kind, those sigmfying short or momentary
existence, have corresponding to them adJecbval forms without
the time indication of the verb.
The verb is concerned with the temporal process as a process,
not with its isolated phases, such as are the third kind of
attribute, hut with the process itself. With the process, how·
ever, not m abstraction from the subject of it, but only and
strictly as belonging to the subject, the subject bemg dis-
tinguished from the process as an identity {quite strictly agam)
in the time of the process, and as that which conducts the pro•
cess or goes through the process. Both of these latter expres•
sions may be comprised under the more general words ' sustain '
or ' maintain '.
The subject then 'sustains' the process, and its identity as
so doing is represented by one word which is a noun in the
nominative case. Thus the verb ' to run ', m the indicative, is
not concerned with running in the abstract but with some one's
conductmg the process of runmng-' he runs •, ' he ran ' ; the
verb 'to fall', not w1th fallmg in the abstract but w1th some
thing's falling or going through the process of falling. The verb
is used with reference to the fact that some thing is the subject
of a certain process in time, and in its tenses has a mark of the
time of the process. Thus, apart from its assertive element, it
N2
18o STATEMENT, THINKING, AND APPREHENSION
differs from the above-mentioned attributes in its objective
reference. Its full use is in combination with the nominative
to represent the assertion that somethmg denoted by the nomi•
native is subject and identical subject of a certain process in
time.
But there are also verbs which, though they refer to the
temporal existence of a subject which is identical in the various
tJmes of ib existence,· seem concerned with a temporal existence
which 1s not really a process, inasmuch as it is not a changing
state but remains constant for some period of time. Such verbs
are for example ' stand ' ' sit ' ' gaze ' ' rest ' ' press ' (when
the' pressure maintains
' an' equilibrium).
, '
' The explanation seems
to be that these states are really conceived of as something
going on, and so m effect as processes, in the following way.
The given state is on the one hand of a transitory kind ; one
which might change at any moment and is entered on by
a changing process which the subject goes through ; on the
other hand, it is conceived as being maintamed or sustained
from moment to moment by the subject to which 1t attaches,
whether this maintenance 1s thought of as self-originated by
this subject or as due to external necessitation. Thus the con-
dition entered upon and maintained, or sustained, unchangmg
through a certam time is practically regarded as a process going
on through the time, a process of which the phases are exactly
ahke. We must of course distinguish from this a metaphorical
use. For instance, two straight lines are said to ' cut ' one
another, where there 1s no process in any sense.
Finally, the verb may appear to refer to something merely
instantaneous. This, nevertheless, is a phase in the temporal
process of the subJect's existence, and the verb refers to it not
as conceived m isolation, and without reference to the subject's
acqumng the attribute-elements in time, but as a phase in
a process sustamed by the subject, and therefore as so sustained
itself. The difference here intended between the isolated phase,
not expressed by a verb, and the phase as belonging to a process
sustained by the subject may be illustrated as follows. A sub-
stance may be changmg colour. Suppose at some moment 1t
becomes red. The attachment of the red colour to it, in abstrac-
tion from the process the subject is sustaining, is expressed by
 TAe Meaning of Gt-ammatical Forms 181

help of the adjective red and so we can say of it that it is red

at a given moment. On the other hand, for the subJect's having
the attribute, as a phase in a process which the subject sustains,
we have the verb-form' to redden', as when we say, he reddened
with anger.
§ 77. We come now to the verb 'to be•.• The adjectives
corresponding to the three kinds of attributes which we have
distinguished from those implied by verbs, are, in order to
express the assertion that the subject of them is what the
nominative case denotes, connected verbally with that nomina-
tive case by the verb ' to be ' in the indicative mood. But this
verb must not on that account be confused with a mere mark
of connexion. It is not a mere 'copula', but it represents the
general form of what we may perhaps conveniently call ' attri-
butive being' The attributes, as we have seen, are particular
forms of the being of the substance or subject in general, and
the verb ' to be', in its ordmary and normal construction, always
stands for the generic form of this attributive bemg. Thus it
is only to be used along with an expression of how it is differ•
entiated m the particular case, that is, what specific form it takes.
Thus, m normal usage, 1t never expresses or asserts the mere
existence of the subJect. That is why we feel an artificiality
about such a sentence as • crocodiles are'. The words '1s' or
' are ' should never c;tand alone, but should bf' followed by the
expression of the particular kmd of attributive being intendf:d.
Accordingly, m ordinary speech, if we hear sor11e one makmg
a statement with the verb ' is ', we expect him to go on and not
to end with the verb. The mere existence should be expressed
by a verb hke ' exist ', as ' crocodiles exist '. Really, to say
' crocodiles are ' 1s a mistake parallel to saying ' crocodiles exist
reptiles '. In fact, the being represented by the verb ' to be '
is not mere existence but ' being so and so '. 1 The verb ' to
' This JS recognized by Anstotle m so far as he says that .r..a, 1n general
.,,,.,.,,,,,,,,a,.,lff tnl•61at" TIN, De Inl 16b 24 [§ 76, note al But he 1s not engaged
m eludlcatmg this meaning of .1,,..,, and what he says 1s merely mc1dental
to h1s mamtaIDlilg the doctnne that Blngle words do not, taken by themselves,
mean that anything exists Elsewhere he usually distmgulShes between ,r,..,
[• There are, of course, even m Engbsh many other verbs which act as
copulas, e g •tum•. •become', •seem•. 'look•. Sweet. op. cit, § 263. In
Hebrew, the grammarians say that there lS no true copula)
 182 STATEMENT, THINKING, AND APPREHENSION
be' has its tenses and refers to the being of the subject as in
time and as having the aforesaid different attributive forms
in time. But, as we have seen, the attributes with which it is
used are those which are considered either as not processes in time
nor phases of such process, or else phases of process as abstracted
from the process, that is, abstracted from their happening.
In the case of the present tense of the verb 'to be', there is
an extended use beyond the strict momentary meaning of the
present. This happens when it is used of those attributes which
(whether rightly or not, doesn't matter for our present purpose)
are supposed to belong to that bemg of the thing which is
identical m the various times of its existence. It is true that
such attributes would exist at any moment which may be
present ; but, nevertheless, the fact is that, when we speak of
those supposed constituents of the identical bcmg of the subject
and say of this subJcct that it is so and so (more especially
when we are defining it), we don't really confine this 'present
tense ' to the actual ' now ', and we most certainly do not
merely mean that those attributes exist now; we are indeed
referring to their existence as 1dcnt1cal m the various ' nows '.
Nor, on the other hand, do we at all naturally or normally look
on their existence as really something maintained in time; that
is as really a process, such as has already been referred to,
wherein each temporal phase 1s like any other, a succession of
homogeneous events. We mean something that does not happen
at all ; we are referring to its existence as identical m various
times, and the ' is ' does not refer to any of the different
momentary ' nows '. The same consideration applies to other
attributes conceived not as momentary but as identical in
various times, whether parts of the permanent being of their
subject or not. We shall return to this matter under the head
of Modality. 1 It is by the asscrtoric inflexion of the verb ' to
be', by its indicative mood, combined with the nominative and
the adjective, that the assertion of the particular bemg of the
subject is represented in language.
hA&. (to be without qualdicatlon) wluch 1s e,astence (properly) of a substance
and 1Trat .,., (to be 1D a particular way) The latter 1s really the proper
formula tor the 111hol1 use of ,7.,a,, and m general the • absolute • use 11 as
artificial in Greek as it 11 m Enghsh. 1 § 93.
 The Mea•ng of Gttllll#llltical Fo,ms 183
So far we have been trying to describe the facts; we have
now to look for their explanation. We are so accustomed to
the attachment of the assertive symbolism to the verb, that we
may not at first realize how striking it is ; but we become aware
of this as soon as the idea of explaining the fact occurs to us,
for the question proves no easy one. It is necessary to attempt
first, as has just been done, an analysis of the objective reference
of the verb-form, in order to see, if possible, whether it contains
any probable grounc;l for the combination with it of the assertive
function. Perhaps the true reason bes hidden in some primitive
stage of the development of language, but 1t may depend on
general principles d1scermblc in the nature of language as such,
and, if so, the following explanation may be suggested.
We naturally look to what may be common to all verbs, but
we have seen that the verb ' to be' differs so much from all
other verbs that 1t stands m a class by itself, which threatens
to make the inquiry harder. An obvious characteristic of all
verbs is the mdicat1on of time ; yet it is not easy to see how
this could be any reason for addmg the assertive function, and,
as we sh,:Lll notice presently, the indication of time by tense 1s
not confined to the strict verb form. Agam, m the case of the
verb 'to be', we have observed that special use of the present
tense, which m a way seems to abstract the distinction of past,
present and future altogether.•
Let us begin with what seems likely to be the most difficult
case, that of the verh ' to be ', for the more diffi ult case often
brings the essential to light. Perhaps the clue may be this.
An assertion corresponds to a doubt or a question ; 1f there
were neither of these, there would be no need for an assertion.
It seems therefore reasonable to expect that, if there be a word·
form in the sentence which represents the part which is really
open to doubt, it would be to this form that the mark of asser•
tion would be attached. In a sentence of the form ' all lead is
heavy ', or ' lead 1s heavy ', there is, in ordmary cases, no doubt
of the existence of the object denoted by the nommative, nor
of the kind of attribute which the adJective indicates-heavi-
ness. For, observe, it is the universal, as heaviness in general,
[• Sweet (I 28g) calls this the neutral present. He compares the neutral
anomic perfect in Latm.]
184 STATE KENT, THINKING, AND APPREHENSION
which the adjective taken alone implies, and not the special
heaviness of a given substance. The existence then of neither
of these need be ass<'rted, and accordingly we should not
expect the assertive mark to be attached to them. What is
doubtful, or matter of question, is the relation between the
subject and the attribute-1.e. whether the subject has or has
not the attribute-element. This is always so, whether the object
denoted by the nominative, or the attribute implied by the
adjective, 1s the logical sub,ect, and in either case the original
and normal tendency of language is to use the same grammatical
form.
Now the word in the sentence which corresponds to this
relation (the relation the existence of which is a matter of doubt)
is the verb ' to be '. For that of which the existence would be
doubtful in the given example would be not lead nor heaviness
but lead's relation to the attribute of heaviness, that is, lead's
having the form of being which is heaviness; or, more pre-
cisely, that its• attributive bemg ',its' being somewhat' should
have this form It 1s the verb ' to be ', then, as properly meaning
not 'to exist' but 'to be somethmg ', which, when combined
with the nom10ative case, corresponds, 10 the manner explained,
not to the subject nor to the attribute, but to the subject's
being subject of the attribute. The verb serves as a copula,
with the qualification already given, because it does refer, not
to mere be10g or existence, but to this attributive being, thus
m itself involving the connexion.
This verb then corresponds to the relation, a doubt or question
about which 1s presupposed 10 making the assertion, and thus
1t is the word to which the symbol of assertion might naturally
be attached. This has nothmg to do with the time indication
common to all verbs. We shall find that the same principle of
explanation would suit other verbs also, and depends on a feature
which they have in common with the verb ' to be '. Where
this verb is used in combination with the adJective, the attribute
and its relation to the being of the subJect are indicated by two
different words, the adJective and the verb ' to be ', and it is
to the word which refers to the ' belonging to the subject', we
may repeat, that the assertive symbolism is given: If, then,
a word combined both these references within itself, 1t would
 The Meaning of Grammatical Fo,ms 185
be the one to which upon the same principle the symbol of
assertion would naturally be attached.
Now, it is characteristic of these other verbs, active, passive
and deponent, that they do combme the reference to a certain
kind of attribute-element with the reference to its being an
attribute of the subject which the nominative case denotes.
This is obvious when the verb stands alone after its nominative
and is what we called m a special sense an attributive verb ;
the grammatical form is then what 1s called deponent or neuter.
For example 'he stops short', 'he walks'.
But this holds also of the relational verbs which do not stand
alone ; of active verbs followed by an obJect, and of passive
verbs, where the noun denoting the active cause 1s expressed
or imphed. It holds also of verbs in general which are parts of
the expression of a relation. They are always the only words
in the sentence which refer to the possession of an attribute-
element (taking this word in its widest sense in which it includes
relation) by its subject. This possession bemg the doubtful or
questioned part, and the reference to 1t characteristic of every
verb, the mark of assertion would naturally belong to the verb.
If this 1s so, the reason of the assertonc function of the verb
would lie m only one of the characteristics which have been
enumerated above, and would have nothing to do, for instance,
with the temporal process as such, conceived actively or pas•
sively. The temporal reference indeed 1s not confined to the
inflexions of the verb proper,• 1t 1s found m thL adJcctival form,
for the participle 1s simply an adjective with tenses. There
seems no reason why every adjective should not have a tense-
inflexion , on the other hand, there is an obvious economy in
the actual plan which the development of speech has followed,
confining in general the tense inflexions to the one word-form
in the sentence, the verb m fact which 1s used to express the
connexion of the attribute-element, to which the adjective
relates, with its subject. If the prmc1ple had been to express
the time in which an attnbute exists or existed by a tense
inflexion in the adjective, the ordinary verb would still have
[• Both noun and adJecbve verbals keep the distinctions of tense and voice
1n Enghsh, 'I hope you are conung ', 'I expect you to ha.ve finiahed •,
• I remember ha.vmg seen bun', 'I heard the dog being called•.]
186 STATEMENT, THINKING, AND APPREHENSION
had the tense inflexion in virtue of what it has in common with
the ordinary adjective.
There is an idiomatic use of the present tense which must
not be confused with the idiom of the present in the verb ' to
be '. 1 When we say 'John plays the violin', we do 1\0t mean
that he is playing now, nor do we mean a playing which, like
the player, remains identical in various times. It is our idiomatic
way of expressing the fact that a given kmd of temporal attri-
bute, whether process or event, is found normally in the subject,
without specifying the time of its existence m that subject. This
again implies the capacity for such temporal existence in the
subject. Now that 1s something which is identical in various
times, itself neither process nor event. Often it is this capacity
which is mainly in view, and accordingly the use of the verb
' to be' in the present is not only possible but seems more
accurate. Thus we say ' John is a runner' or ' John is an angler'
instead of 'John runs' or 'John fishes'. Sometimes, again,
the mam intention is to describe a quality which the attribute•
element has when it appears in a given subJect, without reference
to the time. Then the present tense is used with the adverb
to express this quality, e.g. 'John runs well', and here again
the more accurate form seems to be 'John is a good runner',
or 'John's runnmg is good•. On the other hand, the 'tense-
less ' use of the present may be found in any verb when its
nommative denotes something conceived as having the ' time-
less ' kind of being which we shall discuss later. 11
With reference to the attributive meaning assigned to the
verb ' to be', there are three points which may be noticed.
The common idiom for expressing mere existence by help of the
verb ' to be ', is really an illustration of this attributive meaning.
When we say ' there are such things as crocodiles ' • it is Just
because ' are ' does not stand by itself for mere existence, that
in this idiom it is not used by itself but combined with the
adverb of place ' there ', and by a kind of circumlocution, merely
to exist is expressed by ' to be in some place or other '. Being
here or there is really 'attributive being' of the subject. We
I pp 182-3 I I 93•
c• Cf. s..,., "'"'"'" man,s.1
 The Meaning of G1-a,nmatiotll Fo,,ns 187
have an apt illustration in the German word for existence which
is dasein (being there). Secondly, the verb 'to be• is some-
times followed by an expression (a general noun} which covers
the whole being of the nominative as a subject of attributes
(what Aristotle calls ' kind ' and 'species') and so goes beyond
the partial bemg constituted by a limited number of attributes
expressed by the help of ad1ectives (e.g. 'that is a horse').
But this, agam, is no exception to what has been said, for here
• to be ' does not stand for mere existence, but for an existence to
be defined by the attributes expressed and implied in the general
noun, so that it means to be somewhat. Indeed (the verb ' is '
not referring to the mere fact that what the nominative case
denotes exists, but to the whole kind of existence which it has),
in this particular case we express what this existence is by
identifying the subject of attributes denoted by the nominative
with a subject of the kind referred to by the attributive. This
particular form may be conveniently called 'identifying state-
ment',• its extreme use being where the attributive 1s individual.
This subject will be returned tom a subsequent discussion of
the question whether the verb ' to be ' always expresses identity. 1
Lastly, a s1mdar account holds of the attributive use of the
verb ' to be ', when the general noun following does not refer to
the whole bemg of the sub1ect. It holds too of the case where
the noun following ' is ' is the name of an individual, for in this
case also the verb ' to be ' does not stand for mere ex1stence. 1
§ 78. Hitherto the subjects, the attributes 3 (\\hether repre-
sented by nouns or not), and the relations which we have
examined were all particular. We have now to consider uni•
versals.
The earlier stage of language is able to do without the names
1
§§ 86, 87 • 5 79.
• To avoid misunderstanding, observe that when we say of an attnbute•
element, mot10n, that 1t has velocity we cannot mean the universal by 'motion•,
for the universal has no velocity, but only a particular motion has it. And
even when we say' motlon is either uniform or irregular•, where 'mot.ion•
seems umversal, we can only mean that particular mot.Ions are regular or
uregular , though, as will be explained m the sequel, we do thus express
a universal statement

[• ' The extreme case of 1dentmcation 18 that whl"re the attribute :is indi•
w:lual •.-MS. note Cf. pp. 194, 208, a.nd 349.]
188 STATEMENT, THINKING, AND APPREHENSION
of universals, that is without nouns which mean universals.
And it is worthy of remark that even the most developed stage
of language both of ordmary hfe and of science does not evolve
names for universals of substances or individual things. Such
a word as ' horseness ' or ' animalness ' appears only in the
artificial language of philosophical mvestigations. In ordinary
speech we should use a circumlocution such as 'being a horse'
or ' being a horse in general '. But the simpler language can
dispense with umversal names There, though the umversal
proposition expresses a connexion of umversals (as even Locke
saw), this 1s by various devices expressed m grammatical forms
which belong to particulars.
The most defimtc and adequate method is to take for the
nommative the plural of a particular noun preceded by such
a word as 'all': 'all crabs walk sideways', 'all eqmangular
triangles are equilateral '. The grammatical form 1s that which
attaches a particular case of the attributive to each of the total
of particulars of the universal signified by the nommabve phrase.
Next the quahfymg words, such as 'each', 'all', may be
omitted and the indefimte plural used . -• crabs walk sideways ',
' equiangular triangles are equilateral '. The meanmg of the
grammatical form 1s the same as before.
Agam the singular, with the word 'any' or 'every', or the
singular mdefimte (as in English with the mdefimte article)
may be used ' a crab walks sideways ', ' every circle has
umform curvature '.
These are methods of general application. There are others
which are not. The defimte article may be used with the
singular of the particular noun . ' the crab walks sideways ',
'the circle meets its tangent m one pomt only', 'the vulture is
carnivorous'; but we shouldn't say. 'the man is carnivorous',
' the man is mortal '. An mterestmg idiom is illustrated in :
'lead is heavy', 'man is mortal', 'man 1s endowed with
speech'. But we shouldn't say· 'crab walks sideways', or
' man 1s endowed with speech but fish 1s speechless ', though
we should say : ' fish, when not fresh, is poisonous '. Since
' lead ' is a word applied to every piece of lead, the particular
being denoted by phrases hke 'this piece of lead' {and similarly
for ' man '), it might perhaps be thought that the words ' lead '
 The Meaning of Grammatical Fo,ms z8g
and ' man ' actually denote universals. But this is disproved
by the verbal or grammatical form. For the grammatical form
is that which attributes something to an individual subject and
we cannot in the sentence substitute nouns which clearly mean
universals for lead and man. We cannot say 'manness is
mortal', it is only the individual man who is mortal; nor can
we say ' manness is endowed with speech ', or ' leadness is
heavy '. The true account seems to be that lead is a sort of
collective and represents, so to say, the whole stock of lead.
But when those nouns appear which are the names of uni-
versals, whether substances or attributes or relations, can they
be treated grammatically hke the nouns which represent or are
names of particulars ?
A particular substance uniting elements m itself, such as
weight and a certain shape, has these attached to it in the
sentence by the corresponding adJectlve, cg. 'it is heavy', 'It
is round '. Similarly for an attribute , to express the fact that
a movement has swiftness and straight direction we say 'this
movement is swift and rectilinear '. To these attributes corre-
spond universals, and so we may say that the universal of the
given movement involves m Itself the universals of swiftness
and recblinearity and that 1t is mseparable from them. The
universal itself 1s ' swift rectilinear motion '.
Now, though a particular has certam particular elements, the
universal of that particular has not necessarily the universals of
those elements a.s elements of itself. Thus the universal ' swift-
ness' cannot be an element of swift, rectilinear motion ; for,
if it were, its bemg would be comprised entirely in the being
of ' swift recbhnear motion ', and then a swift motion would
be necessarily rectilinear. Similarly for rectibneanty. Thus,
though the universal ' swift rectilinear motion ' appears as one,
and as a complex mvolvmg the universals of swiftness and
rectilinearity, it is not a complex of which these umversals
are elements or members. Nor is swiftness an element of the
universal ' motion ', for then all motions would have to be
swift. The universal of an emerald, to take a substance, is of
something transparent, green, heavy, and of a certain shape.
As before, transparency, greenness, weight cannot be regarded
as elements of the universal ' emeraldness '. Thus, if a particular
Igo STATEMENT, THINKING, AND APPREHENSION
is a unity of elements or attributes, the universal of that par-
ticular is not necessarily a universal of which the universals of
the given attributes are elements or attributes.
These objections do not apply to such a relation as that of
velocity to movement in general. The being of velocity is com-
prised in that of movement, and movement cannot be without
velocity. The universal of velocity may be said then to belong
to the univers.il of movement. So also the being of volume and
surface is comprised in that of body, and the actual being of
body must always be accompanied by surface and volume.
Suppose then we allow them to be elements in the universals
of movement and body respectively, as universals which belong
to those universals. Whether the universals which a given
universal involves can be said to be elements in the latter or
not, the noun which is the name of the given universal cannot
appear as the nominative to the verb which corresponds to one
of those other umversals, nor as nommative to the verb ' to be '
with the adJectives correspondmg to the other universals, nor
can we employ the circumlocution with 'have'. Thus the
analogy of the case where we are sure that the elements are
elements in the particular thing cannot hf' followed here.
For instance we cannot say ' starness twinkles ', but only
'a particular star twinkles'. Or, to take a universal which
might be considered an element m another (sentiency in animal-
ness, rationality in manness), we cannot say ' animalness is
sentient', but only 'a particular animal is sentient'. We can•
not say ' bodmess is heavy ' or ' bodmess is extended ', for only
a particular body 1s heavy. We cannot say 'bodiness has
surface', nor 'motion in general has velocity'. The ordinary
adjectives and verbs then will not do, and there is no special
set of verbs or adjectives analogous to the ordinary ones to
express the connexion of such a universal as velocity with such
a one as motion. Even when we say ' velocity is an attnbute
of motion ', we mean that a particular movement has velocity
(a particular velocity) as attribute, and not that the universal
of motion has.
Nevertheless we are not at a loss to express the true relation
of universals. We can use the grammatical forms for particulars
with the devices already explained, e g. ' a movement must have
 Tiu MeanlH-g of GraffUMllcal For,ns 191

velocity', and I a velocity must be the velocity of a movement'.

At first sight it may seem an imperfection of language that
a universal fact should be expressed by a statement which refers
to a totality of particulars, but, as will appear later on,1 this
usage of language corresponds to an important truth. The
present discussion may serve to bring out the fact, which is
sometimes forgotten, that the grammatical forms of our existing
language are the expression of particulars; not, however, of
mere particulars {there is no such thing), but of particulars of
universals. It is this which makes it fallacious to use ordinary
grammatical forms unguardedly for universals qua universals,
that is for universals when expressly stated in the form of
universal nouns proper.
1 § Sz ad fin , cf § 71
 IX
THE SYMBOLIZATION OF FORMS OF
STATEMENT
79. • FROM. the discussion of these grammatical and meta-
physical distmctions we return to the problem of the adequate
symbolization of the general relation which the statement
denotes, the relation usually represented by all (some, this) S is
P ; and first we will raise the question how the parts of the
sentence so symbohzed should be named so as to avoid confusing
the objective relation, represented by the distinction of the
nommative case from the verb and what accompanies the verb,
with the relation of subJect and predicate? We were led to
beheve that we cannot retain the ordinary analysis of the
sentence into grammatical subject and grammatical predicate,
for that seemed mcons1stent with any usual or legitimate inter-
pretation of those terms.
The nominative case does not always comc1de with the logical
subject , on the other hand 1t hardly seems correct to call it
the grammatical subject. If there were a special sort of subject
called the grammatical subject, there should be a special kind of
predicate called the grammatical predicate ; but this 1s not so.
Grammarians borrow the term predicate from logic uncritically
and without any mtention of giving 1t a meaning other than it
has m logic ; accordingly they make no special defimtlon of
a grammatical predicate If then there 1s no special grammatical
predicate, we are not entitled to speak of a grammatical subject.
What would be mtended by such a phrase is that for which the
proper grammatical term 1s ' nommatlve to the verb ' in a
prmcipal sentence, and its grammatical equivalent (for instance
the accusative with the infimtive, m Latin and Greek) in de-
pendent sentences.
Now m the case both of those kmds of statement the verb
in which is an attributive or relational verb, ' he walks about ',
[• 'Reconsider use of term metaphysical sub1ect mall that follows.'-MS.
note]
 Stlbject of Attribut-ion Mil A#ribulive i:93
• he watereth the hills ', and of those where the verb is the verb
• to be ', followed by an adjective or adjectival phrase like • he
is troubled ' (and not by a noun or noun-phrase), what is
ordinarily treated as the distinction of subject and predicate is
just the objective distinction between the object denoted by the
nominative case and its attributes or relations.• We may, as we
saw, include relation under attribute, m a wider sense of the
latter term, and then the general distinction is that of subject
and attribute. Instead therefore of the misleading term pre-
dicate, it would be better in the sentences in question to use the
term ' attributive ' for that part of the sentence which corre-
sponds to the attribute or relation which the nominative to the
verb is said to have, what has been called above its ' attributive
bemg'.
The same nomenclature may be applied to those other kinds
of statement in which the verb ' to be ' is followed by a noun,
whether a common noun or a proper name ; for instance, this
box is a cube, Socrates 1s a man, the person approaching is
Socrates. It 1s true that in neither case 1s the relation of that
which 1s denoted by the nommative case to that to which the
noun refers that of sub3ect to attribute, but, in the case of
a common noun, this noun, though not referring merely to
attributes but to a sub3ect as possessing attributes (whether
a part of these, e.g ' a cube ', or indicating somehow the whole
of them, e.g an animal), only serves to convey information
about what the 11ommatlve case b represents through the assign•
ment to 1t of these attributes. The use then of the common
noun 1s attributive. Its affinity to the adJective appears in the
fact that m a certain form of sentence 1t cannot become the
nominative to the verb. b It cannot, with the significance it has
in the sentence where it 1s not the nominative, stand itself m
the nominative case to the verb ' to be ' followed by a. common
noun, or by a verb or adJective which assigns a permanent
characteristic as opposed to denoting an event. We can say
'a whale is a mammal', but not 'a mammal 1s a whale•, for
[• Cf Lotze, Logic. i 2, § 53
b ' nominative to the verb •, • nominative case ' As there 1,no special
nominative mflenon of nouns m Engltsh, this term seems 1nappropnate We
could only say that m such a sentence as • a whale 1s a mammal '. ' whale ' is 1n
a nomtnabval relation to the sentence)
an3•l 0
194 STATEMENTt THINKING, AND APPREHENSION
the latter expression, according to grammatical rule, if it meant
anything, should mean that every mammal is a whale. When
the common noun ' a mammal ' is really a nominative in such
a sentence, 1t stands for ' any mammal ', and it does not stand
for this when in the ' attributive ' of a statement.
The case is different in the remaining class of statements,
where a proper name 1s Joined to the nommative by the verb
' to be ', as m ' the person approaching 1s Socrates '. The
proper name denotes a particular subject and therefore can be
a nomindtlve m any form of sentence. But nevertheless the
proper name 1n such a sentence JS not of use as designating
a subject, for that 1s already implied in the noun itself which
is the nominative ; its use, as servmg to give an account of the
subject, hes in the attribute it implies. In this case then also
we seem justified in calling the part of the sentence consisting
of the verb ' to be ' and the proper name, the ' attributive '
part.
The general account then of all the affirmative forms with
the verb ' to be ' JS this : every such sentence affirms that what
the nominative denotes has a certam kmd of bemg, which 1s
either part of its bcmg (when the attributive has after the verb
the adJectlval form and sometimes when it has the noun form),
or else includes its complete being.
The function of these affirmative forms is quite mdependent
of the question whether the nominative case is the logical sub-
Ject or not, and holds when the nommat1ve 1s not the logical
subject. Thus if, seeking information about elasticity, we ask
in what substances elasticity is found, an answer to the question
may properly have the form • Glass is elastic '. This, though
we did not ask for information about the nature of glass, is the
verbal form of the description of a part of the nature of glass.
Now we often speak of a thing on which an operation of any
kind is performed as the ' subject ' of the operation. In this
way it is natural to call anything about which a statement is
made the subject of that statement. We may therefore define
as 'sub1ect of statement',• with reference to a given statement,
[• Cf 'In gold 1s yellow, the grammatical subJect 1s the word gold, the
subJect 1n the Judgement, the logical subJect, 1s not the idea of gold, but
gold• Lotze, Logic, 1 2, f 37]
 Su'IJjed of Attribution a.ntl A.ttnbwtive 195
anything which has something said about it in the sta\ement,
considered as such. And what is said about it is precisely the
whole statement. The definition of the logical subject which
we have criticized, namely (in effect) that 1t is that about whicb
we make a statement or of which we say something, was too
wide for the log1cal subject, though it mvolved as we saw
a legitimate notion. We see now that, as every element m a
statement has something staled about 1t, we can't speak of the
subject of the statement m this sense. The statement, for
example, ' A causes B ' 1s a statement about both A and B.
By a subject (not the subject) of statement then we do not
necessarily mean the nommat1ve to the verb1 much less the true
logical subject. Nor 1s a subject of statement necessarily that
subject which 1s represented in the statement as the principal
subject of attributes, that namely which we have called the
metaphysical subJect. 1 As this latter subJect 1s correlative to
the attnbutes which the sentence states 1t to have, we may
convemently call 1t 'the subject of attribution', jUSt as we
have called the remainder of the statement the 'attributive'
of a statement. We see then why the tendency of language
on the whole 1s to make the logical subject coincide with the
nominative to the verb. It is because the function of the
verb or verb-form is to describe the nature of what is denoted by
the nominative case.
It is evident that while the nominative case does not neces-
sarily express the logical subject, it does express a metaphysical
subject, that 1s a metaphysical subject m the widest sense as
a subJect of attributes, and not merely m the narrower sense
of substance. We may even say that 1t expresses the 'subject
of attribution ' of the statement, if we understand by this that
subject of attributes, which is directly represented as bemg such
in the statement. As we have seen, every element of reality
referred to in the statement may be a metaphysical subject,
and that any such element is so may be imphed m so far as
what is attribute of 1t appears in the statement (for an attribute
may appear as attribute and so in adJecbval form 10 a state-
ment, and that 1t has itself an attribute may be indicated by an
adverb). Nevertheless 1t 1s not represented as a metaphysical
1 pp I 58, 166, 168
0 2
 1g6 STATEMENT, THINKING, AND APPREHENSION
subject in the sentence or it would have the noun form. Further,
a noun which is not nominative to the principal verb may have
an adjective attached to it, and the form of the sentence may
be such as to imply that an object referred to by the noun
really has the attribute corresponding to the adjective, as in
the sentence: 'we have a mortal body.' But this is only
indirectly and by implication. The verb being the assertoric
word, it is onJ}' the nominative to the verb which is directly
asserted to be the subject of attribution and only the nominative
therefore which is directly represented as being such a subject. 1
Lastly, we may observe that it is the tendency of the earlier
forms of language and of the simpler form at all times to restrict
the subject of attribution to 'substance '. 111 This appears to be
the reason why the logical subject does not always coincide with
the nominative or subject of attribution. It is really the relation
of subject and attribute, m the wider sense of attribute, which
determines the grammatical forms of the sentence : the relation
of subject and predicate only determines a certain stress-accent. 8
§ So. Every statement must have a subject or subjects of
attributes, but there seem to be cases where logical subject and
predicate may be absent.' In a statement which is the state-
ment of an experience the elements in the object to which the
statement relates may be apprehended simultaneously, as
where, m experiencing the shape and colour of an object, we
cannot fairly say that we apprehend the one before the other.
The same thing may happen in an apprehending or thinking
which is not experiencing. In a given case indeed we may have
been thinking somehow beforehand of the general character of
one of these elements and have been seekmg for further know-
ledge about it. I may see that a crystal is of a green colour
and may not be able to d1stmgu1sh in time my appreciation of
its shape and my appreciation of its colour ; but I may have
been thinking already of crystals and be seeking information
about them, and so far crystal might be called the logical sub-
ject. This however suggests the possibility of a kind of simul-
taneity where we could not make the above distinction, where
it could be said that we had not been thmking of either element
1 See note at end of tlus chapter, p 211. I f75
I 1157-8• '194.
 Sytnbolization of Statements I97
of a whole beforehand. If so the statement which expresses
the apprehension as it was for us could not properly be said to
have a logical subject and predicate. We shall find that this
is not a mere surmise, and later we shall have to recognize the
existence of such rudimentary statements.
In the communication of our knowledge, what we put first
in a sentence is generally the logical subject for the listener,
apart from any context which may have determined 1t other-
wise, and so what is the subject for the speaker may well not
be the subject for the listener This may serve to remind us
that in log1caJ investigations we do not perhaps take enough
account of the obv1ous fact that our apprehensions constantly
do not issue m actual statements, and so we fall to recognize
that 1t would of ten be quite art1fic1al to represent them as
involving a statement to ourselves. When we do come to make
them matters of statement, the distinction of subject and pre•
dicate md1cated by the stress may correspond not to the actual
order of our apprehensions but to some interest which has
intervened.
§ 81. We are now in a position to consider the grammatical
form of the special kind of sentence which 1s usually symbolized
by A is B, or all (some, this) A 1s B, the more modern form
bemg S 1s P, or all (some, this) S 1s P.
The reduction of all statements to this form is, as we have
seen, of the greatest importance to the traditional logic, which"
makes the analysis of sentences into subject and predicate
depend entirely upon 1t, and for the purposes of the syllogism
assumes that every propos1bon can be reduced to the given
form The reduction is quite mdependent of the question
whether B is to be called the predicate, and A, or all A, the
sub1ect in this formulation In itself, 1t merely means that every
statement can be put into a grammatical shape in wh1ch the
principal verb is the verb ' to be ' in the md1cative mood, the
nominative case being represented in the symbolism by A, or
all A, and everything indifferently which follows the verb, that
accordingly which expresses what the object denoted by the
nominative case is said to be, being represented by the single
symbol B.
Now this suggests two questions : How do we know that the
Ig8 STATEMENT, THINKING, AND APPREHENSION
form 1s thus universally applicable ? and, what 1s the value of
it ? We find on trial that we can put any given statement into
this form, but that does not guarantee that 1t will suit statements
we have not tried ; some general proof is therefore requisite.
On the other hand the operation 1s sometimes difficult enough
and the result often seems unnatural and strikes us as awkward
and artificial. We want therefore some proof that it really is
always applicable and some justification for it. If the inquiry
had been undertaken by the ordinary old-fashioned logic, a mis-
take sometimes found in it about the copula might have been
avoided. We shall discuss this mistake later
It has been contended above that the modern use of the
symbols S and P so far from being an improvement introduces
an error ; the old symbolism, due to Aristotle, is better, but
1t 1s imperfect as compared with the prec1s1on attained by
modern scientific symbolism According to the prmciples of
accurate symbolism the capital letters A and B in ' all A is B ',
ought to represent the same grammatical forms, but they do
not. They stand indifferently for both adJectlves or adjective
phrases, and nouns or noun phrases Nor 1s any provision made
for the cases where the attribute indicated by the adjective
following the verb ' 1s ' does not attach to the subjects of
attributes comprised in the nominative case, singly, but only
to them as in systematic connexion with one another. This
causes a difficulty when these statements have to be put mto
the orthodox syllogistic forms. There is besides a manifest
artificiality in resolving all statements into a type beginning
with ' all ', ' some ', or ' this ', for there are classes of statement
to which it is quite unsuited, nor would any natural expression
of them ever assume this form. ·
We are obliged then to consider more particularly the ddierent
sorts of expression which the rough symbolism ' A is B ', or ' all
A is B ', may cover. We shall, m the end, be led to a single
generic form with a twofold distinction like the traditional one,
but for this generic form the ordinary symbolism cannot be
retained on account of the defects enumerated.
Since in the ordinary symbolism, • all A is B ', B usually stands
for an adjective and A often for a noun qualified by an adjective
(all horned animals are ruminants), let adjectives be represented
 Symbolization of Statements 199
by roman capital letters, taken from the first half of the alphabet.
Such a capital letter then wm always symbolize an adjective.
Let nouns be represented by small italic letters enclosed in
brackets, what letters wtll_.,ppear presently. In the case of all
general terms their application to a particular will be indicated
by a numerical suffix.
Consider first the representation of a substance with its
attributes and relations. Let the symbol for a substance in
general be {s) : then the symbol for a particular substance will
be {s) 1, (s) 1, &c.
Let a small letter, from the first half of the alphabet, in
a bracket, such as (a), be the symbol for an element in the being
of a substance, somethmg, that 1s, said to belong to an (s),
e.g. the curved surface of a body. There 1s a corresponding
adjective used to mdicate that a substance has the element (a).
In the example taken this adjective might be ' round ' Let
the adjective correspondmg to (a) be A. Thus we say 'this
body has a curved surface', which 1s of the form (s) 1 has (a),
and 'this body 1s round', which 1s of the form (s) 1 is A But
now corresponding to ' round ' there is another noun, roundness,
which does not mean a curved surface, but the possession of
a curved surface Thus, m general, beside the noun {a) and the
adjective A, there 1s a noun which may be symbolized by Aness
as meanmg the possess10n of (a). Aness then 1s the name of
' having (a) ', and represents a universal The possession of
an (a) or the possession of (a) 1 by a particular (s) 1, is itself
particular, and 1s not Aness but the particular Aness of (s) 1•
In ordmary language Aness would generally be called an attri-
bute, roundness being said to be an attribute of the round body.
If this usage 1s followed, some word different from' attribute' must
be used for (a). Perhaps 'element' may serve. 'Element' has
here the widest sense and is not confined to what is spatially
distinguishable. The foregomg distinction exists in the case of
every attribute and can be vindicated even where not fully
provided for m ordinary language. For instance the movement
of a body 1s an element m its being This corresponds to (a),
but there is no word for Aness. Yet the possession of the move-
ment is the true attribute, and not the movement. These must
be different because we can say of the one what we cannot say
200 STATEMENT, THINKING, AND APPREHENSION
of the other. For instance the movement may be circular, but
we cannot say the body's possession of the movement is circular.
This is just like cases where the distinction is more adequately
rendered m language. We can say, for instance, the point of
a needle is in contact with a plate of metal, but not that its
pointedness is so. ' Elasticity ' (like ' transparency ' and some
other words) seems used ambiguously for both clement, (a), and
attribute, Aness. It certainly often means an attribute of the
body as a whole and the possession of somethmg by the body as
a whole, thus corresponding to the adjective 'elastic'; but this
1s not so when we speak of the coefficient of a body's elasticity,
for the possession of something by a substance cannot have
a coefficient. It may be shown that one result of these con-
siderations is that we cannot accurately speak of a substance as
the unity of its attributes but only as the umty of its elements. 1
The symbol (s) 1s general, as applicable to any substance what-
ever, and the correspondmg words are such general terms as
' thing ', ' object ', ' being ' ; e.g. a ' white obJect ', a ' rational
being'. A substance designated as possessmg (a), or as having
(a) for one of its elements, will be represented by (s) A. A
particular substance (sh has no defimteness except through the
elements of which it is the umty Hence with the exception of
statements about thmgs m general, (s) and (s) 1 cannot appear
except in the combinations (s}A, (s) 1A, (s)AB .. &c , (s)iAB
&c.
If we wish to express that a substance thus designated has
an attnbute, this is done by making 1t the nominative case to the
verb 'is' followed by the adjective corresponding to the attri-
bute. Thus the existence of the attribute Bness and the posses-
sion of the element (b) by (s) 1A is expressed by (s)iA is B.
The complete bemg of a substance is the existence of all its
elements m their unity, and its possession of any given element,
that 1s the existence of any given attnbute, 1s a part of its
existence or form of its bemg. Aness then is one of the forms
which the being of a substance (s) takes. Thus then, in the
statement (s)iA is B, 1t is implied on the one hand that Aness
is one form of the being of (sh, and on the other hand it is
1 Tins whole sub1ect requires a fuller discusS10n than the present limits
permit.
 Symbolization of Statenients 201
stated that Bness is another form of the being of the same
substance. The verb ' is ' refers to the attnbutive 1 being of
(sh in general, and one form of this appears in the nominative
~ase to the verb, while another form of it appears in the attri-
butive part of the sentence. Thus the verb ' is ' is qualified by
the adjective B, just as other verbs are qualified by adverbs.
It must be noticed that we are not here trying to determine
what it is that grammatical forms ought to mean, but only to
recognize what they do mean. We do not, for instance, profess
to settle how the verb ' to be' should be, but how Jt is employed
in statements. In 1ts normal use 1t 1s not put by itself to express
the mere existence of somethmg, but always to express a par-
ticular form which such existence takes. 2 It is combined with
an adJect1ve or adJectlval phrase, or with a noun or noun phrase,
to make up the attributive.
The nominative case to the verb may be a plurality of sub-
stances of the same kind (s)A, in the form ' all (some, these)
(s}A ' , or several substances not necessarily of the same kind
may be associated by conjunctions as (s) 1A and (s) 8B are C.
In these cases the same account is to be given of what 1s expressed
about the being of each such substance m the statement.
A given substance, or a given kmd of substance, referred to
m a statement will not always be represented by a word which
(somehow or other) denotes its whole being, but often by one of
those words which distmgmsh and designate 1t by some aspect or
part of its being. The forms then would be (s)A for a subject in
general which has the attribute Aness, and (s}iA for a particular
subject which has this attribute. Nothmg then more definite
1s implied about the nature of such a subject than that it is
something which has the kind of being, Aness But there are
smgle words which are equivalent to the combmabon of words
(s}A. These are nouns which refer to a substance in general
which is subject of a certain attribute and only designated
through that attribute, such as 'a solid', 'a fluid', 'an organ-
ism', 'a crystal', • a cube'. These may be conveniently
symbolized by the roman capital letter corresponding to the
attnbute, the letter being enclosed in a bracket. Thus (A) will
represent a noun which is equivalent to (s)A, meaning a sub-
' I 77. • See§771nit andp 181,foot-note1.
 202 STATEMENT, THINKING, AND APPREHENSION
stance m general distinguished as having the attribute Aness,
and (Ah will represent a particular instance of such a substance ;
e.g. ' this crystalline substance is cubical ' is of the form (shA
is B ; while ' this crystal is cubical ' is of the form (Ah is B.
(A) then can replace (s)A either m the nominative or in the
attributive part of the sentence; e.g. (s)iA is (B)-' this crystal•
line substance is a cube'; (A)i 1s (B)-' this crystal 1s a cube'.
A substance may be designated through more than one attri-
bute so that the general symbols are (s)ABC . . is ABC ...
If a substance 1s to be referred to m its whole bemg, and not
merely as distmguished by certam of its attnbutes, this is not
done by means of a word corresponding to (s) followed by some
word X somehow mdicatmg the complete series ABC ... , that
is, not by something which would be of the form (s)X, but by
a smgle word which is either a proper name or a general (often
called a common) noun, such as plant or animal The latter is
general as applymg to any particular substance of the given
kmd, but it is not the name of the correspondmg umversal (that
would be a word hke plantness) nor is it the name of a particular,
for it does not strictly speakmg denote any particular It is
rather the general form of a name. If these nouns are symbolized
by bracketed roman capital letters taken from the second half
of the alphabet, as (L), (M), then (Lh would symbolize a given
substance of the kind indicated by (L) that 1s a particular of
(L)ness (e.g. plantness, animalness), the numeral corresponding
to some particularizing word such as ' this ' The whole nature
of a given substance, as a substance, is comprised m (L)ness
(e.g. manness) and this gives tht" distinction between the forms
(L}, representing say •man', and (A), which represents say
'cube'; for in the latter case the universal (A)ness only com-
prises one aspect of the being of the substance, namely, its being
a substance which is subJect of the attribute Aness. 1 If the
fact that a given substance is of the kind (L) is to be expressed,
the grammatical form has (L) for the attributive, just like the
adjective, except that in some languages it is preceded by the
indefinite article, for example, This plant-hke thmg is an animal
• For clearness observe that 1£ A stands for' cubical•, Aness 18 cub.ioalness
as possession of a certain kmd of surface, (A) stands for ' a cube • and (A)ness
for 'c11beness • or • being a cube', viz 'being a substance wluch possesses
tbe given kmd of surface •
 Symb(Jlization of Statements
or (s)iC is an (L). We shall return later to the question of how
(L) indicates the whole being of the given substance.
How then do these general nouns, more especially when
particularized, differ from proper names, which also desigpate
the complete individual? We may incline to think a proper
name like John 1s a mere label, while general nouns bke man
and plant have somehow a significance of themselves. But no
word, obviously, has this character, for the meaning or use of
a word is due to convention, and so far all words ahke are mere
labels. The proper name m the strictest sense is a noun which
by convention symbohzes and denotes a particular substance,
without mcludmg any symbol for the kmd to wh1ch it belongs.
The general noun on the other hand does not symbolize any
particular substance, but 1s used, by convention, in a particular
grammatical combination to indicate that what 1t refers to is
a substance, and a substance of a certain kmd. The association
of a given proper name with a given particular substance, as
denoting that substance and distmgu1shmg it from all other
substancec;, and the association of a given general noun (e.g.
man) with a given umversal of substances (e g. manness), as
mdicatmg in the manner described the substances which are its
particulars, and distmgurshing them from the substances which
are particulars of other umversals, are equally matters of con•
vention Nevertheless, 1t might perhaps be urged that the
general noun is significant in the sense that 1t givt>s information
about the obJects to which it is applied, while the proper name
does not We feel there 1s something unsound about this, for
how could proper names be of any use m speech 1£ they conveyed
no mformatton ? The mistake may be shown as follows. The
obJection comes apparently to this, that to say a thmg has
a certain proper name would give no information except that
this name has been conventionally assigned to it; if then we
did not know the convention beforehand, the statement would
convey no information about the thing named But 1t is equally
true that the statement ' this is a plant ' would convey no
information whatever if we did not know what the word ' plant '
stands for by convention. In the case of both words alike,
when we know what they mean, information is conveyed by the
use of them in a sentence. ' The person coming is Socrates '
204 STATEMENT, THINKING, AND APPREHENSION
conveys mformatlon, and maeed much more information than
'the obJect approaching 1s a man'. To distinguish proper
names we might use the 1tahc letter i m brackets, as (i) 1,
(i) 1, &c., 'i' standing for md1vidual substance. 'The person
approaching is Socrates', then 1s of the form (Ah is (ih; 'the
thmg coming in sight is a ship' 1s of the form (s)iA 1s (L).
The sentf'nce with ' 1s ' may also 'itale the relation m which
one substance stands to another, and the manner m which this
1s done has already been discussed 1 The relations a thmg
stands m are parts of its bemg, and consequently the phrase
through which the relation 1s expressed (connected with the
nommative case by the verb ' 1s ', exactly like the adJective)
assert'i a certam form of the bemg of the substance denoted
by the nommatlve case A number of substances m relation to
one another form a system which 1s a umty. Let a Greek
capital with suffix, ~ 1 , stand for an md1v1dual substance, repre•
sented by (s)A1 or (A)i or (L) 1 or (i) 1 Such a system may
he symbolized, following a mathematical precedent, by qi(l:11
l:11, ~.1 • • &c ) The system has attributes which belong to
rt as a unity and a whole, and do not belong to the md1v1dual
members of the system by themselves Accordingly the manner
of expressing the possession of the attributes by the one system
may take the forms descnbed for substances Thus they fall
under the usual formula ' all A ts B ', and under the more
detailed substitutes for this which have been given above.
There 1s, howevf'r, another case, concerned with a system or
group of related md1v1duals, which does not fall under the
formula ' all A 1s B ' ; one for which the traditional method
has made no pro,·1s1on Nor 1s 1t covered by the other formulae
proposed. When a number of md1v1duals form a group or
system in consequence of their relations to one another, 1t may
be that each stands to the rest in the same kmd of relation.
This 1s expressed by associating nouns m the nominative case,
which denote the members of the system, by conjunctions or
some other grammatical device, making them nommative to
the verb ' to be ', and adding in the attributive part of the
sentence an adJective or adJect1val phrase corresponding to the
given relation , for mstance, ' X and Y are equal ' , ' X and
I § 74
 Symbolization of Statements 205

Y are compatible ' ; ' things equal to the same thing are equal
to one another '. The adjective refers to something which
belongs indeed to each of the subjects denoted by the associated
nominatives, yet only in relation to the others. Now m the
ordinary formula 'all A is B ', the adjective or noun denoted
by B belongs to each A separately, independently of any rela-
tion it may have to the rest, so that of each A we can say it
is B, and the sentence is equivalent to ' each of the A's is B '.
But the adjective which forms part of the attributive in the
case before us 1s not of any kmd which B can represent, since
it does not apply to each of the associated nommabves by itself
but only as in relation to the others Consequently also, such
a sentence cannot be reduced to a form m which the adjective
1s attached to each member of the set singly. We cannot say
in the above mstances ' X is equal ', or ' X is compatible ', or
' each of these thmgs equal to the same thmg is equal to one
another'. It 1s obv10us therefore that these cases are qmte
unprovided for m the traditional form 'all A is B ', which is
given as the general form of all propositions with the verb ' to
be ' On the other hand, the fact they express must be put into
language quite different grammatically 1f the ' all A is B ' form
is to be possible How this can be done will be considered later.
As already explamed, 1 when either elements or attributes
come to have names denoting them, as umting elements within
themselves, these also enter mto the same lmgmstic forms as
substances and attributes Thus a movem,.nt unites in itself
velocity and direction For the velocity there are adjectives
such as ' swift ', correspondmg to A, and the attributive noun
sw1ftness, correspondmg to Aness ; for the direction, such
adjectives as ' straight ' or ' circular'. It must be remembered
that it 1s as a particular movement that the noun for it can
have these adjectives attaching to it. 2
It remams to symbolize the analysis of the sentence mto the
subject of attribution and the attributive To av01d confusion
with the symbolism above used for the grammatical distinctions
letters from another alphabet may be used, say Gothic, and the
sentence with the verb ' to be ' may be represented by S is a.

I f75 1 p 187, note 3, and I 82

2o6 STATEMENT, THINKING, AND APPREHENSION
This symbolism might perhaps be called logical as opposed to
grammatical.
It will be more scientific however to write Sa for the state-
ment in general, because this would cover the forms in which the
verb is not the v1.rb ' to be'. Thus ' John runs ' might be
symbolized by S1 a1, and 'John fishes' by S 2 8z, and 'John is
a good runner' by S 3 a 3
But to show the argument m the syllogism we mm,t have the
accurate symbolism of the grammatical forms, and either, as m
the ordmary analysis, mtroduce the words 'all', 'tlus ', &c., or
allow (s)A to stand for ' all (s)A' or ' any (s)A ', and (s)iA, &c , to
stand for ' this (s)A '. In either case the verb ' to be ' should
be kept for the purposes of the syllogism.
§ 82. The above symbolic formulation belongs to cases where
what 1s represented as a sub3ect of attributes (and, as such, is
denoted by the nommative to the verb) 1s particular, whether
sub3cct or attribute Tlus, as we have seen, 1s the general and
normal type m ordmary l,mguage, an<l even umversal proposi-
tions are normally expressed m this fom1. It remains to con-
sider the less frequent cases where the nommat1ve to the verb
1s an abstract noun, hke triangularity, covetousness, humility,
and therefore apparently not a particular but a umversal. What
is the meaning of the forms m which an abstract noun is
nominative to the verb, and to the verb ' to be ' m particular,
and can any statement about the umversals wluch they denote
be always put in this form? Here we must remember that
abstract nouns are not confined to the names of attributes,
though it 1s true that abstract general nouns covering the whole
nature of a particular substance are rare and hardly to be found
except as art1fic1al and techmcal expressions m philosophic or
scientific language
First we must d1stingu1sh a class of cases m which, accordmg
to an idiom of language, the nommat1ve, though the name of
a universal, is nevertheless not the true sub3ect of the attribute
correspondmg to the ad3ectlve, so that the adjective does not
properly attach to the nominative case. In statements like
'covetousness is blameworthy', \\e do not blame the universal;
we blame the md1v1dual persons who are covetous because of
their covetousness. The ad3ective or adjectival phrase then
 Symbolization of Statements
which follows ' is ' does not apply to the universal of the attri-
bute denoted by the abstract noun, but to the particulars which
possess the attribute. The abstract noun is not used in its strict
sense and there is a kind of brachylogy. Again the particulars
referred to may be the true particulars of the attribute-universal,
that is particular instances of the attribute. Thus we may say
' covetousness is a disgrace to a man ' or ' covetousness disgraces
a man ', meanmg not that the universal ' covetousness ' 1s dis•
graceful, but that the particular covetousness of any covetous
man 1s a disgrace to lum Once more the adjective does not
attach to the nommat1vc ' covetousness '. This 1s the ordmary
use of abstrart nouns m normal an<l unartificial language and
the employment of the adjective, which properly belongs to the
particular, shows how the tendency of language to express every-
thmg in terms of particulars mamtains itself even here.
The same kmd of tlung happens even when we have to state
the d1fferent1at10n of a umversal mto its species. This again is
not usually done by means of an abstract noun denoting the
genus, but very commonly by words denotmg the correspondmg
particulars, as ' triangles are either eqmlatcral, isosceles or
scalene'. But, when the abstract noun 1s used, the adjective is
still the adjective which properly belongs to the particular, as
m the sentence, ' movement is either rectdmear or curvdmear '.
The nommative is ostensibly the name of the umversal, but
movement 10 general does not move nor has it any direction ;
only particulars can move 10 a direction, and the more accurate
expression of our thought would therefore be ' every particular
movement is either rectilinear or curvilinear '. In the rarer
cases, where the abstract noun is used 10 its true significance
and is the true nommative to the verb ' is ' in affirmative state-
ment, the attributive part of the sentence followmg ' is ' most
usually has the noun form, and not the adjectival form, and
consequently an identity of be10g is expressed ; for mstance,
' circularity is a species of comcsectionness ' ; ' covetousness
is a form of selfishness ' This is natural, for even 1f universals
are allowed to have umversals as elements there are no adjectives
or (attributive) verbs to express this, and so to follow 10 the
sentence the analogy of adjectives and verbs denotmg elements
in particulars.
208 STATEMENT, THINKING, AND APPREHENSION
Can however the relational forms of expression with ' is ' and
' are ' be used for the relations between universals, the universals
being represented by abstract nouns, on the analogy of the
correspondmg sentences which state the relations between sub-
stances ? The usual and natural way of expressmg what are
commonly called relations between universals is by statements
about their particulars ; e g an equilateral triangle must be
equiangular. This may be put artificially in the form ' equi-
lateral-tr1:i.ngleness necessitates eqmangular-tnangleness '. Such
expressions however are madequate to what is mtended, and for
accuracy we should rather say ' equ1lateral-triangleness neces-
sitates m its particulars eqmangular-trianglencss ' or, again,
' equilateral-triangleness excludes nghtangled-triangleness from
its particulars ', or agam, ' 1sosceles-tr1angleness is m its par-
ticulars compatible with nghtangled-tnangleness '. The accurate
form requires always the reference to the particulars.
Now the relation of a universal U1 to another U2 belongs to
the bemg of U 1 This bemg so, on the d,nalogy of the treatment
of relations between particulars, an adJecbval phrase expressmg
the possession of this relat10n may be used after the verb ' 1s ',
the universal U1 bemg the nommat1ve to the verb. These
adjectival phrases are sometimes very art1fic1al and awkward, as
e.g 1£ we substitute m one of the above sentences ' 1s neces-
sitatmg ', or ' 1s a necess1tator of . . .' for ' necessitates '.
But these forms are comparatively rare , as we said, in scientific
as m ordmary language, the relations between umversals are
normally expressed by means of statements about their par-
ticulars. Tht'3 1s even true of ' identifying ' statements, such as
the 1dent1fication of a umversal with a species of another. 1 For
instance, 1t 1s natural to say ' the circle is a species of conic
section ', or ' the circle is a conic section ', and not natural to
say ' circularity 1s a species of c-omc-sectionness '.
In fact ordmary language reflects fatthfully a true metaphysic
of umversals. Its forms testify everywhere to the mseparable-
ness of the universal from its particulars, and to the impossibility
of knowing or expressmg anythmg about the nature of a uni-
versal apart from its particulars. The reason is that the universal
has no reality except as a universal of particulars. A misunder-
' § 77 (at end).
 Symbolization of Statements 209
standing of linguistic facts has produced a number of puzzles
about the universal, which, mistaken for metaphysics, are
nothing but verbal fallacies. It is characteristic of them to
employ, in regard to the abstract universal, lmguistic forms
which are only proper for the particular 1
§ 83. It 1s now possible to see that every affirmative statement
can be resolved mto the verbal form usually represented by all
(some, this) A 1s B. In this form the symbol B represents the
adjective, or common noun, corresponding to B and (L) in the
symbolism suggested above. The question really is whether an
equivalent, havmg for principal verb the present ind1cat1ve of
the verb • to be ', can always be found for a statement in which
the prmc1pal verb 1s not the verb ' to be '
These other verbs have been d1v1ded mto •attributive' and
• relational '. The attributive verbs serve to assert that the
object denoted by the nominative to the verb has a certain
element (b) m its bemg. Now we have seen that, 1£ B 1s the
ad1ect1ve correspondmg to (b), the same assertion can be made
by takmg B for the attributive after ' 1s ' or ' are '. For the
resolution therefore all that has to be done 1s to find an ad1ect1ve
B corresponding to (b) and, 1f there 1s not one, to coin an
adjectival express10n.
The same holds for the relational verbs, inasmuch as every
relation 1s also an element in the whole being of the thing
related. The resolution then 1s always possible, whether the
verb 1s an attributive, an ordmary relation,Ll verb, or the verb
of being itself followed by a relational expression.
The attributive of the form ' is (L) ' does not arise out of the
resolution of these attributive and relational verbs, because they
do not cover the whole being of what 1s denoted by the nomina•
tive case. Thus the statements, when resolved, will assume the
forms having ' is B ' in the attributive place.
We have shown above II how sentences in which abstract
nouns denotmg universals stand in the nommative case can be
turned into the required form. The adjectival phrases which
have to be invented for the purpose are often awkward and
unnatural enough, but they are none the less legitimate. A
second and special artdic1ality arises from the fact that the
I f 82o
, 1 p. 156, Dote I; §§ 41 S-21.
11 773'1 p
210 STATEMENT, THINKING, AND APPREHENSION
distinctions of time represented by the ddierent tenses of the
verbs, including those of the verb ' to be ' itself, have all to be
expressed by using one and the same tense, the present tense,
of the verb 'to be' , the past present and future reference
being transferred to the adjectival part of the attributive. But
this also has its justification, as will be maintamed later in the
d1scuss1on of modality 1
Fmally, there ts an artificiality sometimes employed m resolu-
tion which may seem at first not merely artificial but an
mfrmgement of the grammatical form The relation of attribute
to subject bemg normally expressed by makmg the subject of
attribution the nominative to the verb and taking, for the
attributive, '1s (arc) ' with the adjective corresponding to the
attribute, the same thing 1s sometimes expressed by turning
the attribute mto the nominative to the verb, wlule the name
of the subject of attribution appears mvolved in the attributive
side of the sentence. Thus, 1m,tead of saying ' this pear 1s
yellow ' we say ' yellow colour ts m this pear', a form of resolu-
tion often employed by Aristotle But though the name of the
subject of attribut10n 'pear' 1s on the attributive !>tde, the
changed grammatical form does not put 1t m the pos1t10n of
an attribute of the given yellow colour, or make 1t a form of
the bemg of the colour The grammatical form indicates that
the noun corresponding to the attributive adjectival express10n
1s not 'pear', but 'the belongmg to pear'. If 1t be objected
that ' belonging to pear ' 1s not an attribute of the yellow colour,
or an element m 1t, because 'belonging to this pear' or 'being
in this pear' means 'bemg an attribute, &c ', and so covers
the whole nature of the yellow colour, the reply \\ould be that
since the sentence means, as supposed, 'yellow colour 1s an
attribute of th,., pear ', this ts a case where the attributive part
has the form (L) as covering the whole being of the origmal
nominative to the verb On the other hand 1t 1s clear that such
artific1ahty is not necessary to effect the resolut10n mto the
given form, for the natural exprcss10n ' this pear 1s yellow '
1s itself of the given form. The artificial form 1s used by
Aristotle as convenient for the presentation of syllogistic
arguments.
I § 93
 The Predicate Adjective in Greek 2n
NoTE TO § 79.
' We have a mortal body ' means that the body we have 1s
mortal. With respect to this case there 1s a confusion in the
grammarians in their use of the word predicate. In Greek
Grammars for mstance it 1s customary, mdeed the rule, to say
that when a substantive qualified by an adjective has the defimte
article and an article 1s not prefixed to the adjective, this shows
that the noun 1s subject and the adjective its predicate
The nature of the 1d1om 1s, however, thus quite wrongly
characterized Consider the examples, TO p.Ev crwµ.a 8v71Tov l'iwavus
lxoµ.fv 1 (Isoc Philip, 134) and ,b-' dp8~s ical Tmalas ~s v,vx~~
(Ta 1TrtVTa) <TV/J-1~ftJov>.71rca (Dem Drat 18. 298) In the first
passage TO 11&,µ.a 1s always said to be the subject and 91'1/TOv the
predicate of 1t, viz ' our body 1s mortal '. But if the sentence
were an answer to the question 'what 1s there mortal in us? '
the log1cal predicate would be not 61 111r<1v but To crilJµ.a, and
01 11To1• would correspond to the logical subject
1

The proper way of statmg the 1d1om seems to be this .-If

the connexion of a given subject and attribute (or attributive)
is presupposed m the sentence and not made matter of state-
ment, the adjective has the article before it-To 61111Tov To 11ilJµ.a,
or, what 1s a case of the same tlung, comes between the noun
and the article (To 1jµ.lupov 81•71r,,v 11ilJµ.a). But, 1f the connexion
of a subject and attribute is 1mphcitly matter of statement,
whether the subject 1s represented as havmg a part1cul::ir attri-
bute, or the attribute as belongmg to a parllcular subject, then
the noun has the article and the adjective not. So m the present
case what 1<; implied is To crilJµ.a ;,µ.wv EITn 61"17ror. Bul either
rrilJµ.a or flv11T&v may belong to the logical predicate •
1 Madv1g, GP'eek Syntax, § r.z
[• Lit 'wlule the body we each have 1s mortal ' \,\'llhon'i, pomt ma:,, L,:
examined m the followmg
i, To{UTl/f EOll<fV OU O'/J&t<p~v ,Ppov,w
.,;, -,ap
/Jctva110'av Tqv Tixv'lv lr<Tf/<T"-/-'f/V• Soph AJaX, 1121-.?
In Wilson's sense /lava110-av 1s sub1ect • You don't talk like a hl,1.,,p' 'A 5Javc
mdeed, not slavish 1s my arl ']

P2
 X
THE COPULA AND THE MODALITY OF STATEMENTS
§ 84 THE mistake of confusing the distinction of subject and
predicate with an objective distinction culminates in an account
of the i:;o-called copula which is often associated with 1t. This
appears in a common doctrmc that the word ' 1s ' does not
signify reality but 1s merely a sign of predication. Thus the
members of the symbolic form, all A 1s B, are familiarly termed
subject, copula and predicate, where the very name 'copula '
implies that the verb ' to be ' is a mere link, properly under-
stood, a sign of the connexion of A and B m the sentence , for
the termmology 1s justifiable only 1f 1t characterizes what 1s
essential m that of which 1t is used. This doctrine 1s entirely
opposed to the view we have mamtamed ; namt'iy, that, in the
form in question, the vrrb 'to be' refers to the being of the
object which 1s denoted by the nommat1ve to the verb
Now we might reasonably suspect the poss1b1hty of such
a change m the meaning of '1s '. It 1s a paradox that such a
simple word should lose the very essence of its meaning, and
still more so when 1t keeps that meaning m ordmary speech.
Ag,un, as a matter of logical formulation, 1£ the copula has
nothmg to do with bemg and is only a sign of ' predication ',
surely 1t would be better to avoid the word ' 1s ' altogether and
to put a symbol for it, jUSt as we put A and B for the (so-called)
subjl•ct and predicate. This bemg so, 1t 1s remarkable that the
verb ' to be' should be employed so commonly and in so many
languages to express this relation and that its use should not
only be allowed but be universal in logical treatises. There
must have been good reason for choosing, consciously or uncon-
sciously, this word in particular. The reason appears to be that
the symbolic formula contaming 1t 1s not merely artificial, but
generalizes ordmary usage, that, m that formula, as we have
tried to show, 1 the verb 'to be' retains its ordmary meaning.
How then could the contrary view have arisen ? Here agam
I§ 79
 The Copula .a,ul Modality 213
we feel that there must have been some ground for accepting
a theory so much opposed to the common uses of language.
The doctrine seems to have arisen m the following way. In
the form 'all (some, this) A is B ', A being called the subject
and B the predicate, as ' what is stated of A ', 1t remained to
find a function for 'is' in relation to the 'predication'. Now
there can be no statement and so no ' predication ' m the given
form without the verb '1s' (or 'are'), and 1t was on this ground,
probably, that the verb ' to be ' was taken erroneously to be
the sign of predication. Thus though Aristotle's analysis 1s mto
sub1ect and predicate merely, we find the copula recogmzed as
a distinct member already by Abelard, 1 who says, ' the members
out of which categorical propos1t1ons are combmed are predicate,
subject and their copula (hnk). The reason 1s that we separate
the verbal notion from the predicate and take 1t by itself .
the interposed verb lmks the predicate to the subJert '.
There 1s somcthmg prima facie hke this m Aristotle's De
/nterpretatione, 2 where 1t 1s said that the verb 'to be' by itself
means no ex1stmg reality, and so m itself 1s nothing, but 1mphcs
a certam con1unction (1 e. of realities) which cannot be under-
stood apart from what is conjomed This however is not part
of a discussion of the analysis of the sentence mto subJert
and predicate, but comes m merely to tllustrate the doctrme
that words m isolation from the sentence, though they have
a meanmg, do not mean that anytlung corresponding exist&.
It is in this connexion that Aristotle says it 1s not even true
that the word for existence, or bemg, itself, when put apart from
any context or sentence, means that anythmg exists. This then
1s so far from bemg a deliberate analysis of the sentence mto
subject, predicate and copula that there 1s no ment10n m it of
predicate or sub1ect and the question of such an analysis 1s not
before Aristotle at all. Conceivably, 1£ Aristotle had h.i.ppened
to reflect upon the connexion of this with the analysis of the
1
Sunt autem membra [videlicet propoi.1bonum categoncarum] ex qu1bu,;
coniunctae sunt praed1catum ac 1,ub1ectum atque 1psorum copula, secundum
hoc sclhcet quod verbum a praed1cato seorsum per i.e acc1p1mus Verbum
vero 1nterpos1tum praedlcatum sub1ecto copulat Abelard, Dialectica, Cousin,
P 246 (Prantl, I c , vol 11, p 196, note 370 , cf id, p 206, note 11 )
AilTI) ,.~,, -ydp oua,,, ltrTI, 'llpG17U7//IIJ'"" 3, u(,116•11/11 T111a, ,Ov dv•u TiiJII avy•••I'•"°'"
1

oil1t In, 11oijaa,, D, Int 3 , 16b 24,

214 STATEMENT, THINKING, AND APPREHENSION
sentence into subject and predicate, he might have developed
what had occurred to him into a distinct recogmt1on of the verb
' to be ' in the standard form ' A 1s B ' as a third element,
neither subJect nor predicate, and might possibly have given 1t
some technical designation, though we may question whether he
would have called 1t a synthesis or copula or by some equivalent
name. But in the whole of his writings this 1s never done,•
which shows how httle he had before him the idea of analysis
into subJect and predicate in this passage and how hazardous
1t would be to attribute the subsequent view of the ' copula '
in any of its forms to him. A comparison of a passage in the
Metaphysu:s 1 in its context with this one shows that by the
synthesis Aristotle at any rate meant notlung merely subJective.
He distinguished clearly enough an obJect1ve synthesis in tlungs
from our apprehension of 1t or our opinion about 1t, and it 1s
to that ohJective synthesis that in our view the verb 'to be'
refers Tins 1s quite sound doctrine in its general form, and
remote from the later view, which we have been considering.
In that view, the verb 'to be' (in the ind1cahve) was conceived
as effecting predication by J01mng the subJect and predicate in
the sentence. Thus 1t was called the copula and came to be
conceived as a sign of predication only
Yet though there 1s no 'predication' (so-called) without
a verb and, in this form, none without the verb 'to be', the
verb 1s not only not merely a sign of predication, but it is not
a sign of predication (in the proper sense) at all It does contam
the sign of assertion, as explained in the discussion of the func-
t10n of the verb, but this 1s not the only purpose 1t serves. In
what sense it might be called a copula has been shown m the
same discussion
§ 85. The view that the ' copula' 1s only a sign of predica-
t10n was the better able to mamtain itself because 1t seemed
confirmed by certain propositions of the given form, alleged m
its favour, where the predicate (or, more properly, the attri•
butive) showed that the obJect denoted by the nominative was
non-existent. An example of the kind of propos1t1on m question
1s ' A dragon 1s an imaginary ammal ' But the true nature
1 Met 8, 10, 1051b II

[ 11 But see D, Int 19b 19-2011 3, 21b 26-33]

 The Copula and Modality 215
of such statements was quite misunderstood. For though the
dragon of course has no being in reality, when reality is dis-
tinguished from imaginary being, 1t has bemg in our imagination.
But being in general obviously includes everything that 1s not
nothing, and therefore includes these imaginary objects. Thus,
m the given statement, this general being 1s conveyed by ' is '
whale the particular kind of being 1s conveyed by the so-called
predicate 1 The statement then about the dragon is a true
affirmative and docs assert a kind of being for its subject.
Examples hke tlus are easy to deal with because the so-called
' predicate' excludes the subject from a particular kind of being
only ; but they suggest a proposition m which the ' predicate '
excludes the subject from every kmd of bemg At first this
might seem to be still easter to deal with, because such a state-
ment could not occur, for what could be the subject (nominative
to the verb} of 1t ~ A word would have to stand for the subject,
but 1£ the subject neither exists nor can be imagined nor in any
way thought, the word would have no meanmg and nothmg
corresponding to 1t , the statement therefore would be meanmg-
less and so not a statement at all
Now m a way tl11s 1s the true answer. But yet there do
occur statements wluch apparently do present this seemingly
impossible form They have a subject which we cannot say 1s
a meaningless word and yet 1t turns out m the proof that there
1s no reality correi,,poudmg and no thought Such statements
occur m science , we may for mstance have an equation to
dctermme the value of x and may by the final reduction get
x1 + 3 =2. This defines x as a quantity such that 1f 3 1s added
to its square the result 1s 2. Such a quantity 1s not thinkable
and yet we cannot say that such an equation m 1t<; final form
when expressed m words as above contams mere meanmgless
words. They correspond to processes m science, and are
illustrated for mstancc by every reductw ad absurdutn argument
10 geometry. Suppose we say 'a square circle 1s a nonentity'.
Here the nominative 1s not qmte meamngless, for square and
circle each have a meanmg. There are mstc1.nces m ordinary
use where the absurdity 1s as great, though perhaps not so
transparent, 1. e. where we begm with a conception which implies
1 Plato, Sophist, 237d
216 STATEMENT, THINKING, AND APPREHENSION
the union of certain elements. These elements in themselves
are known to be real, and we gradually discover that their
umon m the given conception is impossible Thus a statement
of the impossible form can actually occur There will be in it
thmkable elements and it is their umon which we pronounce
unthinkable
We must further ms1st that such statements m their affirmative
form do give their nommatives a kmd of bemg ' X is a non-
entity ' is exactly parallel to ' an anemone is an animal '.
Both indicate formally that their subJect has a particular kind
of being The attributive represents what this 1s. ' Nonentity '
m fact is given as the kind of entity which X has. Thus the
formal mistake of the statement is cancelled m the attributive
through the artifice by which nonentity 1s expressed lmgmstically
as 1f 1t were a kmd of bcmg ; ' X has that kind of bemg which
is not bemg ', which 1s to say ' X has no kind of being '.
We are now able to see what the proper verbal form of the
statement should be The given form 1s an art1ficiahty; 1t
does v10lcnce to the normal forms of exprebsion, and we can see
exactly what the artifice 1c; It 1s the forc10g of a negative
stateml·nt 10to the form of an affirmative ' A square circle
1s a nonentity' really means tlus we have two th10kable
notions, ' square ' and ' circular ', and then a negative state-
ment dependmg upon the difference between them, namely,
' a square cannot be circular'. This shows what gave the
artificial statement, with its fictitious nom10abve, a meanmg;
it shows also clearly that ' 1s ' m that art1fic1al form does mean
bemg In fart we may give a very simple solut10n of the whole
matter thus Just because ' 1s' docs imply bemg, and because
the art1firial subJect or nommat1vc has no k10d of bemg, we have
to put a negative 10 some form or other 10 the ' predicate ' of
this pretended affirmative sentence m order to cancel the be10g
1mphed in the word ' is '. To say now that m the proposition
' X 1s a nonentity ', ' is ' docs not mvolve bemg is obviously
as absurd as to say that m the negative proposition, • A is not
B ', ' 1s ' does not refer to be10g The whole difficulty then
comes from the presentation of a negative statement as 1f 1t
were affirmative, and this artificial operation may be performed
on any negative statement whatever.
 The Copula and Modality 217

§ 86. There is another difficulty about the copula which is

connected with its true meamng. Just as 1t was felt that, in
the form (shA is B, the word ' is ' related to the bemg of
(s)iA, 1t was also felt that B was somehow identical with (s) 1A.
This 1s in an important sense true, but it 1s liable to m1sunder-
standmg and produced a difficulty quite early m philosophy.
Accordmg to Anstotle, a certam Lycophron 1 mamtamcd that,
m a Greek statement of this form, ' 1s ' should always be followed
by a word identical with the subject or nommative, so that
' man 1s white ' 1s a wrong form, because man and white are
not identical He proposed accordmgly to omit' 1s '. We may
reasonably assume that he intended this only for ordinary forms
of statement, where B 1s not identical with (s)iA, and that he
would have allowed the form 'this (md1v1dual) man 1s Socrates'.
The Cynics 2 also are said to have held that the umty of the
subject necessitated the identity of the ' predicate ' with 1t. It
1s characteristic of them and of other philosophers of tl11s penod
of Greek thought that they should have been content with
statmg the difficulty, without trymg to get beyond it. So
Gorgias enumerated his favounte paradoxes and apparently
took delight m mamtammg what, 1f taken seriously, would
destroy the poss1b1hty of knowledge This degenerated mto
a scepticism m which no statement at all was possible except
' man 1s man ' and ' white is white '.
The paradox 1s founded on something true. In the form
(s) 1 A is B, B relates to a kmd of bcmg (Bness) which (sh has.
There must be some unity between (s)i and Bness. Whatever
that unity may be, 1t cannot be mere 1dent1ty ; there must be
difference 1£ it 1s to have any meaning The very obJect of
a statement is to go beyond what is already known of the given
subject. The so-called Judgement of identity, A 1s A, which 1s
sometimes put as one of the laws or forms of thought, is no
thought at all, for as Hegel pomts out,• the propos1t1onal form
as such is not A 1s A, but A 1s B. Such a statement as ' X is
I /Jul al ,.1.. 1'~ l 11 ,- I .. iltf>•i.\or, ll,11.,,,p Av1<6q,pow, ol lJI n)v Al(•" ,,,,.,ppv9µ,(o•, 1,,., cS

6.lllponror ~ AEu1<or la,-,v d.\.\d A•Atvl<OJTGJ., Phys A, 2 , 185b 27

• Plato,Tht 201D,Soph 251B, Anst Metaph 1024b32, [Ross,1,pp 346-7]

[• The Form des Satzes w1derspncht 1hm c,chon o;elbst, dd em Satz auch
emen Unterschled zwischen SubJekt und Pmd1kat verbpr1cht Hegel, Wet'ke
(113.to), v1, § u5 (Wallace, Eng Tr, pp. 213-14)]
218 STATEMENT, THINKING, AND APPREHENSION
X ' has in a way the verbal form of a judgement or opinion, but
a matter which cannot be the matter of a Judgement or opmion.
Difference then bemg necessary, the solution 1s that Bness 1s
not really different from the bemg of the subject of attribution,
(s)iA, but only different from that aspect of its bemg which is
conveyed in that subject Thus, m 'this A 1s B ', Bncss and
Aness are both parts of the bemg of the same subject, each
identified with a part of its bemg, and each different from the
other. Tlus involves identity and difference, an 1dent1ty wluch
without losing its umty ha,; these different forms of bemg.
We thus get away from that false vie"' of identity or umty, as
mere sclf-1dent1ty without difference, which was the fallacy of
the Cymcs and apparently of Lycophron, to the true identity
wluch as identity 1s identity m difference and as umty 1s a unity
of differences, a umty, that 1s to say, such that the elements
which we call different have existence or reality only m unity
with one another; and, v.lulc each of these has m a true sense
a bemg of its own whereby 1t 1s d1stmct from the others, yet
at the sam<.• ttme each hy its own bcmg neccc;s1tatcs the bemg
of those others a
Tlus 1s the general conception of umty and difference covering
every case of statement, but we cannot a priori say anythmg
more about 1t ; its peculiar character can only be seen m
particular 1m,tanrcs, as when, m the statement 'this piece of
chalk 1s round ', I d1stmguish chalk from roundness and under·
stand their umty with one another Or agam take ' this blue
1s a dark blue' The generic notion of blue lhikrs from the
notion of dark and hght which we apply to colours m general ;
nevertheless blue cannot exist except as havmg one of these
shades Here we have another kmd of umty 111 difference
Thus in every statement two apprehensions are given of the
reality of one and the same obJect, on one side m the nominative
and on the other m the attributive Neither, m the normal
case, 1s of the whole bemg of this ob1ect1 but each 1s an appre-
hension of some aspect or part of its being. We see then what
constitutes the umty m d1fference of what 1s stated. The
difference is between two aspects of the obJect, and the umty
[• necessitates v I '1s m necessary relation to the being of those others•
(1907) The 1904 text seems to have got into the pnnted copy by DllStake]
 The Copula and Modality 219

is the unity of the same reality in both these aspects. Further,

it 1s clear that 1f, on the analogy of old usage, we make the
term copula stand for the umty of subject and predicate, that
copula cannot be represented by ' 1s ', nor by any ordinary
significant word as opposed to a technical symbol, the nature
of the umty varymg with the matter of the statement.
The above investigation relates to the form with ' is ' for
principal verb, because 1t 1s here that the difficulty and paradox
about the 1dent1ty of subject and predicate occurs. But of any
statement where the verb 1s not ' 1s ' the same account has
to be given. In th<' nominative, the obJect of the statement
appears designated and d1i;tmgmshed by one aspect of its bemg,
the verb, together with all gr.:1mmat1cally connected with it,
giving another aspect of its bcmg, the obJert itself 1s an identity
in these aspects.
§ 87. The difficulty may now suggest itself that the word
combined with ' 1i; ' in the attributive may perhaps cover the
whole bemg of the nominative, so that the statement would
seem to express an identity Now it 1s the case that there are
attributives of such a kmd, and according to the usage of language
they are combmed with the word ' 11, ' Hence the reference
in '1s' can be to the whole bemg of the subject as well as to
the partial being The difference, however, 1s provided for in
the verbal form, for the attributive when 1t refers to the partial
bemg of (sh, has normally the adJecttv.:11 form, as m ' lemons
are sour ' , while it 1s usual to employ a different form of words
when the attributive somehow covers the whole bemg of the
subJect. We then have a nommal or substantival instead of
the purely adjectival form, e g 'a lemon 1s a fruit', where' a
fruit' covers everything in the subJect.
Is this then not a mere identity or equation ? It would be
well first to dear up the true meamng of equation m algebra,•
for that might seem an obvious case of identity between the
terms represented on either side of the equation We shall find
that an equat10n 1s never a mere identity when 1t represents
any real activity of thought. The equation x2 -2x+ 2 = I 1s
a Judgement {some would call 1t a synthetic Judgement) from
which we get some information about x x2 -2x+ 2 =x2 -2x+ 2
[• Cf. Lotze, Logic, 1 2, f 51)
220 STATEMENT, THINKING, AND APPREHENSION
is a mere identity, for it gives us no mformation about x. It is
an identity but not c1. Judgement. Agam (x- 1)2 =x2 -2x+ I
is an equation which gives mformation, and though it is some-
times called an identity because the development of (x- I )1 is
identical with x 2 - 2x + 1 and gives no information about the
magnitude of x, 1t 1s not a mere identity m so far as it explains
what 1s the expansion of (x- 1) 2 Tlus indeed is obvious from
the fact that we have to prove the equality by mult1phcabon,
and appears more clearly m a complicated mstance hke the
expansion of (x - rt, by the bmom1al theorem Both kmds of
equation have the sign =, but what characterizes the true equa-
tJOn is that, though the subject-matter is the same on both sides,
the aspect of its being expressed on one side and on the other
is different. The sign of equality represents an identity of the
sub1ect-matter m these two aspects In the ordmary arith-
metical or algehra1ral equatJOn the number of units m the
quantity on one side l'l 1dent1cal ",th the number of umts m
the quantity on the other, but the two s1<les represent two
different modec: of const ructmg tlus same number of units and
1t 1s because of tins rhfference that the equatJOn l'l of any use
On the other hand, 1f the method of constructJOn 1s the same
on both sides, the equat10n degenerates mto a mere identity,
and is by itself of no use. Thus 3+ 2 =4+ I 1s a true equation,
for 1t has to be taught, 3 + 2 = 3 + 2 1s a mere identity Generi-
cally an equation represents two d1ffrrent modes of operation
which have the same result, the sign of equality representmg
this 'lameness or 1dent1ty of result , 1t 1s accordmgly not neces-
sary that the equation should represent operations of the con-
struct10n of magmtudc merely ; a ve<'tor equat10n for mstance
represents• l\\o different operatiom, by wluch the same position
is reached m space
§ 88. In those statements of wluch the attributive side some-
how comprises the whole of the sub1ect, the sub1ect, 1f 1t 1s the
[• This should rather be ' may represent' ' The 'lUm of the radn 1Jectores,
drawn from the foe, of an elhp<;e, 1s constant and equal to the ma1or axis•
1s a quantitative vector equation An example of a vector equation which
does not represent an identity m the usual 'lCnse 1s that A+B=R, where
A and B represent two forces or veloc1t1es and R their resultant The 1llustra-
bon was probably suggested by Venn, Sy,nbohc Logic, p 99 Cf Lotze,
Logic, 111 5, § 361 ]
 The Copula and Modality 22I

logical subject, may be something known first under a certain

aspect of its reality and then we may advance possibly to
something which may be said to cover 1ts whole reality. Thus
an object might be first known as something resembling a plant
and only part of 1ts nature might then be given m the subJect
(nommat1ve), but when we get to know that this plant-hke object
(an anemone) is an ammal, the logical predicate covers the
whole bemg of the subject, and 1t ts so far new to us that instead
of bemg a mere 1dent1ty for us 1t 1s sometlung which we did
not even suspect at first. In t hcse quas1-equat1onal 'ltatements
generally there 1s always a umficat10n of different aspects of
the same obJect For even if an attnbut1vc which covered the
whole of its nominative expressed every part of its being (which
1s not really possible) and therefore mcluded explicitly the
elements already state<l m the nominative, there would still be
a umficatlon of differcncm, in so far as the attributive added
new eharactenc,t1r'l to the old But now such ..i.ttnbuhves m
practice do not represent the completed thought of the subject ,
they arc rather the thought of the subJcct as somethmg com-
plete. The defimt1on of a circle, for example, covers the whole
rcahty of e1rcularity and 1s mtended to <lo so , yet 1t docs not
expres~ the whole reality, the dct.uls of "'luch we can only learn
by demonstration. Indeed m tins case the whole rcahty 1s
infimte and therefore mcc1pc1.blc of complete detailed expression
The same tlung applies ,,hen the attribullvc 1s not the true
logical predicate
In the normal use of ' common terms', the whole nature of the
nommat1ve of which they are the attributive 1s assumed to be
1mphc1t m the meaning of the 'common term', but it 1s
explicitly present to the mmd m a part only of its total reality
Thus the common term ' an ammal ' involves, m the first
mstance, the idea of an orgamsm with ccrtam attributes of
motion and hfe , 1t implies also that the full knowledge of such
a notion would be the full knowledge of the thing Tlus accounts
for the artificiality which we find m such expre-,.,10ns as ' So-
crates 1s a wlute object ' , for, according to the natural use of
language, such a common term as is here on the attributive side
ought to mclude the whole reality of the nommatlve, and the
definite notion m 1t ('white') ought to refer to this whole
::z::z::z STATEMENT, THINKING, AND APPREHENSION
being, whereas it refers only to a part. The natural grammatical
expression then would be one m which the adjectival form
replaced the substant1va] form of the common term. Thus the
proper express10n 1s ' Socrates 1s white '. This seems to be
the distinction which underlies Aristotle's theory of certain
(natural) genera and species, certain classes which had some
kmd of prerogative over others The nature of that prerogative,
however, he can hardly be said to have made out His scholastic
followeri, accepted 1t, as usual, uncritically and without eluc1da•
t10n, and through them 1t descended into the ordinary syllog1st1c
logic together with other obscure and unanalysed conceptions
such as ' essence ' 1
Anstotle would say of white and animal as attnbuted to
Socrates that animal 1s a true genus but white thing not The
kmd of explanat10n which he offers 1s that you can say 'Socrates
1s essentially a kmd of ammal ' 2 but he would not allow
'Socrates 1s essentially a kmd of white (thmg) ', 3 though he
does not make 1t dear what 1s the matter with the -;econd
hngu1stJc form I lowevcr he says sometimes of the first that 1t
expresses the essence of the subJect and that the other does not.
Of the first, ammal, he uses the expressions ' that which is
stated of a subject ', or ' genus ' or ' species ', 4 1mplymg that
genus and species proper belong to the essence, whereas the
second, white(ness), ts accident 5 In the Categories we have
another terminology which seems isolated m Aristotle's works
but seems to throw hght on his meanmg Animal m reference
to Socrates 1s said to be 'of a subject ', 6 whereas white 1s 'm
a subject · The prepos1t10n 'of' tells us nothmg, for 1t would
be true to sc1y ' wh1te(ness) 1s asserted of Socrates ', 7 but the
second phrase \\ 1th the prcpos1t1on ' m ' 1s suggestive and we
may mfer that probably ',,h1teness' is thought of as m
Socrates, whereas animal or an animal 8 1s somehow 1dent1fied
with Socrates We should then probably understand the first
phrase as ' of a subject qua subject ' 9
1 See chs 16 and 17, infra • :f, la,-1• /i.,,,p (,;.Iv,.,
1 :I ff7Tlv llrrt p A•u.-ov ,.,, • ,-c} .-al'l,.,,o,u,p.lvoo cw Tel -ylvos cw Tel 1730,
1 ""tAS•/3'1,ws, 1 uB' lnro.-11plvo11 ' ,-c) A111.-c\11 ,w:rv,op,ira, mTcl roii :I
1 (;lov CW (;l6v ,., ,

• ,,nfl' liwo.-flplroo J lnro.-dp,vcw [For the passages referred to above, see

I 7.! l
 The Copula and Modality 223

The real distmction seems to be that the universal which gives

rise to what Aristotle calls a genus is that umversal of which
the particular is what he would call a complete reality (the
md1v1dual substance), while the other umversals, which are not
allowed the same rank, are simply universals of which the
particular ts not a complete reality but an attribute of
a complete reality, that is of an md1v1dual substance.
Thus the particular of whiteness ts not ' a white obJect ',
but the particular whiteness of this particular substance.
Aristotle hardly seems to have realized clearly that thts
was the rationale of the d1stmction, but It IS mdtcated
by his statement m the Categories about the relation of
' second ' and ' fir-;t ' essences. True genera and species are
there termed ' second ' essences because, d.S he says, they
are the essences in \1\,h1ch are the first essences, namely the
md1v1dual substanres
§ 89 '1 he principles developed may now be applied to the
d1scuss10n of some- disputed pomts which relate to statement m
general Smee the time of Kant the ge-neral theory of the
propos1t1on has commonly been discussed under the four heads
of Quantity, Quality, Relat10n and Modality It ts tlus last
head with which, as arismg naturally out of the d1scuss10n of
the verb ' to he ' m the symbolic form A 1s B, we will begm.
By the modality of the copula 1s generally understood a mod1-
ficat10n mtroduced mto the so-called copula ' 1s ' m the form
S 1s P It ought therefore to mclude tense--mod1ficat1on , but
what 1s usually understood 1s the d1vis1on mto ' 1s ', ' may be ',
and ' must be ' , mto what are called assertoric, problematic,
and apode1ct1c propositions respectively. This d1stmct10n of
propositions 1s already indicated m the Prior Analytics, though
the above nomenclature 1s not to be found there In Aristotle
we find that the theory of the syllogism follows this d1v1s1on.
He does not discuss assertoric syllogism merely, but treats m
considerable detail the form which the conclus10n may take
when one premiss, or each premiss, ts problemat1r, or apodeictic.
The usual quest10n which logicians raise about modality 1s
whether this threefold difference should be regarded as belongmg
to the ' copula ', or whether It should be placed m the ' pre-
dicate ', predicate bemg here used m the improper sense already
224 STATEMENT, THINKING, AND APPREHENSION
crittcized. 1 The question then is not whether this transference
of the d1stmction to the predicate can always take place, but
whether 1t is right or not that it should. The most adequate
answer must depend on the view which we have taken of the
copula m the given form, but we shall find that we can m fact
decide the question merely by reference to what 1s ordinarily
meant by subJect and predicate m the doctrine under review.
§ 90. We have mamtamed that the copula means existence;
if therefore 1t 1s d1stmgu1shed mto ' 1s ', ' may be 1 1 and 'must
be', trat would he an obJect1ve d1stmcllon between bemg,
possible l.,emg, and necessary bemg Can such a distinction be
objective~ In any case there 1s one prehmmary obJectton.
We cannot co-ordmate mere bemg with possible and necessary
being mere hcmg would only be mtelhgtblc as the abstract
umverc;al of wluch pos!\1blc and necessary bemg would have to
be the species But, takmg the d1stmcllon as 1t stands, consider
the problematic statement ' S may be P ' What do we mean
when \\-C call an obJcct1ve reality possible? We do not of course
mean that 1t 1s necessary, for the statement ' S may be P '
1mphcs that S may not be P Now, when we thmk of a pos-
s1b1hty as obJecttve, v.e ordm,mly suppose th.it there are certam
co11d1t1ons e::,...1stmg which are necessary to a given result, but
arc not the whole of the necessary cond1t10m, These realized
cond1t1011s appear as a potentiality The development, however,
of the potcnt1ahty depends upon the reahzatlon of the remaining
conditions Whether these arc to be realized or not is a matter
of nece.,s1ty wlnrh cannot be mdetcrmmate If the unrealized
cond1t1ons arc not gomg to he realized, then thcre 1s no possibility
at all. In fact, that only 1s pm,sible which 1s gomg to happen
and 1s m that sense necessary also on the obJecbve side. There
is no objective d1stmctton between possible and necessary being
except 111 a modified sense, easily understood. Hence if the
'copula' denotes obJective bemg, it cannot be dist10guished
into 'may be' and • must be', and the d1stmctJon mto apo-
deict1c and problematic cannot belong to the copula. Similarly
we cannot d1stmgu1sh asscrtonc and apode1ctlc, because what-
ever is, 1s necessarily what 1t 1s
• In the followtn~. except where otherw1~e stated, we shall use • predicate •
as simply meanmg P m the grammatical form S 1s p
 The Copula and Modality 225
§ 91. Between these three forms of statement, however, there
is obviously a real distinction, for we do not use them indis-
criminately. The truth seems to be that they express in their
ordinary use a subJective distinction, a difference not in rcahty 1
but in the completeness of our knowledge. The form ' S may
be P ' according to common usage generally represents the
lowest degree of knowledge ; 1t sigm.fies that, while we do not
know that S 1s certainly P, we either know that some cond1t1ons
f:wourable to 1t exist, or, at least, that we know of nothing to
the contrary. With an increased knowledge of such conditions
our belief may be, as we say, stronger, and we represent this
by the phrase ' S 1s probably P '. Agam, as to the difference
between assertonc and apode1ctic proposit10ns, ' S must be P '
1s the natural form to use when we know the cond1t1ons which
necessitate the fact stated, whereas ' S is P' corresponds to
the case where we arc sure of the fact but do not know its
reasons and thus corresponds generally to what are called
(inaccurately) 'Judgements' of perception. It 1s true that
a man who believes that ' S is P ' on the evidence of perception
1s entitled to say that 'S must be P ', but as a matter of fact
we do not express ourselves m this way unless we know not
only that there 1s a necessity but also m what that necessity
consists. The d1stmct10n then 1s subJectivc, referring to a dif-
ference not in thmgs but m our attitude towards them.
§ 92. This same conclusion may be reached simply, without
any appeal to an obJecbve theory of the copula. The modal
d1stmcbon of propositions implies that, m the assertor1c, pro-
blematic and apodeicbc forms alike, S is subJect and P predicate,
the difference m the propos1tlons lying m the different forms of
the copula On the view before us the predicate, always assumed
to be Pm the form 'S 1s P ', 1s 'what 1s asserted of the subject',
assumed to be Sm the aforesaid form. Now, obviously, in every
assertion something must be definitely asserted of somethmg
else or there 1s no assertion at all, and the true predicate must
be definitely asserted of the subject. If, then, we arc thmking
about two clements of reality Sand P, so long as we arc uncertam
whether P can be ' asserted of ' S, we cannot relate them to one
another as true sub1ect and true predicate m the present sense
of predicate. Thus, in ' S may be P ', 1£ P and S arc true
2773•1 Q
226 STATEMENT, THINKING, AND APPREHENSION
predicate and subject, P must be ' asserted of ' S, and the
proposition must mean ' S is P ', which it docs not. Thus, if
S is true subject, P cannot be true predicate, and conversely.
To find out what is subject and predicate we must ask what is
definitely asserted, and that, obv10usly, is not that S is P ; and
so P is not what is definitely asserted of S Indeed, we are
doubtful whether S is P, though we know nothmg agamst it.
Now it is tlus fact, mvolvmg something about our own state,
that the propos1t10n definitely affirmc; To express this more
dearly the statement may be put mto such forms as the fol-
lowing -· that S 1s P 1s uncertain', or 'probable', or 'my
knowledge nf S 1s such that I am uncertain whether 1t 1s P or
not '. What 1s taken for the true logical subJcct in the analysis
depcnrh on the context, that 1s 1t may or may not be S in the
above examples, but 1t 1s necc&s,iry m any analysis to make
the poss1b1hty a subJcrtlve one Our ronclus10n must be that
the doctrine that the modal <l1stmct1on belongs to the copula
1s erroneous, because S and P cannot be treated as at the same
time respectively subject and predicate in the three forms of
propos1t1on S 1s P, S m,1y l.,c P, S must be P We observe also
that the word ' 1s ' rdam, 1tc; proper force 111 whatever form
the true analysis ts gwcn, and that the ' affcct10n of the copula '
appears as a rule in the predicate, 1£ we mean by predicate the
true logical predicate, but that under some circumstances 1t may
appear m the logical suhJect.
§ 93. Bes1clcs the cl1stmction of posi;,1b1hty, actuality and
necessity, there 1s another which has sometimes been supposed
to belong to the copula, that of past, present, and future time,
sometimes called a km<l of modality \Ve have mamtamed that
wlulc the word ' 1s ' m ' S 1s P ' rcpnscnts hcmg m general
for the subJect, the ' prcchcate' cxprct>scs a special detcrmma•
t1on of bemg \\ luch 1s ac;s1gncd to the suhJect m the statement
Thus the quest10n whether the time d1!:>t111ctlon 1s to be put mto
the copula takes for us the form whether 1t 1s to be put mto the
general bcmg or mto the spcual bcmg represented by the subject
(nommativc) and attributive Now, temporal d1slmctlons are
obviously specific detcrmmations of bcmg, and so, according to
our theory, their place must be m the ' subject ' or ' predicate '
for the same reason that any other special determmatlon of
 The Copula and Modality
being must be there. If then we were to exclude temporal
distinctions from the predicate, we should have also to exclude
such modifications as adverbs of place. Our answer then must
be that as, in this particular form ' S 1s P ', the verb ' is '
stands for the bemg m general of the subject, whether past,
present, or future, the temporal modifications cannot be put
into 1t How then are we to meet the hnguist1c difficulty which
presents itself? Prima facie the tense of the word ' 1s ' 1s present
and therefore docs seem to refer to time. How can we maintain
that 1t 1s without tense, for our conclus1on commits us to that ~
The difficulty is unreal and may be removed as follows. The
general idea of bemg 1s obv10usly wider than the temporal
d1stinct1ons, for it 1s something common to all three times,
present, past and future, in so far as all three are opposed to
complete non-existence. This general notion which transcends
the d1stmctions of time is, nevertheless, accordmg to established
usage expressed by a special 1d1om, the present tense of the
verb ' to be ' So, m fam1har speech we use the word ' 1s '
actually of a series of events considered as a whole, though, of
course, the series as a whole never exists m any one moment
of time which we could call present Thus we say, 'the Eliza-
bethan period 1s one of the finest periods m English literature'.
So of something wluch 1s past we use the word ' is ' with no
sense of domg v10Icncc to its meaning, ' Socrates is the most
stnkmg figure m the lustory of Greek philosophy '. In the very
common phrase 'so and so 1s a thmg of the past', the present
tense ' 1s ' represents bemg generally, while the past, or bemg
in the past, 1s put as a specific form of this general being. The
same present tense 1s 1d1omatically applied to actualities not
really determined m time, inasmuch as they do not happen or
elapse and their existence cannot be represented as a time pro-
cess, actualities for which there is properly no past, present or
future, hke space itself, which 1s nothing that either happens or
elapses. Agam we use the same idiom of laws of nature, which
are not events but universals.
We see, then, that tlus general abstract meaning of ' is ',
this tenseless aspect of 1t, is not a logical convention nor an
artificiality but embodies the usage of livmg speech That is
why such a word as ' 1s ', though apparently refer rmg to
Q2
228 STATEMENT, THINKING, AND APPREHENSION
a particular time, could come to be selected as a part of the
general form in which every statement can be expressed. 1
NoTE. In Greek, thought sometimes attempted to free itself
from the limitation of time by using the past tense for something
not conceived as over and gone but as still in existence. Thus
the imperfect ,jv (lit. was) 1s used 1diomat1cally to express some•
thing generally vahd Thus Aristotle uses for the essence Tl fan
(what 1s) the expression Tu Tl 1j1 ti1•ai. One 1s at first templed
1

to conJcclurc that the contradiction of the tense time to the

prec;cnt n ahty to which 1t refers enables 1t to serve as a sort
of symbol of the abstraction of time altogether. Thts is however
too artificia 1 and we must sec in the formula an instance of an
idiom familiar both m Greek and English The past tense serves
as a ~ort of oratio obliqua. It does not mean that the quality
signified by the verb belonged to its subJCCt in the past, but
that in the past we stated and agreed that 1t so belonged. In
English we might say, for instance, but a right was something
absolut<> ', mcamng we agreed that it is. In German, the
narrative effect 1s often given by the subjun<'hve, as in ' Nun
aber ein Recht ware was absolutes ' The latter is a favourite
idiom for reporting the words of another person in ordinary
oratio obliqua. ' lhr Bruder ware unschuldig ' would mean, in
some contexts · ' She i;a1d her brother was innocent.' Thus m
Greek and English 1d1om the past tense may md1catc something
wluch is understood or taken for granted. Remembering then
the idiom by which Tl Ean (what 1s) 1s ' logically ' equivalent
to o fern (which 1s), Tl ,jv may mean what the thing 1s under-
stood to be-what 1t 1s decided to be.
§ 94. Ccrtam abbreviated forms of expression, utterances like
' it rains' or ' Fire I ' arc rightly seen either to be or to
involve statements, nnd the absence of a definite grammatical
nominative, or rather the absmce of an C'xprcss d1stmcbon be-
tween subJect and attributive in the one case, and the indefimte-
ness of the sub1ect of attribut10n m the other, have seemed to
favour the view that in such statements reality m general is the
subject. Yet the propos1t10n ' it rams ' is understood to be not
a statement that raming 1s somethmg which occurs in general,
• §§ 77 (p 182} and 84,
 The Copula anel M oelality 229
but that it is happening here and now. For a person saying
' it rains I ' would not be understood to refer to the state of
the weather in a distant country. It is therefore a definite
statement of a fact and admits of analysis like any other state•
ment. For instance, 1f it is an answer to the question, 'What
is the weather bke ? ', the subject is 'weather'. ' Fire I'
involves a statement that a fire 1s actually taking place, but
not that only. It has a peculiar form of cxpress10n, for of course
into express10n enter tone and gesture. These accessories are
not represented in the printed or written word ' Fire ! ' and
arc but imperfectly suggested by the note of interJect1on.
Simple as 1t looks, it 1s actually a c-omphcated form. The tone
and gesture arc not appropriate to the mere judgement wluch
the person spec1king has formed for lumsclf, that there 1s a fire
somewhere. His utterance not only implies this, but by tone
and gesture also gives a warning. The accessories then deter-
mine that it 1s not a case merely of a fire somewhere m general,
but of one somewhere near, which interests us and is important
for us to know about. Moreover, the warnmg accent normally
conveys the idea that the fire is dangerous to us. If this were
not the case, the utterance would have a different tone and an
explanation would follow of where the fire was ; but more
probably the mtcrJcct1on form would noL be used at all. If
the speaker says no more than ' Fire ! ' the 1mphc..1L10n of his
tone may be that 1t is dangerous to us
Hence the matter 1s complex, and the statement mvolvcd may
be analysed as follows. We see a man hurriedly and excitedly
advancing and we believe that he has nnportant news. Tlus
then 1s Lhc log1c,1l subJcct for us and lus 111tcrject10n gives the
predicate. If we imagine a listener who has not seen him, there
may be a logical subject and predicate differing from one such
listener to another For one, the idea of fire may come first
and he may 1mmed1atcly afterwards realize the tone of alarm ;
with another perhaps, the appreciation of alarm m the sound
may come first ; and in that case ' something alarming ' 1s his
subJect. Lastly, we must not exclude the poss1b1bty (m the
case of a listener aroused by the sound only) that an order in
time between the clements does not occur, but that they arc
apprehended simultaneously with no time distinguishable. Then
230 STATEMENT, THINKING, AND APPREHENSION

neither element can be prerogative as against the other, and the

statement so communicated or understood has not properly
a logical subject or predicate at all, a contingency already
provided for. 1
In the above and.lys1s we have considered what Judgements
or opinions the exclamation ' fire ! ' implies as formed by the
person uttering it and by those who hsten to it. But the word
is not intended merely to express Judgement and opmion and to
communicate information , its main use is to convey a warning
or exhortat10n, and this is not a judgement nor an opmion.
'flus lead,; naturally to the general quest10n as to the logical
analysis of a wish, command or entreaty. By some these have
been treated as Judgements or propositions, and for the purpose
of logical analysis such a sentence as ' may you succeed ' is
ccms1dcn,d as equivalent to ' that you should succeed is what
I wish ' ; • put that down ' agam 1s made equivalent to ' that
you should put th.it dov. n 1s wh.Lt I command ' Now tl11s
means that such verbal cxpress10ns arc to be taken as state-
ments of facts, ansmg from an apprehcns10n of these facts or
from an opinion about them But this 1s a mistaken view We
must ask what the words arc intended to express, for the mtendcd
mcanmg can alone decide the qurst1011 The one sentence
expresses a wish and the other a command. But a wish 1s not
the appr('hens10n of the nature of a fact, whether the apprc-
hl'm,1011 1s ,L Judgl'llu·nt ur not, nor 1s 1t an opm10n about a fact.
'l he :::.,Lme t<, true of t lil' command A qul'st10n .ilso 1s not
the cxpn c;s11111 o( l,mJ\\ kclgc or of opuuon The ho;tl'ncr to the
question ' ,,.h,lt doc!> tlm, v. c1gh ~ ' <loubtlc,1:,1:, bcl1cvc:, that the
questioner desires mformation, and some "ould say that he
Judges this; but that belief or Judgement 1s not the question,
any more than the qucstlonmg attitude can be the Judging or
behcvmg attitude. E:xprcss1011s of surpnse, agam, and of emo-
tion gcncr,\lly may be treated on the same pnnc1ple
I § 8o,
 XI
SYNTHETICAL AND ANALYTICAL ' JUDGEMENTS'.
THE RELATION OF PROPOSITIONS a

§ 95.b THE doctrme that in 1ts proper logical sense the pre-
dicate m a given Judgement or opimon 1s a new determination
of the sub1ect1 as previously conceived by the person Judging,
leads us to consider a trad1t10nal distinction in modern logic to
which 1t seems opposed, the d1stinct10n of analytical and syn•
thctical judgements tAn analytical Judgement is generally
defined as one m winch the conception of the ' predicate' is
already contained in the conception of the 'subJect ':'} A syn-
thetical Judgement 1s one in winch the conception of the
'subJcct' docs not contain the conception of the 'predicate'.
Now 1£ predicate is taken in what seems its proper sense m
logic, 1t would follow that there could be no such thing as an
analytical judgement
It must be admitted that m a sense the conception of the
sub1ect always rnvolvcs that of tht• predicate, m so far as a thmg
necessarily has of 1lo;; own nature the properties which we rightly
a.!i!i1gn to 1t But that 1s not what 1s meant by an analytical
judgement m the view \vC arc to examine, and for accuracy the
dcfimt10n !ihould be restated as follows . ' an analytical Judge•
ment 1s one of which the predicate 1s contamc<l cxphc1tly in the
subject as the subject 1s conceived by the person Judgmg; m
a synthebcal judgement the predicate 1s new at least to the
person Judging.' If then 1t 1s the essence of a ' Judgement '
(" From tlus point onwards 1t has been difficult to make the termmology
con!.1stcnt Th1!. chapter, for instance, was complet<•d fairly early and I have
left the language ~ubstanbally as 1t stood m the prmted ver~mn of 1913.
The author, 1f ht. h,ul uhted 1t, would have tri<'d to avoid txpreS&10ns hke
' conception of the predicate ' and the use of the word ' Judgement ' m 1ts
current sen&e, not h1i. own He had however usually substituted the word
' statement ' for the verbal exprei.s1on wluch would ordmanly be called
'propos1t10n • Seep 235, note 1
b Cf Lotze, Logic, m 5, t§ 363-4]
232 STATEMENT, THINKING, AND APPREHENSION
which has a logical subject and predicate to give a new deter•
mination of its subject in the predicate, an analytical Judgement
as just defined is a contradiction in terms. It has however been
gravely maintained that every synthctical judgement inevitably
becomes analytical. In a synthetical Judgement of the form
(s)iAB is C, C is the new clement and therefore the Judgement
is synthet1cal ; but it 1s mamtamed that, after the act of judge•
ment, C becomes an element m (s) 1 or rather an clement in the
concept1cn of (s) 1 for us. Thus ' (s)i 1s C' 1s now analytical, is
in fact equivalent to ' (s)iABC is C '.
Now, m the first place, even ,vere this so, the analytical
Judgement would seem to be postenor to the synthetical m
order of thought, for it depends entirely on the synthesis which
precedes 1t. There cannot be analysis unless there has been
synthesis. Secondly, we can show that 1t is an 1llus1on to
suppose that a synthctical Judgement can ever lose its character
of synthesis and pass mto the analyt1..:al m the way supposed.
When would the synthet1cal Judgement c:,,.1st at all ~ It would
follow that the predicate the moment 1t was assigned to the sub-
ject became part of the subJcct, and the Judgement would be
mstantaneously analytical Let us however waive the difficulty
and suppm,c, for argument's sake, that the synthesis takes place
first, and th,1t the Judgement becomes analytical on repet1t1on,
tlus 1s probably \\hat 1s mtcndcd, the temporal d1fiiculty havmg
escaped notice altogether. Now under v. hat c1rcumstanccs 1s
such a judgement repeated ? One obvious reason 1s to mform
some one else, who only knows that (s) 1 1s AB. For him the
Judgement (s)iAil 1s C 1s clearly synthet1cal But "hat of the
origmal author "hen he repeats (s)i 1s C ? Ile docs not mean
that the Judgement 1s vahd because of the identity bet\\een the
C m the predicate and the C he has already thought of as in
the subJect. On the contrary the statement 1s of no mterest
except as a repetlt10n of the ongmal ' synthesis '. It 1s not
mdeed a new truth to the person so Judging, but that does not
make 1t analytical ; and thus he \\ ould give as the reason
why (sh 1s C, not that (sh already contams C but whatever
reason had been found before for Jommg C to something
different from itself m (sh, viz AB. The importance then
of the l>tJ.tcmcnt "hen repeated hes Just m this synthesis
 Synthetical and Analytical Statements 233
of C with elements different from itself, not in an equation of
C to itself.
If we msist on our previous obJection that a Judgement, if
it could become analytical at all, must do that mstantaneously,
then we could not really say it became analytical, but that it
always was both ; we could not say that the judgement had
lost its original synthctical character, but rather that it never
had a purely synthetical character. But such a change of state-
ment would be met precisely as before , for it would still be
true to say that the meaning of the Judgement as such lies in
the associat10n of C with elements other than itself, and never
m an analysis
§ 96. A difficulty may arise from the fact that when we repeat
' (sh 1s C ' we do not always think of the nature of the synthesis
of C with A and D. We seem at times only to remember that
(sh is C, and this may have the appearance of a mere analysis
of a conception m our mmds of (sh as ABC. This difficulty
1s unreal ; what we arc domg m such a case is not to make an
analysis, but to remember that there was a synthesis, though
for the moment we are not attendmg to its nature or may have
forgotten it. In a mathematical demonstration, for mstance, we
often use a formula without rcmembermg the process by which
1t was reached But the use of 1t 1s entirely due to its synthetical
character, and 1t 1s only thus that 1t can be any help 111 our
demonstration No one can get anythmg out of a statement
of the form C 1s C. Suppose that we wanted to prove (s) 1ABC
1s D, • and that we established tlus by showmg that D attaches
to C. Then the last stage of our argument has the form C 1s
D, but (s) 1 1s C, therefore (sh 1s D. Now as the conception of
(sh, as ABC, 1s presupposed, the premiss (~h is C may look like
a mere analyi,1s, but what we really do in tlus proof is to pick
out the attribute Cness from among the other known attnbutes
of (s)i as the one which 1s responsible for D. This shows the
truly synthet1cal character of the premiss we were usmg. We
should not have picked out this attribute at all unless it was
to be distinguished from the others, and that shows we were
thmking of (sh as somcthmg else besides C; (sh 1s not only
[• In the lectures the dlustrallon U&ed here wai. from ' the opposite angles
of a quadnlateral mscnbed m a circle are together equal to two right angles • ]
234 STATEMENT, TIIINKING, AND APPREHENSION

AB, but also C. Thus the interest of the ' judgement ' lies in
the synthesis of the different elements A, B, C. This comes out
clearly if, in that last stage of the argument (C is D, but (s) 1 is
C, therefore (sh is D), we replace the statement (sh is C, by
C is C. This shows us at once that the mterest of repeating
(sh 1s C hes not in any identity between C and (sh, but m some
d1fference. What we really want to say 1s that that which is
AB as well as C 1s also D, because Cncss involves Dness. This
1s equally true whether the presupposed conception (s)iABC is
the rcs•Jlt of a demonstrat10n that Ail necessitates C, or JS an
und1.,monstrated defimtion with which m a particular science
we begm. For the dcfimt1on itr,clf has only mc.inmg, interest,
and use because 1t 1s itself a synthesis of d1fTcrcnt elements.
§ 97. There 1s one general consideration which covers all cases.
If ' (sh JS C ', when repeated, were really an analytical Judge•
ment, 1t would be of the form (s) 1ABC 1s C, where C 1s exphe1t
both m subject ,.ind predicate, the truth of the Judgement would
then he 111 the Jclent1fic,it10n of C w1U1 1tc;clf, 111 other words,
the valt<l1ty of the Jt1<lgemrnt would be clc11ved entirely from
the Jdcnt1ra.l Judgement 'C 1s C '. Dut 1t 1s evident that the
so-called identical Judgement v10latcs the very idea of Judging.
There 1s no such thmg except as a verbal form It JS nothing
for thought, and, inasmuch as the analyt1cal thus rea11y
resolves itself mto an Jdcnt1cal Judgement, we may sc1y of Jt
.\h,o that 1t exists only as a form of wor<lr, and 1s nothing for
thought.
If now \\C give to 'r,ubJCl.t' ..md 'predicate' the mcanmg they
un<loubte<lly h.1, c 111 the or<lmary trc.itmcnt of the syllog1sm,
the subJcct bcmg \\hat 1s denoted by the nominative case to
the verb ' to be', au<l the prc<l1cate the adJcct1val phrase or
noun phr.ise winch follo\\s 1t (.1 <l1stmct10n we have designated
as that of subject and .1ttnbut1vc), the same general argument
"ill apply. For 1t ,viii be evident from what has been said of
the d1stmctlon that the purpoc;e of the attnbutivc 1s to express
for the subject an aspect of 1ls bC"mg chffl·rcnt from that expressed
m the nomm.itivc case The same agam holds m sentences
where the prmc1p.1l verb 1s not the verb 'to be', 1f the term
' predicate ' stands for the attnbutivc, and the attributive con•
sJsts of the verb or verb phrase attached to the nommative.
 Synthetical and Analytical Statements 235
In conclusion then, the repetition of a synthetical judgement
has its meaning and value as recalling an original synthesis: its
meaning never is that the predicate is attached to the subject
because it 1s already there. The doctrine that a synthetical
passes mto an analytical Judgement is a mere confusion, and
strictly speaking there 1s no such thing as an analytical judge-
ment.
§ 98 • So far we have implied as the general form of every
affirmative statement, what 1s commonly calkd the categorical
form (m the ordmary notation, S 1s P), where a so-called pre-
dicate 1 P 1s absolutely affirmed of a subject S, and 1t has been
mamtamed that tlus uncond1t10nal affirmat10n is the necessary
characteristic of statement as such D1stmgmshed however from
the categorical we find m logic another form recogmzed, the
hypothetical, and 1t would of ten be said that by this 1s meant
a form in which the predicate 1s not absolutely affirmed of the
subject, but only under a ccrtam cond1t1on. If so, there would
be a form of statement m which the true predicate was not
absolutely affirmed of the true subject Tlus would seem to
contradict the account we have given, and we must therefore
consider the value of tlus currrnt d1stmct10n and ask whether
we must revise our account of statement so as to make the
hypothetical form co-ordmatc with the categorical.
We may take as the general form of an hypothetical propos1-
tl(m, m,mg the ordmary notation, 'S 1s P, if a certam cond1t10n
1s n·<th.ted ', or ' S 1~ P, 1f X 1s Y ', \\ her<' S 1s P I'> called the
wnscquent ,Llld the ' 1f ' cl..iusc the ..intcxcdcnt Now 1f we
confine our~dves to the idea of a cond1t10n, 1t 1s not necessary
to express a cond1t10n m an hypothetical form at all, for the
ordmary categorical statement of the form S ( =ABC) 1s P, really
' In the ordinary dt•fimtious of the catcgoncal and hypothetical proposi-
tions, the terms 'bUbJect • ,md •predicate• always have the meaning they
h.i.vc in the theory of the 1,yllog1sm, and denote not the logical 1,ub1ect and
predicate, but what we have called • i.ub1cct • and ' attributive • In the
following discussion the terms •subject• and •predicate' will be retained on
the underi,tandmg that tht.y h.i.vc Uub me.i.mng, because the theones examined
arc expressed m thu, terminology For the same reason the oi-dmary sym-
bohsm for the propob1bon will be used

[• Kant based the d1v1btOn on the Category of Relation Kritik d r V,

1st ed, p 178, Logic, § 6o Cf infra, § 102)
236 STATEMENT, THINKING, AND APPREHENSION
ascribes P to AB under the condition C. This is evident when
we compare AB is Q with ABC is P ; the contrast of the two
implying that, whereas we attach Q unconditionally to AB in
the second case, we attach P to AB only when qualified by C.
In this way the categorical form will yield a conditional form,
for from ' ABC is P ' we have ' AB is P under the condition
C ', and from that it might seem, though erroneously, that we
could derive the hypothetical form 'AB is P 1£ AD 1s C '. In
short, it the subJcct of a categorical statement is complex, as
ABC, any member of the complex, or any combmatton of
members, can be taken as subject and the rcmammg members
represented as the condition under which the old predicate is
attached to the new subJcct. Or we may take the whole subject
conception for predicate m the antecedent , thus from ' all
ABC is P ' ,vc should get ' If anything 1s ABC, 1t is P '. The
umversal categorical form thus connected with the cond1bonal
and hypothct1cal would usually be s,.ud to be capable of reduc-
tion to an hypothetical form It 1s mdecd mamtame<l by some
logicians that the hypothetical 1s the truer anrl more accurate
form not only of the umvcrsal but l vcn of the particular pro-
pos1t1on. We shall mamt:1111 1 on the contrary, first that the
categorical cannot be reduced at all to the hypothetical form,
much less be replaced by 1t, because the hypothetical cannot
express the full meaning of the statement wluch 1t 1s to replace,
and, secondly, that the hypothetical 1s not the true form m an
ultimate analysis of the statement or propos1twn, Mncc 1t can
be admitted to be a statement at all only 111 so far as 1t 1s itself
categorical. Thus "e shall confirm the view already taken, in
the d1scus!:i1on of modality, that all st,tlcmcnt as such 1s cate•
goncal, ..mtl den.lop from this the pos1t1on that what is really
to be contrasted with hypothcll1.al Juclgcmcnt or opm1on is
non-hypothetical statement, both bcmg species of categorical
statement.
§ 99. We have seen that ma sense an hypothetical statement
can be derived from a umvcrsal categorical. Nevertheless this
hypothetical cannot express the full meanmg of the categorical
form for wluch 1t 1s substituted, and the categorical therefore
cannot strictly be said to be reduced to 1t. If we compare ' all
A 1s B' with '1£ a thmg 1s A, 1t 1s B ', 1t 1s true that the first
 Categorical and Hypothetical Statements 237
involves the second, but normally something more is expressed
by the categorical. The question is a linguistic one and can
only be answered by investigating the normal habit of a par-
ticular language. When in English we say ' all A 1s B ', we
usually imply that there are such things as A, that the condition
' 1f anything 1s A ' 1s realui:ed. By contrast the normal 1mphca-
tion of the hypothetical form is to leave the reahzat1on of this
cond1t1on undecided. Consequently, when we are not sure that
the condition is realized, we avotd the categorical form and
choose the hypothetical. Otherwise we might cl,llm to tell the
truth 1f we said 'all AB 1s C' though we knew well that no
AB 1s C, for we might contend that the statement means m
strict accuracy ' 1£ there arc any A wluch are B, they arc C ',
and tlus we might even m some cases justify by the admitted
fact that all ll arc C. Thus all we say 1s ' 1£ there arc any
AB, they are C ', while really no A 1s B.
Clearly in common hfc in such cases we should be said to be
either jCStmg or dece1vmg. And in fact a number of elementary
jCsts arc founded upon the circumstance that, while the normal
meamng, apart from any context, of an affirmative statement
ts that its subject exists, there arc exccpt10ns arismg from certam
well-understood 1d10ms which cause no ambtgmty. Thus no one
1s deceived by the statement ' when the hon on the Royal
Exchange hears the dock strike one, he wags his tail '. ' Tres-
passers will be prosecuted ' has been seriously alleged to show
the hypothetical character of the categorical formula,• on the
ground that the notice, so far from implying that there will be
trespassers, is just put up to prevent there being any. But this
1s a special 1d1om, applymg to general statements which arc not
umvcrsal m the true scientific sense, and any generalization
from 1t may be refuted by mstanccs of the same kmd of expres-
sion where the existence of the subject 1s obviously taken for
granted. Thus the noltce 'All trespassers have been pro-
secuted' ts understood to mean that there have been trespassers.
In the case of categorical statements umvcrsal in the true
scientific sense no examples could be found which were merely
hypothetical.
[• A reference, apparently, to Bradley, Logic•, I, p 48, and I, II, § 6, cf.
Bosanquet, Knowledge and Reality, I, § 8)
238 STATEMENT, THINKINGJ AND APPREHENSION
On the other hand, just as the categorical statement in a
certain 1d1om may represent an hypothetical, so also the hypo-
thetical may 1d1omatically represent the categorical Thus when
1t 1s said that if pot:issmm 1s put into water, there 1s combustion,
tlus 1s really undc..rstood as eqmvalent to ' whenever potas-
~mm &c ', which certainly implies that the cond1t1on can be
and has Leen realized Acrordmgly the hypothetical form is
used in certain familiar 1d10ms, \\hen 1t 1s understood well
enough that the rond1t1on can be realized. This 1s very common
for rn~tance m geometry m such express10ns as '1f one side of
a triangle be produc..cd ', and comes probably from thmkmg
of the construction as depcndmg on the reader's will to make
1t The enunciat1011 too of expcmnents m physical scicnce 1s
often expressed in a similar way Normally the hypothetical
statement puts the cond1t10n "'h1ch 1t expresses m such a way
as to make 1t understood that the rcal1zat10n of the condition
is left quite in<letermmate
§ 100 Tlus fuud.tmcntal d1ffrrcncc- bct,,ccn hypothetical and
categorical statement 1s seen still more clearly m the singular
eategonral statement, \\luch 1s of the form 'tlus ABC 1s P '.
If we apply to tlm the method by which an hypothetical is
extracted from the ratcgoncal, \\<C shall get this kmcl of result.-
' 1f any A 1s tl11s UC, 1t 1s P ', or, 'If tlus A 1s tlus BC, 1t 1s P •.
Such unnatural forms of course never appear . their absurdity hes
m putt mg mto a given form a matter which contradicts that form
Consider the form ' 1f any A 1s this BC, 1t 1s P '. This has an
antecedl·nt ,,1th a general subJect Dut now when the subject
of such a clause 1& general, 1t 1s 1mphcd that the predicate of
that clause nught attach to several subjects of the same kmd.
' If any metal has such and such qualities, 1t ,ull attract iron ',
relates the predicate to the class of metals m general. The
form, ho\\ ewr, ' 1£ any A 1s tluc; BC ' contradicts this by
puttmg as 1f grncr.tl \\ hat the mat tcr of the antecedent shows
to be only particular, for the con<l1t10n docs not relate to any
A but only to the particular A wluch 1s this BC. In an 1llustra•
lion the absurdity of the eontrad1ction between the matter and
the form is obvious. Thus ' tlus heavy grey piece of metal
attracts iron•, becomes ',my heavy body attracts iron 1£ 1t 1s
tlus grey piece of metal '.
 Categorical and Hypothetical Statements 239
Further, ' if any A is this BC, it is P ', may be reduced to
• if any A is this A, it is P ', since the individual which is this
BC is by assumption exactly this A. But since the hypothetical
statement as much as the categorical implies a distinction of
the clements called its subject and predicate, the identity m the
antecedent as thus expressed 1s clearly as nugatory as it is m
the categorical form of statement Finally, and this 1s the
fundamental absurdity in this attempted reduction, we feel the
form to be contradictory because there 1s no doubt whether
some A 1s this A, or whether any A is this BC; for we can only
say that 1£ any A 1s this BC, it is P, because we know that this
BC which 1s P is A.
This leads us once more to an essential characteristic of
hypothetical statement. It must, as such, put its condition as
something not certainly lmmvn Thus the reduction of the singular
categorical to the hypothetical violates the very nature of hypo-
thetical statement ' If any A is this A ' cannot represent any-
tlung open to doubt; 1t 1s but the verbal form of an hypothesis
to "'h1ch no hypothetical thmking corresponds The same
crit1C1sm obv10usly applies to ' 1f tlus A 1s tlus BC, 1t is P '.
It 1s thus 1mposs1blc to put the matter of a singular categorical
proposition into an hypothet 1cal form, wl11ch so far vindicates
the categorical form as indepen<lcnt
The discussion has brought to hght an important aspect of
the hypothetical statement, namely, that 1t docs not merely
attach a quality to a subject under a condition, hut leaves 1t
open whether the cond1t10n is realized or not. That 1s the
normal mcanmg of '1f ', and by this meanmg we d1stmgu1sh
the true hypothetical statement from the umversal categorical.
§ IOI. We have seen already a ccrtam relation between the
hypothetical and categorical forms m virtue of which a true
hypothetical statement may follow from the truth of a cate•
gorical. How far arc these forms dependent on or independent
of one another ?
It is easy to sec that the hypothetical form presupposes the
categorical or else 1t could not be understood Thus the cate-
gorical is prior m form. For the expression ' If A is B, C 1s
D ' implies that the meaning and use of the forms 'A is B ' and
' C is D ' 1s already understood. The converse is not true. The
240 STATEMENT, THINKING, AND APPREHENSION

categorical form is intelligible in itself and does not for its employ-
ment presuppose that we already understand the hypothetical.
Does the hypothetical statement then not only presuppose
the categorical by its very form but itself also contain a cate-
gorical element ;, If 1t docs not, how can we call it a statement,
for it will not satisfy the cond1t1on which we have hitherto
assumed as obvious, namely, that there 1s no statement or pro-
position unkc;s c;omcthing 1s defimtely stated of something else.
Now in the form ' 1£ A 1s B, C 1s D ', B 1s not assigned dcfimtely
as the predicate of A, nor D of C Neither of these pairs then
constitutes the true subJcct and prechcatc which make these
clauses together one statement correspondmg to one Judgement,
opmion, &c. W c must ask what 1s decided m the act of thought
wluch corresponds to such a statement In all hypothetical
affirmative statl'ments the affirmation which must be there to
make a statement at all 1s that the consequent 1s the necessary
result of the antecedent. This 1s a categorical statement, one
in which the prechcatc 1s attached to the subJect without any
condition. Tlus 1s offcred not as an arbitrary or possible reduc-
tion of the hypothetical form, but as the statement of what it
is m the complex hypothetical sentence \\ luch really makes 1t
into one statement, rC'prcscntatlve of one Judgement, opmion, &c.
The hypothetical form then cannot claim to be a d1stmct form
co-ordmate \\1th the categoncal m an ultimate logical analysis.
The supposed ddicrcncc between them, as represented m that
dcfimtron of them with wluch we started, 1s due to a comparison
of hypothetical and categorical forms in which the subject and
predicate arc only m appearance the same. Just so the modal
distinction of propositions was, we saw, due to a comparison of
statements \\ hose subJcct and predicate were only m appearance
the same.
We may also sec that tlus 1s true by various modes of logical
analysis, if we take prcd1ratc m the sense of 'logical ' predicate.
Thus 1f our qucst10n had been about the nature of C, C would
be the logical subJect m the statement 'C 1s D if A is B ',
and we should affirm that C 1s such as to be D 1f the condition
holds that A is B; that is, we attach a complex predicate (m the
logical sense of predicate) uncond1t1onally to C. Or suppose the
context were such that the antecedent itself was the subject.
 Categorical and Hypothetical Statements a41
If the question were : ' what follows if A is B ? ', we should
m ' C is D, if A is B ' attach the consequent unconditionally
to the antecedent A is B, and the tone in which we pronounce
the clause 'C 1s D ', as contrasted with that in which we say
• A 1s B ', md1cates that 'C 1s D ' is the predicate.
§ 102. Hitherto we have for convenience described the ele•
mcnts m the hypothetical statement, which arc conceived as
necessarily connected, by help of the' grammatical forms wluch
rorrespond to them witlun the sentence, that is the antecedent
and the consequent But what do these exactly represent?
A common view is that m the statement, ' if A is D, C 1s D ',
a relat10n between two 'judgements' 1s affirmed, viz. that the
Judgement A 1s B ncccss1tates the Judgement C 1s D, and indeed
that 1s one way m which the hypothetical judgement is supposed
to be reduced to the categorical. This 1s entirely erroneous.
In the given form, A is B and C 1s D do not represent judge-
ments, simply because they arc not judged. We are uncertam
whether A 1s B, provided the particle ' 1£ ' is used in its normal
and proper meaning ; that is, when the statement 1s a true
hypothetical. So we do not Judge that A 1s B, and the words
'A ic; D ' (wlurh occur WC' observe not inclepC'ndcntly but
only m connexion with the ' 1£ ') do not represent a Judgement
at all Bemg uncertam then whether A 1s B, we are so far
uncertam whether C 1s D, and so the words ' C 1s D ' do not
represent a judgement either. If ' A 1s B ' really represented
a Judgement, we should be able to say, ex hypothesi, ' because
A is B, C 1s D '.
It 1s a quest10n or a problem to us whether A 1s B, and
a qucst10n whether C 1s D, and the hypothetical sentence states
a relat10n of a certam kmd between these problems, grounded
on our knowledge of the realities to which they relate The
statement 1s that the question whether A 1s B 1s a case of the
question whether C 1s D, and 1t may easily be shown that tlus
entirely accounts for the mfcrence which can be made from such
a statement. Or, to put 1t m a manner which accords more with
the actual cxpress10n, the form of statement (not the statement)
A 1s B necessitates the form C 1s D 1
1 This sub1ect, as well as that of the D1~1uncbve statement, 1s discussed

more fully 1n Part III, chs 5 and 6,

277J'l R
242 STATEMENT, THINKING, AND APPREHENSION
In the general form ' if A is B, C is D ', the hypothetical
clause does not refer to what 1s only hypothetical j there must
be m 1t clements known to be real. What is unknown and
indeterminate 1s whether they can stand in a certain relation,
for instance, A and n may be known to exist, but we may be
uncertain whether B can be attributed to A. Now the hypo-
thetical form bcromes possible because we assign definitely to
the hypothct1c,tl cond1t10n or antecedent the result or consequent
C 1s D How can such an antecedent acquire such a consequent
at all ? The answer 1s · only because \\ e can connect cate-
gorically, 1 c. non-hypothct1cally, with the clements whose
existence we arc sure of (A and B) some other elements of
reality, or, more accurately, can apprehend such a connexion.
And it is m virtue of tlus rclat10n of the known clements in the
hypothetical clause to ecrtam other known clements m reality
that we get the consequent as the result of the antecedent.
Suppose, for instancr, we know m reference to A and B, that
a necessitates A and that n nrccc;s1tatcs ,~. Then through that
knov.lcclge (v.h1ch 1s non-hypothetical) we arc able to say what
is the result 1/ A is 13, namely, that a 1s ,-J, and we can see that
the reasoning by wh1rh we get tlus consequent from the ante-
cedent (1f A 1s B, a 1s fJ) always consists m rccogmzmg the
necessary non-hypothetical consequences of real and non-
hypothetical clements.
Consider for ms tan cc the reductio ad absurdum proof. In this
1l 1s ' assumed ' 1 that two real elements stand in a certam
rcl.1tion. We then argue from what each of these real elements
separately necessitates and so get a statement hypothetical in
form, n,,mely th.1t, 1£ the real clements arc m the supposed
relat10n, certain other real elements must be in a certain rela-
tion. Now we know othern 1sc that these last cannot stand m
the given relation. Thus our assumption is seen to be 1mposs1blc.
Herc \\C notice that m a very important sense we do not argue
from an hypothetical conccpt10n at all (contrary to the usual
rcprcscntat10n) 1 2 for m accordance "1th the cltstmct10n "·c have
1 Tlus use o{ the word • a,-,umcd ' lb d1..,cu,!>cd 1n §§ 296 and 312

• It 1s not d1fhl.ult to show that the mam fallacy of• non-Euchdean space '
1s a complete m1bunderstandmg of h)pothcbcal tlunJ..mg and more especially
of the h)l>0thct1l.il • conccphon • Vidc Part III, ch 7 Another fallacy m
It I!> IIOli<-cd lll § J:?U,
 Categorical and Hypothetical Statements 243
already made 1t is not a conception m the proper sense, but
a problem or question, and the very result of our argument (the
reductio ad absurdum) 1s to show not merely that such a concep-
tion is obJectively mvahd but that we cannot have the conception
at all, since the supposed connexion 1s shown to be unthinkable.
Now we obv10usly cannot argue from what we have not got.
The hypothetical statement then 1s an mfcrcnce, and an
mfercncc m non-hypothetical argument."
§ 103. Our general conclusion then 1s that the categorical
statement cannot possibly constitute a species of statement,
because all statement as such ts categorical. The expression of
every dcfimte judgement, or opinton, &c , is a categorical state-
ment m the sense that a connex10n between two sometlungs is
affirmed absolutely and uncond1ttonally The categorical state-
ment which expresses an hypothetical Judgement, or opmion,
states absolutely a connexion between two problems or questions
about reality. The categoncal statement which expresses a non-
hypothellcal Judgement or opm10n states absolutely a connexion
!Jetween realities.
The doctrmc then that the categorical statement cannot be
reduced to the hypothetical stnctly means that a given non-
hypothetical statement cannot be reduced to an hypothetical.
And we may obt.1111 another proof of this doctrmc from the
account we have given of the nature of hypothetical statement.
For obviously the statement of the connexion between realities
cannot be reduced to a statement of the connex10n between
problems. We must also withdraw the concess10n that the
hypothetical statement (e g 1f anything 1s A, it is B) can be
derived from the non-hypothet1cal, all A 1s B, for this 1s mac-
curate. All that 1s true m the supposed dcnvation or reduction
1s that 1£ the non-hypothetical 1s true, the hypothetical 1s true
also. The one cannot properly be said to be denved from the other
because the hypothetical 1mphes the uncertamty whether there
is such a tlung as A, wherec1.s the categorical 1mphes that this
is not uncertam but ccrtam, and obviously we cannot clcnve the
uncertainty of a thmg from its ccrtamty.
§ 104. The non-hypothetical statement ' all AB 1s C' only
entitles us to attribute Cncss to A under the condition Bness ;
[• 'The hypothetical Judgement 1s always an 10ference.' Bradley,
1 p 407]
C, 11,
R 2
244 STATEMENT, THINKING, AND APPREHENSION
it does not convey the information whether or not Bness is
a necessary condition, and therefore is so far compatible with all
A is C. The same 1s true of the hypothetical form, if A is B,
A is C. For this docs not exclude the poss1b1hty that A may
be C, even if A 1s not B. In both therefore there seems to be
the same ambiguity, and tlus remark applies also to the con-
ditional form (whenever A 1s B, A ,s C), which as such assumes
that thr cond1t1on exists It may reasonably be obJcctcd that
we must here attend to the normal use of language and that
the intention of an hypothetical proposition n, to state that an
A requires, as far as we know, to be under the cond1t1on of
bemg B, in order to be C. And so m the non-hypothetical form
all AB is C , for here again, the normal meaning 1s that we
cannot attach Cness to A without the cond1t1on Bness; other-
w1sc wc should c;ay A ts C. It nevertheless remains true that
even in tlus normal use of language there may be an ambiguity.
When a man says all AB 1s C, he ought to mean that, so far
as he knows, A 1s C not szmpl1C1tcr, but under the cond1t10n B.
But now B may contam more than the necessary c-ond1t1on and
therefore 1t docs not follow that the A which 1s C must neces-
s:mly be D Tlus kmd of amb1gu1ty may correspond to the
speaker's own unccrtamty ; for mstancc, m experimental science,
we may find that an A \\h1ch was without the attribute Cncss
gets that attribute when we mtroduce the cond1t1on Bness We
arc therefore entitled so far to say AB 1s C; but yet we may
be unaware how much of B 1s necessary and may have to
proceed to new experiments m order to eliminate everything
except the true cond1t1on.
Curiously enough,• this amb1gu1ty is found m the exact
demonstrations of mathemat1cc1.l science. It might seem that
this was not possible , that 1f we demonstrated the property D of
ABC, where D depends on A alone, we should derive D from
A. But tlus 1s by no means ah, ays the case. A mathemallcal
proof often enough derives D from the "hole nature of its sub-
JCCt as cxpref:>sed m its dcfimt10n, or at least from more than
the element A. The reasoner 1s not aware that he has introduced
[• ' Matht'mabc1ans do not st'em to l1avc noticed tlus and 1t came as
to B B, when I mentioned 1t to lum."-MS. note Cf Part V, xiv,
a bUT'pn'le
toD.B 71vo9;fromB B.101vo9}
 Categorical and Hypothetical Statements 245
anything superfluous, while the examination of the argument
will not always show the superfluity. For instance, the harmonic
properties of the circle, for the demonstration of which its whole
definition is used, yet follow from a generic clement which
1t has in common with other come sections, a fact which can
be established by a different method of proof; but, though this
is so, the generic element formmg the definition of a conic section
cannot be discerned m the defimt10n of the circle used m the
given proof, nor anywhere m the proof itself.
§ 105. The hypothetical statement, as a general form, just hke
the universal cakgor1cal, cannot be converted simply. From
' all AB is C ' we cannot mfcr that ' all C is AB ', nor can we
from ' 1f A is B, A is C ' mfcr that ' 1f A 1s C, A 1s B '. We
shall show hereafter m dealing with mfcrence that the reason
hes Just m the amb1gmty with wluch the condition is expressed. 1
When the true or exact condition 1s stated, whether in the
categorical form or the hypothetical, either can be converted
simply. Thus ' all All 1s C ' can be converted into ' all C is
AB ', provided that AB is the sufficient and necessary condition
of C. Now, 1f this be true, smce an accurate scientific propos1-
tlon should state the cond1t1011 without the amb1gu1ty, should
give, that is to say, the sufficient and necessary condition, 1t
follows that a scientific propos1t10n, though 1t may be true, 1s
not perfectly accurate m form, unless 1t can be converted simply.
This may be at first unexpected, but mathematical demonstra-
t10ns present contmual examples of it, and it is a tes!. which
every perfectly accurate proof• must satisfy, that we should be
able to convert the conclusion. And if m any case we are unable
to convert the conclus1on, this 1s proof that the conclusion has
not been drawn from the accurate condition, m other words that
the condition states more than is necessary. A consequence of
this prmciple would be the dental of the 'plurality of causes '. 2
The vmdication of the prmc1ple itself will be reserved for the
theory of Inference.
NoTE. The treatment of hypothetical statement in these
sections requires to be supplemented by the lectures specially
devoted to hypothetical thought ; 3 but on reflection I must
1
if 352-3. 8
Part III, chs. 5-7.
[• Cf. § 259. Tlus 1s true only w1t.h the proviso 'sufficient and necessary'.
246 STATEMENT, THINKING, AND APPREHENSION
admit it to be a defect that only that sense of hypothesis is
di~cusscd here which corresponds to the ordinary meaning of
the term ' scientific hypothesis ', and no notice 1s taken of the
case where we know or believe that the condit10n stated in
the hypothetical clause 1s not reahzed-e g. 'if he had known
wha.t would be the consequence of consenting, he would not
have done 1t '. This case 1s directly but not sufficiently con-
sidered m the other set of lectures, a defect I mtcud to remedy.
I shl)ul<l hkc here to explam that the omission docs not mean
accept-inre of the doctrine held by some thmkers that the state-
ments m qucst10n arc not true hypothct1cals. At present I must
he content to say that such a view does not realize the full
difficulty caused by the fart that the only natural way of
expressing them 1s m the hypothetical form, with the particle
' 1£' expressed or understood, cg '1f A were B (as 1t is not),
C would be D '. Any other mode of exprcss10n would be very
difficult, and would stnke us at oncl, as art1fir1al. It seems also
to be qmte forgotten that a reductw ad absurdum proof 1s often
an mst,mce of the h.md, because the condition introduced by
the word ' 1£ ' t!> not conce1vcd and cannot be conceived as
possible , and yet from Am,totlc onwards these statements with
such clauses have been held as promment types of the merely
hypothettcJ.I Fmally the form '1£ A 1s 13, C is D ', which is
the proper form for the case where we arc unccrtam whether
A 1s B, and the form '1£ A were B {as 1t 1s not), C would be
D ', have tlus essential m common that m both cases A's bemg
D would be held to be somethmg conceived merely and not
represented as knov.n to be rc,tl though what this exactly
means requires careful d1scuss1on, and ought already to have
been taken account of 111 the present 10vest1gabon.
 XII
NEGATION OR TUE QUALITY OF PROPOSITIONS •

§ 106. HITHERTO we have treated only of the affirmative forms

of ~tatcment. The negative presents d1fficult1es both m meta-
physics and logic. As 1s well known, negat10n was a source of
puzzle as early as the bcgmnmg of plulosophy m Greece. Of
this philosophy 1t 1s perhaps fair to say that the difficulty 1t
had w1th negation arose from the fact that when once the idea
of negation has created a substantive form m language it tends
to pass mto a pos1t1ve conccpt1on. Not-being comes uncon-
sciously to be treated as a kmd of entity. One result of the
metaphysical d1fficult1es is seen in the Elcalic philosophy and
111 the nature of the plulosophy which succeeded to the Eleat1c.
Parmcmdes found 1t ncccss,uy to maintain that only bcmg
l'x1sted and that not-being could not exist. That he should have
found this apparently tautologous statement important is s1gm-
r1cant. Negation must have caused some very serious difficulty.
l'armemdes argued ag..unc,t tl1c d1v1dedness of bemg that d1v1s10n
woulcl mvolve somehow the mtcrvenl10n of not-bemg between
the divided parts That ,rnuld give the mtervcnmg not-bcmg
a title to reality lie probably had before lus mmd an image
of material obJccts m space with empty spa<.e (or as the phrase
1s ' nothing ') between them He d1d not however mfcr that we
arc mistaken in speaking of empty space as nothing ; he seems
rather not to have que,t10ncd the pos1t10n that empty space 1s
properly not-bcmg. The only alternative left him was to deny
such empty space I le might then alternatively have said that
the world of sense 1s really full, though this is not apparent to
the senses, or that the world of sense is an illusion because 1t
presents to us empty space. He chose the second The plulo-
sophers who followed did not accept this denial of the sensible
world but were sufficiently affected by the reasoning of Par-
[• The chapter 1s headed • Iba.nt obscun (sola sub nocte per umbram,
Perque domos D1tis 11Q1;uas et mama regna •. -4,n. v1. 268}. MS. nold.]
248 STATEMENT, THINKING, AND APPREHENSION
memdes and his disciple Zeno to believe that empty space was
somehow impossible and involved the ascription of existence to
non-bcmg. Thus Empedocles and Anaxagoras supposed space
to be entirely full, 1.ccountmg for motion by the hypothesis that
the apparently empty space was filled by very small particles
which were able to move because the particles in front of them
moved away Continuity was preserved because the place of
a particle which moved was 1mmed1atcly occupied by another.
Motion m the universe was l1kc- the motwn of the particles of
a fluid witlun itself Leuc1ppus and Democritus, 1f correctly
reported, appear to have reverted to the older view, allowing
the existence of a ' void ' as well as a ' full '. The important
cffcct of Eleat1c speculation ..1.nd the hold 1t took upon Greek
plulo!.ophy 1s shown by the faLt that even Plato in lus scientific
explanat1011 of the physical world, m the Timaeus, retains this
doctrine 111 a modified form. lle lumself however made a con•
tr1but1on to the theory of ncgat10n m the Sophist, by pointing
out that not-bcmg often llll'<ll1<J relatwc not-bemg, somethmg
that 1s wlmh has a bcmg of 1ls own but not some other kmd
of hm1g- Thus Pbto sho\\ed how m that sense there 1s no gulf
bctv.cen lwmg anll not-bemg One kmd of bcmg mvolve<J
another, 1 c. its O\\ n not-bemg One kmd of bemg !Ms its own
bemg wluch 1s not that of another ; every kmd of being involves
its other, i.e relative not-bemg This ts the fam1har modern
doctrme om ms determmatlO est negatio.
The student of Plato must however notice that, m the Republic,
not-bemg, the extreme opposite of bemg, with the 'mter•
mediates' lymg bctv.een, 1s not the relative not-bcmg of the
Soplust It is the <1bsolutc negation of any kmd of bemg, the
correlate of l1yvo,a (not-knov.mg). This 1s not ignorance m
the ordmary sense but mere absence of conscrnusness. The
not-bcmg of the Sopliist 1s obv10usly an obJect of knowledge
(m its widest sense) Thus the ancient difficulty was meta•
physical rather than ll)g1cal and arm,e mamly from the question
whether anythmg obJcct1ve corresponds to negation. The logical
difficulties relate to the fonn as such ; for instance, whether it
is a different species from the affirmative, or whether the latter
is in some sense the form of all statement; and again whether
the negative symbol belongs to the so-called copula, Difficulties
 Negation or Quality of Statements 249
arise also about the negative conception which are not purely
logical. Certain of these hardly appear in Plato and Aristotle,
and even in modern times we observe an unconsciousness of
some of them. Thus 1t is usual to d1v1de • propos1bons or
'judgements' mto affirmative and negative, without even asking
\\hat is the genus of which they arc the species. To say that
m Judgement (or in a proposition) something 1s affirmed or
dcmed of something else is the elementary fallacy of definmg by
an enumerat10n of species instead of a statement of the genus.
The difficulty is lost sight of m a definition which enumerates
the t,vo.
We shall mamt,un th.1t the negative proposition is a distinct
species and cannot be brought under the general affirmative
form, and shall try to determine the general characteristics of
slalcment as common to the two species
We shall further mamtain that the affirmative and negative
forms involve one another, and that, though d1&tinct, they are
not properly co•or<lmc1.tc, the affirmative bcmg pnor
As regards the mctaphy1,1cal question<; we shall argue that
negation 1i:, not merely subJective but belongs to reahty. Fmally,
\\ c i,h,tll clu,cw,s ctrt.1111 i,pcual chfficult1es presented by that
nl'gdt1vc sL1tement m winch the negative does not unply, even
indirectly, a pos1llvc determination of 1ts subJect, but its entire
non-existence
§ 107 The negative statement with the verb ' to be' can in
..ippcarance be reduced to the affirmative by om1tt1ng the
negative with the verb and putting a negative expression, in
the form of an adJecbve or common noun, mto the attributive,
so that • (~hA 1& not B' becomes • (shA is not·B' or • (s)iA
is a not·B '.
Now If we treat the negative and affirmative as two di.ff erent
species, there arises the problem of finding a form to represent
that which is common to the affirmative and negative state•
ments. This cannot be '(s) 1A is B ', or • (s)iA 1s not B ',
smce the common genus cannot be either affirmative or negative.
Thus to overcome the difficulty which meets us1 if we are looking
for something analogous to these formulae, we might possibly be
tempted to suppose that the affirmative 1s the true generic form
[• e.g. Mill, S;yst,m of Loc,i;, J. 1v, i§ J and :z)
250 STATEMENT, THINKING, AND APPREHENSION
and that the solution ts to reduce the negative to the affirmative
in the manner described. It will be necessary, therefore, to
consider the meanmg of the verbal transformat1011 by which the
negative statement 1s m a manner reduced to an affirmative
form.
In the first place, there is a very important difference between
tlie new quasi-affirmative and the ordmary affirmative ; the
negation tb not chmmatcd after all, and the statement docs not
as such give any mformat1011 of a positive character about its
log1cctl sub1ect or the subject of attribution In consequence 1t
doeis not give a further positive determmation to our conception
of these sub1ccts, like the ordinary affirmative statement. In
some except10nal cases it may seem as 1£ this were really pos-
sible; thus 'five is not an even number', • five ts a not-even
number', 'five 1s an odd number'. But now 'odd' is as
positive as ' even '. Th1b however 1s made possible not by the
negative statement itself but by something else which we know
about the particular matter m hand The affirmative statement,
five is an odd number, 1s really an mfcrcncc from the negative
five 1s not an even numucr combmccl with ,lllother ancl that an
affirmallve st,itcmcnt, namely that all number 1s either even or
odd. Tlus is not of umversal applicat10n bccJ.use we cannot
always get such an affirmative d1s1uncttve premiss If I say,
' tlus man 1s not a Mohammedan ', I cannot thereby <lctermme
his rehg1on, nor even that he has any at all
What 1s characteristic of the quas1-J.ffirmative form to which
the negative propos1t10n 1s reduced 1s the expression, com-
~

pounded with .t negative, \\ h1ch 1s made to do duty as an

a<l1cct1ve or common noun and has the grammatical form of
one. Tlus quasi-ad1cct1ve of the form • not-B' 1s often supposed
to have a somethmg correspondmg to 1t called not-Bness, JUst
as Bness corresponds to the true ad1ect1ve B. Buess is a kmd
of bemg and &o ob1ect1ve; but, as 1t 1s a umversal, 1t gets
called, in the maccurate conceptuahst1c phraseology which 1s so
prevalent, a concept10n or a concept. Similarly the assumed
not-Bness 1s called a negative conception or concept, and is
treated, in effect, as 1f it were a umversal.
We have then to examme the negative expression not-Bness
and ask whether any true umvt:rsal corresponds to it, whethrr
 Negation o, Quality of Statements 251

it is not rather only a verbal form like that of a universal. Can we

m fact in any reasonable sense of ' conception ' say either that
1t 1s a conccpt10n or that. there 1s any conception of 1t ? And
again does 1t serve to transform a negative propos1t1on mto
a real affirmative~
§ 108 a It 1s true, and a commonplace, that the recognition
of a pos1t1ve character m anything involves acts of comparison
"1th that which 1s not itself, and so far mvolvcs ncgat10n ; so
that to rccogmze anythmg as of a definite character means
cfo:,lmct10n. We recogmze some feature m the tl11ng wl11ch
enables us to identify 1t ac; one among other thmgs with which
we <lo not. confuse 1t. llut now we do not know a g1vrn some•
thmg as merely contr.istcd with somethmg different, but by
\\hat we call its own nature, through wluch alone the contrast
1s possible ; and thus the sweepmg statement omnis determinatio
est negatio b 1s m1slcadmg, for 1t 1s (taken strictly) an identi-
fication of the determinate and definite with the negative, and
;ipparcntly at all events an obJcction to any distinct10n between
them But the negative itself depends for any <lcfimtcness 1t
c..i.n have upon the pos1t1vc dcfi111te ch..i.ractcr of the contrasted
poc,it1ve clements.
Tlus leads to an important distinction between the (supposed)
merl'ly negative 11ut10n and the positive, although the latter, as
we have seen, mvolves negat10n. If we allow the ordmary way
of spcakmg about concept10ns, \\C may say that the positive con·
rrpt1011s get their negative side by contrast with other positive
conc.cpttons, by contrast, that 1s, with conceptions as positive
m themselves as the former, and these m their turn are not
known as merely other than those which we arc contrastmg with
them So we may rightly say that we d1stmguish these from
one another by their own positive nature, by what they are and
not by what they arc not Now with the supposed purely
negative conceptions the case 1s exactly the opposite ; the
negative conceptions can only be d1stingu1shcd from one another
and have only an 10div1dual or d1stmct character through what
they are not, or more accurately through a definite pos1t1ve
(• • Rewnte all this with a view merely to "conception" and phrasing',
MS note
b Usually but IIllstakenly ascnbed to Spmo.za )
252 STATEMENT, THINKING, AND APPREHENSION
something which the objects they concern are not. In fact to
every so-called negative conception such as ' not-Aness ' there
corresponds an indefinite sphere, consistmg of what is not A,
which could not be dealt with at all except through the definite
and positive ' conception ' Ancss, exclusion from which con-
stitutes the ml,anmg of the negation. Thus ' not-A ' and
' not-B ', ' not-Ancss' and ' not-Bncss ', can only be distm•
gu1shcd from one .mother through ' Ancss ' and ' Bness '.
This 1s one way of showing th.it the so-called negative concep-
tion 'not-A ', 1f 1t C}..lstcd, eould not be equated to or identified
with any pfls1t1vc conception whatever.
If we discard tlus apphcat10n of the terms • conception ' and
' concept' and use language "lurh seems to smt the facts, we
may say that whereas Ancss and Bness arc distinguished by
what they arc, the supposed not-Ancss and not-Bncss could only
be d1stingmc;hed by means of the d1stmction between Aness and
Dnc!>!>, that 1s hy what they arc not Thu,; the statement
' (s) 1A 1c; a not-B ' rc111<imc, cs5tcntully d1c;tmc.t frnm a statement
of the affirmallvc form '(s1 1A 1s C ', and 1s mdcc<l dependent
on tlus :l.ffirmat1vc for its 111l'a111ng
It seems then that 1£ there "ere such a thmg ac; a negative
conception not-Ancss, 1t could not be equated to a pos1t1vc one;
but we may ask whether m the proper ml.anmg of conception
there is any true conception at all corrcspondmg to the verbal
form not-Aness.
The conception of Ancss, when 1t 1s not \\ hat we shall call
a problcmat1c concept10n, 1s the· apprehension of Aness Can
there be something not-Ancss prc5tcnt 111 all the mstanccs of
what 1s nc,t-A, wlm h arc each c;,.ud to be a not-A, and can it
be a true umvcr!>al, and apprehended as such~ The universal
is somethmg common to all ,ts particulars, and the only
tlung common to all the <;o-callcd not-A's 1s cx1stcncc, smce
not-circle for mstancc includes a triangle, a sound, a theft,
and so on ad mfimtum. But existence 1s not the same as
not-Aness, so that not-Ancss 1s not the universal common to
the not-A's
We have, of course, to meet the objection that, while the
not•A's have notlung positive common to them except being or
existence, they JUSt have the negative not-Aness common to them
 Negation or Quality of Statements 253
all, since each is a not-A. Suppose then not-Aness ( =' not•
being-an-A ') were a true universal. A true universal can only
exist in certain forms called its differentiations which its own
nature necessitates These arc species in the strict sense and
the given universal 1s their genus. Thus a lme must be straight
or curved, and lmearity of itself necessitates the existence of
rectilmeanty and curv1hneanty as forms in which it must exist.
• Equilateral four-sided figure ' on the contrary 1s not a diffcr-
entiat10n of eqmlateral figure ; for four-sidedness 1s not neces-
sitated by eqmlater::i.lness but by the universal ' figure ' of which
it is a differentiation.
On the other hand, if a umvcrsal can only exist m certain
forms, these ,, 111 be its d1ffcrcntiations and true species. If
a given sometlung 1s not A, and so accounted a not-A and a
particular of a supposed umvcrsal not-Ancss, it 1s a not-A not
as a mere negative but as having a positive quality or nature
Bness different from Aness, and only because 1t has such a nature.
Thus its not-Aness will consist m its Bness, and not-Aness only
exists m it in the form of Bncss; and, m general, not-Aness
can only exist in the form of the positive umversals Bness,
Cness, Dncss, &c, wluch are such that their particulars are not
A. It follows then that each such positive universal as Bness
must be a differentiat10n of not-Ancss and a true species of it.
But this is impossible. The whole nature of the spec1es-umversal
is comprised m the genus, takmg genus and species m the strict
sense. There 1s nothing m the quality of the species which does
not belong to that of the genus and is not involved m it as bemg
a necessary form which the genus must take in existence. Thus
there is nothing m blue but colour, and blue is one form which
colour as such must take There is nothmg m rectilmearity
wluch 1s not hnc::mty. rectilmearity 1s one necessary form of the
being of lmcarity In this way the universal constitutes the
whole nature of its spec1ec; or d1fferenti.1.tion. Ilence whatever
Aness and Bness may be, provided they are different so that
a B 1s not an A, not-Aness would have to comprise the whole
nature of Bness. But this 1s impossible . for not-Aness 1s simply
the absence of a certam quality and mere absence of a positive
quality cannot constitute the whole nature of any positive
quality. If we turn from symbols to real instances the absurdity
254 STATEMENT, THINKING, AND APPREHENSION
is manifest. We should have to say, for instance, that theft was
a species or kind of want of circularity and circularity a kind
of absence of theft, or that the whole nature of theft was com-
prised in want of circularity.
Contradictions also may be developed. If Ancss, Bncss and
Cness arc all chlTcrent m the 11cnsc that every A 1s not B, every
B not A, &c., 1t \\ould follow as before that Cncss was a species
of not-Ancsc; and at the samr t1mr a !>pcc1c,; of not-Bness But,
1f a umvcrsal 1s a d1ffcrentiatH,n of t\, o different umvcrsals,
either these two arc related themselves as true genus and species
or they mutually mvolvc one another so that each necessitates
the other. Thuc; though not-Anesr; and not-Bness exclude one
another, either one 1s a 11pcc1cs of the other, or each mvolvcs the
other, but both these alternattvcc; arc self-contradictory There
arc further absurd1ttcs m the t\\ o cases In the latter case not-
Aness and not-Bncss can only necessitate one another 1£ Aness
and Bness necessitate one anothrr (m the sense that every A is
B and every B is A), smcc the nce;at1ons, as ,, c have seen, can
only he rrlated through the positive elements of wluch they arc
the negations This must be trur whatever Aness and Bness
may be, so long as they differ from Cncss m the manner described
Tlus 1s absurd, for then, to take an example, smce a circle 1s not
a sound and not a theft, a sound would have to be a theft
and a theft a sound In the other case \\ here not-Anesc; must
be a d1fferent1at1on of not-llness or vice versa, there could be
nothmg to determine which was the true alternative, and each
\\ould have to be a d1fTerentiatton of the other But this ag:un
1s a contrad1cbon These conttad1ct1ons would occur m the case
of any two positive umvers..il.;; whatever \\ luch exrludc one
another m the given !>t'nc;e, licc.wr;c a tlurd umvcrsal can always
be found \\ luch exclude!, both Fc1r mstance, a theft must be
a kmd of c1rclc and a Circle a kmd of theft.
The mistake committed m mah.mg not-Aness a universal (or
a ' concept ') 1s the fallacy of treatmg the difference of one
quality from another ,ls 1£ 1t ,, ere 1t,;elf a quality, like the qualities
wluch differ, and belongs to a familiar class of mere verbal
fallacies m \\Inch tlungs are put m 1mposs1Lle relations to them-
selves. Relations more especially arc treated m effect as 1£ the
same m kmd as the terms related, "1th the general result of
 Negation or Quality of Statements 255
producing impossible unending series. 1 In the present case this
sort of absurdity arises as follows. Since the quality Bness
excludes Aness, this relation of the two is taken itself as a quality
called 'not-Aness ', and Bness becomes a kmd or species of
not-Aness. But now not-Ancss, taken thus as a quality, is
a quality which excludes the quality Aness. Hence, as before,
this exclus10n in turn 1s not-Ancss and, as before, not-Aness
bewmcs a species or kmd of not-Anccis, that is a species of
1tciclf. Not-Aness then bemg a species of not-Aness, 1t 1c; a species
of a species of not-Ancss, that 1s a species of a species of a species
of not-Aness and so on without end.
§ 109 Consider now not-Ancss from the side of apprehension.
Not-Aness cannot be apprehended as mere negation, for the
mere negation of anythmg cont:uns nothmg to apprehend. If 1t
1c; obJected that we do apprehend not-Ancss or absence of Aness,
because we apprehend that somcthmg 1s not A, 1t must be
replied that tlus 1& not the apprehension of mere not-Ancss, but
the apprchcns10n of a positive something the nature or quality
of which, for mstance 13ncss, 1s d1fferent from Aness It would
hardly then be replied that mere absence of Anecis can be appre•
hendcd, when there 1c; nothmg but absence of Aness, for that
would be to confuse the absence of an apprehens10n with the
apprehcns10n of absence. It being then impossible to apprehend
mere not-Aness or mere difference from Aness, we must in any
negat10n of Aness apprehend somcthmg positive, of a nature
Dness different from Aness, thus the apprehension necessarily
mcludcs that of Ancss and that of Bness. That 1s, the only
possible apprehens10n of the negation of Aness is m the appre-
hension of Aness together with the apprehension of a quality,
say Bness, different from 1t, and the apprehension of each as
different from the other 2 The apprehension 1s of two chfferent
qualities and not of a quality of difference.

' TI1c5e fall,tcu '> an- '>otm tlml'<, t,tk<'n very ,cr10u'>ly Thu, ma W< J1-known
book the general m1,tnk<• about n•l,ttlon (wluc.h can he c.,t,1ly <'XJ'll>'>t.d) 1'1
m..ide part of the h,:1<,1~ of a ml'taphy,1cal theory, and the muc tangle• of
a verb.ii fallac.y 1s quamtly taken to be a 5ort of '>Clf-wntrad1cbon m the
nature of Thought 1t,clf LV1dc Part V, §§ 433-7)
• Th1!> latter apprch<'n~1on 1~ a part of the forml'r, and not <,omcthmg added
to 1t For Bness bemg d1flcrcnt from Anc55 we nccl'&5anly apprehend this
"'hen we apprehend both Ancss and Bness, otherwise we should confuse 1t with
256 STATEMENT, THINKING, AND APPREHENSION
It may be objected that though not-Aness, or difference from
Aness, can only be apprehended in the apprehension of An•ess
and Bncss, and of Ancss and Cness, &c , yet the apprehension
of it may be d1stmgmshcd from the apprehension of them, and
1t can be appreh:.-nded as a umty m the difference of Bness from
Ancss, and the difference of Cnesc; from Aness, &c. ; m short
that a rclat1on, e g equality, can m general only be apprehended
along with the apprehension of tli.e terms related, and yet that
this docs not prevent our abstracting 1t from them, and appre-
hending 1t m abstraction as a umty m the different instances.
But surh abstraction 1s not always possible. One quality may
so depend upon and presuppose another that 1t cannot possibly
be separated from that other by any kind of abstraction For
mstancc, straightness can only be straightness of a hnc and
cannot be apprehended apart from lmeanty or in any abstrac•
tlon from lmearity Thus there 1s not, properly speaking, any
' conception ' of straightness in abstraction from lmearity. So
also oddnrss prcc;upposec; number. There 1s no conception of
oddncsci in itself, in abstraction from numbrr Such an abstrac-
tion could not be performed Similarly the negation of Ancss
can on the one hand only be apprehended as the difference from
Aness of some positive quality, say Bncss, and on the other
cannot po.,s1bly be abstracted from such pos1t1vc quality, Just
because negation is d1fTercncc of posillvcs. 1 Thus of not-Aness
in the abstract, or of difference from Aness in the abstract, there
is no apprehension or 'conception' whatever, any more than
there could be of oddness or straightness m abstraction from
lmc or number.
Further, "1th regard to the rase first alleged (the abstraction
of a rdatwn like equality), 1t 1s obvious" c cannot m our thought
abstract the concept10n of any rclat10n from the conception of
the kmd of terms it n.·latl's Of tluc; general principle negation
1s a particular case But JUSt as 111 the case of oddness \\C get
a true umvcrsal and true conception by takm~ m the nature
of number to wluch oddness belongs and which 1t presupposes,
Ancss Thus there 1s no app1ehens1on of the difference of Aness and Bness
other than the apprehension of both Aness and Bnc~s together
1 This 1s m effect the I'latomc doctrme, m the Sophist, that negation 1S

• othemess •, but with the add1tJon that there can be no conceptmn of other-
ness m abstraction from poS1bves whJ~h are 'other· than one another
 Negation or Quality of Statements 257
could we not similarly in the case of not-Aness, by taking in
the nature of the positive being to which not-Aness belongs and
which it presupposes, get a true universal and a something
which could be apprehended in abstraction ? Is there not an
abstract unity in fact, apprehensible as such, which could be
called 'positive bemg m general which 1s not Aness'? We
shall find that the analogy does not hold, and that there is
again no reality to correspond to this verbal form of a universal.
We have to abstract from the not-Aness of special forms of
being, such as Bness, Cnesc:;, &c , determined Ill each case by
a positive character such as Bness and as inseparable from it
as oddness from odd number, and h.1.vc ex hypothesi to include
abstraction from the pos1t1ve natures of Bness, Cncss, Dness, &c.
Moreover this positive something, which 1s abstracted along with
the not-Aness, must be one and the c:;ame in all the mstances
Bncss, Cness, &c.
What will this abc:;tract1on be ? The pos1t1ve clement cannot
be positive bcmg in general, for that would mcludc Aness. It
must be a kmd of positive being somehow specially dcfin<'d.
But, as we have seen, there 1s no c:;pecial kmd of being common
to all that wluch is not-A, nothing, that ts, common to sound,
triangle, theft and so on by wlm.h these are all not-circular.
Thus the required abstract10n cannot be found.
The matter may be put also m tlus v. ay. The positive clement
which must be included in the proposed abstraction, as mdicated
m the verbal form, would have to be the special kmd ot positive
being which differs from Aness. Now difference from a given
somethmg Aness can only exist as caused by the positive
character of what 1s d1stmguished from the given somethmg.
Thus the distinction of the kmd of bemg to be abstracted as
being different from Ancss must depend on some positive quality
which 1t has Thus a positive quality must be common to all
the bemg which 1s not-Aness, which however 1s 1mposs1ble
This makes 1t evident what the fallacy of such an abstraction
consists in. It 1s the endeavour to define a special kmd of being
by the mere fact of its distinction from something else, which
is an inversion of the relation between defimtion and distinction.
For it is pos1t1ve definition (including the recognition of a positive
quality which 1s sui generis) which alone makes possible a dis•
s
258 STATEMENT, THINKING, AND APPREHENSION
tmction of the thing defined from other things. Such distinction
therefore presupposes a definition to explain it. The conclusion
arrived at is attested by the fact that every one would think
it absurd to institute any sort of inquiry into the nature of
not-Aness as something common to everything distinguished
from Aness, into the nature, say, of non-circularity as something
common to a triangle and a theft For this fact would be
unaccountable 1£ not•Aness were a true umversal,-were any-
thing of wlu<'h we could seriously be c;a1d to have any concept
or roncept10l'l, or wluch could itself be called in the ordinary
confused phraseology a ' concept '.
The mere verbal fallacy of bemg misled by the artificial form
' not-Am·ss • into supposing that there as a universal not-Aness,
or that there 1s a concept not-Ancsc;, or that there 1c; a concep•
tlou or ronrept of 1t, ha'l led to unworthy puzzles which have
been gravely taken for plulosopluc truths It 1s excusable to
find difficulty 111 getting at the exact nature of the mistakes 111
verbal fall,tc•1ec;, muced 1t 1c; often really difficult, but 1t 1s quite
inexcusable not to 5ee that they mu,;t be nonsense, and actually
to mistake !-Uch mfinmtics of mtellcct for subtleties of meta-
physic.
§ I 10. Smee not-Bnec;s does not represent a umvcrsal, or kmd
of bemg ,the reduction of (s) 1A 1s not 13 to the verbal form (s) 1A is
a not-B rannut c;cn e to reduce the negative statement to a true
affirmative form llo,,ever, 1t nught be contendrd that the
statement (s)iA 1s not I3 puts (s) 1A 111c;1de thl' sphere of reality
which 1s outside what 1s B, and that tlus gets ,·erbal expression
in fact m the form of statement (s) 1A 1s a not-B. Now it is
quite possible th.1t i,omc c,nc who chooses tlus latter form
deliberately does mtend that (s)i has some other place in the
reality outside B and that may bl· precisely has reason for
choosing it. Tlus would at once suggest that he chooses this
form because the usual negatav<:> c;tatement has not naturally
that imphcataon That 1t has not got this 1mphcatlon 1s indeed
the truth, the intention of the form b<'mg normally only to
exclude (s)iA from "hat 1c; B and not at all to assign (s}i a place
m reality. But more than tlus. It 1s not true that the negative
statement as such, (s) 1A 1s not B, necessarily implies that (s)i
has a position m that other sphere of reality at all ; it is com-
 Negation <W Quality of Statements 259
patible with the statement that (s)iA as (s)iA has no being
whatever. If I find that this page is not in my book, it does
not follow that the page, as a page, 1s anywhere. The page
may have been burnt, and, as a page, have absolutely no
existence. ' The pam 1s not in my head now ' means that the
pain has no existence, not that it exists somewhere else. To
avoid misunderstanding, we shall return to this sort of example
in the discussion of statements wh1rh assert complete non-
existence
To return, (s)iA 1s a not-B al all events gives us no mformation
about the pos1t1011 of (s) 1 m the 111fi111tc sphere of reality outside
what is B Whatever (s) 1 may be pos1tivcly remams unknown,
since merely to put (s) 1 in this mfimte unknown (for we are here
to omit the only thmg by wluch 1t can be known, namely, its
exclusion from what 1s B), is to give it no positive determination
of being whatever. We have already seen, m the example of
'five is not an even number', that when such a positive deter-
mination seems possible, 1t 1s only because wr have combined
with the negative statement a d1sJunct1ve affirmative which
confines not-B to one or more defimte positive realizations,
thereby really lumtmg the sphere of not-B to a portion of the
infinite not-B This 1s what 1s done mall ehmmative arguments.
In them we seem to arrive at a kno" ledge of what (s} 1 is by
'ltatcment'l of what 1t 1s not. Yet these ncgatJvrs only serve to
give (sh a positive detcrmmat1011 because they cancel certain
pos1t1ve alternatives wlurh are given to (..,) 1 111 a disJunctlve
statement. Thus, 1f (s)i 1s either B or C or D, the negatives,
(s) 1 is not C and (s) 1 ts not D, m rombmat1011 with this dis-
Junction, give us finally (s) 1 ts B. This 1s an ehmmative argument
and is the analys1c; of what Bacon and Jue; modern followers
understand by mductlon
Finally, then, the negative statement taken by itself cannot
give pos1t1vc detcrmmabon as the affirmative docs. It remains
essentially different m spite of the apparent reduction, which is
merely verbal Moreover, although 1t l'l true that ordinary
negative statements (cg. nobody m the next room can read
Greek), hke ordmary affirmative ones, normally presuppose the
existence of their subJects of attribution, this existence is not
asserted by the negative statement as such.
82
 260 STATEMENT, THINKING, AND APPREHENSION
§ I I I. It will save fruitless effort if we recogniz~ at this point
that we cannot profess to explain negation and its correlative
affirmation. Negation and affirmation may be contrasted, but
they cannot be explained or derived from one another, nor
understood through anythmg but themselves. Anything given
as the basis of such an explanation, to be mtelhgible, would
have to presuppose the thing explained They are not in fact
among the things to "luch thl' idea of explanation can have
any application
• The most direct and adl'quatc tlung we can say 1£ we wish
to dr-sa1be the chffrrcncC' bet WC'C'll affirmative and negative
statcmentc; secmr;; to be tlus. An affirmative (s)iA ic; B giver;; m
its attributive a kind of bemg wluch the subJect of attribution
has, while the' negative gives a kind of bemg which this subJect
has not Further, a negative cannot give a kind of being which
the subject has without presupposing a strictly affirmative state•
ment. Conversely, the affirmative cannot perform the func•
t1on of a negative, and tl'II us wh:1.t sort of being a subJect has
not, except by the assumpt10n of ncgat1vr statements. If
I know (s) 1A 1c; B, I do not know (s) 1 1s not r, unless I have
the negative :-tatement 'what 1s B 1s not C '. Now both
statements do somcl10\\ advance our knowledge and make our
conceptions more dcterminate.b If then we merely define the
affirmative statement as one "hich gives a further determination
to the conception we have of the subJect, that might be under-
stood of the negative statement also Suppose I know (s) 1 as
A, and \\tshmg to extend my knowledge ask v.hl'ther (sh 1s B,
or C, or D, so that (s) 1 1s the logical subJect The question
1mphcs that for me the' conccpt10n of (s) 1 1c; indeterminate with
respect to B, C, :mrl n, and thus my concept10n of (s)i includes
an unknown fidd of pos..,1h1ht1C's, bemg mdccd so far problematic.
Now when I arrive at (s) 1 1s not B (m \\ h1ch logical subJect and
subject of attribution eomc1de) my conception of (s) 1 is more
determinate on the one hand than 1t was, because a certain
field of possibility is definitely excluded On the other hand it
is more adequate to reahty, for I no longer thmk it possible
[• 211d para ' Rephrase all through \\here the word " conception " occurs •
MS not, The reason for this will appear from Part II, ch xiv
b 'Alter dre,,,m1nt11,11n language', MS notr, repeated at§ 112)
 N egatioti or Quality of Statements 261

that (s) 1 may have an attribute which it cannot have. It follows

then that in the negative statement we have acquired a new
determination for our conception of the bemg of its subJect.
We cannot therefore define the affirmative statement simply as
giving such a new determmation, for that will not distinguish
it from the negative.
In this wider sense of determination, both the negative and
the affirmative statement give a new determmatton of the con•
ception of the logical subJcct and, generally, a conception of the
subJect of attnbut10n different from that which is given of it in the
nommative case. Further, the d1stmct1ve function of each 1s to
give a kind of determination wluch the other does not give.
Thus neither 1,; a form of the other and neither 1s the general
form of the st,tt cmcnt or propoi,,1t1on. The general form itself
cannot be symboltJ:ed by ' (s) 1 1s B ' or ' all A 1s B ' or ' (s) 1A
1:. B ', or anythmg of that kmd, any more than the universal
of number can be expressed numcneally, for such a u111versal
cannot be either odd or even \Vhat then 1i,, the common form ~
From the side of apprche11s1011 or conceptwn 1t 1s simply
the idea of a general dcterm111at1011 for the given concept1011,
,,luch ts brought about by the act, whether of knowledge or
opinion, "luch finds cxprcsston 111 the i:,ta.tement. Now, in
scekmg a nc,, tletenmndtton of the Iog1cal subJcct, \\e must
hc1,ve before u<; some pos1llvc conc.ept1011s of a kmd of bemg m
rclatwn to \\l11d1 the dctcrmmatwn 1s tu be got, and tlus deter•
mmation d1flerent1.i.tc& itself ,tt onLe 111 two w.ty'> either the
bemg to "luch the pos1t1ve concept1011 rctcri,, belong'> to the
i:iubJect, or 1t docs not The ,t%ert1011 of the firi:it 1:, the .i.ffirmc1,tive
&tatement, of the i:,econd the negdtl\ e
From the side of \\hat 1s c1pprehended \\e may represent the
matter thus. A given rc.i.hty 1s definite and determmatc, as
opposed to havmg being m general or as opposed to some
universal \\Ider than 1tsclf, by havmg certc1111 dcfimte kmds of
being and not havmg others Its determmatencss therefore neces-
sarily has these two aspects the pos1t1vc one, of the bemg
which it has, and the negative one, of not havmg the other
kmds of bemg. An act of knowledge, whether Judgement or
not, 1s the apprehension of some detcrmmat1on of a thmg other
than a deternunatton which we hc1,vt: already apprehended in it.
262 STATEMENT, THINKING, AND APPREHENSION
This is what determines the generic definition of statement of
knowledge. As to the two special forms, the statement is
affirmative if we apprehend another kmd of being as possessed
by the thing ; that is, if we apprehend the positive determina•
t10n, and negative 1£ we apprehend some other kmd of being
as not possessed by 1t ; m other words, 1£ \\ e apprehend the
negative dctcrmmation Opm10n and its statement are to be
treated an..ilogously
Again, 1£ the sul>Jcct of attnbution, th..il 1s, the nominative
case to the, erb, 1s not the s,1me ai. the log1cal sul>Jcct, 1t 1s evident
what c]1,wgcs should be made m the .ibove treatment so as to
cxpnss the d1ffcrcncc bet,,ccn affirmatI\c and negative state•
ment 111 terms of the !-UUJCCt of attnbutJon, smcc m every case
the statement gives m It!> .1ttnbut1,·c J. different conception of
the subJect of attnbul1u11 from th.it represented in the nommat1ve
case to the verb
§ J 12. We have no,, tu ,1!:tk "hethcr 111 vie,,· of th1!> more
gcner.il rlt.count ,LdJptcd to both c;pccI.Ll forms of statement we
ran relJIII our pre, wuc, d1:.tmt tton ot log1c,1l subJcct and pre·
d1catc1 \\luch ,,.ii. l>,ti.cd on the .1ffirm.1t1vc form of the state·
mcnt; and, 1f i.o, \\hcther 1l c,m l>c mJdc gc11er.tl so as to suit
both the form:. of st.itemcnt
Let us take the case "1th the Ycrb ' to be' as pnnc1pal verb.
In the first pl.ice, ,, e m:1y follow the :uulogy of the trad1t1onal
usc1ge accorthng to \\ Im h Dncss 1-. the predicate, both m the
affirm.1t1,c /s) 1 A 1-, B, ,,nd m the m·g..itiw (\') 1A I!> not B, sup-
posmg (s)i to be the h1g1cal sul>JLlt, l>ut \\e 'lh.ill have to alter
our defi.111tto11 of prcd1c..1tc \Ve may s,iy that the predicate 1s
that part of 1t•,1ltty 1!1 rd,Lt10n to wlm:h the subJcct receives its
nc,, dctermmation m our c-011cept1011 of 1t Then 1t will be
Bness in both <:aSl'S 1 ln., mdu<lcs our dcfimt1on of the predicate
m an affirmative statement, th.it Bness is the new kmd of bemg
assigned to the subject, 1f l.1Ji 1i. the logical subject. But,
secondly, "c may define the prcd1c-Jte not as the positive part
of being, m respect of \\ h1ch the new dctermmat1on takes place,
but as the "hole of the new determmatlon. {Tlus, though 1t does
not suit the traditional usage, really follows better on our usual
notion of ' predicate ', in virtue of "h1ch a certam difficulty 1
1 It lb Ldl>)' to bhow the confub1on of the tr.i.d1t1onal logic in tlus matter
 Negation o, Quality of Statements 263
is always felt in calling Bness the predicate in a negative pro-
position of the form (s)iA is not B.) If so, in the affirmative
form of statement the words ' is B ' really correspond to the
predicate, and m the negative the words ' is not B '.
Here it might reasonably be obJected that since the new
determmat1on, whether negallve or pos1t1ve, 1s represented as
a determination of the being of (sh, and smce we have mam•
tained that ' 1s ' represents the general bcmg of (s) 1, we ought
to retain this in our treatment of both forms and so make the
predicates 'Bness' and 'not-Bness '. Then 111 a sense the
negative would be reduced to the affirmative form Our reply
would be that, even 1f we did this, \\ c should not be making
the affirmative the gener.il form, for the general form would be
divided into the two, 111 respect of their predicates 'Bness' and
'not-Bness ', o\\111g to that difference between them which we
have already rccog111zerl, and thus would remain the general form
as wider than affirmat10n or ncgat10n The sufficient answer how-
ever seems to be that this 1s not the normal meamng of the
words 111 actual statements. In saymg (shA 1s B we do not
tlunk, m the general conccpt10n of ' 11:, ', of a mere general
determination which may be realtzed as bc111g something or not
bemg sometlung That 1s, 1t docs not stand for the concept1011
of mere determ111abon 111 general, but ,,c do tlunk of 1t as the
general form of bcmg \'v luch ts to be rcahzcd 111 Bncss. And so
agam in the negatn,e statement 1tsclf, ,,c do not tlunk of 'ii:,'
ai:, being 111 general, to be presently chffercnttc1ted mto a Pegativc
form, to be d1fferent1ated, that ts, 111tu '11ot-bcmg '. On the
contrary, we fed that the ' 1s ' goes closely "1th B as repre-
senting a particular kmd of bcmg, and the negative word 1s
mtcnded to cc1nccl that Tlus 1s brought out by the fact that
in speaking we cannot, 111 the negative statement of the form
(s) 1A ts not B, properly make a pause after the word' 1c; ', for we
do not do this normally unless we intend the ' 1s ' to be followed
by a positive determmation Consequently, 1£ m speakmg we
did stop at ' 1s ', the listener "ould feel surprised 1f we con•
tinued the sentence as a negative, and indeed a special rhetorical
effect may be produced m tlus way. Lastly, on the analogy of
our previous assert10n that in the affirmative, (sh 1s B, the word
'1s' represents the general bcmg of (sh, while B represents
264 STATEMENT, THINKING, AND APPREHENSION
a special form which this takes, we may say that in the fonns
(s)iA is B and (s)iA is not B ' is ' and ' is not ' correspond to
the general positive determination of (sh and to the general
negative determination of (sh respectively. And in each case
the particular form which the determination takes is given by
the addition of B, which corresponds to the kind of bemg which
gives determination to (s) 1, m the one case as identified with
a part of, or \\ 1th the whole of, the bcmg of (s)i, and in the
other case as d1stmguished from that bemg.
It 1s uot difficult to apply the above analy:.is to forms where
the prmc1pal verb 1s not the verb ' to be ', or to see what should
be the analogous extension given to the distinction of subject
of attribution and attributive This part of the investigation,
however, rcquirL·s discussion, "Inch must be omitted for the
present. We also reserve a consideration of cases where the
nominative of the negative statement 1s the true logical subject,
wluch prci.cnt ccrtam fr.1tures peculiar to themselves.
§ I 13. Though the aflirmat1w and negative forms then arc
ddkrent, they im olvc c,tch other and yet arc not co-ordmatc
m lhl· strict senc;r of the term.
• 'I hl' affirm,1t1\·r c;tatctnrnt (s)i 1s B uwolves ,t negative, for
1t 1:. not 111tdhg1ble unles<, B 1s d1<,tmguishcd from other clements
111 reality .ind tluc; chstmctwn "ould be matter of negative state•
mcnt Ag,un the negat1\ c m its very naturr presupposes the
affirmative \Ve c..annot, 111 an 111qu1ry, .i.mve at the result (s)iA
is not B without ftrst putt111g before ourselves the question
whether (~) 1A 1s B, and at any rate the negative st.tlement gets
auy dcfu11tcn<.sc; th,1t 1t has from the affirmative, for without
that the negative g1, cs no <lctcrm111.tt10n at all. The expression
'1s not B' hac; its meanmg too only as the cancellmg of '1s B '.
The affirmatlw then 1s 1n tins sense prior, and 1t 1s prior also
in order of time, for "e cannot brgm "1th mere negation. It
is quite true that knov.ledgc must bcgm \\1th distinctions which
involve negation, but 1t 1s also true that the negation 1s only
intellig1ble here through the pos1t1ve conceptions. To know
anything about blue, I must be able to d1stmguish it from red
l• 2nd para • mod1ly ', 3rd para • alter a little m form•, and 'm the
order of tune the ucgatl\e t.,mnot prc~cde the affinnat1ve for we can't begin
'\\llh mere negat1011 •. !,IS 11ote.s,J
 N eg«non o, Quality of Statencents 265
and so perceive that blue is not red, but this distinction is only
real for me, as opposed to the mere empty idea of distinction
m general, through the positive character of the elements dis•
tinguished. It is only because of the positive characteristics of
blue and red that one cannot be the other and that we recognize
their distinction. Notice further that in the negative statement
implicated in the affirmative the qualities mcompatible with the
predicate, or with the attnbutlve of the affirmative statement,
arc not necessarily all kno\\ n to us, and arc at any rate not
postulated in a determmatc form. They cannot therefore condi•
tion for us the nature of the predicate or of the attributive itself.
In the negative statement on the contrary 1t is necessary that
the determmate character of the positive conception to wluch
the negative form 1s opposed should be known, because it is
only through this that the negative gain,; any meaning.
§ I q. All kno\\ mg and opmion ,, hie h issues in affirmative
und negative statement 1s subJectivc as the act of the thinkmg
~ubJect, but the sigmficancc of the act hes m its effort to appre-
hend obJcctlve reality. There arc, however, reasons which may
mclme one to the idea that there 1s somctlung specially sub•
Jl'Ct1vc about ncgatJVe statement ; that the negative is some•
tlung merely for us and not for the object, JUSt as uncertainty
and "ondcrmg arc merely subJcct1ve The negative statement,
denvmg a<i 1t docs its meanmg from the corrcspondmg affirma•
t1ve1 might seem subJectivc as rcprescntmg our mere failure to
pronounce the affirmative In an mquiry \\ luch ends ,11 the
statement (s)iA 1s not B, we first ask whether Bness can be
attributed of (s)iA Tlus rons1derat1on 1s obviously for the
tlunkmg subject only If \\ c decide that we cannot affirm (s)iA
1s B, that is the failure of an attempt of ourc;, and •;o may seem
to be nothing for the obJcct But such failure, though pre•
venting us so far from saymg that (s)iA 1s B, doer,, not justify
the statement ' (s)iA 1s not B ' , the failure bcmg subJect1ve,
the possibihty remains over that another attempt might succeed.
A more real difficulty is this: the negative statement may
attribute non-existence, and such an attributive cannot correspond
to redhty ; on the other hand, it ir,, difficult to undersland how
something merely subJcclive could c-..tend our knowledge, as
propositions of this kind certainly do.
266 STATEMENT, THINKING, AND APPREHENSION
Let us first consider the case where not all existence is
demed, but only a certam kind of existence. Now this simply
means the recognition JD the statement ' (shA 1s not B ' of the
distinction between (s)iA and Bness, and the distinction is
obviously obJecttve. For, 1f we make 1t merely subJective, we
should take all variety out of the real and end in a sort of
Eleatic paradox, that being 1s one without any manner of
difference.
Often the d1stmct10n between t\\o clements of reality is not
merely stated but elucidated positively by statements of the
pomts of d1ffcrcncc , hut m negative statement proper the fact
of difference l!t just i.tated without saymg what 1t consists JD.
l!.ven when the latter is stated, the elucidation of the pomts of
d1ffcrem.e may have to end, 1f everything 1s fully expressed, 1 m
such bare negation as ' Aness 1s not Bness '. Thus we say
'thrse t\\O flowers differ in colour', one being blue and the other
red. But of blue and red "e could only say ' blue 1s not red'
and 'red 1s not blue', and not wh..tt co11!tt1tutcs the difference
between them as we perceive them.
§ I I 5. The ord111.1ry negative statement therefore ' (s)iA is not
B ' concerns t \\ o clements "hich belong to reality, {s)iA and
Bness, and 1t !tlates th,tt the nature of (s) 1 1s different from
Bncss. Thus (s) 1 bcmg excluded from what 1s B, however much
we increase the sphere of D, provided 1t docs not become large
enough to mdude all reality, the negative !ttatement can retain
the charactern,tac \\hllh \\e claim for at. But when the difference
bet\\eeu B and ,ill reality v<1mshes, 1t seems as 1f the same
account could no longer be given \\"c luven't now .i d1stmc•
t1on !.,et\\ cen t \\ o re ah ties, for the statement mean'> that (sh is
nul rec\! ,ll all \\ hen Bm·ss then ts thus extended to cover
the \\ltok of reality \H' arrive .1t ,1 kind of hm1tmg case of the
negative i:,tatcmcnt
To solve our difficulty \\l' need but reft:1 to what was developed
m the d1scuss1on of the ' copula.' If 1t be said that ' (s)iA is
altogether unreal ' state~ a relation bet\veen reality and unreality,
and so gives a kmd of bemg to the non-existent, it seems enough
to reply that there can be no such statement because (s)iA
would have to be a mere word ,, 1thout any thought even to
• ~e the c.l1:1cu:.b1ou or the n.1tur.Al cxprcblo1on of negative thmkmg, § 117.
 Negation o, Quality of Statements 267
correspond. This in a sense is the true answer, but it requires
eluc1dat1on. There are statements where we cannot be quite
certain at first that any thought corresponds to their subject,
and such statements actually occur in the exact sciences. (s)iA
contains thinkable elements, which however cannot be thought
in that umty wluch the verbal form imphes for them. The
negative statement which verbally seems to deny all reahty to
the complex (s) 1 only dcmes thc1t the aforesaid elements can
stand in the given umty, 1t docs not deny reality to the clements
themselves Let (s) 1 be A ,rnd B and let the statement be that
(s)i is 1mposs1blc and unthinkable. Aness and Dness "ill be real
clements and the statement (s)iAB 1s 1mpm,s1ble will simply
mean that an A cannot be B. The except10nal form of the
limiting case thus disappears and ,,c recover the normal form
of the negative statement
Lastly, as to the complex &ubjcct 1b,clf, it :,cems as 1{ there
&hould be a conception of 1t, 1£ 1t 1s the subJect, and yet smcc
the conJunction AB 1s 1mposs1bk, even for thought, there 1s no
l om,cpt1011 111 the true &cnsc. 'I lu-, d1ffirnlty ctgam 1s solved by
a d11,tmct10n alre.i<ly md<lc between a wncept1011 m which the
clcmcnt5 arc thought or apprchrndcd and thc1r connexion also
thought or apprehended, and th.it form m which we tlunk the
elements but, 1mtcad of thmkmg the 11.iture of their conncx10n,
have only before u:, the problt•m or que:,t1011 whether they can
be connected Tlrn, lctlter km<l of conccpt10ns mc1y be con•
ven1ently r.illc<l' probh·mallr ', on the an.iJogy of the problematic
&t,1tcmcnt By llus <l1stmrt 1011 \\ c can undcrst:tnd how it may
be possible to bcgm with a seeming conception of (s) 1AB and
yet to discover th.1t no such thmg as (s)iAB 1s even conceivable.
It 1s the want of thi'> d1i;tmct1011 bet ween the defimte or deter-
mmatc concept10n,-the c-oncept10n 111 the proper scnse,-and
the problematic concept 1011, "Im h 1& largely responsible for
the d1fficult1es famih.ir both m ancient ,md modern times about
non-existence (or not-bemg) and the 1mposs1ble, and c1bout the
contrast of the merely unagmary with the real.
We have been cons1der111g mstances 111 wluch the clements
supposed to be combmed in (s) 1 presented either at first or after
investigation an obvious contrad1ct10n ; there arc, however,
statements of a more usual character which seem to affirm
268 STATEMENT, THINKING, AND APPREHENSION
non-existence and where we should not say that the grammatical
subject represents something never existent or wholly incon•
cc1vable.
Take such a one as ' the partnership of A and B is no longer
in existence '. The characteristic of such an example is that
,,c cannot call the partncr~lup inconceivable masmuch as it once
existed. The statement C'Cprcssei; the fact that A and B arc
not now partners, or m the relation of partnl'rslup this is the
ordmary negative, and fulls under the rule v..lm.h we have given.
But \\-e may treat the example othcnusc 1£ we use reality in
the wide sense a!> mcludmg the past and future as '\\ell as the
present, a me,mmg repre!>cnled by the tenselcss use of the verb
'to be'. The parlncrslup, being in the past, belongs generally to
rc.ihty, 1s ,111 clement m rc.ihty , and 111 the negative statement
\\c d1slmg111~h the time of its existence, a real tune, from the
prc!,cnt t nnc. llcre then again the account holds that we arc
d1!,lmgu1:.hmg 1111c p,irt of reality from ,rnotlicr • \\,'hen I say
'page 40 h nol 111 my book', 1£ Jt C}I.J~t::, no lnngcr as a page,
the same account m.ty be g1vrn of 1t ,lb of the d1s~olvcd partner-
ship. \\'e ob<,cn·c, hO\\C\Cr, th.1t m th1:, extended mcanmg of
o,ii,tence or reality, c, en 1£ the p,tgc a& ,1 pJgc 1s no longer
c:-...1stcnt, 1t nught uc said that the statement 1mphcs that the
page is m the '>phcrc of reality "luch 1s other in place than this
book no,,, ,, hcthc1 other m tune or not But 11evc1 theless m
making sud1 .i. i,tatcmcnl ,, e ,trc not tlunkmg of tlus general
scn.,c of C'\.l!>tcnrc, but of prc5cnt cx1stc-11rc, ,mcl our statement
docb not m·n·ss1tatc th.it the page should be m tht' sphere of
prcbcnt cx1:,tcntc outs1uc the hook ln .111)' ca'>c, the mtcntion
of the i:.tatcment 1~ uot to ckcl.1rc that the page has any kind
of cxi&tt:ncc but to d1i:,tmgu1sh 1l from ,t certain kmd of existence
§1r6b From thr stanc.lpomt "r ha,c nO\\ reached ,,c may
appropri,ttely l'On<;1dcr the trul' 11,1turc of the 50-callcd' 1magmary
quant1tie5 ' m m..ithcmatu''- Suppo&c 111 a given problem we
are asked what quantity can s.tt1sfy certam cond1t1ons It may
happen that there 1s no such qu..intity, and that to assume it

[• 'When I SoiY" page 40"' • Rcron.1dcr ', ~IS 11ote

b This lS one of the l>CCbom, the hnal po-,1tlon of which was doubtful
'fl1c author thought of putting it at th" i,n<l ol Part III and before' Induct,on •.
I tluuk he printed 1l bcrc to hlx•rak hi. ~oul Cf pp cxu-xm J
 Negation or Qualit,y of Statements 26g
would be to assume an impossibility. The problem may be
expressed in the form of an equation with the unknown quantity
sought for represented by a symbol as x. The calculus can only
show 1t 1mposs1ble that there should be such a quantity by
conducting to an operation which we recognize as impossible,
e g. x = sin•1 b' where i is greater than unity and therefore xis
determined to be an angle whose sine is greater than unity, or
agam x = ✓ - 1. Such results are answers m the negative and
assure us that no such quantity can be found ; but they have
a value and function beyond that. It 1s found that calculations
can be made from these 1mposs1blcs "h1ch lead to correct results.
Thus they are not mere negations but have a positive character.
If we equated them to zero \\C should get a twofold false result.
In the first place, to equate x to zero docs not show that what
we are seeking 1s 1mposs1blc Zero m the problem may refer
to quantity measured m some definite way and, 1f the quantity
fiought 1s the distance of a pomt from a given pomt, e g. from
the • ong10 ', the answer ~ = o docs not mean that the position
of the supposed point u, impossible, and so that there 1s no such
pomt, but that its pos1t1on 1s the same as that of the given point
or • ongm ' itself. Secondly, 1f we equate 1mposs1blcs to zero,
we thereby equate them to one another, and the result of th1-,
m algebra \\ ould be that all real quant1t1es \\ ould have to be
rquated to one another In the calculus itself these 1mposs1bles
arc otherwise found to differ from one another, which is rnough
to prove that they cannot be regarded as mer'- blank negations,
a mere answer ' O() ' to the question proposed 10 thr problem.
Thus 1t looks a-, 1f 10 unreality itc;elf we were able to distmgu1sh
elements different from one another, and so the realm of unreality
seems to have gamed a kmd of bcmg. Now this gets expression
m mathematics m so far as mathematicians have long been
accustomed to call these 1mposs1blc quantities (rather • 1mpos-
sibilities of quantity ') • 1magmary ' quanllt1cs, and all true
quantities, m d1st10ction from them, ' real ' quant1t1cs. The
nomenclature itself is unfortunate, for, as we shall see m the
examination of geometrical reasoning, 1t 1s only by what might
be called the test of the 1magmat1on that the mathematician
can decide at all whether any gcometr1cal relation 1s real. Thus
~70 STATEMENT, THINKING, AND APPREHENSION
there is a grave implicit contradiction. Besides it is charac-
teristic of such quantities that they cannot possibly be imagined.
Yet clearly the erroneous distinction is intended to save these
expressions from being treated as mere ' nothmg ', and so far
is m the interest of an important truth. Fortunately also the
wrong termmology causes no error m ordmary mathematical
procedure, ,vhich treats the thmgs so misnamed m a legitimate
manner.
The account W<' have given of negative statements and
espel'ially of those which assert imposs1b1hty seems to give an
adequate solution of the difficulty The negative statement 1
alV\•ays presents two reahtJes as d1ff ermg from one another, and
the statementc; which assert 1mposs1b1hty for their subject of
attribution are all reducible to this form, namely, a distinction
between two reahtie<; Now the difference of two realities, or
their incompatibihty, may be itself different from the difference
or mcompat1b1hty of another pair of realities Agam such
incompatlbihty or difference is exactly what an 1mposs1bihty
means It 1c; because Anesc; and Bncss arc mcompatible that
1t is impoc;s1bk for an A to be a B, or, what is the same thing, 1t
is because of tlus incompat1b1hty that an AB 1s an 1mposs1b1hty.
Impossib1hty 1s only the other side of necessity. for if what
1s A must be C and what 1s C cannot be B, then an AB 1s an
11npos<i1b1hty and, moreover, a defimte kmd of 1mpossib1hty
bccau-,C' of the defimtc positive character of Anc-,s and Bness.
The d1fkrencc of Dncc;s and Eness may be of a different kmd
and so be the ground of another kmd of 11nposs1bihty. The
d1fferC'nce then between t\\o such differences corresponds to the
difference' between two 1mposs1b1ht1es The mystery therefore
vamshes, for "c thu<i understand that the 1mpossibilit1es a.re
not ' nothmgs ', but have so much defimteness that they are
distinguishable from one another. Through this they get a
determinate character, "h1ch however docs not make them real
objects or even imaginary objcrts Their determinateness,
neverthelesc:., 1s grounded m reality, for they mean actual
and real differences bet,,cen real elements If Aness is really
different from Bnesa so that an A cannot be a B, it is also true
that 1t is reall·y impossible for an A to be a B, in other words an
a I n5.
 Negation o, Quality of Statements 171
AB is a real impossibility. But that does not make impos•
sibilities such as the impossible points in Geometry into real
objects, i e. into what is not impossible.
The' imaginary' points m Geometry, in which two curves which.
do not really cut are said to cut one another, are the expression
of the particular way in which the given curves fad to cut.
Their defimteness, measured by the definiteness of the mathe-
matical expression of their co-ordmates, does not mean that
they are m any true sense imagmary, but merely that the
impossibility of the one curve cuttmg the other is a dcfimte
kmd of impossib1hty, wluch can be dcfimtely measured, and
mathematics only determine the ddimte measure. Vlc can now
see that there is no paradox at all m the fact that from these
so-called imagmar1es, which arc 1mpossiblcs, \\C can derive true
statements about realities. For to every reality there corre•
sponds a set of necessities which dctermme it and make it what
it 1s To these again correspond a set of imposs1bihties, which
lead back to the necess1ttcc; from wluch they arc derived, and
therefore equally well serve to define the given reality. As the
set of necessities wluch define one rcahty d11Tcr from those which
define another reality, so also doc" the set of 1mposs1b1hties
correspondmg to and definmg the one reahty differ from the set
correspondmg to the other reality Tlus 1s the simple rationale
of the so-called '1magmary quantities' and of the mathematical
treatment of them. It 1s qmte futile and a serious misunder-
standing to try to JUSltfy the use of 1magmarics by certain
methods of ' mterprctatlon ', which merely mean that by giving
the symbols another significance a given formula ceases to
represent an imposs1b1lity altogcther. 1
§ 117 The apprehension of the definite character of anything
involves, we have said, distinction from other thmgs, and so
111volves what may be matter of negative statement To recog-
nize red as red, I must rccogmze 1t as a colour d1stmct from
other colours, distinct from those I have seen, and this carries
with it, on reflection, that 1t is distmct from other" I have not
• The above analys1oa; was supplemented by the 11lustratlons given in lecture.
It may be added that on the pnnc1ples here mamtamed 1t 1s possible to
supply an admitted desideratum 1n pure geometry ~d make use of the
so-called imagmary pomts and Imes without dependmg upon the results of
co-ordinate or algebraic geometry.
272 STATEMENT, THINKING, AND APPREHENSION
seen. Nevertheless distinctions thus necessitated by the definite-
ness of what is apprehended do not normally give rise to
negative statements. Thus though the apprehension of green
involves the apprehension of its distinction from red, and, say,
sweet, it is not natural and normal to put this m words m the
form of the statement ' red is not green ', or ' red is not sweet '.
It may be done for some rhetorical purpose but not for the
purposes of knowledge. Such forms strike us as no less nugatory
than the so-called analytical judgement, and that is why we
don't use them.
This fcclmg 1s well-grounded, and these verbal negative state•
ments have a real affinity to the analytical Judgement in one
rec;pect. We cannot 'form the idea of' (i c. apprehend) the
definitenec;c; of anything (e g. of red) apart from distinctions
from certam other things (e g. colours). Now these distinctions,
which are thus necessary to 'forming any idea' (the apprehen•
sion) of the thmg in question, we do not express m the verbal
form of statement, becauc;c they are presuppoc;cd in the idea
(apprehension) of the logical subject, \\hereas the obJect of
a verbal c;tatcmt•nt I'- to convey something new about the logical
subject of the c;tatement, and not somcthing which 1s necessarily
known already Thus, though we distinguish 'straight' and
' curved ', and ' straight ' and ' hot ', we do not naturally say
' straight 1ci not curved ' or ' straight 1s not hot '.
When then 1s the \. erbal form of negative statement natural
and normal ? When do we naturally say (s)iA 1s not B?
Clearly when our eoncept1on of Anc% does not necessarily
im•oh·c for us the d1stinct1011 from Bness, or the absence of
Bnes1; \Vlwn the expression ' (s)iA 1s not B' 1s natural we
may d1stingu1sh t \\ o cases The statement may correspond
either to the apprehension of something m (s) 1A wluch excludes
Bness, or to the mere observation of the fact that Bness is
absent from (~) 1A The first case 1s of the form (s)iA is C,
where Cness excludes Bness , ' this substance does not show
blue colour m the ftamc of the blowpipe'. We arrive at this
by observing that the colour shown m the flame 1s, say, red.
Why then ha\'e the negatn·e statement at all, and not the
affirmative which tells us more and is fully adequate to the
thought behind the expression ~ The negative 1s not adequate,
 Negation o, Quality of Statements 273
for if I say the colour is not blue I do not say what colour it
is and I omit besides something which I know, which also is
the reason for what I say.
The ground of this is one which we shall meet with again in
the case of the particular affirmative proposition. It depends
on our subjective interest, the object with which we conducted
the observation or experiment or inference. We state only what
matters to that oh1ect, "hat 1c; relevant from our point of view.
I may want for some purpose to use some particular kmd of
substance. I find a substance with some of the apparent pro•
perties of the given substance, but one of the indispensable
tests, say, is that it shows a blue colour m the flame of the
blowpipe I find that it shows a red colour, but all that is
important to me is that 1t 1s not a blue colour I am not
interested in what the obJect 1s, now that I know it is not of the
kind I want. And similarly 1t may be sufficient in my argument
to state only that (shA is not B, though of course 1f I am chal-
lenged as to my grounds I must give the affirmative statement.
The second case is that m wluch (s)iA is merely observed to
be without Bness, an attribute compatible with Aness Since
this case belongs to empmcal observation \\ c cannot always be
sure of the absence of the given quality ; we sometimes are
sure only of the absence of any observation of 1t Yet we may
not fully realize this and so may come to assert without quah•
fication that (s)iA is not B. When we are sure that Bness is
absent do we observe mere negation? We have seen already,
10 the discussion of the ' negative conccpuon ', that this is
impossible. We should apprehend an A, viz. (s)iA, which was
without the attribute Bness This so far 1s not the apprehension
that Bness is absent, any more than it is of the infinite field
of the other qualities wh1rh are absent We need to have our
attention somehow directed to Bness and then we become aware
of its absence. 1 Now this means that we apprehend the nature
of the particular A before us, (s)iA, and the nature of Bness,
and see that Bness is different from that ,vJ11ch (s)iA 1s. Thus
the apprehension of the negation and of absence is after all the
apprehension of two positive realities as different from one another.
1 To find out whether Atkins 1s m the ranks, we have to observe each
rank and file and see that he 1s not Atkins.
1773•1 T
 XIII
ERRONEOUS ATTEMPTS TO DEFINE 'JUDGEMENT'
§ 118 • WE \\ere led from the doctrine that the 'copula.' is
the sign of a subJechve artiv1ty, called pred1cat1on, to a doctrine
which explains 1t by the obJective side of thought. For we
maintained that the \\ord '1s' refers to the bemg of the object,
and decided the question about the identity or difference of the
subJect and predicate by reference solely to their objective
meaning We now return to the subJechve side and ask the
question whether m apprehension there 1s some sub1ective syn-
thesis of subJect1vc material ·Judgement', not m its proper
sense but in its erroneouc; use for a fict1t10us 1 mental act sup-
posed to correspond to every c;tatemrnt, I" not infrequently
resolved 11110 <onrrpt10ns, c;ometimrc; called ideas, and then
defined as ..i. r,ynthes1s or combmatwn of ideas or ronceptlons;
the latter bemg regarded as clements of Judgements, not them-
selves Judgements ~ In tlus there 1s c;omethmg true, but the
account given 1s most m1sleadmg and inadequate It 1s often
said that the verb,11 expression of a propoc;1t1011 1c; a symbol or
representation of thought, ..ind thought agam 1s somehow made
to represent thmgs, but 1t ,s not an adequate account, even
of the verbal form, to say that 1t represents thoughts; the
value of the words oftenest h<'!> m their mcanmg thmgs. 'Glass
is elastic' does not st,md for .i i;ynthe!>1s of symbols 111 c;ome
one's mmd: 1t means that real glass has real clast1C>ty 'The
spade 1s m the ground', to uc;c- Mill's 11lustrat10n, 3 docs not
mean that one of my 1clras 1s m another Yc-t, wlulc the' Judge•
1
ff 38, 41, and 44
s For early forms of th1-. kmd of v1t>w i.ce §§ 84 and 1::?u
• Mill, A Sys/rm of l.og11:, Bk I, th 5, § 1
[• Cf • Jenem bloss subJektlv'leyn'!ollcnden Smne des Urtheds als ob Ich
emem SubJektc em Prii.dtkat beilcgte, \\1den.pncht der v1elmehr obJecbve
Au'!druck deo; Urtheils die Rose 11,t roth, Gold 1st metal! u s f ; n1cht Ich
lege 1hnen et\\a'! erst be1' Hegel, Log1k, Enr, § 167 (Works, VJ, p. 329)]
 Attempts to Define 'Judgement'
ment ' is not a mere putting together of ideas, we cannot say
it is a putting together of realities. It might then be suggested
that the truth lies between these extremes and that, as a sub•
ject1ve act, ' judgement ' is a combmatlon of ideas wluch means
a combination in reality. Here we should have to ask what
these ideas are, what is meant by combmmg them, and what
is meant by' means'. If we push our qucst10ns home, we shall
often find that the idea or the 'ideal content' turns out to be
an mdiv1dual mental image or an imagined 111d1v1dual. Now,
if I say 'the 'lpade 1s m the ground', 1t 1s true that I may
have mental pictures of thr ground and of the c;padr, and these
may be combmed m so far as my picture of the ground mrludes
my picture of the spade But I do not mean to state that when
I state my ' Judgement ', nor 1s 1t these ' 1deac; ' that I am
thinking about. No doubt I probably thmk that the reality 1s
somehow ltke them and yet I do not rely on the pictures even
as representat10ns of reality. But, 1t nught he obJccted, these
ideas mean reality What do we here mean by ' mean ' ?
A word means usually sometlung unlike 1tself Is 1t mtenderl
that these p1cturrc; mean reality m the senc;e that they arc like
reality? If it be said that the idea 1s hkl' reality then we must
reply· that 1c; not the meanmg of 'meanc; '. If one horse 1s
like another hor:,,c, we cannot :,,ay that the firc;t means the
second. Nor agam can we say that these 1dea'l mean reality
111 the sense that we take them for reality for we do not. It
may now be admitted that such ideas and their combi11ation
are not what I am tlunkmg about ancl that they do not mean
rrahty, yet 1t may be suggested that these very 1dras m their
combinat10n are my thought about reality. Now that would
mean that to thmk about reality 1s to have such ideas or mental
pictures before us ; but that agam 1s not so, for the ideas
whether m combinat10n or not are of no use unks'l I do some-
thing more than have thcm-i e unless I at lcac;t think that
reality 1s somethmg hkc them. It turns out then that 1t 1s
essential to what 1s here mtendcd by 'Judgement' that there
should be some activity of thought other than the presence to
the mind of ideas in the sense of mental pictures in combmation.
It 1s important to notice that thic; indispensable activity is quite
unprovided for m those views of the nature of thought which
T2
276 STATEMENT, THINKING, AND APPREHENSION
are common in empirical philosophers, like Locke and Hume.
For them, on their presuppositions, there could be nothing
but a series of mental pictures: there could be no place for the
activity of thinking or of judging. But of this they are often
unrono;ciouc;, and 1t 1s to the crecht of Hobbeo; • that he clearly
rcahzcs that accordmg to }11s dortrinc thought could he nothing
but the surression of mental 1magec;
§ 119. It 1r,, essential to r,,talement and to ,,hat ic; styled
'judgement ', that it should be true or false, thr combination
then of ideas, 1£ it 1s eqmvaknt to the statement or the opinion
thereby expressed, must be true or falsr If I think wrongly
that ' Williams 1s in h1,; rooms ', no doubt I have before me
an 'idea' of the room,; and of W1lhamo; in them as mental
pictures. Still my m1c;takc dori; not Ii<' in tl11s presentation to
conc;c1ousness, but in somethmg elsr, that 1s m my belief that
reality somc-ho,..,, correspondc; to th1c; combmat1011 Thus truth
and falsehood would hrlong to snmc-thmg which 1c; indeed in
thought, hut wl11ch whakvc-r cI.,r 1t may be is neither these
ideas nor their rombmat10n But 1t may well be urged that
truth and falsehood do bC'long to ideas, m c;o far ac; these stand
for somethmg d1stmgu1shed from reality , 1{ then such ideas are
not true or false 111 their combmation, what 1s true or false m
the matter~ Again, we may he remmded that when a 'Judge•
mc-nt ' 1s wrong or improbable 1t 1,; often said to be only an idea
or a 'me-re idea of ourc;' If this meanc; that the combination
of mental pictures 1c; wrong or untrue, we must reply, as before,
that truth and falsity rannot apply to that at all We must
ronredc- of rour<1c- that "hat 1c; "rong may be naturally and
propc-rly railed ' ufr, ' , 1t 1,; indc-cd ideal, but it 1s precisely
that act1v1ty of thought "lurh 1s other than the combmation of
thec;e mental p1cturc-., It 1s the belief that certain real elements
are combined, or, 1f we w1<:h to re-late tlus to the mental pictures,
it 1s not their c-ombmation 111 the mmd, but the belief that there
1s a combination m reality somehow s1m1lar to it This 1s what
we mean by ' idea ' when we say that our idea 1s wrong. It is
[• e g ' the perpetual ansmg of phantasms, both m sense and 1magmation,
is that which \\e commonly call discourse of the mmd' and' Dlllcrenharum
autem observabo non est a sen'ltone propne dicta sens10 altqua distmcta'.
Elem Phsl, Part IV, ch 25 (l\foles\\orth's ed, Eng 1, p 399, Lat 1, p 325)
Cl H11man Naturt, ch 1v, § 1, 1b Eng iv, p 14)
 Attefnj,ts to Define 'Judgement ' 2'17
not a 01ental picture and cannot be reduced to any terms of
mental pictures. 1
If "c now drop the term ' mental picture ' and say mstead that
the 1dcas combined in the ' Judgement ' arc our ' concept1ons '
oi realities correspondmg to the subJect c1.nd the attributive
(leavmg the term ' conception ' for the moment uncriticized), 1t
m.iy be that Judgement or apprehension or opm1on mvolve1,
thc1,e ct1:1 d1!>tmguisbablc elements m gcncr..11. But we cc1.nnot
represent the Judgement (m the proper sense) by merely !>aymg
that 1t 1s puttmg such ideas together We ..1rc bound to say
what sort of putting together "c mean , for the expression
'puttmg together' 1s m itself too Vd.gue to tell us anything,
bemg only a metaphor derived from putting obJects together in
space. Now such puttmg together of ideas as we here really
mean 1s simply Judgmg that the obJcct to wluch the one idea
refers possesses the kmd of bemg to wluch the other refers ; so
that, 1f we .isk what kmd of puttmg together judgement is, we
bc1.ve to Ube 'Judgmg' to explam it, and thus come m the last
resort precisely to the md1v1!>1ble act itself of Judgmg, as appre•
hem,ion after mquiry, an act wluch cannot be further elucidated
or described, cannot be resolved mto parts nor represented d.b
m.ide up of them. The not10n of Judgement H, umquc, 1t cannot
be reduced to any other dcnonmMt1011 We must simply rccog•
mte it 1n its um'vers.i,l chard.l.ter through mstancci. Ill wluch we
c.ll.l.:rcisc it. A bllllllc1.r account holds of opmion.
§ 120. We may now treat the question from the pomt (If view
of what has been said previously of the relaLion of thought to
itb obJect. For clearness let us take (s)iA ii. B m the case when
we know that (s)iA is B. Can tlus be rightly called a combma·
t1on of our concept10ns of (s)iA and B or Bness ? What arc
these conceptions ? Mental images we have excluded. Suppose
it be said that they are some entitles, though not images, cxiating
in the mmd. !t may be answered thc1.t nothmg could be found
1
To avoJ.d. nusunderstandmg, observe that an cxa111pk hai. been taken
where • mental images ' have a real place , the error m the statement
' Williams ii. m lus roolllll ' usually regards a statement made when we are
away fron1 Wdbanu.' rooms, so that it does not express a pre11ent experience.
When the same statement cxpreases .i present expcnence it has notbmg to
do with the mental images. They belong to the <..~e where we are presumed
not to be in tbe roollll,,
278 STATEMENT, THINKING, AND APPREHENSION
to correspond to these words except just the mental images.
But, even if verbi causa we allowed these nondescripts, the
attempt to represent 'Judgement' as a combination of them
,., ould fall under precisely the same kmd of criticism as that
which has been passed on the combmation of mental pictures.
What mtelhgiblc scn!le can ,~e give to the phrase 'conception
of (s)iA ' m tlm Judgement of knowledge ? It 1s m the case
of kno,., ledge Mm ply the ,tpprche11;;10n 1 of the reality or object
(s)iA itself ,111CI, if what is apprehended be called the content of
the .1pprd1ensio11 or conception, this content is Just the reality
itself. Now that bcmg the kmd of meanmg wluch the concep•
tions of (s)iA and B or Bness would have, how can v. e combme
them and of \\-h.i.t kmd ,.,ould the combmatlon be~ For com-
bmat10n mm,t be of i.ome detimte kmd. We cannot m the act
of knov. mg (whether Judgement or not) be combmmg the
' contents' of the apprehension, for these arc the objects (s)iA
and Bnei.!:i, .and the i.t,ttcment of our knowledge does not mean
that v. c have effected .i.ny 1..ombmat1on of objects But now 1f
,,c .ibstr.ict the content of tl1c ,1pprchcns1ons, 1 e. 'what 1s
apprehended ', there ib nothmg left to combme, for apprehen•
s10ns without anytlung apprehended arc entirely empty.
§ 121. We lllclY arnvc at the S,Lme rc!>ull m another \Hty,
Con!>i<lcr the conu~pt10u m 1b fullnec,s a!> really v. hat it 1s, with•
uul ab!:itraclmg the c1pprchcn!>1011 from \\-h..it 1s apprehended,
.tb the .1pprchcn!>1011, that is, of the n<1.tures of (~hA and ll.
Bow can \\e combmc !>uch .tpprchcns1011i. 111 their full nature?
·1 he combm.ttwn \\ oul<l Juve to be of some defimte kind ,u1d
1t \\ould be our i.uLJcct1vc act Now \\C might combme them
m the &ense of tlunkmg one after the other m tune, or it m.ay
be thmkmg them somehow simultaneously ; we may, that is,
hc1ve the apprehension of (s) 1A along with that of B. But now
clearly neither "c:1y of combmc1t10n i!> havmg the Judgement th.i.t
(s)iA is B. Ck.i.rly ,Ll!:io, .i.s the 11cLture of the object 1s mseparablc
from the apprchens1011 (as its so-called content), any other pos-
sibdity of combination, except that merely temporal one, must
depend upon the nature of the obJects or content, and on the
1 This IS the conception whlch occurs 111 knowledge there is another Jaud
which we have called • problcmat.Jc •, but the c.ons1derabon of the first IS
enough for the present purpose.
 A'ttempts to Define 'Judgement '
combination of which they are capable. But that is an objective
fact independent of any subjective act of ours, and so it is not
a combination which we make. It can only be recognized or
apprehended by us, and the Judgement of knowledge, or other
such act of knowing, 1s such apprehens1011. Apprehension itself
1s obviously ultimate Everythmg we can say about 1t, or
mdced about anytlung eh,c, prc&upposcs 1t, 1t 1s futile therefore
and J. mere fallacy to profess to explam the ..ict of apprehension.
The question may be asked ' 1s there any sense 111 wluch the
conceptions of A and of B may be said to be clements m the
thought corresponding to (s)iA 1s B ~ ' Only 111 tlus sense, that
to form the thought, the Judgement or opmion, I must .i.pprehend
(s)iA and Bness. But my thought docs not consu,t of these
apprehensions , it 1s not a mere comb111.1t1on of them. It 1s
a new J.pprehcns1on , m fact the apprchens10n of a relation
between the obJects of the afore&,ttd apprehensions In any
leg1t1m.i.te sense then of the word combmatlon we cannot com-
hme our conceptions or .i.pprchens1ons at .i.11, much less represent
'Judgement ' or other act of ours as such an act1v1ty on our part.
The &tatement none the less does refer to some combmat,on
or other, to some umty or other of different clements, or the
m1stcl.kcn vww would nc, er h.ive ,mscn And what it docs so
refer lo 1s obviow,ly ,1 combm.1t1011 of the nature of (s) 1A with
that of Bness, or more correctly a umty of some sort between
them. Now, wlulc we apprehend (s) 1 A and Bness, that 1s, have
our conceptions ot them, "e also .ipprehend their relation and
umty. Tlus 1s the Lorrect d.ccount ; not ~h.it we unify our
J.pprehens1ons of A and B, but that we apprehend the umty of
A .incl B, a umty of .i ccrt.im kmd. This m.iy be put shortly,
thus -In the Judgement of knowledge and aLt of knowledge
111 gener ..Ll we do not comb111e our .i.pprehcns1011s, but apprehend
a combmat10n.
§ 122. We may no,., turn to the cons1derat1011 of ,1 modern
theory of so-called 'Judgement•, which has enJoyed considerable
vogue, and illustrates the fut1hty of the attempts to explain or
define 'Judgement' m terms of anytlung but it!>elf, and pat•
ttcularly m terms of ' ideas '. Though apparently h1ghlv modern
in form 1t is only the result of a mistaken tendency m philosophy
which 1s quite clnc1ent.
 28o STATEMENT, THINKING, AND APPREHENSION
This modern theory of the ideal clement in judgement defines
judgement 1 as follows: 'Judgement is the act which refers an
ideal content (recogmzed as such) to a reality beyond the act.' 2
The ideal content 1s then identified with what 1s called the
' logical idea '. ror the purpose of tlus definition idea has in
this theory h,o &1gmfic,it1ons. (1) A mental image, e.g. of a
particular horsc-.,omct1mcs called 'the psycholog1cal idea'.
(ii) The 'mcamng' of tlui. ment.il unage which=' the log1cc1.l
idea'.
The mc.inmg of (1) {wluch mcc111111g = 11) 1s said to be a parl
of itself, e.g. the mmtal image of a particular horse means
horscncss, '"luch 1::. supposed to be a part of the content of (1).
All tlm,, .mcl the theory built upon 1t, depends upon an
erroneous an,1lys1s of i.uch terms a& • s1g11 ', • symbol ', and
• mcanmg '. • Sign ' we hncl defined thus . ' Any fact th.it
has a meanmg, ,mcl mec1mng consist::. of a part of the content
L1.c of the &tgn itself, or the fact 1to;clf], cut off, fixed by the
111111d, aud con::.1dcrccl ,tpart from the c:i,..1stc11cc of the sign.'
In the Jirst pl.ice, tlm, cxpl.m..1.lion h,u, notlung to do with
' mcanmg ' or ' sign ' or ' symbol '. So far as it describes
anything .1t all, tt dcscnbec., an .ict of ..i.bstract1on wluch may
be either (a) the c.1bstr.tct1011 m \\luch \\e thmk of a universal
apart from the p.irt1culdr 111 ,vluch it ii. mamfei,tcd-horseness
.Lp.trl from a pdrticul,tr horse, (b) that 111 \\]111.h we tlunk of
c1.n md1viduc1.l clement m .m md1v1duc1.l "hole apart from the
other clements-cg. the horse'& tail, the surface of a given sobd
apart from volume, "eight, &c.
Secondly, the last clau::.e of the definition of sign 'considered
apart from the existence of the sign ' 1s self-contradictory :
when we han re.illy to do with a s1gn "hich has a meanmg,
the mcanmg can only be thought of as a meanmg m reference
to that which has 1t as .t meanmg-v1z. the sign In short
' meaning ' 1s meanmg of
Tlurdly, the sign .is c1.bo, c defined turns out m the sequel to
be a symbol or natural sign 3 cl.& opposed to an arbitrary sign.
1
Judgement 1s used throughout in the loose sense which has been cnt1cu:ed
1n preVIous lectures.
1 F. H. Bradley, Pmuiples of Logic, 1, ch 1, § 10 (ut ed, p. 10)
1
Tim. 1mphcs that symbol and uatur.il sign are convert1ble terms, and yet
\\c imd a{Lcrwards th.tt a natural ~,gu 1~ m.i.dc \\1d1.r t.han a symbol.
 Attempts lo Define 'Judgement ' 281

This naturalness, apparently, can only mean that the significance

is a part of the sign's own nature, e.g. cunning-of the fox's
nature (an example given by the writer). The mathematical sign
x h.t.s not this charc1.cterist1c, but to oppose on that account
'natural' to' arbitrary' 1s to be misled by an amb1gu1ty. The
symbol 1s not natural in the &ense m wluch natural 1s opposed
to arbitrary: but a i:.ymbol ai:. ..i. symbol lb arb1tr,try All signs
and symbols have a meaning only by our mb1tr..try dec1b1on:
none of them mean anytlung of themselves · 1t 1s we who 1nea1i
~omething by them. A picture or image of J. fox docs not in
itself ' mean ' cunmng , we mciy arb1tranly gwe 1t such a
meamng, .i.s 111 "' political c.i.rtoon, .md "c arc understood
bec.i.use people arc already aware of such wnvc11t1ons In
another context the fox picture might be understood to have
been chosen for some very different purpose-c.g. lo symbolize
J. colour.
Fourthly, we may ..i.sk-\\hJ.t part of a content lb the me.i.nmg?
e.g. why should the meaning of a horse-image be horseness-
a leg 1s a part of the content why 1s not horse-imcLge the
bymbol of a leg? We notice that, 111 the 11lustrc1.t10n of the fox,
the fox-image happens not to be t.i.kcn for the symbol of foxnei:.s
but of cunnmg. To the quest10n why tlus 1s so no direct am,wcr
I& given, though 1t 1s pcrhap1:, 1mphcd or may be got from the
l.iter d1stinc.t1011 of J. <,1gn wluch '>t.i.nds directly from one wluch
:,t.i.nds indirectly for its mcanmg, e g 1t might be i:.a1d that the
fox-unage &tands directly for fo:i..nes1:,, c1nd 111d1rectly for cunnmg.
WhJ.t such a d1st111ct1011 could mean the author does not even
attempt to elucidate, and the truth 1s that the d1stmct1on
between standing ' directly ' and ' indirectly ' for a meanmg will
not bear any cxammatlon. What could be meant by ' standmg
md1rectly for'? Agam, what is the maximum part of the con•
tent, which can be the meamng or a meanmg ot the given
mental image ? Why, mdecd, 1f the mental image can mean
a part of its content, should 1t not mean the whole ?
Now let us consider the appbcat1on of the given theory to
the mstance of horseness.
(1) Horscness is obviously not the meanrng of the mental
image, and 1£ horseness is a part of the content of the mental
image of a horse, it ii. so only as an abstraction from lhe particular
282 STATEMENT, THINKING, AND APPREHENSION
in which it is manifested. But now that is impossible, for
horseness is manifested in real horses, and not m the mental
image.
(ii) The image could not ' naturally ' mean this abstraction
from itself. If IL did, m any given case, 1t would be because
we had arbitr.mly chosen that it should. But this never takes
place: \\ e do not mc.m universal horsencss by the image of a horse.
But now let us nevcrthclcsb asi.umc these untenable distinc•
t1011s and apply them to what lb said of the Judgement.
Fu:st, the 1de,L \\ e c,illcd i\O (11), or 'me.mmg' as = what
i& mc.i.nt, 1s of course the reality. This 1s dear from the state•
ment that ' we never assert the fact 111 our heads, but something
el:.e which the fact stands for '. That • somcthmg else ' 1s the
mcamng, =what 1s meant. Now 1t 1s said m the context that
there lb no Judgement \\lthout ideas, because Judgement is
true or fa.be, ,md ' truth ,md falsehood depend on the relatio1t of
our ideas to reality '. \\,'llctt CJ.n ' 1de,i ' me.i.n m such a btate•
menl? Not idea No 1, nor docs tht· ,iuthor :.uppose 1t to me.in
llus. The relation of these to Rcahty c.mnot detcrmme truth or
falsehood , th,tt can only be done by my Judgement about their
relation to Re.1hty. C.1n 1t then be idea No 11 ~ Now that idea
1s the meanmg, ...:reality meant, but m tlm, wntext the idea
bas to be d1stmgu1shcd from the rcc1.hty, otherwise "c should
have tlut ' the 1clJ.t10n of l{eJ.lity to Reality 1s true or false'.
Neither then of the t\\ o sense& of idea 1s possible, and the
theory rdute:. itself TJ111-, 1:. .in e'\.ecllent 1Ilubtrat10n of the
truth th.it 1£ \\C push home the que1ot10n of ,.,h.i.t the ideas really
arc by "h1ch Judgement 1s l'xplamed we find in the end nothmg
but the mental image , evcrytlung elbe turns out to be mere words
aud to resolve 1t!iclf mto the mental image
If \\e ,t1ok, ho,\cvcr, ,\h,1t reasonable sense the word 'idea'
bas 111 such c1. statement ,is th.i.t "luch "e arc crit1c1Z1ng, the
ans\\cr dearly 1& that 1t :.t.inds for Judgement or op1mon;
neither mental pictures, "c m..1.y repe.it, nor what they are
supposed to mean, can dctcrmmc truth or falsehood-only Judge•
ment or opm1on can detennme that. Thus to say that Judgement
is 1mposs1blc without such ideas 1s to say that judgement 1s
1mposs1ble without Judgement (accordmg to the ,u1tcr's use of
tbe word Judgement).
 Attempts to Define 'Judgement ' 283
Again, 1t 1s said that we do not assert the fact in our heads,
but what it 'stands for', and in the same context ,ve find
thc1t the idea lld.s to be the idea of some existence Here we
find the same interpretation necessary for ' idea '. My idea
of &omcthmg 1s not the mental image, nor its mcanmg (if il had
any) : my idea of A 1s the Judgement or opinion that the real
A has the rcc1l quality B, and the relation of thc1t to the mental
image 1s obvious .md lus c1lre.tdy been dn,cussed.
Agam, notice a certam confm,1011 mtroduced mto the second
mc,uung of idea-1de.i. No 11 It 1s the me.inmg r,ymbohzed by
No. i idea and must therefore be the reality. 1low, then, can
1t be called 1dcc.1. ~ For throughout the oppositwn of idea to
RcJ.hty 1s mamtc.1.mcd, and we arc told th.i.t these ideas (1 c. of
No. u kmd) arc the ideas wluch Judgement re.illy uses, and they
therefore must be the 1dec1s winch detl'rmine truth and false-
hood by their relation to Reality, and c.1.s such arc d1stmgu1shcd
from Reality.
This may have come about through ,t Lonfu&1on of t\\o ::,en&es
of' mcanmg'
(a) The tlung me.mt In the c:,,.prcsi,1011 ' these t\\ o senses
of idea a'> the symbol and the symbohtcd ', "What has
Leen called the meaning 1s no\\ c.i.llcd the i,ymbohtcd,
,111d therefore =-what 1s meant.
(b) the mcanmgncss of the f;1g11-1 c the fact tL.tt the sign
h.u, a mcanmg-the property of the sign of havmg
,l mcanmg Tlui, 1;, at lc,u,t so far on the wJ.y h being
an idea that 11 1;, not the reality mL..i.nt, but .i property
of the symbol wluch 1s idea
ln the account of Judgement we ,ire told that ' we do not
use the mental fact !or 1magl'], but only the meamng ', That
perhaps agrees with the statement that ' \\ c never assert the fact
m our heads, but sometlung else which the fart st.i.nds for'. The
question arises, what 1s the use of the mental 1mc1ge at ..11l, for
it seems from the above that we do not use 1t ; and clearly
according to the second of the ,~bovc i,tatemenls (' we never
assert, &c.') we get at 'what 1t stands for', and make our
judgements about that. Again, therefore, \\e must a"k what is
the use of the mental image? Upon the statement that we do
not use the mental fact but only the me,.uung of 111 there arises
284 STATEMENT, THINKING, AND APPREHENSION

another difficulty. The meaning of the mental image is, accord•

ing to the general context, properly the reality which it is said
to stand for. Now m Judgmg we do not use reality or realities)
though we Judge about them, m any mtelhgible sense of the
word ' use ' ; aud i,o \\ c cannot use the former sense of the
meanmg of mental 1m,tgc, because 1t 1s reJ.hty. The statement
then, "hether true or not, i,ecmi, only mtelhg1ble 1£ that other
::-eni,e of mc,111111g 1s mteudcd-' the meJ.nmgnesi, of the mental
image ', \\ luch, J.'> hJ.l> been i,,uc1, 1s prubJ.bly wufused some•
tune~ "1th the tlung meant. As "e do not ' usL ' the ment,11
uuagc, accordmg to our author, and \\C do not m,c the mcanmg
of 1t, as whJ.t 1s meant, 1t seems only to rcmam that \\-C should
ui,e the mcamngness of the mental 1magc-1 e we use the fact
that 1t has ct me,uung, and tlus 1s, perhaps, nearest to the
,u1ter's mtcnt1on, though he hai, not rc.ihzcd 1t with any cleJ.r•
nest.. But the meamngncss of the idea (=mental image= No i
idea) 1i, prec1i,dy as such of the idea, and c,urnot be abstracted
from 1t Thus, 1f we use the mc.inmgnes-, of the 1111Jge-1dc,1
(tJ.k111g tlus to ut· No I idea, or the 'p::.yeholog1cal 1de,t '), we
certainly u::.c the m1ctgc-1dca 1ti,clf 1 e. \\e must ui,e 'the
mcnt.al f.tct ' "luch \\ c arc said not to use
Further, 1t 1i, :.,ml th.it the meanmg \\ Inch \\ e ui,c Jl> the
' log11..c1l 1dc,1 ', ,md tlu::. agam ' 15 "h,1t we pred1eJ.tc, and we
predicate 1t of rcJ.hty '. Nov., 111 the d1fticulty lc1st Luns1dered,
11l order to make ::.cnsc, we h.id to equate the log1cal 1deJ. to the
mcamngncss of the mcnt,11-nnagc idea If we keep tlus s1gmfira•
t10n, the statement about the logical idea and the prcd1cat1on
of 1t becomes ' Ill Judgement "c predicate the mcanmgness of
the 1mage-1dca, and v. c predicate 1t of reality ' But obviously
the mcJ.nmgncs!> of the mcntal-nnage idea (1f 1t has any) can
only be prcd1c.ited of the mcntctl image and not of reahty.
Agam, as to tlus prcd1c,tt10n of the ' logical idea ', 1 1f 1t 1s the
meaning as= the t hmg meant, as 1t obviously should be, then
that is reahty, and a!> "c ha\e been cannot have the name idea
at all, 1£ idea 1s d1stmgu1shed from reality But now the context
1mphes 1t must be so d1stmgu1shcd, because a reference 1s made
1
It 1S a sign of the confw.10n of the doctrme that tills ' idea ' :.hould be
c.illed ' logical '. \Vhat could be the me,mmg of ' logical ' 1n such a reference ?
(Fiction of a logauan ~) And why 1:, the other idea called ' psychological ' ?
 Attempts to Define 'Judgement' :285
back to the previous statement-' without ideas no judgement',
where the ideas in question are distinguished expressly from
reality.
Finally, we come to the definition of judgement which ends
the discusc;1on. 'Judgement is the act which refers an ideal
rontent to a reality beyond the art'; ' reference' 1,; a vague
term, and we mu,;t ask what kmd of reference rs intended. It
simply means that, m A 1s B, Bness is r<.'"ferrcd to A, and, 1£ we
ask how rt 1s referred (for the judgement ,., more than that),
referring can't m<.'"an actually g1vmg B to A Thus the only
reply can be that we Judge that the reality A has the reality
Bness. Thus ' referring' meanc; Judgmg and once more the act
of Judgement is defined by itself. What is referred-viz the
ideal content, falls under our previous criticism If a content
i~ ideal, as somehow d1stmgmshed from reality, 1l 1,:; merely
absurd to say that \\C predicate this idea of ours of reality---
, c. attribute it to reality The truth 1s that tlus ideal content
1s not idea No. 1 • 1 e not ' the psychological 1clca ', or mental
image If, then, 1t rs to be idea No 11-the 'logical' idea-,
that we have seen 1s only ideal at all 1£ 1t i:;tands for the• mcanmg•
nei:;c;' of the mental 1mag<.'", and then 1t ccrtamly cannot be
predicated of reality. On the other hand, 1f rt has the former
significance of meamng (=what the mental image meam,) it ic;
indeed true that what 1s supposed to be meant by the mental
image, but really 1s not, 1s predicated of reality But that 1s
only berause 1t 1,; not ideal-1 e not an ' ideal coritent '
at all.
§ 123. This elaborate theory has a fairly c;rmple ongm. It
is unconsciously grounded on the same principle as the old-
fashioned copymg idea theory-viz that it io; our ideas which
are true or false, and Judgement rs true or false because 1t
somehow involves them The ideas themselves are true or fals<',
according as they agree ,,1th reality or not, and 'agree' means
to be hke reality Our author's form of tlus principle we have
quoted already; there 1s no Judgement without ideas, because
Judgement 1s true or false, and 'truth and falsehood depend upon
the relation of our ideas to rcahty '. Here the vague word
' relation ' is substituted for ' agreement ' or ' likeness '. In
developing this the writer simply modifiec; or alters certain
286 STATEMENT, THINKING, AND APPREHENSION
features of the old copying theory, and always for the worse.
He is in precisely the same difficulty as the ordinary copying
theory, and what he does is to substitute for the comparatively
clear statements in 1t confused and self-contradictory statements
which depend 011 verbal confusions. The primitive theory is
what 1s fairly represented m Aristotle 11 It 1s a putting together
of ideas which arc like realities, and the Judgement is true when
their puttmg together 1s al,;o like the combmation of the corre•
sponding realit1ec; The correspondence then 1s likeness. The
first change that the present theory makes 1s to alter the relation
of idea to reality from ' copymg ' or ' likeness ' to ' meamng '.
The mental image ( =fact m my head =psychic fact =psycho-
logical idea) 1s said to ' stand for ' the existence or, more
defimtcly, to ' mean ' ex1stcncl' This 1s quite an 1mposs1ble
use o{ ' mean ' and the wnter falls mto 1rretncvable confus10n
in his attempt to e)..plam what he means. Thus for the com-
paratively clear idea of hkene,;<1 or copymg I'> substituted some-
tlung unmtclhg1blc. Agam, m the pnm1t1vc copymg theory the
Judgement 1s simply the posc;ec;<,1011 of a synthc<11c; of ideas which
' correspond.:;' to reality m the way of bemg like a c;ynthes1s
m reality The d1ffilUlty 1s that such possession 1s of no use
unler;s we know that the likeness exists, winch must be by a new
Judgement. But 111 the new theory before uc; the Judgement
does not appear ar; the mere possession or ronsc1ousnec;s of the
ideas which arc said to mean the reality, or m<lee<l a<; the posses-
sion of idea'> at all A change 1s made, no doubt to av01d the
above difficulty of the copymg theory, and the change 1s of
a twofold character. First, the Judgement 1c; not placed m the
possession of idc,ts but m the use of them. Secondly, the
mcamng of the mental nnagc- or idea 1s 1t,;cli called an idea,
an<l the 1d(•a wlurh I'> ' uc;c<l ' 1s the l.1tter and not the former.
The reason we shall presently c;ec.
As to the ' use ', we have seen --
I. That it 1s a contrad1ct1011 to say that the meanmg of
the image can be used ,Hthout usmg the image.
2. The use itself turns out to be referrmg the meaning of the
mental 1mage--called also the ideal content and the
logical idea-to reality . and tlus referring, we have
c• D, 1,11 16A I.:?, 19a 33, Jfet 8, 1051b 3 er mfra, p 297]
 Aue1npts to Defint 'Judgement• 287
seen again, is nothing but judging the ideal content
true of reality, and so judging is defined by itself.
This set of changes therefore substitutes confusion for com•
parative clearness of statement, and adds the fallacy of the
circulus in definiendo.
Finally, how is this new theory situated as regards the diffi-
culty which seems so fatal to the copying theory ? It is here
we shall discover the real reason for introducing the confusion
about the meaning of an idea, and for eallmg this meaning
another idea. The difficulty for the copymg tht>ory is that the
possession of a complex of ideas hke the complex m reality does
not help us unless we know the hkencss , and for that we must
know the rcahty and thus we should not need the idea. Or
else, 1f we can only have the idea, we could never know the
reality and never know the • correspondence ' of the ideas to
reality The difficulty applies mall essentials to the new theory .
If an idea ' means ' reality or existence . 1f the • fact in my
head ' stands for existence, that 1s of no use unlcso; "e know
it, and that we cannot do without knowmg the ex1!>tence, which
makes the meanmg-1dea superfluom, Or else, 1f "c have only
access to ideas, we could never know that the mcntal-1mage idea
had a • meanmg ', 1 c W<-' could not have the ..,ccond kmd of
idea-the meamng-1dea- at all. The difficulty 1c; unconsc1ously
concealed and evadl'd by confusedly blending two different senses
of the word ' mcanmg' 111 one Tlus confusion we have already
analysed The idea wluch 1s the meaning of the idea called
'mental image' (' psychic fact', &c) 1s treakd m, -
(a) That which 1s meant =existence, and ac, -
(b) The meanmgness or representative charartcr of th<' mental
1mage-1dea
As (b) this idea may be 'ideal content'. Ac; (a) 11 1o; of
course the existence meant. And thus the gap brt\\-ecn idea
and reality 1s bridged by th1c; mconc;1c;tcnt cloublt• nwanmg of
the 'logical idea' or the 'meanmg '. In conformity with
the vague statements that there 1c, no Judgement "1thout ideas
and that Judgement uses 1<leas, the predicate of the Judgement
1s made an idea or ideal content, becaui;c ' meaning ' 1s taken
in the second sense. Judgement then becomes the referring
288 STATEMENT, THINKING, AND APPREHENSION
( ""'judging) of an ideal content (which is the predicate) to some-
thing. Now this is saved from being a mere subjective act
within the mere subjective material of ideas, with no guarantee
that it has anything to do with reality, by taking the idea
predicated ( =the meaning) in the former sense of meaning-
viz. what 1s meant, 1 c. reality. And so now in the formula
• Judging i11 the referring of an ideal c-ontent to something ',
we can c:;ubst1tutc the word 'reality' for 'something' since
the ideal content no\\= reality And so we finally arnve at this
portent of definition-' Judgement 1s the act which refers an
ideal content {recognized as such) to a reality beyond the act'.
There is no such complication of confusions in the old copying
theory.
§ 124. The confusions which lurk in this modern theory will
be plainer 1£ we now examine more closely Mr. Bradley's use of
the words ' sign ', ' symbol ' and ' meaning '.
He disrusses the meaning of symbol, wluch (as the remark
'for logical purposes ideas are nothing but symbols' shows) ic:;
of the greatest importance to him He supposes himself to be
giving a mere commonplace, which everybody might know, but
instead of that he has given a curiously confused and false
representation of the meaning of symbol and symbolism, sign
and significance.
He omits the most ess<.>ntial feature of 1t, viz. that it is an
arbitrary convention.
He speaks, without any safeguarding expression, as if a fact
could in itself mean something else. No fact can do that. It
is entirely due to our C"onvention that a fact has a meaning
other than itself. Signs do not mean anything, it is we who
mean, and we mean something by the sign It 1s obviously our
convention (cg) that x m alg<.>bra stands for an unknown whose
value 1s to be sought.
What, however, he means by a symbol (he would call x, ,.,, and
:: signs, or rather the kmd of signs which arc not symbols, as
will presently appear) would rather be illustrated by such a fact
as that an evergreen tree may suggest constancy to me (Oh
Tannenbaum, oh Tannenbaum, w1e treu sind deine Blatte I) and
may become a ~ymbol to me through thi,;. The reason is that
 Attempts to Define 'Judgement' 289
the continuation of the greenness in contrast to other trees, in
spite of changing seasons-a thing which is pleasant to us, has,
in its permanence and resistance to external change, a common
element with a steadfast character, which also 1s a pleasant thing
to us. This it certainly has of itself, and not by my convention
merely. Further, I may say that one reminds me of the other,
and this again, though subjective, 1s not due to my convention.
But that does not make the one a symbol of the other, does not
indeed decide (e.g.) whether the character 1s to be a symbol of
the tree or the tree of the character. Most certainly one of them
does not ' mean ' the other If, however, I srnd a card with
a fir-tree painted on it to my friend as a token of my constancy,
that is a symbol to him and means constancy, 1f we have agreed
upon it. Suppose there is no such convention bct,veen us, still
I must ha,·e arbitrarily given the tree its meamng In itself it
can mean nothing The picture, however, might be said to mean
the tree. If I intend my friend to interpret 1t, he must at least
have something to suggest to lum that 1t is a symbol, so that
he may think I have decided 1t shalI be a symbol And then his
course would undoubtedly be to consider what qualities the tree
has, and which of them have anything 111 common with what he,
from lus knowledge of men, may conJccturc I wish to symboltze.
There is therefore here no exception to the arbitrary and con-
,·cntional character of the symbol
In the case (e g ) of an algebraical symbol the meanmg 1s no
part of the content, and doubtless Mr. Bradley means that what
he calls a symbol is the particular kmd of sign m which the
meaning happens to be part of the content. Even when the
meaning is related to a part of the content, as in the case of
the fir-tree, or the fox taken as an emblem of cunning, 1t cannot
possibly be defined as he has defined 1t, for the process so
described \\ ould only be the act of abstraction ; in the case of
the fir-tree, c g , 1t would simply be to thmk of evergreenness
or unchangeableness generally apart from the particular tree.
The abstraction \\ ould m no proper sense of the word become
a ' meaning ' of the tree.
The confus1on is further illustrated, as follows : • In con•
trast with a symbol a sign may be arbitrary.' This implies
that a symbol is not arbitrary, which is of course impossible.
2 773 I u
290 STATEMENT, THINKING, AND APPREHENSION

Apparently he makes sign wider than symbol, and the division

1s like this

arbitrary sign sign not arbitrary ( =symbol)

Then presently he seems to imply this division :
sign
(a) with natural meaning (b) not with natural meaning.
'Symbol' as bcmg a sign must come in one of these species,
but t'ie author says ' a natural sign need not be a symbol '.
But symbol cannot be m class b because b must equal arbitrary
sign Therefore ' symbol ' must be m a This gives the
following
sign
(a) natural sign (b) conventional sign
_.___ = sign not with a natural
not-symbol symbol
meanmg.
Further symbol, though a natural sign, 1s called a ' secondary '
sign, because it does not stand directly for its meaning. This
seems to imply that the class of natural signs 1s divided into
• symbols ' which stand md1rectly for their meaning, also called
' secondary ' signs, and those \\ hich, standing directly for their
meanmg, \\ ould presumably be called ' primary ' signs. The
formrr, though he docs not say so, are probably the mental images
whose meaning as the 'logical idea' 1s used m judgement, at
least they probably belong to this class, whether there are any
other kmds of members in it or not
V'-l'e may suppose that while fox • md1rectly' means cunning,
the mental image of a fox ' means ' foxness as that of a horse
' means ' horseness But no explanation whatever is given of
this, and the theory prevents a kmd of explanation which would
have been possible, 1£ horseness had not been called a part of
the content of the horse-image Horseness really in the true
sense 1s not a part of the particular horse, but everything in
the particular horse 1s included m horseness. Thus it might be
said that while foxness includes the whole bemg of a particular
fox, cunnmg includes only a part of 1ts bemg. This would have
given some chance of disbngu1shmg primary and secondary
 .Atumpts to Define 'Judgement ' 291

signa. But this way is not open since both cunning and foxness
are considered as part of the content of fox. And of course it
is absurd to suppose that the image of the particular fox either
• means ' or ' symbolizes ' or is ' a sign of ' either cunning or
foxness.
The final classification appears to be :
sign
(a) natural sign (b) conventional or arbitrary
sign
(a,) secondary
-----~ ----
-=sign which has not a natural
meaning.
standing viz. symbol (no indication whether to be
directly standing subdivided into primary
for its indirectly and secondary)
meaning, for its
meaning,
e.g. fox e.g. fox
symbolizing symbobzmg
foxness cunning
But what possible sense could be assigned to standmg ' m•
directly ' for 1ts meaning ~ ' Direct ' and ' mdirect ' are here as
impossible of application as straight and crooked, and the use of
such terms betrays an extraordinary confusion
The only chance of getting any meaning into such an expression
seems to be to make the sign, which 1s nothing but a sign, stand
' directly ' for its meaning For instance + stands for division
and is nothing but a sign of div1S1on, while fox is something in
itself, besides being taken as a symbol of cunning, and so by an
abuse of language might be said not to stand directly for cunning.
But then it would be the arbitrary sign alone which could be caUed
direct and all the natural ones would have to be called indirect.
This account of sign and symbol and meaning is in fact utter
confusion
f 125. It has been pointed out above that strictly things can•
not mean anything, that it is we who ' mean ', and we mean
something by some sign or symbol which we use for the purpo,e.
There are two such uses .
I. We use a symbol, e g a word, for communication of our
U2
292 STATEMENT, THINKING, AND APPREHENSION
thought to others. But of course no such use can be made of
our mental images, which can only be in our own consciousness
and not in that of others.
II. (i) We use symbols for the operations of our own thinking
as in algebra. :But clearly there is no use of mental images
analogous to this.
(ii) We usc language 111 our own med1tat1on and thinking and
certainly thmk by help of 1t The symbol 1s here merely con-
\"entional and so far there 1s no analogy between 1t and the
mental images :'.\or <lo they (the mental images) enter mto any
system of constructc·d rcJatwns like the symbols of algebra, or
\\ ords m their grammatical construct1on.
There is howe\ er a certain analogy In thmking by means
of ·words, we use "ords and we are not thinking of them but
of things, of tlungs or m1httes meant by these words. In
thmkmg about realities, md1v1duals not present to us, we use
mental images and we are not thmkmg about the images but
1:1bout those reabtles But there 1s this great difference, that we
do not use the images as signs or symbols of the realities; on
the c-ontrary they are our imagination of "hat the reality really
looks hke Wt> might rightly say 'that 1s how we suppose the
reahty looks ' and tlas 1s language wluch \\ c would never apply
to a symbol.
Besides there 1s an important use made of mental images
when we thmk of um versa ls. Agam 1t may be said "e use the
mental images Mr liradley mdced 1s stnctly hardly entitled
even to the help, i,uch as 1t 1s, of tlw, fact, for lie says 1t 1s the
mental image as suc-h \\luch \\l' do not use. We are not thmkmg
of the 1ma~cs but t'{ the umversals. But here agam they arc
not used ai. S) mbols of the um\'crsals. On the contrary, we can
only thmk of the Ulll\ ersal as realized m a particular and the
mental image 1s of use only as the image of such a particular.
We imagme ourselves to be actually contemplatmg a particular
m which the universal 1s realized
To put 1t in the simplest language When we are thinking
of particulars absent from perception, "e are imagining the
object itself thought about (the obJect we are thinking about),
not imagining a symbol of 1t : in thinking of a universal we
 Attempts to Define 'J1«Jgement ' :z93
imagme a particular realization of the universal, we do not
imagine a symbol of it.
There is a grain of truth m what Mr Bradley says in so far
as the analogies above pointed out obtain, but they do not
conduct to symbolism and he has almost entirely mistaken the
real condition of the facts and the function of the mental image.
He speaks as 1£ the mental image were for purposes of thought
related only to a universal wluch is 1ts meaning. The mental
image of a horse has to do "1th Judgement only as havmg the
universal horseness for its' meaning'. We' use' only the mean-
mgs and the meaning is for him (as shown m his dlustrat1on of
the horse image) a part of the content, namely 1s a universal.
We think, he says, or rather assert the fact the image stands
for, not the image In that case ,, e should make assertions only
about universals and our mental images would relate only to
these. But ob'\iously our mental images constantly relate to
md1v1duals, and c1.rc not there merely to help us thmk umversals,
but constantly to help us tlunk particulars He says, indeed,
"e always predicate the ' logical idea ', which strictly implies
that the predicate 1s always universal. Tlus 1s not true, the
predicate (the true Iog1cal predicate) may be particular. There
1s in this connexion an important difficulty, which 1s not pro•
v1ded for m his imperfect analysis. It does not however affect
the present cr1tJC1sm. We thmk of the 1mage as hke the object
and we say that the obJect looks hkc what we are imagining.
But an image 1s not like an obJecl as an obJeCt 1s hkc an object.
One horse 1s hke another m the real umversal ' horseness ' both
are mamfestat1ons or reah,mt1ons of ' horscness ' and that 1s why
they are ahke But an 1magmary horse 1s not a realization or
mamfestatlon of real horseness, it 1s simply (and the language
1s accurate and can hardly be improved) the 1magmat1on of
such a reahzat1on Thus the 11nagmary horse 1s uot like the real
horse m the real umvcrsal horsencss: its bkeness to the real
horse does not consist m its bemg, hke that, a reah.tat1on of
horseness.
Now when the author treats ' horseness ', which is what we
assert and so is the real (what the mental image of the horse
stands for), as a part of the content of the mental imaie of
294 STATEMENT, THINKING, AND APPREBiNSION
a horse, he baa implicitly made horseness equivalent to real
horseness, present in or realized in the mental horse-image.
Thus he can in no way distinguish the imagination from the
experience of a horse. The truth perhaps is that he has not
reflected enough on the matter to be fully conscious of what he
is doing. It 1s perhaps significant that, in his account of the
Association of ideas, he commits himself to an explanation which
confounds experience with imagination.
 XIV•
APPREHENSION, CONCEPTION, AND STATEMENT
§ 126. THE subjects to be considered, according to the ordina,ry
phraseology, m this chapter would be described as the relation
of Conception and Judgement; more particularly (1) the differ-
ence between Conception and Judgement, and the questions
whether (i1) Conception precedes Judgement and (iii) truth and
falsehood belong to Judgements only and not to conceptions.
But the very form in which these problems are raised implies
the acceptance of certam distmctions without critical examina·
tion, which must first be disputed. The questions themselves
cannot be clearly proposed untll the presuppositions of their
ordinary formulation have been weighed.
In the judgement of knowledge and the act of knowledge, we
do not combme our apprehensions but apprehend a combina•
tion. That 1s, while we apprehend (s) 1A and Bness (that is, have
our conceptions of them), we also apprehend their relation and
unity. Now obviously the conceptions themselves are not the
given judgement nor are the apprehensions of them the appre•
hension which constitutes the Judgement. The question then
naturally arises as to the difference, 1£ any, between concr.ption
and Judgement. In our account "e have used the term appre•
hension sometimes for one and sometimes for the other. Are
then the apprehensions of (s)iA and of B or Bness themselves
judgements ? The question 1s of importance and applies to judge•
ment m its strict sense as well as m the erroneous sense which
we have criticized Let us first consider how 1t has been treated
m modern logic and in some early writers, ancient and modern.
In the modern substitution of terms of thought for terms of
language, judgement tends to be represented as a synthesis of
conceptions. The verbal form 1s considered to be the expression
of a subjective act of thought called Judgement, and the single
[• II 126-141 (except 138-g) were revised m 1914-15 The tenmnology,
however, especially the vexed word' Judgement', was not alt~red consistently.]
2g6 STATEMENT, THINKING, AND APPREHENSION

words of the sentence, or sometimes complex groups of them,

are taken to express the several conceptions of which the judge•
ment 1s said to be a synthesis, ,, 1th no great clearness perhaps
as to whether these words, or groups of words, mean or denote
the conceptions or, rather, correspond to them m some way
without meaning them
This representation of Judgement, as having conceptions for
its elementc; and as bemg the putting together of these elements,
favourc; the tendency to look upon conceptions as something
preceding Judgement, as ,l material out of which the Judgement
1s formed. Such a tendency, to mention a smgle instance, 1s
found in the logic of Lotze • If this were so, conception would
be the umt of thought a11('1 pnor to Judgement The alternative
\\Ould seem to be that the umt of thought 1s the Judgement (on
the g1,·en hypothcs1i, as to the meaning of 'Judgement'), con·
cept1ons havmg no c:x1stencc except as clements m a Judgement.
This then 1c; one of the problem., which ansec; on tlus modern
view of 1udgcment, exprcs1>cd 111 the tcrmmology proper to that
v1e\\ A Sl'Con<l p1 ohlem "lm.h hm, ,m..,cn 1i, "hether truth and
falsehood belong to <.011ccpt1ons, or only to Judgement, conccp·
tlon being neither true nor false
In the pl11losophy of Locke, "h.1t he calls ideas correspond
fairly to ronreptmni, m, abo, c understood They appcdr also
as elements of the st.,temcnt, 111 so far as he defmes kno\\ledgc
as the perception of the agreement b or disagreement of t\\o
1deac; Locke's ans" er to the t\\ o problems "h1ch we have
named 11> cle,,r In the first place 1t 1s cY1dent that with him
thought bcgmc; ,, 1th ideas, that these arc the true u111ts of thought,
,md that propos1t1ons am,e oul uf ,l I omparison of them
Secondly. truth ,md falsehood r belong only to propos1t1ons and
not to c;mglc 1dlw, taken by thcmscl\'es But here he 1s mcon·
s1stent ,,1th lumsclf, that 1~ with \\hat he says about real and
fantast1c.:1.l d 1dc,1s and about true and false e ideas, though he
endea\ ours to ckar lumsclf of the charge at the begmnmg of
Bk II, ch. xxxu, ' Of true and false ideas '. The fact 1s that
he 1s necessarily driven mto contradiction by his theory of the
origm of ideas, that is by his theory of perception. It 1s vain
[ 1 e g Logic, 1 1, § 8, b LoLkc, Essay, I\' 1, H 1-.z
c 1b., IV. 5, § 2 d 1b., II JO, • 1b., 11. 32.J
 A,pP,ehension, Conception, Statement 297
1n his philosophy to define knowledge as the perception of agree•
ment or disagreement be~·een our ideas, unless the ideas some•
how represent reahty. He becomes aware of this and admits
that the condition may not always be satisfied. Hence he has
to distinguish between ideas which adequately represent their
originals and those which do not. The former are called real
or true or adequate, and the latter fantastical or ch1mer1cal, or
inadequate or false
One of the earhest form!> of such theories and a kind of text
for them 1s found m Aristotle lumself. According to the De
/1iterpretatio11e 1 the "ritten word 1s symbol of the spoken word
and the spoken \\Ord symbol of states of the soul. Next a "ctis-
tinct1on 1s made between thought which 1s neither true nor false
and thought which 1s The latter 1s statement, to which no
special name 1s given m tlus passage. It involves synthesis or
d1v1s1on, the first meanmg affirmation, m which certain elements
arc put together, and the second negation, m wluch certain
elements arc d1v1ded or d1stmgu1shcd from one another. The
clements m question are called thoughts which arc without
synthe!>is and d1v1s1on, and are neither true nor false. There is
a parallel d1stmct1on in the corresponding verbal expression
Aristotle says, that is, that the statement has elements m 1t
\\ h1ch correspond to the thoughts wluch arc without synthesis
and d1vis1on these elements arc the nouns and verbs which
taken by themselves arc neither true nor false. In a .,econd
passage, in the De Amma, 2 he says of the form of thought to
which both truth and falsehood can attach that 1t 1s a synthesis
of thoughts, from \\h1ch we might mclme to mfer that 'thought'
1s the term corresponding to conception In the first passage
before us, however, the thmking "'luch corresponds to the state-
ment 1s itself designated as a kmd of thought, namely that
which involves truth or falsehood, "lulc the conceptions are
thoughts which arc without truth or falsehood and without
synthesis and division The smgle "ords taken by themselves,
l Ion ,.•• ow Ta b Tp """"U TOW iv Tjj iJ,uxi ,ra9"P4Ta>II IJVp/3oM ,cal TG "11"1¥1"110.
Tcii11111 -rj '/>Cml, ••• w•pl -ycip 11t'w8ot1iv 1<clll 3111ipt11i11 l1JT1 TO if,tii3or l<al Tu oAf/lir. nl
I'll' ow wd,-aT'fJ. dnl 1<tu Tei !'11107'1 1011<• T9/ ci11t11 1111116t1Jflilf ,ml IJ1G1ptt1fM ro,ipin1
iii• 3-4 and 12-14. Cf pp 286, 314
z '" of, Si ,ral Tcl \£tiiaclr ,ral Tel 6.11..,a,., """'""'' TIS ~a,, IIOf//AGTOm liHl••p ,., /ivroiv
43ua z7.
2g8 STATEMENT, THINKING, AND APPREHENSION
though neither true nor false, have a meaning and are called
speech in distinction from assertion and denial. The sentence
1s called ' word ', and the enunciat1ve form of it, that is to say
the statement, 1s called ' word declaratory '. 1 Thus there does
not seem to be any name m Aristotle for an activity of thought
which might be supposed to correspond to the statement (like
the term Judgement erroneously used m modern logic), at any
rate there 1s none m common use m his writmgs. His usual
word for a propos1t1on 1s premiss, 2 which designates 1t from the
point of view of its position m an argument and not from its
meaning
Itmust be observed that Aristotle here makes the same
mistake as the logicians who use Judgement and conception in
the sense we have been discussing. He takes the sentence to
represent a sub1ect1ve state (states of the soul 8), namely the
thought which 1s either true or false, JUSt as the moderns make
1t represent Judgement S1m1larly he makes the verbal elements
of the sentence taken singly represent the thought without
synthesis and d1v1S1on, or without speaking truth or untruth,'
just as m the modern theory the words of the sentence are made
to represent conceptions The nature of this mistake has been
already pointed out 5 The sentence which 1s a statement clearly
describes the nature of a thing or obJect, and the verbal elements
of the statement do not represent elements m our thought but
elements m the thing or obJect
From the Anstotehan description and 1ts modern counterpart
1t might be supposed that the parallelism between the Judgement,
with conceptions as its elements, and its verbal statement, with
smglc words or phrases as its elements, 1s such that the words
which are <'lcments of the statement denote the conceptions
which arc clements of the Judgement But thts 1s not correct.
It very often happens that there 1s not m the verbal statement
any word to denote a conception or conceptions supposed to
belong to the so-called Judgement For instance the statement,
lead 1s heavy, supposed to represent a Judgement which 1s
a synthesis of the conceptions of lead and heaviness, contains
1 >..S,,os c1 rOl/>flPT1Kclr De Int 17• 2 and 8
1 wp6Ta.rm, cf supra, p 83 • "°"711«T11 .-;;s 1},l'xijr
• •6',pa cwcu a111elaceis, a,,,11 arwfliac111s iral Bu11piou1r, a,,111 .-oii dA?ftum• • \f,c~tafa,.
I I us.
 A-/'P,ehenlion, Conceptton, S'4Ufflent agg
indeed a word which denotes one of the supposed conceptions
(lead) but has no word which denotes the other conception
(heaviness), although heavy certainly presupposes heaviness. 1 It
would however be sufficiently accurate for the purposes of the
erroneous theories before us to say that the verbal elements of
the statement either denote or imply the conceptions contained
in the judgement.
§ 127.• What 1s presupposed then in the two questions, whether
conception is prior to judgement and the true unit of thought,
and whether truth and falsehood belong to Judgement only and
not to conception, 1s that the statement always expresses a Judge-
ment and that the judgement itself is some sort of putting
together or ' synthesis' of conceptions These presuppositions,
as we have endeavoured to demonstrate, are erroneous.
In the first place it has been contended not only that the use
of the word judgement is quite incorrect and that it does not
always correspond to the mental attitude which issues in the
statement, but also that there 1s no one general form of thought
at all to correspond to the one general form of language called
statement. Secondly, the verbal statement, so far from implying
Judgement, does not, in general, signify anything subjective at
all. Thirdly, neither Judgement nor any other of the forms of
thinking or apprehending which lead to statement can be repre-
sented as a combmat1on of conceptions
It does not however follow that there are no such problems,
even 1f the way in which they are put involves fundamental
mistakes It may happen that these problems do correspond
to some real demands of our reason and that they only want
restating ; in their essentials they may prove not to depend
upon the mistaken theories which their usual formulation implies.
We shall endeavour to see whether this 1s so by returning to
a consideration of the meaning of conception as contrasted with
1
The relation of such an ad1ective as heavy to the noun heavmeu iii
COllSldered m § 81

[• A d1gress1on on the worth to the philosopher of normal hnguJStic usage

has been transferred from here to my Postscnpt The idea was perhapa
onginaUy suggested by Curbus, who m the preface to hls G,,ek G,11,,.,,...
says, • Rightly examined our sense of language 1s here, aa often m my belief,
a far more certain guide than the halr-sphttmg combmataons of philosophy ' l
300 STATEMENT, THINKING, AND APPREHENSION
judgement and with the other forms of thinking irreducible to
judgement, which isr,ue in statements.
The word conception is a word in ordinary use and not
a technical expression, at least not origmally. The word concept,
on the other hand, is technical. What is the proper meaning of
conception as it occurs in ordinary use ? There seem to be two
main senses. In the first place we find the expression conception
of something, as for instance my conception of a solid body or
of a fluid. In psychology and m philosophical reflection there
1s a tendency to regard the concept10n of an obJect X as some•
thtng entirely mental and distmct from X This, although so
far not altogether incorrect, 1s commonly associated with the
quite erroneous tendency to treat this mental something hke
a mental image. If in ordinary hfe we asked a man what his
conception of a fluid was, he \\Ould tell us ,,hat he thought was
the nature of a fluid, that 1s to say he would tell us what he
thought a fluid "as. Sm11larly my ronccpt1on of X 111 general
means "hclt I conceive X to bc-, though \\C shall 5C'e that this
phrase requires some cluc1dal1011 Conce1nng here may mean
thmkmg and, m that case, my rnnrept1011 of X is "hat I think
X to be, "hcther thl' thmkmg 1s apprehension or only belief or
opm1on Thus my conception of a body may be that 1t 1s
a congt:nes of atoms If then "e merely take the phrase ' what
,,e thmk body to be' as a descnpt1011 of the conception of
a body, 1t might be said that a congeries of atoms 1s what we
think body to be and therefore 1s the roncept1on of body. But
this 1s dearly ,uong, for thC' phrase a congeries of atoms does
not mean anythmg subJect1ve or mental but something objective.
The concept 1011 ot body here 1!> not a congeries of atoms but
' that 1t 1s a congeries of atomc;' I may say 111d1fferently that
my coneept1on of X ll> th.it 1t 1s Y, or,, hat I tl11nk of X ts that 1t
1s Y For c1ccuracy therefore ,, c reqmre to expand the phrase
what "c think of X, or "hat "c thmk X to be, mto ' what we
thmk X to be, taken c1s thought by us of X '. For mstance, in the
example given, my conception of body 1s not a congeries of atoms
but 'a congeries of atoms thought of as bemg what X 1s '. It
\\ ill be observed here that "hat 1s said to be my conception of
X cannot be represented verbc1lly by a noun, or a noun-phrase,
but ah, ays requires a ,·crb or 1mphcs .t sentence m which a verb
 Apprehensioti, Conception, Statement 301

1s used ; implies, that is, from the point of view of language

a statement, and, from the point of view of thought, knowledge
or opinion. For instance, if I say my conception of X is that
it is Y, this really contains the statement X is Y as made by
me. Again, if I say ' my conception of X is my thinking {the
change of phrase will be explained presently) of X as Y ', my
statement that X is Y 1s implied. Thus my conception of X
either contains or 1mphes m itself knowledge or opimon of mine
(what 1s erroneously called Judgement) corresponding to the fact
that its verbal expression includes or unplies statement. If now
"" say 'my C'011C'ept1011 of Xis that 1t 1s Y ', what 1s it that has
to be understood before the word that? It would probably be
ans\\ ered that 1t 1s my opinion, my behef, &c. This would seem
to 1dent1fy conception with opinion, behef, &c , so that its verbal
expression would be statement or affirmation. Again, 1£ we ask
the questions what 1s my opmion, &c., about X and what is
my conception of X, the answer may be the same to both
questions. that 1t (1 e X) 1s Y. This agam serves to identify,
dpparently, conception with a form of thinking, such as opinion
01 kno\\ ledge, which corresponds lo statement.
lf tlus were so, 1t would seem as 1f conception were a redundant
word quite unnecessary m the language My conception of
a thmg would simply be my opmwn, Judgement, &c , about 1t
Would 1t then not be a kmd of perversity m language to use
the phrase' concept1011 of', about \\hich there JS a kind of vague•
ness, so that 1t tends to be regarded as a mental image, instead
of the stra1ghtforn ard and unambiguous terms knowledge,
op1111on, Judgement, which cannot possibly be confounded with
mental images ?
§ 128. We shall find that language 1s after all vindicated once
more m regard to its normal employment of a special word
conception, as d1stmct from opm1on, Judgement, &c. The
forming of an opm1on or Judgement 1s a defimte act ; for
instance, 1f we form the Judgement that X JS Y by proving it,
the act of Judgmg is the act of provmg The accurate form
therefore of the question before us seems to be, ' Is having
a conception the same thmg as the act of forming a Judgement
or an opinion?' , m other words, does having the conception of
X as being Y mean precisely that I am forming the opinion
302 STATEMENT, THINKING, AND APPREHENSION
that Xis Y, or judging that Xis Y? The answer seems clearly
No I and this answer would be the natural one in accordance
with the normal use of language. We should be inclined to say,
perhaps, that our conception of X as being Y was not the
judgement or opinion that X is Y, but rather the result of it.
But now what is the nature of this so-called result ~ One result
may be a change m our mental image of X in cases where
a mental image 1s possible and relevant Yet as we have seen
the conception 1s not properly a mental image at all. There is
however another kmd of result \Vhen ,,e have been through
the process of formmg the opm1on that X 1s Y, or of judging
the same thmg m a proof, our thought about X is changed ,
we no longer tlunk of 1t as merely X but also as being a Y.
After we have formed the Judgement (or the opimon)-then,
when we think again of X, we may thmk of it, that is treat it
in our thoughts, not only as X but as a Y This 1s clearly not
the or1gmal Judgement that X 1s Y, for v. e need not have that
judgement before us It is enough that we remember, in the
case of a proof for instance, that \\ e proved X to be Y, without
going through the proof aga111 , mdec<l we have sometimes for-
gotten the proof In accordance with this we may go on to
prove somethmg of X wluch follows from its bemg a Y. We
treat X then as a Y, and yet are certamly not Judgmg that X 1s
Y, because that would mean proving that X 1s Y, and we have
not the proof before us. This example shows that there is
a thinkmg of X as Y, which is not strictly speaking Judging
that X 1s Y, though 1t depends on this judgement. We must
here guard against supposing that this thmkmg is a memory of
the preceding act of Judgmg, that 1s of the proof that X is Y ,
for this would really mean that we were proving 1t over again.
Memory is necessary but not memory of the nature of the
proof ; it is the memory simply that there was a proof. The
same is of course true 1£ we did not prove that X is Y but had
an immediate apprehension of 1t m a past experience , we should
think of X as Y, but tlus would not be the formation of the
apprehension (often wrongly called Judgement) that X is Y, for
that could only be the ongmal act of perception. Now this
thinking of X as a Y is that true meamng of conception for
which we have been searchmg, and its nature becomes clear in
 Af>lwelle,uion, Conception, St4'ement 303
such examples as those above taken. This enables us further
to understand how conception, in the sense before us (conception
of X as Y}, may appear as a kind of element m an act of judge-
ment or of opinion, for instance I may form the Judgement that
X 1s Z because it 1s Y. Here I thmk of X as Y, but do not
properly Judge X to be Y, the only Judgement bemg the judge·
ment that X is Z In accordance with this the only thought
which has the verbal form of statement 1s that X is Z. That
X 1s Y 1s implied, 1t may for instance be represented by some
,·erbal form hke xy, but 1t is not stated explicitly.
§ 129 We are now m a positmn to restate and answer the
ordmary questions about the conception, when conception means
conception of something as something, or conception of X as Y.
In the first place, while it 1s true that conception m the present
sense may appear as an element m a Judgement or opmion, or
generally m any of the activities of thought which issue m
a statement (an element m the manner Just explained) ; yet
1t always implies such an activity of thought (judgement,
opmion, &c ) Thus it cannot be an ongmal element of thmking
prior to Judgement or opmion. To avoid confusion we need to
observe that the Judgement, opm1on, &c, which the conception
of X as Y, or the thmking of X as Y, implies, has not necessarily
been explicitly stated For example, we may have perceived
X as bemg Y 10 a number of mstances without needmg to make
a statement about 1t, though our frame of mind each time is
an apprehension which might issue m statement. We may next
treat X as if 1t were m all cases Y, that is have implicitly formed
the induction 'all X 1s Y ', without explicitly reflecting on this.
The reality of the implicit mduct1on shows itst.lf by the fact
that we treat all X as 1f we thought it were Y Further, we
habitually thmk of X as Y, that is to say we have the ' con•
ception' of X as Y.
From this case we must distinguish another, which 1s very
similar to it but not to be confused with it, the association of
ideas. We may experience X and Y often together, but that
is not the experience of X as be10g Y but merely an experience
of their association. For instance, I may associate m this way
a certain kind of walking-stick with a certain man, but I neither
auppose that the man is the stick or the stick is the man, nor do
304 STATEMENT, THINKING, AND APPREHENSION
I treat the man as if he were a stick. Thus, though it may happen
that I never think of X without thinking of the Y associated
with it, I do not think of X as being a Y, and so have no con-
ception of X as a Y. So, for instance, if I either see or imagine
the stick, it may be that I cannot help thinking of its owner,
but I do not think of the stick as being the man, and have no
' conception ' of the stick as a man.
The next question 1s whether a conception in the present sensr
<'an be true or false The rrply here is quite clear. The con-
reption of X bemg Y 1s based upon the opinion or judgement
that X is Y. But our opm1011, or our belief that X 1s Y, may
bt' true or false, and at least 1t 1s ah\ays admitted that a
i,tatement and the thought corresponding to 1t must be one or
the other. But, 1f it is false that X is Y, then our conception
of X being a Y is also false, whereas it 1s a true conception, 1{
the statement that X JS Y JS true.
\Ve are thus led to a doctrine wluch 1s directly contradictory
to the ph1losoph1c view that conceptions as clements of Judge·
ment (Judgement 111 the erroneous sense of a mental activity
common to all statement) are neither true nor false; for con·
reptton m the present sense 1s one of those elements m an act
of thought corresponding to a statement. Here agam ordinary
language 1s JUSt1fied, for \\e constantly speak of having a true
or a false conception of something.
§ I 30. But now there 1s another meanmg of conception m
non-phtlosoph1cal thought, "h1eh yet 1s a usage belonging to
the stage "hen thought has become reflective, though not m
the technical sense philosophical This 1s illustrated m such
expressions as the conception of duty, of will, of mass, of evolu•
tion. In this case also the \\ord conception 1s followed by the
preposition ' of ', used however m a d1fferent sense. Thus the
conception of X would no\\ mean the conception X simply,
the preposition ' of ', according to a familiar idiom, expressing
equivalence, as when "e say the city of Oxford The conception
of duty would then mean the conception duty, not some con•
ception formed of duty as bemg so and so. The conception of
force again does not mean some conception I have formed of
what force really 1s, for this would be expressed in words different
from the word force. Another good example would be the
 A ;p,,1,ension, C0HC1,Ption, Staument 305
conception of chemical combination or again of a fluid. As we
shall presently see, this use of the term conception is peculiar
and restricted. In the most modern {I will not say the most
advanced) thinking the technical term concept 1s commonly
used for the sense of conception which we have been illustrating.
Concept is a term that we shall have to consider presently ; at
the moment we are only concerned with the fact of this use of
1t. These co12cepts do not apply to the whole matter of the
science or department of thmkmg to which they belong Thus
for instance 10 mathematical physics \\e should hear of the
concept of force, but not of the concept of electricity. In
chemistry we should hear of the concept of atomic combination,
but not of the concept of oxygen or hydrogen. In morals again
people will talk of the concept of duty, but not, commonly at
least, of the concept of temperance or of courage There is
a book entitled The Concepts of Science, which certainly does
not mtend to deal with the conceptions of everything with
which science has to deal. An instructive example of the limita-
tion m usage 1s that, whereas 1t would not be at all natural to
say the concept of a crystal, 1t would be qu1t1.: natural to speak
of the conception of a crystal ; mdeed we feel tlus latter designa·
t1on to be appropnatc Yet 1t 1s mterestmg to observe that the
conception of a crystal seems properly to belong to tins second
meaning of conception , we mean, that 1s, the conception crystal,
Just as we mean by the conception of force the conception force.
If this is so, 1t shows that concept docs not cover the whole of
tlus second main use of the word conception ,, hich we are now
discussmg. If we were to ask what the concepts of the sciences
meant, we should probably be told that they were the ultimate
notions, or ideas, or conceptions of a given science, conception
being taken in the second mam sense of the word By ultimate
conceptions are here understood the conceptions which on the
one hand are peculiar to a science, and on the other hand pre-
supposed by that science The science, as such, does not discover
them but assumes them, nor does it even, properly speakmg,
investigate them in the way of estabhshmg either their meaning
or their validity , that is, in general, the business of philosophy.
So also m that thinking about morals wluch 1s practical and not
philosophical, and so far corresponds to science though it is not
X
3o6 STATEMENT, THINKING, AND APPllBBENSION
science, we do not investigate the conception of will but IIIMlllle
that we can will ; we do not discuss the conception of duty but
ask what our duties in particular cases are.
I I 3 I. Let us return now to this general sense of conception
of which concept is but a particular species. What is meant by
calling force a conception ? The last word means, at least, some•
thing subjective. But now m mathematical science, for instance,
force does not mean a notion or idea or conception of ours : it
means nothing subjective, it means an objective fact. Thus, for
example, m dynamics the mathematician is not dissecting a con•
ception nor treating of a conception at all in his theorems about
forces ; he is thinking about the ob1ect1ve fact of force, whether
generally or m particular Similarly we are constantly hearing
of the conception of cause, that 1s (m the present acceptance
of conception), of the conception cause, but cause or causation
does not mean the conception, it means an ob1ective fact There
1s often associated with the terminology which we are discussing
the expression content of the conception, and 1t is its content
which 1s supposed to d1stmguish one conception from another
Observe that this content as content of the conception must be
something subjective, because the conception 1s subjective. If
we ask m any particular case what 1s the content of a given
conception, we shall get an answer which seems to designate
something obJectlve and not subjective at all. In some cases,
where there is a complex involved, we can get an answer verbally
different from the conception about which we asked the question.
For instance, the content of cause might be said to be the
necessitation of an event ; but this clearly aescribes a general
fact or what we may call a umversal. In the case where there
1s no such complexity m the conception, the answer can only
be a repetition of the word by which we have designated the
conception, or else a mere synonym Thus for the content of
the conception of mass we should only get mass over again ;
yet even so the word naturally denotes an objective general
fact It 1s very suspicious that this content, which should be
mental and subjective, can only be described by a word which
properly belongs to the objective. We must suspect here the
influence of the old copymg theory of knowledge. The content
is really somethmg vaguely and confusedly thought of as if it
 Af>1>,MMSion. ConufJnon. Sllltement 'Pl
were some mental entity, corresponding somehow to the objective
fact.
If we now ask what js the true account of the relation of
objective and subjective in the case, for instance, of the so-called
conception of cause (where, by hypothesis, the conception of
cause is the same as the conception cause) the correct answer
seems to be that, on the one hand, there 1s the fact of causation
which is an objective reality, and, on the other, our apprehension
of this fact, which is subjective. The apprehension may possibly
be accompanied by some imagining of particular instances in
which we think causation 1s operative, but such imaginations
or mental images are not our apprehension nor can cause be
imagined There 1s no room therefore left for a mental some-
thmg called the content of conception and this 1s altogether
a fiction.
As already indicated, the confusion which 1s 1mphed in such
expressions as the conception of cause, meaning the conception
cause, has lurking m it, especially where associated with the
fictitious word content, the old copying theory, though probably
with not full consciousness: but the mam m1sch1ef seems to be
caused by the popular modern conceptualism. The inability to
recognize that what are called abstract expressions (such as will,
or heaviness, or necessity) signify real objective facts has led
to a conceptuahsbc view of universals ; and this agam, perhaps
without full consciousness, has become engrained in thought and
expression. Hence a universal is habitually, one may even say
always, m ordinary reflective writing describe.I as a conception,
as if particulars (as m the well-known conceptualtstic theory)
were the only obJective reality, and universals only thoughts
of ours.
This second main sense then of conception seems merely
a piece of confusion.
If this is so, it is obvious that the two mam problems about
the priority of conceptions to Judgement, &c., and their truth
or falsehood, cannot arise according to this use of conception.
I 132. How then is ordmary language situated as regards the
confused and untenable use of conception 1ust elucidated ? Does
it ever commit itself to such expressions as the conception of
mass or involve itself m the d1fficult1es which attend them ?
X2
308 STATEMENT, THINKING, AND APPREHENSION
Does it endorse the false idea of content, and the conceptualiam
implied in treating force and kindred notions as conceptions 1
The reply seems to be that ordinary language lends no authority
to such errors. It speaks of force as force and not as the concept
or conception of force, of duty as duty and not as the conception
of duty, and, what specially deserves our attention, it never
treats these universals as subJectJve or merely mental, but
always as realities. It 1s only m reflective and philosophical
thought that those difficulties arise which dnve men into con•
ceptuahsm The significance of this pomt can hardly be over-
estimated ; it seems constantly, perhaps always, to be lost sight
of m the discussion of the familiar metaphysical difficulties
about the universal The same 1s true of science proper. When
science is about its proper business we hear nothing of con•
ceptions, and the theorems about force are about force as an
objective fact The only exception 1s to be found in the prefaces
which mathematicians sometune~ prefix to the actual mvest1ga-
tion. In their prefaces they often fall mto reflective thmkmg
and profess to be mtroducmg the learner to what they call the
conceptions of the particular mvest1gatJon This is merely the
ordinary mistake , but 1t disappears as soon as we come to
the proper business of the science
The question now arises whether there is any legitimate use
of the word conception m relation to such realities as force,
mass, will, &c, wluch have been themselves erroneously treated
a.s 1f identical \\1th conceptionc; It tmght seem to us before
reflect mg on the meamng of the "01 d that 'conception of' must
be of universal apphcat1on, so that we could speak of the con·
cept1on of anythmg "hatever But "hat should we say 1f any
one asked us ,, hat was our conception of blue, or of greater? We
should find 1t difficult, m fact 1mposs1ble, to answer and should
ask what the question meant It 1s not therefore quite evident
that the expression 'conception of' 1s of universal application.
Where we found the expression legitimate we had to do with
something with a d1stmct nature X, thought of also as being
a Y, so that X and Y were d1fterent and d1stmgu1shable aspects
of the same reality. Thus I may say that my conception of
body (X) is that it is a congeries of atoms (Y), where body
corresponds to the popular idea and congeries of atoms 1s some•
 APl>~ehension, Conception, Statement 309
thing which has to be learnt from science. On the other hand,
m that use of conception which we have found so confusing,
the conception of X was not of X as being something different,
say Y ; it was simply and solely X. So then the answer to
the question 'What is the conception of X ?' would seem to be
that 1t is X, and therefore nugatory. This is quite obvious in
the cases, hke blue and greater, where X 1s simple, that is, not
some umty of d1stingu1shablc elements (for observe that though
greater implies a distmct1on between at least two d1stmgu1shable
thmgs, it is not itself any unity of elements). In certam other
cases, ho" ever, those namely ,,Inch arc not simple, 1t may seem
at first as 1f an answer w.i.s possible that was not nugatory,
becausr we use words d1ff erent from the word X, or the expression
X But this 1s an 1llus1on, as an dlustrat1011 will show. The
d1stmctton is merely verbal and the answer only gives a synonym.
We might say that force 1s the cause of a body's state of rest
or mot1011, or give some other s1m1lar defimtton. Mathematicians
hke to say that force 1s that wJuch changes, or tends to change,
a body's cx1stmg state of rest or motion, but they '\\-Ould be
\ cry hard put to 1t to explain what they meant by tends to
ch.mge. Or agam one might say that cause 1s the neccss1tat1on
of an event. But these expressions arc only an explanation of
the meanmgs of the "ords force ,rnd cause Cause does not m
fact stand for anythmg chtterent from 11ecess1tat10n of ;ui event,
and sundarly for the am,\\ er given m the case of force This
sho,, s that m the ans\\ er "c arc merely g1vmg what 1s verbally
different, m the form of a synonym for the ., ord used m the
question, to denote the obJect mqu1red about Thus 1t 1c; really
quite as nugatory to ask \\ hat 1s the conrcpt1on of force or of
cause as to ask what 1s our conception of blue.
We must return to the question whether in these cases there
is any proper use of the word conception Conception 1c; al
least somethmg subJect1vc What 1s the subJect1vc element in
relation to cause, force, duty, &c ? It seems to be only the
c1pprehension of them, or (what 1s not relev.i.nt to our present
purpose) the 1magmat1on of particular cases to which they would·
apply. If we ms1sted on using the word conception for such
instances, we should have to make 1t mean our apprehension
of them. But this is by no means a natural employment of the
3ro STATE KENT, THINKING, AND APPllUtHENSION
term conception ; no one, either in ordinary life or in the
sciences, would normally call his apprehension of a thing his
conception of it. Perhaps it would be right to say that in
a given case, when people speak of the conception of a thing,
what is really in their thoughts, though they do not clearly
recognize 1t, is the apprehension of the thing. We are thus led
once more to the conclusion that there is no proper place for
the word conception m such cases as we have just been con•
s1dering.
It may be asked however whether we should always use the
term apprehension ; for, strictly speakmg, apprehension is of
reality, and, though in the illustrations given apprehension may
be rightly used, since force, cause, &c., are realities, it may be
objected that this is not always so and that we might properly
speak, say, of the conception of magic, magic being no reality
but an illusion to which the term conception, as subjective,
seems quite appropriate. This question will be considered
presently in connexion with a restatement of the whole problem.
As to the peculiar word concept, there is another use of it.
In that modern logic which substitutes judgement for proposi•
tion, it is customary to replace the old word term by the word
concept, and as in the older logic the proposition was considered
to be a combmation of terms so here the Judgement also is said
to be a combmation of concepts. In this usage concept seems
equivalent to that second meanmg of conception where the
conception of X ordinarily means the conception X.
§ 133. It has been contended that, whereas several different
forms of thinking {knowmg, formmg an opmion, and believing)
may issue in -one and the same form of statement, there really
is no common universal corresponding to the common word
' thinking ' of which they are species, and that the unity of the
various activities of the mmd hes in their relation to the activity
which is knowledge. We shall begin with the consideration of
knowledge because it 1s upon knowledge that the other activities
depend and through knowledge only that they can be explained.
If we started from the statement itself we should be led in
the same direction ; for the proper function of the verbal fonn
called statement is to describe something known to be true ;
that 1s to say 1t is the verbal expression which properly corre-
 Af,;relu,,,sion, Conception, SuiteMMd 3n
spondl to knowledge and to knowledge only. The meaning of
the statement is not the thought which gives rise to it, it does
not mean our thinking as such, nor does it describe anything
subjective whatever. It describes the reality which is appre-
hended as a matter of knowledge, the object thought about, and
it may be said to mean what we know of this object. That is
why the verbal forms in the statement are forms of the object
only. In opposition then to what seems a common tendency
it must be pronounced futile to try, as Kant actually does,• to
find forms of thought by any analysis of the verbal form of
statement. Our analysis of it can only conduct to forms of the
reality thought about, not to the forms of the think-
ing about it. Let us consider then the statement and the
thought, the act of knowing, which normally corresponds to it,
and let us inquire, with the help of the form of the statement,
what elements are distinguishable in the act of knowing and
whether they are prior to the whole act of thinking, the act of
knowledge which issues in the statement.
In the analysis of the grammatical forms of the sentence
which expresses a statement 1 it has been maintained that these
forms express m the first instance the relation of substance and
attribute and m the next place relations between substances:
also that, as thought advances, the statement 1s extended to
the more general relation of subject and attribute, where
subject 1s not necessarily substance. In such a sentence
the subject 1s represented as the umty of 1ts attributes. This
1s a description then of the real obJect as the umty of such
elements. These elements though distinguishable from one
another only exist m the unity of the subject. They are
elements m the subject of the sentence But, though this is
so, it is clear from what has been said ot them that they are
not elements which are prior to the reality here called the sub•
ject i for mstance, in the famihar example of matter and form
(in the literal sense), neither matter nor form can exist indepen•
dently, they exist only in their umty with one another.
The act of knowledge corresponding to the statement is the
1 I 76 et seq.
[& e.g. Prol. I 39, Krihlt d r. V., Trame Analytic, Metbod of the dis-
covery, &c., rat ed., p. 76 ]
312 STATEMENT, THINKING, AND APPREHENSION
apprehension of the nature of the object as expressed in the
statement and therefore contains the apprehensions of these
elements m the object, and these apprehensions are in this way
elements in the whole thought or whole apprehension which is
the act of knowledge corresponding to the whole statement. 1
But now as the elements m the object have no independent
nature but one "luch has bemg only m connexion with the
other elements, so also the apprehensions of them cannot exist
apart from one another, for mstancc we cannot apprehend
surface except as the boundary of the solid. It 1s true that, the
elements m the obJect being different and distinguishable, we
can concentrate our attention on one of the elements m what
1s ordinarily called the act of abc;tract10n ; nevertheless we can•
not possibly apprehend the separate nature of the given element
save as implying that from which it is inseparable, just as odd
cannot be understood apart from number. Thus then the appre·
hensions which appear as elements m the total apprehension of
the obJect cannot be prior to this total apprehension and can
only be had as apprehensions m the having of this total appre-
hension Hence these apprehensions, "hich are clements m the
total apprehension and correspond to the clements in the object
named in the statement or implied m It, cannot be prior in our
thinking to the total apprehension, for they can only be had in
having 1t, Just as the elements of the whole object can only
exist in the object The unit of thought, where thought 1s the
apprehension of a subJect as havmg attributes, is therefore always
the apprehensmn of a unity of a plurality of clements and there
is no such thing here as what is called simple apprehension.
This is the exact contradictory of such a doctrine as that of
Locke,"' which rcmams m essentials the popular view.
* *
§ 134. In the preceding discussion we have attempted to get
at the true nature of that d1stmct1on an imperfect understanding
of which has produced the erroneous distinction of conception
and judgement. The total apprehension 1s what corresponds to
the incorrect term Judgement, mcorrect m the case of knowledge,
1
Tlus however is not a sensation of ours or any subJectlve state. Cf.
infm, p 313 [N B feeling and sensation are not d1stiogullhed by the author.]
[• EutJy, II I,§ 3, et pa.aSim. At the end of§ 133 a lacuna 1s marbd.]
 Af>fwehension, Concep'tion, State,nem 313
because not every act of knowing is an act of judging. We do
not distinguish between judgement and conception but between
the apprehension which is complete in itself (what we have
called the total apprehension, the apprehension of a unified
plurality) and apprehensions which are incomplete and cannot
be had by themselves but only as elements in the aforesaid total
apprehension. In accordance with tins It should be noticed that
the part of the statement which corresponds to the incomplete
apprehension is a word and not itself a form of statement,
a point to be returned to. 1
To the result arrived at certain objections might be made.
It might be said that this account of knowledge, whether judge·
ment or immediate apprehension, is based upon statement,
whereas the question really 1s whether there are not apprehen-
sions which are entertained by us without the exercise of what
the objector would probably call judgement , complete apprc·
hensions which arc not expressed in statement. Besides, the
statement referred to relates to things or objects which present
this umty in plurality and It might fairly be asked "hether the
question would not be different m the case of something purely
subjective, such as a sensation of heat or smell. It 1s true, it
might be said, that the sensation only exists m myself and has
no independent being apart from me, but 1t is also true that
when I have a sensation this fact is not before me, I simply feel
the sensation; this is complete in its own quality and this 1s
all that I experience. To this we nught reply once more that
having a fcelmg 1s not the same as apprehending 1t. 2 But of
course the only satisfactory answer must be got from an examina·
tion of the fact of this kno" ledge itself If I am to know a given
sensation, I must recognize m 1t a special quality which makes
1t the definite sensation 1t is, "hich in fact distinguishes it from
others and prevents me from confusing it with others, as the
quality of a smell distmguishes it from the quality of a sound.
This does not mean that we explicitly compare the given sensa•
tion which we arc apprehending with others; rather the com•
parison is implicit and we have the sense of the definite quality
of the sensation as distinct from others of which we have had
experience. In order therefore to make our apprehension of il
1 p. 316. ' p JS, note 1; p 31~, note 1
314 STATEMENT, THINKING, AND APPREHENSION
possible there is really a complexity in our thinking and a com•
plexity of a certain kind in the object, and there is no appre•
hension of the given object, or given element in reality, possible
except as an element m this completer apprehension conveyed
in the comparison, whether explicit or implicit, with other
elements of reality. Thus we do not get a simple apprehension
but one which corresponds to a statement (and statement is
necessarily complex), the statement, although not necessarily
expressed, being one of relation 1 But besides this, if we appre-
hend the given sensation, we must apprehend it as our sensation,
which makes 1t an element in our being, and so we come really
to the d1stmction of subject and attribute, or element corre•
sponding to attribute, as m the former case, and to the appre•
hens1on of the sensation as necessarily an element in a reality
upon which 1t is dependent. The apprehension also itself is an
element m a wider apprehension. The objections therefore seem
to be answered.
§ 135. Confinmg ourselves still to the statement of knowledge
and the thought which corresponds to it, that is to the act of
knowing of which judging is one form, we must now ask whether
the apprehensions of the elements m the object, themselves
therefore elements m the apprehension of the whole object, are
by themselves true or false This corresponds to the incorrectly
phrased question whether conceptions are true or false considered
m themselves
If we call these apprehension-elements themselves apprehen•
s1ons, as we have done, we may seem to have answered this
question 1mphc1tly already ; for truth 1s nothing but the appre•
hension of reality, and thus, 1£ these elemental apprehensions
are apprehensions at all, they must be true. This kind of
language implies that the given apprehensions can be separated
from the whole to which they belong and that, so separated
and in themselves, they are true. They would then be what
Aristotle calls • thoughts apart from synthesis. 8 He says that
taken apart from the synthesis m thought, the thought that is
1 Cf p 312. • Cf. f u6.

[• V1de p. 297, note 1 The reference 11 to ri ar&paM alml nl ri ,,,,,.,. he••

T¥ 41'111 a-wfla-tG1s nl 81acplO'tG1r a,D4pe1T1.
DI ltd, 16& 14 Cf D, Aft. 43(1A 26.]
 .A.Hw1lumi.o,i, COHtlPlion, S ~ 3:r5
which corresponds to the statement, they are neither true nor
false, though as we shall see he came to modify this doctrine.
But now, in strict accuracy, these apprehensions, which are
rightly so called, cannot be separated in apprehension from the
wholes to which they belong. This gives us the answer. The
elements of the object of which they are apprehensions have no
independent nature and no reality at all except as elements in
the reality of the whole object. Similarly these apprehensions
themselves have no existence at all as apprehensions except in
the act of thought which is the total apprehension of the object.
The result then clearly appears to be that these thoughts which
are elements in the whole thought constituting the act of know•
ledge and which correspond to what are erroneously called
conceptions in the ordinary distinction of conception and judge-
ment are true, but not true as taken by themselves, for they
cannot possibly be taken by themselves. They can verbally
appear to be so, but there is no real thought corresponding to
the verbal expression. They cannot really come into our thought
at all except in the wider apprehension to which they belong.
The question then : ' Can these elements when taken by them•
selves be true or false~• illustrates once more a point, on which
I have often insisted, that a question itself may be fallacious
m form and thus imply at the very start some erroneous view.
The fallacy here is that the question implies the false view that
the elements can be taken by themselves.
There are some qualifications to he made ; they need however
cause no difficulty. In a statement the nommat1ve case to the
verb may be the name of something which has been apprehended
as having a certain nature before the given statement was
arrived at. The statement, then, expresses some new knowledge
of it, that is, the apprehension of something new in it. Thus
the object as at first apprehended may seem hke an element
in the object which is the subject of the new apprehension, and
the apprehension of it therefore may appear to be an element
in the new apprehension. If this were so, it might seem that
the first apprehension of the given subject could not be had by
itself, according to the principle developed above. But this is
not wholly correct. If the subject of the sentence is a dependent
element in the existence of some object, it is not true that it
316 STATEMENT, THINKING, AND APPREHENSION
is an element in the ob3ect which is the subject of the new
apprehension, for it is identical with this object. Nevertheless
it remains true that the apprehension of 1t cannot be had by
itself, because the apprehension of it, which 1s independent of
the thinking that issues in the given sentence, was gained m
another and preceding apprehension in which it was an element
and therefore not to be taken by itself The case however is
different if the subject of the new apprehension is an object
which can be apprehended by itself, a substance m fact. This
again 1s not an clement m the obJect of the new apprehension
but 1s identical \\ 1th 1t, and, as bcmg a substance, involves an
apprehension of itself which can indeed be taken by itself but
1s precisely the apprehension which represents a whole. The
same is true of any other noun m the statement which 1s the
name of a substance. This at once suggests the question as to
how we can d1stmguish such a noun from a sentence, and the
thought which corresponds to 1t from the apprehension which
corresponds to a sentence, smce both the noun and the complete
grammatical sentence seem to correspond to the apprehension
of a substance. \Ve may be qmte sure that the difference in
grammatical expression has a real importance. Suppose we have
a statement of the form X 1 is Y, where the nommat1ve case
means a substance already apprehended as having the charac-
ter1st1cs symbolized by X (or 1t may be by the full X 1). Then
there arc two cases. First, "e may actually apprehend this
character of X 1, and from that go on to the apprehension of 1t
as Y. This 1s expressed m the statement X 1 1s Y. Or, secondly,
we may merely remember that we apprehended X 1 as X, but
not be actually conductmg the apprehension. In the first case,
the verbal statement which accurately corresponds to our
thought 1s ' this (1.e X 1) is X and, as bemg X, it 1s Y '. In the
second case, the apprehension of X 1 as X 1s presupposed and
we are not concerned to state it. It appears only ac; an element
presupposed m the c1.pprehens1on which we are concerned to
state as being the nc,v apprehem,1011 Thus X 1 has not the form
of a statement, bemg a noun or noun-phrase in the sentence,
while X 1 is Y has that form. 1 In the doctrme we are criticizing
X 1 would be said to be a conception, and X 1 1s Y the judgement
1 § 134,
 Ap;welunsion Conception, Statement
1 317
in which the conception X 1 is an element. As contrasted with
this, the result arrived at above may be put shortly thus :
the verbal expression X 1 implies an apprehension and, in the
manner explained, involves the memory of 1t, but does not state
it, whereas X 1 is Y states explicitly what is apprehended.
Here we have perhaps come across one of the main reasons
why X or X 1 1s called a conception. As we have already
observed, we should not naturally call anything which we recog-
nized as an apprehension by the name conception. Now X 1
does not here represent an apprehension which 1s being realized
(as for instance an act of perception), but an apprehension
remembered to have been experienced (associated in most
cases with an imagination or memory-image). There is an
apprehension of the obJect m an md1rect sense, in so far as we
remember that there was an apprehension, but the reference to
the object 1s remoter than it was m the apprehension of the
object which is remembered to have been experienced. The word
conception lends itself to this, because it seems something in our
mind detached from the obJect and so is regarded as something
· conceived ' as opposed to · asserted '. Probably also the asso-
ciation of the mcmory-unage helps to conceal the vagueness of
the terms conceiving or conception. If we pressed the question
as to what the conception was, we should find that, while there
was virtually a confusion of 1t with a mental image, the reply,
when attention was once awakened by our qt1estlon, would be
that the conception was not a mental image. It is nevertheless
thought of, in some obscure way, as though 1t were a kind of
special mental entity This fiction, for it deserves no better
name, seems to have greatly influenced psychology, and we
regularly find m manuals of that science a chapter on the
• formation of conceptions ', in which, one may be allowed to
think, there lurks a fine confusion of the mental image with
reminiscences of conceptuabsm and even of the old copying
theory of knowledge.
The same considerations apply to nouns other than those in
the nominative case, and to any other words in the sentence
and the thought corresponding to them, when the apprehension
of that to which they refer has preceded the act of thought
which corresponds to the given statement. In ' lead is heavy '
 318 5TATJUIENT, THINKING, AND APPREHENSION
heaviness would be said to be the conception synthesized with
the conception lead.
I 136. So far we have confined ourselves to the act of know•
ledge and to the statement in which it issues, inasmuch as it is
from the act of knowledge 1 that the other forms of thinking
have to be explained. It will now not be difficult to deal with
these other forms, namely opinion and belief. 2 For the present
purpose belief may count as opimon. Suppose for simplicity
that, in the instance X 1 1s Y, X 1 and Y refer to realities previously
apprehended and therefore true m the sense explained, but that
the statement itself 1s only a matter of opmion to the pers~,n
pronouncing it. After this he will thmk of X 1 as Y, so that
we might symbolize a new op1mon by ' X 1 (Y) is Z '. Here it
will be observed that we have a parallel to the first sense of
conception which we discussed m the case of knowledge, the
conception of X 1 as Y. The difference 1s that, while in both
cases X 1 1s thought of as Y, the first depends on the true
apprehension X 1 1s Y, whereas the second depends only on the
opinion that X 1 1s Y. In the first case the connexion of X 1 with
Y is known : in the second case 1t is not known and, though
surmised, remains a problem The consequence is that, if we
allow the expression conception of, the true conception is that
which corresponds to the true apprehension X 1 is Y, while the
second is not a true conception m the full sense, but only in the
qualified sense represented by the phrase problematic concep•
t1on. Where then this use of ' conception ' corresponds to
a matter of knowledge, the elements in the obJect signified by
X 1 and Y are understood and known, as well as their connexion.
In the case of opinion, the connexion 1s only a matter of con•
jecture more or less probable 3
The problematic conception seems rightly called untrue if the
opinion that X 1 is Y on which it is founded is itself untrue. In
the example taken the untruth does not lie in X 1 and Y, or
rather in the thoughts corresponding to them, but in the assump•
tion of their combmation. This naturally leads to the question
whether the problematic conception, which may be untrue in
the sense explained, must always show such a complexity, and
consist in thinking of a combination of real elements as possible
1 ff JO, 47 • Part II, ch 3 • pp. a6o and io7.
 Af,11,eluftsfaff., Co•"1'tto•, SkdMM11t s19
which is in fact non-existent and impossible ; or whether it i•
possible that a thought without any such complexity, what is
usually called a simple idea or a simple conception, can also be
false. If we were concerned with truth and knowledge only,
the so-called simple conception, or simple idea, would be either
the apprehension of something simple and without complexity
m the object, or the thinking of it as real through the memory
that we have apprehended its reality. And here it must be
remarked that the existence of such simplicity is not affected
by what has been said above of the multiplicity which even
a sensation, however simple it may seem, must involve, both in
the comparison of it with other sensations and in the reference
of it to its mental subject for notwithstanding these references
the element is properly called simple if it is not itself a unity
of discernible differences. An extreme case would be a relation,
since it implies two different objects at least But a relation
may be perfectly simple m the sense that its nature is not
a combination or synthesis of different elements. Consider for
instance the relations greater than, or equal to. Though equality
must be between two different things, its own nature is perfectly
simple and presents no unity of elements. We may see the same
again m an instance like the quality of white colour, and such
examples could be multiplied.
§ 137. We return to the question whether the so-called simple
idea or simple conception can be false. This concept1onal
terminology is not only vague but hides the true difficulty
which occurs in the putting of the question. We want apparently
some state of consciousness, concerned with something or other
which is quite simple, parallel to the state of consciousness which
consists in the apprehension of a simple object, or rather perhaps
of something simple in an obJect. But now what could this be
and what is it that could correspond to the simple object ? Let
X be this simple somewhat for which we are looking. By
hypothesis X cannot be anything real, for the apprehension of
it would necessarily be true, and the mistake, if possible at all,
would consist in taking this X to be real, though 1t is not any-
thing real. We then still want to know what this. X is, for it
clearly will have to be something for consciousness. If it
were a sensation, for instance, it would be a reality in the widest
320 STATEMENT, THINKING, ,•_.'ND APPREHENSION
sense of the word real as opposed to the imaginary. Nothing
seems to be left for it except that it should be an imagination.
But, if it were that, we should recognize that it was so, for
we never confuse an imagination with an experience, and we
should not take X to be something real ; and hence the supposed
mistake could not arise. This should be, in philosophy, a decisive
argument, but it "ill not convince those who, like Hume, make
a very common mistake about imagination, more or less con•
sciously confounding the unagination with something which is
of the same kind as the corresponding reality. 1 The imagination
is in truth totally different in kmd from the real object imagined ;
the imagination of a sound, for instance, is not itself a sound
and cannot be mistaken for one. If it is objected that in a dream
we mistake an imagination for a reality, the reply is that we
have now to do with the wakmg and not the dreaming state.
Moreover, the dream-obJect is somethmg quite different from
the imagination. It 1s mdecd probable that the two are usually
confused m speculation under the one term mental image, but
their difference is a matter of consciousness , we know that the
most vivid imagmat1on for instance of a band playing is some-
thing totally different from our dreaming that we hear a band
playing.
But, if we allow this false account of imagination, so that the
imagined sensation for mstance would differ (as Hume thought)
from the real sensation only in vividness, the imagined sensation
would be an experience JUSt like its prototype and JUSt as real;
so that here again there would be no room for the so-called
false conception. Fmally, 1t 1s a commonplace that the mind
cannot itself create a new sunple idea. The assertions of Locke
and Hume, for instance, to this effect are familiar According
to them we can have no 1magmation of the absolutely simple
of which the original is not an experience, and 1t will be admitted
that their doctrme, m spite of its overstatement and whatever
the defects of its expression, is in some important sense true.
§ 138. No one supposes, of course, that the imagination or
mental image of a house is the same as the experience of an
actual house. Nevertheless there is a tendency to think of the
imagination of a sensation as if it were itself a sensation, though
1 Tms i1 a m1Stake which Mr, Bradley makes with fatal conaequezacea.
 A,PP,-ehension, Conception, Statement 321
fainter in intensity than the sensation of which it is the image.
Thus with Hume the idea (imaginary image) of a sensation
differs from the impression of the same only as something less
vivid. There is doubtless a great temptation to this view, for
the imagined sensation, e g. when we remember an experienced
sensation, 1s said to be like the one previously experienced and
so may be called a mental image of 1t. Thus the two might
seem the same in kmd, for 1t might seem that what was like
a sensation must be itself a sensation. On this view, when we
remember a sensation of heat (e g.) we should imagine it and
this would mean that we were conscious of an actual sensation
of heat ; not the former one, of course, but another sensation
of heat resembling the former one in the quality of heat, and
that we were conscious of the resemblance. If this were so, there
could be so far no reason why the mental image of the previous
sensation should be any less vivid than the previous sensation :
yet writers, bke Hume, who practically hold the view before us
are careful to say that it must be less vivid , indeed Hume says
that there is a degree of vividness which no idea can attain.
In the theory as such there is no justification whatever of
this, but 1t i5 easy to see ho" the necessity for the assumption
arises.
Suppose \\e think of a pain under which we suffered much.
We imagine it and, 1£ the mental image we1e itself a pain of
the same mtens1ty, we should suffer as before ; but we do not.
We remember a dazzling light, but our mental image 1s not
accompamcd by the pam and mconvemencc of dazzlmg. It is
thus necessary to make the mental image at least a less intense
sensation than the original, 1£ it 1s to be a sensation at all.
But the truth seems to be that tlus modification 1s not enough,
for the mental image would still have to be pamful even though
less painful and there must be some suffermg even though little ;
but there appears to be neither the one nor the other.
It seems however possible in the case of some kinds of feelings,
though apparently not in all, that the memory of the past may
affect the bodily organs and produce a new present actual sensa-
tion like the previous. Thus, if I sec something the eating of
which made me sick m the past (e.g. green elderberries), it does
seem as 1f I not only remembered the past sickness but actually
1773•1 y
322 STATEMENT, THINKING, AND APPREHENSION
experienced a sensation of sickness It would be a confirmation
of this if vomitmg actually ensued and it would be interesting
to know 1f this ever really happened to anybody.
But m the first place this actual sensation is not the mental
image of the old one and the latter is not pamful. Moreover
this is not a mere contention, for we should recognize in our
ordinary expencncc the difference between this feelmg of s1ck-
nesi, and the mental image of the one previously experienced.
If, however, this should be disputed, there arc other considera•
hone; which seem decisive
Suppose the image of a past sensation, such as is present
when we remember a past sensation, is itself a sensation, then
of course 1t 1s uc; real an experience as the former sensation.
The nature of its cause makes not the slightest difference to
itself · m brief 1t 1s an experienced sensation and as much
experienced as the original :5ensat1on of which 1t 1s said to be
the ' image '. The analysis of the act of memory would then
be this We had an actual sensation S1 , we have now another
actual sensat1011 S2, and we remember that we had the prev10us
sensation S1, and we say to ourselves that 1t was hkc S2•
This leads to two d1fficult1es, e.1ch fatal to the theory. (1) By
hypothcc;1s the- so-railed mental image of S1 1-; Sa, and S2 thus
'represents' S1 Clearly then, as S2 represents S1 m quality,
winch means that S 2 ac; a sensation must be of the same kind
or quality as S1 , so also 1t must represent S1 m intensity, and
must therefore, to do tlus, be itself an mtensc sensation. Or,
to put 1t otherwise, suppose S1 of appreciably less mtens1ty than
S1, and l': 1 a previously experienced sensation of intensity not
appreciably diffrrcnt from S11--an experienced sensation may have
of course any degree of v1v1dness and mtensity less than that
of S1-then S2 would represent I 1 and, JUst as a sensation of
less intensity than S2 would represent not a sensation .I1 but
a sensation ~2 less m intensity than ~1, so a sensation of greater
mtens1ty tl1an Sa (and not S2) \\ould represent the sensation S1 •
Thus the mental image of a previously experienced intense
sensahon, if. itself a sensation, must be an intense sensation.
If then I remember an mtensc- toothache, I must have an actual
sensation of intense toothache and suffer accordingly, which is
contrary to experience.
 Apprehension, Conception, Statement 323
Suppose again we had what we call an actual toothache
but of low intensity, we should not say that this was like the
previous sensation, except m mtcns1ty, but should say it was m
a (for us) important characteristic not like 1t while we should
estimate a toothache of similar mtensity as hke 1t. Now, as the
imagined sensation 1s, by hypothesis, itself a really experienced
sensation, we must think exactly m the same way about 1t and
thus we should tlunk that the 1magmed sensation (which must
be of such low mtens1ty, 1£ a sensation at all, that we do not
suffer from 1t) was not hke the past sensation We should m
fact only think ' 1magmed sensat10ns ' to be hke past experiences
when they were appreciably of the same mtens1ty. And, if it
happened that 'unagmcd &ensat10ns ', bcmg also (1t must be
remembered) actual sensations, could not exceed a certain degree
of intensity, the result would be that we should have mental
images of all scnsat10ns previously experienced up to that degree
of mtcns1ty, but of none appreciably exceeding 1t. Thus, if the
presence of the mental image were necessary to memory, we
should not even remember sensations which pass beyond a
certain mtens1ty, an mtense toothache for mstance could not
be remembered
(2) The mental image S2 is an experienced sensation, as actual
as the previous sensation S1 of which 1t 1s called an image, m the
theory now m question, J.ntl it would not be an ' image' at all
unless we not only experienced 1t but estimated that 1t was like
the past scnsat10n Now th1s, bemg memory, is not a mere
consciousness that we have experienced something of the sort
before without knowmg what sort-not the mere recognition
that this kmtl of sensation 1s already familiar-but involves
a knowledge of what it was that was hke the present sensation.
We say the past sensation was like the present and we therefore
compare the two together. We have not got the past itself
however to compare. We must have some representation of 1t.
This representation must be the mental image of S1 ; thus S2 ,
which is to be compared with it, is distinct from 1t Hence either
S1 1s not the mental image, or we have a second mental image S 3
which must equally be a present actual sensation : and this
process will obviously go on ad infinitum.
Or we may put it like this. If we are to compare the past
Y2
324 STATEMENT, THINKING, AND APPREHENSION
sensation S1 with the mental image S1, itself a present sensation,
we must remember the past sensation and 1t 1s what we remember
of it that we compare with the present sensation S2• Certainly,
1f we do not remember it in any way, no comparison will be
possible. But, by hypothesis, memory is the possession of a
mental image together with the 'Judgement' that it is like the
past. It 1s then a mental image which has to be compared with
S1• Thus. as before, S8 cannot be the mental image or we are
Jed to the mfinite process.
It seems then to follow that the imagination of a past sensa-
tion cannot be identified with a present actual sensation of the
same kind
§ 139. There arc also some special cons1dcrat1ons which con-
firm this In the case of some sensations the memory of a past
sensation, however mtense the sensation may have been, does
not seem capable of producing a real present sensation hke the
feeling of sickness already instanced-a sensation which, it
must be remembered, would be distmct from the mental image.
For instance, we may remember some great volume of sound,
as such a fortissimo passage of an orchestral piece. We may say
we ' fancy ' we hear the blare of the trumpets, but we do not
suppose in the least that we have an actual sensation of sound
(' Heard melodies arc sweet but those unheard are sweeter').
There 1s indeed nothmg here analogous to the feehng of sickness
which seems to exist m the case we have considered.
Sound gives a further and more dec1S1ve confirmation of what
we are mamtainmg It is of the essence of the perception of
melody as opposed to that of harmony that a relation should
be perceived or apprehended between successive sounds as suc•
cessive. For this 1t 1s necessary that a past sound should be in
some sense retained 1t must be m some sense remembered, for
clearly without memory we should only be conscious of the
sound at the particular moment and not of its relations to
the sound which was 111 the past and has ceased. Now, if the
retention were the retention of a memory-image of the sound
and this last were an actual present sensation of sound, the
comparison and memory necessary to melody, or rather the per•
ception of melody m a series of sounds, would result in the
simultaneous presence to consciousness of the sounds constituting
 Apprehension, Conception, Statement 325
the series, or rather of actual sounds equivalent to them. This
of course would yield the perception of harmony, if the interval
between the successive notes were consonant, and of discord, if
they were not ; it would not yield a consciousness of melody.
The perception of melody then necessitates that the retention
of the sounds in the memory, after they have ceased, should
not be, nor involve, the presence and consciousness of actual
sensations of sound The imagination of a sound 1s not itself
a sound. The mental image, or the mcmory•1mage1 of a sound
is not itself a sound.
It must be observed that tlus has a serious consequence for
a theory of memory which seems much favoured by psycho•
logists. It 1s supposed that to a given expenence-f>ay a sensa
tJon of sound-corresponds a certain definite act1v1ty of the
nerves, to which they arc excited by the experiential stimulus.
Memory 1s explained by supposmg that the nerves are somehow
(1t cannot be by the experiential stunulus) stmmlated again to
the same kmd of act1v1ty The nerves m short have a certain
function to which they are stimulated m experience and they
arc in memory stimulated to perform 1t agam. But, 1f so, the
c1.ct1vity being the same m kmd each time, if the first act1v1ty
conveyed the consciousness of an actual scnsat1011, so must the
second ; and the memory of a <,ensat1on, or the imaginary
sensation, would itself be a sensation This 1s of course 1mpos·
s1ble and therefore so also 1s tlus theory of memory. The
difficulty is doubtlesc: glossed over by assuming ti.at the activity,
m the case of the memory image, attributed to the nerves, 1s
weaker or less intense than m the case of the real experience,
but as the forego mg sho\, s th1& makes no difference.
Mr. Bradley, though he does not offer th1., phys1ological
explanation of memory, falls mto the same error. 1 In Assoc1a·
tion what he says happens is tlus The soul in an experience
has two elements m consciousness, :.c1.y A and B. This experience
is called a ' function ' of the soul. In • Assoc1at10n ' 1t 1s sup·
posed that the function 1s for some reason or other repeated
and thus the image (or reality) of A 1s followed by the image of
B. But as the original function was an • experience ' so must
the repetition of 1t be. If A and B were sensations, the repetl·
1 Logic, II, 11, ch I, §§ 30-31.
326 STATEMENT, THINKING, AND APPREHENSION
tion of the function must mean the presence of actual sensations
hke both A and B. The confusion 1s most obvious when, on the
second occasion, an experience of A actually occurs but (by hypo•
thesis) not an experience of B.
§ 140. We will return later to this question of simple ideas, &c.,
in order to devote to them some more thorough consideration. 1
For our present purpose 1t 1s enough to say that an entirely
arbitrary creation m imagination is impossible and to concede
that our power 1s limited to effectmg m the imagmat1on new
combinations of simple clements, derived from experience.
The result we appear then to have reached 1s that in any
statement the ' simple idea', or ' simple conception', must be true.
Observe that the formulation of the supposed result is qmte
easy as long as we make use of this vague language about con•
ccptions and 1deas • But, 1f we substitute what seems to be the
true attitude of thought confusedly represented by these words,
we shall find 1t scarcely possible even to put the question ; so
that the answer 1s really nugatory What 1s the simple con-
ception? It is c-lear at any rate that 1t derives its meaning from
the simple property of the obJcct The only possible answer
seems to be that the simple conception 11:, the apprehension of
the simple property. But that 1s necessarily true and so the
question could not arise Now, as we have seen above, we can't
here treat the concept10n as bcmg the thmkmg of the thmg to
be so and so, or the trcatmg 1t as so and so The only remaining
possibility then secmr,; to be that 1t is the memory that we
experienced the given quality, or else the mental imagination
of the given quality But no one, m ordmary life, would call
his memory of X 1 lns concept 100 of X ; and the other alter-
native, that by the conccpt1on of X we mean its mental image, 1s
evidently false Once more then the word conception has
properly no apphcatwn here at all. The ordmary natural
language bears us out, for no one would ever use such expressions
as my conception of blue, of hot, or of loud. If 1t be objected
that, inasmuch as we do meet with such expressions as the con-
ception of force or of duty, we arc here neglecting the evidence
of language, we reply that the normal language 1s entirely in
our favour ; the apparent exceptions belong only to reflective
thought wh1d1 has advanced to the plulosoph1c stage. Indeed
1 Part III, ch. S
 Apprehension, Conception, Statement 327
this use of the term conception is mainly due to the false
philosophic theory of the universal called Conceptualism.
One may be allowed to repeat the opinion that the writers
who use the term simple concept1on in such ways as the con-
ception of force, of cause, or of change are thmkmg of a mental
something dangerously hke that utter fiction the mental image,
and that psychology in seekmg for an account of the formation
of conceptions is mamly, if perhaps not consciously, affected by
the same :fiction. In spite of chapters devoted to the supposed
psychological processes by which this conception 1s arrived at,
we must stoutly pronounce it a confused illusion.
§ 141. If the foregoing mvestigatlon is correct, there 1s no
active apprehension which is of the entirely simple. The
ordinary forms of the statement probably began with the dis-
tmction of subject (substance) and attribute, the later and more
reflective thmkmg developing at last the form m wluch the
attribute itself may appear as nominative Probably also the
form which states relations, between substances m the first
instance and afterwards between elements of reality m general,
was developed soon after the form of statement concerning
substances Now m our present developed experience it always
happens that the substance, or subJect, itself has been appre-
hended before, and s1mtlarly the d.ttnbutes. And, 1f WC' look at
such previous apprehensions, we find the sctme tlung must be
said of them, and so a further apprehension such as corresponds
to statement must be presupposed But tlus r..:gress cannot go
on for ever. A common way of avo1dmg the difficulty is to sd.y
that the regress ts not always to a statement (or rather to the
thought corresponding to a statement) but to certam simple
ideas such as white, straight, or hot, and that i,tarting from
these as material we build up statements from them. This 1s
the very doctrine we have been denymg No doubt the infinite
regress of which we seem m danger, on the view we have been
defendmg, greatly strengthens the popular view that we begtn
with simple ideas in the manner of Locke, who here as usual
is elevatmg the popular view to philosophy. Once more we have
occasion to state an important philosoph1cal prmc1ple. If we
find on careful analysis that a given attitude of consciousness
involves and presupposes a constderJ.hle complexity of what are
erroneously called conceptions (more properly apprehensions),
328 STATEMENT, THINKING, AND APPREHENSION
a complexity very much exceeding our expectations ; if that
attitude of consciousness further implies a much more developed
phase of consciousness than we were prepared for, one perhaps
difficult to reconcile with some theory which we have been
accustomed to rely upon, we are m great danger of being led
to whittle away the facts to suit our prcJud1ces and theories.
This docs much harm to the development of either sound
psychology or sound philosophy What we have to do uncom-
promismgly 1s to try to find out what a given act1v1ty of thought
presupposes as imphc1t or explicit Ill our consc1ousncss, without
allowmg these precious results to be mtcrfcPed with by any
preconceived opinion I may give an mstance of a very simple
character, the cond1t1on of anger Plato has already told us •
that anger mvolvcs a certam amount of reason, and 1t really
involves a very considerable development of consciousness ;
more mdeed than we might expect before mqmry, masmuch as
anger seems to be a rudimentary emotion, bemg found m the
very young and, generally, m the dullest and most unreflective
consciousness But, when we contrast 1t with the mere con•
sc1ousncss of pain suffered, we sec that 1t 1mphcs the idea of the
causality of another person. It implies therefore not only that
the consc10usncss has already apprehended the fact that there
arc other persons, but the category of causality too, because the
hurtful act is attributed to the other person, or there would be
no room for anger. But we must go even further than this;
the categones of will and of mtention are both implied. There
would be no anger unless 1t was thought somehow that the
aggressor had caused, willed, and intended the m1sch1ef suffered.
This 1s perhaps more than w~ expected, but we must not on
that account suppose any of the clements absent. They are
absolutely necessitated by the fact of anger and, 1f they are not
there, there 1s no anger. This 1s the true method, we must try to
get at the facts of consciousness and not let them be overlaid as
1s so commonly done with preconceived theones. We may know
nothmg whatever about the first beginnings of a given state of
consciousness, but we may perfectly well know what it must be
whenever it does appear. S1m1larly m the case before us, we
know next to nothing of the beginnings of the apprehending
consc1ousness, yet we are perfectly well able to say what the
[• The reference 111 to Rep. 440, prei.umably.]
 Apprehension, Conception, Statement 329
act of apprehension presupposes, provided only it is a true
apprehension. So far we are not concerned at all to say how the
first apprehensions came about, but we know what they must have
been and must have presupposed when they did come about.
It will have become evident that the regress ordinarily supposed
to the simple idea is quite impossible, because as we have seen
the simple idea cannot be apprehended by itself. If it be true,
as psychologists seem all to assume, that consciousness begins
from sensations only, we have to reply that there is a d~nger
here of a confusion m the use of the word consciousness. We
must repeat that that consc10usness which consists in having
a sensation ts not the same as the consciousness which consists
tn the apprehension of a sensation. If agam we assume what,
unless I am greatly mistaken, is the opinion of psychologists,
namely that the first begmmng of apprehending consciousness
is the noticing of a sensation, then "'C should contend that such
noticing implies a contrast more or less dimly between the given
sensation and the context of previous expenences. And further,
the beginning of the apprehending consciousness, not being m
the apprehension of a single sensation, hes either in the appre-
hension of the different quc1hty of two sensat10ns, m an act of
consciousness m which they each for the first time become
matter of apprehension, or m the apprehension of the quality
of the g1ven sensation in vague contrast with the precedmg
context of sensations. This latter docs not mean that the vaeue
apprehension of such a general context precedes the experience,
but that the context itself only becomes matter of apprehension
m any sense however dim m this sc1me act in which we ' notice '
the given sensation. If this noticing of the quality of a sensation
tnvolved its reference to ourselves, wc should have the appre·
hensions of subject and attribute, neither, as before, presupposed
but both becoming matter of apprehending consc10usness m one
and the same act. It must not be forgotten that the first kind
of apprehension we spoke of is not naturally expressed at all in
a. statement ; the reason of this has been explamed already. It
really also mvolves what one might call the unconscious use of
the distinction of sub1ect and attribute, m so far as the two
contrasted sensations appear as members of our general appre-
hending consciousness, however vague the apprehension of the
two sensations, or of a given one of them, may be.
 xv
THE QUANTITY OF PROPOSITIONS AND
THE UNIVERSAL
§ 142. PROPOSITIONS arc usually distinguished m point of
quantity mto Universal (all A 1s B), Particular (some A 1s B),
and Singular (th1s A 1s B) propositions
The d1stmct1on of the particular from the universal is some-
times erroneously stated thus · the umversal judgement attaches
its predicate to the whole of 1ts subject wlulc the particular
attaches its predicate to a part of its subject. Now clearly
a statement would not be a statement if 1t dtd not attach its
predicate to the whole of its subJect. In the form, some A 1s
B, B is attached to a part only of A, or more accurately to
a part only of what 1s A. But some A 1s the subJcct, not A;
and to the some A wluch 1s meant, the true subJcct, to the
whole of that the predicate 1s attached What 1s common to
the statements, all A 1s B and some A 1s B, 1s the class to which
their subjects belong, the class A But the subject of the first
1s the whole class A and of the second only a part of 1t. The
cause of the 1llus1on, that there 1s a common subject, 1s that m
both statements ahkc the only defimtc part of the conception
of the subject is the same, namely the concept10n of Aness, the
rest bcmg, m the case of the particular c;tatemcnt, mere indefinite
particularity. But further, we must call m question the right
to make such a d1c;tmct1on at all between the umversal and the
particular statement The sentence, some A 1s B, at once pro•
vokes the question, what A 1s B ? If we arc starting from the
particular in some A 1s B, we must know what A is meant and
so must those whom we arc addressing 1£ the statement is to
be accepted. We see that m this case the verbal expression is
imperfect and that the true form, which would correspond to
the actual thinking, 1s CA 1s B, or all CA 1s B, which 1s a universal
statement. But now there is a second way in which we can
reach the statement, some A is B, namely, by immediate infer•
 The Quantity of Statements 33I
ence from all B is A. Here it might seem that the further
determination of some A need not be possible. Now observe,
first, that here the thought does not really start from an indefinite
particular but from a universal statement as before, from all
B is A and, secondly, that some A has not really the indefiniteness
which 1t seems to have in the verbal expression, for 1t really
turns out that it 1s that part of A which constitutes the whole
class B, and which 1s therefore known under a definite universal.
It is again the form of the expression which is imperfect, for
the 1mmed1atc inference should be that some A constitutes the
whole of B But, agam, we may appear to get the particular
form, some A 1s B, from the universal d1sjunct1ve all A is B or
C or D (which we must notice 1s a universal statement, in the
true sense of universal) Yet even m this case we may be able
to say after all what A 1s B In geometry, triangles being either
eqmlateral, 1sosccics, or scalene, we do not say some triangles
arc eqmlatcral ; at all events we do not naturally use tlus form
of expression, and we never start from such a form. We know
that <'qmlateral triangles result from a certain definite construc•
t10n, and in Euclid that 1s the only way in which we learn that
triangles can be equilateral, for !us first proposition of the First
Book 1s not merely a way of constructing something that we
already know to be possible, 1t 1s really our first information
that the thmg 1s possible ; the definition of equilateral triangle
which has preceded being, until the demonstration appears, as
some say, merely nommal, or as 1t may be Letter put, pro·
blematic For the statement, some triangles are equdateral,
the mathematician has no use, the proper sc1ent1fic statement
which corresponds 1s that three equal straight Imes may be so
placed as to form a triangle. Thus 1t 1s clear that the state-
ment, some triangles arc cqmlateral, lb only derived from this
universal statement by conversion, a case which has already
been considered
This account docs not smt such an example as ' some Imes are
straight ', Imes bcmg d1v1ded into straight and curved, because
we cannot give a general construction of a straight line. A
general account may be given which covers all cases. The
statement, some lines are straight, doesn't occur naturally ; if
it is made, we ask what lines are str,ught, and the only answer
332 STATEMENT, THINKING, AND APPREHENSION
possible is, straight lines arc. We must go behind the verbal
expression and ask how the statement can be got ; what the
thought is to which it corresponds. We get it in the appre•
hens1on represented IJY the d1sJunctive statement 'a hue must
be straight or curved', or' linearity must take the form of rectl·
linearity or of curv!lmcarity '. This is not a statement about
something mdefimte, some undefined cases of a universal, but
about the quite defi111te u111versal lmear1ty. Some Imes arc
straight, therefore, only represents an clement m the apprehen•
s1on or Judgement, expressed m the d1sJunctlve statement. It
can only be ;udged m that statement and not by itself, and
is really an imperfect verbal expression of the thought which
underlies the d1s1unct1ve statement. This is why the statement,
some Imes arc straight, does not arise naturally either 111 science
or in ordinary statement. It 1s an mcorrect form of expression.
It follows therefore that v. c must mamtam that the d1stmctlon
of umvcrsal and particular statement cannot be recognized in
this sense, for the particular always turns out to be the nnpcrfect
verbal expression of a umversal statement.
Why then docs the particular form e>.ist in ordinary language ?
The facts seem to be these Suppose for example we have 'all
CA 1s B ' ; we may put this m the form some A 1s B, simply to
characterize the relation of our statement to another statement
which concerns A and B only, namely, all A 1s B This we do
when one statement, CA 1s B, 1s considered as evidence in favour
of 'all A 1s B '. W1sh111g to prove that all A 1s B, or at least to
find out whether this 1s so or not, we put our knowledge, all
CA 1s B, in the form some A 1s B, om1ttmg the C because our
ultimate obJect is to omit the C. We mdicate by the word
' some ' that our knowledge 1s not yet sufficient to do this. Or,
1£ we desire to refute the statement no A 1s B, we put all CA
is B in the form some A 1s B because that 1s enough for our
purpose. The C as such 1s md1fferent, though we should be
obliged to supply it 1f we were asked to substantiate our state·
ment. This is the explanation of the under-statement of our
knowledge. The statement we make 1s relative to the object
that we have m view and yet 1t ultimately depends for its
validity on the special distinction which we have chosen to omit.
Lastly, we must call in question the usual way in which the
 The Uni7lersal 333
distinction between the universal and singular statements is
expressed. They do indeed differ, but the difference between
them is not properly one of quantity. All A 1s B normally, and
m demonstrative science always, means not merely that the
multitude of a class, exhaustible or inexhaustible, has an attri•
butive ascribed to every member of it, but that the universal
nature of the class necessitates the corresponding attribute. It
is not a matter of quantity and is m no sense a progress by
any kind of quantitative addition. This A is B, in its normal
use, is employed by us in cases where we are not able to assign
B to any universal manifested in ' this A ' ; the universal state•
ment all A is B <lifters from the singular, this A is B, not by
addmg more A's, or telling us that other A's are B, but by telling
us the reason why this A is B , this A is B because it is A, which
means that A as &uch necessitates B The distinction between
the forms is thus the distmcbon of the ' fact ' from the ' reason
why', not the distinction of the one (or some definite number)
from the all.
§ 143 Suppose we have the statement every X 1 1s Y, or all
Xi's are Y, where X 1 and Y refer to ohJects of ordinary expen•
ence or of scwntific knowledge and not to things of which we
come to have a conception for the first time m philosophical or
reflective thmkmg We arc accustomed to say that Vness 1s
a quality which all the X 1's have and that it 1s something which
they have ' in common ', or that 1t is · common to them all '.
This Yness we then call a universal Just bet:am.! of the fact of
its being common to all the particulars which are X1 • Thus we
think of it as a unity m a multiplicity and m the whole of
a certain multiplicity. Hence it 1s called a universal, and we
treat this characteristic of 1t as if it were its essential nature,
so much so that its name ' universal' is derived from this. It
would seem therefore as 1£ this bemg common to a number of
particulars was essential to its nature.
Yet 1£ we consider any particular mstance of Yness, as for in-
stance any X1, it has this quality Yncss m itself, and the distinc•
bon of \\hat we call 'universal' and 'particular' seems to be in
it, taken by itself. For instance, suppose Yness 1s circularity, then
every particular circle has what is called' the universal' (equality
of radii) particularized as the equality of its particular radii.
, 334 STATEMENT, THINKING, AND APPREHENSION
This is confirmed by the fact that through consideration of
a particular circle we arrive at truths about circularity m itself,
with no reference whatever to the particularity of the circle
before us (for example to the special length of its radii) and,
what is of more importance, with no reference to any other
circle m which ' circularity ' would appear. The distinction
therefore seems to exist withm the particular itself, and to
depend m no way on the presence of the so-called ' universal '
in a multiplicity of particulars. This is confirmed by the fact
that it makes no difference whatever to the d1stmct1on Yness
(circularity, for instance) recognizable in a given particular that
the number of other particulars showmg the same characteristic
should be diminished. It would make no difference if all the
rest ceased to exist; the given particular would have its nature
in no way altered, and yet what we have called the universal
element recogmzable m 1t would no longer be present m or
common to a mult1phc1ty. For instance, a sport might appear
once in the world of plc1.nts and be destmed never to be repeated,
and yet we should recogmze its general type. Tlus would be
impossible 1f umversahty as such was necessarily presence m
a multiplicity.
It 1s true on the other hand that the recognition of what we
call the same quality m different particulars may lead us to
distinguish its nature and recognize 1t m abstraction, as we say,
from the particular,; themselves, and 1t 1s on this account
probably, nay certainly, that we tend to look upon this umty
m a multiphc1ty as its necessary and mam characteristic. But
the cons1derat1ons to which we have been led seem to show that
this is quite erroneous, that the peculiar nature of the universal
does not at all he m its presence m many particulars, that on
the contrary 1t 1s because 1t has a certain nature which is inde-
pendent of presence m a mult1phc1ty that 1t can be the same
m a mult1plic1ty. Thus m Anstotehan phrase this presence m
particulars (many) 1s not of the essence of the umversal, but an
'accident'. (The usual terminology(' umversal '), mcludmg that
of Aristotle (KaTa '11'<1JJTos), unfortunately gives the contrary
impression.) This being so, the puzzle of the class with only one
member ceases to be a puzzle, and we may find the result just
arrived at of great help m solvmg some of the familiar difficulties.
 The Universal 335
What we have now to do, clearly, is to appreciate what the
so-called universal essentially is. Possibly, among other things
we may then discover why it is that we sometimes feel a difficulty
about calling something a universal although it can appear as
the attnbutive to many different nominatives.
§ 144. The universal is not a mere thought of ours. It is
something we ascribe to reality itself and suppose to be identical
m particular realities. In experience and m science (as Plato
saw) we always treat it as 1f strictly ob1ectme. Our difficulties
only arise in our philosophic rcflect10n upon this fact. Now
just as it is tl1c very nature of the universal to be a umty which
must take specific forms (number ar such must be odd or even),
so also 1t 1s its nature to be part1culanzcd The universal 1s
the umversal of particulars, and its reality cannot be separated
from them any more than its unity can be separated from its
species The d1stmct10n of universal and particular is ultimate
and self-evident; the nustakes and difficulties made about it
result from trying to explam 1t in terms of something different
from itself It reqmrcs no explanation; nothing can make its
meanmg plamer, and what that meanmg 1s we realize m particular
cases. It 1s above explanation, for we are constantly using the
distinction, and m any explanal10n of anythmg whatever must
use 1t and presuppose 1t.
In the above, the expressions ' it 1s the very nature of the
universal to be a unity which must take specific forms ' and
' 1t is its nature to be particularized ' are chose 1 to brmg out
the fact that diffcrentlat1on or different species of the genus,
and md1v1duahzat1on or the individuals, are nothmg outside the
nature of the universal and therefore do not require to be
reconciled with 1t. The universal as genus 1s not something m
the specific universals with the d1fferenba added to 1t as some•
thing outside 1t, so that the two together constitute the species
(as though the species agreed m the genus only and differed m
something which was not of the nature of the genus). Green
and red agree m being colour, but what they differ m 1s precisely
also colour , their differentia is not outside the nature of colour,
nor are difjerentiae in general outside the nature of the genus
they belong to. That 1s expressed simply by saymg that the
species are forms which the genus-universal takes. This unity
336 STATEMENT, THINKING, AND APPREHENSION
of the genus m the manifold of its species we understand in the
particular instances of it, thus number is either even or odd.
It is a unique kind of umty m multiplicity and we never think
of explaimng it or of expressing it m terms other than itself.
Plato, it is interesting to observe, though recognizing the fact
of the plurality of species of one genus, never made any difficulty
about it. We say without feeling a difficulty ' this is a kind of
colour ', ' this is a kind of surface '. Similarly the individual
or particular has not the universal in it and something also
beside the universal to make it particular. As the whole nature
of the species is covered by the genus-universal, so the whole
nature of the particular is covered by the universal. In this
particular colour there is nothing but colour. in this particular
number there is nothmg but number. The expression 'parti•
cularization of the universal ' has been used instead of particular,
or particulars, m order to emphasize the fact that the nature
of the md1v1dual is nothing but what belongs to the universal
itself. Particularization is of it and in it as much as differentia•
tion. The universal ' conception ', we can now say, is nothing
but the apprehension of the universal, that is of the umversal
as objective reality For want of understanding this there has
arisen a confused theory of some special mental something called
the universal conception, or the universal ' concept ', which
is not a mental image ( =imagined individual). This is a mere
fiction; there 1s no such mental entity. Just as the universal
cannot be, except as particularized, so we cannot apprehend it
except m the apprehension of a particular, either experienced or
remembered. 1
§ 145. In a universal statement the universal is obviously
apprehended and 1s referred to m the verbal expression ; but
the same is true, man important sense, of the smgular statement.
Aristotle affirmed that even perception is of the universal, in the
sense that we perceive an individual never as merely individual
but as individual with a quality, implying that such a quality
is somehow universal, so that with him the formula of a singular
statement of perception would be ' this is so and so '. He
1 Thi.a 1s the apprehension of memory, for in memory, even when of a
particular as sueh, we apprehend a. previously expenenced reabty by the help
of memory-wages, for memory is not merely having a memory-image (see
H 137-8).
 The Unive,sal 337
merely asserts this truth without any proof, thinking doubtless
that it could be justified by verification in the case of any
singular proposition. 1 As the particular then is inseparable from
the universal, so our apprehension of it, as expressed m the
singular statement of experience, involves the apprehension in
some way of the universal. The apprehension of any particular
as definite necessitates, as we saw, its distinction from other
particulars within the sphere of some kind of being which is
common to them and so universal. Thus we may distinguish
individual red, blue, or green colours. But further: we cannot
apprehend the definite quality of anything (say redness) without
apprehending what we should be prepared to recognize agam
as the same in a different case, m other words a universal. This
would be quite 1mposs1blc 1£ we had apprehended something
merely particular and entirely confined to the particular instance.
Even 1f we have had but one experience of a given quality, when
we recall 1t in memory we can think of other instances of 1t as
being possible. This, put merely in the strict form of enter•
taining the question whether there could be other instances of
1t, would be 1mposs1ble unless we apprehended it as more than
merely individual. The same thing follows from our power of
imagining different instances of 1t.
§ 146. Thought 1s often said to be the faculty through which
we apprehend universals, as distinguished from perception, the
faculty of apprehending individuals. Tlus distinction cauF.es
difficulty when 1t is recognized that perception so1.1ehow involves
a universal. Thought 1s a term applied widely to any activity
of the mind concerned with apprehension, which is not experience
or perception, and so may mcludc even memory and 1magina•
tion. It may on the other hand be restricted to reasoning and
the apprehension of universals, toge-ther with the questioning
state of consciousness. 2 There 1s a danger of overstatement in
two contrary ways. If, distinguishing thought from perception,
we make the universal the obJect of thought alone, perception
appears as 1f it were of the mere particular without any universal.
If, 90 the other hand, recognizing that the universal is somehow
apprehended in perception, we still make the universal the object
of thought alone, there is danger either of confusing the distinc•
• Pp 4S, note a, and 355, note 1. • Part I, ch. 2,
~~I Z
338 STATEMENT, THINKING, AND APPREHENSION
tion of thought and perception, a distinction which we feel must
somehow be maintained ; or else of putting thought side by
side with perception in the act of perception, without any real
unity between them, so that perception still remains distin•
guished as bemg in no sense apprehension of the universal. The
subject is difficult, but the difficulty does not concern the dis•
tinction of umversal and particular , 1t concerns the nature of
our apprehension of this distmction. The total being of the
universal is not its unity and identity m particulars, but the
whole of the particulars as the particularization of this unity.
When thought 1s said to be the faculty of apprehending uni-
versals, the umversal 1s meant as a umty m distinction from the
plurality of its particulars. Now the being of the universal
extends beyond the bemg it has m the particulars perceived m
experience, and 1t 1s this fact which is at the bottom of the distinc-
tion of thought, as of the umversal beyond the particulars, from
perception, as of the particular; whether this d1stmction 1s rightly
conceived or not 1s another matter It 1s felt of course that thought
as of such a universal takes us beyond the present perception.
Consider the apprehens10n of the particular in perception.
Though the particular 1s not the whole bemg of the umversal,
yet 1f we apprehend 1t as a particular of a universal it seems
necessarily to follow that we arc m some sense apprehending
the universal ; that is to say, we cannot apprehend the narrower
bemg in the particular as narrower without bemg somehow aware
of, mother words apprehending, the wider bcmg. Now it may
be said that there 1s an important sense m which we are not
apprehending the wider bemg, because it may be that we are
not apprehendmg any of the other particulars m which that
wider bemg consists. It may also be said that we do not appre•
hend, in the apprehension of the particular, the fact of the bemg
of the universal m other particulars ; its bemg in the other
particulars is not in this particular. (Otherwise this particular
itself would be m all the other particulars.) We may help our•
selves by a simple analogy : we perceive a particular surface,
for instance we see it; but we do not perceive (see) the volume
within or behind 1t. Nevertheless we cannot apprehend the
surface as surface except as the surface of a volume ; thus we
do apprehend that the volume is there and has an existence
 The Universal 339
which we are not perceiving. In this sense we do apprehend the
volume. Now similarly when we perceive the particular (appre•
hend it in experience}, we can apprehend that it belongs to
a universal of a wider nature without any further apprehension
of that nature, except as something that 1s here particularized
and can be particularized elsewhere ; in short, as something
which has a nature beyond what we are perce1vmg. That is
what we do when we apprehend the particular as particular.
Now such an apprehension is an apprehension of the universal.
If we consider those apprehensions which are called • judge-
ments ' of perception and are expressed m language (' that stone
is blue', • that silver object is a pencil case'), 1t 1s not difficult
to see that here we do apprehend the universal m the way just
described On the other hand, in what 1s called abstract thmking
(excludmg symbolic thmkmg for the present), though we appear
to have our attention merely directed to the unity and identity
of the universal as agamst the particulars, we can only do this
m the apprehension of particulars (experience or memory).
Moreover, if we- merely tlunk of the umvcr:,al as something
common to these and to other particulars not perceived (which
L.orresponds perhaps fairly to the phrase • mere abstraction'),
such apprehension of 1t 1s actually not of anythmg more of 1t
than "c have already apprehended 111 the perceptions. This
' thought ' therefore 1s not only 1mposc;1blc without perception,
but has no more ' content ' than we find m the percepti,,n ;
the difference hes 111 our attcnt10n to this side vf the universal
and our mtercst 111 11, as the obJcct, for mstance, of some mqu1ry
to be undertaken.
There 1s however another case m which there is a greater
d1ffercncc. If we arc able to apprehend, besides the nature of
the particular instance and also the fact that 1t 1s a particular
of the umversal, somethmg more which belong& to the universal
however particularized; for mstance, 1f we sec not only that
this triangle is a particular of tnangulanty and that its exterior
angle 1s greater than either of the mter1or opposite angles, but
that thu, property must be c;o m any other particular triangle ;
then this ' thought ' 1s of somcthmg beyond the particular
experience, and m practice we unhes1tatmgly call that' thinking'.
Yet the result of this apparent pas&mg beyond the particular
Z2
340 STATEMENT, THINKING, AND APPREHENSION
experienced has been to exaggerate this distinction and to lead
to quite false theories of such an investigation as the above.
The result has been to reason as though mere thinking effected
the distinction, and as though thinking could be conducted
without the apprehension of particulars and did not depend on
it. Now even thic; tlunking, as we <,hall c;ee hereafter, is quite
impossible without the apprchcns1on of the particular and we
shall find that tins 1c; really the secret of demonstration.
§ 147. Consider now the beg111111ngs of apprehension, which
differ from these' judgements' expressed m language. Language
is umvcrsal, 'cstal>hsh~d names', as Aristotle says," arc' general',
and to apply a term to anything imphes that somcthmg of the
same sort has been recognized already, and so the tlung 1s con-
ce1ved as the particular of a umvcrsal But m the beginnings
of apprehension we are noticing things we haven't noticed before
and thus the,;e begmmngs are necessarily without language. We
have here not to try to explam how we begin to notice or
apprehend anything, but to determme what the act of appre•
hendmg anything must mYolve to be an act of apprehension at
all. It will accord with the use of language to employ the word
' notice ' for the most primitive form of apprehension {babies
are said to be 'begmnmg to notice'), meaning by this the
apprehension not of mere difference between the obJect noticed
and others, but of Just enough of its positive quality and of
theirs to make the perception of difference possible. Noticing,
however elementary, necessitates d1i,tmct1on of what 1s noticed
from something else, which 1i, therefore also so far notlccd. 1
Now, 1£ we notice a particular A1 , we cannot apprehend it as
a mere ind1v1dual, but as having some d1stmct1ve quality A,
this quality being md1viduahzed m A1 To notice the quality
as distinct we must d1stingu1sh 1t from some other quality; thus
we need to notice at least two md1v1duals, A1 and B1, where
A and B are distinct. W c arc ex hypothest concerned with the
case where A and B are noticed or apprehended for the first
time. If A1 is a particular colour, B1 is not necessarily another
• Cf. Anstotle, An. Po. b. I 9 (passim), esp. aranos 'YGfl TUW ,l&wf>l,,-, lv&s, •parrov
iO'TClTCll (ms &v TG d1ffpij n-rfi IRII ,-cl d6Aov,
plv '" tji ,f,u}{j lfO.loA.ov • •• n>Jp , .. TOWOIS
Joo" J.,,
[• Td ,.,;,,.,,,,. ,ro,,,d rafl',v, Mt!laplt w40• r 1 (VIZ, are common to each of
a number of individuals).]
 The Universal
colour j B1 might be for instance a sound. Such distinction
classuies the distinguished and involves the apprehension of
a. universal, as the bemg common to them, even if only the
universal of something-in-general. Each as distinguished is
apprehended as a bemg.• Now this is not the universal of the
quality Aness m A1, not for instance of redness m the colour
noticed, the universal, that is, which Aristotle has m mind when
he says ' perception is of the universal'. He refers to the quality
which makes the md1vidual ' so and so ', and that is not mere
being We recognize difference and identity of bemg m A1 and
Bi, because we apprehend in them two diffcrent forms of bemg.
But, m thus not1cmg A m A1 for the first time, we have ex
hypothesi not more than one mstance of A before us m apprchen·
sion. What we arc apprehending as A is mdeed a universal
quality m a part1cul.i.r and so far Aristotle 1s nght, but it docs
not follow that we apprehend 1t as such, that we have, 111 his
termmology, 'perception of the universal qua umvcrsal '.
Observe that a umversal 1s not merely 1dent1cal m its part!·
culars but 1s somethuig definite identical 111 its particulars; and
this implies the distmctlon of the somctlung definite from its
bcmg identical 111 the part1Lulars, and it 1s this somcthmg dcfimtc
wluch d1stmgu1shes 1t from other umversals What tlus somc-
thmg defimtc is can only be understood m examples, for mstance
red m the particulars or mdiv1duals of redness (wluch by the
Wd.Y are not red tlungs, but mdividual red colours), animalness
in the particular ammals. We sec m the rxamples that tlus
something defimte is not the same as univcrsahty or parti-
cularity. When regarded ,ts partieulanzed m the md1viduals this
sometlung dcfimtc 1s often called their quality (their distmct1vc
quality), plant, for mstance, and annual, 111 this plant and tlus
ammal. They are not mere mdividuals but we may say they
arc d1stmgu1shed by their quality as plant or animal In
ordmary language tlus something definite is represented by the
mere adJectivc, as red or c1,mmal, opposed to redness or ammal•
ness on the one hand and on the other to this red colour or
a red colour and so on. In our symbolism, 1t is A as opposed
[• Cf. • 1llud quod pnmo cad1t m apprehens1onc est ens, cu1us 1ntellectus
includltur m omnibus quaecunque quis apprehendlt '. St. Thomas Aq.,
Summa Th 1-2, xc1v .i.d ::.ecu11tluu1,]
342 STATEMENT, THINKING, AND APPREHENSION
to Aness and to A1. Now though in the case of a given universal
there is a word corresponding to this something definite (as the
adjective red), there 1s no general term for it for all universals,
that is, there is no technical term. We must remember too
that universal itself, as we employ 1t in philosophy, is really
a technical term and does not belong to ordinary language. As
a term for the something defimte, 'quahty of the umversal'
suggests itself, smce this, as 111 md1v1duals, 1s their quality, and
distmgu1shcs mdiv1duals of one class from those of another.
But E>uch a use of quality would only he analogical; strictly
quJ.hty 1mphcs reference to the ind1v1dual Perhaps we may say
simply 'characteristic bemg of the universal', with the express
undcrstandmg that 1t 1s a techmcal term, only to be understood
by realizmg what corresponds to 1t m the examples given.
Similarly we can give no account of number m general, except
through the mstanccs of 1t m wluch we realize what 1t 1s : and
so also of space and timc. 1
In the case before us, we have apprehended A1 and B1 as
particulars of a universal, but not as respectively particulars of
Aness and Bncss. We cannot say that because we have appre·
hended A1, so far, as a particular, we have therefore apprehended
it as a particular of Ancss (wluch of course would mvolvc the
apprehension of Aness as a universal). The truth seems to be
that while A and B, or more accurately the apprehension of the
difference of A and B (when they thus appear as different
qualities 111 A1 and B1), have mdccd enabled us to d1stmgmsh
A1 and B 1 as p.irt1cular bcmgs and so to rccogmze for mE>tance
the particularity of A1, we have not so far made any d1stmction
between A and the particularity of A1 • ThuE> v.c should appre-
hend A (the ' characteristic bcmg' of the umvcrs.i.l m question)
in 1ts difference from B, but not apprehend 1t as havmg an
1 Perhaps • mtnns1c being ' v.ould suggest more readliy what 18 meant, but
there 111 danger lest some more mysterious meaning should be assigned to
mtnns1c. The absence of a general term from ordinary language and the
nt'Cess1ty of inventing a technical expression 1s m itself no difficulty Not
only 15 umversa.l a techmca.l term but so 1s quahty , 1t 1s not a necesS1ty for
ordinary speech It 1s s1gmficant that adJect1ves hke red get to be used as
substantives The painter says he puts red mto his picture ; the mathe-
matician studies the circle not circularity, in ordinary language, and yet he
doesn't mean the md1v1dual circles On the other hand, the adJective hot 18
not used for heat.
 The Universal 343
existence beyond A1, that is not as a universal in particulars ;
nor, though we apprehend A1 as a particular, should we appre•
bend A1 as a particularization of A. In short, what we apprehend
is a particularized universal of ' character1st1c being ' A, but
what we apprehend m it is this 'character1st1c being', neither
as universal nor as particularized It 1s this mdeed which is
what Aristotle calls ' first or primary m the soul ',1 but it 1s not
there as universal. Nevertheless 1t is a universal; 1t 1s some-
thing identical m particulars and that 1t 1s which makes it
possible to identify 1t m another mstance. When we have done
this we have recognized the universal as a universal. As already
mamtained, 2 1f we had apprehended A, in the particular form
A1 , merely as confined to the particular, 1t would be quite 1mpos·
s1ble that we should recognize A as m another particular (which
1s what we actually do), Just as 1t 1s 1mposs1ble m the case of
two md1vidual books before me to recognize or 1dent1fy the one
individual book as in the other.
Now, 1f this analysis 1s correct, even m our developed con·
sc1ousness and when we possess language, 1t would seem quite
possible to be sometimes apprehending the characteristic bemg
of a universal without our consciousness bemg fully awake, so
that we could not be said to be apprehending either its uni-
versality or its bemg particularized. Thus I may be enJoymg
a beautiful colour, without reflecting at all whether I have seen
1t before or not, and mdced I may not even be thmkmg of any
being of it beyond the time of perception at which I am per·
ceivmg 1t ; I may not even be reflecting about 1t as m an
individual time of perception. 1£ this 1s so, the exbtcncc of this
fact of our consciousness, the nature of which is not clearly
apprehended, 1s probably what mclmes us to think of perception
as of the mere 10d1v1dual (because here at least there would be
no consciousness of the universal) and causes us, even when we
realize that the umvcrsal u, somehow apprehended m perception,
to have a m1sgivmg still and to think that this perhaps does not
belong to the actual perception as such, but only to an after•
reflection upon 1t. Now in this what we are really doing is to
confuse the definiteness of the being of the universal, 1ts charac•
1 ,rpwT011111 Tji1f,11xfi, p 340, note r.
I § 145.
344 STATEMENT, THINKING, AND APPREHENSION
teristic being (A), with individuality or particularity. Finally,
if this is a true presentation of the facts, it shows us not only
how important it is to attend to the scruples of the ordinary
consciousness but also that those who hold that perception is
always of a umversal and those who hold that as perception it
is of the mere particular, are both m error ; the former because
not everytlung we apprehend m perception, though it 1s a
particular of a universal, 1s apprehended either as particul.ir or
as umver&ctl, and the lcttter because nothmg is apprehended as
mere particular.
§ 148.• In the present section I wish to strike at the root of
all the fallacies about Umversals. I shctll attempt this by a more
precise consideration of the real meaning of the unity of the
universal, of what 1s meant by speaking of it as something
common to particulars and 1denbcal 10 them. In our ordinary
hfe and in scientific thinkmg, the facts which, owmg to their
apparent difficulty, have been seriously disputed and very fre•
quently denied altogether by philosophers are always, though
generally unconsciously, assumed without any hesitancy what•
ever to be real. We always, in the ordmary, treat the umversal
as somethmg quite obJect1ve, as an absolute umty. Thus there
is one and only one universal of ' circularity ' and we only
consider one 10 mathematics Such umversals are moreover
treated as something entirely obJect1ve. The umversal is no
mere thought of ours but a real umty in obJects, and further
sometlung identical 10 the particulars, which identity cannot be
done away by substltutmg the term similar for same. No
difficulty whatever 1s made about the identity of the universal
m particulars, on the contrary, this 1s always treated, though
of courbe not explicitly and philosophically, as a quite obvious
fact. In short the umty, reahty, and 1dent1ty of the universal
in the particulars 1s presupposed in every sentence that we
utter.
It is indeed astonishing that m our philosophic criticism we
should overlook so startling a fact But yet 1t is and has been
[• § 148 was, hke the two succeeding sections, a separate paper left by
Wllson m typescnpt HIS view of the universal seems to have been suggested
by Green, from whom also much of lus cnt1C1Sm of Anstotle and the sylloguim
.i.ppears to have started (bCe Green, Wo,-ks, m, p. 56) J
 The Universal 345
universally overlooked. Plato indeed realized a very essential
side of it, namely that thought and language would be quite
impossible without assummg this unity, identity, and reality of
the • Idea '. Yet even he hardly realized the great point that
the matter is without any suspicion of difficulty, m all our
ordinary life and conversation. This stands m strange contrast
to the hopeless puzzle which metaphysics has made of the
subject. Moreover, any theory of the obJcctivc existence of
universals is at the present day very commonly, mdeed nearly
always, looked upon as the quite peculiar product of a very
subtle metaphysic, while the mere and sheer truth 1s that the
reality of the umversal 1s the unconscious assumption both of
all our ordmary speech and thinkmg and of sc1cntdic thinkmg
also. What I have therefore come to feel 1s that the solution
of this celebrated problem ought to be specially easy mstead of
specially difficult, 1f we only go about 1t the nght way ; that
our d1fficult1es are entirely of our own makmg The distmct1on
of universal and particular is indeed absolutely necessary for
the explanation of anything whatever. Any explanation of
anything presupposes this d1stmct1on. If then the d1stmction
is not explicable (m the sense, that 1s, of an explicable puzzle)
since 1t 1s itself the key and prcsuppos1t1on of all explanation,
every explanation would fail with the failure of this key. Once
more then the conviction 1s borne m upon me that the puzzle
is somehow wholly artificial and that we may have the courage
to l,eheve we can entirely solve 1t.
Since the ordmary language 1s quite clear and 1s always
understood, the clue hes probably m the tr.mslation of all
the philosophical techmcaht1es into the normal cl.nd ordmary
language about the matter to which they relate. Consider the
simple example of triangle and tnangulanty. We say that this
triangle is a particular of triangulanty, that it has the quality
of triangulanty and that the tnangularity is common to all
triangles. Now what docs the abstract noun triangulanty mean ;
or, at all events, what docs the ordinary language, which corre•
sponds to it and mto which 1t would have to be interpreted,
mean ? If we ask what tnangulanty 1s, the answer would
probably be that it is having three sides. It refers in any case
to the particular sides of a particular figure, for 1t is only
346 STATEMENT, THINKING, AND APPREHENSION
particular sides which can be three in number, and these must
belong to a particular figure. Thus then, more accurately,
' tr1angularity' means 'having three particular sides '. Now,
since only particular figures can have three particular sides, the
still more accurate mterpretat1on of the formula would be
' a particular figure's havmg three particular sides '. If we
think of a particular triangle A, the tnangularity of A is ' its
having its particular sides three in number '. This property 1t
obviously has to itself. It 11:, not something common to it and
another triangle; 1£ we tried to make 1t so, we &hould be identi•
fymg the sides of one particular triangle with those of another.
We clearly must have something more general than 'triangu•
lanty ' m the sense of • tr1angulanty of this figure '. But now
tnangulanty as such can only be tnangulanty of particular
figures, for 'figure m general' has no sides at all. We have to
find the more general acceptance of triangulanty, though 1t
cannot possibly fail to refer to parllcular figures and particular
Imes. All then that remains 1s to d1stmgu1sh tnangularity m
particular from tnangulanty m general by making the one • the
three-sidedness of this particular triangle ', and the other ' the
three-sidedness of any particular triangle '.
Let us now substitute this meaning of' triangularity m general'
for tnangularity m general 111 the sentences wluch express our
puzzle or difficulty about 1t. ln mak111g the substitution v.e
must be a!> accurate as 111 a mathematical formula and must
not alter any part of the ltnguist1c expression. Consider now
the common assertion that the umvcrsal, which 1s the name
given to triangulanty, 1s that wluch the part1rnlar triangles have
m common This will now have to become 'the three-sidedness
of any parl1cular tn.mglc 1s that \\-h1ch 1s common to all the
particular triangle& '. This is clearly nonsense. We seem to
arrive therefore at a very strikmg result, namely thdt • tr1angu•
lanty' docs not mean what 1s common to all particular triangles.
If tlus is true, the consequences arc momentous, and we must
for safety mquire how that which 1s common to the particular
triangles is really to be expressed m language. The ordinary
language feels no difficulty and 1t would be replied without any
misg1vmg that what is • common ' to the particular triangles is
' that each of them has three sides '. Another answer might be
 The Universal 347
that what is common to ,particular triangles is the number
of their sides, three, the same for all. Consider the first expres-
sion, that each of them has three sides. If ' common ' be taken
in its ordinary and literal acceptance, it would mean something
which was m the nature of each particula.r But now this could
not be that each of them has three sides, for the resulting
sentence 1s agam nonsense. The fact is that m such a sentence
the word ' common ' has not its literal s1gmfi.cat1on. What it
does mean is determmed by the words associated with it.
Havmg somethmg m common m tlus c.ise 1s each of the particular
triangles having the same number of sides. This 1s what common
really means, and 1t 1s the confusion of this true meaning, per•
fectly mtelhg1ble m itself, with the more literal meaning which
is largely responsible for the philosopher's difficulty. Consider
now the second expression, that what 1s common to all the
triangles is the number of their sides, three, which 1s the same
for all. Observe here how the word ' same ' 1s absolutely neces•
sary, and that 'similar' would be quite wrong. As we before
mqu1red what was the real meaning of 'common', so \\C now
have to ask how the word ' same ' 1s m,c<l m this special context.
The answer aga111 1s qmtc clear. What 1s the ' same ' 1s the
number 3, that lb to say .tgam 3-ncss, of which the number of
the sides m each case ,,, a particular So that it 1s this absolutely
well understood d1stmct1011 of umversal and particular which
cxplams the mcamng of the words ' common ' and ' same ',
mstead of these latter \\ ord,, bcmg any t>xplanJ.t10n of the
distmctlon of umversal and particular. Now substitute for
' what 1s common to the particular triangles ' the expression
we have Just found ' the u111versal of wluch the triangles are
particulars ' ; we then gel ' tnangulanty 1s the universal of which
the triangles arc particulars'. This at first looks correct enough,
but, 1£ we substitute as before for ' tnangulanl y ', we get ' any
figure's havmg three sides ii, the umvcrsal of which the triangles
are particulars'. This 1s agam nonsense We thus seem con-
ducted again, strange as 1t may appear, to the conclusion that
we cannot correctly say ' tnangulanty ' 1s a umversal. Yet,
after all, this is the conclusion which will be md1cated hereafter,
when we shall show that 1£ we say ' tnangulanty 1s a universal',
we get not only mto the folly of the ' member of itself ', but
348 STATEMENT, THINKING, AND APPREHENSION
also into absolute contradiction in terms. The statement that
' triangularity is a universal ' is thus seen to be a linguistic
impossibility, which is disguised by the grammatical form of
the abstract noun which is supposed to represent the universal.
If this reasoning is correct, we have arrived at a decisive
settlement of the puzzles about the universal, whether the
ancient and rational ones, which arc still unsettled in serious
modern metaphysic, or the puenhties of certam paradoxical
recent authors. Moreover, our foreboding that the true solution
must be sometlung sunple turns out to have been fully Justified.
It wdl be easy to show how many very sdly verbal fallacies are
all of them easily and dcc1S1vely met from the position we have
arrived at. Tlus proof that we cannot say ' Aness is a umversal '
is indeed analogous to the proof which will be given that uni-
versals cannot be counted. In each case the proposition arrived
at destroys the whole mass of fallacy corresponding to it, namely
the logical and metaphysical absurdities of things represented
as members of themselves and the iantastic mathematical non-
sense which has 111 recent years been developed about number.
It will be found for instance that such difficulties as arise from
treating 'particularity' as a umversal di,;appear at once, as well
as the nom,ense c1bout the wonderful ' cJass of classes '. I have
not c1.llowed my,;elf, 111 this mqutry, to be diverted from the
facts by theoretical difficulties made about them. I have simply
concentrated attention on the reahty of the facts about the
obJectiv1ty, umty, and 1dent1ty of the umversal and shown that
whether difficult to understand or not, they must be fully and
frankly rccogmzed Tlus 1s so tar m the spmt of Plato m his
Parmemde~ Further, I 1:,cem to have discovered that the true
source of our metaphysical d1fficult1es hes m the attempt,
a mistaken attempt too frequent in philosophy, to explain the
nature of the umversal m terms of sometlung other than itself.
In fact the relation of the umversal to the particular is some·
thing sui generis, presupposed m any explanation of anything.
The nature of the umversal therefore necessartly and perpetually
eludes any attempt to explam itself. The recognition of this
enables one to elucidate the whole puzzle of the Parmenides
of Plato.
§ 149. The ordinary forms of language before the appearance
 The Universal 349
of reflective and abstracting thought were intended to express
the nature of particulars and are still in everyday life employed
for this purpose. Thus the nominative case to the verb was
originally (and very commonly still is) a word designating an
individual substance. 1 A misunderstanding of the original
meaning of grammatical forms ha~ produced the fantastic
puzzles, as one may be allowed to call them, which, mistaken
for metaphysics, are fallacious thinking, mere verbal fallacies.
The ground of the mistake Jres in applying to the abstract
universal, forms which are proper to the particular, as I shall
now endeavour to illustrate by examples. In the normal form
of speech, where the sub;ect 1s an ind1v1dual substance, we have
such expressions as ' this rs a hyacinth' (this is an A), or 'this
flower rs a hyacinth ', the general formula for which 1s X 1 1s
an A, where the nominative is eqmvalent to A1, that is to
a particular A. Now m this sentence what the nommativc case
stands for 1s a particular m the strict sense of the word, and
a particular substance It is a particular of the universal repre•
sented by ' Aness '. Aness 1s not a mere universal, it has
a special quality or character corresponding to the symbol A.
The particular sub;ect, corresponding to the nominative case to
the verb, 1s said to have the quality Aness. In this form of
sentence, with the indefinite article followmg the verb, the
quality or character covers the whole nature of the substance
A1 ; that is to say there 1s nothing m its nature which is not
comprised m its havmg the quality Ane~s 2 Observe that this
relation must obtain between an individual and a universal, if
the ind1v1dual 1s a true particular of the given universal. For
instance ' this flower ' is a true particular of ' hyacinthness ',
whereas, in the sentence ' this flower ts blue ', ' this flower ' is
not a particular of ' blueness ', since blueness does not cover its
whole nature : the true particular of blueness is the colour of
the flower, not the flower itself.
Suppose now that m the above form of sentence we substitute
an abstract noun for the name of the particular substance (the
nominative case to the verb) so that the sentence 1s of the form
' U1 is an X ', can we assume the same relation as before between
the nominative case, the universal in question (U 1), and a uni•
1 Part II, ch, 8 and § 79
350 STATEMENT, THINKING, AND APPREHENSION
versal corresponding to the attributive (say Xness) ? Consider
a particular example. ' Circularity ', as common to all circles,
is their true universal. Now take the form ' circularity is
a universal '. Can we say, on the analogy of the sentence in
which the nominative case 1s a particular substance, that corre-
sponding to ' a universal' m the attributive part of the sentence
there is a univC'rsal to be called ' universalness ', and that the
nominative cac;e, which here denotes a universal, is a particular,
or a particularization of ' universalness ' ?
In the fallacious thinkmg referred to, the sentence in which
the nominative case 1s a universal is treated as 1£ it were in all
respects hke that in which the nominative case is an individual
substance. There 1s no consciousness that any distinction has
to be made and no critical cxaminat10n (such as is absolutely
necessary) of the meaning of the grammatical forms. The con-
sequence 1s to make one universal, c g. • circularity ', a parti-
cularization of another umversal namely ' universalness ' ; that
is, one universal 1s made to be a µarticular of another ; and
indeed this other, univl.'rsalncsc;, would have to be such a uni-
versal that its particular,; could only Le universals In the
ordinary and correct use of umvcrsal, different universals may
indeed have a umvcrsal common to them, but they are never
particulars of tlus common universal, only differentiations of 1t.
For instance • come section ' 1s common to ' circle ' and ' para-
bola', but the latter arc its d1ffcrentiations and not its parti-
culars. Thus this new use of parbcula.r1zation of a universal
ought, prima fac1e, to excite our suspic10n.
Now observe what ha.ppens •Universalness' 1s made a uni-
versal. But, by hypothesis, ' umversalncss' 1s the universal of
which all universals are particulars Thus ' universalness ' as
a universal must be a particular of • universalness ', that is
a particular of 1tsclf. Tlus is obv10usly absurd, and the proper
inference from it 1s that the treatment of the form of statement
with the universal for its nommativc must be a fallacy. The
next step then ought to be to try to prove otherwise that it 1s
a fallacy, and to discover the ground of the mistake. But instead
there arc modern writers foolish enough to take this obvious
contradiction for a piece of subtle metaphysic.
The contrad1ction arrived at is really decisive and quite
 The Universal 351
enough to prove the fallacy of the whole procedure : but we
shall now endeavour to show otherwise that 1t is a fallacy, and
upon grounds which even these writers themselves cannot but
admit, that is by proving that it involves a contradiction in
terms.
In the normal form of sentence now in question, ' This (i e. A1)
is an A' or 'X1 (1 e. A1) is A', that which is designated by the
nominative case possesses the quality of Aness and its whole
nature is comprised m Aness. It is said, in ordinary phrase,
to be an instance of Aness. If then' U1 is a universal' is treated
in the same way, e g 'c1rcularity is a universal', the whole
nature of circularity has to be comprised in universalness, that
1s we should have to say that its nature has the quality of
universalness. This bemg so, consider what happens m the case
of ' this figure 1s a circle ' ' This figure ' must have circularity
as its quality and must be an instance of circularity. But, by
hypothesis, universality belongs to the nature of circularity, for
mdeed all the nature of circularity has to be comprised in
' universalness '. Thus 1£ tlus figure, that 1s this c1rcle, has the
quality of c1rculanty, 1t has to have the quahty of universalness
and to be an mstance of universalness. That 1s to say ' this
circle' must be a universal, a sufficiently absurd contradiction
of course.
This shows that a sentence in which an abstract noun 1s the
nominative cannot be treated hke the normal sentence of which
the nommativc case 1s a particular substance, and explains the
origin of the fallacy which we arc d1scussmg. It turns out that
universalness 1s not a true universal, and this 1s not surprising,
for a universal must have some definite quality, what I have
called 1 the' mtrms1c character' of the universal Universalness
has no such quality or intrinsic characteristic and so cannot be
a universal.
The fallacy then of treating universality as itself a universal
is the consequence of treating the universal as if 1t were a
particular. Every universal then 1s treated as a particular of
universality, not as a species or d1ffcrent1a of a universal but
only as a particular of it. This involves the hypothetical
existence of a universal, 1.e. universality, which is such that the
1 See p 342, foot-note 1.
352 STATEMENT, THINKING, AND APPREHENSION
only particulars of it are universals and of true particulars (or
individuals) it has none at all.
§ I 50. Difficulties have arisen in the singular development of
modern mathematical speculation about the nature of number
which appear to be due to a misunderstanding of the true
function of universals. The result has been the development of
a system of new conceptions of number which is a mere fantastic
chimera.
Consider what a mathematician would call • the series of
natural numbers' I, 2, 31 41 5; what are really the numbers,
for there are none others, notwithstandmg the fictions of the
theories just referred to. These numbers arc properly speaking
the 'universals of number'; for instance 3 stands for 3-ness,
as seen in any particular group of 3. Further, as universals
they cannot be added to one another , and this is probably the
mcanmg of Plato's unaddihlc numbers." Moreover, they are not
only not addiblc (2-ness plus 3-ness 1s not equal to 5-ness) but
they cannot be counted, that 1s to say we cannot count the
universals of number in the above series and say there are five
of them, if we stop at the umversal 3.
Some mathematicians, and probably most people who think
unguardedly, treat these universals of number Just as if they
were particulars. This mdced 1<; the general form of most of
the fallacies which belong to the kind of pseudo-speculation we
arc discussing, the treatment, that is, of umvcrsals exactly as if
they were particulars. Let us then see what happens if we
attempt to count them. I, 2, 3, 4, 5, we should say 1s a group
of five universals, as such 1t is a particular of 5-ness. But the
number 5 m the series is itself 5-ness, thus 5-ncss together with
the precedmg umversals of number 1s a particular instance of
5-ness, and !,O a particular mstance of itself This 1s obviously
absurd. Lest 1t should be taken for metaphysics, as it probably
would be by those who mistake paradox for philosophy, we
must try to show othcn,ise that 1t 1s a fallacy by a reductio ad
absurdum m the first instance, and aftcrwards by explaining the
impossibility of counting the universals of number from the
nature of umversality itself.
[ 1 See Clasncal Revi,w, June 1904, vol. XVJ.il, No. 4, pp. 247-60, where
these points are developed J
 The Unive,sal 353
If the universals of number can be counted, they must have
something in common in virtue of which they are particulars
of the same universal. Now here, in the nature of the case, this
universal must be ' universal-of-numberness '. Consequently
since each particular of Aness has the quality Aness, the universal
of number must have the quality of ' universal-of-numberness '.
This being so, the particulars of this last number-universal must
have this quality of ' universal of numberness ' Thus they
must be themselves universals Take an instance. A partirular
group of 5 as being a particular of 5-ness, which is itself a parti-
cular of 'universal of numberness ', must itself be a universal,
namely a umversal of numberness. This 1s of course a flat
contradiction : a particular set of 5 cannot be a universal.
We must now have a direct argument based on the nature
of the universal. In order that umts should be c.-apable of being
counted they must not only be of the same sort but they must
be mutually exclusive. But now are the umversals of number
mutually exclusive? Consider 5-ness. Every instance of 5-ness
is the sum of a particular 2 and a particular 3, i.e. every parti-
cular of 5-ness is the sum of a particular of 2-ness and a particular
of 3-ness. Evidently then the nature of 5-ness is mseparable
from the nature of 2-ness and the nature of 3-ness. It is impos•
sible therefore to say that, for example, 5-neo;s and 3-ness are
mutually exclusive. If wc arc now asked how the nature of
5-ness, if not excluding the nature of 3-ness, does involve this
latter nature, or what it has m common with this latter nature,
our answer must be simply the fact above stated that every
particular of 5-ness involves a particular of 3-ness.
The universals of number then cannot be counted, for they
fail to satisfy the conditions of countablcness. The form of
counting is quite inapplicable to them. Thus, \\ hen we see
what counting means, we see that it cannot apply to the
universals of number . we might as VI ell try to persuade them
as to count them. The form of persuasion cannot apply to such
objects as the universals of number and so they cannot be
persuaded.

1775•1 Aa
 XVI•
CLASSIFICATION
§ I 5I. THE recognition of the universal has given rise to
important questions "luch conrcrn metaphysic and psychology
ct<, well as logic The contrast between the universal and the
particular presents certain difficulties to thought. Real existence
seems all particular, although the universal is necessary even to
a smgle statement or apprehension Hence there is a tendency
to represent the umversal as a subjective product, as a mere
idea. T}u., leads to a further difficulty because 1t is found that
the universal as such cannot be presented to imagination any
more than to perception. Thus ongmated the well-known con•
troversy between conccptuah-,m, realism, and nommahsm, a sub-
ject which belongs to metaphysics and psychology rather than
to logic To tl11c;, though properly metaphysical, we shall
return , at present we shall concern ourselves only with classi-
fication and the cognate subJect defimtion, recognized parts of
the traditional log1r.
§ I 52. Clasc;1ficahon 1s an activity of the human mind m
knowing obJcctc; ; an act1v1ty which logic does not create but
can only reflect upon. Formal logic has in tlus branch of the
study made the same mistake as m the case of the syllogism ;
1t has tned to gmde the mmd by warning 1t against fallacies
which 11 could not possibly commit. Thus 1t has enunciated
the grave dictum that 111 dividing a class we must not pass from
one fundamentum divisioms to another, but must divide by one
surh principle only. The reflectivt' consc10usncss which is logic
assmulates itself in fact too closely to the scientific consciousness
which 1s not reflective but directs itself to obJccts. As that
consciousness lays down rules, so logic tends mistakenly to
[• See notes 011 the 'IOurccs • fh1~ requires to be entirely rewntten and
the word roncept everywhere taken out ' MS ,iott He had intended to
embody also H 46o-76 in the course The foot-note references are D10Stly
.upplied conJecturally )
 Classification 355
regard itself as testing the value of the rules which it examines
and even as producing correctness in them. Now what it is
examining are just the rules of thinking, rules therefore that
cannot be broken. Thought cannot justify its own rules or even
criticize them.
§ 153. The study of classification comes quite at the beginnings
of logic, implicitly in Plato,• explicitly in Aristotle.b This is
natural ; for the most elementary act of knowledge must involve
classification, and reflection upon that knowledge is logic Our
knowledge begins, doubtless, by attention to particulars m
experience and 1s stimulated by practical needs, not originally
by a desire for knowledge for its own sake. Mankmd, faced by
practical needs, seeks practical rules so as to know how to deal
with nature ; frames principles to go by m action. Such a rule
or principle is a universal We seek something which we can
count on at all times and m all the variety of the different
cases, and this is a umty by contrast with such variety. Even
if the knowledge be of a particular object, what is sought is
knowledge of the behaviour of that obJrct not at one time only
but at all times, and so 1s universal as compared with the different
times in which the behaviour ic, manifested The interest of
knowledge, whether theoretical or practical, will go further and
ask for a reason, and a reason or an explanation 1s, m the nature
of the case, umversal. Indeed even the attempt to represent
ourselves as knowmg one mdividual by itself inevitably involves
the universal. To know an obJect as something defimte, it must
be distinguished from other objects and is tl1us necessarily
related to them, must have something m common with them.
To distinguish 1s also to umfy. The characteristics we assign
to the individual have also a universal character as we recognize
something in it which might be applicable to other individuals. 1
1 Cf §§ 141 and 145 on thought as always umversabzmg 1ts ob3ect, and
Anstotle's generahzabon of the fact oi perception, d "14P _, l,nw ,t ala"1t11r ,-oij
'l"OCoiia• .nl pt} ,-oual .,.,,,o,, l'Ven 1f sense perception 1s of the ' so and so ' and not
of some particular ' t!J1s ' An Po 87b 28, cf 100a 16

[• Especially m the Phaedrus, Philebus, Sophist, and Polit1C1U Defimticm 11

considered expl1C1tly m the Meno
b e g. Tofncs, vi. 6; De Part An. i 2-4, Mdaph Z, 12. Dichotomy 11
examined in An. Po. 11. 13 and De Part. An. 1 ;: and J J
Aa2
356 STATEMENT, THINKING, AND APPREHENSION
f 154. This rudimentary classification involves a recognition
of different common elements, a variety of universals. Naturally
the same process of unification is adopted with these and
a community in them recognized. To such a new community,
not of individuals but of universal elements, logic has given the
technical name of genus, calling the more particularized uni•
versals, species. But such a genus, when other genera have
been formed with which 1t has community, becomes itself
a species as opposed to a class which includes it. Thus we get
the fam1har d1stmctio11 of genus and species, and, for the manner
in whirh a species differs from its genus and from other co·
ordinate species m that genus, the technical term differentia.
The question then arises how far the process can be carried
in both directions. The descent to lower and lower species
terminates in infimae species, the ascent terminates in what are
called sum.ma genera. Further, the fact that a given genus
includes a plurality of species suggests an inquiry as to the
complete division of a genus, and thus arises the logical theory
of Division. The question is also suggested whether the summa
genera can be exhaustively determmed ; this is the philosophical
investigation of the Categories. •
§ 155. A very common view of the relation of genus and
species seems to be fairly represented as follows. If two or more
individuals m which a complex of elements is distinguished have
a common element, the notion of that element is called a species
or class notion and the mdiv1duals are said to form a class or
species, as unified by the common element. If the common
element is itself complex and, when compared with similar com-
plexes, 1s found to have some element in common with them,
this new common element 1s called a generic notion, and the
species unified by it constitute together a genus, the differentia
in the case of each species consisting m the elements which it
has besides. the common or generic element.
Thus, if the species notions be AB and AC, A will be a genus,
B the ddferenha of the first species, AB, and C of the second,
AC. Again, a group of elements contained in the notion of
any universal is called the intension of the notion ; the aggre-
gate of individuals to which the universal belongs is called the
extension o{ the notion. From the point of view of the intension,
 Classification 357
the genus or rather the intension of the genus, appears to be
a part of the intension of the species. From the point of view
of extension, the aggregate of individuals corresponding to the
species is a part of the aggregate of individuals corresponding
to the genus, and thus the extension of the species is said to
be a part of the extension of the genus. Genus then in relation
to species 1s either a whole or a part accordmg as we look at the
extension or the intension. A consequence of this is the familiar
doctrine that the extension of a term vanes inversely as the
intension.
§ 156. But this very representation of the meaning of genus
and differentia makes these terms relative. There would not
necessarily be a fixed element in the species which makes the
genus, but one element or another would be genus or differentia
indifferently. Which was which would s1mply be determined
by the species with which we compared the given one; e.g. in
the species AB, A would be the generic element if we compared
AB with AC, but B 1f we compared AB with DB. This mode
of representation has the further tendency to make us suppose
that the elements of the species are a mere aggregate with no
order or precedence. Thus the notion A would appear to be
differentlated into the species AB b)' the addition to it of the
element B, when we are considering AB and AC ; while B appears
to be d1ffercnt1ated by the addition of the element A, when we
compare the species AB, DB. This same sort of mdifference of
the elements to each other affects also the division of the class.
As we shall see presently, the division of the class is effected
by the ddierentiation of the elements of the class notion. Sup•
posing then that, m the class notion AB, the differentiated
species of A are A1 , A2, A3, and of B, B1 , B2, B3 ; as A seems
inddl'erently combmed with B, so also 1t might seem that we
arrive at the determmation of the species of AB by any arbitrary
combination of a differentiated element of A and a differentiated
element of B. We shall endeavour to show that this is an
adequate account neither of classification nor of the relation of
genus to species. Instances of terms wh1ch might seem to
correspond to this view of classification might certainly be
found, for example a golden sphere. Here the elements golden
and spherical seem indifferent to each other, and which is taken
358 STATEMENT, THINKING, AND APPREHENSION
for genus and which for differentia appears immaterial. Even
here, however, we notice that 1t is not natural to represent
golden sphere as a species either of spherical object or of golden
object. We do not, in ordmary language, call it a kind of the
one or of the other.
I I 57. However that may be, there 1s another class of universals
which present a great contrast to this one. In golden sphere
we may ignore the sphericality and abstract the quality of the
material, or conversely. But now take, for instance, redness
and blueness, which we naturally call species of colour. If we
eliminate all that is meant by colour, nothmg whatever 1s left, or,
if we suppose some differentiating element left, it would have to
be something different from colour. Thus the difference between
red and blue would not be one of colour, whereas 1t is colour
in which they agree and colour in which they differ. We cannot
give verbal expression to the d1fferentia which constitutes the
species, except by using the spedes name itself, red or blue.
Consider next an expression where we have a name for the
d1fferenti.a. Compare, for instance, (plane) rectilinear figure with
figure on the one hand and tnanglc on the other. Here the
genus is figure and the d1ffcrent1a rectilinear 1s now distinguish•
able in thought and represented by a word m a manner not
possible for red and blue. But as blue stands for blue colour
the divergence is only apparent ; for here also, if we abstract
figure from rectilinear figure, we either leave nolhmg, since the
whole bcmg of rectilmear figure is comprised in figure, or, if you
say we leave rectilinear, this 1s unmtelligible save as a deter•
mmation of figure. Plane figure necessarily 1mphes boundary
and this agam necessarily 1mphes that the boundary 1s of
straight or curved Imes. Thus we cannot ehmmate rectdmear
as a separate umversal I l has no existence even for our
abstracting thought apart from figure. Agam, take the most
general notion in tlus department, namely, space. The sphere
is not a determination of the general notion of space by some•
thing else not space which makes 1t a sphere and not space in
general. On the contrary space cannot be differentiated by
anything else than what is spatial. Space is not a genus formmg
part of the being of the whole sphere : on the contl."ary the
whole being of the sphere is spatial. Again, odd and even
 Classifieation 359
numbers are said to be species of number, but we cannot
eliminate number from them and leave something non-numerical;
for odd is intelligible only as meaning odd number. The charac-
teristic of all these mstances is, first, that they do not allow the
d1tierentia to be treated as something different from the genus
and added on to it, and, secondly, that the relation of genus and
differentia is not a merely relative or reciprocal one ; the same
term cannot appear indifferently now as genus and now as
differentia.
§ 158. We have thus reached a very different idea of the
relation of genus to species ; one also which seems to have an
affinity to the natural use of the correspondmg words m languc1.ge .
. . . We must seek to dctermme more precisely the nature of this
relation. In the instances taken we cannot represent the genus
as a part of the whole specific conception. The genus mcludes
all there 1s m the species, there 1s nothing left over. The
d1fferentia cannot be separated from the genus as somcthmg
added on to 1t, 1t comes from within 1t. The species 1s a necm,·
sary development of the genus , even and odd are not outside
number but necessitated by its nature as number, lmc as hnc
must be straight or curved, 1t doe& not wart to receive this
determination from something outside itself. The genus in fact
ii, only an mcomplete abstract10n, when separated from the
species. We may say if we please that rt 1s a common element
m the species, but this formula 1s a dangerou& one because rt
suggests that the genus 1s m the species with something other
than itself. It ,s c1. common element, but lhen in the species
there are nothmg but special mamfestat1ons of the nature of
tlus common element itself. It 1!> more than any one of the
species, m the sense that any one of them 1s only c1. part of its
total reality, of the complete meamng of the genus. In tlus
sense we may say the genus is the species, and the species arc
the genus. The umversal is less than any one of the species only
when represented in our subJectrve and incomplete abstraction.
In reality the genus and species are entirely mscparablc. We
must, however, avoid the mistake of 1dcntrfymg on this account
the genus with the mere sum or aggregate of its species, or with
any one of them. That would be to omit their community.
The genus is rather the umty of which they are the necessary
36o STATEMENT, THINKING, AND APPREHENSION
manifold expressions, a unity which necessitates its own plurality
from within.
I I 59. We may help ourselves to grasp the relation intended,
by such phrases as that the species is a realization of the genus,
and that the genus is potentially the species. This is one view
taken by Aristotle m the Metaphysics,i.. where the genus is
represented as the matter or potentiality of its species. This
thought requires to be earned further, to the relation of the
species to the ind1v1du.1I The ind1v1du,il is the fully realized
universal, the reahzat10n of the bpec1es, as the species 1s of the
genus. As the umversal as genus demands of itself its realization
in the species, so 1t 1s the nature of the species to demand of
itself reahzation m the ind1v1duals The totahty of the indi-
viduals 1s the rcahzat1on of the species and tlus 1s the answer
to the ancient difficulty, winch 1s ..tlso a modern problem, as to
the relation of the one universal and the many mdtviduals.
Tll!S formula of potent1.i.l .1nd .ictual (real) may stand 1f we
remember to abstr.tet from 1t any 11nphcc1tion of time as such.
The umversal or genus cannot cxH,t except as realized in the
species but 1t 1s not somcthmg prior to them m time, itself an
actuality contauung them only potentially. That is no doubt
the 1mphc,it10n in common speech of these words, as when we
bay the acorn 1s potentially the oak But the umvcrsal is not
what physical potcnttahtics arc, an actualizing potentiality in
relation to somcthmg future; actual m itself but contammg the
species only potentially. It has no actuality except m the
species It is true that the species 111 which the genus is realized
may be temporal and may appear m a time order after a previous
reahzcition of the umvers.i.l But tins docs not make the umversal
prior to such later reah1.at1ons, but only makes one realization
of 1t prior to another Shortly, the relation of time subsists
between the rcahzcit1ons of the umversal and ts not a relation
mto which the umversal itself enters at all m contrast with its
species.
It is perhaps advisabfo to choose some d1stmctive word and

[• ti o~i, n) -,,.,or 1.lirAws "~ l11T1 mpa T4 tlir .,,.,011, .ra,, • d ln1 pir, ells llA,, ,,. i11Tl11
(t l'i" -,np ct,a,i,q -,1..os iral iiA'I •••) 1038& 5, cf. 1058• 23]
 Classification
call the species a determination or "differentiation of the genus;
for although this explains nothing, it may guard against a
misunderstanding of that peculiar relation which we have been
trying to recognize. We recognize rather than explain it, for
explanation usually means the reference of the thmg to be
explained to something similar to but not identical with itself.
This relation is sui generis and therefore defies explanation, is
not expressible in terms of anything but itself.
f 16o. We have then two kinds of classes. Symbolizing the
elements by AB , m the one kmd, A, though it may involve
B, is not a determination of B, nor B of A, though it may
involve A , 10 the second kind, one of these elements is a deter-
mination of the other. It 1s this difference that really determines
the usage of langu.1ge. In the first case it is neither natural nor
normal usage to represent AB as a kind or species of A or of B.
No doubt it is somehow felt that sphere, for example, does not
include, in its own nature, golden sphere On the other hand
we have no difficulty 10 speakmg of red as a kmd of colour, or
of triangle as a k10d of figure The tendency of language then
1s in favour of recogmzmg the second .is tJ1c true application
of these kinds of words The obv10us mark of the distinction
between the two ktnds of class 1i. the md1fference with wl1ich
in the one case either may be taken as genus or d1fferentia,
whereas 10 the other 1t 1s 1mposs1ble to reverse the order. The
ultimate agreement between them is the contrast between the
wider and the narrower class, where wider ~nd narrower refer
to the extension. The class A mcludes the md1v1duals of
the species AB, whether B 1s or 1s not a true d1fferentia
of A.
§ 161. The logical theory of d1v1s1011 originates m the fact that
the genus includes its species ac; a plurality and thl' species its
sub-species. This suggests the question whether any general
rule can be laid down a priori for the subd1vis1on of a genus.
Let A and B be the elements of a given general conception.
One may be either a differentiation of the other or not. In the
latter, case we may further distinguish two kinds. First that
in which A and B seem indifferent to each other, and secondly
that in which A involves B, but not m the way of differentiation.
Thus m three-sided (closed) rectihnear figure the possession of
362 STATEMEN:f, THINKING, AND APPREHENSION
three sides involves the possession of three angles, but neither
of these two attributes 1s a differentiation of the other.
In the d1vis1on of a class then we have these three cases to
consider. The data of the division are, first the differentiation
to which each of the elements A and B 1s hable in itself, and
secondly the fact that A and B are not alone but in combination.
Case I. Take the case where neither of the elements A and B
necessitates the other. Let the d1ffercnt1ations of A be a1 a11
a 3 , &c., and of B, b1 b2 b3 , &c Though the nature of the one
element does not necessitate the other we cannot assume that
any member of the series a 1 a 2 &c can be combmed with any
member of b1 b2, for A and B are m combination and this may
hmit the possibihttes of combmat1011 of their respective differ-
cntiae. Thus we have prima facie two cases that m which
the d1fferentlat1on rcmams free m spite of the combination
of A and B, and that m which the combmation of A and B
prescribes certain limits to the combmat1on of their differentiae.
Case II. Let A and B mvolvc one another Herc 1t might be
supposed that the poss1b1htlc!, of combm,ttlon were limited,
inasmuch as the elements involve one another. For instance,
the possession of three sides by every triangle necessitates the
possession of three angles, and the differentiations of the sides
in respect of ratio cond1t1on the d1fferentic:Ltlon of the angles,
i e. their magnitude. The d1fferent1c:Ltlons of the one then are
not free from influence by the d1ffcrent1at1ons of the other. Bul
this is not umversally Lhc ca!,C A certain closed surface neces-
sitates a certain enclosed volume, but the various ddferentia-
t1ons of the surface may be combmed indifferently with any
differentiation of volume m respect of quantity. Thus a given
volume may be enclosed by a surface of any shape whatever.
But, in this case, there is nevertheless a limitation of the freedom
of diff erent1at1on 1£ we choose a dtff erent prmctple of division.
Although it is true that the magnitude of a given volume cannot
determine the shape of the surface which encloses it, yet, if the
shape be given 10 kmd, the volume does determine the differentia•
tion of certain of the elements of the shape. If, for instance,
the shape is to be spherical, spheres are d1fferent1ated by the
length of their radii, but a given volume determines one radius
only. Or, if the shape is to be a parallelepiped, the magnitude
 Classification
of the volume enclosed does not indeed fix any of the dimensions
of the enclosing surface but it does determine their differentia•
tions taken all together. If we choose arbitrarily a certain
length and breadth, we cannot choose any height we please.
The three dimensions together are controlled by the rule that
their product must be equal to a constant. In certain directions
then the differentiation of the two elements is independent, and
so again the differentiation of the body enclosed is independent
m respect of quabty of all differentiations of the surface whether
of kind or magmtude.
In each then of these two prmc1pal cc:l.ses we have a priori
the alternative poss1bd1tics of a free or a determinate differentia-
tion. The information necessary to settle what 1s actually true
of any particular class, whether the d1fferent1ations are free or
not, and, if so, by what law they are controlled, cannot be supplied
by logic but must be got by experience and from the special
sciences Logic c.i.n only formulate a priori the different general
cases and the two possib1htics of free or hnuted differentiation.
Case Ill. Fmally, where B 1s a true d1ffcrentiatton of A, we
have only to do with the free differentiations of the one cle•
ment A.
§ 162. In the precedmg we hc:l.ve stated the data for dividing
a class. If we wish to d1stmguish the species, we must do it by
cons1dermg the various d1fferenttattons of the elements m a given
umversal and the law of their combmatlon as affectmg in them
the combination of these different d1fferenhae... D1v1sion of
a class, however, is not necessarily understood to be an enumera-
tion of all possible species Sometimes "e only seek fur a division
which may exhaust the species m this sense, that the classes we
assign, while mutually exclusive, contam all other possible
species under them. Such a div1s1011 mc:l.y be reached by
differenbatmg one element only of the general conception, for
clearly the species formed by combmmg the other elements of
the original notion with a complete set of differentlae of the one
chosen must be exhaustive By a complete set of d1fferentiae
is here meant not all differentiations of A but the complete set
[• A species m modem botany includes ' all md1v1duals winch resemble
each other sufficiently to make us conclude that they are all, or may have
been all, descended from a. common plant ', British Flor11, •, Bentham &
Hooker, p. xi Cf. Essay II 1n Dr. Poulton's Essays 011 Evolution (Oxford, 1908) ]
364 STATEMENT, THINKING, AND APPREHENSION
allowable to it when combined with B and C. For instance, if
ABC is the original notion and a 1 a1 a3 are the complete diif~-
entiations of A for this purpose, an exhaustive and exclusive
division of ABC is clearly a 1 BC, a 11BC, a 3BC. The element thus
selected is the so-called fundamentum divisionis. The ordinary
rules about the fallacies to be avoided in division need not be
here discussed. It is obvious that a complete differentiation of
the class proceeds by differentiating simultaneously all the
elements and takmg such combmattons of them as the rule of
the original combination permits, and it is here that the idea
of co-ordination arises. Let the elements of the genus be A and
B. Suppose the first determmations of A are a1, a 1, 4 8 ; these
are said to be co-ordinate as bemg 1mmed1ate differentiations
of A. Let the immediate differentiations of a 1 be a11, a11, a18 i
these a.lso are co-ordinate with each other but are not co-ordinate
with 4 1, a2, or a 3• Thus co-ordmat1on 1s the relation of certain
species to one another, as being aU ahke members of one
differentiation of a common element They are said to be
subordinate to the common element ddforentiatcd If now we
have species of the form a1 b1 and a 2 b2 (where b1 and b8 are
co-ordinate d1ff crentiations of B), since these two complexes
result by immediate diffcrent1at1on of A and B are they also
co-ordinate ? The answer accordmg to the normal use of the
word co-ordinate would seem to be ' No '. Co-ordmation appears
normally to refer to the differentiation of one single element.
Thus a 1 b1 and a2 b2 would be co-ordinate only as subordinate to
A and through the co-ordmate differentiation of A mto a 1 and a 1•
§ 163. There 1s another relat1011 usually recogmzed between
these immediate d1fferent1ations of a common element. They are
somet1mes sa1d to be opposed So Aristotle defines opposites •
as the members of the same genus wluch arc at the farthest
distance from one another But m some genera we do not find
it natural to speak of the species as opposed to one another at
all. We do not, for instance, naturally ask what 1s the opposite
of a square. Agam it is often impossible to determine what

t• Cf besides Caugories 61- 1'l, p 366, note r, TA '" mw,jl 'ffllfl W'AltrrOP
3'11fP'P""Tfl illflll'ria, Metaph. 1055• 28, al. We cannot translate ' ccmtranea •
tbough that is nearer to wbat Wllson means than the genenc word • opposrtel •.
He wrote ,,,_;a m his MS ]
 Classification
members of a classification, even if we do not actually use the
word opposition, have the maximum difference from one another.
The species sometimes form a series which either terminates in
neither direction or only in one. Of the former kind is a series
of differences in intensity ; 1 of the latter kind is the infinite
series of rectilineal figures beginning with triangle. It must be
noticed that we are dealing with a question partly linguistic and
partly technical and the difficulties which concern the use of the
word • opposite ' apply also to its Greek correlate.
We also d1stmguish opposition and contradiction. Contra•
diction is the mere negation of a given species and the con-
tradictory therefore mcludes all the species co-ordinate w1th the
given one. The opposite falls within the contradictory, and is
considered, m some sense, as the extreme of difference within
the given genus. What does this exactly mean and to what
kinds of classes 1s it applicable ? If in a given classification we
cannot get extremes of which we can definitely say that they
have the maximum difference, it does not follow that we cannot
distinguish degrees of difference. We may still be able to say
that the species A is more opposed to the species B than to the
species C. Perhaps it might be said that the meaning of opposi-
tion is that the opposed species are entirely incompatible and
cannot coexist m the unity of the same subject For instance,
moral good and moral evil might be opposite, m this sense. But
this is obviously true of all co-ordinate species ; each one
excludes every other Isosceles triangle 1s clearly more akin to
equilateral than to scalene triangle, yet properly understood the
isosceles triangle cannot be equilateral. Orange is more akin to
red than to green, but the same colour cannot be both orange
and red ; and so of the notes of a musical scale This therefore
will not do for a defimtlon of opposition
Perhaps the true account of what 1s implied m the normal
use of opposition m reference to species of the same genus may
be as follows. Suppose the generic clement is differentiated
.according to a principle which produces a series of species in
1 Cf. Plato, Phslebus, [24 n,•) quahtJ.es hke hot and cold.
[• The reference seems to be that given in the foot-note "P"X°'fM' .,dp iral
.... ,.,_ ff Tf lrpplrr,po,, dtl iral Tl> ,fn,xp/,Ttpol' cliaumr , , , l.-.11pov "l('Y"O'T" 4,, N
fJ,pplrr,po,, •2 TOWIIJ'Tiov 4114 ]
366 STATEMENT, THINKING, AND APPREHENSION
a certain order. (We may leave the kind of order undetermined,
provided only that it be an order ) Then one species as a member
of this order, if not the first or last in order, will have one before
it in the order and one after it It will be nearer to these, in
the sense of the particular order, than to any other member of
the series. Thus in the series of natural members 4 is nearer
to 3 and to 5 than to any other number. And in general, in
the order ABC, we may c;ay there is a greater difference between
A and C than between B and A or B and C This seems to be
what is meant by comparative opposition, when we say one
species is more opposed to another than 1t 1s to a third. If
there is a definite beginning and end to the series, the beginning
species and the end species arc clearly farther from one another
than are any other pair. 1 a Tlus corresponds to the idea of
absolute opposition, Aristotle's ' opposites'. 2
§ 164. But it does not follow that relative or absolute opposi-
tion should be found 111 every class. Neither will be found in
the division of a genus unless the prmciple of differentiation is
one which produces an order. Thus m certain divisions of plants
and animals we do not thmk of usmg the term opposition. If
there is an order, but the series hac; not a definite end and
beginning, there 1s only relative and not absolute opposition.
Thus the series of integers has a begmmng but no end and there
1s no absolute opposition. Again, any geometrical series may
progress ad infinitum m both directionc; from a given member,
and the series of finite portions of a straight hne is infinite in
a positive and m a negative direction from a given finite portion.
The subJect is well illustrated by observing that 1f we differentiate
a genus by two different prmc1ples, producing two different
orders of the same species, we shall have different ideas of
opposition. If we d1fferent1ate ' come section ' by the angle
which the cuttmg plane makes with the axis of a circular
cone, we shall have the hyperbola, the parabola and the ellipse,
as the angle is less than, equal to, or greater than, half the
vertical angle. Here hyperbola and ellipse are most opposed.
1 -rd wAE&ffov d.\,\17.wr llfltT'7ltOTa , , •• lvavTla 6pl,011Ta& Cal, 6& I 7.
1 ia,a.,.•r'a

C- Rather • any other two opposed species'. On the general subJect see
Hegel, Logic, (Encyc ) f r 19.]
 Classification
If we differentiate however by the distinction between central
and non•central curves, ellipse and hyperbola will be grouped
together and opposed to the parabola.
In the mathematical illustrations the order is definite in such
a way that we can state definitely the difference between the
species ; but we may be able to place the members of a genus
in a certain order and to recognize a greater affimty between
one pair than between another and yet have no definite state•
ment of the nature of the difference between them. Thus in
the order of colours, \\e rccogmze that orange comes between
red and yellow There is an affimty between red and orange
and between orange and yellow and more in each case than
between red and yellow Thi,;; re,;ults from an immediate per•
ception of colour, not reducible to terms of anything else. So
with the order of musical notes in c1. d1atomc scale. We recog•
nize an order of pitch and are certain that in that order the
mediant, for instance, hes between the tome and the dominant,
and that the sub-med1ant follows the dominant, in the ascending
order. We can say no more of 1t and need to say no more than
that it is an order of pitch "h1ch we recognize as certainly as
an order m space.
§ 165. There remains the method of dichotomy. This takes
one differentla and divides the whole field of a class into the
members which have this and those wluch have not. Thi,;
method, as Aristotle says, 1s exhaustive ; but then the negative
class as such admits of no further divu,1on or, as he puts it,
' there are no d1ff erences of negation, merely as negation ' ; 1
mere negation contains no principle of differentiation and there-
fore there 1s no real gain m the apparent exhaustiveness.
§ 166. It follows from what has been said that the complete
and true method of dividing the universal 1s by co-ordinate
positive diffcrentiae. 2 But now the question arises how we are
in practice to effect such an exhaustive division. If we suppose
(as the common logic seems often to do) that we begin with the
perfectly determined mdividuals (with the maximum, that is, of
intension) and then proceed by abstraction to universals and
1 -ETI 11Tfp{/rff& ,.i., <Wa-ywoi, &,u,-ii, #tu &a,paiia"' ol a,xoTOfll)iii,Ttr, 0(,11 ' " ' Ii
&a,t,opd. 11Tff"l/lfO,S, ; 11Tlflll111S' davl'UTOI' -,a.p .ra., ,r.a, Toii ,.~ iwTOS De PMt. A,mn.
64:zb Zl, [Cf. II 460-476)
1 r!.rr,ll&vpr,,.la,a, &a,t,upnf, Top 141• 36
368 STATEMENT, THINKING, AND APPREHENSION
thus to complete classification, division will appear merely as
the inverse of that process, implying the previous process to
give it its material. This however does not correspond to the
actual process of thought. We are presupposing that which is
wanted as already done, and all we should be doing would be
a mere matter of arranging species in an order of subordination
and co-ordination This we could not fail to do, provided we
had them at all. Classification and division are not a mere
fom1al, or so-called logical, operation performed on a given
matter. The process is not formal at all but has to be effected
in the sciences as distmguished from logic and depends entirely
on the particular subject-matter. Thus the differentiation of
plane triangle mto acute, right-angled, and obtuse depends in
Euclid on many propositions which terminate with the con-
clusion necessary to the classification, namcJy a proof that the
three interior angles of any plane rcctdineal triangle are equal
to two right angles The species of triangle then are necessarily,
one with three acute angles, one with a right angle and two
acute angles, one with an obtuse angle (an angle greater than
one right angle) and two acute angles. Similarly, elsewhere, no
a priori rule can be given to determine the classification of
a given subject-matter. It is, however, possible to lay down
one general prmciple. Exhaustive division, not m the sense of
complete enumeration of every possible species but in that of
a division mto classes winch mclude all, is not attainable in the
case of objects so far as they are known only a posteriori. We
can never assure ourselves from mere observation that the list
of species of a given genus is complete. Completeness is
possible only m the case of universals which we can deal with
a priori, where the mind so far sees into the nature of the
universal that it can discern the determinations which it involves.
This may be immediate, as where we divide lines into curved
and straight, or may be reached immediately by a proof or
series of proofs, as where we divide plane triangles into three
species. This is possible in (pure) mathematics. The process
by which the differentiation 1s then carried on is nothing less
than the mathematical demonstration itself and thus dependent
on the particular character of the science. Empirical science
can only ain 1 at this as an ideal and can only classify such
 Classification
material as it has got. It must always have before it the
possibility of the discovery of new differentiae and can never
regard its division as exhaustive save in the barren way of
dichotomy.
f r67. We have seen that the word' kind' seems linguistically
more appropriate to those species in a classification which are
true differentiations of a genus. This seems to be the simplest
distinction of kind, as when we say blue and red agree m kind
because they are both colours, whereas heavy, blue, cold, double,
crooked, quick is a group lacking a common basis of classifica•
tion. These adjectives, we should say, differ m kind and our
meaning would seem to be that there is no common element of
which they can be regarded as determinations except the entirely
empty notion of ' being ', which is common to all notions what•
soever. But language often opposes a difference in degree to
a difference in kmd, and here the word kind has not quite the
same signification. In this use universals are said to differ in
degree or quantity when they are determinations of some
common element in respect of the more or less ; whether this
variation is measurable as m geometrical quantity, or whether
it is a mere more or less, not admitting of definite measurement,
as in the intensity of our sensations. A difference m kind or
quality opposed to this does not mean the absence of a common
element but only that the determinations of the common element
are not deterrmnat1ons of degree, that is, m respect of the more
and the less. In this sense red and blue would be said to differ
in kind, because the differentiations of their common element
colour are not in respect of quantity or degree A brighter and
a fainter red on the other hand would usually be said to differ
in degree ~nd not m kmd.
§ 168. In this connexion we meet the paradox that a difference
of degree may sometimes amount lo a difference in kind. The
statement is paradoxical because it seems to identify two sorts
of difference ; yet we observe that the phrase stops short of
absolute identification. For ' is ' we have ' amounts to '. We
shall find in the cases to which this paradox is apphed that it
is not true that as one element varies in degree there comes a
point in the variation at which a difference of kind appears in that
same element. Sometimes two elements are concerned which
~UI Bb
37" STATEMENT, THINKING, AND APPREHENSION
are somehow connected with one another and, while one varies
in degree, there is a corresponding series of changes in kind in
the other. The paradox comes from a confusion of the one with
the other. To variations in degree, for instance, in the physical
occasions of our sensations there may correspond differences in
kind in the sensations themselves. Thus, in the theory of colour,
the wave length vanes quantitatively and the corresponding
colour varies m kind. But the variation in degree is confined
to the physical causes and docs not enter into the colour series
at all. What 1s of mterest m such a case is that variations of
degree in a cause may determine not variations in degree but
variations in kind m the effect. Again, 1f the difference in
curvature between a closed and an open curve be considered
a difference in kind, the change m kmd of curvature of the
conic sections as the position of the cuttmg plane varies in degree
of angle, a variation of kind accompanying a variation in degree,
belongs to this head. We have already noticed this case and
the analogous case of the change from a curve of two branches
to a curve with one branch. 1
There is another group of cases where there do not appear to
be two but only one element with changes thus corresponding.
In the case of a sensation of warmth, as the degree of tern·
perature-the physical condition of the feehng-d1minishes, it
may be thought correct to say that the feeling is approaching
the sensation of cold. The feeling diminishes m degree until the
zero point of warmth 1s reached. As the physical conditions
continue their variations in degree, we may gradually begin to
feel cold and the sensation of coolness may increase in degree of
intensity. The truth here seems to be that in the variations
of the sensation of warmth, there is a common element in respect
of which the sensations may be said to differ in degree. But
the series of decreasing sensations of warmth do not pass into
the series of increasing sensations of cold, for these two have
no common element, except sensitiveness to heat and cold.
The diminishing series ends in an absence of the sensation of
warmth, that is there is no sensitive feeling at all, either of
warmth or of cold. Thus the state arrived at is not in an
identical series at all and so is not a sensation which belongs to
l I 164.
 Classification 371
and connects both series. When we enter the series of sensations
of cold, we enter a new series and not a continuation of the old
one. This new series consists of variations in degree of a common
element of cold, which is different in kind from the element
common to the first series. These two common elements cannot
be regarded as varieties in degree of one and the same element.
They are species of sensitiveness to heat and cold and differ in
kind, not in degree. The thing left out of account which contri-
butes to a confused view of the question is that the physical con-
ditions on which the continual change of sensation depends have
varied all the time m the degree of one common element while the
two series of sensations were varying in the degree of two suc-
cessive elements. The termination m zero of one series no more
leads to, or begins, a new series of sensations of cold as such than
it leads to, or begins, a series of sensations of pain, hke a smart
or a toothache.
§ I 69. • From an early period in the history of logic distinc•
tions have been made between different sorts of classification.
There has been a tendency to regard some classifications as
truer, or more important, or more natural, than others and to
deem some genera as of lugher rank than others and as more
natural kinds. Thus 111 Aristotle there are genera and species
which seem to have a special cJa1m to be so called. They are
thought of as the truest genera and species and the tendency
seems to find something bke formulation m the phrase 'secondary
essences' as used m the Categories. 1 There hac;, however, always
been great obscurity as to what this higher value or rank con•
sists in. In this spirit a distinction is sometimes made between
a natural and an artificial classification, m the writings of
Linnaeus for instance and in some more modern botanical and
biological authors 2 But we look in vain 111 such writers for any
clear idea of what they mean by a natural as distinguished from
1
Cf. fl 438-50.
1 See Whewell, History of tht Indm:twe Snencts (New Ed 1847), Bk. XVI,
c:hs. 1-5.
[• The author's cnticism here appears somewhat pedantic ; contrast I I So.
The greater naturalness of the later classdicabons, compared with the Llnnaean.
lies in the fact that the latter's arrangement was numerical The lnstory of
systems of botanical class.dication is very clearly set out in Tht Eletnents of
Botany, by Adrien de Jussieu (trans J H Wtlson, 1849), pp. 509-37]
Bb2
372 STATEMENT, THINKING, AND APPREHENSION
any other classification. It is impossible to construct a clear
idea from what they say and they seem never to realize clearly
what they want. Thus they tend to define the word natural by
itself. Whewell, for example, says that certain classes are natural
in~smuch as the division employed brings together those plants
which are naturally related,1 and Linnaeus' own confusion is
illµstrated in Whewell's History of the Inductive Sciences.•
We may find some help by cons1dermg a botanical system
presented to the world as a natural classification, that of the
two de Jussieus. 8 We see that one of the main characteristics
which determined the classification was the number of coty•
ledons. It 1s a fact that monocotyledonous plants agree with
one another m a vast number of details besides this character
and differ in the same particulars from dicotyledonous plants.
There 1s, however, no apparent connexion between these details
and the character chosen as a d1ffercnba.
§ I 70. This suggests the explanation that the principle which
appears more natural 1s one to the diffcrentiation of which
corresponds the greater number of other elements in the objects
to be classed Behind this 1s the idea of necessary connexion,
however dimly realized. Just as a variety in the general defini-
tion of a come section necessitates every other variety and
difference between the species, so the ideal principle which
modern scientific students are feeling after is that of a universal
which includes the whole nature, not merely a part, of the thing
classified, in the sense that the particular form which the
principle takes in a particular species determines necessarily
every particularity of that species. Clearly the secret and
imperfectly understood reason for prcferrmg one principle to
another in such cases 1s that one prmc1ple-the one which
seems natural-is in this way far more of an index into the
nature of the obJect than the other.
This difficulty which the empirical sciences have in deter•
mining what is a natural classification in a given department
is due to the fact that the advance of knowledge alone can show
that one principle is, at least as far as we have gone, of wider
embrace than another. The investigation being a posmiori
1 Whewell, Hiskw)' of tht Inductive Sciences (New Ed. 1847), XVl. 3, I 2
(vol. iu, p. 313).
1
1b XVI. 4, § S • 1b. XVI. 5, p. 369.
 Classification 373
gives us no insight into the necessary connexion between the
variations of the given element by which we are classifying and
the concomitant variations of the other elements. As long as we
have no insight a priori into such connexion we cannot be sure
that the element chosen includes in the above sense the whole
nature of the thing which is to be divided. The ideal can only
be obtained in an a priori science where we can understand the
necessary connexion. In the empmcal sciences the ideal can be
only approximated to and the form of the supposed natural
classification m any such science is lic1.ble to continual modifica·
tion as the science advances
§ I 71. The tendency to look on some sorts of classification as
specially real, by comparison with others which are artificial,
has found expression in Mill's doctrine of recll kinds. There is
a great contrast, Mill thinks, between classifications of things
by such attributes as blue or heavy and those by such universals
as plant or animal. The former he thinks arbitrary or artificial,
the latter he names real kmds and he thmks it would not be
wrong to say that wlule real kmds are <1. d1stinct1on m nature
itself, the other kinds or classes arc made by us for our own
convenience. The d1st111ct1vc characteristic of a real kmd seems
to be that it comprehends an inexhaustible number of attributes,
so that the members of a real kmd agree with one another in
an infinite number of characteristics and differ also m the same
manner from the members of other kmds. The characteristic
of the kmd which is not real is that the attributes 011 which it
depends are easily exhausted. The class of blue things differs
from that of red only m the fimte attribute colour. We have
then to ask, is a mathematical notion hke triangle a real kind
or not? It would seem from the examples that Mill takes that
it ought not to be a real kmd because, although the attributes
of a triangle are inexhaustible, yet they are exhaustible in the
sense that they arc all derivable from a fimtc number of attributes
in the definition of a triangle. Thus infinity of attributes does
not constitute after all the nature of a real kmd, although it 1s
the characteristic Mill most insists on. We must have infinity
of attributes together with the impossib1bty of deriving that
infinity from a finite group of attributes.
Now, in the first place, we naturally ask what evidence an
374 STATEMENT, THINKING, AND APPREHENSION
empirical philosopher like Mill has of the existence of such kiads.
Observation could never show in regard to any supposed real
kind that its attributes were anything but finite. The reality
of an infinity can be known only a priori. Moreover, how could
we tell that the fact that a certain number of attributes had not
been derived from a given finite group was anything but an im•
perfection m our knowledge, and thus that, even assuming an
infinity of attributes, real kmds would not pass into artificial
kinds 1f we only knew enough ? This brmgs us to a second
difficulty and we ask what reason there 1s for the preference
shown to the real kinds. The infinite attributes of triangle, an
artificial kind, all follow from the fanuliar simple definition.
The real kind, w1th 1ts mfintte attributes, 1s a sort of chaotic
ununified infinity. For our purposes, at all events, the advantage
would seem to be altogether with the non-real kind. And how
does Mill know his real kmds ? He must assume that we assign
individuals to a definite kind on the ground of some attribute
or group of attributes, which is definitely determinable and is
the mark, as logicians term it, of the kind. But this mark is
useless unless its nature necessitates the presumed inexhaustible
number of attributes and, m consequence, that infinity would
be the consequence of this finite mark and the kind would cease
to be real. Mill's idea then seems altogether self-contradictory.
We may perhaps gather what was in his mind from the followmg
extract: 1
'A hundred generations have not exhausted the common pro•
pert1es of animals or of plants, of sulphur or of phosphorus . . .
while if any one were to propose for investigation the common
properties of all things which are of the same colour, the same
shape or the same specific gravity the absurdity would be
palpable.'
It may be replied that geometry undertakes to investigate
the properties of objects that have the same shape, and that in
one sense at least the investigation described as absurd is the
procedure of all those sciences which isolc1.tc certain attributes
of things for the purposes of study. Mill, however, clearly means
something of this sort. It would be absurd to form a class of
objects according to their shape and to propose to inquire into
1 MJJJ, Sysle#t of Loi'& •, I. vii, f 4.
 Classification 375
their common properties in general, without any reference to
what depends on the shape, to ask for instance for the chemical
or physical properties of spheres. What however would the
investigator do who examined the common properties of sulphur?
He must classify pieces of sulphur by some common mark and,
by hypothesis, this common mark is not to necessitate the
presence of the other properties of sulphur or sulphur would
cease to be a real kind. He would, therefore, be examining
a class of things to find out common properties m them which
have no connexion with the mark or marks by which he classdies
them. This is the precise mistake which to his mind makes the
absurdity attaching to the non-real types of classdicat1on ;
classifymg things, that 1s, as blue and then askmg for the geo•
metrical properties of blue thmgs. It 1s clear that scientific
investigation is only Justified by the knowledge, or by the
suspicion, that the properties we are mvest1gatmg are necessarily
connected with the properties which serve as the basis of our
classification.
§ 172. We feel, however, that there is a difficulty, though we
cannot admit that Mill has succeeded m discovering its true
character. We may perhaps offer the following as an explana•
tion. In our ordmary experience, ,,hether rightly or not, we
distinguish things and attributes, things as independent realities
and their attributes as partial or dependent rcaht1es. Or the
attribute may be represented as a part of the existence of
a thing or, at least, as not exhausting the nature of the thing.
Corresponding to these we have two kmds 01 universals, for we
must avoid the mistake of supposmg that the attribute is
universal, and the thing itself particular The particular then
which corresponds to one kind of universal is the ind1v1dual
thing. To that corresponds the habit of language m which such
universals, as common terms, have a quasi-noun expression. The
other kind of universal has for its particular what seems only
a part of a complete reality, what is expressed m language by
an attributive term. Here then the universal 1s understood not
to comprise in itself the whole nature of a particular thing.
This is why what Mill calls real kinds seem, as the phrase is,
to go deeper into the nature of a thing, and this is really all that
he tries to express when he speaks of these kinds as in some
37'> STATEMENT, THINKING, AND APPREHENSION
way specially real. This again accounts for a certain mystery
in his description of their inexhaustible number of attributes
and of how the attributes cannot be got at through one universal
from which they can be derived. Such universals of individual
things arc of a problematic nature : they contain this infinity
as the thought of a potentiality which may be developed without
limit. They do not m fact expre,s the nature of a thing ; they
are rather the idea that there is a particular thmg and that it
has within it this mfinitely developable nature. But Mill fails
continually through not distmguishmg the universal which is
definite from the problematic umversal which implies that there
is something definite without determining what it may be.
Such umversals, m one way, do not include and, in another, do
include the total nature of such a thmg. They do not include
it, in so far as they do not give a defi.mte universc1.l by differentia•
tlon of which one could arrive at the thmg itself. They do
include it problematically because they are the universals of
particular thmgs conceived as possible deternunations (1£ we
only knew how to conduct the process) of a umfied umversal.
The more artificial character, m Mill's view, of the other classi-
fication simply means that the umversals there arc abstractions
of particular aspects of the thmgs and artificial m the sense
that m Nature they arc not so separable. This view agrees so
far with what seems to have been the Aristotelian tendency.
In the Organon the secondary essences 1 are described as those
which arc the proper genera and species under which the primary
essences 2 arc found These essences then arc •animal' and
'man' on the one side, and Socrates and the md1v1dual animals
on the other. It 1s not to be understood that the Aristotelian
logic arrived at a quite dear idea of what the d1stmcbon meant,
but on the whole the d1stmcl1on really mtcnded seems to have
been between umversals hke ammal, wluch may be truly said
to include the whole thing under them, and universals like
whiteness,• which arc such that the thing is sometimes classed
under them, as when we say Socrates is a white (being), 1 yet
are understood not to include within themselves all that the
thing in its fullness means.
1
a,,.,,pai o6o'la.. • rrp&rra& oiHtla,, • S. ian Ae-S. f'I,

[& The word ' whiteness ' as sub&tdutcd for the neut. adJective AtlNM' whic;h
the a.uthor uaed,]
 XVII
DEFINITION
§ 173. DEFINITION seems properly a subject belonging to logic
as a study of thought since it concerns a relation between things
and classes and their constituent elements, considered generally
and without reference to the particular nature of any individual
relation of the kind. In the majority of cases a given attrtbute
is recognized to belong to a subject and yet to express only
a partial determination of that to which 1t belongs. Hence we
are naturally led to ask whether a group of attributes can be
found which embraces the nature of a thing or of a class com•
pletely. To this search corresponds the statement of Aristotle
that definition 1s of essence or being. 1
Again, the attempt to distinguish one thing or kmd froqi
another involves the recognition first of some common element;,
and secondly of something m which the thing or kind differs ·
from other tlungs, somcthmg peculiar to the thing or kind.
Hence arises the question whether there is any attribute or group
of attributes wluch will serve to distinguish a given thing or
kind from certain others. The statement that defimtion is to
be by genus and differentiae 2 is the outcome of this second
question. The two questions correspond to the manner in which
the search for prmc1ples of defimbon developed in practice.
That the second probably attracted more attention at first we
may gather both from the derivation of the word definition,
which in Greek and Latin means fixing of boundaries, 8 and from
the fact that the Anstotehan dictum 4 that defimuon is to be
by genus and di.ffcrent1ae was accepted m the schools and from
them has passed into the traditional logic. The other impulse,
the search for a complete determination, does not so directly
suggest the distinction of genus and diffcrentia, but does in fact
lead to the same inquiry in the end. 6
1 A,u,p~s ,u11 ~ ~ ,.; IN, ""! wt1l11r. An Po. 90b 30.
1 & &f1U1,,J,.le 'fl11ow .au &atf,opG,,,, Top. 103b 15. Meltlj,h. 1024• 26, 1037b ag
,,,__,.., -y4.,.).
' ¥C•• &p,C4dm, definite.
• Nute :z (above). • S257.
378 STATEMENT, THINKING, AND APPREHENSION
I 174. The object of defimtion is sometimes said to be the
thing or objective reabty itself and sometimes our own concep•
tion, our subjective idea. The very idea of definition in the
second case seems difficult, if not paradoxical. It may seem
that we cannot seek to define our own conceptions because we
must know what they contain before we begm, or else what is
called the definition of a conception 1s either the definition of
a word, and nominal, or only the arrangement of elements,
already given and known, under the heads of genus and ddfer•
entiac. The latter process, at any rate, might seem to be merely
fonnal and of little interest or value. All th1s it may be seen
implies that the idea must be clearly, i.e. exphc1tly, before us.
Again there is a kmd of definition which seems to relate to
the obJect, for appeal is made to instances 10 experience. This,
however, seems to be equally trifhng ; the formation, that is,
of a definition by abstraction of a umversal from particulars.
This case 1s really the same as the one we have just been con•
sidering, because the end which is sought for is really the deter·
mination of a universal as such. This process is formulated by
Aristotle, 1 who represents 1t as starting from a set of similar
individuals and comparing them with a second set, distinct in
species but identical m genus, and so arriving at a genus by
successive ehminat1on of ddferentiae. Such a process seems to
be wholly artificial and to presuppose throughout what it seeks
to attain. In order to abstract magnammity, we must know
that the mdividuals taken arc examples of 1t, and how are we
to know th.1t without already knowing what magnanimity is?
Aristotle himself betrays the difficulty, for he says that, if we
abstract from the members of one species of magnanimous
persons and find a universal and then abstract from another
group of individuals m another species, we must again compare
these and abstract the common clement from them : and, if
there is one element common to all the species, this will be the
common definition, but, if we do not find anything in common,
we shall not end with one genus. 11 This of course implies that
the process may fail, and is an mherent contradiction, but

1 An. Po 97b 7, seqq.

I d H ""U", Ho .ra., a. .r,, ..... p.rpA.Olfuxta,, 1b. 97" 24.
 Definition 379
Aristotle neither recognizes this nor raises the questiOA which it
suggests, It is clear that the particular facts from which the
abstraction is made are for the purposes of the abstraction
ultimate. Hence arises the question of how we got at this basis
and what we can do, if we do not find that we can unify the
species under which these particulars seem to fall. The difficulty
showa that there may be something wrong tn the choice of these
particulars as ultimate, and this makes it absolutely necessary
to ask what 1s the guarantee that we have started with the right
instances as a basis at all.
I 175.• It may seem that \\c have only to reject this process
as absurd, but the fact that such a theory could have been
held and have become so popular cannot go for nothing. There
is probably some process which does correspond to 1t, though
the process has been inadequately formulated. The answer
cannot be got out of mere general notions or the form of thought.
We must look at the specific matter to which this process has
in practice been applied. This matter seems to be our moral
experience. It was the Socratic abstraction of moral notions
which probably occasioned this logical doctrine of abstraction,
and 1t 1s this process which Aristotle is represcntmg in his
account. The facts of this abstraction have already been suffi•
ciently examined in earlier sections. 1 The truth 1s that the
process is not one of analysis of a given c.omplex, that is, it
does not proceed from the more determined particular to the
more abstract umversal ; it only appears to do so because the
materials from which we start are fully determined. What
we really do 1s to make determinate and cxphc1t what was for
us previously indeterminate and implicit. The general idea of
affinity becomes the more determinate idea of what the affinity
consists in. This explains the apparent paradox. It shows that
there is something corresponding to the Aristotelian doctrine
and how the mismterprctation of that doctrmc arises. In con·
clusion it is useful to observe that we see the same thing in the
working of our consciousness in other departments. In some
1 Part I, ch 2, ff 11-12.
[• I have onutted here a long d1acuSS1on of Socratic definition. Tbe aub•
stance was later embodied m Part I, ch 2, f u Wilson's attitude wu DO
doubt ongmally auggeated by Green (Wc,,-ks, vol.111, pp. 55-8).]
38o STATEMENT, THINKING, AND APPREHENSION
cases we are even obliged to remain in the stage of recognition
of affinity without being able to characterize that affinity further.
Thus we recognize that colours have an affinity which is so
definite that it distinguishes them from sound, but we are quite
unable to say what the affinity of colours or sounds consists in.
§ 176. The attempt to find clear distinctions between objects
produces the kind of definition which is said to be by genus
and differentia and implies an explicitly known material to be
arranged. There 1s another tendency, also in the direction of
clearness, represented by the Socratic definition, which seems
to be the effort to become explicitly conscious of a universal
implicit in certain particular statements. There is a third
tendency, somewhat different, the attempt to get at a universal
which shall unify a manifold ; to grasp in fact all the variety
in the attributes of an object under one notion. This corre•
sponds to the Aristotelian search for essence. 1 The very idea
of this is, on the one hand, a comprehension of the whole thing
and, on the other, a comprehension which is by no means the
enumeration of all the possible attributes of what 1s defined.
It is clear at once that tlus has an ob;ect1ve meaning and does
not relate merely to the arrangement of matter already given.
Whether such an idea as that of cssenc.e 1s valid or not, it 1s
clear that 1t forms an important element in the effort of the
mind to gam a mastery over the object. It may be noticed that
this effort is wrongly formulated m the scholastic account of
essence : but 1t must not be rejected on that account, or we
may fall into the same mistake as that of Mill, the error of
cr1tic1zing the formula and not understanding the activity of
thought which has produced 1t. If we ask what essence 1s for
Aristotle as comprised m a defimt1011, we do not find any clear
account, m so many words, of what essence is itself and this
may explain in part the failure of the schoolmen. It is with
him rather a fixed idea, by the help of which he defines the
attributes contrasted with 1t. Thus m the Topics 2 essence,
property and accident arc distinguished. Essence and property
are both necessary to the thing defined and convertible with it,
but to distinguish them we are merely told that the property
1
dpcO'pt i"" .,a,, Toii ,-[ IO'T& • oiiula, An. Po gob 30; cf. Melfl/)h. 1037b :15.
1
T~ Tl ,Ji, tfl'fl1, r&°", t111/l,lltll'ltc6s, Top, 103b 7-19.
 Definition
is that which being convertible with a thing is not of its essence.
The distinction, however, of essence and property is the basis of
Aristotle's theory of science in the Posterior Analytics,• and his
use of these terms there implies that essence is that group of
the attributes of a thing from which all other attributes follow
or are derived, the essence being itself underivable, primary,
and hence its own ground and, from the point of view of know-
ledge, self-evident. Further, this derivation is apparently
objective as well as subjective: there is an order in the being
of the thing as well as m our apprehension of it.
§ 177.b This conception appears at first sight to agree with
the facts of mathematics, and this is perhaps why it seemed
satisfactory to Aristotle. But really it involves two serious
difficulties. It presupposes an absolute order of thought;
implies therefore a certain dependence of one part of truth upon
another. In actual demonstration, however, we find that '\\e
can reverse this order of essence and property, at least some-
times ; that is, 1£ we prove that A necessitates B we may be
able to show also that B necessitates A. Aristotle was aware
of this in so far as he teaches that the property is convertible
with the thing of which it is a property. Yet he never faces
the question which this suggests and never asks himself whether
this does not destroy any absolute precedence of one element
to the other. One may conJecture a reason for this. Possibly
he had not before him instances where, in the icciprocal relation
of two elements A and B, an independent start was made from
A to arrive at B, and conversely. When c1. reciprocating con-
dition 1s proved by Euchd, a reductio ad absurdum proof is
normally employed, which assumes that the first condition has
been established. When the proof is not ad absurdum the second
theorem (as in i. 48) may still proceed by the heJp of the former.
A second grave difficulty is that it would seem impossible to
ascertain what is the true definition or essence, because to know
the essence as essence we must know that 1t accounts for all
the properties, and therefore it would seem that we must first
know all the properties. Yet the latter condition cannot be
[8 Especially Book II, passim
b This section is inconsistent with Wilson's later view~, see sections referred
to 111 the foot-note, p. 382.]
382 STATEMENT, THINKING, AND APPREHENSION
realized even in mathematics, to which this kind of definition
seems applicable, because the properties may be infinite. Besides
we have this paradox, that in mathematics instead of knowing
the definition in this way from the properties we arrive at the
properties from the definition.
But for the existence of mathematics the second difficulty
would not be enough to threaten the ,·alidity of the idea of
definitional essence. We might possibly never have an absolute
definition, in this sense, but we might have a relative one ; we
might, that is to say, unify the facts which we know about
a given subject by dividing them into two groups, one of which
conditions the other, and this assemblage of the complex of
conditions we might term the essence. Growing knowledge of
a thing might then sometimes widen and sometimes narrow this
essence. New elements might be discovered which we could not
derive from the old essence , or, agam, some of the elements,
at first included in that essence, might turn out to be derivable
from others of the group. Thus the complex of the essence
would be diminished. Definition then would be an ideal by the
aid of which we umfy what we know, and this corresponds fairly
to the present position of the empmcal sciences. But it is
otherwise m geometry ; there we appear to know the essence
first and to know that 1t is the essence without having to go
all through the properties. Moreover, there no new knowledge
can disturb the validity of the definition, because the very
extension of our knowledge to new properties consists in their
derivation from the definition itself. The reason of this lies m
the special nature of geometry and consideration of it must be
postponed to a later secbon. 1
§ 178. The true method of dividing a class has been seen to
be a developmen_t in an orderly series of the manifold under-
lying its unity.VIn definition by essence there is also a unifica-
tion of the manifold. The essence is to be such an element or
group of elements as may account for the manifold or even
infinite attributes of the thing to be defined. Definition by
mere classification wdl not satisfy ~he requirements which are
implied in the search for essence nor necessarily conduct to the
essence at all, but it can be applied to the essence itself and it
1 See Part III, ch. 3, esp. H as7-8.
 Definition
is important that it should be so applied, for the following
reasons. It removes the individual from its isolation and
indicates its place in the system of reality to which it belongs
by representing its relation to other individuals and species
whether in the way of co-ordination or subordination. More•
over, it tends to give us a clearer grasp of the subject of the
definition taken by itself and may even lead to a more complete
knowledge of the relation of essence and property withm the
individual. For example, it is true of every isosceles triangle
that its interior angles are together equal to two right angles,
and this might be proved of all isosceles triangles independently.
But, as we know, it is a property not of isosceles triangles as
such but of all triangles, that is of the genus. We may represent
this symbolically as follows Suppose that A, an element of the
essence, has the determinations a 1 and a 2, of which a2 1s a further
determination of a 1• We may not have recognized the fact o{
this differentiation and, having the element a1 before us, may
have either demonstrated the attributes xyz of a8 or have recog•
nized their presence in a8 by observation. Of these attribute,;
x may depend on the most general form A, y on a 1 , and only z
upon the complete determination a2 Now 1£ we i.ucceed m
effecting the determination of A into a 1 and then mto a 1, it is
at once suggested to us to find out which, 1f any, of the attributes
x:,,z depend upon a more universal form of a 2, say a 1 or even A.
A problem 1s thus suggested to the science, which the science
itself must solve. It must not be supposed that the mere act
of differentiabon enables us to make these Jistinctions in the
properties.
I 179.• Science thus gains in clearness and completeness. We
know what we did not know before, the true conditions of
x and :,,. Science gams also in extent ; we discover that such
a property as x is independent of a particular d1fferentiation
and belongs to every species of A. This development of know•
ledge has been conspicuously illustrated m the modern develop•
ment of mathematical science. With this is connected another
important matter, the possibility of converting a scientific pro•
position. When we have disentangled the real condition of an
[• In lecture this section wa~ illustrated in 1893 from certa.tn anaJosoua
properties of curves of the second degt"ee ]
384 STATEMENT, THINKING, AND APPR£HENSION
element we have always a convertible proposition : B condi;
tions A, A conditions B.1 And this gives us a test or sign as
to whether our knowledge has the true completeness and the
conditioned is referred to its proper condition. If the proposi•
tion is not convertible, we have still to seek the true condition
and must analyse a1 further. This analysis may result in such
a differentiation as the one described or, on the other hand, in
the discovery of a complex of elements within a1, a part and
not the whole of which conditions the attribute in question.
Science gains a further advantage by a distinction of genus and
differentiae within the essence. The thing defined is not only
related to others but new attributes may be discovered in this
way. Suppose that in two things which are otherwise very
different we discern elements which are analogous to one another,
in the sense that they are differentiations, whether co-ordinate
or not, of the same genus. Suppose that, in the ,one case, we
have demonstrated a property connected with one of these
analogous elements. This suggests that we should look for an
analogous property in the second, and we shall be prepared to
find that the analogous property will differ from the property
demonstrated in the first instance, partly because the analogous
elements are not identical, and partly because they have such
a different context. This may seem a very simple consideration,
but, when applied to a given subject-matter, it has resulted in
remarkable discoveries. Thus, remarkable analogies have been
discovered between attributes which, apart from such a classi•
fication as is made 10 the definition, may appear so different
that they would not in themselves suggest that we should look
for any affinity. The suggestion is made m the manner described
when we discover analogous elements in the essence of both.
Moreover, we are led to look for new attributes in an analogous
subject-matter where the investigation would otherwise not have
occurred to us. There are instances of this in both the mathe•
matical and the empirical sciences. The search for analogies
has been very fruitful in biology m regard to the life of plants
as well as of animals.
§ 18o. These considerations perhaps explain the question pre•
viously raised I as to what is the real aim of scientific writers
l 1259• If 169,
 Definition
when they recommend some particular classification of their
subject-matter as pre-eminently natural and as superior to other
possible classifications. They may also explain why classifica•
tion as such is useful to science in facilitating discovery, and
not merely as a convenient or neat arrangement of the facts
discovered. In the empirical sciences every advance shows the
impulse to unify the manifold of what is known by the discovery
of some ideal method of classification, which seems really to be
the application of the true differentiating method to the essence
itself. Certainly the guiding conception of so-called natural
classifications seems to be that the elements chosen are such
that to their variations correspond the maximum of variations
in the total of the elements which are known in the subject-
matter considered. The presupposition doubtless is that what
is taken for the class universal necessarily cond1t1ons {though
we may no•know how) everything which there 1s in the indi•
viduals studied, just as the mathematical essence conditions
everything there It is clear, from the considerations we have
adduced, that such a classification gives a greater grasp over
the matter of the science. Such classifications have then a two•
fold use and meaning. They are a step in the discovery of the
essence, which 1s mev1tably presupposed as determining the
elements which vary together, and they help us to discern
the unity, which underlies the different manifestations of the
real world, by extending the relation between the subJects of one
science to those of another.

11713•t cc
 XVIII
DENOTATION AND CONNOTATION
§ 181. MoRE than once I have referred to the confusion which,
I hold, besets the subJect of denotation and connotation. The
mischief seems to me to be due to a serious want of clearness
about the nature of meaning, some other results of which are
pointed out in my criticism of the doctrine that the true subject
of every existential judgement is the ultimate reality. In the
following investigation I shall devote myself to a search for
a positive answer to the general problem to which the theory
of denotation and connotation seems to belong This is a positive
supplement to negative crit1c1sm of traditional or current views
advanced by me elsewhere. I shall follow this by another
negative criticism which illustrates the method to be pursued
in examining any discussion of denotation and connotation
which may be found in the ordinary manuals.
In the case of any given writer one has to ask how he defines
these terms, or, failing an explicit definition, what definition he
really presupposes. One has then to inquire whether he is
consistent with the definition, 1f he gives one, and whether he
really employs more than one principle for the distinction he
makes or assumes. If he explains the distinction only by giving
instances of it or by describmg the general cases of its applica•
tion (and this is pretty much what Mill does), we must ask
whether the terms have any meaning except as labels for the
relations indicated If they are vindicated as simply meaning
those relations, we must insist on knowing why these terms
have been chosen as technical terms to designate the relations,
instead of words which have no significance otherwise. Again,
if they are mere labels, we can test their applicability in a given
[• This is redrafted from a manuscnpt note-book of uncertam date. The
chscussion 1s coloured by polemic against Mdl and the o,d,11417 logic. It
represents fairly the VJew which Wilson always maintained in informal
instruction Cf Part II, ch 8]
 Denotation and Connotation
case only by substituting for them the account in ordinary
significant speech of the relation of which they are the labels.
The result of this will always, I think, be found disastrous. If,
however, the word denotation is used as m ordinary speech, we
shall find most extraordinary confusion to result. Any attempt
to give a significance to ' connote ' by making it equivalent to
' imply ', as Mill tends vaguely to do, leads also to confusion.
I 182. The distinction between denotation and connotation
has to do with the meaning of words and with the relations
which meaning bears to subjects and attributes, a grammatical
distinction which is presupposed; to put 1t more generally, it
has to do with the relation which a word may have to some
attribute or subject with which it may be concerned. In the
discussion of the distinction a word is sometimes said to be the
name of something, where ' name of ' has a sense which must
be nxed by ordinary usage. Again, words are said to mean or
to signify or to denote something, and here again we have non•
technical expressions, the sense of which is fixed by usage The
words connotation and connote are, however techmcal, employed
in a novel sense. 'They cannot therefore be presumed to be
equivalent to the words implication and imply, which are termc;
of ordinary speech. If they were, they would be superfluous,
there would be no need for them Now we arc entitled to use
technical grammatical expressions, hkc subject and attribute,
because it is precisely word-forms which we are to examine and
word-forms distinguished into nouns wluch are proper names,
nouns called abstract, nouns called concrete- general names, and
adjectives. But our question being what 1t 1s we must at first
use no other technical terms. For, in the first in-;tance, we have
simply to inquire what the linguistic facts arc and how these
word-forms are related as regards meaning to the distinction of
subject and attribute in their actual use We must not assume
and employ the words connote and connotat10n, which are
technical, because that would be to beg the question. The
question is the legitimacy of the distmction conveyed by these
words. as distinguished from denote and denotation, and it is
just because this is not realized that the d1scuss1onc; of the
question (so far as I know them) arc quite futile. The question
generally put as the problem is whether a given kind of word
CC2
388 STATEMENT, THINKING, AND APPREHENSION
has connotation or denotation, or both, and what is the denota·
tion and what the connotation of a given word. This pre•
supposes the legitimacy of the distinction (the very thing which
I call in question) and, if the distinction is in fact false and
illegitimate, attempts to answer such questions must result in
confusion. We might as well ask whether this quill pen can be
mended by kindness ao; well as by reproof, and, if so, by what
sort of kind treatment. Thus, it is usually, perhaps always,
taken for gr.intcd that every noun and adjective must have
either denotation or connotation, if not both It 1s at all events
assumed that every noun has denotation at least, the dispute
turning on whether all, or only some, have connotation. But
supposing it should in fact be true that some words of the above
kind have no denotation, how particularly puzzling and con-
£using must be the mqmry as to what 1s the connotation and
what the denotation of such words I
§ r83 We have therefore to inquire first mto the lmguistic
facts mvolved rn the problem before us and to abjure any new
technical \\ord until the facts compel its use, that is, until wc
come acroo;s some new relation for which ordinary speech has
no accepted word Moreover, wc must first consider words as
they actually are used and not in abstraction from the use made
of them in a particular sentence ; \!\'hen we do consider them
in the latter and abstract way we must expressly say so or
confusion will certainly arise. Indeed the confusion m the
doctrine of denotation and connotation is caused partly by
failure to remember this distinction. Now the relation of a word
or grammatical form to any subject or attribute with which it
is concerned is a matter of usage. It is not fixed by the gram•
marian or the logician, it depends on no logical or grammatical
theory. We ought then to be able to ascertain definitely the
relation as a mere fact of linguistic usage.
§ 184. Let us therefore first consider the use of ' name'. Mill
treats all nouns and adjectives as the names of something.
Consider then such nouns as John, man, stone; such adjectives
as virtuous, human, heavy ; and such nouns as humanity, virtw,
heaviness, weight. We will first ask whether these nouns and
adjectives are the ' names of ' subjects or attributes, according
to the normal meaning of ' name of ' in language. John as
 Denotation and Connotation
used ordinarily in any given statement is the name of .a. parti•
cular individual,1 and we should say that this word John also
means the given individual whose name is John. It is the name
of an individual subject and not the name of any attribute of
his, as (for instance) his height, his weight, or his virtue. Singular
proper names then are in usage always the name of one person
and mean only one person.
§ 185. Consider now 'concrete general names' hke man. If
we suppose that a word has always, hke a proper name, to mean
that of which 1t 1s the name, of what 1s man the name, and of
what is stone the name ? Tree m a given sentence may refer
to a particular tree, but it does not follow that it 1s the name
of that tree. In ' this tree is hollow ', 1t 1s only the combination
'this tree' which means a particular tree. In this sentence then,
if anything could be the name of the tree meant, 1t would be
this tree. Yet 1t would be contrary to linguistic usage to call
such a combmatlon a name of the thmg meant, to say that ' this
tree' 1s the name of the obJect pomted at. The same holds of
the tree. If again we abs.tract stone and tree from their use m
given sentences, we cannot say that they are the names of
a given stone or tree, for they \\ould by hypothesis have then
to mean that stone or this tree and no other, and this they
certainly do not. Tree then 1s not the name of any particular
tree. Neither however is 1l the name of the umversal of tree,
for 1f that has any name it would be such a "ord as ' lree-ness '.
Neither therefore m use nor m abstraction could \\e say that
such common nouns or general names a& tree arc the names of
either subjects or attributes, 1£ a name has to mean that of
which Jt JS the name, nor the names of the umversals of subJects
or attributes. We may, however, note 1n passing c1.n ordinary
usage by which general names arc called names of somethmg ;
as, for example, when we say ' Gill JS the name of a certain
measure ', ' Ibis is the name of a certain kmd of bird '. This
has interesting 1mphcations which need not detain us now.
§ 186, We will next consider abstract nouns bke u1eight, heavi-
ness, virtue. Heaviness is naturally and rightly said to be the
name of the universal, and 1t means that of which 1t is the name
1 The pnmary appbcat1on, m the development of language, of the 'W01'd

' name • i,. probably to proper name:..

390 STATEMENT, THINKING, AND APPREHENSION
and nothing else. This is so, if we consider it 1n use, for though
an expression like this heaviness refers to and means the weight
of a particular thing, the word heaviness in the combination
does not ; it still means the universal, since this heaviness is
equivalent to this particular mstance of heaviness. The sawe
holds 1f we abstract heaviness from its use in a particular
sentence ; it is m general a word which 1s always to be used
for the universal, as meaning 1t and as its name. Thus abstract
names like heavmess, whiteness, courage, are nanres as strictly
as proper names. Indeed an abstract noun is a name even more
completely than a proper name 1s, for the word John in abstrac•
tion (as it might be found in an English-French dictionary, for
instance) cannot be said to be the name of anything at all.
Strictly it 1s only a word such that when used m a given sentence
it 1s the name of an individual male person. An abstract name
is a name whether considered m use or in abstraction.
§ 187. Consider next adJect1ves. Mill and other logicians speak
of them, without any scruple, as names. Of what then is an
adJectlve like heavy the name ? Mill would answer that heavy
is the name of heavy thmgs, whereas heaviness 1s the name of
the attribute, weight, and so for other adJectivcs. But now
heavy 1s not the name of heavy tlungs. If anythmg 1s, it must
be precisely the "ord combmat1on heavy things. In use, heavy
1s neither the name of heavy thmgs, nor of any heavy thing i
it means ne1thcr lhe heavy tlung nor its heaviness; it is not
the name of its heaviness nor of heaviness m general. In use,
then, it is the name neither of a subject nor of an attribute
(particular or universal) A fortiori it is not such a name in
abstraction, nor 1s the account of 1t, 1n abstl action, that it 1s
a word which m use, that is, m a particular sentence, is the
name of a particular heavy thmg (subJect) or a particular heavi•
ness (attribute) In no legitimate sense, then, of • name of ' is
heavy the name of anything. The same must be said of adjectives
which do not refer, as 1t might be argued, to attributes because
they refer to the whole being of a thing, as subject or substance.
Such adjectives, for instance, as human, which in ordinary
untechnical speech 1 are rare, are not properly names of anything.
§ 188. Consider now the relation of signify and mean to the
1 Sec § 78 on the cxprcoo1011 of u111vcn.ab 111 language.
 Denotation and Connotation
subjects and attributes with which word-forms· are concerned.
To decide whether a word means some given thing we must see
whether some word which is acknowledged to have that meaning,
that is to, mean the given something, can be substituted for the
word in question in a sentence. The proper name John has to
do with individual subjects, men in fact. Such subJects are, as
such, subjects of attributes. What does the word 'John mean ?
How again is its meaning related to such subjects and such
attributes ? In a given sentence John means a particular man
who is a subject of attributes and is certamly thought of as such
in the use of the word. The word John does not mean any
attribute of the man, does not mean, for instance, the height
of the man nor his virtue. Again, John clearly does not mean
any group of the attributes as attributes, nor even all John's
attributes as attributes. It certainly means a subJect which has
such attributes as we know with an indefinite number of others,
but that is not the same as meaning the attributes.
To take an dlu'!ltrat1on from other matter, tlus clock means
the object I am pomtmg at, understood (as a clock) to have
a pendulum, weights, and wheels But this clock does not mean
the pendulum nor the weights, nor the pendulum and the
weights and the wheels ; it means the clock as the organism
which has these organic parts. Thus, to mean something which
1s necessarily thought of as having distinguishable parts or
elements is not to mean any such element C\r any group of them,
or even all of them as clements, but is to mean the nrgan1c
unity of them. Just so, then, of the elements d1stmguishcd m
a subject as 1ts attributes The word which means such a sub·
ject as John in a given i.entcncc means an md1vidual man as
subject of attributes and not the attributes as attributes. This
is then what the given word John docs not mean If then it
does not mean attributes, does it mean a subject ? It certainly
does, and it means moreover a subject as a subject. Here,
however, we must guard against a misunderstanding. The word
John does not mean a subJcct as abstracted from and dis•
tinguished from its attributes, but a subject as havmg attributes,
as being the subJect 1t 1s, having the attributes 1t has. But
a. subject as a subject, or as the subject 1t 1s, 1s a subject ~s
subject of attributes or as subject of the attributes it has. T-his
Jg:I STATEXBN'l", :fHtNKlNG, AND APPREBENSIOK
then is the meaning of the given word John, in the giwn
sentence. It does not mean attributes as distinguiahed from
1ubject, nor sub,ect as distinguished from attributes, but the
subject as the unity of the attributes.
§ 189. This seems also to be a complete account of the relation
of the attributes and subject to the meaning of the word, and
one doesn't see where any new terminology is required. For
instance, we can't correctly say that something (say attributes,
or say subJect) is directly meant and something indirectly. The
distinction of direct and indirect cannot be applied at all to
meaning ; the attributes, for instance, are not indirectly meant,
nor is their unity as subject. No meaning could be given to
the expression • indirectly meant ' m the case of the attributes.
John may be a person I know as red-haired and courageous,
but John does not mean the red hair nor the courage, either
directly or indirectly ; John means the man who has these
attributes.
§ 190. Nor, agam, can we find any use for primary and
secondary as applied to the meaning. For instance, it is not
true that the subject as opposed to the attributes, or the subject
as havmg the attributes, is primarily meant and the attributes
secondarily. Neither the subject nor the attributes are meant
at all in abstraction from one another. What is meant is the
whole of which the attributes are elements, and the whole is
meant as havmg these elements To say that because the
attributes are elements in the whole thmg meant they are
therefore meant secondarily is quite nugatory and nonsensical,
for the only explanation we could give of the phrase ' secondarily
meant '-and this would certainly need explanation before any
use could be made of it-would have to be that they are elements
in the thing actually meant and meant as having them. Similarly,
if we said that they are indirectly meant, this would be a tech•
nical phrase comed by us, and, to explain it, we should simply
have to repeat the above relation of the attributes to what is
really meant. We might as well say that the whole was seriously
meant and its elements playfully, and then explain our technical
phrase playfully by the above method. John, then, in the given
sentence denotes a subject (i.e. a particular subject) as having
attributes i for in the ordinary use of language denote is equi•
 393
valent to .mean and denotation to meaning. Thus the werdt
tl,uta and unoUltion are the only ones wanted, for all that i,
meant is that the thing to which the proper name applies is
denoted by the name. To use the clock as an illustration. This
clock would never be said in any sense to mean its pendulum,
whether indirectly or secondarily.
§ 191. The use of imply and implication will be found equally
fallacious in application to a subject and its attributes, whether
the opposition is between implication and meaning, or impbca•
tion and direct meaning. We might say, m the ordinary use of
language, that 'he remarked" there's the door"• had, besides
1ts strict meaning, the implication that I was to go. But this
is merely the distinction between the ordinary meaning of the
phrase and its meaning m a particular context. It is just as
much meaning in the one case as the other. Clearly such
a distinction 1s not here relevant ; we should never call parts
of the thmg actually meant, or elements of 1t, 1mphcatJons, the
thing itself being meant quite directly as havmg the parts.
When we call a thing a clock, we do not mean {or mean directly)
something by clock, whereas we only imply that 1t has wheels ;
we mean quite directly that it has wheels, for the word clock
does necessarily mean a machme with wheels. As before, if we
persist in saymg that the attributes are 1mphed in the name of
the sub3ect, are its implications, not its meanmg, we shall be
departing from linguistic usage and g1v1ng 1mphcat1on an art1•
ficial sense, to be explained necessarily as ::JO arbitrary usage.
Indeed we shall have to explam our use of implication by the
very distinction we propose to elucidate by it. This would be
entirely nugatory, and we had better com a new word altogether.
It will, at all events, be admitted that me.1mng and direct
meaning are complete, in the ordinary and natural use, without
the addition of implication, whether the latter be used to cxplam
a sentence like the above, ending' there's the door•, or to mean
the consequences necessitated by what is {directly) meant. In
the case we are studying, the subJect 1s meant as having the
attributes ; the meaning then is quite 1mposs1ble without
the so-called imphcation. Even if we allow such a distinction
aa that between directly, or explicitly, mean and imply, we
cannot say that when a word means a whole which is a unity
394 STATEMENT, THINKING, AND APPREHENSION
of elements and as a unity of these elements, it implies these
elements ; that which is an explicit element in what is directly
or explicitly meant cannot be an implication in any intelligible
sense.
I 192. We will next consider words hke heaviness, as when we
speak of ' the heaviness of this stone', or say that ' heaviness
is due to attractive force '. In a given sentence ' this heaviness '
certainly refers to the heaviness of a particular subject as such,
but does 1t mean it as such? This heavmess, when fully expressed
is equivalent to ' the heaviness of this body ' and, to be under•
stood at all, must be so interpreted. The expression therefore
means or denotes the particular attribute of a particular body
and as its attribute. The belonging to a particular subject is
a part or clement in what is denoted, or explicitly and directly
meant, an element therefore in the denotation. The given
expression, this heavmess, does not mean the p~ticular sub•
Ject ; that is not a part of what is meant by this heavmess,
but the words 'this body', which mean the subject, are a part
of the compound verbal expression, when fully set out, ' the
heaviness of this body '. Consider now the sentence ' the heavi-
ness of this p1ece of glass 1s remarkable '. Here the heaviness
may seem to mean a particular instance of heaviness, but that
is scarcely a correct interpretation. This heaviness conveys
meamng at once, but 1f we merely say the heaviness, and no
more, we cannot mean anythmg It 1s as if we only half pro·
nounced a word ; 1f we complete the utterance we know what
is meant. Wc must supplement our incomplete words, then, by
an addition bke ' of tlus piece of glass '. The heaviness of this
piece of glass, then, means a particular instance of heaviness,
whereas heavmess denotes and signifies the universal of the
given attribute. A similar account may be given of I the heavi-
ness' m the sentence 'the heavmess of gold is one of its remarkable
properties'; here 'the heaviness' refers to a universal and the
compound expression refers to the particular attribute of a
particular subject ; we are referring to a universal, but to
a species of the universal, ' such heaviness as this'. 'The heavi-
ness ' then is a word-form which is part of a complete word·
form meaning either the particular heaviness of a partic1,1lar
body or a particular kind of heaviness, according to the words
 Denotation and Connotation 395
by which ' the heaviness ' is supplemented. Apart from some
sentence in which it occurs, we cannot assign any meaning
whether of subject or attribute to the word-form.
Consider finally a statement about heaviness in general; for
instance, ' heaviness is due to attractive force '. It might be
argued that we here have abstracted all consideration of the
particular subjects which are heavy. If 'heaviness', then, meant
or denoted these subjects, such abstraction would be impossible.
Wishing, then, to make a statement about the attribute and not
about the nature of the subJects of it (except that they can have
the attribute), we use a word which means the attribute and
does not mean the particular subJccts. If 1t did mean these
subjects we could not, m using 1t, make abstraction of them.
But though we can in th1s sense make abstraction of the sub•
jects, we cannot make abstraction of the fact that the attribute
belongs to subjects. This therefore is not abstracted in our
thought of heavmess but must be explicitly present.
An attribute is an attribute of a subject. lleavmess is the
universal of an attribute of subJects ; how then 1s its meaning
related to these subjects ? In such a sentence as ' heaviness is
due to attractive force ', the abstract noun denotes or means
the universal of an attribute, as havmg a certain quality, and
as the universal of an attribute of subJects or as an clement m
a unity and as being necessarily such This refercnce to its
subJects therefore belongs essentially to the denotation of it.
So 111 ' the heaviness of th1s piece of glass', the word which
means a particular attribute must mean or denote it as belongmg
to a subject. The meaning or denotation is impossible without
the reference. This reference, then, does not take us beyond
the denotation ; it 1s entirely withm the denotation.
Clearly, then, it would be nonsense to say that the subject
is indirectly meant or implied. No one would naturally say that
the expression ' the weight of this body ' implied this body, any
more than he would say that ' yonder cage with a canary m it '
implied the canary. If he were pressed to say whether the
expression implied the canary, he would say, 'no, the canary
is directly mentioned '. If then it were pointed out that the
whole expression did not mean the canary, he might be puzzled
and induced to say that it was aftcr all in1plied, supposing
3g6 STATEMENT, THINKING, AND APPREHENSION
vaguely that anything which an expression did not mean, but
referred to or presupposed, must be implied, as the only alter•
native to msan. He might, however, give what seems the correct
answer, that part of the expression, though not the whole, does
mean the canary, and that therefore the canary is not implied.
In fact, if we employ the tecbuical word • imply ', we shall have
to repeat the explanation which has already been given of the
meaning of • this heaviness ', ' the heaviness ', or • heaviness ',
in a given context. Again, if we allow the term ' direct meaning ',
the reference to a subject 1s as directly meant as anything else
and, if we distinguish explicit somehow from 1mpbcit, the subject
is quite explicitly meant in the sense of being an explicit clement
in what 1s explicitly meant.
§ 193. Our conclusion, then, 1s that the word which denotes
or means the universal of an attribute or attributes of a certain
kind must denote or mean a universal which as such contains
the universal of such reference to a subject as an explicit clement
in itself. This reference is essential to the denotation, meaning,
or significance, and cannot be separated from 1t. We leave
nothing out when we say that the word heaviness denotes the
universal of the attribute or the attribute universal. The denota•
tton would be indeed nothing without the reference aforesaid.
It belongs to the denotation as much as anything else does,
certainly a& much as what we may call the quality of the
attribute denoted. There if> then no need \\hatsoever for
a word ddfering in sense from denote or denotation. We see,
moreover, that the denotation, meaning, or significance of these
abstract nouns cannot possibly be elucidated in the usual
manner, that ts, without reference to their use in sentences, and
this, 1t may be added, applies to all the words about which
logicians use this language of denotation and connotation. It
1s a question of word-forms which is before us and of the use
of these word-forms in the sentences of normal speech.
I r94. This becomes still more evident in the discussion of the
question so far as it concerns adjectives. An adjective like
heavy clearly has something to do with the attribute heaviness
and with the heavy things which are the subjects of that
attribute. By the theorists we are criticiz:ing, heavy is said to
de111Jte heavy things, the subjects of the attribute heaviness, and
 Denotation and Ccmnotation 397
to conn11tt this attribute itself. Now first it is clear that heavy
docs nDt denote heavy things, the verbal expression which does
that is just ' heavy things '. Heavy does not denote any sub•
jects. As then what is to be denoted or connoted is either
.subject or attribute, does heavy denote the attribute • heavi-
ness 1 ? Heavy can only refer to heavy things through their
heaviness, so that surely heaviness should be the denotation of
heavy. But clearly the adjective, heavy, does not denote the
attribute, the substantive • heaviness ' does that. This is so
obvious that the connotationists were driven into their theory
that heavy denotes the subjects, the heavy thmgs. But this it
certainly does not. It denotes then neither subject nor attri•
bute, and as denotation 1s either of subJect or attribute in this
theory it can have no denotation. We may say then that if
heavy does not mean or denote any subJect or attribute, and
if meaning tnust be one or other of these, heavy can have no
meaning al all. The conclusion should m fact be that since
words like I heavy' have no denotation, they have a fortiori no
connotation. They can, .at least, only have connotation, and
since the only definite thing that is said about connotation 1s
that it is implied meaning, or words to that effect, adJectives like
'heavy' \\'ould be words not with direct but only with implied
and secondary meanings.
§ 195. Now clearly a word must mean something, and we have
therefore to consider what we mean by meaning. Ordinarily,
to say that a word A has a meanmg 1s the same as to s:.y that
the word A means something or that A means B. Of what
then can we say that 'heavy' means this) It might be answered
that 'heavy' means 'having heaviness', but that would of course
involve a confusion with another sense of meamng. One form
of words is often said to mean another, when the other 1s a verbal
equivalent. Thus I discomfort ' means uneasiness, and arbor m
Latin means tree. But the thing meant, which we are concerned
1
with here, is the thing which the word symbolizes. Discomfort '
is not the symbol of the word 'uneasiness'. Heavy is only the
adjectival form which 1s the word equivalent of the participial
form • having heaviness•. It does not, of course, mean ~he
objective fact symbolized by the noun form '(the) havang
heaviness'.
 398 STATEMENT, THINKING, AND APPREHENSION
t 196. The right and only method then is to inquire what is
the use of the word• heavy' in a sentence. The obvious answer
is that the word• heavy' is the word-form {adjectival) which
we use in a sentence when we desire to state that a given specified
individual subject has a given attribute, or when we want to
represent, in general, such a subject as having the attribute.
The first we do by taking the noun which means the particular
subject as nominative to is or some similar word, and adding
after it the adjective 'heavy', 'this piece of glass is heavy'. The
second we do by representing in the sentence a subject as having
the attribute, though not stating that it has it, by adding the
attribute adJective to the noun without a verb, as 'this heavy
piece of glass has fallen '. This we might express in a kind of
shorthand by saying that the adjective is the word-form for the
attachment of the attribute to the subject. The word-form, the
adjective, is then necessarily relative in its use to an expression
in words which has meaning, in the sense that it is an integral
part of such expression in words. The expression in words
contains a noun which means a subject of the attributes to
which the attributive refers in one or other of the manners already
described. In the first case 'this piece of glass is heavy', the
expression 1 1s heavy' of which 'heavy' is an integral part, cannot
accurately be said to mean the subject possessing the property,
for an expression having that meaning could not contain the
verb. What it normally means is that some one thinks that
the particular subject has the particular property. In the other
case, 'this heavy piece of glass has fallen', the expression of
which the adjective 1s an integral part has meaning, and it
means a particular subject having the attribute heaviness. The
relation of the meaning of this expression itself to subject and
attribute, which is our general problem, is given directly by
this account of it.1
I 197. We can now state the grammatical function of the
1
We are not here concerned to elucidate all the grammatical forms, though
the subject under discussion can only so be completed, we note, however,
that the verb with the ad1ective does not mean the object to which 1t relates.
In ' this can leaks ', leaks does not mean • the leaking of the can ', but these
last worda themselves mean that. The verb m the sentence means that
the obiect to which 1t refers Jq taken to ex1~t See further on this
subiect I 76
 Denotation and Connotation 399
abstract noun hla'lli111ss and distinguish it from the function of
the adjective heavy. When we wish to name a particular subject
which has an attribute or to state that it has that attribute,
we use the adjective along with the noun ; when we wish not
to state that a given attribute is found in a given subject, but
to consider the attribute without considering the nature of the
particular subject to which it belongs, or rather without con-
sidering what particular subjects it belongs to, and to state
something about that attribute itself, and not about the parti-
cular subject it belongs to, we use the abstract noun corre•
sponding to the given attribute, because it has the meaning
above described. We then make it, by suitable linguistic devices,
the subject in the logical sense of subJect We can of course
use artificial forms of expression and say 'heaviness is in this
piece of glass ', or ' this piece of glass has heaviness ', but the
normal and proper expression is 'this piece of glass is heavy'.
Again, ' this heavy piece of glass has fallen' is the normal usage
and not some artificial substitute like ' this piece of glass which
has heaviness has fallen '. 1
§ 198. Consider next such concrete general terms as ' tree '.
The word tree does not mean some particular tree or trees and
therefore does not mean subjects, nor does it mean any attribute.
Thus tree denotes neither subject nor attribute Nor again does
it denote the universal of a subJect, for it does not denote tree•
ness, nor the universal of any attribute, nor of all the attributes,
of a tree. Thus, from the pomt of view about meaning a1ready
criticized, tree has no meaning and no denotation In sentences
it is used with the definite or indefinite article or with some
word like' this'. With the indefinite article' a tree' may mean
a particular tree, as in ' a tree stood near the house '. This
meaning, however, it often has only as qualified by some other
words, e.g. ' a tree which had been marked for cutting down'.
Here ' a tree ' means a particular subJect as having attributes,
and the same account of it in respect of denotation may be given
as we have already given for proper names. With a qualifying
adjective, e.g. • an elm tree ', ' a tree' may stand for any tree of
a certain class of trees. Again in such a sentence as ' a tree
gets nourishment partly from the air ', 'a tree' stands for any
1 On these artificial forms see I 82,
400 STATEMENT, THINKING, AND APPREHENSION
tree. In the former case, when • a tree ' stands for a given
particular tree, we observe that the word ' tree ', not itself
meaning the particular subject, is part of a compound expression
'a tree' which does mean a particular subject.
In the second case we have something stated about all trees.
Here 'a tree ' does not denote a given particular tree, for then
the proposition would be true of that tree only, nor does it
denote tree universal. For ' treeness ', the universal, does not
get nourishment Nor does ' a tree ' mean or denote ' all trees ' ;
for a. word meaning all trees should either be followed by a plural
verb or, if followed by a singular, should mean all trees as an
organized group, and only as such. But it is not meant that
trees as a group get nourishment, but that each tree separately
does. Nor can we say that 'a tree' here denotes any tree, though
1t 1s the equivalent in words of 'any tree'. For, if 1t denoted any
tree, it would denote this one which I have in front of me now,
and, if it denoted that tree, it would denote no other. The
article is indefinite, so that no particular tree is named or
specified, and the statement is understood not to depend on the
particularity or peculiarity of any given tree. Thus the state•
ment 1s equivalent to a statement about a universal. We
must indeed recognize that we have here to do with a special
idiom, a linguistic device for representing a universal statement
•by one that 1s singular in form and is constructed exactly as if
a wee meant or denoted a given particular tree.
§ 199. These considerations show that the ordinary logic
manuals (perhaps all treatises of logic) have been written with•
out the nature of the problem, which is approached with the
formula of denotation and connotation, being in the least
realized. There is no sign of a sense that the problem requires
careful and detailed analysis, and in consequence the writers
are unable to say anything which is not hopelessly confused of
the denotation and the connotation (as they are pleased to call
jt) of a word bke animal or wee. We cannot really ask what
a noun or adjective denotes, or whether a noun denotes a subject
or an attribute, wjthout going into grammatical details. Indeed
the whole investigation seems properly to belong to grammar,
not to logic, and the need of it 1s seen in grammar jn the defini•
tions given of the ' parts of speech '. Certainly it doesn't seem
 Denotation antl Connotation 401
to belong to lpgic as logic, nor is any use or any application
made of it in the rest of what is called logic, For what we
account the main problems of logic 1t is clearly of no use what•
ever. It does not, however, follow that it 1s not important to
the logician to investigate and understand the use of grammatical
forms. He deals with thought through its expression.
I ioo. There is another way of putting this distinction between
denotation and connotation or of applying 1t. This way, which
has not necessarily for its basis the distinction of subJect and
attribute, is that in which the distinction 1s most often met
with perhaps in ordinary philosophic discussions. The members
of a class are said to be denoted by something or other, and
that in virtue of which they c1,re members of the said class, their
common characteristic, 1s said to be connoted by the same
something, This common chara.cterist1c may be an attribute,
as ' heaviness ', but it may be something which covers the whole
nature of the particular, as blue or blueness covers the whole
nature of this blue colour. Here it is not natural to call blue
or blueness an attribute of the blue colour. It would probably
be called the ' quality ' of the particulars. In any case the
common characteristic 1s a universal, and the universal, whether
quality or attribute, is apparently what is supposed to be
connoted and its particulars denoted.
§ 201. What then 1s 1t that connotes the universal of a class
and denotes its particulars ? Sometunes 1t would be said that
the ' concept ' had connotation and denotat10n, and som~times
the 'term'. Under the confusion which attaches to the word
• concept ', 1t seems at least that those who use it mean by 1t
some sort of thought. If then denotation and denote hc1,d the
meaning they have in ordmary language, a concept could not
possibly denote anything. Many writers mdeed speak of a
• universal concept ' and not of a ' universal ', and treat the
universal as if it were necessarily a concept. If then the
universal is what is connoted, it would be the universal concept
which 1s connoted, not the universal concept which connotes.
According to the traditional phraseology, however, the right
expression seems to be the denotation and connotation of terms,
that is, of certain words. We have here to do with wo~ds, m
fact, not with thoughts. But what term connotes the umversal
1773•1 od
402 STATEMENT, THINKING, AND APPREHENSION
and denotes the particular? Is it the noun 'heaviness' or the ad-
jective 'heavy'? Is it the noun blue (that is, blueness), or the
adjective ' blue', as in the sentence ' this colour is blue' ? Is it,
in the class of trees, the noun 'treeness' (an artificial term) or
the word ' tree ', which is treated always as a noun and called
a concrete general name ?
Now ' denote ' and ' connote ' are either technical terms,
whose meaning is not to be interpreted from the ordinary
meaning of denote at all, or else denote means here what it does
in ordinary usage. The former hypothesis seems ridiculous
enough ; for, if we adopt it, it is absurd to employ a word
' denote ', which already has a significance, with no reference
to its accepted significance. The second hypothesis then is the
natural one, and though we can say nothing about the meaning
of ' connote ', we can use the meaning of ' denote ' to answer the
question before us. 'Blueness' does not denote the particular
blue before us but the common quality of all blues, and so for
'treeness ', 'animalness', &c. Further, since to connote is at least
not to denote, and' blueness' and 'heavmess' denote the universal
quality or attribute, they cannot connote this universal. There,
fore on both grounds it cannot be 'blueness' which denotes this
blue colour and that blue colour, and connotes the blue of them.
'Heaviness' cannot connote weight and denote the heavy things.
' Blueness' does not denote the blue colours ; if any word can
denote them, it would be just the compound word-form • blue
colour'. Again, heavy does not denote heavy things nor a
particular heavy thing, as we saw. Thus the only nouns and
adjectives which, according to the theory, come into queiltion,
cannot denote the individuals of a class. It looks as though
those who use this phraseology {applying 1t without thinking
about it or discussing the distinction} think of heavy things as the
denotation of 'heaviness', so that the connotation of 'heaviness'
would just be the universal it means, that 1s, what it denotes ;
the difficulty is, however, concealed by some other verbal account
of the meaning of 'heaviness', by saying, for instance, that • plant•
ness' connotes 'life and growth'. This, however, is an impossible
use of 'denote', for we have seen that 4 heaviness' does not denote
heavy things.
t zoz. On the other hand, Mill makes the words which denote
 Denotation antl COffflotation
$Ind connote, the adjectives' blue', 'heavy', &c., and the common
terms, 'plant', 'animal', and 'man'. Here what Mill really
means by the tknotation of heaviness is simply the particulars of
the universal heaviness, which are designated by an utterly false
and anyhow useless technical terminology, for they are not the
denotation of the word 'heaviness'. His use of connotation is
Jimply a perverse and falsely phrased way of saying that the
word ' heavy ' 1s attached by the word ' is ' to each heavy
object referred to; the words rneanmg the heavy objects being
nominative to the verb ' is ', which verb is followed in the
sentence by the adjective ' heavy ' because of the common
quality ' heaviness ', or to indicate that they have the common
quality heaviness. In either case, if we take' denote' as having
the only meaning 1t Cd.n properly have, there is no word at all
which has the individuals of the class for its denotation ; no
word, that is, of the only nouns and adjectives which are sup•
posed to come into the question. Whatever, then, connotation
may be, the distinction is an absurdity since the proposed
application of denotation cannot be made to the words in
question.
§ 203. The distinction then can be maintained only by
adopting the former hypothesis and by making denote and
dlnotation have a meaning which is not the meaning of these
words in ordinary speech. But, then, what can the words
mean? We can only guess from the instances given. The
denotation of the term heaviness would mean the particulars
of the universal which the term 'heaviness' means or denotes.
Clearly the terminology is not of the slightest use. The con·
notation of the term heaviness would be just the universal which
it means or denotes. It would be as useful and instructive to
call the one abra and the other cadabra. In the case of adjectives
the denotation of heavy would be the particular heavy things,
to the words really denoting which the adJective is connected
in a sentence by the word ' is ' in order to state that they have
the attribute heaviness, and thus the connotation of 'heavy'
would be universal' heaviness' (and similarly when the universal
is not an attribute, but a quality in the sense explained). A little
more sense might perhaps be got into the tenninology by inter•
preting it to mean that •heavy' is technically said to denote Jiea vy
Ddi
404 STATEMENT, THINKING, AND APPREHENSION
things, because it is pa.rt of a word-phrase which does denote
these subjects.
§ 204. It is waste of time to ask, on this view, whether a proper
name has a connotation, for the distinction could apply only to
. an adJective or a common term. Obviously we could not repeat
the above formulae for the proper name, which therefore could
have neither denotation nor connotation. The same holds of
the first case. Observe further that this can be a matter of dis-
pute only because the disputants are not really agreed on the
defimtion of these terms. The problem therefore is not how
the terms should be applied in a particular case, but whether the
distinction is a legitimate one at all.
§ 205. An explanation of the interest taken in the question
whether a proper name has a connotation may be hazarded.
The question is perhaps confused in practice with a real and
important question of which it 1s a mere parody. This is the
question whether the apprehension of a particular involves any·
how the apprehension of the corresponding universal. The
question we have been examining is solely a question as to what
a word, or group of words, actually does mean. ' This house,'
for instance, docs mean the particular as an instance of the
universal because of the general term 'house' ; the form of the
expression shows this. Thus if what 1s to be connoted is thought
of as a universal, those who ask whether a proper name has
connotation are asking whether we can think of the person John
save as a particular of a universal. But the true question (smce
particularity 1s drstmguished from umversahty) is whether the
name itself means the reference to the umversal or whether the
meaning of the name includes this reference, e.g. d a dog's name
is' Shot'.•
§ 206.b The meaning of connotatio in medieval logic is an
important point which some modern interpreters, I suspect, have
missed and so have got mto confusion. In the first place it is
enough for the purpose to note the general meaning of supponn-1.
[• I thmk tho dog's name was intended to be Slgllificant apart from its
denotation. My own dog's name,' Shot', 1s substituted aa preferable m tbia
counexion to Wilson's Cf. p. 203.
b §§ 206-7. Redrafted from a letter, g.iJ. o6, to C. Cannan and from
Wilscm's manuscnpt (Diacu88iOl18, No. I). Cf. An Introdtldio• lo Logie, 1
H. W. B. ]OIOPh, pp. 140-2.]
 ·Denotation and Connotation
If A and Bare so related that we can say 'this A is B ' 1 then
B is sa.id supponere p,o A. • Quandoque concretum aliquam
rem significat, vel connotat, sive importat, sivc dat intelligere,
pro qua etiam. suppottit, quam abstractum nullo modo significat
nee aliquo rttodo supponit pro eadem ut " iustus " et " iustitia "
et similia, nam iustus verc supponit pro homine quando dkitur
Iustus est virtuosus, non enim potest supponere pro iustitJa....
Iustitia supponit pro qualitate et non pro homine.' 1 The idea
of connotation is ultimately based upon the distinction of sub•
stance (not mere subjert) and attribute. The correlative of
a connotative term is an absolute term. An absolute term,•
or a merely absolute term, is one which has a primary meaning
(significat primo, pr1mario 1 prmcipaliter) and no secondary
meaning (really the name of a substance as the fuller context
in Ockham seems to show). A connotative term is one which
has both a primary and a secondary meaning (secundario), such
as album (also intellectus, potentia, &c ). Observe that he gives
this as what the scholastics frequently employ. This 1s eluci-
dated by a passage from Albert us de Saxoma, an Ockhamist, 3
and one from Pierre d'A11Iy,' who 1s m general agreement with
Ockham, but it ts sufficiently elucidated by the fuller context
of Ockham. 6 D'A1lly's words show that pro quo supponit albus
is the white substance, e.g. this man. The double meaning of
albus depends on its reference of an attribute to "' substance,
and the two meanings are the substance and the attribute. One
would naturally suppose that the primary meaning of album was
albedo (whiteness) (album 1s in one place said clearly to signify
or mean albedo), and the secondary the white obiect, but the
words of Albert suggest that 1t was just the other way, \iz. that
al.bus means primarily id pro quo suppomt, that is, homo, or
1 Ockham, Summa totius logicae {Oxford, 1675), Pars I, cap 5, p 9 Quoted
in part by Prantl, Geschicllte der Logik im Abendlande, m, p 16l, n. 827.
• Ockham (Prantl, I c , hi, p 36i, note 831 ), Summa, Pa!"'I I, cap x, ad init
quae non sigmficant ahqu1d pnnc1pahter et ahud vel idem <iecundario
• Prantl, 1v, p 62, n 232, where Albert 1s not d1scuss1ng connotat1on,
• Petrus de Aylhaco (or de Alyaco), Prantl, 1v, p 109, n 459
, Ockham, 1. c., Pars Prima, cap x, pp 21 and 22 (cf. Prantl, m, p. 364,
n. 83J). Though Ockham speaks o.f the d1stmcbon a.i, common with the
Scholastics, Prantl knows nothing earlier than the passage he quotes from
Duns Scotus, gu. sup An. fw I. 16 (Prantl, m, p. 134, n 598) and quotes
notlung between Duns Scotus (ob. 1308) and Ockbam (ob. 1347).
4o6 STATEMENT, THINKING, AND APPREHENSION
hie homo, because you can say correctly hie homo ,st albus i
thus the white substance is its primary meaning. Its secondary
meaning is whiteness, connotat aliquid quod non est de essentia
sirnificati per subiectum, ut ' homo est albus' ; .•. ' homo e&t
rationalis ' . • . ' rationale ' connotat formam substantialem
'hominis.
The two passages do not, as a matter of fact, introduce
primario and secundario, but taken with the passages from
Ockham naturally seem to imply the above interpretation.
But it must be observed that, for a reason which will appear
immediately, the fact that the attribute is said to be connoted
does not prove that it was the secondary meaning. The full
context of Ockham throws no light on the question which was
primary and which secondary.
Now besides the adjectives albus, iustus, the nouns albedo,
iustitia are connotative, and for the same reason. 1 The question
therefore arises : What 1s the primary and what the secondary
meaning of albedo and iustitia which is connoted ; and what
1s that pro quo suppomt such an abstract term ? The answer is
to be got from two very important passages m Ockham.• The
abstract term ' whiteness ' ' suppomt ' for a quality or attribute
(qualitas accidens vel forma) in the concrete thing. You can say,
that is, this quality in this body, or this colour in this body, is
whiteness; whereas album 'supponit' for the thing (hoc corpus).
E converso, as he says, an abstract term hkc ' ignis ' supponit
for the substance, a particular fire, presumably. 8 Also a con•
crete may supponere for a part of the whole substance. The ex•
ample given is that anima is a part of man, and supponit pro parte
cius ; homo est animatus et ita animatum supponit pro homine
toto. The two meanings which make albedo connotative (like
albus) are the white thing and its white colour (aliquid habens
albedinem and albedo). The question as to which is connoted
implies a mistake. In the nature of the case it is possible that,
1 Other instances in Prantl, m, p. 364, n 831, potentia, actus, intellectua,

• Prantl, iu, p. 363, ns. 826 and 827.

1 Ignis is thus called abstract both in the letter to C. Cannan and in WiJaon's

MS. notes. He is referring to a passage not m Prantl (111, p. 363, n. 827)

• nam ignis supponit pro subjecto, et 1gneus quod est concretum supponit
pro forma vel accidente eiusdem. Dicimus en1m quod calor est igneus, et
non ignis • (an adJective stated not of the subJect but of part of it). Ockham,
S11111t1U1, Pars I, cap. v, p. 9.
 Denolation anti Connotation
since the tenn combines the two meanings, 111th of these should
be said to be connoted. This was the actual fact. This is
proved quite clearly by two pa.ssages, one from Ockham's Quot•
libna, 1 and one the passage beginning ' quandoque ' quoted
above.1 The latter means that album connotes the white thing,
and we see from the other passage that it also connotes
whiteness.
This throws light on what would otherwise be a difficult
statement in Thomas of Strasburg, 8 Scientia connotat respectum
ad scibile ... et ideo non ohstante ilia connotatione, qua cernit
respectum ad scibile. . . . Thomas is thinking of science not as
an attribute of the man who knows but of the sc1tum ; scientia
alicuius, like quantitas alicuius, is connotative. This is quite
in agreement with Ockham. Scienba refers, hke all connotative
words, to a subject and an accident or attribute of it, and
connotes the subject just as whiteness does. It also connotes
the attribute. We may compare Ockham's account of intellectus
as a nomen connotativum. He gives 'these kinds of connotative
terms:
(1) all nomina concreta primo modo dicta;
(2) nomina relativa, e g. simile;
(3) pertinentia ad genus quantitatis ;
((4) unum, bonum, intellcctus, &c )
Thus, for connotative words, their significare = connotare, and
the answer to the question, whether the attribute or the subject
is connoted, 1s that both are connoted. There appears no trace
of the clumsy modern distinction of denotation from connotation
in medieval logic, though there is possibly something like it in
Thomas of Strasburg · 6
' Sic potentia generativa simpliciter et principaliter dic1t divi•
nam essentiam, quamvis connotet respectum ad gemtum, puta
paternam proprietatem, qua connotattone non obstante, simpli-
citer loquendo potentia generandi debet dici cssentia divina, sive
res absoluta 'et non relatio, nee debet dici esscntia ct relatio,
sed magis essentia sub relatione.'
' Prantl, iu, p. 363, n. 826 • Supra, P 405
• .Prantl, iv, p. s, n r7.
& Cap x {Oxford ed ), p 22, 9 lme<i up Cf Prantl, iii, p 364, n. 831.
1 Thomas ab Argentina, Comm. in IV l,bros senl Genua J 585, L1b. I,
Dist. VII. Qu. 1. Art I Concl 3, fol. 49 v, after the unportant passage
quoted m Prantl, 1v, p. 5, n. 17.
 4o8 STATEMENT, THINKl~G, .AlfD APPREH1£N'SION
The passage is not quite clear, .but it seems to show tha~ -the
primary meaning of potentia {which is a. conaotative word
according to Ockham, or would be with some thinkers according
to Ockham) was the quality in the divine substance, just as
scientia is said, simplic1ter loquendo, to be a qualitas absoluta.
· There is therefore a complication introduced, for, in the case of
scientia, we have besides the proper 'subject' of it, the mind,
the 'object·• of it, the thing known. The first two elements of
meaning are all we have had hitherto in reference to connota•
tion. The new element in Thomas doesn't seem to have occurred
to Ockham as connotatum, and thus the last passage is not so
valuable as 1f he had said : Scientla connotat respectum ad
animam, and, potent1a generativa connotat respectum ad gener•
antem. It must further be noticed that Thomas 1s not talking
specially of this subJect of connotation ; his remark is incidental.
One thmg we certainly do get out of the passage is that he
called the • attribute signified ' itself the primary meaning of
the abstract noun corresponding. This confirms the view adopted
above.
§ 207. But the question remains as to which of the meanings
was primary and whether the same meaning was primary for
both abstract and concrete attributive terms. These logicians
would be likely to get into difficulties here, for they do not
appreciate the fact that they are dealing with forms of linguistic
expression, which are distinguished by the kind of purpose they
serve in a sentence. The grammatical form of the adjective
albus referring it to the subject homo (habens albedinern = albus)
might well lead to homo as the pr1mano s1gnificatum, and the
passages quoted seem rather to favour this view. But the noun
substantive albedo would cause a difficulty, for it would seem
unnatural to say that whiteness meant primarily the white object.
This feeling seems to be betrayed by Thomas of Strasburg in
the passage partly quoted above. It runs thus : ' Scientia
connotat respectum ad scibile, quamvis ipsa sit in prima specie
qualitatis, •.. Scient1a enim non dicit qualitatem, et relationem,
sed dicit qualitatem sub relatione : et ideo, non obstante ilia"
connotation,, qua cernit respectum ad scibile: ipsa tamen sim•
pliciter loquendo dicitur qualitas absoluta.' 1 Ipsa would natur•
1 It immediately precedes' sic poteatia •• &c., p, 407, above.
 Denola,wn and Connotation
ally designate a primary meaning, not a secondary, and the·
obvious tendency of such a passage would be to make the quality
or attribute the primary meaning of albedo and the white object
or, as Thomas, probably moved by a scruple, says, the reference
to the white obJect the secondary meaning. The resulting theory
that the primary meaning of abstract attributives is the attri-
butes is really not inconsistent with the v1ew that -the primary
meaning of concrete attributives 1s the subject. On the contrary
there would be the same principle in both, viz. that the primary
meaning is id pro quo supponit, for, according to Ockham, the
concrete supponit pro the sub1ect and the abstract, also according
to him, suppomt pro the attribute. Whether the view repre•
sented in Thomas was Ockham's, or whether Ockham formu•
lated any view on the subject of the primary meaning of albedo,
I cannot say, but there is a passage in Ockham which shows
a strong tendency in the direction of Thomas's view. He sayi,,
• Aliqua sunt concreta ct abstracta, quae sic sc ha.bent, quod
concretum significat ahquam rem et supponit pro ilia, quam
nullo modo abstractum signdicat nee pro 11la supponit : exemplum
est de justo et JUstitia.' 1 In the passage already quoted, 1
where he adds what it is that forms the pro quo supponit of
the abstractum, he uses exactly the same language. The words
nullo modo i.igmficat go, however, rather further and are not
compatible with making the subJect or substance the seconda.ry
meaning (i.e. significat sccundario). However, with him there
must be two meanings for albedo, for it must be connotative
like quantitas, potentla, actus, and jUstitid, If he thought it
out at all he might therefore have taken for the secondary
meaning, not the subject, but the reference to the &ubject, whict
is exactly Thomas's formula.
Sfocc the • pro quo supponit ' of a given predicate is connected
with the predicate by the verb est (hie homo est albus : hie
homo est homo habens albedmem), which seems to identify the
subject and the predicate, it is more natural that the pro quo
supponit should be made the primary meaning.
a Prantl, lll, p. 363, n. 826, from Quotl. v. qu, 9,
• Supra, p. 405, line 2.
END OF VOLUME l
 PART III
INFERENCE

u
 I
THE GENERAL NATURE OF INFERENCE
§ 208 THE main ob1ect of the part of logic which precedes
Inference 1s to study the forms and characteristics of propos1•
t1onc; and terms (or, as i5 sometnnes said, of Judgements and
ronct>phonc;) \\hen they have been arrived at, rather than the
manner 111 which they arc attained, though the latter may
require some consideration as sub•nd1ary to the main mqmry.
Wl• have given rcasonc, for prcfcrrmg to speak of Judgement,
op1111on, and apprd1en!,10n in this connexion, and have main•
tamed that the d1~trnct1on between 3udgmg and mferring is false
alld, if assumed without srrutiny, lea<ls to difficulties 1 Logic
is, however, also roncernc<l with the way m "luch we arrive
at Judgements 2 ,111d op1mons, not w11h every way-not, for
instance, ,oth the processes of perception-but with that which
1s c-allcd mfcrrmg
Before we begin the ia1tudy of logic we arc familiar with the
idea of mfercnrc It 1s an operation we conduct in everyday
life and m the !>Ciencec; \Ve do not learn 1t from logic nor did
1 Sec the critic1~m of this doctrine 111 Part II, l-h r, §§ 39-41

• On the use and abuse of the word Judgement see Part II, ch 2

Chapters 1-4 were part of the Logic Lecture~. though from time to time
given a& a ~eparate conr~e \\1th the chapters on Induction This accounts
for some confusion m the opening which was never properly reconSidered
The chapters were continually revised 1n manuscnpt note-books of pupil~.
and new &ect1ons added The &tructure goes back, however, to an early
period and retain& marks of what may be called the Kantian penod of Wilson's
thought The cntic,~m of the syllog1sm and of modem mathemaneal theories
(Chapter 7) 1s, however, as Wilson would llave left 1t m pnnc1ple Chapters
s-7 are from a separ<Lte cour~e on Hypothetical Reasoning Substantial
,tdd1tions and mod1hcations were made m 1906, I have supplemented them
from the or1g111a.l manuscnpt
Wilson nearly always put the minor premiss of a syllogism before the
maJor I have altered 1t throughout as he nowhere Justifies the breach of
con,ention, and it 1s very confusmg where, as in tlus part, the argument 11
largely formal
 The General Nature of Inference 4I3
logicians invent it, though they have often affected to teach it.
Words like conclude and conclusion, proof and prov~, are not
part of a technical vocabulary invented by philosophers ; they
belong to the inferences of ordinary life Although, then, some
of the technicalities of the schools do make their way into the
language of ordinary men, most of these terms are the familiar
product of a certain natural logic and come simply from the
fact that men have gradually been led to reflect upon the opera-
tions of their own thinkmg.
§ 209 If we consider imtanccs in which we should naturally
c;;peak of concludmg or mfcrrmg, we shall find that they are
not instances of expericncmg Our attitude in experience seemc;
to us to be mamly receptive, but in inference we appear to
exercise an originative activity either m the discovery of truth
or in probable conJecture It 1s tlus origmativc activity of ours,
as opposed to experience, which 1s one of the main charac-
teristics of inference
But we do not suppoc;e ourselves entirely originative. In the
first place, we do not tlunk that we originate truth but that we
discover 1L Secondly, we do not thmk that our method of
discovery, 111 the process of mferrmg, 1s entirely independent
and unassisted from without ; we thmk rat her that our activity
consistc; in denvmg the truth discovered from something already
known and ultunately from experience This brmgs us to
another main characteristic of mferring ; the knowledge it gets
1s derived from other knowledge. Jlence we again have the
idea that this kmd of kno\\ ledge ir;; m some way dependent.
Thus, from the c;ubjective c;;1de, inferred knowledge comes to be
looked upon as indirect, as not immediate knowledge but
mediated, while on the obJecttve side the inferred facts are
sometimes looked upon as dependent for their existence on those
from ,vh1ch they arc inferred We somet11ncs even tend to put
inferred knowledge on a lo" er level, somehow, than that from
wluch it 1s derived. Thrs 1s a natural outcome of the idea of
dependence. It happens that an inferred opmion may be called
a mere inference , and, with an imphcahon of defectiveness, we
may say of some kind of knowledge or opinion that it is only
inferential, implymg a want of something we think better and
should probably call direct apprehension of the subJect-matter.
B2
 INFERENCE
Thus, it 1s often c;airl, whether rightly or not, that we can only
infer the existence of other people and do not directly apprehend
their being. And here 1t must be pointed out that the deprecia-
tion of an inference as a mere inference seems to be confined
to cases where the inference 1s not a certain one but a probabihty
because of the uncertainty of the matter, so that the defect is
not here caused solely by the injerential character of the process.
It is otherwise m cases like our belief m the existence of other
people, for there we do not doubt the truth which we say we
arrive at only by inference, and yet we do feel a "anl of direct
apprehrns1on.
§ 210. Now this Ycry dir,sati:.fact1on may so react upon us
when '\\e reflect upon 1t that we may rnchnc to thmk that the
inference m such a case 1!! after all not really certain. We may
reflect that it is somehow not hkc mathematical demonstration
and suppose that therefore 11 1s not certam. Yet, to take an
everyday dlustrat10n, we should under ordinary circumstances
be sure from the exprcsc;1on of a man's face that he was angry
and show thl' firmness of our conv1ct1on Ly actmg upon it,
lOmm1tting ourselves perhaps m consl•qucnce to some serious
line of conduct. Herc we should naturally call our thmkmg an
inference. We inforred his state of mmd, where \J,e could not
have the direct knowkdge "h1ch he himself has of his o" n
emot1oni.
On the other hand, we fincl no such deprcciat1011 as this in
the <;c1cnces There t lw h1ghl•c;t Yaluc 1s given to what claims
to be proof, that 1s, to "hat 1s rightly inferred. Nevertheless,
m reficct1on upon the method of !ic1cnce, though the inferred
knowledge 1s not depreciated as bcmg mkrred, the 1dea, never-
theless, of a dependent character m ,,hat 1s inferred often
persists, a<; wl.'11 as the 1dl.'a of a ccrtam 111d1rectness and want
of immediate apprd1ens1011 1
Whether all these 1deas and tendencies are Justified or not
we !!hall be better able to judgl' when we have advanced further
m our study of the subJcct. We must at the threshold be
prepared for the poss1bihty that the problems we have indicated
may involve different ktnds of inference, and that the answers
to them may therefore be divergent. It might be a question,
' This sub1ect 1s resumed m ch 4, infra
 The General Nature of Inference 415
for example, whether there is the same directness, or even the
same sense of the word, in a probable inference about an object
of experience (say the inference that the pea 1s under the
thimble) as m certain geometrical proofs. The inference about
the pea and the thimble seems to point lo a real difference in
kmd between inferential opinion or Judgement and another
method of frammg a conclusion on the same subject-matter.
§ 2 I I. The most general account, then, of the problem of logic
111 Inference is perhaps this : to study the forms and the nature
of that act1v1ty of the nund by wluch we advance from given
Judgements or op1mons to new Judgements or opm1ons neces-
~1tated or made probable by the former, not by experience but
by some other operation of the mmd. Yet here also there are
d1fficult1cs Tlus operation "ould usually, perhaps always, be
c..illed tlunking as opposed to cxpencncmg ; yet there is often
co11s1dcral>lc vagueness ar, to what ' thought ' should precisely
mean, 1 and d1fficull1c-, .mc,c 1£ 1t 1s realized that percept1011 1s
somehow mt1matcly connected w1th what would 11c1.turally be
called thought, though the nature of that conncx10n has not yet
been cleared up or mvest1gated This at all events is true, thc1.t
the advance 111 knowledge wluch 1s meant 1s, as we saw, not
one made by getting some new perception to add to the given
statements from which in some sense the advance 1s made. For
when mfercncc 1s supposed to be clue lo the operation of thinking
\\<e find a tendency sharply to d1stmgmsh this thmkmg as pure
thought from both experience and 1magmation. Tlus is un·
doubtedly the c,isc \\-1th the formal log1crn.ns, who suppose that
formal rear.omng, which really means rcc1sonmg from the mere
forms of statement without any matter, proceeds entirely by
pure thmkmg.
Y ct imagination 1s ab~olutely necessary to the processes of
those sciences which seem to be the most perfect type of what
can be produced 111 the way of new and ccrtam Judgements by
inference. Geometrical thmkmg, ancl indeed all mathematical
inference, is impossible without 1magmation, and 1t is even true
that the abstract reasoning treated of in formal logic (which
consists of syllogisms m which the premisses r.eem to be mere
forms of statement and can therefore be expressedsymbohcally),
1 Pc1rt I, ch .l
416 INFERENCE
reasoning which is often taken as the type of pure thinking, is
quite impossible without the use of imagination. Again, imagina•
tion depends upon experience, and thus inference stands in
essential relation to cxpcnencc. But notwithstanding this con•
nexion with expcrwru e and imagmat1on, it still remains true
that what 1s really meant by inference 1s that the new knowledge
we reach '"c do not rcad1 simply by getting fresh experience.1
In view of these very real d1ffieult1es the best way to form
a corrcc.-t 1dt•,i of mfcrencc appears to be to study first those
inferences "luch arc certain and "hich constitute knowledge,
or wlm h at lea.st we suppose to be ccrtam, for it seems obvious
that "e can, in a given subject, only understand the imperfect
type from a cons1derat10n of the perfect We shall accordingly
bcg111 v. 1th a consideration of this kmd of mference
§ 212. It will be observed that the above general account, m
common with the m,ual defn11t10ns of mfercnce, has an important
presuppos1t1011 which 1s not alv. aye;, 5uffic1cntly reflected upon
Thought 111 I lus mfomng process 1s spoken of not as .,bsolutely
ongin,it111g truth,• but .1s !>tartmg from some truth, whether
that Is given m experience or not. I11 the tccluucal language
of logic the Judgements or optn10ns from ,, luch the thmkmg
process thus starts arc ralled premisses The Judgement or
op11110n to wluch thought advances by 1ts own operation is
called the co11clus1011 Now It 1s obvious that the concluston
must be different from the prcm1ssec; ctnd 111 some sense really
new Thr propo<;1tion" ' .111 A ts B ' and ' some A 1s B' arc
d1tforcnt, but the latter 1s not new 1£ \\e have the former. That
this should be so t'> self-ev1dcnt, but that it ts ctlso recognized
explic1tly m log1cal tre..1.t1scs 1~ shown, for mstancc, 111 the obJeC·
tton to the syllogism that 1t 1s a petitio prmcipii. For the
mcanmg of th.it t'> that the conclus10n docs not seem to be new
.is compared with tJ1c ma1or premiss. It should be observed
that tlus demand is made even in an empirical logic hkc that
of Md! b He will not allow an argument to be a true inductive
1
E, 1de11cc of the reallr distinct character of inference as a form of appre-
henb1on may be drawn from the hypothetical statement Tht-re we have at
an} rate knowledge or op1mo11 e"pressed which can be got only by mfcrence.
Cf, §§ 102 and 298
(& Cf Ill) note t~ § 39
b ' Coll>e:. of inference 111 the proper .i.cceptatiou of ihe term, ihoi.e m wh1U1
 ~ The General Nature of Inference 417
inference unless it contains more than is contained m any single
premiss. Now this involves a presupposition which he has not
reflected upon • it implies that the new knowledge is not the
result of experience and must therefore be due to the inferring
process itself. Thus, the conclusion is unavoidable that in some
important sense a mental process which ls not experience can
originate knowledge It is futile to object that the mmd merely
works on the material wluch 1s given m experience, for this
unphes that we are able m the process to get on to new know·
ledge not 111 the material. This then must be due to the mental
process which brmgs the new result. Such or1gi11at1on con·
trad1cts the very foundation of an empmcal philosophy hkc
that of Locke and 1\1111
Here we arc not so much concerned with this criticism as with
the hght "Inch 1l throws on the important principle presupposed
111 the ordmary idea of 111ference, m so far as we find that 1t
forces 1t5clf even upon those philosophers whose doctrines make
1t, one would have thought, 1mpos'l1blc to admit 1t.
§ 213 Inference 1s usually di\ 1ded mto mediate and unme•
<l1c1.te. To Judge from the mstancec; by which the d1stmct1on 1s
illustrated, the idea at bottom of 1t r,;ecms to be that in immediate
inference "e pa,;s directly from one prcmi~s, 1 e. from one given
judgement or op1mon, to the conclusion by a mere reflection on
the given premiss. In med1c1.te inference something comes
between a g1Yen prmuss ancl the conclus1on, that something
being another judgement or prenuss Thus, 1mmed1c1.le mfcrencc
has only one premiss, mediate has at least two We may
represent this mediation m another way. In a given premiss
1t would be said ordinarily that we have two concepltons in
a certain relation To be more accurate, the premiss states a
relation bet\\cen t,"o obJccts of apprchem,1011 1 The inference
is 1mmed1c1.tc 1f "e arnvc at some other rdation merely by
cous1dermg the given c.oncepbons themselves, by considering,
that is, what we apprehend in thc'>e two obJccts m the act of
apprehension rl!prescnted by the given premic;s. It 1s mediate
1£ we get a new relation by the help or mediat1011 of some new
' See Part II, ch. 14, on the use of the term conception
we set out from known truth, to arrive .i.l other~ really d1&tinct from them,
SJ,stem of Logic, II. 1, § J ]
 INFERENCE
conception not contained in the premiss ; by the help, rather,
of the apprehension of something else not contained in our
previous apprehension.
Now clearly this latter can only be done by relating the given
conception (so called) to this other conception, and such acts
of relation are Judgcmenb or opmions, and hence we get more
than one prenus'>. We get exactly two 1£ we represent the
proce:,s as follows we find ourselves unable in a matter of two
concepl1ons, on the strength of the knowledge wluch makes
them what they arc for us c1.t the moment, to relate them m
some particular way. Thus (we should say), we cannot connect
them directly. We then mediate their connexion by a new
concepl1011 to which each 1s related . each relation 1s the matter
of a Judgement or op1mon, and thus we get l\'-O premisses.
§ 214 The above .H·counl of mediate and 11nmcd1atc 111fcrencc
1s not offered .t'> c1. <,at1:,factory definition of what such terms
ought lo mc..1.n, or even as 1mply1ng that mfcrcncc l'> properly
:.o di\ Hled It is mtcndcd rat her ,1,; ,111 e~planat1on of what the
lrad1l1011al d1stmet1011 amountl> lo Thub, if ,, c find 1t stated
t'iat an 11nmed1c1I c mfcrenec 1.., one m wluch a Judgement follo,,s
1mmedi,ttcly from another Judgement, 1t might well seem a more
natural construction to put upon the cbstmctlon of mediate and
1mmed1ate inference to say that 1£ a Judgement A necessitates
another judgement B directly, that 1s 1mmcd1ate mference , and
1f that Judgement B 10 turn neceso;1tatcs another Judgement C
directly, the mforence from A to C t!'> mediate, because the
connexion of A and C 1s only acquired through B. But that
1s nol the tra1ht1on,,l ..,em,1• of mctl1,Ltc mfcrenee, for, m the
ordinary deduct1,e logic, the '>yllog1sm 11, the m,L111 type of
mediate mfcrcncc, and such a. defu11t10n doer,, not sutl the r,,yl-
log1sm bcc,tusc m the i;,yllog1s111 the hrst prcnuss 1r,, not !'>upposed
to necessitate the second. In the end we shall depart so much
from trad1t1on as to show reason for calhng all mfercnce in an
important sense 1111med1ate. No doubt such a view seems at
first sight paradoxical, 1f we arc under the 1mprcss1on which the
account of 11nmcd1ate mference 111 formal logic naturally makes
upon us , for we !ihould not expect that such so-called mferencc
1s real mfcrencc <1.t all, but that immediate inference only gets
its 11.i.mc by c1. kmd of J.nalogy.
 The General Nature of Inference 419
§ 215. The syllogistic logic in treating of inference has only
before 1t the general form S 1s P or S is not P, with the quanti•
tat1ve distinctions all S, no S, some S, and this S. What it does
m effect in the case of 1mmed1ate inference 1s to ask, given
a certain judgement or proposition tn one of these merely general
forms, what relation can be inferred between the subject and
predicate conceptions (more accurately, between what corrc·
sponds to them m rcahty) or their negations, & besides those
stated m the given proposition ? Further, 1t is an element in
the problem as usually conceived that the conclusion of the
inference should be restricted to the form tn which the so-called
ongmal subJect and predicate conccpt1011s and their negatives
,ire to make up the subJect and predicate conceptions m the
conclusion For example, from all A is B, we may tnfer some
B 1s A, or no A ts not·B, but not, for 1m,tance, that A and B
are compatible; for the predicate ' compatible' does not occur lll
the premisses. Agam, the mfcrencc from all A 1s B to some
B is all A is not recognized as an 1mmed1atc mforence, because
.t rcstnctlou 1s made that the so-called predicate 1s to have the
ongtnal adJectlval form, or at all events 1s not to be preceded
by a quantitative word like all, or some
Such restrictions .1re clearly artificial, but so, too, 1s the whole
.1ccount of 1mmed1.i,te mfcrcnce. It ,., not meant that these
restrictions arc formulated and consuously made , they arc at
first simply the result of .i.n uncritical tr,1d1tion '1 be mfcrence
B ts les,; than A from A 1!. greater th.an B conforms mdeed
entirely to what 1c; essential m the dcfimt1on of immediate
mferencc a!. 1t ~eems to be understood III syllogistic logic, and
we note al">o that 1t 1s this kmd of unmedutc mfcrcnce \\luch
1s most frequent ,ind unport,rnt m ordmary ltfc, and m the pro•
ccdure of the science<; Accordmg to the dot tnne of the syllo•
g1stic logic, 1t would be said that m such tcl.'~es thr 1mmcd1ate
mference 1s from the matter of the propos1t10n, not from the
form, whereas m the 1mmed1ate inferences treated of the mfer•
ence is from the form and not from the matter.
Now even if this were true, as 1t 1s not, 1t would not be any
reason for neglecting the ' material ' 1mmcd1ate mferenccs, unless
it could be shown that they are comprised under various ktnds
l" vu, what Ioimal lugn, name& their wnlraditlory term& J
420 INFERENCE
of immediate inference from the form merely. Now clearly they
are not so comprised; the relation of A to B in the proposition
all A is B is neither the same as A is greater than B nor does
it include this relation, for m 'A is greater than B ', A and B
cannot be represented as subJect and predicate respectively, nor
as subject and attribute, nor could the given inference from
A 1i; greater than B be derived from rt by any rule of immediate
mforencc given m the syllog1st1c logic
Nor, agam, can these material mferences be represented as
syllog1st1c ; the attempt to represent them as such would result
f,Olely m i,,t.1tmg as a premiss of the given inference that 1f A is
greater than B, then B 1s less than A, and next m subsuming
under rt a particular A and a particular B; \\hercas of course
tt 1s not the propos1llon that tlus B 1s less than this A which
we are concerned with. The mference rs the alleged premiss itself
The fallacy of such verbal reductions to the syllog1stic form
will become more apparent "hen we criticize the traditional
reduction of ..i.11 demon-.trat 1vc rc,i-.0111ng to the syllogistic form.
There 1t will be shown that tlw; torm 1s but one among many
of a cerlam class and thcJ.t the other members of the class cannot
be reduced to 1t
\Ve can now sec what the general notion 1s '"luch ts common
lo both of these kmds of 1mmed1ate mference, and we can see
that the type of 1mmed1c1.te mfcrence studied m formal logic is
but one among many. The general type 1s, that from some
given relation bet" ccn two dement!> X and Y \\ c mfer directly,
,ind without the adrht10n of .my other statement about them,
some other n•l,tt10n behH!en X .111d Y. Kow the :.yllog1st1c logic
wnfine-. 1tsrlf to one only out of 111.tny possible rcl,itiou!>, that
wluch l'i u:,ually expressed by the term pred1cJ.t1011 The term,
howc, er, 1s u.,cd uncnt1cally, for subJect and predicate do not
here mc,m log1ral '>UhJcct .111d logical pn·d1catc proper. In this
refereme, the propos1t1011 I!> &upposed to have the form S 1s P,
or Sis not P, J.nd 1t 1s meant that in this form P 1s the predicate.
Now the true relation of S and P (the rel<1.t1on which the syllo•
gisttc logic has here m view) 1s tlus: P-ness 1s a kmd of being
which S, the !>O·callcd subject, 1s stated to have or not to ha_ye.
The relation therefore 1s an obJcct1ve one. 1 But any other
1 P.Lrt 11, § 68.
 The General Nat1,re of lnferc1tcc 4u
relation between two such obJects upon which an immediate
inference can be grounded 1s equally entitled to a place in any
theory of immediate inference.
To return now to a previous pomt, it might be said that our
ability to mfcr 11nmed1ately from the statement A is greater
than B that B is less than A depends, as it really docs, upon
our knowledge of the mailer of the relation, in other words, on
our knowledge of the special character of the relation greater
and less, whereas the other kind of inference depends upon the
general form of the Judgement or statement without any matter.
Now tlus would be a false d1stmctlon. The one kmd of inference is
as material as the other Just as from the statement A 1s greater
than B, our inference that B ts less than A depends upon our
knowledge of the relation of magnitude, so also our inference
&ome B 1s A from ' all A is B ' depends upon our knowledge of
the nature of the spccul obJcct1ve relation which has here,
through a confus1on, got the name of prcd1catlon (which pro•
perly drs1g11ate, ,t merely subJcct1vc rclal10n). Moreover,
.,., w<- shall sec hereafter, the rules of the ,yllog1stic logic
<lepend upon Lile special nature of till', rel,1t1011, and, m tlus
sense, arc as much material mfcrcucc as .i.ny other. The
'!UbJecl of 1mmcd1ale mfcrcncc 1c; treated m the ordmary
logic mo:1.nuals under such heads as conversion and opposition.
We ,trc not, ho\\ ever, here concerned to follow this in detail,
but only to consider 111 general the relation of the conclusion
to the prenusc; m ~urh 1mmed1atc mfercnccs, to ask how far
they constitute ..i real f>tcp m tlunkmg .i.nd what claims they
kL, c lo be c<illed mfcrcncc at all \Ve sllclll also have to
spl·.a.k of other kmtls of immediate mfcrence besides these so•
called ' formal ' ones
§ 216 The statement f>omc A,., B might be called an immediate
111fcrcnre from .i.11 A 1s B, but 1t doe'! not satisfy an essential
d1aracter1st1c of mfcrence 1t 1s lc'!s than \\-e knew m the
premiss. Consider, however, the relation of some B is A to all
A is B (the only form of immediate mference by conversion
from all A is B allowed 10 formal logic). The inference in the
form some B 1s A appears no longer to be, hke some A is B,
a mere part of what we knew before, in the sense of being less
than what we knew before, for the true full mfcrencc is some
 INFERENCE
B is all A.• W c ask, then, the verbal form being different, is
there a difference of meaning such as we require for inference ?
In the facts exprec;sed by statements such as all A is B, or A is
equal to B, or A is a friend of B, the realities to which A and
B refer stand m a reciprocal relation to one another, and the
nature of each 1s affected by its relation to the other in the
obvious sense that the hemg of each mcludes the relation. But
the relations mcluded m the complex fact of the reciprocal
relation of A and B arc various. Sometimes they arc different
m species, and then the difference 1s seen at once, as in A 1s
the father of B and so B is the son of A. But the same is true
even when the relations arc the same m species ; if A and B
arc friends, A's fr1endsl11p for B is different from B's friendship
for A. Even U1 such an instance as A 1s equal to B the same
holds. Each of these d1ff erent relations 1s expressed by a separate
verbal form of statement, e. g by all A 1s B and some B con·
st1tutes all A. Even if we confine ourselves to the traditional
form some B is A, that expresses a reld.bon of J3 to A different
from the relation of A to B Such difference of relation we may
illustrate more concretely, thus: A 1s half B; in that way A 1s
related to B; an 1mmed1ate inference is that B 1s twice A,
representing the relation of B to A. These two relations arc
obviously different, the half and the double. But now the rela-
tion of A to B and the relation of B to A being d1fferent, they
nevertheless necessitate one another and the act of judgement
or opm1on involves m either cai.e, for the person forming it, the
other relation, though he docs not express 1t verbally. The
objective fact, indeed, to wluch the first statement relates is
d. unity, havmg two sides represented fully by two statements ;
the two sides arc not merely parts of an aggregate, but are
inseparable ; the complete fact to which the statements refer
being their umty. The expression, however, in either of the two
statements is as expression one-sided ; h which side 1t will happen
to take depends on what we have taken as our starting con•
ception or logical subJect.
We sec, then, that the so-called inference 1s in a sense not
[• Tiu11 1s to adopt the extreme view of one school of formal logiC111D11,
But then, in Wilson's sense, there 1s no inference smcc all A 1s some B and
some B 1s all A arc statements of 1dent1ty
b • Conwder the 1mpl&cauonb.' MS nole.]
 The Gene,a/, Natu,e of Infwuice .f.23
new ; and this is what causes us our difficulty and makes us
doubt whether the inference should be called inference at all.
It is not uew ; the conclusion is involved in the act of thought
which makes the premiss ; the truth being that the premiss
brings out one side of the act of thinking, while the conclusion
brings out the other side. Yet we cannot call one of these
statements identical with the other, nor is one a part of the
other: each of them involves the whole, which gets its expression
verbally in both together. Now 1t 1s because of their difference,
which, as we have seen, is not that one 1s merely part of the
other but that they express two different relations, that such
inferences are entitled after all to the name of inference. They
satisfy this definition, that the one is different from the other
and necessitates the other. The process from the one to the
other is not on that account the less inferential that it is so
simple, and it wdl actually turn out in the end that this imme•
diate necessitation by one element, or complex of elements, in
a whole, of another element, or complex, in the same whole is
what always constitutes mferencc.
Inasmuch, however, as the simplicity of the process tends to
mislead us and to make us think that in the immediate inference
there is merely a repetition and at most only a change in the
verbal expression, 1t is useful to observe certain examples. Con•
sider the immediate inference a from all A is B to no A is not-B.
Suppose we start from all A is B : that necessitates no A is
not•B, and the latter states exphcitly an element in the whole
thought to which the first statement, as a partial statement,
corresponds. For we ask first whether A can, or cannot, be B,
and have before us the possibility both of all A is B and of some
A ts not-B. In deciding for the first we exclude the second.
The two, then, are inseparable aspects of the same truth, but
the difference between them appears at once when we consider
how each can be got. Though each ts necessary to the other,
they are not in our thinking co•ordmate, for we find we can
only possess the universal negative in the form of an immediate
inference from the affirmative. We might think at first that
we could get each independently ; that, while we get all A is
B by finding that the nature of A necessitates B, in getting the
[• • Rewrite this more clearly' MS. note. See Part II, ch. 12 ]
 INFERENCE
negative we might, though ignorant of the reason why all A is
B., have arrived at no A is not-B simply by finding that there
wu no A outside B. But, as we have already seen, in many
instances (and indeed in every instance of a scientific universal),
this last process cannot be performed independently, because
the area of not-B is infinite, and hence we can only tell what
is in this sphere or not by considering positively the nature of
A and the nature of B. Jn the cases where we seem to form
no A is nol-B without the affirmative all A is B, we reany
depend entirely upon affirmations, which divide up exhaustively
the indefinite sphere of not-B. Thus, the exclusion of A from
not•B rould not be an mdependent act 111volving mere negation;
there 1s no ,vay of arriving al no A 1s not-B except by estab·
hshmg all A 1s B. Hence, we cannot arrive at the negative
statement m question at all except as an immediate inference
from all A is B. All A 1s B may be immediate m the sense
that I see immediately that A necessitates B, no A is not-B
cannot be itself an immediate apprehension, but can only be
acquired as an immediate inference
§ 217 a It is sometimes said that the inconceivability of the
contradictory of a statement 1s the test of its truth, and specially
that 1t 1s the test from which we really derive the certainty of
axiomatic truth Now the preceding discussion of the relation
of no A is not-B to all A 1s B gives us one criticism of this
theory. The theory at bottom a'lsumes that we can start with
the negative A cannot be not-B, and upon that are able to
ground the statement that A must be B ; for the assertion that
we cannot conceive the contradictory of A must be B really
means that we affirm that A cannot be not·B and that we
apprehend that directly, together with the implication that this
judgement is acquired 1mmed1ately. That is what is meant by
calling it the 11.lt1mate test. But, as we have seen, we can only
pronounce this negative Judgement because we have already
seen that A must be B. That is to say, this inconceivability
of the contradictory supposed to be a test of the axiom is only
the consequence of our already having apprehended the truth
of the axiom. All that is true in the doctrine is this : that, if
[• Referring to Mill's controversy with H Spencer SJ•stem of Logic, ii,
ch. ; ; cf. infra, I§ 355 seq]
 The Gene,al NalUrt of I nje,ence 41S
we really see the necessity of a thing, we cannot conceive it
otherwise,
I 2 I 8. We may now give another kind of example of the
reality of the step made in immediate inference, In what is
called pure formal reasoning-the simplest of all-we sometimes
find it either necessary or convenient to draw an immediate
inference from one or more of the premisses in order to get our
conclusion. Now this shows the reality of the process that we
go through ; that 1t must be something more than a merely
verbal change. It will indeed sometimes be found that a com•
plex of formal premisses, though obviously dealing with the
simplest relations possible, causes us considerable difficulty, and
we see our way through a complicated system of premisses by
help of a number of mere immediate inferences. Indeed such
difficulty is sometimes found with quite a few premisses. For
instance, given that no AC is B, and no D which is not-B is A.
The solution becomes quite easy if we first transform by imme•
di ate inference the second premiss, and put it in the form
All A which is not-B is not-D ( = No A·not-B is D).
The original first premiss, then, being transformed into all AC
is A-not-B, we have a simple syllogism in Celarent with A-not-B
as the middle term, and the conclusion is that no AC is D.
Or, again, given that all AB is C, and all A-not-B is D. This
is solved simply by immediate inference from the first premiss
to all A-not-C 1s not-B, that 1s, all A-not-C 1s A-not-B. This
gives us with the second premiss as major a syllogism m Barbara
of which A•not-B 1s the middle term, and the conclusion is see'
to be that all A-not-C 1i,, D.
§ 219. There are certain processes which we ~hould on reflec- ·
bon not be inclined to call inference (nor are they usually
recognized as such m logic) which yet have the verbal form of
an inference and, if judged by the test which we have just
applied to immediate inference, appear entitled to be .,called
inferences because the argument seems to require the first st\
which· is preceded by the word ' therefore '. Moreover, they
seem sometimes to exhibit in the conclusion a real difference
from the sum of the premisses. For instance, if we know that
a.fl is C and find first that A 1s a. and then that A is /J, it follows
that A is C. This would usually and naturally be expressed as
 INFERENCE
follows : A i'i II and A is /J, therefore A is 11/J. But o.fJ is C and
therefore A is C. Here, while the step A is o.fJ appears necessary,
it yet seems to be only the two premisses together. Again,
suppose AB exists, and we have A is II and B is P, therefor,
AB is o./J ; but we know 11/J is C, therefore AB is C. The step
AB is 11/J seems necessary. Further, it seems this time to differ
from the premisses, because from 1t given alone we could not
get the premisses. It 1s compatible, for mstance, with A is fJ
and B is a Euclid m the first proposition of his first book uses
an argument of the form A = B, and C = B : therefore A and
C = the same thmg B ; but things equal to the same thmg =
one another : therefore A -= C.
Herc the first inference seems a mere restatement of the
premisses, yet the step is in fact made and also seems necessary
to the complete argument.
Agam, 1t differs from the premisses either singly or together
because the term ' the same thing ' occurs in neither, and this
seems got by a comparison of the premisses. This again seems
a new act and not a mere restatement.
§ 220 To solve such d1fficult1es certain distinctions have to
be made which ought to be preliminary to any theory of
inference, and are yet commonly, perhaps always, neglected, to
the con!usion of certam parts of the subject.
We must distinguish first between the thought which the
verbal form given to a Judgement expresses and the whole
thought wluch produced the expression, for the former may not
be the whole of the latter As the Judgement really 1s the whole
thought which produces the imperfect expression, 1£ the latter
is taken as the true <"xpre,;sion of the Judgement, there arises
the fallacy, common m logic, of distmguishing the Judgement
as a result from the thought said to produce 1t ; an impossible
abstraction, for this thought is the full Judgement. What 1s
called the process of arriving at the Judgement 1s really the act
of judgmg. Thus, m B is C, A 1s B . therefore A is C, A is C 1s
represented as a Judgement resultmg by inference from the other
two. But tlus mference is exactly the Judging that A is C, and
thus A is C expresses only a part and not the full Judgement.
The full expression 1s A is C because B is C and A is B. This
solves the difficulty Just raised about the proposition AB is 11fJ1
 The Gene,al Natu,e of Inference ~7
namely, that the premisses cannot be got out of it. AB is a.p is
only judged on the ground A 1s e1 and B is P, and so the full
expression of the judgement (that is, of the thought which is
necessary to make the verbal expression AB is e1/J possible) is
AB is aP because A is a. and B is {J. Thus the premisses must
appear in the only way m which AB is a.{J can really be a judge•
ment, and the difficulty raised is a fallacy caused by the false
abstraction of a judgement from the way 10 which it is judged.
We must again distinguish between the apprehension and the
fact which 1s apprehended. It will be found that much depends
upon the question whether the premisses are taken to represent
the one or the other.
Tlurdly, we must distinguish between our apprehension of
a fact and our memory that the fact was apprehended (which
1s not necessarily a memory of the apprehension itself). Observe
that if the apprehension was an expenence, the memory of the
experience (not the mere memory that it was experienced) is
not itself an experience and is not a repetition of the previous
experience. If, however, the apprehension was a process of
proof, the memory of the full proof 1s itself the process of proof,
or we may call 1t a repetition of the proof
By help of these distinctions we shall sec that in the cases
under consideration the fact represented by the conclusion is
equivalent to the facts represented by the premisses in conjunc-
tion and not something different from them necessitated by
their conjunction, and that m this sense the conclusions are not
inferences. On the other hand, we shall see that the appre•
hension represented by the conclus10n, or corresponding to it,
1s not the same as the apprehensions represented by the so-called
premisses, nor is it the same as thec;c m conjunction, and the
difference 1s not merely one of verbal expression
§ 221. The difficulties may be resolved m this way. Consider
first the premisses as representing the facts apprehended. The
fact of A's being a, and the fact of A's being /3, that 1s, the
co-existence of these facts, does not necessitate A's being e1/3 as
something different from itself. On the contrary, 1t 1s the fact
that A is both a and /3. And the other cases may be treated
similarly. Consider next the premisses• as representing acts of
(• 'Change the example, smce a here necessitates /J.' MS. note J
11773•a C
 INFERENCE
apprehension. The facts A is ci and A is /J are in themselves
not separate, but the apprehensions of them may be separate
and, more than that, in some cases it may be impossible to have
these apprehensions together. For example, the triangle formed
by the diameter of a circle and by two straight lmes drawn from
its extremities to a pomt in the circumference 1s a right-angled
triangle. It 1s true ot the same triangle that the square on the
diameter 1s equal to the sum of the squares on the other two
sides. The apprehension of the first property is the proof given
by Euclid in Proposition 31 of his third book. The apprehension
of the second property 1s the proof given in the 47th Proposition
of the first book, and nothing shorter. Neither of these appre-
hens1om, contains the other, nor can we have them simul-
taneously, as we cannot conduct two proofs simultaneously.
Here then the judgement A 1s both a and /J, which we un•
doubtedly ground somehow on the apprehension of A as ci and
the apprehension of A as {J, is not the same as these apprehensions
nor the same as their conJunction. It 1s therefore m some sense
a Judgement which 1s different from them but which they neces•
sitate. It is on this account that the process has a resemblance
to mferencc, and on this account also 1t 1s natural in the state-
ment of the argument to add to A 1s a and A 1s {J the statement,
therefore A 1s a/J.
To see whether there really 1s an inference we must ask what
exactly the apparently new Judgement, A ts o. and {J, is. Suppose
we prove that A 1s o., and then prove that A 1s {J. It would
probably be said that m the proof that A Js {J, or at the end
of 1t, we remember the result of the proof that A is a, though
we have not the proof. The v.ord result JS somewhat mJsleadmg:
1t rather 1mphes that we remember the mere fact that A is a.
without reference to the proof, because 1t 1s somethmg different
and resulting from the proof. But this is qmte impossible; the
accurate expression 1s, that we remember neither the proof that
A is a nor the mere fact A 1s a, as a result, but the fact that we
proved A is ci. To put it otherwise, we are not really appre-
hending A's bcmg a but remembering that we once did apprehend
1t. 1 In the proof of A's bemg {J, if it is all before us, we have
the apprehension that A 1s {J. If we have this together with
1
Obspn•e, not remembering the apprehen1110n.
 The General, Nattwe of Inference 429
the memory that A was proved to be a., this would be verbally
expressed in the two inadequate formulae of judgement, A is
a. and A is fl simply. And now is the judgement A is a.fl an
inference from the apprehension A is fl, and the mtmory that
we apprehended that 'A is a.' ? It is not such an inference, for
1t is obviously nothmg but having these two judgements together.
The same is true if the verbal form of judgement A is ft also
represented not the proof that A is fl but the memory that there
was a proof. But again we may ask, is it an inference from
the original apprehensions or apprehension ~ If 'A is fl ' happen9
to represent the second apprehens10n, and 'A is a' the memory
that we had the first, the difference is only in what 'A is a.' stands
for, and this 1s only the difference between apprehending that A is
a. and remembering that we once apprehended 1t. But now no
one would call the memory that somethmg happened in our
apprehension an inference from the apprehension of 1t, though
1t 1s true that the apprehension here conditions something
different from itself, that is, the memory that 1t happened. For
an mfercncc is always understood to be from what we have
before us, what we arc now apprehending. Now by hypothesis
the given apprehenc;ion 1s not before us. The same is obviously
true if both 'A is a.' and 'A is fl' represent memories that there
were proofs of them ; and 1f one or both of the original appre-
hensions was an experience and not a proof, exactly the same
treatment apphes Thus, finally, though 1t would follow that
the judgement 'A is a.{3' 1s not pi:operly called an inference from
'A io;; a' and 'A 1c; /J', we seem to have the explanat10n why it
1s natural to put in this step (therefore A ts a.fl) expressly. The
reason is that when 1t 1s naturally introduced 1t represents
a stage of consciouc;ness which 1s different from the apprehen-
sions (in the proper sense of the word) that A ts a. and A is /J,
1s indeed a step necessary to the proof
§ 222. The preceding mvec:;t1gat1on seems to bring out the
reason why we hesitate to call certam processes inferential even
when the step taken seems necessary, or at least natural, and
therefore not a mere restatement of what has preceded. In
the processes which we do not hesitate to call inferences the
facts or fact apprehended in the premisses necessitate the fact
apprehended in the conclusion as a fact different from t.hem-
c 2
430 INFERENCE
selves, and the latter fact is apprehended as thus necessitated.
Now we have seen that this definition at once decides the cases
under consideration not to be mferences. Again, it is these
processes (1.e such as correspond to this definition) which a.re
those actually recognized m logic as inferences, though without
a clear consc1ousnc,;s that this is so and of all that it implies,
Now a prmc1ple "h1ch ts correctly used in particular instances
is not al\\-ay,; rlearly recognized m the abstract or correctly
formulated. Tl11s has happened with regard to what seems the
true prmc1ple of inference ; the apprehension, one may repeat
shortly, 01 one fact as necessitated by a different fact or facts.
For inference m general is sometimes incorrectly represented as
the necessitation of one Judgement by another or others. If
this were so, a memory, as necessitated by a given apprehension,
should be regarded as an mferC'nc-e But, as we have seen,
nobody tlunks of mamtaming that; the real reason being that
we arc gmdcd m tlus particular mstanre by a sound instinct
and arc usmg the true prmc1ple For though m the given case
the memory 1,; neccss1tatc-d by the apprehension, the memory
1tc;elf 1s not an apprehension of this necessitation.
§ 223. The account given of the process A 1c; a. and B is /3 :
therefore AB 1s a./3 would, from one point of view, make an
inference in the third figure of the syllog1sm into an immediate
inference.
The- form of the figure 1c; M 1s P, M 1s S, t hcrcfore some S 1s P
(or some P 1s S).
The condus1on follows because the same thing M is both
S and P. We have then M ti. P, 1\1 1,; S, therefore M is both S
and P. The latter we have ruled not to be an inference from
the premisses, but merely a statement of them as both holding
together (which, we must notice, does not differ from the simple
statement of them), and the inferences from 1t of some S is P
and some P ts S arc unmediate. Thus there would be nothing
in the third figure but 1mmcd1ate mfcrence. The syllogistic
logic ts committed to this anyhow, if 'M 1s PS' is regarded as
one judgement, because that logic docs not recognize 'Mis both
S and P' as an inferenre. But now tlus analysis is not confined
to the third figure. Consider the first figure : all B is C, all
A is B, therefore all A 1s C. The premisses, all B 1s C and all A
 The General Nature of Inference 43X
is B, may be combined in one statement, just as in the third
figure. Thus some B is all A and is at the same time C, or
some B is C and all A, and from this the inference to all A is C
is obviously immediate. This, however, 1s only an anticipation
of what will be maintained later, that the relation of the con·
clusion to the complex of premisses (whether syllogistic or not
makes no difference) is always immediate, is not mediated by
anything intervening.
§ 224. We have been led to recognize a principle which holds
not only in the non-inferential processes we have been con•
s1dcrmg but m inference proper What we have said of two
proofs each with a single conclus1on, that the app'rehensions
wluch they constitute may possibly not be present together,
may hold w1thm one of these proofs. For it may have parts
which cannot be had as simultaneous apprehensions. The verbal
form of a premiss A is B, used in drawing an mfcrence, may
correspond not to the apprehension of A's bemg B, but to our
memory that we have had such an apprehension, so that m
the strict sense we arc actually not Judging the judgement
'A 1s B '. This, though not the exception but the rule in the
vast maJonty of proofs, seems quite ignored m the usual treat•
ment of 111fcrencc. It has important consequences and, among
other tlungs, it seems to be a part at least of the key to the
poss1b11ity of a kmd of error m the exact sciences which is
a stumbling-block to theories of knowledge and error.
§ 225. We may return to the contention that the Judgement
A and C arc equal to the same thmg differs from the judgements
A = B and C = B, because the term ' the same thing ' does not
appear m either of these two latter, which 'may be accounted
premisses. It u. true that it does not appear m the verbal
expression, but 1t is contained m the thought which corresponds
to this expression. For m Judgmg C = B, we must recognize
B as the same B which = A or we should not use the common
term B at c1.ll. If we had forgotten the Judgement B = A this
would not be true, but by hypothesis we have not forgotten it.
The reason for mtroducmg the step m the verbal expression is
that which we have already given for the introduction of the
step A 1s a./J or AB is a./3.
§ 226. Inference is often spoken of as if 1t were essentially
 INFERENCE
a connexion of judgements, as though the premisses were judge•
ments which necessitated our forming another judgement. Shall
we say simply that the Judgement of the premisses (that is,
the judging of them) necessitates the judgement of the con•
clusion? We have seen already that this, whether true or not,
is too wide to be taken for a dcfimt1on of inference, for it would
include memory.
Now first observe that "hether this is true or not 1t 1s in any
case not the mcamng of our verbal statement. If we say A is
C because B 1,; C and A 1s B, we do not mean that our Judge•
ments B 1s C and A 1s B necc!:is1tate our Judgement that A 1s C,
but \\C mean that the facts B 1s C and A 1s B necessitate the
fact 'A 1s C '. In science again the value of such inference 1s
not that somebody believes the conclusion because he believes
the premisses, but thc1.t ,u1 obJect1ve 11ecess1tatlon 1s actually
apprehended. Thus it 1s the connexion of the obJects appre•
bended \\ h1ch 1s meant, 1t 1s this which is apprehended m
mference and which is of mtercst and importance to science.
Secondly, we may truly say that I apprehend the conclusion,
or have the apprchcns10n represented by the conclusion, because
I have the apprehensions represented by the premisses In this
sense one apprl'hcns1011 m.t.y be said to necessitate another,
But, though true, tlm statement has a form which may mislead.
Apprchens1on, ,,e h.t.ve seen, cannot be abstracted from what
is ,lpprchendcd, and there is ..i. d.t.nger 111 scparatmg them, as we
have seen m d1scm,!:img the a priori view of knowledge .t.nd the
theory of the inconcc1vab1hty of the opposite as a tc5t of truth.
There we hc1.vc m.t.111l.i.1ncd that c1. ncccsbary apprchcnb1on 1s only
intclhg1blc as mcanmg an apprehension of an obJect1ve necessity.
We shall find this account verified in the case before us. How
1s 1t exactly that the apprehensions B is C aml A 1s B cc1.n be
truly bald to cc1.use or necessitate the apprehcns1011 A is C ? Only
li1 this \\ay. I apprehend the facts B 1s C and A 1s Band then
apprehend these facts as necessitating A 1s C Now that means
that the necess1tat1011 of the apprehension A is C by the appre•
hensions B 1s C c1.11d A is B 1s after all just the apprehension
of the necessitation of the fact A 1s C by the facts B 1s C and
A 1s B. Thus, as before, the necessary apprehension is only
11ccessc1.f) because 1t 1s an apprehension of a necessity, and the
 The General Nrlture of Inference 433
question as to how necessity in the thinking can correspond
with necessity in the object cannot arise.
Finally, suppose that one or more of the premisses of an
inference is not an apprehension proper of the fact that B is
C or that A is B, but our memory that we apprehended the
fact (experienced it) or proved it. If ,ve remember proving that
B is C and A is B, even though it be said that memory is
uncertain, yet at least we know that, 1£ we remember rightly
that B is C and that A is B, the facts would necessitate A's
being C.
§ 227. The definition of mference that we have given docs not
&tate that the connexion on which 1t depends is always one of
universals, and yet we know that we always find 1t to be so."
Now this 1s not because we define inference beforehand as only
deahng with such connexion. We find m any instance where
we could be said at all reasonably to apprehend the necess1ta·
t1on of a particular fact by another particular or particulars,
that such necessitation 1s only a partlcularizat1on of a necessary
connexion of universals, and m tho inferences which we arc
about to examine we shall always find that we have to do with
this universal necessitation. Aristotle recognizes the universal
character of inference m so far as he makes the reason or
ground 1 to be universal, but he recognizes this fact as familiar
without reflecting on it and v.ithout reahzmg the necessity of
asking any questions about 1t, as is often the case v.1th facts
w1th which we are fam1har. Clearly the question must be asked
why mference should have tlus charctcter. There 1s a danger of
avo1dmg the difficulty by somehow mcludmg umversahty in our
definition of mfcrcnce, as some modem treatises do,b thus over·
1 nfTuw, irc6oA011. Cf, § 2 37

[• ' Con~ider and embody the tact that we ~ecm to 111£er i,ometimes from
what is peculiar lo the mdnidual · i\l~ note I have put in a foot-note
(p 434), from a hasty scribble, what wai, the difficulty Wilson felt. Cf p 481
b I am not sure of the reference ' The universal 111 its differences LS
then the basis of mediate Judgement or inference' Bosanquet, Logic, II 1 1.
' The general pnnciple on which the validity of every concluston rests may
be expressed by the formula . " What falls under the condition of a rule,
falls under this rule itself".' Kant, Logic,§ 57. Cf § 58, (See ali:10 infra,
§ 262,)
The statement about Aristotle 1s ob11Curc The general rc!crc11ce may be
 lNFERENCE
looking an important point and evadmg a difficult investigation.
We may defer the quest1on untd '\\e have further examined the
nature of inference m sciencc.1
1 § 262 The nature of the umversabty 1s rv1dent when we argue {rom
a mark of what 1& of a certam kmd But there are lllStances where we argue
from what 1s c1. mark of a particular md1v1dual alone, so that if there 1s
a universal propos1bon 1riphed 1t 1& of c1. different sort from the other, and 1t
1& important to i.ec whether the ,1cw that mferencc 1s universal can here be
vmdu.ated
to ir &warm (uv>.>.o-y1up9i) 3,1 TU tta.8/,>.011 l11rapx~II' (An l'r I 24) Tlus l!I the
formal rule e'li mere parllc1uarib11s nrhil uguitur, but the reference to the
cause appears to relate to the doctrmc of the Posterior Analytics, that
thr aim of i11uT,jp1J lb to connect the predicate with the subJect through the
proxunatc 1.ause of tl1c predicate The maJor prcnuss, Aristotle therefore
m~1~ts, must be •al/,>.ov Moreover, &c1ence 1s the i.earch for pnmary, 1 e most
um\<ersal, caui.es (An l'o 72a 4-5) But, on the other hand, the J>ostmo,
An11lyt1es lb full of refled:mn on the quei.tmn whether demoni.trc1.tion 1s or
1~ not universal, and on the supenonty of true demonstration becau&c 1t 1s
universal (see cg An. Po 1 13 and 24) ln fcl.1.t the obiter dict11m would
perhaps have been reconb1dercd ]
 II
THE SYLLOGISM

§ 228. THE problem which a syllogistic logic proposes seems

to be to discover the general forms under one or other of wluch
all demonstrative argument must fall. This may obviously
become the problem, from this point of view, of dctermmmg
the general type of inference from two premisses. Now, 1n
order apparently to get a quite general solution, tlus logic makes
abstraction of the so-called matter of the propos1t1ons and deals
only with what 1s called the form of the propos1t1orn,. The
premisses and conclu&1on, that 1s, are treated only under the
general form of the relation of subJeCt and predicate (where
subject and predicate have the special meamng which has been
cnt1c1zed above), and appear only m these shapes: all S is P,
no S is P, some S 1s P, and some S 1s not P Smgular judgt·
ments such as this S is P rank as umvcrsal. But the generallty
thus sought for is by no mec1ns attamed by this method. In
actual \\orkmg the syllog1stic logic has unconsciously taken
quite a bm1ted problem about a special kind of inference and
of that problem itself hc1s given but a hm1tc<l solution.
§ 229. Even 1£ the problem proposed were correct, namely, to
find general forms for .i.11 mferencc (and we shall see that it is
not), yet we shall find that the rules of inference laid down as
prebmmary to the discovery of valid syllogu,tic forms show that
the solution actually offered 1s not the most general solution of
the problem which is proposed. For instance, a rule of syllogistic
logic 1s that no conclusion can be drawn from two negative
premisses. Yet from no P is M and no S 11, M we do get a rela•
tion of S and P not given by either premiss alone, and that
relation might be 1mportant. Agam, 1f the m1ddle 1s not dis•
tnbuted m one premiss at least, we arc supposed to get no
conclusion. Thus from all P is M and all S 1s M there should
be no conclusion. Y ct here a.gain a relation is cstabbshed
 INFERENCE
between S and P which may be of importance for some further
inference. The fact is obvious that wherever there is a common
element in any two Judgements, that must serve to relate the
elements in the two Judgements to one another. 1 The truth is,
that this logic does not inquire what inferences m general can
be drawn from two premisses m the given general forms. It
really asks. given two premisses with a common term (i e.
havmg the 5amc conception as subject-conception or predicate•
conception 111 each), when CcUl \\e conclude from these to a J)l'O·
po111tion of ,,luch the sub1cct-concept1on 1s one of the subject·
conceptions or pred1cate-conceptlons m the given premisses while
the predicate 1s the rcmainmg subJect-concept1on or predicate•
concept1011 Thus from no P 1s M and no S 1s M the problem
really 1s to hnd a conclusion such thc1.t the subJect-conception
of that conclusion 1s either S or P and the predicate-conception
either P or S Now there 1s no conclusion from the given pre•
m1,;ses which can hc1.ve the given form, an<l that is the only
JUstdicatlon of the rule that inference from two negative pre-
misses is impossible. The same may be said of the rule about
an undistributed middle With this prow,o, then, this special
restriction as to the form of the conclusion, the rules of the
syllog1st1c logic are correct, but as usually stated, without
the necessary proviso, they arc incorrect because they violate
the elementary pnm.1plc oi .ill reasomng whatever, that things
related to the same tlung c1re thereby related to one .1nother.
§ 230. The syllogu,tlc theory is often spoken of ai, 1f 1t were
an analysis 2 of the forms of demonstrative rcasomng, but the
stnkmg thmg c1.nd one which has not been sufficiently observed
by log1c1.i.11s 1s that JI~ method Ji, by no me,rn,; what is usually
called analyt1cc1.l. It docs not take arguments and abstract from
them their umversal formc::, \\ luch 1s what would be understood
by analysis Indeed, 1£ 1t did so, 1t could not have the kind
of completeness and certainty within its own hm1ts wh1cb it
1 Observe that m § 219 the contention was that A 1s both II and fJ 1s not

a real inference, not that there 1s no mference as to the relation of II and /J.
In fact II and 13 are related 1n the conclusions some 11 1s /J, or some /J IS a, m
consequence of their relation to A, as m the ordmary view of the third figure
of the syllogism What was pomted out about this was that the inference
is unmechate from A 1s both a and fJ. not that there 1s no inference at all.
• Cl the Aristotelian term Td artz.\1mn.
 Syllogism 437
undoubtedly has. It might show by examples the evidence for
a form of argument, but its judgements would be empirical,
wanting m universality and necessity. Its actual method is as
much a priori and ' constructive ' as that of any pure mathe•
matical science. It starts with the general conception of a pro•
position, with a d1stmct1on of subject and predicate ; it then
distinguishes the possible varieties of proposition exhaustively
a pnori and not by any analysis or empirical examination of
actual propositions. Then, agam, 1l determjncs a priori all
possible combinations of two premisses and determines from
them a priori all possible varieties of conclusion of the bm1ted
kmd described This h, exactly parallel to the method of a
mathematical science, and 1t will become clear as we go on that
the determinatton of the rules, figures, and moods of the syl•
logism, which occupies so large a part of this logic, is no part
of true logic whatever, though vahd enough m itself, but 1s
.i. science m the same sense as pure mathematics.
§ 231. It must not, however, be supposed that this a priori
tlunking wluch we have been descnbmg can proceed by pure
abstraction only, although 1t docs deal with forms which to
a certain extent arc abstract, We may perhaps tlunk that m
this kind of logic we work with the general form of the syllogism
from the first and thc1.t we dcnvc from that any application to
pc1.rticular cases Now that 1s altogether impossible ; we cannot
understand these forms except by takmg dcfimte instances to
show what the symbols mean, that 1s by havmg matter as well
.i.s form. Take, for cxc:1.mplc, all M 1s P, all S 1s M, therefore
.i.11 S 1s P. To sec the vahd1ty of tlus we must take a parttculc:1.r
syllogism with actual propos1t1ons, and m that instanLe \\-C must
see directly the proper conclmwn, which is as specific and
definite as the premisses themselves. We must further see on
reflection how the general form of the conclus1011 depends on
the general charactenst1cs of the form of the premisses. The
first step, namely seemg the conclusion m a particular case, 1s
the condition of our bemg able to reason at all m the particular
way in question ; the second, namely the abstracting process,
is the condition of our being able to make the general logical
abstraction • of the syllogism It is also directly self-evident
L• ' "logical" reconsider.' :vIS, note.]
 INFERENCE
to us that the form we are abstractmg is universally valid,
because we can sec that nothmg m 1t depends upon the matter
peculiar to the instance. Its method therefore is the appre•
hension of the universal 111 the particular, and we see how both
imagination and perception are necessary to that abstract
invest,gat1on, a priori a& 1t 1s, which determmes the syllogistic
rules. This proccdi.rc docs not mdced prove any rules of
inference ; 1t 1s simply the immediate recogmt1on of them.
§ 232. The syllog1st1c rules, being but abstractions from the
actual procedure of lhe human reason, cannot be described as
rulc,s discovered by the log1c1an and laid down for the guidance
of our reason : they tell us what reason necessarily is, m
.i. ccrt:\m hm1tcd department, and we cannot prescribe rules for

what cannot be otherwise. Tlus 1s not realized by the syllogistic

log1c1am, when they <lra\\ up a list of fallacies to be avoided.
As reason has rules, they seem to thmk that these rules may
be broken and sometimes expressly say so, and then they
imag10c that th1& 1& confirmed by the fact of error. They arc
rules, though the phrai.c 1s m1sleadmg, but they cannot be
broken ; 1f they could, no logic could ever instruct a reason
capable of such error. On mspcct1011 1t will readily be seen that
the so-called formal fallacies arise from a misrepresentation due
to the log1c1an himself of certam aclual trams of argument ;
these he treats as though they \\ ere m.tcnded to be demonstra·
tlons, whereas they are but probable arguments, which no one
supposes to have proved tht'tr conclui.1ons. The precept, for
instance, about the d1stnbut1011 of the middle m syllog1stic
argument 1s absolutely nug.i.tory ; no one can perform the
intellectual feat of arguing \\1th .1n undistributed middle and
supposing thc1t he hc1s a necessary demonstration.
§ 233 a The ord111ary .i.na.lys1s of a proposition on which the
syllogistic logic 1s based, all S 1s P, 1mphes a type of proposition
m which the predicate 1s c1ttachcd to each md1v1dual compre·
bended 10 S, or each case of the umversal Sness, mdependently
and not as bemg 10 some relation to the other individuals.
Take the statement that all triangles have any two of their
sides together greater than the third ; here we can say that
l" ThJS secbon 11o markf'il • to be rewntteu •. I have found 110 redraft and
left 1t m 1t11 p1Cl$eut ,•cry lc&mc form.J
 Syllogism 439
each triangle has any two sides together greater than the third
because each has the given property quite independently of the
existence of other triangles. But now take the statement that
things which are equal to the same thing arc equal to one
another. Here we can get no single subject expressed at all
which will do for our purpose We cannot, for instance, make
'each thing' the subject and 'equal to one another' the pre-
dicate. Thus the ordinary form of the syllogism JS adapted to
the first case but not to the second. In this second case 1t
normally requires some trouble as well as considerable peri-
phrasis to force the statement mto the proper verbal expression
required for the syllogism. We might put a given argument
into this form. Every group of things such that its members
are equal to the same thing JS a group of which the members are
equal to one another But AB 1s a group of which the members
A and B are equal to the same thmg. Therefore AB is a group
of which the members (A and B) arc equal to one another.
This conclusion we then have to interpret into A is equal to
B or A and B are equal to one another, the natural modes of
expression. The necessity of this form of reduction comes from
the fact that A and B do not appear 10 the statement as
separate subjects of a given predicate, but only as related
members of a group. The art1fic1ahty of the reduction is obvious
and the reasoning is dearer and simpler m the ordinary non-
c;yllogistic statement of the argument. We shall see presently
how such artificial forms arise, and that the true form of reasoning
in such a case is not syllogistic at all. But besides this there
are certain processes which concern statem,ents of this kmd
(which we may define as statements of which the grammatical
subject is a group of particulars considered as related m a certain
way to one another), which the syllogistic logic leaves altogether
out of account. We have already had an instance where A and
C arc equal to the same thing is derived from the two state•
ments A = B and C = B. Now, on the one hand, this process
explicitly appears in the syllogistic presentation of certain argu•
ments in geometry and, on the other hand, there is no account
given of it in the deductive or syllogistic logic , for obviously
it cannot be a syllogism that will give such a result. This
shows how little the syllogistic logic jg based upon an analysis
 INFERENCE
of given arguments ; for the very first proposition in Euclid's
Elnnmts contains an example of reasoning of this kind, with the
difficulties of which, from the point of view of syllogistic logic,
we have been dealing.
§ 2.34. The syllogistic logic then has not solvt!d its own
problem completely We shall now show that its problem is
not the complete problem of demonstrative reasoning. We
observe that m the c;yllog1sm the subject and predicate terms
in the conclusion arc related through their relation to one and
the same term; and this relation 1s of the special kind called here
prcd1catlon. Now this is clearly a species of something more
general, the relation of two terms to one another in virtue of
their relation to a third term. The general problem, then, if
conceived on the analogy of the problem of the syllogistic
logic1an'l, would be to find the general rule'! "h1ch can be laid
down for dell'rmmm~ the relat 10n between two terms which
follows from their relation to a third term In the nature- of
the rac;e the answer mu'>t be that there arc no general rules.
For obv10uc;ly the rule of mfercnre m the case of each relation
must depend on l he particular nature of the given relation itself.
It must be got by our knowledge of the '!pec1al subject-matter
and rannot possibly depend upon any general forms of thinkmg
or mference, ac;, for instance, on what 1s called formal thinking,
wluch 1s supposed mdt."ed (though, as we have seen, wrongly) to
make abstraction of all such matter.
Take, for mstance, the argument most B is C, most B is A,
therefore ,;ome A 1s C This argument depends upon our special
knowledge 01 intuition in the department of quantity, that 1f
we take of a. given whole a quantity more than half and if we
t ..ikc in this same whole .mother quantity greater than half,
there must be something common to the two parts This is
a matter of absolutely direct mtuition, and no reason can be
given for it other than itself. Clearly the rules of the syllogism
would not g.ive uc; the required conclusion. If we force the
argument itself mto the verbal form of the syllogism we shall
find that we can only do so by a verbal transformation in which
we make the argument itself its own major premiss. Thus :
Two parts of the same whole which are each more than half of
it must have a part m common, A and C arc such parts of
 Syllogism
a whole B, therefore some A is C. But the major premiss is
nothing but putting into ordinary words what is meant by the
symbolic expression If most B 1s C ( - BC} and most B is A
( = BA), some (B)A is (B)C, for B stands for any whole what•
ever, and (B)A and (B)C for any parts of it greater than a half.
Again, the plane figure A 1s inside the place :figure B, the plane
figure B inside the plane figure C, therefore A 1s inside C. This
time the inference depends on our spatial apprehension. It is
self-evident. The two premisses are quite sufficient by them•
selves, yet no syllogistic rule can get the required conclusion
out of them.
§ 235. What now 1s the relation of the syllogism to this very
general conception of inference ~ The answer makes clear the
limitations of the syllogism. The syllogism does deal with
a definite relation usually, though maccurately, expressed as the
relation of subject and predicate. This predicate is m the
affirmative statement a kmd of being wlurh the subject has,
either covermg the whole of it or but a part of it, in which
latter case the relat 10n 1s called that of subject and attribute,
and the same holds, mutatis mutandis, for the negative state-
ment. It 1s our consideration of the special character of this
relation that gives us the rules of the syllogism , m fact we
recognize the rules of inference here as elsewhere because we
have a d1rcrt mtu1tion of the character of the special relation
before us. We must, nevertheless, avoid the error of supposing
that this relation is one which covers all other relations. It
does not, for instance, cover the relation expressed by A is equal
to B, where the rclat1011 of A and B is not that of predication.
This relation of predication has perhaps been unconsciously
confused with the general form of every relation because every-
thing can stand m such a relation. But what docs this latter
exactly mean ? The relation of father and son can, for instance,
be put mto the so-called pred1cational relation, 1£ we say, The
father 1s father of the son. Y ct the predicational relation here
obviously is not the relation of father and son, nor the genus
of which the latter is the particular, for the son cannot be
predicated of the father or the father of the son. These two
terms do not stand in the relation of predication.
§ 236. All this mediate inference has a middle term, but it is
 INFERENCE
clearly not always that of the syllogism. Its definition is simply
that it is a term to which the others are related, and as this
relation is not always that of subject and predicate, so the
middle term cannot always be the syllogistic middle term. We
may easily verify the fact that m scientific investigations we are
mainly concerned in seeking relations between given terms which
are not in that relation here called predication. There is another
cons1derat1on which will brmg out the fact that the syllogism
deals with a special relation and therefore with what is called
matter in reasoning as opposed to form We shall always find
that the premisses are not the mere abstract form of the pro•
position. They are very abstract, 1t 1s true, but they are not
all form ; there 1s left m them just enough matter to make
reasoning possible. In Figure I, for instance, there are two pre-
misses, each of which may represent a umversal affirmative.
Now they could be mere form only 1f they represented the
universal affirmative in its simplicity. But of that there is only
one form (all S 1s P). Thus 1£ we arc really dealing with
abstract form only, both premisses must have exactly the same
form and identical symbols Thus m tbe expressions all M is
P and all S 1s M we have not abstracted all the matter. The
term M is to be the same in both premisses ; P and S are to
be different m general and different from M. This, then, 1s
matter which has not been abstracted. If 1t were abstracted,
both premisses would be reduced to on<." and the same form
In fact all M 1c; P and all S 1s M, as a general form, 1s the general
form of two propositions "h1ch are different from one another.
Thus the matter m them 1s that the subJect•conception M in
the one is different from the subject-conc-ept1on Sm the other, and
the predicate-conception P in one from the predicate-conception
Min the other, and finally that the subJcct-conceptlon M of the
one proposition is identical with the prcd1cate-conception Min the
other. This matter which has been left 1s precisely what makes
the inference possible, for without 1t there would not have been
two premisses but only one. We may illustrate this by an analogy
from geometry A theorem about two circlesmtersectmg is univer•
sal, and 1t is right to say that it is universal because it treats of
the relations of universals. But there is a sense in which it does
not treat simply of the universal of the circle, for there 1s only
 Syllogism 443
one such universal, and we can't have the universal circle cutting
the universal circle. That about which the theorem is demon•
strated is the universal of two particular intersecting circles, or
the universal of which the particulars are groups of two circles
intersecting one another, and this depends indeed ultimately
upon the universal of the circle.
§ 237. It is no doubt supposed by some to be decisive for the
claims of the syllogism to be the general form of demonstrative
reasoning that all demonstrative arguments can be reduced to
the syllogistic form. This is taken as a fact readily verifiable
though without an attempt to show that it must always be so.
If we inquire why it really 1s that this apparent reduction should
be possible, we shall discover the nature of the fallacy. If we
take the syllogistic argument all M is P, all S is M, therefore
all S is P, we can represent this as the application of a general
principle to a particular case ; we apply, that is, the general
principle that all M is P to the particular case of the M's which
are S. Now every particular inference which concerns a given
kind of relation can of course be represented as the application
of a general principle to a particular case, that 1s, as the applica-
tion to the particular inference of the rule of inference which
belongs to the given relation. Thus we can bring the argument
verbally into the form of a syllogism. This obviously involves
making the rule of the inference the maJor premiss in this
syllogism, and the particular apphcation of the rule of inference,
the conclusion, will be represented as inferred from the rule of
inference itself, as if the latter were a premiss. This in fact we
shall find to be exactly what 1c; done in the so-called rcduct10n
of what are really non-syllogistic arguments to the syllogistic
form. Moreover, 1f the inference is put in a certain form which
is really general, though employing special symbols, the reduc-
tion will take the form of which we have already had examples,
viz. the inference to the general form, disguised m symbols, from
itself stated without symbols. And 1t is clearly a fallacy to
represent the rule according to which an inference 1s to be drawn
from premisses as one of the premisses themselves. We should
anticipate that thas must somehow produce an infinite regress,
and that this 1s so can easily be shown. But now observe
that there are general forms of argument to which the proposed
2773•2 D
 444 INFERENCE
reduction cannot be applied because the major premiss (in
the reduction), wluc11 is the rule of inference, is nothing but the
general form of the mference or argument itself. Thus the
argument B is greater than C, A is greater than B, therefore
A is greater than C, would be reduced to the syllogistic form by
taking as major premiss the followmg statement, when of three
magnitudes the first is greater than the second and the second
greater than the tlurd, the first 1s greater than the third, and
as minor, A, B, and C stand m this relation to one another.
But the expression, 1f B 1s greater than C, and A is greater than
B, A is greater than C, is perfectly general, because taken to
l>r true whatever A, B, and C may be, and 1t 1s because it 1s
general that 1t can have this c;ymbolir form Thus it is the
exact equivalent of the supposed maJor prenuss from which it
is pretended to be deduced. It may here be remarked that the
theory of the so-called dictum de omni et nullo bemg an axiom
of the syllogism or a canon of c::yllog1st1c rf.'ac;onmg 1s only an
amusmg mstancc of this same fallacy Th1<; dictum is nothing
but the syllog1stic figure of which 1t 1s supposed to be the axiom
written down 111 ordmary words instead of bemg partly written
in symbols. To return to our supposed reduction. 1f we hke
to give the symbols a more particular meaning, so that A1
represents not A m general but a particular A, we shall get by
the proposed reduction an mfimtc regress Thus let the argu-
ment be• A1 = B1, B1 - ell therefore A1 = cl. The rule which
has to be put as major premiss 1s, thmgs which arc equal to
the same tlung arc equal to one another. Under this we sub-
sume A1 and C1 arc thmgs equal to the same thmg, and so draw
t lw conclu<;ion that they arc equal to one another This 1s
syllogism I. Now syllogism I, which 1s of the form MP, SM,
SP, m its turn exemplifies another rule of mference which 1s
the so-called dirtum de omm el nullo This must now appear
as a maJor prem1s,; The rcsultmg syllogism may be put
variously ; the following shorl form will serve. Every mference
which obeys the dictum 1s correct , the mfcrence of syllogism I
obeys the dictum; therefore 1t 1s correct. This 1s a new syl•
logism (II) '\\ lurh agam has for rule of inference the same
dictum; hence a new syllogism (III) and so on in saecula
saentlorum.
 Syllogism 445-
The foregoing is an indirect refutation, of the nature of reductio
ad absurdum, A direct refutation may, however, be given a,
follows. In the above procedure the rule of inference is made
a premiss and a particular inference is represented as deduced
from it. But, as we have seen, that is an inversion of the true
order of thought. The validity of the general rule of inference
can only be apprehended in a particular inference. If we could
not see the truth directly in the particular inference, we should
never get the general rule at all. Thus it is impossible to deduce
the particular inference from the general rule, and the so-called
reduction 1c; merely verbal with no corresponding process of
thought.
It 1s of interest here to remark that there is a natural and
justifiable way of speakmg which promotes that confusion of
thought which has suggested these verbal reductions to syl•
logistic form. Take, f.or example, B 1 = C11 A1 = B1, therefore
A1 = C1• If we arc asked why A1 = C11 we may answer, because
A1 and C1 are both equal to B1 , or because A1 and C1 are equal
to the same thing. Now 1£ c1 statement preceded by the word
because 1s given as the reason for another, we tend to assume
that 1f the reason 1s a true one the said statement must be
a premiss from which the other 1s deduced. And this is in
effect the doctrine of Aristotle. Yet we have seen that that is
quite 1mposs1blc 111 the particular examples we have examined.
It is not true that the word because, however valid the reason
it introduces, must introduce a premiss. What then does it
precisely mean? In the given case we see that A1 - C1 in
virtue of the fact that both are equal to the p~rticuJar thing B1 •
That 1s the first step m our thought. But we reflect further
that the equality of A1 and C1, as apprehended, docs not depend
on anythmg else m them except that they both are equal to
B1, nor on anything m B1 except that it is the one thing to
which both are equal. Thus we see that the inference may be
generalized and that we can say that thmgs which are equal to
the same thing are equal to one another. This therefore, ac;
being the only t'sc;cntial, we truly call the reason, and it is rightly
preceded by the word because. Yct, as our analysis has shown,
we can only get this generahzation by an act in which we
recognize it in the particular instance, an act which cannot
D 2
 INFERENCE
possibly be preceded by the generalization as a premiss. We
recognize then that the true reason of a fact which we appre•
bend is not necessarily a premiss from which we deduce the
apprehension of it, and we arrive at the very converse of
Aristotle's view that in the proof the fact must be deduced
from its cause (arnm•) as premiss 1
This ic; the result of one of those principles which we have
maintamed to be a nrccssary preliminary of any account of
inference, the d1stmrtton, namely, between what a given premiss
means and our apprd1ens10n of 1t. To sum up, such arguments
.is B = C, A "' B, therefore A = C, or, most A is C, most A is
B, therefore some B 1s C, or :my other reasoning which proceeds
by rclatmg t\\O terms to one and the c;;ame term, are, in this
general symbolic form, on the same footmg with respect to
logical analysis as the syllog1stlc forms, &uch ac; M 1s P, S 1s M,
therefore S 1s P, themselves. The analysis 1s as ultimate m the
one case as m the other, and 1£ the first kmd requires further
analysis, m the sense that the rule of mfcrcncr 1c;; to be explicitly
stated, so also dors the second; the truth bemg that in either
case 1t 1s nugatory, for 1t turns out to be a mere restatement
of the argument itself In fact m every one of the inferences
in quest10n, the rule of mfercncr 1s explicitly stated, requires
therefore no furthrr &tatement, the rule being precisely tlus
statement.
§ 238. We have now to mqmrc what 1s the relation of the
conclusion to either prem1c;s of an mference taken by itself. We
will consider generally the mfcrencc in wluch the terms, say,
A and B are related to one another through their relation to
the same thmg M. A convement symbolism 1s Br1M, Ar1 M,
:. Ar3B, where the precise nature of r3 depends on a knowledge
of the relation r (the general form of the rl'lation hac;; the same
main symbol r because the relations must obviously be of the
same kmd). The question how the condus1on 1s related to each
premiss occurs m the ordmary logic in a special form which
concerns the syllogism only, smce it is the syllogism only which
1
An11totle never distmguishes the obJecbve reason of a fact from the order
of apprehennon in our thought , he speaks as though the apprehension of
the reason could precede the apprehension of the appbcation to a particular
case
 Syllogism 441
is considered there. In the syllogism the form of argument is
such that a special difficulty has been felt about the relation of
the conclusion to one of the premisses, the major. It has been
objected that the major premiss, all M is P, must include the
conclusion all S is P as a part of what it states, inasmuch as
S 1s part of M. Thus instead of proving that S 1s P, we seem
simply to assume a wider statement, which includes 1t, and this
statement 1s the major premiss. This is what is meant by saying
that the syllogism is a petttio prmcipii. Let us, however, con·
sider the more general form, wluch includes the syllogism. If
the act of thought wluch gives us Br1M 1s not such as to show
us the side of M in wluch A stands m the relation r to 1t1 then
t}us act of thought Br1M certamly cannot yield the relation r1
between A and B. It requires to be supplemented by the other
premiss, Ar2M, winch shows us that aspect of M to which A 1s
related, before a conclusion is possible. The same is true if we
c:1.sk whether the conclusion 1s contamcd m the premiss Ar11M.
Thus the argument requires both premisses, winch proves that
the conclusion cannot be sa1cl lo be cont,uncd m one alone.
Once more we observe the 1mport.mce of d1stingu1shmg between
'"hat c:1. premiss mc..tm, ob1cct1vcly and our apprehension of
this
§ 239 Although the c:1.bove rea!>onmg mcludcs the syllogism,
we may consider the latter separately Let the major premiss
be all M 1s B. This premiss may be got m one of two ways.
There are cases where the group M 1s exhausted ; here the
predicate B may be attached to each member separately, giving
the general i,tatement ,1,ll M 11, B. Consider now a member A of
the group M. The act of thought by "h1ch B was attached to
M m general was by hypothesis one whereby B was attached
directly to A. Thus the cond1tton of mfcrencc 1s not realized
,m<l t.here 1s no true syllogii,m. The supposed conclusion A is
B is an explicit p,ut of the Judgement or opinion all M is B.
But, in the first place, no one m such a case ""ould seriously
give A 1s B as the conclus1011. These exhaustible groups are of
httle importance The umversals studied by the sciences have
infinite possibilities in the way of indiv1duals 1 and even when,
in the case of exhaustible groups, we have attached the predicate
B to each member, we do not get the scientific judgement that
 INFERENCE
we require till we have found the common reason present in
each member which causes its Bness : only then has all ceased
to be an all of enumeration.
In a second group of cases all M 1s B may be arrived at 011
the ground that the universal nature of M necessitates B, and
that in two ways. We may be able to see immediately that
the nature of M, as sut.h, independent of particular manifesta-
tions of 1t, necessitates B ; that a three-sided rectilinear closed
figure, for instance, must have three angles; or, again, this
necessary connexion of M and B m thek universal nature may
be inferred as tl1e consequence of some other fact.
But, secondly, all M is B may be based on a so-called inductive
argument. For our present purpose this 1s m exactly the same
position as the previous case, which may be called either the
a priori intuition or the proof that Mness as such necessitates
Bness. For though tlus statement 1s based upon part1culc1r
instance!> of M, these instances arc only of vc1lue because a com•
par1son of them has led us to believe thc1t B does not depend
on their spec1,1.I J.nd 1nd1vidual character, but upon the M which
they have l1l t.0111111011. In either ca!>e then the stc1tcment all
M is B represents J. connex1on, known in the one case, believed
probable in the other, bet\\ccn the umversals Mncss and Bncss.
Now this act of thought 1s clearly not an attachment of Bness
separately to each md1v1dual , for, in the case of what may be
cc1lled a priori connexion, no ind1v1<luc1l 1s con!>idcrcd as individual
at all, wlulc m inductive connex1on the value of our mduction
depends upon our being able to apply Bncss to cases which we
have not considered, and this "e do because we suppose the con•
nexion to depend solely on the umversa.l chc:1.rc1cter -of M however
manifested. Thus, in the case of a priori 1ntu1t1on, the thought
of M not contammg all the species of M, we may require the
nunor premiss to mform us th.it there ts such a species of 1t
as A. In the second case the thought of Mness has not included
every one of its realizations, and we may make abstraction of
the particularity even of those we have observed. The con-
nexion is between universals just as much as if we had never
argued from any particulars. Thus when we apply a universal
judgement or opinion got inductively to instances we have not
observed, there 1s as before a true syllogism ; we require in fact
 S,,UOgiMH
a minor premiss to inform us that M is realized in the new
instance.
§ 240. The conclusion 1s often said to be drawn from the
premisses by an act of inference. This would seem to imply
something more than the possession of the premisses, some
operation in fact performed upon them. The operation may be
described as, first, the combination of the premisses, and,
secondly, the getting the conclusion from the premisses. What
is in the first place clear is that there 1s an 1mmed1ate connexion
between the fact represented by the conclusion and the fact
represented by the premisses taken together. There is, then,
no reason why the premisses necessitate the conclusion, for they
are themselves the reason. To put it otherwise, we can interpose
no link between the conclusion and the premisses. Thus, what
JS nghtly called from one pomt of view a mediate inference (as
when I proceed from B is C and A is B to A is C) is nevertheless
essentially immediate as regard& the connexion of the conclusion
with the complex of the premisses ; and, m this sense, all
mediate inference is as immediate as that kmd of inference to
which the name immediate is usually confined. It is obvious
that we must always have 1mmcd1atc nccci,s1tatio11. Suppose
that A and C cl.re connected mediately through some hnk B.
This, we might s.i.y, necessitates that A is connected with B and
B with C, and so, 1f all conncx1on were by !mks, we must interpose
somethmg again between A and B, and so on ad infinitum. We
can state then the true chd.ractcr of the necessitation of which
inference is the apprchcns1on, as follows . ' one clement of
reality, ,vhethcr simple or complex, immed1ately and of it&elf
necessitates the existence of another clement, tlhe clements being
ddforent.' It 1s this ultimate fact wl11ch makes what 1s called
a synthetic umvcr:.al Judgement possible, and we may add that
we may call those universal synthetic Judgements, which arc not
acquired by a tram of inference m the ordmary sense but are
self-evident, inferences on this account, because they simply
mean the immediate necessitation by one thing of something
different from itself. So far therefore they share the nature of
inference; whether they arc really inferences we may consider later.
Let us now consider the subjective side. Since the objective
facts corresponding to the premisses necessitate the fact expressed
by the conclusion, with nothing intervcmng, it might seem that
 450 INFERENCE
when we have both the premisses we must have the conclusion,
and that there can be no place for an act of inference to draw
the conclusion out of the premisses, given the premisses them•
selves. This does not agree with the familiar usage of language
when we speak of drawing the conclusion from the premisses,
and yet ordmary speech 1s, in an important sense, Justified.
The possesr.;1on of the ct1nclus1on 1s certamly nol merely havmg
the prem1s1,es m their or1g111al form. The act of thought which
gives us such a premiss as Mr1B may not show us that side of
M of wluch we apprehend Mr11A (and s1mllarly for the act of
thought which gives Mr2A) Merely, lhen, lo have these pre•
m1ssci, Ill their or1g111al form \\ould no more conduce to the
inference than 1f we had them alternately We must have
somethmg morl'. lt would usually be l>atd that we have to
combine the prem1sse1, or put them together Tl11<i 1s no help ,
these terms arc mere melnphors, ,,e }1c1vc l>een their msuftic1cncy
already m their apphcat1on to processes of thought, \\hen d1s-
cussmg the l0mb111at1on, or puttmg together, of ideas. We must
look mto the naturt· of the proce:;,s itself to cxplam them. What
u, right, howe, er, 1s the rccog111t1on thc1t some act of umficabon
11, required. An,tolle rccogmzed the ncceb1,1ty of the umficabon
of the premisses ; m<lecd 1t could hardly escape notice when
logical reflection had begun at .111 ; but he did not happen to
realize the necessity of elucidating 1t, and llus causes at least an
appearance of contrctd1ct1on bct\\een the Prior Analytics and
the Nicomachean Ethics 1 In the Analytics 2 he quite clearly
held that a man may have both prcnusses and yet not put them
together, and so not have the conclu-,1011 There 1s something
in t.lus, but 1t 1i, not quite correct, and 1t require:;, careful cxamma·
t1011, \\ luch 1t did not get from Aristotle, and 1t 1s tlus which
produced the app,ucnt Lontra<l1ct1on
Suppose the order of apprchcns1on 1s Ar 1M, Mr2B Since the
1
Iu the N1wmachean J:.tl11t,, Book \'II, ch 3, 10 1147a 24 ct seq, he
11,Sl,Ume,, throughout h1-i tl1,cusb1011 of the p1<1.(,t1cal syllogism, that 1f you
have t\\o prem1sbeb :i,ou must have the conclub1011. For his argument m the
Eth,,s, the pomt of 1,,1ew of the Analvtus would have suited admirably I am
forbidden to t.i.ke bweets, tlm1 1s a sweet thmg , the necessary conclusion 1s
not drawn because 1t 1s m the interest of 'appetite' to refuse to combme
the prcm1sbeb
• p,,,o, A nall'llcs, 11 • .z I, especially, otillE• llt 1<cuAun tlll1m1 ml &,, ,;, A 0A91 ,;, B
'""'PXfl t<al WMII' 'TOVTO , ; , r, oi'lfiii,ac ,.~ lnrapxECII' ,c) A 'Tf r • •• oil .,ap i,riO'TCln1 O'TI
Til A Tf r, p~ t111v6uapiiw Tu tmfl itr6.,,par t,;a 33•
 Syllogism 45:I
second act of thought does not include Ar1M, that is, Min its
relation to A, we may not in the second act of thought, which
corresponds to Mr1 B, be thinking of M as in the relation , 1 to
A at all, and, if that be so, no inference can result, We therefore
have to remember that M, which is r1 B, was apprehended as it
is m Ar1M. But this is not the apprehension that corresponds
to the first premiss ; it 1s the memory that it was apprehended,
which memory, however, is itself a way, though another way,
of apprehending the original fact (Ar1 M). Now this act of
memory need not necessarily supervene on the two premisses.
It may be present at once with the second. The judgement
therefore Mr1 B ts modified m the sense that we now judge
that M which appears in the second prcnuss as r1 B is that
which stands in the rdation r 1M to A m the first premiss; that
is, we have the complex judgement Ar 1Mr8B. The next step is
that we sec, because of our knov. ledge of the relation r, thc1.t
Ar1 Mr1 B immediately ncce&sitatcs Ar3 B. Now Ar1Mr1 B 111c1.y be
considered as a mod1fication of Mr2B by the substitution for M as
there apprehended, viz. as apprehended m the second premiss, of
M as .i.pprchended m this present premiss Ar1Mr2B Similarly,
1f we start from Mr2B, we arrive by a smular substitution at the
same Ar1 Mr1 B. Clearly then this substitut10n is what makes
the reasonmg possible. But how 1s such modification of the
premiss JUst1fiable? Ar1M was true of M as presented in the act
of apprehension which gave us that premiss How can we
Justify the substitution for M in that premiss of somethtng
different from M as there presented ? And yet, if it 1s not
different, what could be the use of the substitution? It would
not be Justified 1£ M were confined in the first judgement to the
particular nature of M referred to m that Judgement. But 1t
is both understood that M has an indefinite margin of unknown
possibilities, and it is also understood that what is known m
the first judgement is such that 1t cannot be affected, m the
sense of bemg contradicted, by any further knowledge of the
undetermined poss1biht1es of M. The Judgement Ar1M then
cannot have its truth impaired by any further mformation we
may get about M as, for instance, Mr1B, and this is the justifica•
tion of the substitution.
The last consideration provides us with an adequate criticism
of the view that truth lies a in the complete agreement and
[• Tho referC11cc 1s to Mr. Joauum's The Nature of Truth. Cf. f§ 117, 545 J
452 INFERENCE
mutual support of all Jtems of experience, and therefore that
the truth of anything which seems to be known is liable to be
modified by future experience. The very nature of all inference
is seen to contradict this doctrine. That this is so may be seen,
in its simplest form, m any syllogism ; and, shortly, the con•
tradictton of this theory is a presupposition of any mediate
inference whatsoever, however simple.
The foregoing is very simply vindicated m the case of the
syllogism Suppose we start with the minor premiss, all A is
M, and the act of thought by which we apprehend that premiss
does not by hypothesis contain the apprehension that Mness
involves Bness, or that all M is B. Suppose next that we
apprehend this latter, the maJor premiss. We now substitute
in the mmor the information which we have about M in the
major, so that A is M gives way to A is MB. This provides
us with the so-called conclusion We may say that the original
prcm1s::. A 11, M docs not remain m consciousness in its original
form, and this is due to the substitut10n for M, as known m
the ongmal form of tlus premiss, of M as known m the maJor
prcnuss We have therefore substituted for ::.omething m the
ongmal !>tJ.tcmcnt d. new form of M The quc&tion then i& how
to Justify the suost1tut10n For if we only substitute something
which, though verbally different, is quite 1dcnt1cal with the
original Mas ortgmally apprehended, 1t is futile It must there•
fore be ::.omething different to be of any use How then can
it be justified? The reply u, that, 111 the original proposition,
our notice of M was certamly not confined to what we happened
to be apprehending m it. We were aware that there was an
indefinite margm of reality in M which \\'C were not appre•
bending ; and when we say that A is M, having a certain
defimtive apprehension of M before us, our thought 1mphes that
A has this definite character of M together with any other
definite quality m the unknown margin of the reality of M. We
obviously presuppose then, m this very simple syllogistic argu•
ment, that the truth that A 1s M cannot possibly be m any
sense impaired by anythmg which we may subsequently learn
about M, so that this simple presupposition of syllogistic argu•
mont absolutely contradicts the metaphysical theory of which
we have spoken.
§ 241. It is pa.rt of the trad1t1onal theory of the syllogism that
 Syllogism 453
Figures II and Ill can be reduced to Figure I. This goes back
to Aristotle, who, in the Prior Analytics,1 shows the validity of
conclusions in these figures by such a reduction. But, as we
have seen, we must not abstract a judgement from the judging
of it. The conclusion therefore of a syllogism must not be
regarded as a separate judgement, a result forming a judgement
by itself apart from the premisses. The full judgement of the
conclusion contams the Judgements of the premisses and con•
scquently must vary 1f those judgements vary. It is interesting
to notice that this truth is in one passage 2 virtually recognized
by Aristotle, who there insists that the definition of a property
which he supposes to be arrived at by Figure I must contain
the cause of the property This cause is the middle tenn of the
1,yllogism and therefore, as he rightly observes, the definition is
nothing but the syllogism itself written down in other words.
Now this 1s, in effect, to recognize that the conclusion is properly
&peaking the ,1,,hole syllogism 3 The result is that we cannot say
that one argument is reduced to another so long as the premisse&
c1rc different There will only be ,m appearance of reduction,
in so far as the conclus1011s have the same verbal form. But
the thought corresponding to the verbal form 1s really d1ffercnt1
and it is cc1sy to show, for mstJ.nce, that the mood Darapti, in
Figure 111, which contc11ns two umvers.11 affirmatives, cannot
be reduced to Dari1 1 a in Figure I
1
A.11 Pr 1 7 • .1 nal l'o~t u /; 1 § 220

[• ' Darapti to Darn ' WJl~on nowhere explr1.m~ wh.it he mean& by tlui..
Pre&umably he mean& thc1t the minor prem11,1, 1s converted to Y and not to I,
I.,{ s 216]
 III
PRI~CIPL~f:> A~D METHOD OF THE PURE
DEMO~STRATIVE SCIENCES
§ 242 a Ii wa!> long held from Aristotle onwards thJ.t the
syllogt!>m \\d.!> the method of cidvancc in geometry and mathe-
matics gencrc1.lly. These sciences were supposed to start from
certain gcncr..il prmc1plcs which they do not themselves demon·
strate, prmc1plc!> arrived c1t by 111tu1tion or induction, and the
progre!>s from the!>C prmciples was supposed to be made by
syllog1stlc infcrenLe Tlus theory i!> the direct consequence of
th.it. a priori theory of dcmon!>tr.i.tion wluch we have described.
It 1s owmg to such a view th.1.t we have c1t the bcgmmng of the
vcmouc:; book,; of Euchd'c; Elements a collect1011 of defin1t1om,1
axiom:., c1.11Cl post ulc.1.te ... , for these represent the effort to collect
.i.11 t.he supposed startmg prmciples or .issumptlom, We might
su&pect tlus theory on the ground that the advance in science
would then rest on no specific knowledge of the matter of the
science, beyond what was contained m the prefixed collection
of principles. It would depend solely on the use of syllogistic
forms supposed to be common to all demonstration. But this
we know is directly contrary to the fact. AgJ.m, the supposed
assumptions bemg taken to be !muted m number, their com•
binations in an argument mmt be hnutccl and so the conclusions
must be hnutecl 111 number. ln consequenLC geometry should
have been cxhc1ustcd long ago. The fact seems to be that those
[• \V1h.011 ai,i,ume, throughoul that Euclid develop:. h1., geometry syllo-
gu,ti.cally. Thu:. he :.peak-; repeatedly ai, 1£ the ax1om1> wluch Eucbd quotei.
1n h111 propo:.1tio111, were mtended to be prem1i.be, 'fh1:. 1:. a curious m11>take
It was much l<lter that Euchd'i, Elements were attempted to be set out syllo-
gist1cally An example 1s given 1n Manscl's edition of Aldrich, A,-t,s Logicae
Rudimenta, Appendix L Generally Wdo;on regarded lus own view as derived
from Kant (e g Jfritik, 1st ed. pp 707 et seq) It 1s m fact more hke
Schle1ermacher's (see e g Dialogues, § 327, quoted by Ueberweg, Logic, § 101).
Leibntz, Meditationes de cognitione (pp 540-1), puts the other view, • non
contemnenda ventat1s enunc1abonum cntena sunt regulae commums logicae,
qu1bus eti.am Geometrae utuntur firma autem demonstrati.o est quae
praescnptam a logica formam scrval • ]
 P,inciples and Method of Science 455
who hold this doctrine of demonstration have not accurately
realized what their own doctrine should mean. It ought to
mean that the original principles or premisses being granted,
any further conclusions are got by combining them into syl-
logisms. That would be the whole operation of thought in the
way of discovery and inference. We may add accordingly that
this is refuted by a very simple experiment, for if we take the
definitions, postulatt"s, and ax10ms of Euclid and try to operate
upon them with the syllogistic forms, we shall make no discovery
at all; we shall not get one of Euclid's theorems; we shall get
no conclusion whatever worth having.
§ 243. If we examine any discovery in geometry and note
down everything that is necessary, we shall find that we can
do nothing at all unless we first draw a figure. The process
advances by making new constructions, the validity of such
constructions is a matter of immediate spatial intuition, and so
also are the consequences inferred from these constructions in
the way of other geometrical relations which they necessitate.
For the discovery of the right construction there cannot be
any rule , it depends upon the manner m which the particular
case suggests the steps we have to take. Thus the finding of
the construction is the real difficulty in the discovery, and there
is a great difference between one mind and another in the
capacity of seeing what the problem suggests. When the con•
struction is found the problem is solved and the proof also is
complete , we do not need to add to it a chain of argument
such as we find in Euclid, for though we can thus explain to
a learner the connt"xion of the premisses and the validity of the
conclus1on, the best way is to retrace the process of discovery.
Further, such procedure as that in Euclid may sometimes not
show at all how the construction is found. For instance, though
in Pythagoras' theorem (Elements, 1. 47) we see by means of
the argument both that the conclusion 1s justified and the use
of the construction, the real process by which thought advanced
in the discovery of this constructwn is kept out of sight. The
process of discovery then is not an analysis of propositions
already given us ; it does not consist in putting into syllogistic
form any given material of premisses. We add to the given
material by the constructive acts, through which we apprehend
 INFERENCE
something other than the material, that is, something which
that material necessitates. We do this in geometry by con•
structing new spatial relations the validity of which we see in
the act of construction, and thus we apprehend the universal
in the particular.
Now, as these constructions are in the nature of the case
particular, how is the result universal ? The anc;wer lies in the
nature of our mathematical faculty itself In constructing the
particular we see immediately the universal validity of our con•
struct1on. The mam thought here, notwithstanding differences
of expos1t1on, 1s due to Kant 1 It seems clear that Kant thought
the method ronfined to mathematics, but that is not really so :
it is universally true that "c can only apprehend the relation
of universals m apprehending the corresponding particulars.
Though 1t may be difficult to understand this theory at first
owing to preconceptions derived from current theories of know•
ledge, 1t is nevertheless the fart, and can be dearly verified m
any actual demonstration m which we sec that we nowhere use
the particularity of the figure as such in our argument. The
ultimate fact therefore 1s that m geometry and mathematics
generally we are able to see the universal in the particular and
so to arrive at a umversal Judgement. It 1s this transcendence
~f he particular, though the particular must be used, which
rbi\ces mathematical method what is called a priori. In the
empirical sciences, on the other hand, we cannot thus see the
universal in the particular , m other words, given the connexion
of two elements m a particular experience we do not see the
universal connexion, that is, we do not apprehend the reason.
§ 244. A kmd of paradox remains to be considered : the figure
on v. hich everythmg depends may be imperfect and not an
adequate realization of the universal. It is on this acrount that
students of the subject have so often failed to recognize the
importance of the figure and have tended to suppose that the
reality of the process lay somehow in pure thought ac; opposed
to any particular and imperfect constructions. It is this fallacy
which 1s mainly responsible for the chimera of non-Eucltdean
spa<'e. The writers who mamtmn th1S do not realize the quite
1 1,,,.,,1, d , l"er,11111(1 (1~t ed, p j12; ~nd ed, p 740], Thel"E' 1s an
antic1patton, however, of the tntth in An'!totll', .11,fttaphysus, 1051• 22,
 Principles antl Method of Science 457
fatal objection which lies in the impossibility of any construe•
tions to correspond to such space, This is because they do not
realize the place of the figure in Euclidean geometry. They do
not even suspect that it is precisely the construction of the
figure which is the true process of the advance of thought in
geometrical discovery. In truth a thing cannot be ' thinkable
but not representable '. The paradox we are now considering,
the possible imperfcctton of the figure, cannot really be removed
by supposing that we imagine a perfect figure. A figure in
experience might be m itself perfect. A falling rain-drop, for
instance, might be a perfect sphere or an accurate ellipsoid, yet
we should have no means of ascertaining that it was perfect,
and so its perfection would not be experienced by us. Now
there is exactly the same difficulty about any figure in our
imagination ; we may want it to be accurate but we can no
more be sure that it is so than if it were actually found in our
experience. We certainly must thmk of a perfect figure, for on
that our proof depends, but that is not the same thing as saying
that the figure in our imagination is perfect ; we apprehend,
that is to say, a particular, and in some sense we apprehend the
perfect particular. That apprehension is not an experience and
not an imagmat1on, but, nevertheless, it 1s an apprehension of
what the nature of the particular must be. This is undoubtedly
true, and 1t remains to attempt to characterize this kind of
apprehension. We seem to arrive at a new and important
distinction m our faculty of apprehension
We must clearly apprehend the nature of the perfect individual
figure, for on that our proof depends. But the apprehension,
as we have seen, is not an experience, because though a particular
given figure may mdeed be perfect we cannot apprehend 1t as
such with certainty in experience. Nor 1s 1t the apprehension
of an imagined particular, and for the same reason. It is an
apprehension which we have with such experience and with such
imagination. This peculiar kind of apprehension does not
appear to have been recognized in any theory of knowledge or
of inference. The fact 1s that the difficulty above described
has not been recognized and the solution of it has therefore not
been suggested.
§ 245. It may be objected that in actual geometry the axioms
 INFERENCE
do appear as major premisses and that although we must admit
that the process of discovery is not syllogistic, in the strict sense,
yet when the theorem has been found and the connexion is
presented as something to be understood, it consists of syllogisms
as in the ordinary Euclid. We must distinguish two kmds of
major premisses. Axiomatic premisses, which are not demon·
strated, includmg defimtions, and premisses which are them•
selves products of demonstration.
To take the former first. Euclid shows (Elements, i. 1) that,
in the triangle ACB, AC and BC are- equal, each of them, to
AB, and being equal to the same thmg arc equal to one another,
because of the ax10m. Here the axiom appears in the last step
as the maJor premise;. This we have explained already. The
truth of th<' axiom is only apprehended m the individual instances
which we here pretend to derive from 1t. In the case before us
we see the truth directly m the particular mstance or we should
never see 1t at all, and therefore to represent 1t as deduced from
the axiom is an mvcrsion of the actual order of thought. We
recognize, however, that the truth we see has a universal
character, not depending on the particular magnitudes, and this
recognition of the umversal is prer1sely the axiom itself and
follows after that partJcutar act of thought which is wrongly
represented in the syllogistic procesc; as derived from it.
§ 246. We do, however, find m geometry cases where there
seems to be a real syllog1sttc process; where, as it would be
said, the result of a prev1ouc; proof 1c; applied to a particular
instance Consider the proof (Elements, 111. 31) that the tnanglc,
of which the bac;e 1s the diameter of a circle and the vertex
a point on the circumference, has a right angle at the vertex.
Jf this theorem ic; subsumed as a mmor premic;c; under the truth
of Pythagoras' theorem ac; a maJor premiss, we infer the same
property for this triangle. This may be put thus . If two
straight lines are drawn from a point m the circumference of
a circle to the extremities of a diameter, the sum of the squares
on the two lines is equal to the square on the diameter. The
act of thought, viz. the whole proof, which here gives us the
rnmor premiss, does not give us the major, we don't thereby
prove Pythagorac;' theorem. Again, the act of thought which
gives us the maJor premiss does not give us the minor, does
 Principles and Method of Science 459
not, that is, prove Elements, iii. 31.1 The act of thought there-
fore which gives either premiss does not give the other and thus
a condition of valid syllogistic inference seems to be satisfied.
Here, then, there seems to be a syllogistic process. It is of the
simplest character, an operation found in all departments of
thinking, causmg no d1:fficulty, and not depending, be it observed,
on any knowledge of the special matter of geometry. On the
other hand, what is most valuable to the progress of the parti-
cular science as such is not this mere obvious subsumption of
the particular under the universal, but the process by which we
discover the universal in the particular which is subsumed under
1t. That depends not on the syllogism but on construction with
apprehension. Now when we examine more carefully the nature
of the major premic;s, we shall find that m the actual process
of thinking whereby we know 1t fully and do not merely
remember the result of the demonstration, m other words, in
the full apprehension of the Judgement represented by the
major premiss, the syllogistic subsumption disappears. Strictly
the maJor premiss, 1£ the Judgement expressed m that premiss
is properly Judged, can only be applied by our reproducing the
whole proof. If we do this, we find that m the case of the triangle
m the semicircle we first demonstrate by construction that the
vertical angle 1s a right angle and then go on by a system of
constructions (which is nothing but the proof of Pythagoras'
theorem) to prove that the square on the base is equal to the
sum of the squares on the sides In this complicated geometrical
proposition, JUSt as much as m any of the more primitive ones,
provided every judgement used 1s treated m the same way, we
shall find the whole process of thought consisting of these con·
structtons with the corresponding apprehensions. We shall find
no process of advance by syllogism here any more than there.
The syllogism also itself will agam have the same explanation,
as arising from the recogmt10n of the universal character of our
construction'> and their consequences.
§ 247. If, then, instead of conducting the proof in the particular
case that A 1s B and the proof in the particular case that Bis
C, we merely remember mther result or both, and so do not
fully apprehend the facts correspondmg to the premisses, we
1 The angle m a sem1c1rcle 1'1 a right angle.
qn~ E
 INFERENCE
apparently have the syllogistic- form, all B is C, all A is B,
therefore all A is C ; and that is where the syllogistic form
would seem to come explicitly into geometry. But we must
not suppose that the syllog1stic form is related only to imperfect
apprehension and due only to that. It corresponds to some•
thing m the romplete proof, that is, in the complete act of
apprehension. Consider the complete proof when the syllogistic
form does not appear. In the case of a particular triangle in
a semicircle represented by the figure before me, where ABC
represents the triangle itself, I apprehend by Euclid's proof that
the triangle must have a right angle at the point represented
by B. Then adding the construction of Pythagoras' theorem,
I appr?hend by Euclid's proof that the right angle at B neces•
s1tates that AC1 = AB 2 + BC2• I thus apprehend that the
particular triangle m the partirular -;em1circle has this relation
be-tween the squares on its sides. Bui now the process is seen
not to depend upon the particularity of the figure, and thus we
apprehend the universal, namely, that any triangle m a sem1-
c1rcle has a nght angle at the vertex and, bcrause o{ that, the
relation bet\\ <"Cll th<" squares on its sides The verbal expression
of the generahzatmn \\hich we have thus apprehended m the
consideration of the particular 1s that any such triangle 111
a sem1c1rcle has a right angle; any triangle which has a right
angle has the Pythagorean property; t hcrefore any triangle m
a semicircle has the aforesaid property Thus m the properly
compl£>te apprehensmn we have this syllog1sm, not, however, as
the process by means of winch "e have discovered and inferred,
but a-, 'lomethmg the truth of "htch has only been fully appre-
hended 1n the precedent apprehensions concerned with the
particular figure In other words we only apprehend this
syllogism fully by the prorr'ls m question. Now suppose I have
neither proof before me. I remember that I proved that the
triangle in a semicircle has a right angle, and l remember that
I proved that a right-angled triangle bas the Pythagorean pro•
pcrty. Even 1f it 1s obJcctcd that memory 1s not trustworthy,
I know at least that 1f the facts are as I remember them, they
necessitate that any triangle m a semicircle has the property in
question Thus I have a true hypothetical judgement. But it
would be affectation to say that tl1erc are no cases m which
 Princi,ples and MNhod of Sciencs .¢t
I do not absolutely trust my memory. I know, for instance,
that it was proved that the triangle in a semicircle has a right
angle ; then the facts represented by the premisses are known ;
each is apprehended, the apprehension consisting in a memory
that it was proved. But clearly the knowledge of the syllogism
is incomplete as compared with the knowledge of it in the
complete apprehension above described and depends in the last
rec;ort for its validity on the fact that we can replace the memory
of the proof by the proof itself. We thus see how the possession
of the syllogism in the one case differs from the possession of
it in the other, although the difference does not appear in the
verbal expression. It follows, then, that the syllogism is not
due merely to defective apprehension ; it exists anyhow but is
only fully comprehended in the case where we have the proofs
of which it is the generalization.
§ 248. The foregomg 1s not confined to geometry or mathe-
matics, but holds of any syllogism in which one or more of the
premic;ses 1s a universal proposition, the truth of which is known.
For such a universal, if really known and not merely probable
and based on mduchve evidence, can only be apprehended in
the particular; that 1s, we can only apprehend that all A is B
by apprehending m an instance A, that A 1s B as being A. This
dependence then upon the particular is not confined to mathe-
matics as Kant seems to have supposed.a
The conclusion to which we are led may be summed up as
follows. The true syllog1stic process in rcatfoning being the
possession of a major premiss followed or preceded by a minor
which is not included in 1t cxphc1tly ; the apl>rchension of the
minor again not including that of the maJor, and a conclusion
following which is necessitated for our thought by the possession
of the premisses, we never have such a process in mathematical
science, unless the major premic;s ic; possessed, not as the appre-
hension of the facts stated m "it, but as the memory merely that
1t has been apprehended. Thus the syllogistic form as a process
ra 'Nur die Mathematik mcht aus Begriffen, sondern der Konstruction
derselben . 1hr Erkenntmss able1tet . . Das philosoph1sche Erkenntniss
das Allgememe Jederze1t. rn abstl'aclo (durch Begriffe) betrachten 111uss,
indessen dass Mathematik das Allgemeine ita concreto (in der einzelnen
Anschauung) und doch durch reme Vorstellung II pnon erwtgen kann.'
I c, p 734 (762, 2nd ed ) ]
E2
 INFERENCE
of thinking is followed in mathematics only when knowledge in
the most complete form is not present, that is, when the facts
stated in the maJor premiss arc not apprehended in the foll and
strict sense.
Such a syllogism 1s not complete in itself, because it depends
on the absent proof of the major premiss. Thus such a syllogism
could not be an absolute startmg-pomt in knowledge. Now 1t
can easily be shown that tlus must always be so in the case of
universal statementc; m matterc; of knowledge, not merely in
rt1athemat1cs, as Kant sel.'ms to have supposed. Obviously the
only rase we have to ronsidcr 1s that in which the major premiss
is not d<'monstrated. (Induction 1s excluded bcrause it cannot
give knowledge but only a high degree of probab1ltty. Even 1£
1t could give knowledge, that knowledge would be demonstrative
knowledge and th<' syllogic;m would depend upon it for 1ts major
premiss) The maJor then in question must be self-evident.
But the universal fart represented by the universal c;tatement
can only be matter of appr<'hens1on 111 a partirular, for we rannot
possibly apprcht>nd tht> nature of any umversal except by
thinking of a particular in which 1t 1s manifested. Now if the
universal 1c; sclf-ev1dent, it 1s seen in any partirular whatsoever,
and therefore our apprehension of the nature of any of its
particulars involves that of the universal. It cannot then be
preceded by the apprehenc;1on of the umversal so as to be sub•
sumed under 1t lf, for inc;tancc, the umversal 1s all B is C, and
the mmor all A is B, the A's arc particular mstanres of B; the
apprehension then of any of them involves the apprehension
that A as being B 1s C It 1s only after reflection on this that
we see that the thmg's being C depended only on its bemg B;
whirh 1s just what produres the universal statement represented
by the major premiss.
It 1s well to add here for clearnci:;s that the formula 'seeing
the universal in the particular ' means that m the case, for
example, of a particular A, though I am obliged to think of
a particular, I apprehend that 1t 1s B only as being A, or only
because it is A, and not through anything particular or confined
to itself.
f 249. The validity of the general form of the syllogism, that
is, of the syllogistic rule, is like any rule of inference recognized
 Principles and Method of Science 463
in any particular syllogism, and the particular syllogism is, ae
we have seen, not deduced from it. The question then anSQli
whether every particular syllogism is not itself similarly the
generalization of a particular and the apprehension of a universal
m a particular. Suppose that tn the syllogism B is C, A is B,
therefore A is C, the statements BC and AB are full apprehen•
s1ons ; we have then a truly known universal and we can only
apprehend, as was said above, the universal all A 1s C because
r11l B 1s C and all A IS B, by considering a particular. The only
question is how this 1s done. Take the particular A1 . I appre-
hend that A1 as A must be B, and so 1s B1 (as m the geometrical
example). Agam, to get the universal ' all B 1s C ' when I can
really have such a statement as certarn, I apprehend m B1 that
.ts B it must be C. thus I apprehend A1 as being B because
1t is A, and as bemg C because it is B ; but to c1.pprchend A1 as
B because it is A, is to apprehend A1 to be B not as a particular
but only through its umversal ch,uacter A. Thus we apprehend
that all A is B or any A 1s B. The above apprehension then is
apprebendmg cl.ny A to be C because any A 1s B and any B is C,
which is the syllogism, and the syllogism therefore is a universal
apprehended thus 111 r1 particular.
Next, suppobe we only remember that c1.ll A is B was proved,
that is, w.is apprehended 111 c1. particular instance, and so for
all B 1s C It 1s not enough to say I think of c1. particular A,
and imagme 1t B, tor that is not the Judgement any A must
be B which 1s required for the syllogism I remember that A1
as A was shown to be B. This 1s an c1pprehcnsion through
memory tllclt A1 db A tb B (viz. B1) S1mtlarly we have an
apprehcnb1on through memory thc1t B1 as B lb C Thus though
the nature of the apprehension is d1fterent, the general form 1s
the same as in the prccedmg case c1.nd can be treated in the
same way. Thus \\C agam have the syllogism as a universal
apprehended m J. purt1cular case. Lastly, suppose I am uncer-
tain whether A 1s B and whether B fr, C, yet I know that 1£
A is B and 1£ B 1s C, then A is C. I tlunk of a particular A,
viz. A1 • Now this means that I apprehend the nature of an
A as particular, while 1magmmg a particular (1f I am not
experiendng or remembcrmg an actual particular), and I appre•
hend the fact that 1£ a particular A, ab A1, were B, and if as
 INFERENCE
B it were C, A1 would be C, wluch 1s apprehending also that any
A, if it were B, would be C, if all B is C.
1250. We can now determine more precisely the nature of
what is called the application of the untvcrsal to the particular,
a phrase by which the syllogism is sometimes characterized, and
whether or not this 1s ever an inference. We shall be able at
the same t1me to see what 1s the ground and excuse for the
difficulty m.ide about a petitio principii in the major premiss.
Suppose, as m the example given from Euclid, we prove m the
instance B1 that all B 1s C, which is equivalent to saymg that
we apprehend that B1 1s C because it 1s B. Suppose we apply
this, let us say, to a new mstance Ba If now m so doing we
fully apprehend that all B 1s C, that means we apprehend that
B1 as B is C, and this will be a repetition of the former proof.
lfore there 1s no mfercnce that this B ( = B11) 1s C from all B 1s C.
But now suppose we do not thus prove that B 1s C but remember
that we proved it ; we might express the process thus all B
was proved to be C, therefore tlus B ( = B2) is C. This reasoning
has the verbal form of an mfercnce. In proving that B was C
wc- proved rcaJly that any particular D was C, for we found
a particular B1 "ai, C, only because 1t was B, and in tlus we
recognized that any B was C. Thub th<' rewgmt1on that tlus
B ( = B2) is C, bccam,e 1t H, B, 1s no new act of apprehension
but 1s truly comprised m the previous apprehension. Now this
1:. the true (though not understood) ground of the feeling that
the maJor premiss really contams the conclusion, for however
the Judgement all B 1s C 1s got, 1t is felt to mean that any B
whatever, as B, 11, C. Now the feeling would be right 1£ B1 was
apprehended merely as a particularization of B, but this is not
the case. B8 1s apprehended as a particularization of another
umvcrsal A, so that B2 1s also A2 , and A2 m fact 1s 1dcnt1cal
with Ba, The recogmt10n of that 1s the mmor prem1ss. Now
the apprehension that any B, ai. B, 1s C, represented by the
major premiss all B 1s C, did not m any way contam the appre•
hens1on th.it any B 1s A, and so not the apprehension that A11
is B, and consequently not the apprehension that Az, as B, is C,
which 1s the conclusion. We thus come back to our former
solution,
§ 251. The first prmc1plcs of geometry arc ui:.u.i.lly divided
 Principles and Method of Science 465
into definitions, axioms, and postulates ; the last appearing as
practical rules stating what constructions are allowed. Properly
speaking, however, the science of geometry as such is not con•
cerned with the practical construction of geometrical figures. 1
The real value of the postulates is that they represent an imper•
feet recogmtion of the necessity of constructions, that is, of
particular figures, either m experience or m imagmation. They
should then be a classification of various kinds of theoretic
construction. The classification as given 1s by no means ex•
haustive. It does not even exhaust the constructions m the
Elements of Euclid ; nor 1s this surpnsmg, since the classification
was made without any consciousness of its real significance.
But, though it might have been exhaustive of what 1s actually
found m the Elements, it could not possibly be complete for
geometrical science. The various kmds of construction only
become known in the solution of various problems and they
cannot possibly be anticipated beforehand Thus there is no
prov1S1on m Euclid's Elements for the various fam1har construe•
hons of conic sect10ns.
§ 252. Whether m geometry or elsewhere, axioms are usually
regarded either as simple mtu,it1ve truths and contrasted as such
with truths which are mferrcd ; or as not intuitive mdeed but
the result of an mductive process. This second view 1s artdic1al,
ansmg solely m the interest of a given theory. Even m the
second altcrnallvc, ho"'cver, they arc still considered to be
sunplc by contrast with the complexity of demonstrated truths
and not mferred by a process of demons tr atlon strictly so
called. We shall mamtam that the axioms are, on the one hand,
not the result of induction, and, on the other, are, though
certainly intuitive, not to be distmgu1shed from demonstrations
as these arc usually regarded The demonstrat1011s themselves
are all mtuitively apprehended, and 1£ they arc got by some
construcllvc process, so cire the a..""-1oms Every axiom is syn•
thetic and not a mere 1dcnt1ty Thus it contains d1fferent
clements whose conncx1011 1s apprehended, and this apprehen•
s1on takes place m a -construction. We must, that is, have
a particular :figure imagined or perceived, and the umversal
validity of the axiom is seen 111 the particular construction
' V1de Plato, llepubhc, Bk. VI, 510 D.
 INFERENCtt
because we there see that the connexion apprehended is inde•
pendent of the part1culanty of the instance chosen. But this
is exactly the process that we h.1.vc described already for
demonstration The construction for an axiom 1s simple; in
demonstration 1t 1c; more complicated. The usual d1stmction
then between ,txtom" .md dcmonc;trat1ons 1s quite unreal ; the
act of thought 111 ,duch we .tpprchcnd an axiom or a demonslra·
tion 1-, cxartly the !>.tmc m kmd m both cases. Agam in
dcmonstr,1t1on, the '>Uccess1vc constructions arc all mtmtive and
may all be t crmed axiomatic ; the mference from one end of
a ,;enc& of con'ltruc-t10n'i to another 1,:; 1 "hen generalized, an
.1.x1om lt'l meamng, hov.cver, 1& rccogmzed only 111 these very
acts ot con,,trurtJon .
§ 253 In E111'!1d we find two kmd& of dcfimt1on. Both refer
to rc,thty, but one kmd '\11,C mc1.y call nommal as m some sense
g1v111g the meaning of c1. name without the construction necessary
to decide "hether any re.ti obJcct corresponds to 1t or not.
Thus the defttution.., of obtu.,c•, acute-, c1.nd right-angled triangles
.i.rc .t.l fir&t 110111111.tl, hung only Jtl',l1ficd when Euclid has proved
th1.t the t hn·e mtcnor angle.., of ,1 t n.1.ngle are together equal
to t\\o n~ht angle'> Tluo; restricts the number of right or
obtus<· c1.ngk!> to one ,tnrl allows ,l tri,mgle to contam three acute
,tngJ<'.,, Tlrn!> dcfiml1011 1.., of no u,;e ltll 1t 1.., Jtl';t1ficd by construe·
t10n, m other ,,ord-,, by dcmon'>tr,tt10n There 1,:; .t scLond kmd
"Inch t!> not nomm,tl . for ex,.unplc, the defimt1ons oi a tri.i.nglc
or ,t urdl', bt•c.tU'>l' \\C sec the poss1b1hty of those figures at
onLe, ,m<l \\t' m,1v think ,,c <lo tlus befon· conducting .i process
ol micrcnce, !>UCh a"> 1s ncccs5,tr) 111 the former kmds of dcfimt1on.
But the truth t!> tlMt \\C sec tlm poss1b1hty only by a particular
conslrul't1011, "l11ch a-.sures U!> m the ca,:;e of a triangle, c;ay, that
&p,tcc l .in be cndm,cd by three .111d not by two straight Imes.
So for the definition of a circle We see that the revolving•
pomt, at a hxed radius, must return to the same pos1t1on and
so form a closed figure It 1s the ,;1mpltc1ty of these constructions
"Inch c.im,es them to be overlooked, and thus the fah,e idea
hc1s ,mscn that a dehmt1on 1s merely the description of a 'con•
cept1on' (.1.11 expression which will hardly bear exammatlon) and
distinct from the acts of thought which involve proof. The
necessity for a con!>tructivc process becomes obvious in the
 p,-i,nciples and Method of Science 467
definitions which we have called nominal, and this helps
us to see that there is construction involved in the simpler
cases also.
§ 254. In the empirical as compared with the mathematical
sciences we observe a very important difference in the position
of definition. In an empincc1.l science we cannot start from
a definition and advance from it as we appear to do 10 mathe-
matics. The diffcrcncc cannot be accounted for by the ordinary
theory of definition, which 1mphC!, more or less unconsciously
that defuut1on only describes the content of conceptions ; a view
which depends on a false dntlthcs1s, a wrong distinction between
obJects and conceptions As to tlus ant1thes1s 1t is supposed,
for example, thdt w1th10 limits we can make conceptions at will
and so make definitions It 1s c1.dm1tted that there are elements
which we c.mnot mdke 10 thought but must accept from experi•
ence; given these, we arc still supposed to have a freedom of
combmat1on. Dcfimt1011 thus gets a subJectJvc arbitrary
character and cannot be treated as somethmg from which we
can derive mformatlon about reality If this were 10deed &o,
we could not underi,tand the posit10n of definition in mathc•
mat1cs. Tlus conceptualist1c view depends upon a m1sunder•
standing which may be illustrated trom Mill's Logic. Mill th10ks,
as we have seen, that "e cc1.n make a dcfimt10n of a dragon,
that this like dll dcfimtrnrn, 1& merely nommc1.l, thc1.t 1L only
cxpresi,c1., the meamng of a name, \\ luch he 1.,omet1mes appears
to distinguish even from an idea. It turns out, however, that
there 1s somethmg behmd the name. He means that to 'dragon'
corresponds an idea or tonccptJon so mdependent of reality that
we do not believe that a correspondmg obJect can exist. He
then seems driven to cxplam why the procedure of mathematics
results m no mere 1magmary theorem<;, by supposmg that to
a mathematical defimt1on 1s .tddcd a postulate of the reality of
its object. But, more than this, Mill's expressions would commit
lum to the view not merely that no deduction ' affecting matters
of fact' 1s possible from the dcfimtJon but that no deduction
at all is possible from the definition unless the postulate is
added.1 That he docs not realize this is probably due to the
strange confusion by which he d1stmgu1shes the view of defini•
1
.Mlll, A System uf L~gic, I. Vlll, §§ 5-0.
 INFERENCE
tion as a • statement and analysis of the mere meaning of a word•
from definition as • the statement and analysis of the content
of an idea '. What from his pomt of view could ' meaning ' be
but idea ? In the second book, \\ here he is discussing mathe-
mat1cc1.l demonstrc1.t1ons, he has to assume that deduction can
be made from the content of the idea itself apart from the
postulate. 1 The former doctrme, however, 1s obviously absurd,
for ckarly the poss1b1hty of deducmg anythmg from a conception
hes m the 'content' of the conception itself, and, supposmg the
content (ac; 1s presupposed) were somethmg different from reality,
the postulate of existence could m no way fac1htate the process
of deduction The only effect it could have would be that,
while the mfcrent1c1.l process proceeded from a conception, exactly
as 1£ no postulate had been made, we should merely add at the
end of every deduced conception the words ' and this conception
correspond!> to a real obJect' The existence therefore of a
postulate would not explam the peculiar proce!>s of mathematics.
Mill's account I!> a part of the old nustake that mathematical
mference 1'> ml'rcly formal, cons1stmg of syllog1stic deduction
from g1vc11 prenw,i;e!>. It 1s amusmg to observe that Mill thus
mvolves l11mself m a characteristic cont_rad1ct1on, for in the next
chapter, 10 order to meet some difficulty about ma.thematics, he
m..i.mtams that re.ii obJccts do not correspond to mathematical
defuutious, !>O that thl· po!>tulc1.tc after all 1s false and mere
hypothc~n, ll
§ 255. We have seen that m the e,a!>e of the dcfimt1on of
a dr,1gon it has been m,\intamcd that \\e can h.i.ve an 1magmary
conccpt1011 to \\lmh no real C'bJel·t corrl'sponds. To that we
ans\\ er firbt th.it these rnnu.:pt10n!> turn out on analysis not to
be conceptions at all Ill the strict sense They have clement!>
supposed to correspond to parts of reahty and in thmkmg of
them \\C v10latc no rule necessary to reality. The dragon's body,
for mst.i.ncc, els c:.\.tended m space 1s never thought of as v10latmg
a gl·omctrical prmc1ple. But the wmbmat1on of these clements
(hvmg tisl>UC and fire) 1s not really performed m thought and
1t 1s quite untrue that "e have combmed or can combme m
I 1d , II v, §\ 1-.Z
1 It l'i trul' that he trn:~ (um,uctt.'l>li!ully) to p,n e the w.i.y for t1u11 by
a change of front 111 llook I, , 111, § 6.
 Princi.f,les and Method of Science 469
thought what we know or believe cannot be combined in
reality.
Secondly, if the theory is put in the form that we can conceive
objects which we know cannot exist, it is easy to show that this
is self-contradictory. We can only know that these elements
are not combinable by our knowledge of what these elements
are, that is to say, by our ' conceptions ' of them. It is the
conceptions, therefore, which have this incompatibibty, and that
JS cqmvalent to saying that we cannot combme these elements
m conception.
The question also 1s easily settled as a corollary of the account
already given of the meamng of conception as properly appre•
hension of reahty. 1
§ 256. In empirical science there appear to be two courses
open to us, m the matter of definition, if we assume certain
clements as apprehended 111 experience.
We might define an ob3ect either .ts combuung certain of
these clements, without refcrrmg to experience to sec whether
the combmat1on occurred, or as combmmg clements which we
have observed combmed. In the first alternative there could
be 110 security that something real corresponded to our definition.
Yet, if a not uncommon view of dcfimtJon m mathematics were
right, we could sec no reason why we should not construct
a body of hypothetical physical science with the same prec1S1on
.t!:> m geometry. But we know that this 1s not possible The
second alternal1ve, where we at least know that a real object
corresponds, seems tl).crefore preferable But though this 1s real
knowledge, Jt would not enable us any better to draw conclusions
from the defimt1ons The reason 1s one which ts common to
both defimtions. As we have seen, a conception 1s not com·
pleted when we have stated the clements of it and added merely
the demand that they arc to be combmcd. It JS obviously not
completed until we know that they are combmed. This ts the
position in empirical definition. Even when we apprehend m
experience certam clements, we do not sec how they arc com•
bmed. In geometry, on the other hand, we not only apprehend
the elements but also how they arc put together ; we understand
the nature of their umty and we realize 1t to ourselves in an act
I Put 11, 1.h 14, Cl>J,JCI.Jally p. 315.
470 INFERENCE
of construction in which the unity of the elements is as much
apprehended as the elements themselves. We know exactly
how they go together Now this 1s jUSt what we cannot do with
the elements of J.n empmcal dcfi111t1on. There 1s something in
language wluch corresponds lo this. In a. statement of what
an empmcal object 1s, \\C find attributes m general only unified
by their refcrenle tu the r:,amc bUbject ; they appear as so
m.my beparJ.lc ,idjccllvc~, CJ.ch atlJ.chcd mdcpende11tly to a sub·
stanl1ve-· Um, body 1s h,ud and round and sweet', &c. In
a mathematical dcfin1t1011, on the other hand, the words form
an orga.mc grc:1mm,itu. al umty and not a mere aggregate of
clements Jomed by 'and'. We <lo not say a triangle has three·
foldnc:.s and sides c.1.nd lb rcctibnear and encloses a space.
Knowmg how these clements arc put together, we express their
umty thu& : ,t lnangle 1s a three-sided rect1hnec1r closed figure,
or (better), c1. triangle has three re1..l1lmear sides wluch enclor:.e
a. r:.pa1..e. 1
~ 257.• Tlll' usu..11 theory 4 that lhc exact r:,uenceb stc1.rt from
defimt1ons of their obJectb, ass1g01ng them crrtam c1.ttnbutes,
and then derive other facts about the objects from these defini•
t1ous, involves a sharp <l1stmct1on Lclwccn the attributes com·
prised in the dchmt1on, and those • derived ' from them by
demom,trJ.l1on. Tlus d1stmct1011 1s connected with a technical
distmct1on of ci,i,cnle from property , essence l>emg the attributes
constituting the dcfuut1011 A further d1stmcl1on 1s to be made;
not everythmg which 1s necesl>1t.i.tcd by the essence 1s property
m the stncl techmcal sem,e. A property 1s a necessary attribute
wluch 1s derived from the cr,r,enrc alone .i.nd not from anything
else. Thus, 111 tn,mglc, that the mtcnor angles are together
equ.tl to t\\ o nght .inglcs 1s a property, because 1t 1s derived
from nothmg ebe th.in the essence m the defimt1on. But that
the exterior angles formed by producmg the sides successively 111
the same d1rectio11 arc cquc.1.l lo t.\\o nght angles, though necesi,ary
to J. triangle, 1s not a property, because 1t follow& from some·
llung lesi, th,m the whole essence and 1s true of every rectilinear
1
U § z59 ad tin • Cf. §§ 17 3-4

[• 011 AnstoUe'b d()(.tnne, ~ee Mr Joachim's .iristotle on Commg-tu-b, and

Pasu11f-11W11y, lntroductton ]
 Principles eintl Methoil of Science 471
figure of a certain kind (viz. without re-entrant angles). The
distinction goes back at least to Aristotle. For him property 1
as something both necessary to the subject of the essence I and
peculiar to it is distinguished from essence, but he nowhere
explains how it is that of two kinds of attributes, both necessary
and both peculiar, one kind should be called essence. We may,
however, derive an Aristotelian definition from the actual use
which he makes of the d1stmction. What is everywhere pre-
supposed in the Posterior Analytics comes to this · the essence
1s a group of attributeo; of the thmg defined, such that no one
of them is derivable from any other or others, while everything
<'lse necessary to the tlung ic; derivable from this essence. In
that treatise the terms 3 for the,;e derivatives which are pro-
perties suggest an existence in some way dcnvative and depcn·
dent. The distinction was very likely suggested by geometrical
science, and may seem to acco,rd with the procedure of Euclid ;
yet the conception, ,f we examine it, seems almost paradoxical.
For how can we know the essence except by knowing that it
acC'ounts for everything in the object? And how can that be
done unless we know everything 111 the obJect and sec that the
essence accounts for everything ? Surh knowledge is clearly
impossible The answer might sec-m to be that we may at any
rate approximate to the e~scnce. We may divide what we know
of a thmg into essence and property, essenC'e bemg the group
of attributes which we cannot derive-, and from which we can
derive the rest of what we know. Essence then would be
relative to the amount of our knowledge, increasing when we
found new attributes v.h1ch we could not der1veirom our original
essence, and dccreasmg when, through improvement m our
knowledge, we found some part of the supposed essence to be
really derivable from some other part. A perfect definition of
essence, then, would never be attained ; essence would m fact
be an ideal to wluch we are always trying to approximate, and
that fairly agrees with the actual condition of the empirical
sciences
§ 2 58. This answer to the difficulty does not, however, suit
the mathematical essence That is not provisional ; we are sure
l i3,o,, I OUfJ(a.
• tr~,-/1•/J'l"i' ttal' alrr6, 1"'1a, or 1tt1l'II'" d' a.bT6.
 INFERENCE
that it entirely covers tl,e object to which it relates, and yet
we could never verify our belief by seeing that it does in fact
account for all the otper attributes, for these are infinite in
number. Agam m mathematics we appear to get our know•
ledge of the properties from the essence instead of depending
on a priori knowledge of the properties to determine the
essence.
Now how can we be o;ure that the definition does cover every•
thing ? The answer, though in a way decisive, seems at first
a mere tautology When m geometry we define a given figure
we mean by the object of the definition, an object which has
only the attributes assigned in the definition, and those which
necessarily follow from them. This account is obviously con•
firmed by geometry 1to;elf, and m this way the sufficiency of the
essence io; guaranteed from the first. Yet it seems to involve
a very great difficulty. We naturally ac;k why we c;hould not
secure the same completeness in physical science by statmg that
the object defint>d 1s only what 1s determined by the definition.
Now o;omettmes this could not be clone, the essence of gold,
for instance, 1s something given to us as a problem; 1t 1s not
constitut<'d by, or apprehended 111, our construction. We may
answer, however, that m geometry we are not concerned with
the total reality of an md1vidual thmg as we are when studying
gold, but only with a part of it, in other words, with its shape.
But in physical science we cannot even make a definition of an
object, that 1s, a part of reality, which does not embrace the
whole reality of the object. We do not know, given the clements,
how they arc combined In geometry we can take any elements
whatever, the umty of which m a \\ hole we can apprehend in
a construction. Whenever this can be done m an immediate
intuition realized in a construction \\e have an obJect which is
a unity of different elements and have thus completely deter-
mmed an object, not necessarily the total reality of an individual
thmg as in physics, but still an object. Such essence 1s an
aggregate of elements o;een to cohere directly 111 a umty, without
the intervention of conne<'tmg lmks We must here define more
precisely what this coherence means. Suppose the elements A
and B in this umty necessitate C : we should not include C in
the definition, though A, B, and C cohere. What we do mclude
 Princi,j,les and Methotl of Science 413
are the elements which we can see may form a unity but do
not necessitate one another, for example, a three-sided rectilinear
figure. Threefoldness does not necessitate threefoldness of sides ;
three sides do not necessitate that space should be enclosed, and
a three-sided closed figure is not necessarily rectilinear. Now
such a construction leaves the elements indeterminate in some
directions. In the definition of a triangle we see the coherence
of the elements in the unity of one object without determining
the magnitude of the sum of the interior angles, though they
must have some determinate magmtude. So again we leave
indeterminate the relations of the sides as depending upon the
relations of the angles, though these again arc in themselves of
course determmate. This indetermmateness, however, and the
ignorance which 1t implies does not prevent us from appre•
bending the determinate character of the unity of the elements
so far as we have apprehended them. We see, for instance,
that space can be enclosed by three straight Imes, though we
are as yet ignorant of the properties JUSt referred to. Now
what arc called the properties of a figure as opposed to its
definition or essence are simply the further determinations of
the parts left indeterminate m the combination represented by
the definition, which determination'! are cond1tioned by the
possibility of further determination in each clement taken by
itself and by the nature of the umty m which they are put.
These new determmations arc discovered by that process of
construction which we have described and which corresponds
to what is usually called the demonstration.
§ 259. The idea of essence as the undcmonstrated from which
all else in the object can be demonstrated conveys the impression
that the essence 1s something absolute. In Aristotle the distinc•
tlon is absolute, yet we find that different definitions may be
given of the same geometrical figure, for example, of the ellipse.
Each bcmg an independent starting-point for the investigation
of the whole figure, and each demonstrable from any other,
property becomes essence and conversely. Now this can be
explained from the account already given of geometrical defini•
tion. The definition is nothmg but the apprehension of the
unity of certain elements, the pomt being that we really do
apprehend this unity and do not merely suppose it to exist. This
414 INFERENCE
we may put subjectively as follows : it is the conception made
by combining certam elements and we really do combine them.
This subjective expression seems to answer to the act of con•
struction but 1s m1slcadmg ; 1t is not true that we make the
combination, we apprehend or know the nature of the combina•
tion m the empirical act of drawing the figure. Thus the elements
and their combmatu,n arc m no sense arbitrary, but necessary.
The further procedure 1s simply the apprehension of further
connexions by further construction Now, 1£ m this proces!I of
demonstration or construction we come to another group of
attr1butcc; of such a kmd that we can start from that also (con•
sisting, that is, of clements whose unity 1c; a matter of direct
mtubon and therefore independent really of the process by
which we have: arrived at 1t from the previous construction),
th1c; c;erond group will form a df'finihon JUSt as much as the
origmal group. Further, 1f the second group of attributes can
bf' sho\\n to rond1t1on the fir'!,t, 1f, that 1s, we can not only start
from the second group as a first prmciple but by a constructive
process derive the first group from 1t, the second group will be
a definition of the same obJect as the first Thu& either of the
two groups can appear as csc;cnce or property altcrnat1vcly.
In this way we arrive at a conclusion not anticipated m the
Anstoteltan theory, or m the later systems derived from it,
namely, that the same tlung may be both self-evident and
demonstrable But we may go further . the conditioned when
properly understood a (that 1s, as the 1e•hole of what 1s necessary
to determme the cond1t1oned and 110 more than is necessary)
always rec-iprocally ncrcss1tatec; the cond1t1oning ; a pomt which
we shall vmd1catc m the d1scuss1on of the plurality of causes. 1
Thus each necessitates the other reciprocally, and m geometry,
ac; in any truly demonstrative science, essence and property are
always rcc1proral All propos1t10ns therefore which arc properly
'~ 152
I• \\'1l~on ah,a\l> m~i-1:ed upon t.h1~. but the truth appeal"'I to be that many
propo~1bons 111 geometrv (for example) are by their nature not convertible
Thf' Greek geometer,,' views a~ to geometncal conversion are explamed
by Sir Thomas Heath, The Thirteen Books of Eudul's Elements, vol 1, p 256
• It 1~ clear that a conversion of the leading type mu,;t be umque, while there
may be manv partial conversion'! of a thf'Orem accordmg to the number of
hvpothe'II!, fwm whu.h 1t ,;tarts' (p 2~i) l
 Princi,ples au Method of Scimce -.1s
proved, namely, those in which the fact is proved from the
sufficient and tw:essary condition, have their converse true.
Thus Euclid, Elements, i. 48, is the converse of i. 47 (Pythagoras'
theorem). A triangle has its three interior angles together equal
to two right angles ; this truth having been properly proved,
that is, made to depend upon triangularity as the sufficient and
necessary condition, is convertible. Every rectilineal figure
whose interior angles are together equal to two right angles is
a triangle ; we might prove this property of an equ1lateral
triangle, but we should not have assig11ed the true condition,
we should have put in too much, and the proposition is therefore
not convertible. It is important to add, however, that though
two elements of reality A and B may reciprocally necessitate
one another, there nevertheless may be a priority of the one over
the other in the order of our apprehension. Suppose we can
start independently with the construction of A and demonstrate
from that the existence of B. Suppose again that in proving
the converse, that B necessitates A, we presuppose the con•
struction of A. Then, though we have a real converse proof,
B has not formed an mdependent startmg-pomt and cannot
form the material proper to a definition. This is what constantly
happens in geometry. There are two forms · First, suppose that
starting from A we derive B, then in the proof of A from B we
may either use the proof of D from A without a reductio ad
absurdum, as 1s done 111 Euclid, Elements, i. 48, or alternatively
we may use the former proof (B derived from A) in a reductio
ad absurdum. Thus Euclid first proves that if a quadrilateral
is inscribed in a circle the opposite angles are together equal to
two right angles. The converse he then proves by reductio
ad absurdum, 1 the form of which is that he assumes a quadri•
lateral whose opposite angles are equal to two right angles and
assumes that 1t is not inscribable in a circle. Then, that being
assumed, a quadrilateral inscribed m a. circle 1s proved not to have
1 It can, observe, also be pro,·cd without such a method a

r• Wilson alway~ maintamed that a direct proof of all theorems 111 possible.
11ua does not appear to be the case , see, for example-, Euchd, El.,,,mls, 11i, 1
(Heath, 1 c , vol. ii, p. 8), Wilson spent great energy over the attempted
conversion an the last part of tlus section. It IS clear from his manu9Cnpt
that he was not satisfied with the result reached in the text ]
2773°2 F
 INFERENCE
the given property of the opposite angles bemg equal to two
right angles, and this 1s contrary to \\< hat was proved in the
original theorem In some cases then we fi.nd that we can make
a proof of A from B independently of the proof of B from A,
and that we can also have another proof m which we make the
demonstration of A from B depend upon and presuppose the
demonstr.1t1on of B from A But now let ui. take the proposition
that the 111ter1or ,tnglcs of a triangle are together equal to two
right angll!s and attempt to prove the converse, which 1s true.
We shall hnd that "e presuppo,;;c that there is such a figure as
a triangle and that we do not start from the mere conception
of a closed figure v. hose mtenor angles arc equal to two right.
.J.ngles For, 1f "c undertake the construction necessary to
vindicate that conception, we fi.nd we presuppose the possib1hty
of three straight Imes enclosing a space Any use of the axiom
of parallels involves this, because the axiom presupposes the
construction of a triangle What 1s the reason of this ? The
relation of the magmtude of the .ingles to the sules of a figure
{1f we do not introduce the cond1tion that the fi.gure 1s closed)
remains mdetermin.itc unless we hke to admit the hm1tat1on
that any one angle cannot exceed two right angles Thus, if
we fuc the magnitude of the sum of the angles of the figure
(estm1atcd on what might be called one ' side ' of the figure) to
be two nght angles, we shall detennme neither the number of
the sides nor v.hether the figure 1s dosed or not What effects
the determination 1s the closing of the figure ; that 1t 1s which
determmes the relations of the angles , that is, the various
determmat1ons of that closmg c1re due to the determination of
the magnitude of the mterior angles. It 1s on this account that
we find ourselves compelled to start from the determmatlon of
the closedness m respect to the number of sides.
§ 200 The science of pure quantity, though so abstract,
depends as much as geometry on a perception or 1magmat1on
of particular md1v1duals To prove the elementary propositions
which are the l>.is1s of this soence we do not thmk merely of
numbers in gcner .J.l but \\ e must represent defimte groups of
numbered tlungs For the theorems m ar1thmet1c we must count
definite md1v1dual instances, and for more general theorems, as
in algebra, ,, here we consider not merely the sums or multiples
 Principles anti M ethotl of Science 4n
of definite numbers but sums and multiples in general, we still
require groups of units which are of definite number, though in
our argument we do not take into account the precise number
which they have. Our results are universal here as in geometry
because we see that they do not depend on the particularity
of the instances. Our attitude is the same as in geometry ; in
the individual instances which we construct empirically we see
the universality of the results. The latter are self-evident and
we never tlunk of appealmg for confirmation to new particular
mstances.
§ 261. In the case of the science of pure quantity the procedure
above described 1s supplemented by another process, that of
symbolic representation and symbolic operation, so that whereas
we start with defimte quantities and operate upon these, when
symbols are used we seem to operate on symbols. This is the
difference between algebra with its kmdred sciences and the
direct procedure of pure geometry, and requires special con·
sideration. Clearly nothmg can be effected through mere symbols
taken by themselves ; they depend always for their use on our
power of constructing or apprehendmg what they symbolize,
and thus we might mclme to underrate their importance, thinking
of them as a convenience rather than a necessity. But in truth
symbolism 1s the necessary rnstrument of the sciences to which
1t belongs and great advances are made by the discovery of new
symbols for new problems The first purpose of symbols m
algebra 1s to d1st111guish various quantities from one another as
merely distmct. quant1llcs without defimtc number. It ts 1mpos•
stble for us to work with the mere abstract notions of quantities,
conceived as merely different from one another, without some
perceptive or imaginative units representing them and so enablmg
us to keep them distinct and in our memory Such perceptive
units are supplied m algebra by the letters of the alphabet. By
their difference m form they represent the difference of the
abstract quantities and, as they are made to differ m a way
easily recognizable, they are not confused with one another.
Thus we remember what quantity had what operations per•
formed upon it. Secondly, a most important function of sym•
bolism 1s to represent conveniently not only the quantities but
the nature of the operation performed. They serve here the
F2
 INFERENCE
important purpose of helping the memory and enabling us to
hold together with certainty a great variety of mental construe·
tions and operations which we should otherwise inevitably forget
or confuse Thus they render possible a train of argument
which would be quite impossible without them. Generally
speaking, any symbols will do, 1f sufficiently distinct in form ;
those being most suitable which, while simple, are sufficiently
distinct to prevent confusion A good deal again depends upon
convenience m wntmg or printing and representation to the
eye. If we must appeal to the touch, as "1th the bhnd, that
may produce a very important difference m the nature of the
symbols. Thirdly, the symbolism it5clf serves 111 a remarkable
manner to save the effort of the several mental constructions
which are necessary, because to each construction correspond
certain definite changes m the symbolism These changes are
easily remembered as valid, without the necessity each time of
going through the actual proce'>ses of thinking which make them
valid. Thus a rhfficult tram of argument is reduced in great
part to a number of merely mechanical chang<'s m our symbols
which we perform by rote, not apprehendmg their validity each
tim<' but rcmembermg that a proof has been given which estab•
lishes it 1 However, mcchamcal as tl11s process may seem, it
depends for its worth on the possibility of our making those
mental constructions,\ luch the chang<'s 111 the c;ymbols represent.
Fm,llly, the proc-c<;o;e<; thu,; symbolucd ,ire real constructions and
never have a merely an.ilyt1c.il character. They represent, even
the simplest of them, '>ynthct1c Judgements so called.
1
Compare what \\a& "aid m § 247 on the u~e of 1ememberecl theorems in
geometry,

[NoTK -The use of the word• construction' 1n tlus chapter requires to be

mochfied 1n tbe light of the author's more mature views See, e i , pp. 522
and 524]
 IV
THE UNIVERSAL CHARACTER OF INFERENCE AND
ITS RELATION TO IMMEDIATE APPREHENSION
§ 262. WE now return, a!> we promised,1 to the question of
the universal character of all inferring. The question is, can in-
ference involve particulars only, without reference to universals;
can it, m other words, be the apprehension of a particular fact
as necessitated by another, without the apprehension of a neces-
sary connexion of umvcrsals ? It has actually been asserted
that all mferencc 1s from part1cul,1rs to particulars, 2 that from
parl1cular instances oI A observed to be B we infer that other
particular mstanccs of A arc B Tlus t'i cc1.stly refuted. The
real proccc;s of thought 1c; tlus From th<• exa11110at1011 of the
part1cul.tr instances of lhc u111\'cr!>al A \\e find reason 111 the
nature of the instc1.nccs lo thmk 1t probable that Aness 1s a con·
dilion of Bncss. "re then mfcr 111 the ca!>c of the second set
of A's that they arc B's, bccau<,e "c judge thal probably all
A is B. The essent 1al tlung then 15 lhc supposed connexion
of the um verb.tis Ancss and Bncss. If "c take mfcrcnce which
1s certam and not merely probable to be the apprehension of
a fact as necessitated by another apprehended fact, then, if a
umvcrsal necessitates another,3 th1!> carncs with 1t the necessita•
tion of a pdrt1cular fact by .i particular. Thus, if Aness neces•
s1lalcs Bncss, a parllcular A as such must be a particular B.
The question 1s whether the converse is true; whether, if one
particular fact necessitates another, this can only be because one
umversal necessitates another. Suppose the particular fact F 1
necessitates the particular fact F2 ; suppose that what in F 1
necessitates F 2 1s A1 and that it 1s particular. A1 cannot itself
be merely particular ; it must be the particularization of a uni•
1 § 227. 1 AF. by Mill, System of Logic, 111. 3, §§ 4, S, and 7.•

• For an explanation of the proper sense of necess1tat1on of one umvenal

by another (not connexion), see § 82.
[• Cf. Locke, Essay, av. 17, t 8.J
480 INFERENCE
versal of Aness. It cannot be the mere particularity of Ai
which conditions another definite particular ; that would give
no ground of any definite determination, for, as mere particular,
it would be bke every other particular. It can only determine
anything as bemg a particular of definite kind or quality, that
is, as a part1culanzat10n of Aness, a universal. If, now, Aness
necessitates Bness, ~ particularization of Aness as Aness neces-
sitates one of Bnesb, mdced tlus 1s the meaning of the expression
that the umven,al Ancsb necessitates Bness Can, however, the
particularity of A1 cond1t1on somethmg other than what the
nature of Aness thus necessitates? Clearly not, for, as before,
mere partlculanty, apart from the universal particularized, con•
tams no ground for anything Tlus may be put otherwise. We
have mamtamcd that the n.iture of the particular 1s entirely
covered by that of the um,·ersal, and there 1s nothing m the
part1<·ular but what the nature of the universal necessitates.
Thus an} t hmg "luch 1., m·rc,,1tcttcd by the nature of the
particular must be due to the nature of the umversal. This
will be made cle.trcr 1f "c consider a possible objection. It
might be said that a part1cul,tr has Ill 1t bOmcthmg more defimte
than bcmg c1,n mc;;tanre of ,t g1vl'n u111versal \Ve can say more
of this circle th,m th.it 1t 1s tlus nrcle or a circle As a particular
1t has a defimtc position and dcfimte relations, a relation, for
mstancc, to another nrclc \\ h1ch 1t may touch or mtersect.
But v.e answer th,tt tlus belong'> really to the umversal itself
and 1s after .111 but a p,1rt1rnlc1nzat1011 of 1t, or pc1.rt of its
part1culanzat1on. It belongs to the nature of circularity, for
mstancc, that p,1rl It ul.ir urde, should be capable of mtersectmg
or toudung 111 gL·m·r.11 ,Hl) rcl,tt1011 R 1, 111 "luch t\\O particulars
A1 and B 1 of t\\ o um, erbals stand to one another, 1s itself
a particulanzat1on of .1 universal l{ and depends on the natures
of the umvcrs,11, Anl'ss and Bncss The kmd of relation m
\\ hich a p,1rt1cul,tr of a ccrtam o;ort stands to another particular
must depend on the 1.md to \\luch that other particulc1r belongs
as well as on the kmd to which the first particular belongs.
This therefore must be true of the relation of necessitation as
obtammg bet\\ccn l\\o particulars. This accounts for the fact
that every actual inference, "e find, depends on a relation
between umvers.ils, and \\ e c.mnot make .i.n inference which is
 Inference and Immediate Apprehension +8t
"not a particular instance of a general principle, which principle
again is a connexion of universals. The general principle is
what we have called the rule of the inference; the inference
drawn is simply a particularization of it. This particularization
is not derived £rem the rule, nor subsumed under 1t as 1ts
premiss. The rule is only apprehended in its particulanzation
and only as particularized. Thus both rule and validity of
a particular mfcrencc arc apprehended together ; the apprehen•
sion of the rule cannot precede the apprehension of the parti•
cular1zat1on. The fallacies which spring from subsuming the
inference under its own rule, taken as a premiss preceding the
particular mferences, have been already expl:uned and refuted. 1
Shortly, they amount to mfemng the mfcrcnce itself.
Fmally, there are cases where we might &ay that we inferred
and yet we cannot feel sure that there 1s anything which would
be fa1rly called a connectmg universal, g1vmg rise to a true
conclusion. These are cases where, observing m an md1v1dual
(say) two features peculiar to 1t, seemg one I infer the other.
If I know no necessary co11nex1on between the two, the inference
for me contains no universal prmc1plc But, for the same reason,
it 1s not a ccrtam inference What has been said is true of all
certain inference and concerns real necessitation.
§ 263.a We now return to ccrtam associat10ns of the word
Inference wluch were mentioned m the mtroductory sections to
this Part. 2
Inferred knowledge 1s considered to be something dependent,
as distmgmshcd from the knowledge from wluch it 1s derived.
It is conceived as md1rcct in contrast agam with that same
knowledge, which 1s itself regarded as d1rcct. Fmally, there are
cases where it 1s accordmgly regarded as mfenor. So we speak
sometimes of a mere inference It 1s true that dependence and
md1rectness do not themselves seem necessarily to convey
infenor1ty in respect of certainly. Yet there are cases where
we thmk that what 1s inferred would be better apprehended
otherwise, and 10 such cases our Judgement is about different
ways of apprehending the same fact and doesn't depend on the
1 §§ 209-IO

[• § 263 to end of the chapter was never revised, nor apparently properly
reconsidered j
 INFERENCE
comparison of a fact not got by inference with one derived from
the same fact by inference.
f 264. The judgement of inferiority arises in the case of
particular facts apprehended m experience, where inference
implies a kmd of knowledge different from that by which the same
facts are directly apprehended. The completest apprehension
of a particular fact m the present 1s perception ; the completest
apprehension of 1t as a past fact 1s our memory of the perception.
When we mfcr to a fact which belongs either to past or present,
though we 1magme at, we yet have not this completer kind of appre•
hens1on J Ierc abo our inference can at most be only probable,
based as 1t must be on particular observations about which we
cannot be certain; so that 1t does not constitute an apprehension
of the fact at all I may mfer, for instance, that another person
has a sensation of a certam kind ; but my mference is not
ccrtam \\htlc lus perceptive knowledge is certain.1 Agam, the
exact character of tus sensation he apprehends while I cannot
pos!>ihly apprehend it Here I am shut off from the only com•
pletc knowlcdg<', \\hu.h 1,; perception.
This uncertamt y, while often passed over in ordinary hfc-
so much so that we treat the highly probable as 1£ 1t were
certam-ic; yet forced upon us m cases of extreme practical
importc1.nc<', notably m \\ hat 1s called c1rrumstanhal evidence
m cnmmal trials Herc we become aware of the fact that our
most careful deductions from what 1s given m experience cannot
be equwalent to the Judg('ment of an eyewitness. In the sciences
again, surprise<; about opm10ns and theories wluch rested on
strong empmcal c, 1dcncc arc not mfrequcnt.
§ 265 In the dcmomtrattye sciences, whether or not inferred
knowledge t!> thought of c\!> mdcpcndent or md1rect, its certainty
1s not depreciated, but rlc<'m<'d equal to that of the premisses
themseh es. Here we arc not trying to apprehend particular
facts as particular facts, we are concerned only with the con•
nCX1on of umversals Yet the difference is not that there 1s no
consideration of universals m the former cases. We have already
tried to show 2 that 1t is only through universals that any
inference can be made, and this truth may be verified in any
instance whatever The distinction is that in these demon•
1 Compare what wa::, ::,d1d in ~ zo9
 Injef'ence and Ifltfltelliau Apprehension 483
strative sciences we are not concerned with the nature of any
particular as that particular ; not therefore with singular judge-
ments but with the connexion of universals, and therefore only
with universal judgements.
Now the apprehension of the universal in this connexion
cannot be perception at all, though 1t takes place in perception
or imagination. And so we have not here the contrast of appre•
hending, or trying to apprehend, the same thing in two ways
which are so different m kmd as perception and a thinking
process which is not perception. And when we examined the
nature of geometrical argument we saw that all the apprehen·
s1ons, both of the premisses and of what they necessitate
inferentially, came m the same way, namely, in a geometrical
construction. Accordmgly what was apprehended as neces•
s1tated by a prem1ss or premisses was as certain as lhe premiss
or prem1sses apprehended That is why in geometry we do not
have the oppos1t1on of inferred knowledge to what 1s not inferred,
as of less certain to cerlam. The question before us therefore
concerns the dependence and mdirectness of mfcrred knowledge
and the sense m which these terms are rightly or wrongly
applied. Further, supposing the same fact to be apprehended
without mference, the question is whether there is any reason
for preferrmg the one kmd of apprehension to the other.
§ 266. We have mamtamed that mfcrence depends upon the
obJective fact that one element of reality, simple or complex,
may necessitate another d1stmgu1shablc from it, and we have
considered inference to be always the apprehens10n of one fact
as necessitated by another, though not assertmg that every such
apprehension 1s accounted an mfercnce m thll ordmary use of
language. The characteristic of the cases which arc not called
inference is lhat in them a relation between two elements of
reality is apprehended immediately from our expcricncc of them,
not by the help of rclatmg them to something else ; that is,
not by the help of additional knowledge consisting m the appre•
hension of their relation to somethmg else. But we shall see
that the converse 1s not true m the ordmary usage ; for not
every such apprehension is accounted non-inferential. Again,
if two distinguishable judgements necessitate a third; if, that is,
we apprehend that the two different facts apprehended in the
 INFERENCE
two different judgements necessitate something else, this appre-
hension is usually accounted an mference. But to this also we
shall see there may be exceptions
Suppose we know one tlung as A and another as B and from
this do not see the relation r 3 bet\\een them. Suppose that,
when we gel the additional knowledge that A stands m a rela-
tion r 1 to M (Ar1M) and B m a relation r1 to M (Br1M), we see
that this of itself necesc;1tates that A stands m the relation r 8 to
B (Ar3 B) Tim \\ould be called an inference on both of the
above grounds, both because of the relation of A and B to the
third thmg and because of the two Judgements necess1tatmg
a third.
Here the fact Ar1M and Br2M, as a complex, necessitates
without anylhmg further (ex hypothesi) the fact Ar8B, and the
apprehension of tlus complex Ar1Mr 2B 1s enough (ex hypothesi)
to give at once the apprehension Ar3B. There 1s no md1rectness
or med1atenec;o; c1ther 111 the rclat 1011 of nece&c;1tat1on or m the
connexion of our knowledge of thl' one with our kno,\ledge of
the other ta pomt ,, h1ch has already bcen m:.untamed).
As to the dependence, ,ve have seen that m the mathematical
sciences the neccc;s1t,1t1on 1s reciprocal Tlus 1s evident enough
Ill a non·lemporal wnsequence, and \\e have &een thc1t m general,
when "e take the true cond1t1on, the existence of the cond1t1oned
necessitates that of the rond1tion Yet there may seem to be
m a true sense a certain dependence of an effect m time on
previous cond1t1ons bec,tuse they exist fin,t But an event and
its temporal cond1t1ons J.re part1culanzat10ns of umversals
(called, some of them at least, m such reference ' laws') whose
nature 1& not somet hmg wluch hJ.ppens m lime and whose con-
11ex1on also 1s somethmg non-temporal The propm,1ttons m
science concern these u111versab and their connexions. For
a scientdic propos1t1on 1s that surh and such a temporal con-
d1bon C has such a,id such an effect E, and \\C have seen that,
1£ C is the true cond1t1on, the umversal E necessitates the
umversal C Thus we can say briefly that 1£ X, as the true
cond1t1on, necessitates Y, whether temporally or not, neither
fact 1s mdependent of the other, for neither fact can exist save
in 1ts relation to the other There 1s, however, a dependence
of one of our apprehensions on .i.nother, 1£ we only know the
 Infe,ence and Immedt.tde Apprehension .Ss
fact Y by knowing first the fact X and seeing that it necGSsitates
Y, and if we are not, at least at the time, able to apprehend Y
in any way which is independent of our apprehension of X.1 •
Whereas there is m one way a directness in our apprehension
of Ar8 B, as necessitated by Ar1Mr1B, there is in another sense
an indirectness, because the relation of A to B 1s not apprehended
by considering A and B alone, but by comparing each of them
with M.
Now, first, 1s the knowledge of the relation of A to B, when
derived from the apprehension of A and B alone, to be regarded
as direct and non-mfcrcnt1al ?
Secondly, 1£ Ar3 B 1s not derived from the conceptions of A and
B alone, but only by apprehending their relation to a third thing,
1s that knowledge always to be called md1rect and inferential ?
Neither principle 1s stnctly observed m practice.
In Euclid, Elements, i 4, the two triangles arc not related to
a third, but the process 1s undoubtedly looked on as inferential
and as a demonstration. But the truth 1s tlus · the magnitudes
compared, m tlus instance the bases of the triangles and the angles
at their bases, are produced by a construction wluch 1s one and
the same for both and can only yield one length for the lines
and one size for the angles. Consequently the magnitude must
be the same m the case of both triangles The relation between
the two tnanglcs 1s therefore got m the most direct manner
possible, got, that ts, from what we know of them, which is their
construction. The so-called demonstration by the application
of one triangle to the other 1s only a way of putting this (and
1s not always applicable m the case of sphencal tnangles).
Here, then, ,,c
have a case v.hcrc the immediate necessitation
of one fact by another 1s accounted an inference ; or, more
accurately, the apprehension that A and B stand in a relation
r 1s seen directly from A and B without reference to a third
somethmg and 1s yet c1.ccounted an mferencc.
This, however, 1s not done consciously, and so we cannot say
that the prmc1ple which we mamtam to be the ground of all
inference-viz. that one fact necessitates another different from
it-would always be held to make an apprehension an inference.
1 As e.g 1n Euclid, 1. 35. See below, p. 487.
L• • Expand th11,' .US. not,.]
 INFERENCE
In ibe other case we might anticipate no exception. But
pouibly what "l\'e know of A and B may necessarily contain
a relation to M, and this may be inseparable from A and B in
our apprehension, i.e neither A nor B may be definable at all
without the relation, or, mother words, A as such may be Ar1M,
and B as such Br1M Now suppose 1t to be a consequence of
these relations to M that A stands in a relation r3 to B. Since
this, though 1t has the verbal form of what 1s called indirect or
inferential knowledge, is after all got immediately from the
knowledge "e have of A and B, can 1t rightly be called indirect
and inferential ~
In pr.l<:lice we find that 1t sometimes 1s and sometimes is not.
Thus in Euclid, Elements, 1. 1, AC and BC are only known as
each constructed cq ual to AB. Y ct we find that their consequent
equality to one another 1s looked on as an mfcrence. On the
other hand 1t 1s assumed "1thout proof m geometry m general 1
that the diagonals of a four-sided ' convex ' rcct1lme<tl figure
cut one another within the figure. Now each d1.tgonal 1s entirely
deternuned by 1t5, rdat1on to the given four-sided figure and can
only be apprehendt'd m that relation. Their mtersection a is
the immediate consc·quc·nre of this rcl.tt10n. Herc the propos1•
hon 1s taken as self-evident and not treated ,lb ,m mference
What is d1shnct1ve of such cases 1s that we do not ..tpprehend
A ab A and also as ArM W c can apprehend A only as ArM,
bccclu,;c objcct1vdy A lb only <h·fmcd a:. ArM And similarly
for B. The pos1t1011 ,rnd length of t hl- d1agonal, "h1rh are
required for its dcfuullon, arc only clctcrmmcd by 1b, bemg
,\ d1~1gonal, 1 t•, hy 1h Jommg t\\O oppo-,1tc angles of the quadn-
1.itcral S11111l,irly the hnc._ AC ,ind BC of Eurhd, i. 1, arc, as
we s.iw, constructed only ,is equal to AB. ln c.ises such ,ls
t hcse, then, the• kno\\ ledge must be c.i.lled dirert, for obviously
in the n.tt ure of the tlung there could not be any knowledge
more direct b
Herc an important d1stinctlon has to be made. Suppose that

' See e g Euch<l, ble111e11t~. 1v, 9, where the diagonals of a square are
assumed to cut one another w1tlu11 the square

t• Modtrn geometry \\Ould require a proof of this.

b • Expand this • MS not, ]
 Inference anti Immediate Apprehension 487
A, as A, is Ar1M, and B, as B, is Br1M: in consequence A may
stand in a relation r 8 to B. Nevertheless we may not apprehend
Ar1M and Br.M as directly necessitating Ar8 B. It may be that
M 1s so related to N that we apprehend in the case of A (i.e. of
Ar1M) the fact Ar,N, and of B (Br1M) the fact Br5N. We then
see that Ar,N and Br6 N directly necessitate Ar8 B.
Now this satisfies the definition of indirect or mediated know-
ledge of the relation between A and B, because it depends on
some additional knowledge not contained m the definitions of
A and B. For, though the relations to M are contained in these
definitions, by hypothesis the relations to N arc not. An instance
of this is the argument in Euclid, Elements, i. 35 1 where the two
parallelograms AC and BF are constructed in relation to a
common base and the two parallels BC and AF ; their relation

to the base and the line parallel to it (the latter complex = M)

belongs to the clcfi111t10n of each of them. But the relation is
not enough to give us the relation of their magnitudes to one
another. We must further apprehend a relation of M to the
figure ABCF ( = N). In this way we apprehend the relations
of the parallelograms to the same figurc,1 and these relations we
see 1mmed1ately necessitate their equality. One might indeed
dispute the apphcat1on of direct and indirect to this reasoning,
because the knowledge wanted 1s precisely about the relation
of the two areas to somct hmg d1stmgu1shable from them. Thus
the 'equality' of the parallelograms means that a umt of area
can be found which would be contained the same number of
times m both. Tlus 1s the same thing as to show that one can
be cut up into pieces which will exactly cover the other.
On the other hand, m some instances, 1f we can have the
mediate apprehension we can also have the immediate. In the
last case there was no choice of alternatives ; in this case we
1 The last steps are: 6FDC =- 6EAB, :. ABCF-FDC = ABCF-EAB.·
 INFERENCE
have both alternatives. Thus, starting with one definition X of
an ellipse, we prove Y of it. Y turns out to be itself a definition
of the same figure We have proved that there may be such
a figure as Y in consequence of the existence of such a figure
aa X, but we can also see that such a figure as Y is possible
immediately.
We may say prov1sionally that the knowledge of Ar8 B is
direct and unmcd1ated when what is contained in the definition
of A and B is seen of itself, 1 e. without any other knowledge,
to necessitate Ar3 B, but 1s mediated when what is contained in
these definitions requires to be supplemented by some other
knowledge. Hence, whereas the relation of two thmgs to one
another by virtue of their relations to a tlurd thmg may be
mediated knowledge, 1t 1s not so 1f the relation to the third is
a necessary part. of the defmit10ns of the first two.
§ 267. The question wluch faces us now is whether, when two
distinguishable facts necessitate a third fact distmgmshablc from
each of them, the apprehension of this third fact should always
be accounted an inference, and, 1f an inference, whether it 1s to
be always cons1dcrcd indirect 111 the ~t>nse "e have given to
indirect.
We have seen that the apprehension 1s not always accounted
an mferenee. Two straight Imes cannot enclose a space. Here
are two facts, the e:,,..1stences of the two straight Imes, neces-
sitating a third, viz ' that they do not enclose a space '. That
is, they either do not meet at all or do not meet twice. The
apprehension of this 1s not taken as an inference, but as self•
evident, and therefore as nol mdirecl.
We might argue that there 1s no apprehension of necessitation
which cannot be reduced to this form
Let A of itself necessitate B and suppose this known 'directly'.
The existences of A and B arc two different facts and the appre-
hensions of them two different apprehensions. The relation
between them, 1t might be said, is a third fact
Here we must beware of a verbal fallacy. In some cases it
1s at once obvious that the reduction would be only verbal.-
For the fact of the necessitation of B by A may not be a fact
distinguishable from what we must define them to be. In other
words, A and B cannot be defined apart from the given relation
 Inference and Immedi.ale Ap,P,ehension 489
between them. A three-sided rectilinear figure must have three
angles. The angles, it 1s true, are different from the straight
lines ; it is not, however, as angles merely that we compare
them but as the angles between the lines forming the figure.
In general, if we apprehend A and only apprehend B as neces•
sitated by A, the necess1tat1on of B by A cannot be apprehended
as a third fact d1stmgu1shable from A and B as we apprehend
them For our apprehension of the fact that A necessitates B
cannot be distinguished from our apprehension of B.
But suppose A and B can be apprehended as realities apart
from the necessitation of either by the other. Can we then say
that the apprehension of the fact that A necessitates B is the
apprehension of a third fact, different from either of them, which
they together necessitate ~ We reply that the necessitation of
B by A is part of the whole nature of A and therefore is not
something different from the fact of A's existence, nor a third
fact different from A and B as facts. Nor can we intelligibly
say that the nature of A and B necessitates the necessitation
of B by A-a mode of expression md1spensable to the view we
are critic1zmg-for the nccess1tation of B by A 1s mcluded in
the nature of A and B and we should thus be describing the
necessitation as necessitating itself Again, 1t 1s only a verbal
fallacy to represent a co-operation of A and B as necessitating
the necessitation of B by A For the nature of A 1s enough.
If we make the verbal change, the above becomes 'the nature
of A necessitates the necessitation o{ B by A'. This, however,
1s only tntelhg1ble if 1t means that the nature of A necessitates
B Thus, 1f A necessitates B, the necess1tat1on of B 1s not
another fact necessitated by A in addition to the B which it
necessitates, and to say that A necessitates the necessitation of
B is either unmtelhg1ble or a way of statmg that A necessitates
B. So also, obviously, 1f we have two really different facts A
and B necess1tatmg m combmat1on another fact C, the neces-
s1tatmg of C cannot truly, but only verbally, Le described as
a fourth fact necessitated by A, B, and C.
, i 268. The foregoing mquiry enables us to understand in what
sense the distinction ordmanly made between axiomatic and
inferred or demonstrated truths can be vindicated.
We understand also that m the ordinary treatment of what
 INFERENCE
is in itself a justifiable d1Stinction, there are certain incon•
sistencies. We were concerned before in showing that in an
important sense the knowledge of axioms is of the same kind
as the knowledge of demonstrations. This knowledge we main•
tained to be an apprehension of a necessary connexion of uni•
versals in the construction of a particular case, the apprehension
being always direct and the axioms requiring a construction as
much as the theorems In this sense, moreover, the knowledge
was seen to be all self-evident
But now, this bemg so, a relation between A and B which
is apprehended from the defimtions of A and B alone, without
comparison with a tlmd thing M, 1s different from a relation which
is only apprehended by the relation of A and B to an element
M, not comprised m their definitions. The first may be called
self-evident and direct m reference to our knowledge of A and
B, and corresponds to what 1s called axiomatic knowledge. The
second is mfcrential and indirect m relation to the same know•
ledge.
Yet the premisses Ar1M and Br2M necessitate the conclusion
Ar3 B without rrference to any other hct to mediate the con•
nexion or necessitation, so that thh, necessitation 1s self-evident,
direct, and axiomatic. Tlus gets an important recognition even
in the ordinary use The necess1tabon of Ar3 B immediately by
Ar 1M and Br2:\I 1<i, as we have seen, a rule of 111ference. Now
some of the axioms st..1te<l as axioms arc nothing but rules of
inference. 'Equals to the same thing arc equal to one another',
for mstance, 1c; only the rule of mference, expressed otherwic;e
in 'C ... B, A = B, then A = C '
Finally, m answer to the question "hether direct or un•
mediated knowledge 1s superior to mediated, we reply that in
the demonstrative sr1cncec: at lea<il there 1s no difference 1n their
certainty. We ought to have both, for neither is complete
without the other, each belongmg to the nature of the facts
before us.
 V
SIMPLE AND COMPLEX IDEAS AND
PROBLEMATIC CONCEPTIONS

§ 269. LANGUAGE develops differences in form to correspond

to differences in thought : and the normal meaning of a given
form is that form of thought for which it was developed. But
in idiomatic use the form gets extended beyond its proper and
primary sense This 1s possible without c-onfus1on when the
context or the suhJect-matter sufficiently indicates the true
meanmg This very transference may be taken advantage of
to deceive . for the form may be used with nothmg to suggest
that its use 1s not the normal one, the hearer so understands it
and the speaker mtends he shall, whereas he himself really
means the sense which 1s not normal-a Jest or a prevarication.
The normal meaning of '1f A 1s B, C 1s D ' and of 'if any
A exists, it is D' implies that 1t 1s uncertain whether A is B, or
whether any A exists This is the true hypothetical statement 1 ;
the normal meaning of the so-called categorical statement, 'all
A is B ', 1s that A certainly exists
The latter, m a familiar idiom, may be extended to cases
where the existence of A is uncertam. Nevertheless, the normal
meaning of 'all A is B' involves the existence of A, and it is
just because of this that the form '1£ a thmg i:I A, it is B' has
been developed for the cases where this existence 1s uncertain.
A similar extension happens to the hypothetical form. It is
not true, however, that '1f there is such a thing as A, it 1s B' ;
'tf A exists, it 1s B' , 'if anything 1s A, it 1s B ', are ever, at
least naturally, used to mean 'all A 1s B' , but rather the
hypothetical form may seem (for a reservation must be made}
to do duty for what may be properly called the conditional
1 On the use of the word • statement • instead of • Judgement •, see Part II,

particularly § 33 The hypothetical statement, when taken to be true and

not an opinion, 1~ a real 3udgement, because it is a decision on evidence, for,
as shown m § 298, 1t 1s an inference
2n~2 G
 INFERENCE

form. •A's being B is :he} condition of C's being D' is a true

conditional and indicates normally that A is really B. So also
the conditional may have the form w1en where} A is . B, C is
. D, an d
1
has in that form the same implication.
The hypothetical iorm appears sometimes to be used in the
same sense. For instance, m giving the description of a scientific
experiment and a d1rect1on how to perform it, or even in stating
the •construction' of a geometrical figure, we use the idiom :
• If an equilateral triangle be described, &c.' In such cases the
possibility of the experiment, or of the construction, 1s not left
uncertain nor supposed to be doubtful. Yct tlus is not a mere
eqmvalcnce to the conditional form The hypothetical character
is not lost. The question left open 1s not whether the construe•
tion can be realized but whether the reader or hearer will realize
it and it 1'! really on tlus account that the '1£' is used. The
condlltonal form, on the other hand, may stand for a true
hypolhet1cal 1 a pomt to \\ luch \\ e shall return 1
§ 270. It is a common doctrine that the categorical can be
reduced to the hypothetical form Hern by the categorical is
meant •all A 1s B ', a!> opposed to the hypothetical '1£ A exists,
it is B '. This is a mere mistake about the usage of speech,
for, whereas 'all A is B' implies normally that A exists, '1£
anything 1'! A, 1t is B' normally implies that the existence of
A is uncertain. Thus tht> proposed reduction is impossible.
Aga.111 a d1st111ct1on 1s usually made between the categorical and
hypothetical statements as follows A categorical statement is
one wlurh attaches its predicate to its subject uncond1tionally:
in' all A 1s B ', B 1s attached uncond1tionally to A. An hypo-
thetical statement attaches its predicate to its sub1ect only
under a cond1t1011 Thus m 'A 1s B, tf A 1s C ', A 1s made subject
and B predicate (uongly) But every statement must attach
its predicate to its subJect definitely and unconditionally · and
the hypothetical statement is only a statement at all because it
does this Just as m 'A is C', C is attached to A without
condition, A itself bemg a sufficient condition of C, so, in •if
A is B, A is C ', the connexion of the apodosis with the protasis
'§§271-2.
 Simf,le a1Ul Com'f>l,x Ideas 493
jg affirmed without condition, the protasis itself being a sufficient
condition of the apodosis. Or we may put it thus : a consequence
is affirmed to fGllow immediately and unconditionally from the
hypothetical connexion of A and B. What this hypothetical
c-onnexion is, is the main object of the present inquiry: it is
enough, however, for our present purpose to describe the hypo•
thet1cal statement as above ; for this at all events is evident,
that in it something 1s attached uncond1tionally to somcthmg
else ; something affirmed uncond1t1onally of something else.
With the above dcfimtion, then, of 'categorical', every state•
ment 1s categorical and ' categorical statement ' == ' statement'.
Thus 'categorical' can only serve to distinguish an expression
which is a statement from one that 1s not, and its proper opposi•
tion is to 'interrogative', 'optativc ', 'imperative', &c., not to
'hypothetlcal '. Consequently, when we maintain that the hypo·
thet1cal cannot be reduced to the categorical, categorical is used
111 the hm1ted sense m which 1t only means the form 'all A is B'
111 opposition to '1f A exists, 1t 1s B '.
If, then, we make categorical expression = 'statement', what
word shall v. c use for the form 'all A is B ' ? According to the
ordinary use of the words ' real' and ' hypot hct1cal' and their
usual opposition to one another, we might call the one statement
'real', and the other 'hypothetical'. Tins, however, might be
taken to imply that the hypothetical statement was not about
reality whereas 1t will be seen hereafter that it i'l 1n a sense
about reality and is mdeed only possible through a conrcption
of what certain realities are
The distmction of statements mto •hypothetical' and 'non•
hypothetic-al' would avmd tluc:; difficulty. The classification,
then, is
Statem<.>nt m general = categorical statement

non-hypothetical statement hypot hettc-al statement.1

If, however, we prefer to kcrp 'categorical' for 'all A is B ',
we must abandon the uc;ual defi111t1on of the word categorical
and make it suit the implication of 'all A 1s B' m its normal use.
§ 271. If we classify these c;tatements as above, then the
conditional statement proper comes under the head of the non•
1
D1sJ1111ctive statements are considered m §§ 305-6.
G2
494 INFERENCE
hypothetical, as we have abo,•e described it. The causal pro•
position, as it may be called, has the form 'because A is B,
C is D '. Now this not only affirms the relation of condition
and consequence, but asserts the realization of the condition.
How then does it differ from the conditional statement as
above described ? As a matter of verbal expression the causal
puts exprec;c;ly what the conditional form implies. This has for
its consequence that the causal form cannot in language be
extended to include the hypothetical wlule the conditional may.
'Wherever (or, whene\·er) A is B, C is D' may be used hypo•
thctically, although, apart from exceptional circumstances, this
is not natural. If a man were really uncertain whether any
A 1s B, and wished to be accurate, he would hardly use the
conditional form If, however, he was inclmcd to think A could
be B, cc;pecially if strongly mclmed, and did not aim at strict
accuracy, he might choose the conditional form. It might seem
easier for the conditional form 'A's being B is a condition of
C's being D' to be used for a. true hypothetical. But even here
it would not be natural to say 1s ', but rather 'A's being B
1

would be a condition of (or, would condition) C's being D'.

This, ho\\e\ er, 1s not the only difference between the causal
and the cond1t1onal forms otherwise there would be no
proper use for the cond1t1011al. 'Because A is B, C 1s D' means
accordmg to the matter or context either 'because A is always
B' (or, 'because all A i<; B '), &c., or 'because A is B in this
case', &c. The conchttonal form 'whenever A 1s B', how•
ever, docs not imply that A is always B, or that all A is B, but,
while normally 1mplymg that thcre a.re surh cases, leaves it an
open question whcthcr A 1s alwayc; B. Thus once more the
conditional cannot as such refer to the present case of A's
being B.
The hypothetical c;tatemcnt is not the only form wluch arises
from our attitude of uncertamty. We may not know whether
B belongs to A, but we may know that either B or C belongs
to it, and this is the disjunctive statement' A is either B or C'.
The disJuncttvc statement, however, does not always arise out
of, or express, uncertainty. We shall return to this hereafter.1
§ 272. A statement, as we saw above,• may have the hypo·
I § 305. I I 269.
 Simple ,1111l Complex Ideas
thetical form verbally while its matter is not hypothetical. But
more than this, a statement may be hypothetical if taken in
one sense, but not if taken in another,
Suppose that A is a universal known to have been realized,
Then the statement 'if anything is A, it is B ', when referring
to tune in general, is not hypothetical, because it means that
all instances of A (in whdtever p.1rt of time) are B. Or again,
though referring to an unknown future or present, the speaker
may really presume that there are or will be further realizations
of A. In such cases the statement is not intended by him to
be hypothetical, though 1f questioned he might admit that it
ought to be put hypothetically.
In other circumstances A may indeed belong to reality as
havmg been realized, yet 1t may be a matter of conscious
uncertainty to the speaker "hcther A is realized for the tune
to which ' 1s' refers m lus statement that '1f anything is A, it
1s B' ; or where he says '1f anything will be A, 1t will be B ', he
may feel doubtful whether A \\-tll ever be realized m the future;
and so far his proposition 1,; hypothetical. For 1t 1s the essential
character of the hypothetical statement that the matter of the
clause introduced by '1(' should be uncertain, 1.e. this is the
normal mi:aning of the hypothetical form
Nevertheless, even here, there 1s a non-hypothetical statement
involved, form the given cc1sc, where A 1s realm the sense that
1t has at least been realized, the hypothetical statement 'if
anythmg is A, 1t 1s B' depends on the implied assertion that the
nature of A necess1t.i.tcs B.
'Whenever A 1s B, C 1s D' has a non-hypothetical form as
compared with •1f A is B, C 1s D' : and mdeed normally implies
that A's being B does belong to reality as being realized at some
time or other. It may be non-hypothetical altogether, or again
not intended as hypothetical ; but it may be hypothetical,
because any further realization of A's being B may be uncertain.
§ 273. There is a '"ay of representmg the statements' if A is B, C
is D', and 'whenever A 1s B, C 1s D' which is very misleading.
This arises with logical reflection. Thus it would sometimes be
said that 'if A is B, C is D' means that the proposition A is B
has for its consequence the propos1t1on or judgement C is D.
This is done in order to analyse the hypothetical proposition
 INFERENCE
iato subject and predicate. Accordingly 1t is said that the
hypothetical judgement or propos1t1on expresses a relation
between two Judgements or propos1t1ons.
Now this way of spe.iking 1s only the product of logical
retlection: no one 111 ordmary hfe would express himself thus.
Indeed 1t would nghtly be answered that we do not mean any•
thing about the relation of propos1t1ons as propositions ; we
mean that A and B and C and D are such m the nature of
tbmgs or 111 rec1hty1 nol 111 our mmds merely, that 1£ m reabty
A were B, 111 re..i.hty C would be D. The mode of expression 1s
not merely art1fmal but fallacious. It does not even follow from
the aboYc that, wlulc the true meamng of the proposition 1s
what has been dc&cnbcd, a consequence of 1t 1s the aforesaid
relation between 'propositions' or 'Judgements'.
A Judgement 1!> not a Judgement unless 1t is Judged. The
same 1s true, mutatis mutandis, of a statement. Thus the words
'A is B ', "lm.h may represent a statement, do not represent
a statement 111 ,my given context, unless th..i.t context represents
acts of thought m \\ h1ch the statement 'A 1!> B' 1s made. But
in •1£ A 1& B, C 1s D ', used to express .c. true hypothct1cal state·
mcnt, we do not state t' Judge') that A 1s B, for we are ex
liypothesi uncerl.irn "hether A 1& 13 ~untl..i.rly ,.,.c may make
C's bemg D dependent on \\hclhcr A 1s B; m that c.isc 'C 1s D'
h, unccrtam and \\ e do not state th..i.l 'C 1s D '. Thus neither
'A 1s B' nor 'C 1s 1)' represents c.1. &tatcmcnt m the hypothetical
form of statement 111 question.
A similar cntic1s111 apphcs to tr ..i.n:,l..i.tmg •v, hencvcr A 1s B,
C is D' mto •whcnc, er the propos1l1011 or Judgement "A 1s B"
is true, the propos1t1011 or Judgement "C 1s D" 1s true', but
here there 1s an add1t1onal fallacy The 1111phcation 1s that
a statement can be some-tunes true and somctuncs not true.
But this 1s a fallacy. If a statement 1!> true at all, 1t 1s always
true. Truth 1s not something \'vlmh can vary m time. It 1s
upon tlus fallacy that Boole's whole calculus of mfcrcnce as
applied to the relations of propositions 1s b.iscd.1
An hypothetical statement then 1s never the statement of
a relation between two statements.•
I § J84.

[• The notion that the propo!i1t10111> arc connected by ,elation came into
common use from Kant Cf note to § 98 J
 Simple and Comf,la I leas 497
§ 274. In the true hypothetical statement there is a kind of
paradox. It relates to something the reality of which is uncer•
tain and yet something definite 1s said about it-definite and
positive, not merely negative, for instance, not merely the
definite statement that 1t 1s uncertrun.
This uncertain something is the hypothetical element in the
hypothetical statement, and we must investigate its precise
nature. With the idea of the hypothetical we do not connect
merely uncertainty. To be uncertain about a. thing is not to
treat it as an hypothesis, nor as the hypothetical part of an hypo•
thetical statement. There 1s something apparently more positive
than mere uncertainty. Whether rightly or not, we often think
of the positive side as consisting in some fiction of ours. (Fiction
here is taken m the sense of conscious fiction, and not of an
unconscious mistake. A 'fictitious idea' sometimes means a
mistaken idea.) Or again 1t may be said that the hypothetical
element is a mere idea of ours, which really comes to the same
thmg.
Without analysing for the present (though 1t greatly needs
analysis) the idea of fiction, we can see that the hypothetical
element m a statement cannot be wholly fictitious and cannot
be merely uncertain. It must contain somcthmg known to be
real and something about which we are certain.
§ 275 If we examine any hypothetical propositions we make
m ordmary life or m science, as d1stmguished from phtlosophic
reflection, we find them to be of the type '1f A 1s B, C 1s D'
where A, B, C, and D are none of them regarded as fictions but
as realities. It would not even be correct to say that A, B, C,
and D are 'concepllons' , but, even 1f we do call them con•
ceptions, they arc supposed to be conceptions of realities and
so not fictitious. If there ts supposed to be any fiction in such
statements, 1t would probably be said that '1f A 1s B' represents
or is something fictitious, or mere idea , or that the hypothesis
that A is B (or the hypothetical case of A's bemg B) 1s a fiction.
But the important thmg 1s that A, B, C, and D are not thought
of as fictions.
It is, however, a current doctrine of popular philosophy that
the mind has a power of constructing fictions which are not
hypotheses in the above sense. It 1s supposed that the mmd
 INFERENCE
cannot makt simple ideas, but that it can arbitranly makt
complex ones, because it can arbitrarily combine simple ones
into complex. This doctrine quite ignores a consideration which
should lead to c1. re-exammation of 1L and, mdeed, to its rejection,
that idea:, m.iy Just ab much be impossible to (..Ombine as to
make. We (.,.,innot, for mstancc, make the complex idea of
a closed rccl1lmc..1r 1igurc wluch hcts three s1del> .ind only two
interior ,.mgle&. If the dm,lrme ,,ere true 1l would imply that
we nught hJ.vc, m a st.itement, au clement AB (.is m '.ill AB
1s <.: ', or, •1f AB 1::, C, 1t 1s lJ '), which w.is a fictitious idea or
conception, beside:, &uch fiction as may be supposed to consist
m au hypothes1& If &o, the form of ,m hypothetical statement
would not necessarily be of the type '1£ A 1s B, C 1s D ', where
not one of the clements A, B, C, .ind D 1s a fiction. We have
therefore to consider" helher there really can be fictitious com·
plcx 1dc..1,; Tiu:, ag,un suggc:,t& the question, whether there can
be ,my 1it.L1t10uo:; &nnplc idem,
But the usual d1:,tmct1on bet" cen simple and complex 1dec1.&
1s not clear; indeed 11 1s o;ometlmcs confused and contradictory.
As a prehmmary therefore we must first discuss the distinction
itself.
§ 276. The locus classicus for the d1slmcl1on of •snnple' and
'complex· 1dec1s 1s Locke's Essay, Bk. 11, especially Chaps. 1i
c1.nd xu.
The ideas peculiar Lo e.it h of the five senses he calls simple,
Tac;tes (sweet, sour), Sounds (loud, Jo"), Sn1clls (sweet, stmkmg),
Colours (white), Touch (h..ird, sott, hot, cold). There are others
which we shall return lo presently, but those named arc such
as we should naturally takt' to be simple ideas, 1£ any are simple.
But this common-sense v1c\\ 1s threatened by a consideration
wluch had not occurred to Locke and 1s fam1har m modern
philosophy. The apprehension of red as a d1stmct colour,
different, say, from white and different from a sound, necessarily
involves the d1stmction of it from other colours and other objects
of the sensitive consciousness. The rccogn1t1on or apprehension
of •,-,d' as a d1stmct and definite colour 1s only possible for
a consciousness which has other objects to compare 1t with.
Thus every such 1dea as red, hot, sweet, loud, necessarily involves
a plurality to which it 1s related. Hence, if the involving thus
 Simple and Complex I tleas
of a plurality made an idea complex, these simple ideas of
Locke's, which correspond to the objects peculiar to the five
senses, would all be complex.
Whether this is a real obJection to the distinction or not (and
we shall see that it is not), it docs hold against Locke's presenta•
t1on of 1t, and he can be sho\\ n to contradict it himself.
All ideas of relation he makes complex. So of course 'equal',
'greater', .i.nd 'less'. This, as his own language shows, is because
.i. relation involves the different obJects bet\\ ecn \\ hich it obtams
and thus mvolves a plurality 'The second (of the acts of the
mind wherein 1t exerts its po" er over its simple ideas) is bringing
two ideas, whether simple or complex, together and setting them
by one another so as to take a view of them at once, without
uniting them mto one ; by which way 1l gets all its ideas of
relations.' 1 It 1s clearly the presence of the t" o ideas thus
compared which makes Locke call the corresponding idea of
relation complex. He expresses hunsclf mdeed rather con•
f usedly : for the first of those acts of the mind is 'combining
several simple ideas mto one compound one and thus all complex
ideas are made' ; 2 whence of course the relations ought not to
be complex ! Yet elsewhere he says d1stmctly that relations arc
complex ideas.3 His account of the formation of the idea of
relation 1s obviously confused At .ill events he thus 1mplicitly
admits that an 1dra 1s complex 1f 1t involves a plurality of ideas
compared. He therefore contradicts himself, for the ideas of
red, &c., as we have seen, also involve a plurality.
The same thing results from his account of modes, and here
he cxpltc1tly contradicts him'>elf Thus he says that modes,
which include both 'simple' and •mixed' modes,' are complex
ideas; they arc 'such complex ideas wluch, however com·
pounded, contam not m them the supposition of subsisting by
themselves but arc considered as dependencies on, or affections
of, substances' 11 ( e g. triangle). That 1s because a mode as an
attribute of substance involves the reference to the substance,
the 'complex' clearly being itself and the substance (accordmg
1 Essay, II. Xll, § 1. • 1b
I 1b II. Xll, § 3 • 1b. II. Xll, §§ 3, 4
• ib. II xu, § 4 , cf xm, §§ I seq , ' 1111xed ' or ' compounded ' modes are
a species of the complex ideas called modes.
.500 INFERENCE
to Locke's way of thinking, that is, for he probably doesn't
think of its relation to its substance as an element in the com•
plex). Here again the unplymg of a plurality makes the idea
complex for Locke and he 1s involved in 1mplmt contradiction.
But the contradiction 1s also explicit because what he has called
simple ideas must have precisely this kind of plurality. Properly
speaking, colour must be a mode and therefore complex ; for
colour is the colour of an extended surface and so not 'subsisting
by itself'. Ile would, however, probably answer that colour is
not this but a mere sensation hke heat At least he would have
to admit, then, that no sensation can subsist of itself, but is an
'affection' of ourselves, 1 c affection of our spiritual substance.
He expressly makes solidity a simple idea ; but it does not
exist by itself . 1t 1s exactly an aff ect1on of a substance.1 There
are other explicit contrad1ct10ns, for he makes •succession' a.
simple idea, but 'succession' 1s obviously an idea. of relation
and therefore.-, accordmg to himself, should be a complex idea.
Again, •po,,cr' 1s a •sunple idea' ; 3 yet obviously it involves
relation and thmgs related, a.s much a.s any of what Locke calls
ideas of relation. It should therefore be an idea of relation and
•~omplcx'. Moreover, po,,cr, as an 'affection of substance'
and not subs1stmg by 1tscl!, should be a simple mode, if a sunple
anything, and therefore a complex idea 'Umty' 3 1s also a simple
idea, but umty 1s 'umty OF', and so m1phes plurahly.
§ 277. The reason of Locke's confusion 1s that he didn't see
the ncccss1ly of dctermmmg clearly whal he meant by s1mphc1ty
and complexity. Ile unconsciously used these distinctions some-
times m one sense, sometimes 111 ,mother.
By a complex 111 general 1s meant something which consists
in a unified mamfold, or m a umty of a. manifold or plurality.
But cverytlung depends on the kmd of plurality. We do not
as a rule mean that a thmg 1s a complex because it necessitates,
involves, or 1s connected "1th a plurality from which 1t can be
distinguished, but because 1t itself consists of a plurality which
1s umfied. Thus from complexity 111 a thing 1s excluded the
necessitation of plurality other than and distinguishable from
itself. (Examples will make this clear.) Locke's mistake is that
1
Essay, II w, § 1, where 1t is a. mode and therofore a. complex 1doa..
1
1b , II vu, §§ 1-8 a 1b., II vu, § 1.
 Simple and Com1'J,x Ideas 501
he takes the latter for complexity, and hence relations and modes
are with him complex.
But sometimes he thinks of complexity in the true sense and
so makes succession, for example, though a relat1on, simple.
Power again, though both relation and mode (it never occurs
to Locke that a relation may be a mode), is also for him simple.
An example will make the issue plain. Consider equality.
Equality is a relation. It is complex, according to Locke,
because it is a relation between two magnitudes. It is certamly
true that in itself 1t 11eccss1tates two magnitudes and therefore
a plurality. But 1f equality itself 1s a complex, 1t must be
a unity of different elements m itself What makes it a complex
or plurality must be the plurality of elements or members which
constitute it. The two magmtudes, then, as making it a com•
plex, must be elements m 1t and it must be their unity, in the
sense of the whole system which they constitute. But the two
magnitudes are not parts or elements of the relation of equality
between them. A solid, we can say, consists as a complex of
surface and volume ; we cannot say c-quahty •consists' of the
two magnitudes which arc equals. The two magnitudes, as
1,tandmg m the relation of equality to one another (or in any
other relation), arc thereby umfied mto a system, i.e. a complex.
That is true enough. TIHS system 'consists' of the two magni-
tudes ; but this umficat1on of the two, i.e. this system which
they constitute, is obviously not their equahty. The plurality
of clements which makes a lhmg complex must be such that
those elements m their umty-of whatever kmd-are the thing
in question.
Thus, if the plurality of the two magmtudcs made their
equality complex, the two magnitudes m their unity must be
(or constitute) the relation of equality between them, wluch is
obviously absurd. The relation of equality, then, unifies them
but is not itself the umty of them, m the proper sense of 'unity
of them', 1.e. a umty of which they are elements, the magnitudes
themselves as unified.
The phrase •umty of' is ambiguous. That wluch gives a
plurality unity and systematizes 1t produces a system which is
a unity of the plurality, but the principle which gives it unity
is not itself the unified system ; it 1s distinguishable from it.
 5oz INFERENCE
The children of the same parents are unified into a group, called
brothers and sisters, by their relation to the same pair of parents.
They form a umty of which they are the members ; but that
unity is not the relation v. luch unified them. They are members
of the family group, not members of their relation to their
parents
A complex 1r:, nol d1stmgu1shable from, and other than, the
totality of 1tr:. members So, ,my plurality from which it can
be d1stmgmshcd cannot be of members of itself, and so cannot
mdke 1t .i. comple:x But equality 1s obviously distinguishable
from and other th.i.n the magnitudes between which 1t obtams,
and so their plurality does not constitute 1t a plurality. Hence
equality is not mcide a complex for the reason which Locke
gives, and this applies, generally, to any other •idea of relation'
whatever.
We have then to ask whether equality in itself as distinct
from the plurc1ltty of the equc:tl magnitudes 1s a unity of elements.
If we take 1t 'm itself' 1t seems clear that "e cannot make it
a complex "luch 1s a umty of members; we could not assign
any members the umty of which constitute!> or makes up
equality. It 1s m this sense, therefore, simple. A sign of this
is that equality" ill not yield a complex attributive or a plurality
of attributives, as every true complex idea should. We say
of the magnitudes of two bodies that they are equal, and
the attributive 1s simple and will not yield several attribu-
tives. But 'ammal • yields both bvmg orgamsm and sentient
organism
It is because succession shows no such complexity within
itself, has no members, that Locke cc:tlls it simple. If consistent,
he would have seen that relation as such 1s not necessarily
complex.
§ 278. But there 1s another kmd of plurality or manifoldness
in 'ideas' which 1t is not natural to call 'complexity'.
Consider the particular red colour of tlus object. This is an
individualization of the species red colour, which again is a
differentiation of the genus colour. The genus is d1stingu1shable
from its differentia, and the species as umversal from the
individual. Hence there is plurality, whether red means the
species red colour in general or thlS red colour. This plurality,
 Simple and Complex I fleas
however, is not one of what could be called elements or I members'
of some colour, while 'colour' and 'red colour' are not elements
of or members of this red colour.
Thus the equality of these two magnitudes is a particular case
of equality, which is a species of relation of magnitude, which
again is a species of the genus relation. Relation is then
differentiated into relation of magnitude, and this again into
relation of equality of magnitude.
Now in equality agam, the elements, relation in general,
relation of magnitude m general, and relation of equality of
magnitude, are not naturally called members ; the genus 'rela•
tion' is not a member of a relation of magmtude, nor are these
(i.e. 'relation' and 'relation of magmtude ') members of the
relation of equality of magnitude.
It is therefore also not natural to call 'the relation of equality
of magmtude' complex For the clements of a complex are
'members'. Members arc normally understood not as different
JO degree of umversahty, but as co-ordinate either as universals,
and of the same degree of umversahty, or as individuals.
Hence, whlle relation and relation of magnitude are not
members of that relation of magnitude which 1s equality, a
relation of magnitude m general and a relation of weight in
general are, as classes, members of the genus relation. The
species, equilateral, scalene, and isosceles triangle, are members
of the genus, but, even so, not of the genus or abstract universal
present JO them, but members of the group of classes which
constitute the whole of the species of the genus.
A complex is conceived as a unity of members which are
co-ordinate in universality, so that each member 1s a part of
the whole in the sense at least that the whole contains something
besides itself.
Now colour is not a member of red colour, for red colour is
not a whole of which colour is a part JO the sense that red
contains something besides colour. On the contrary, red con•
tains nothing except colour. Colour comprises everything in it.
Red is not colour together with something which is not colour.
In equilateral triangular figure, figure is not a member of equi-
lateral triangular figure as a complex. For everything in the
equilateral triangle comes under figure : it is altogether figure,
 INFERENCE
and not figure together with something which is not figure. Nor
is triangle a member of equilateral triangle, for equilateral
triangle is altogether triangle, not tdangle and something else
besides triangle.
The plurality of manifoldness in red colour due to the distinc•
tion of genus and species in 1t 1s not such a manifold or plurality
as can make 1t a complex.
In this way, then, 1 e. in the accurate sense of 'complex',
equilateral three-,;idcd figure is not a complex idea, and, 1£ simple
means that which is not complex, what 1s not a unified system
of members, 1t is certainly simple If, however, we mean by
simple that in the nature of which no plurality is present, and
that which, if related to a plurality, 1s d1stingmshable therefrom,
then red and triangle are not simple. For red m itself has colour
and a d1fferentiat1on 'red', and tnanglc in itself involves the
plurality of figure and three-sided · ' m itself' because red in
itself is colour and can't be distinguished from colour, as some·
thing not colour, any more than three from number, whereas
surface can be d1st111gu1shed from volume as something not
volume. 1
On the other hand, the universal has plurality inherent in
itself, because 1t 1c; its nature to be necessarily chfferenilated into
species. It 1s not some one l111ng present m all the species along
with something else, viz a d1fferentia, constitutmg with 1t the
species. The d1fferent1a 1s entirely mcluded m the universal.
But this plurality 1s nevertheless not romplcx1ty. A universal
may have true comple'C1ty, 1£ 1t comprises elements which are
not differentiation!:. of one another. In the general idea of a 'red
and transparent body', redness and transparcncc are not related
as genus and species. Thus Locke's examples of equality and
succession are not complex but have this manifoldness of genus
and species. ' Equal in magmtudc and heavier in weight ' would
be a proper instance of a complex relation.
§ 279. There is a danger in some cases of putting complexity
into an 'idea' (so called) which relates to a complex, when the

• We ca11 assign to red, as red, no attnbute which i~ not an attnbute of

it as ~olour. We can aSSlgn to colour, m an important sense, no attnbute
which is not true of red (colour) But we can state of volume attnbutes
wluch are not true of its surface
 complexity in question does not belong to the idea. Consider
the number 'three', in Locke's language 'a complex of three
units'.1 •
Suppose we divide a multitude of units into sets of three each.
Each set is a complex of three units. Each unit in the set is
a member of that complex. But this does not cause the number
three to be complex. Threeness does not consist of the three
units : none of the units 1s a member of the number three but
only of the group of three.
Their threeness also does not consist in these three units :
they are not members of the threcness, but of the set of three.
The threeness of the set doesn't consist in 1ts three units ; it
consists in havmg three umts. As such 1t shows no complexity :
the plurality 1t does show 1s that of genus and species and
particular, quantity and number and number three.
The same is true of triangle, viz. three-sided rectilinear figure.
In any triangle the three sides are unified m a complex whole :
and the triangle consists of these hnes m their unity (which is
their relation of contiguity), but the tnangularity of the figure
does not consist in these three Imes, 1t consists in its having the
three sides ; they arc not members of its tnangulanty.
These cases show that the members of the complex (set of
three; the triangle) wluch, m their umty, constitute the complex
(this set of three; this triangle) are not members of their pro-
perty of having this particular complexity (not members of the
property of bemg a set of three, threeness; not members of
the property of tnangular1ty) and therefore their plurality does
not make it a complex.
This three, m fact, ABC, or tlus set of three, consists of A,
B, and C m their whole nature : their threeness is only an
aspect of their nature ; 1t therefore cannot consist of them in
their whole nature. Thus they cannot be members of their
threeness.
Similarly for triangle. This triangle consists in the whole
• Locke, Essay, II. xvi, § 2; cf II xi,, §§ 3 and 5,

[
4
The author wrote, ' Locke's callmg this a mixed mode amounts to
111akmg 1t complex '. See the passages m the foot-note. Locke calls numbers
simple modes, but all modes "" complex ideas. Three 1s called complex
because 1t ,s reac:hed by repeating the simple idea of unity (cf.§ 282),l
 INFERENCE
nature of these three lines, as bounding this particular piece of
plane surface. The fact that they form a three-sided figure is
only an aspect of their nature, not the whole of it. They are
therefore not members of their triangularity. The same is true
for triangularity in general. Triangularity in general does not
consist of three Imes.
§ 280. Tlle foregoing distinction is not that between a complex
and the idea of a complex. Our idea of anything is our idea
of what it 1s m 1tsclf : 1t 1s what we think of it considered as
thought by us ; and what we think of it is what we know or
conJecture it to be. In the first case, then, our idea of a thing
is simply our apprehension of its nature or a part of its nature.
It is not an imaginary picture of the thing-'1dea' in that sense
-for that 1s not what we think the thmg 1s, nor 1s it the thinking
what the thmg 1s. In the second case, our idea of the thing is
our conJecture of what its nature really 1s. This, though helped
by imagination, is neither the mental image nor the imagining
of it.
We are here conrerned with the firc;t sense of idea of a thing
as our apprehension of its nature or of a part of its nature.
Our idea of a complex like the triangle or set of three is then
simply our apprehension of it. Our apprehension of a complex
involves apprehensions of the members of the complex and so
may be, m this sense, called complex apprehension.
In the case of the triangle we have distinguished between the
complex triangle and its attribute of triangularity. So we dis-
tinguish now between the idea of the complex (the triangle or
set of three), and lhe idea of its attribute (tnangulanty or threc-
ness) ; that is, between our apprehension (as subjective) of this
set of three and our apprehension of its threencss.
The distmction then of this three from its threeness is an
obJective distinction in the obJect apprehended, and not a dis•
tinction between something subjective (our apprehension of it)
and somethmg obJective (the object's nature), and is parallel to
a subjective distinction in the corresponding apprehensions.
§ 281. We have excluded from complexity the plurality of
genus and differentia in the species. This is not, however,
because the species necessitates the genus of which it must be
the determination. Elements in a manifold may necessitate
 Simple and Complex Ideas
each other, not in the way of differentiation, and may constitute
a complex. A line has both length and direction (direction is
either constant, straight, or changing, curved or crooked) ; neither
of these is a differentiation of the other, though each necessitates
the other. Thus a line 1s complex.
A rectilinear boundary (i.e. of a finite plane space) must have
a number of straight Imes, not less than three. The attribute
of triangularity involves the attributes of linearity and recti•
linearity; but the threefoldness of the sides (though some
number or other 1s necessitated by rectilinear boundary) is not
a differentiation of boundary or of recttlinearity or of linearity,
and conversely. In this sense the attribute of triangularity,
whether universal or particular, 1s complex. We have seen
already in what sense 1t 1s not complex.
Contrast threeness. Threeness has no such complexity. The
plurality in it, number, odd number, number Three, is one of
differentiation only.
§ 282. If A necessitates B, the system AB is complex, sup•
posing A other than B. (This does not include differentiation ;
for the species necessitates the genus and is not something out•
side the genus, it is the genus m a certain form ) Now it may
be said that 1t is a part of the nature of A to necessitate B.
But this does not make the 'Aness' of A complex. For, by
hypothesis, the quality of A as A-its 'Aness '-is distinguishable
both from B and from the necess1tation of B. B is not a member
of the 'Aness' of A. The necessitation of B 1s not a member of
the 'Aness' of A. For A doesn't consist, in its nature as A, in
the necessitation of B, or rn B and the necessitation of B, for
that leaves out its quality of Aness. We should therefore have
to add this and say that the Aness of A consisted in the neces•
sitation of B together with Aness, which is absurd. Thus the
extension of a line necessitates pomts as its boundaries or
extremities. But the extens10n doesn't consist m the points
or in necessitating the pomts, nor m both, nor in both together
with extension.
The A quality in A, then, is not made complex because it
necessitates B which is not A.
We may apply this to show that an attribute is not complex
because it necessitates a subject : by substituting the attribute
H
508 INFERENCE
for A in the above proof and its subject for B. Surface is aa
attribute of a solid body. It does not consist in being the body, or
in body together with being an attribute of the body, or in the
latter. Any one of these leaves out its quality of being a surface,
And we cannot say surface consists in bemg an attribute of
a body and in bemg a surface.
This may be shown otherwise. Belonging to a subject is
undoubtedly of the nature of an attribute. Now 'belonging to
a subject' in general 1s not made complex by the reference
to a subJect-the subJect 1s not a member of 'belongmg to
a subJect' or of 'being an attribute '-and all that remains for
it is 'being an attribute', that is, an identity.
Thus 'being an attribute' is not complex because the plurality
it involves is not of clements which can be members of it. But
any particular attribute is merely a species of attribute in
general. Thus its mamfoldness is that of differt:ntiation and so
not complex. Surface, e. g., is a species of attribute. a
An attribute agam may necessitate other attributes, but this
does not make 1t complex. Surface necessitates another at•
tribute of a body, viz. volume, but, a!. surface, doesn't con•
sist in volume or m 'necess1tatmg volume', or in both ; and,
as before, not in necessitatmg volume and bemg a surface,
This is sufficient to disprove Locke's view that a mode is a
complex idea.
§ 283.b It 1s important to distinguish two kinds of complex
idea or complex conception For clearness we return to the
statement that the conception or idea of a complex is the appre•
hension of a complex ob.1ect, and that this 1s the only mtelhgible
sense of complex idea or complex conception. Observe that
apprehension does not here necessarily mean perception. What
is apprehended must be present to thought but it 1s not neces•
sarily present to perccpt10n. Ordinary accounts of memory tend
to confuse this. What we remember of a thing is certainly not
the memory image, 1t 1s something in the thmg. Thus what we
[• A lacuna 1s marked here in the manuscnpt
b The language here, as often above, 1s unnatural because the author has
mechanically substituted the words ' statement' and 'state' for ' Judgement •
and ' Judge ' m his old draft The act IS surely not an act of statement wlu.ch
is 'performed' but a companson which IS made, together with an 1dentification.
I have sometimes had to reintroduce the words • Judge ', &c ]
 Simple anti ·Complex Ideas 509
apprehend in remembering is the nature of.the thing which is.
not present to us in perception.
Suppose I remember that certain words were spoken by
a certain man at a certain time. I may imagine myself hearing
them, but the fact which I say I remember, which therefore
I am thinking about and which I apprehend, is not these
imagined sounds. But it is just the apprehension of the fact
which is necessary to memory ; otherwise the mental images or
imaginations would occur as the images m a mere reverie. When
I say of them, in the case of memory, that they are like the
past experience, I am in some sense (other than havmg this
image) apprehending the past experience; I could not otherwise
perform this act of statement.
The complex concept10n, or idea, bring then the apprehension
of a complex object, the true conception, or the conception m
the most proper sense, is, obviously, the apprehcns10n both of
the members or elements of the complex and of the nature
of their unity. Now in the case of a given complex object we
may in perception apprehend its members or elements, and we
may apprehend the fact that they are a umty, but yet not
apprehend the nature of their unity because, as we say, we do
not understand how they arc unified In a solid, for instance,
we do apprehend both volume and surface, and their unity also,
for we see how one necessitates the other. So m ' triangle ' we
understand the unity of the three sides with the three angles,
for we see exactly how the possession of three sides necessitates
the possess1on of three angles. Suppose, however, that we
apprehend m perception the fact that m a crystal substance
a certain chemical quality is combined with a certain geometrical
form. This apprehension does not contam the reason why this
chemical substance crystallizes m this particular shape, and
accordingly we have not an apprehension of the precise nature
of the unity. There 1s therefore an undetermmed and mdefinite
element in our apprehension. Our attitude, then, 1s that we
apprehend the elements of the complex and the fact that they
are unified but are uncertain about the nature of the umty, If
we call this whole attitude 'conception', then, as contrasted
with the full and proper sense of conception, we may call it an
indefinite conception, or a partly determinate, partly. indeter!
H2
510 INFERENCE
minate, conception. Inasmuch as the nature of the unity is
a problem to us, this 'conception' might be called 'problematic',
though that word is more natural for a second case of indeter-
minateness.
In this second case certain elements are apprehended but not
even the fact of their unity ; on the contrary, we are uncertain
whether they can be unified in a given way. Here, then, the
possibility of the elements being unified at all in the given way
is a problem to us. If, then, this 1s to be called ' conception',
it may be called a problematic conception. Both kinds 0£
problematic conception have the verbal form of a definite con-
ception, i. c. conception in what seems the primary and proper
sense of the word. Thus in the statement 'a talking canary
would fetch an enormous price', 'talking canary', which may
be a problematic conception of the second kind, has exactly the
same kind of verbal form as 'tangent hne' or 'right-angled
triangle', which are definite conceptions.
It 1s true that m what we here call a definite conception (like
triangle) there may be said to be an indctermmatc element (and
indeed this is true of all 'definite' conceptions), because we do
not know many thmgs which triangle involves. For instance,
before we apprehend the proof, we do not know the magnitude
of the interior angles. But this indeterminateness is not relevant
to the present distinction. We do not indeed apprehend the
full nature of the three lines unified in a given triangle, or of
the corresponding universal, but we have apprehended some-
thing in them definitely, and we also apprehend definitely the
unity of these clements which we have apprehended. We
know, in short, how these elements are combined or put
together.
This distinction of problematic conceptions from others is
obscured by the verbal form, since this may be identical m
both cases. It 1s of the greatest importance for understanding
the nature of hypothetical thmkmg. The want of it causes
serious confusion in logic, as, e. g., m the theory of Definition.
A type of the ordinary confusions thus caused is Mill's account
of Definition. It is the same confusion which is at the bottom
of the somewhat naive illusion, which affects even distinguished
mathematicians, that it is possible to conceive a space other
 Simple anrl Complex I tleas
than three-dimensional. 1 These problematic conceptions, jf con•
ceptions at all, must be classed as complex conceptions.
§ 284. The foregoing investigation was necessitated by the
common doctrine that simple ideas cannot be made at all by
us, whereas complex can, and that the latter may be fictions,
while the former cannot. Having discussed the meaning of the
distinction of simple and complex as applied to ideas or con·
ceptions, we can now begin discussing whether either kind can
be fictions.
The common doctrine above referred to, whether in popular
form or as held by philosophers such as Locke and Hume,
implies that the simple idea (or conception) is some mental and
subJective existence different from the thing to which it refers
and existing apart from 1t. This 1s treated as if it were somehow
a mental copy of the thing ; and, if we press for an exact
meamng, 1t reduces itself to the mental image m 1magination,
as it is explicitly in Hume. 2 This image, however, we have seen,
is not our 'idea of' or 'conception of' the thmg.
On this account the doctrine involves two thmgs which it
either implicitly or explicitly treats as different : first the
existence m the mmd of the mental copy or counterpart of
the thmg, and secondly its validity in relation to the object
or the reality, 'correspondence' or whatever else it may be
called. So when it is said that the simple ideas are not fictions, 3
or that we do not make them," it is meant that their existence
m our mmds is not our arbitrary creation. On the other hand,
the way m which they are caused in us, the source of what we
passively receive and do not make, guarantees their validity-
this source 1s experience. This is their only source, so that the
same cause makes them both not fictions and 'valid'. Con-
versely, as the idea 1s taken to be a mental something quite
apart from its object, or the real thing, so an idea arbitrarily
made by us would be, for these empirical thmkers, entirely
independent of any relation to an obJect in the way of 'validity'
or 'correspondence', and so with them fiction and invalidity go
I §§ 317-19,
• 'By iiuas I mean the faint images• Hume, T,-eRtise, I. i. 1. In Locke,
the pnmary qualities really come to tlus.
1
e g Locke, Essay, IV. iv 4.
• 1b,, II. 11. 2,
 INFERENCE
together. This indeed is not the strict consequence of their
view ; it would only follow that such a fiction might not be
·'valid', not that it could not be. As to the 'source in experience'
there is much confusion. If pressed, the theory in its usual
form involves the tenet that the object not present to conscious-
ness ca.uses in us the idea of 1t which is present to consciousness,
whence a.rises the familiar difficulty that neither the validity of
the idea could be guaranteed nor even the existence of any
'object' of it. Or, as in one aspect of Hume's philosophy, the
object itself may be made mental, an 'impress10n' existing only
in the mmd hke the idea : the idea, then, becomes its fainter
copy; and certamly such famter copy, the same m kmd as the
impression, is not our conception of the ongmal or idea of the
original (which latter 1s itself, according to this theory, directly
apprehended). Even if we allow this 'idea' (1 e of Hume's) to
be true imagmahon and not a fainter mental slate of the same
kmd as the ob1ect or experience (which cannot possibly be
imagmat10n) the result is the same, this imagmat1on is not 'idea
of' or 'conception of' the thing or object. Smee, however,
experience naturally means direct apprehension of the object,
the above difficulties are often obscured or avoided by ambiguous
language such as 'got in experience', a 'got from experience',
'given m experience', and the hke: where advantage is uncon·
sciously taken of this natural meanmg of experience.
§ 285. The general statement of the doctrine is, then, that
experience is somehow the origin of the existence of the simple
ideas, as these are understood by empiricists (and generally in
the confused way which we have found m Locke), and also gives
them validity. Now whatever may be meant by this, whatever
'being given m experience' means, the facts which the Empiricist
refers to in this doctrme arc such as what we call the experiences
of an actual colour, of an actual sound, of an actual shape (with
some Empmcists). And they mean that everything of which
we have an 'idea' is experienced, in the sense in which colour,
sound, and (with some) shape arc experienced (seen, heard, &c.).
Now in this sense of 'origin in experience' there are 'ideas'
which do not thus originate in experience, because that of which
[& 'This relation (causation) arises entirely from expenence, when we find
that any particular ob1ects are constantly con1omed ' Hume, Enquiry, 1v. 1]
 Simple .and Complex Ideas
they are said to be 'idea' 1s not experienced in this sense-as
not seen, felt, heard, smelt, or tasted ; nor are these ideas
combinations of such as are given in experience in this way.
This being so, whether they are simple in themselves or not does
not matter. Such ideas are, e.g., those of substance and attri•
bute, cause and effect, necessity m general. The development
of the empirical philosophy begun by Hume was due to a growing
consciousness of this fact. The development, however, failed to
explain the origin of such ideas, and only ended m self-contra-
diction. The association of ideas, for instance, the favourite
resource, either does not give any new ideas beside those found
in experience, which form its material (and this is indeed the
true alternative), or 1£ 1t 1s supposed to do so, the admission is
involved that the mmd can make new ideas which it has not
found in experience, and thus the premisses of the theory are
directly contradicted.
In view of the doctrme that 'experiencing' of the kind meant
(and best described by reference to the facts as above) 1s a source
of certain 'ideas' and guarantee of their validity (waiving for
the moment the question whether origm and validity can be
separated m the manner described), we have to ask how such
ideas as Substance, Power, &c., arc related to experience con•
ce1ved in tlus way.
§ 286. But it 1s useful first to consider a set of ideas of which
also 1t may be said that they arc not 'given' m this experience
although they arc not m the same relation to that experience as
substance, cause, necessity, &c. These latter (1. e. substance, &c.)
are what caused the first difficulty and misg1vmg to Empir1c1sts; 1
but there is an abundance of other ideas in ordinary thinking
about which there 1s a similar great difficulty, though the
Empiricists are either not conscious of it at all, or only very
imperfectly. As Plato pointed out, we do not, for example,
observe or perceive 'equality', because we do not perceive two
perfectly equal magnitudes. That 1s, m the language of these
Empmc1sts, we do not 'experience' equality, or perfect straight·
ness, or perfect circularity. Some Emp1ric1sts a have indeed
1 Cf Locke's controversy with Stilhngfleet, Hume on Ca.use.
[• e. g. • Since, then, neither in nature, nor m the human nund, do there
exist any obJects exactly correspondmg to the definitions of geometry.' Mill,
 INFERENCE
so far realized this that they are driven to say, like Mill,1 that
there are no real objects possessed of the attributes studied in
geometry, which must involve for them the conclusion that there
are no ideas whatever corresponding to the words. This con-
clusion is actually admitted by Mill, but he conceals from him-
self the contradtctlon m which he is thus entangled by devices
of language and restores agam to the mind as 'hypothesis' what
he had banished from 1t as 'idea'.
But the usual answer (and the very one to which Mill himself
has recourse, 1mphc1tly, later on in his discussion) would be that
though we do not perceive perfect straight Imes, perfect equals,
&c., we do perceive the unequal and approximation to equality,
the curved and approximation to straightness, and the imperfect
suggests to us the perfect, 1.e. the equal and the straight. But
now suggest10n, understood as these Empiricists would have to
understand 1t (they have never really thought out what they
mean by 1t), could only mean that what was suggested had
been already experienced, and this 1s the real ground of Plato's
doctrine of Remm1scence, into which, or something essentially
like it, the Empme1st would be dnven
The true answer, which 1s 1mposs1ble from the point of view
of such empmc1sm, conducts us necessarily to the recognition
of a kmd of apprehension which 1s not perception and therefore
not 'experience' as understood in the view we are examining.
It 1s granted that we observe or expcnencc unequals and
inequality But, for tlus, we must apprehend the things so
observed to be unequal. Now mequahty 1s unintelhg1ble without
the idea of equality, and therefore m such statements we must
both have and use an '1<lea ', 1 c equality, which by hypothesis
we do not observe or experience, and 111 this sense do not find
m experience, 1 <' m whc1.t we experience. In this sense the idea
ts not given m experience. Now to have the idea of equality
m this active way 1s to know what equality 1s. This again does
not mean to have a mental picture or an imagination of equality
as d1stmgmshcd from apprehension of actual equality, for that
1 System of Lo~ic, 11, 5
System of Logic, 11. S, § 1 'It 1s customary to say that the pomts •.. ex1St
1n our conceptions merely,' 1d. 1b , and the note m the same section, • thlS
unreal exactitude might be called a fiction' (9th ed., vol. 1, p. 261),]
 Simple antl Complex I tleas
is not equality itself. To know what equality is we must apprt•
hend equality itself ; and thus the idea of equality is simply
the apprehension of equality and nothing else but this appre•
hension. Thus the experience or perception of inequality itself
necessitates the apprehension of equality, which the experience
does not present. Further, if the experience guarantees the
validity of the 'idea of inequality' (using the language appro•
priate to the theory we are criticizing), it certainly guarantees
the validity of the idea of equality smce the former necessitates
the latter. And so for other 'ideas', such as 'straightness'.
The relation of such ideas to experience may be put also in
this way : the apprehension of the obJects of these ideas (1. e.
strictly the apprehension of them which constitutes what are
called the 'ideas' of them) 1s a necessary part of the whole
apprehension of which the apprehension called perception or
experience is the other part. Just on this account it is not in
any way necessary to assume that we have the ideas m question
before all experience, or that we could have them. They are no
more actively present to us without perception than are colour
and heat. If we like to call the whole act of apprehension above
referred to experience, then we experience the objects of these
ideas of cqua!Jty, straightness, &c, as much as the unequal,
crooked, &c. But this 1s not only not a fair use of the word
'experience' but tends to obscure an important distmction.
Wh~t, of course, 1s nght m the view of these Empmcists 1s their
feelmg that we have none of these ideas without experience, but
they entirely fad to understand their relation to experience.
§ 287. The ideas above considered, or rather their objects, have
corresponding to them m experience contranes to them more or
less remote, which may sometimes be accounted approximations
to them. But others, also not •given m experience' (such as
cause), have no such counterparts m 'experience', and this is
their distinction from the first set We do not, e g, experience
something approximatmg to causation or contrary to it. If we
keep the meaning of the expenenced which is proper to the
empirical theory before us, Time 1s as much an 'idea' of this
kind as is Cause. The Empiricists arc strangely unaware of this.
For instance, in the empiricist sensational Idealism, Time is
always presupposed and yet, on their account of the data of
 INFERENCE
experience, it could not be among these data and would have
to be accounted for by the Association of Ideas. This difficulty,
however, never occurs to them. Time is not seen, felt, heard,
or experienced in general, in this sense of experience ; it 1s
apprehended along with the apprehension called experience, and
as apprehended 1t is an apprehension of particular time. Here
again, as m the case of the former set of 'ideas' or objects, we
cannot have the one apprehension without the other. Properly,
neither apprehension is more 'vahd' than the other, m the sense
that one 1s more certainly of a real object than the other. But
if 1t 1s held that the apprehension of experience 1s 'valid', or
guarantees the corresponding 'idea' as of a real object, then it
guarantees Time. For that which is necessitated by what is
taken to be real must itself be taken to be real. Now the objects
experienced necessitate Time. We, on the other hand, apprehend
this, and our apprehension 1s necessary as bemg of the necessary.
This proper account is inverted when we represent it as a neces•
sity of thinking as distmgu1shed from the obJcct, and so virtually
as primarily a subjecuve necessity. In recognizing or appre-
hending Time we are simply apprehending side by side with
experience a side or aspect of the object we are experiencing,
which aspect 1s not experienced. Time, then, is not an 'idea'
or 'conception' given in ~perience as simple ideas are supposed
to be, but must be as 'valid' as that which 1s given in experience.
Causality 1mphes Time and it is not mere necessitation of
events, or of their order, but implies the d1stmct1on of events
which happen from objects to which they belong and which do
not happen. The fact again of the existence of such objects,
identical in change, could not be experienced in the sense in
which we are now usmg 'experience' ; and this contention is
borne out by the failure of Locke to get the ideas, e.g. of
substance and power, out of experience as he understands it.
Some Empmc1sts might perhaps hold that we do experience
permanence, or identity in change, when we apprehend the
movement of a body. But, 1£ this is denied to be experience,
the events apprehended m experience at all events necessitate
this belonging to an object, and this, we contend, we do appre-
hend even if we do not experience 1t.
Causality will ultimately be found to mean that the events
 Simple 11,n,J, Complex Ideas
belonging to an object, or to a system of objects, have a definite
order, that is, therefore, a necessary order. There is a reason
why a certain event happens in the particular order it does in
relation to the other events of the system. We apprehend this
necessity as belonging to the order of events : we apprehend it
as much as we do the events, though we do not apprehend it in
the way of experiencing it. And as before, therefore, experience
guarantees it, if it can guarantee anything. Here again the true
relation is inverted and the apprehension, because it is held not
to be experience, is represented as m the first place a necessity
of thinking (and hence arises a quite false problem as to how
reality has to agree with lhis necessity of thought).1 There is
in this case a special temptation to the fallacy, which we shall
come to presently. We repeat that our thought about the object
js necessary because it is the apprehension of something ncces•
sary. Necessity must he in the object apprehended: for appre•
hension as distinguished from that has no content whatever.
Whence necessary thought, as what we are obliged to think
about an object, can only be the apprehension of a necessity
in the object.
But, now, we neither apprehend m experience, nor do we
apprehend otherwise, the particular way in which a particular
event is caused : we apprehend a necessity in general, i. c. that
the event must have some cause or other This is a universal
and there is therefore a special temptation to make it subjective,
because of the common mistaken tendency to regard universals
as merely subjective existences. The example of the con·
trasted set of 'ideas', such as straightness and equality, shows
that there is nothmg strange m the statement that we can
apprehend something in an object of experience, which 'some·
thing' we do not experience, and that we can thus apprehend
a universal. Form the experience of two unequals we apprehend
their relation to equality ; this relation m them we do not, e,:
hypothesi, experience ; we must also apprehend equality which
we do not experience, and this 1s not a particular equality but
a universal.
1 A ialse problem is one 1n which we put as a question what cannot possibly
be a matter of question It therefore always admits a reduclio ad absurdum,
which could easily be given in the present case.
 518 INFERENCE
If it be said that what we apprehend in the object of experience
must be a part of our experience of it, then inevitably it would
result that we experience time, the identity of a subject in its
changes, or of a system of subjects, and that we experience the
fact that events must be 'causally' connected. And if experience
is of the real, these will all be experienced realities. But as the
example of equality shows, we cannot properly call experience
all that we apprehend m an object of which some aspects are
'experienced '.
So far, then, from it being difficult to relate time, causality, &c.,
to objects of experience or to account for their vahd1ty1 they must
either be considered to be as direct obJects of experience as
anythmg else, or to have their validity guaranteed by that of
experience, if experience can guarantee anythmg
The real reason why people make experience the ground of
knowledge is that they think we directly apprehend in experience
real obJects, and this, as already noticed, 1s the natural meaning
of experience. But though this is the only thmg which makes
their appeal to experience mtelhgible, they often either obscure
1t, or are actually not entitled to hold 1t, because of their theories
of perception.
Our conclusion, then, is that if experience is o{ real obJects
and their properties, or if the obJccts of experience are real, it
inevitably follows that both these groups of ideas, equality,
perfect straightness, time, cause, &c , arc properties of real
objects.
It must be remembered that m any case the object cannot
apprehend itself : we must do that. Therefore such appeals to
experience, and indeed any appeal to experience, imply that our
apprehension of the obJcct 1s final and authoritative, in some
respect, whether 1t be of the quality or kind of quality of the
object or at the lowest of the existence of the objects. It follows
that our apprehension mu::,t also be authoritative m respect of
that which we apprehend as necessary to the objects of experience.
§ 288. We return now once more to the question whether
a simple idea can be made. A simple idea may, as we saw,
necessitate a complex idea, so that the one question involves
the other, whether, that is, the corresponding complex idea can
be made. The sense of 'making' which seems most intelligible
 Simple and Comf>lex I leas
is that of composing something out of pre-existing materials
which are not made. But if ideas were so made they would be
complex, and thus we are anyhow, apart from the theory to
which we have been referring, led to consider whether we can
make new simple ideas.
For our purpose, which is to find whether there is any element
in an hypothetical statement which may properly be called a
fiction, or fictit1ous, we might be content with the fact that no
one believes it possible to make a simple idea, and that there is no
record of any one ever havmg made one. At first sight it might
seem enough to say that the absolute origination or making of
an idea out of nothing 1s mconceivable and unintelligible.
Yet it is worth while considermg the question for its own sake,
These are vague words. 'Inconce1vable' may mean that we
don't know how the thing could happen, or that we know it to be
unpossiblc as contrad1ctmg something known. 'Unmtelligible'
may mean that we do not understand it, or that it is contrary
to reason, which again means contradictory to something we
know to be true. There 1s always a danger of passing from one
of these meanings to another. As to' ongmation ', 1t 1s important
to remember that we arc accustomed to attribute origination to
ourselves, and yet the way m which this happens may be said
to be mconce1vable and umntelhg1ble m the first of the above
senses of these words We suppose ourselves to originate the
movements of our limbs ; but we do not understand how we
do it, nor can wc even form a conJecture. If this is immediate
origination, then at all events immediate or1gmation, 'out of
nothing' as the phrase 1s, would be a fact. But what is more
important and more to our purpose 1s that we certamly suppose
ourselves to origmate trains of imagination and trains of thinking,
where, again, there is immediate or1gmat1on or, if not, we at
least do not understand at all how they arc originated. It is
decisive to say that a thmg is mconcc1vable only if this means
we know it to be impossible because we know 1t contradicts
something that we know for certain. Thus m the case of this
example, if we are to pronounce 1t inconceivable, we ought to
know that the nature of the thing itself does not admit of
making or ongination.
If we consider some ideas that we have, we certainly do see
520 INFERENCE
that they are incapable of being made or originated at all in
time ; this is the ordinary way of speaking ; it is more accurate
to say that their objects are incapable of it, Unity, plurality,
the number three, &c., are universals which as such cannot have
any kind of beginning m time. In the nature of the case they
a.re unchangeable. And it might be contended that for any idea
whatever, supposed to be made, there must be a universal which
could not be made. But, for our present purpose, 1t is not
necessary to msist upon this. The question cannot be properly
discussed unless we get rid of the confusion caused by the false
representation of 'idea' as a mental existence which we can have
apart from any reference to or apprehension of an object,
A definite idea or conception being nothing but the apprehension
of a definite object, the question whether an idea can be made
arbitrarily is the question whether the apprehension of an ob3ect
can be made, :'l.nd this, agam, really means whether the object
apprehended can be made by the apprehending mind. Object m
the wider sense mcludes states of ourselves, feelings, emotions,
which can be apprehended, and objects not states of ourselves,
the latter, narrower, sense bemg the usual one. From these,
both of which arc called obJects of experience, arc distinguished
imaginations ; 1magmcd sensations, and imagined obJccts 1 in
the narrower sense. For clearness we may repeat that the
definite idea or conception of an object is not the 'imagination'
of it, but the apprehension of what it is itself
We may apprehend an 1magmat1on, or mental image, as an
imagination, and this apprehension of its nature is the 'idea' ot
conception of it. To make an idea, then, we must make 1ts
object, Suppose we could make an object in the narrower sense,
\t would of course be real and the idea of it, i. c, the apprehension
of it, the apprehension of somethmg real. If we could ongmate
a sensa.tion, the same would be true of its ' idea' or conception.
As to imagination we might here agam be content with the
universal adm1ss1on that we cannot imagine anything simple
except as the imagination of somethmg already experienced.
But suppose we could imagine something quite new, it would
be something existing as our imagination ; the idea of it would
be the apprehension of 1t as something ex:1stmg in our imagina-
• On the dulerence between imagmat10n and experience, cf § 137.
 Simple and Complex Ideas
tion, By 'fiction' is usually meant something mental, not
corresponding to reality. In the first two cases the thing made
(an 'object' or a sensation, e. g.) would be a reality and there•
fore not a fiction and the idea of it (i.e. the apprehension of
it) would be in no sense a fiction. In the third case the imagina-
tion would not have any known reality ( = 'object' or sensation,
c. g.) corresponding to it and, in that sense, would be called
a fiction ; but it would certamly be an cx1stcncc, and the idea
of it, though not the apprehension of 'reality' m the above
sense, would be the apprehension of an existence, Thus even
here the 'idea' or 'conception' would not be either untrue or
without relation to truth (that 1s associated with fiction and
the fictitious), nor would it even be the conception or idea of
something untrue
Now that we see what 'making' an idea would have to mean,
it is clear that 1f a definite simple conception could be made-
and in the nature of the case all simple conceptions must be
definite-it could not be something false or even doubtful, or
a pretence that somethmg was real which was not. In short, it
would have none of the usual associations of the word fiction
and fictitious. This, then, is the important matter ; not whether
it can or cannot be made. Thus m an hypothetical statement,
or indeed m any statement, there can be no simple idea which
can be called false, or doubtful, or a fiction, in the sense of the
pretence that something exists which does not exist.
§ 289. We come now to the complex ideas. It is both popularly
and m some philosophical systems ta.ken for granted that the
mmd can •make' complex ideas, m the sense that it can arbi•
tranly con301n simple ideas, irrespectively of whether the cor.re•
sponding conjunction (to use the language of these systems) has
appeared m experience or not. It would follow m this case that
there could be uncertainty about any correspondence in reality
to these complex ideas. Such ideas, then, might be called
fictitious, as indeed they usually are.
We shall, however, find the truth to be that we have no
arbitrary power at all of thus making complex ideas, and that, in
the only sense in which we can be said to make them, the making
is necessary and the result not only not fictitious but necessarily
real. For 1f we really put together two or more simple con~
 INFERENCE
ceptions into a conception which is truly one, the conceptions
must be put in organic connexion. But such organic connexion
depends on their own nature, and we manifestly could not put
them into a connexion of which their nature did not admit.
Observe that the connexion cannot be merely their being some-
how together ; 1£ we thmk the connexion at all (and we do not
put them into connexion unless we think the connexion} we
must think of it not as connexion in general, but as a particular
form of connexion. When, for instance, we connect the posses-
sion of three angles with the possession of three sides, we see
how the one fact is connected with the other. The connexion
of them will be one which their nature necessitates and therefore
as real as these elements themselves which are connected. This
is illustrated in geometrical demonstration. For in this we really
advance to new complex conceptions, not by experience but by
our own rcasonmg processes ; and, in this sense, viz. as not
merely takmg from experience but advancing to a new con•
nexion of elements by our own thought, we may be said to
make. Moreover, what we make is accepted without hesitation
as true and as real ; at all events, taken to be everyway as true
and real as the clements from which we start.
We have obviously, on the other hand, no power to put
elements together either in a way known to be contrary to their
nature or m a way about which we are uncertam whether it is
possible for them or not. For, in the first case, we should be
thinking the unthinkable ; and, m the second case, we have not
really put them together. In the second case, at the best, we
only know that we see nothing in the conception which prevents
the connexion, but we know no more. Our frame of mind is
not conception or statement, but mterrogatlon or wonder. We
merely wonder whether the connexion 1s possible. The second
case, then, 1s clearly that of conceptions which we have called
problematic: they have only the verbal form of a conception.
The common error 1s to take them for conceptions proper, which
involves considering them as fictions which we have made with-
out reference to reality. But it is now obvious that it 1s a mistake
to suppose them 'made' . they are not made at all and are
therefore not fictions. Instead of their representing a connexion
made between known conceptions, they merely represent the
 Simple and Complex I leas
state of mind in which the connexion is problematic to us ;
a state of mind in which, on the one hand, we know nothing
against the connexion and, on the other, do not see its possibility
(i.e. necessity). We do not here 'make', we are only m doubt
whether the connexion can be made. Particular examples show
this at once. Thus in no case is it really possible for the mind
to produce a merely fictitious idea, whether simple or complex.
In the sense in which a complex idea can be made, the idea
cannot be a fiction but must be as real, or apply as much to
reality, as any other idea.
I 290. The preceding argument is expressed in the language
usual about conceptions and, as such, is of polemical use,•
because it does not involve an appeal to the doctrine about the
true nature of an idea or conception, which we ourselves main•
tain. It requires to be restated in the terms of this doctrine,
for the ordinary language is, in the end, likely to mislead by
suggesting the common confusion about the mental existence of
ideas and conceptions.
A complex conception or idea is nothing but the apprehension
of a complex object in which 'simple' elements A, B, C, &c.,
are united or unified m some particular way. The apprehensions
of A, B, C, &c., severally may be called simple ideas as appre•
hensions of simple elements whether umversal or mdiv1dual.
Obviously, then, the complex 'idea' is not properly speaking
a 'combination' or 'complex' of the simple ideas corresponding
to it, but rather the idea of the complex which the objects of
these simple ideas form; not, that is, a complex of these appre-
hensions of the simple elements but the apprehension of the
complex formed by these simple elements. At all events, to
call the complex idea a complex or combination of the simple
ones is, at best, an obscure and misleading way of saying that
the obJect of the complex idea (or apprehension) is a certain
kmd of complex of the objects of the simple ideas (or appre-
hensions).
Suppose, now, certain simple ideas a, /3, "I are given. These
are respectively apprehensions of objects A, B, C, &c. The
question, then, whether there can be a complex idea combined
[• In truth, I believe, an embodiment of his own ea.rher treatment of the
subJect. Seep 814 ]
2773•2 I
 INFERENCE
of these simple ones is only intelligible as being the question
whether the objects A, B, and C form a complex m a certain
way, i. e. with a certain kind of umty, and whether we can
apprehend this as a definite conception , can, that is, both appre•
bend the elements and the nature of their umty. Can we see
how they form a unity of the given kind?
But whether A, B, and C form a complex of the given kind, or
can form it, depends entirely upon their own definite nature and
not on any act of the apprehending mind. The nature of the
clements themselves determines necessarily what unity they
have or can have, and thus the mind has no power whatever
arbitrarily to effect a combination of them (the necessary pre-
requisite of the complex idea), since that is only the apprehension
of such combmat1on. Thus the mmd has no power whatever
to make a complex idea, and therefore, in any intel11g1ble sense
of the words, no power whatever to make a complex idea out
of simple ones
As to the sense in which we may be said to make a conception
or a demonstration in the exact sc1ences, 1 thi'l agam is an
inaccurate and m1slcadmg use of words In the reasoning pro•
ccss, though as such 1t 1s not experience, wc are simply appre•
bending what the obJccts, with the apprehension of which we
start, necessitate through their own nature and we no more
'make' what they necessitate than we make them. If we do
not know the nature of the obJects sufficiently well to know
whether they can be umfied m a complex of the given kind, the
poss1b1hty of surh umficat1on of them is a problem to us and,
if our attitude 1s to be called idea or concept1on at all, our
conception must be what we have named problematic conception
or quest1on-coneeptton.
1
Compare § 289, p ;22, begmnmg ' ThlS 1s illustrated .
 VI
HYPOTHETICAL STATEMENT AND
HYPOTHETICAL ARGUMENT

§ 291. UNCERTAINTY, then, which is a condition of the hypo•

thetical element, and not fictitiousness, m the sense explained,
attaches to problematic conceptions. Nor can we possibly think
them really valid, although we may incline to the belief that
they are, or, as the phrase 1s, believe they are, or thmk it very
probable that they are. Where the belief is not justified this
appears to introduce a fictitious element mto the 'conception'
(as it is called). But this is a mistaken way of looking at it.
The behcf does not any the more enable us to effect a connexion,
whose poss1bihty for the elements we do not know ; we have
not thereby 'made' the conception and there is no fiction or
error in it. The error hes in the belief.
Thus, if A and B be conceptions m the primary sense, that
is, definite conccption!i, whether simple or complex, and we
mistakenly thmk of their connexion m a unity, 1t 1s the belief
that A and B are compatible m a certain kind of unity which
is erroneou,;. This we may symbolize by AxB, putting ,c (as
a symbol of the unknown) for the connexion, or by A ? B. The
conception AxB 1s not erroneous and therefore made wrongly
and therefore a fiction ; for the conception corresponding to
AxB has not been made at all
The consideration therefore of the problematic conception
reduces itself to the consideration of a judgement (so called)
about probability, or rather of a belief (for we shall have once
more to question the use of the word Judgement for belief at
all). 1 What we have to consider are such statements as 'all
(some) A 1s probably B'; 'it 1s probable that A and B can
be combined in a certain kind of umty' ; 'I believe that A and
B can be combined in a system C' ; or, as will be seen, the true
character of an hypothetical clause such as '1f A is B '.
• §§ 48 et seqq
I 2
sz6 INFERENCE
I 292. The hypothetical statement has for its hypothetical
clause 'if A is B' or ' if A exists'. When A is a simple conception
or a proper complex conception, as opposed to a problematic
conception, the latter form (i.e. if A exists) is not admissible,
because it cannot be a question whether an object of the con•
ception exists and so the hypothetical form is merely verbal.
When A, then, is a conception in the true sense, whether simple or
complex, the only true form of the hypothetical clause is .' if
A is B '. When, on the other hand, AB is a complex conception,
where the connexion of A and B is problematic, 'if AB exists'
is a correct form : yet not only is this equivalent to 'if A is B ',
but this latter form is the simplest and most natural verbal
expression. If AB exists presupposes in fact 'if A is B ', in
cases where AB is not only a problematic conception but is put as a
problematic conception. Thus there are two types of hypothetical
clause involved, '1f A is B' and '1£ AB is C '. The latter,
again, 1s equivalent to ' if A is B and C' , or 'if B is A and C '.
Thus, in every case, the hypothetical clause can be expressed
without any problematic conceptions whatever. Moreover, this
expression is the most natural and adequate because the pro•
blematic conception, not being a 'conception' m the normal
sense but only simulating a conception m its verbal form, ought
to be replaced by the hypothetical form which is adequate to it.
It follows therefore that the hypothetical clause cannot refer
to what is merely fictitious or even to that the existence of
which is merely uncertain. It always relates ultimately to con•
ceptions of whose validity we are certain, and the uncertainty
is not as to whether the obJects of these can exist, but as to
whether they can be combmed, or, in general, stand in a certain
relation to one another 1
§ 293. There may of course be a belief that a certain pro•
blematic conception is realized. Then we may have a statement
of non-hypothetical form 'AB is C ', where the speaker does not
intend to put AB, though really problematic, as problematic,
in accordance with his behef that AB is really existent. From
this point of view, then, the statement would not be resolved
1
We might equally well have taken for the general form of the hypo•
then.cal clause '1f A stands in the relation R to B' We should have arrived
at the same result
 Hypothetical Statement and A,giument s21
into 'if A is B, it is also C '. Yet for strict accuracy and for
the purpose of scientific thinking, the only true form is the
hypothet1cal. Thus for 'AB is C' must be substituted in strict-
ness 'if AB is possible, it is C ', or still more accurately '1£ A is
B, it is also C', or 'if Bis A, it is also C', or 'if A is B (in any
case), that w h1ch 1s both A and B 1s also C'.
Similarly m the hypothetical clause • if AB is C ', when AB is
really problematic, the speaker may believe that AB is realized
or realizable, and so, from his point of view, this hypothetical
does not presuppose '1f A 1s B '. But nevertheless, as before,
1t must 1n accuracy be resolved mto '1£ A is B and C ', or 'if
Bis A and C', or 'if ! :: ~}, and that which is both A and B
is C'.
In both cases, to express the full thought a statement should
be added to express the bchcf that the conjunction of A and
B is possible, or that A can stand m the given relation R to B.
§ 294. a Thus, m every case, the consideration of the meaning
of the hypothetical clause in an hypothetical statement resolves
itself mto a consideration of the meaning of the form 'if A is B ',
where A and B arc both conceptions proper, of whose validity,
therefore, there is no doubt.
We must now try to determine precisely the nature of the
thought, or clement of thought, of which this form (if A is B)
1s the expression.
It 1s necessary to the hypothetical clause '1£ A is B' tha.t
there should be uncertamty whether these two elements of
reality are compatible, but 1t 1s clear that in the hypothetical
clause we are not concerned merely with this ; there is some•
thing more positive, and our question is what does this more
positive something consist in ?
Take 'if A is B, C 1s D' as the general form of hypothetical
statement, the elements A B C D all being real. Let us ask
first: 'What is it about which there 1s a definite statement?'
It may perhaps be said that we 'predicate' of the statement
'A is B' that the statement 'C is D' is a consequence of it.
The point has already been settled, but we may add the following
consideration. The inadequacy of such a representation will be
[• Re•wntten m 1900. It is termed by the author ' Aporematic chscu11&on •
 INFERENCE
·evident if we compare the statement 'because A is B, C is D 1,
-which implies the truth of 'A is B ', or the realization of AB.
-Of this it might equally be said that 1t asserts that the statement
1
C is D' is the consequence of the statement 'A is B' ; and in
this way the same explanation would be given of two very
different forms of very divergent meaning.
But perhaps 1t may be said that this is so, only that the
latter form with 'because', mstead of 'if', 1mphes that the state-
ment 'A is B' is true. This mvolves what seems a paradox,
that we can assert something of a statement quite apart from
its being true or not ; as though a statement could be a state-
ment irrespective of considerations of truth or falsity. We have
already seen that this paradox and mistake come from supposing
that 'A is B' and 'C is D', in the form 'if A 1s B, C is D' 1 &c.,
represent statements ; and we may here add that even the
causal form ' because A 1s B, C is D ' docs not really assert
a.nythmg about the relations of statements as such.
In such a statement we apprehend, or recognize, or suppose
ourselves to apprehend or rccogmzc, a fact , the fact is not our
recognition of 1t. Now m the form 'because A is B, C 1s D' we
obviously assert that the fact that A 1s B (not our statement
that A is B) necessitates another fact, that C is D. Thus the
statement expresses a relation between the facts, A's bemg B,
and C's being D, between two objective realities, and not between
two statements wluch are our subJect1ve acts-our subjective
recognition of the facts. It does not even follow that the state-
ment 'A 1s B ', as our subJectivc act, necessitates that we should
pronounce the statement 'C 1s D' ; for though, if we pronounce
the first statement, we ought to pronounce the other, 1t does
not follow that we ~hall. The connexion may be a matter of
demonstration which we have not got. It 1s the fact recognized
m the statement 'A is B ', which necessitates the fact of C's
bemg D, whether we recognize the latter m a statement 'C is D'
or not.
Again the statement 'if A is B, C is D' can appear as the
ground of 'because A 1s B, C is D '. For if the hypothetical
statement is true, and if in any case 'A 1s B ' is true, it will
follow that 'C is D' will be true. Thus 'if A is B, C 1s D ; but
A is B : therefore because A 1s B, C is D ', Smee, then, a state•
 H')l,Pothetical Stateme#t tind A,gument 5.zg
ment which eJtpresses a relation between facts can be deduced
rightly from the hypothetical, it follows that the hypothetical
must also somehow express a relation between facts. The same
also follows directly from the statement itself. For its truth
obviously depends upon and 1s entirely cond1t1oned by the real
nature of A, B, C, and D, and can only be arrived at by a con•
sideration of what their real nature necessitates.
Nevertheless the statement does not mean that the fact of
A's being B has a certain consequence, for it is implied that we
do not know whether A is B. The 'if' relates somehow to our
subjective point of view, as there 1s no possibility of alternatives
in reality. Can we, then, say that in the hypothetical statement
the two sides are so combined that the nature of the facts, as
we know them, and the relations between them determine rela-
tions between our sub1ect1ve statements or suppositions ? For
instance, does the statement mean that our supposition that
A 1s B involves the supposit10n that C 1s D? In the ordmary
meaning of 'suppose' we may suppose 'A is B ', but this is not
necessary to the hypothetical statement : for m pronouncing
'if A is B, C 1s D' we may not suppose A is B but incline neither
way, or we may actually suppose A is not B, as often when we
are hoping to find a reductio ad absurdztm argument to confirm
our own pos1l1on. Docs it, then, mean 'IF we suppose A 1s B,
we must properly speaking also suppose C 1s D '. Here, how-
ever, we are repeating the lF m what should be an explanation
of it. Again, though the statement would be true, if the hypo·
thct1cal statement corresponding to it were true, it seems clear
that 1t is not the primary meaning of the hypothetical statement
but a deduction from 1t. The primary meaning, if we keep the
IF which the proposed equivalent also keeps, 1s that 'IF A m
reality 1s B, C m reality will be D ', and 1t 1s on account of this
primary meaning that 1£ we also suppose A 1s B, we should also
suppose C 1s D.
The hypothetical, then, does not seem primarily to express
a relation between suppositions. Does it, then, express a rela•
tion between statements ? This question has been answered in
the negative, but we may add here that to obtain any relation
of statements we should have to introduce the IF again, as
before : ' If the statement "A 1s B" 1s pronounced, so must the
530 INFERENCE
statement "C is D" •, and the same criticism holds as in the
former case. Notwithstanding the IF, then, the hypothetical
statement '1f A is B, C Js D' does not primarily, it would seem,
express a relation between statements nor a relation between
suppositions, understandmg 'supposition' as above.
We may here repeat and apply a former consideration: If
we compare again 'because A is B, C is D' with 'if A is B,
C is D ', the first expresses something which is the consequence
of the real natures of A, B, C, and D, and is a statement about
the nature of reality, not about our subjective attitude to
reality ; and the statement is . 'The fact that A is really B
necessitates the fact that C 1s really D.' So also the second
expresses something which is a consequence of the real natures
of A, B, C, and D. It means m its simplest and most natural
expression that A, B, C, and D arc such 111 reality that, if A
were really B, C would really be D. This, too, JS a statement
about the nature of reality.
The IF, as already said, seems to hc1.ve a subjective reference,
but so far we have not found a. subJect1vc reference that would
suit the meaning which we know it to express. We have still
to ascertatn the meaning of the IF and to find what 1s the exact
attitude of mmd, corrcspondmg to the hypothetical clause 'IF
A1s B'.
§ 295, 'If A is B' is obviously not the sc1.me as' if A is supposed,
or assumed, to be B' The idea of supposition or assumption
(we are not ident1fytng these two terms) must therefore be found
in the IF, 1! anywhere. Consider the form '1f A 1s B, C 1s really
D' or 'if A were B, C \\ ould really be D' Substitute for the
hypothetical clause the words 'suppose that A were B' and we
get 'suppose that A were B, C would really be D '. This is
obviously not justified, but only 'supposmg A were B, C would
have to be supposed to be D '. Thus we cannot get the true
form of the hypothetical statement by such a substitution for
IF. The same happens 1£ we substitute for the 'if' clause the
words 'assuming A to be B '.
However, the words 'assume', 'assuming', and 'suppose',
'supposing' are ambiguous. They may mean '1£ we suppose,
&c.\ and this would reintroduce the IV, instead of giving a sub•
st1tute for 1t, and further 'if A 1s B' 1s clearly not the same as
 Hypothetical Statement and A,gument 531
1
jf we suppose A is B '. They may mean 'whenever we assume
A is B ', or 'in the cases in which we assume that A 1s B ', and
then the above criticism will hold.
The hypothetical may be correctly expressed thus : ' If A is
really B, 1t follows that C is really D ', it being uncertain whether
A is really B.1 Thus it expresses that something follows from
something or that something necessitates something. Now in
the statement 'if this triangle 1s rectilinear', &c., substitute for
the hypothetical clause the words : 'from the supposition that
this triangle is rectilinear', &c. This 1s clearly untrue : it would
only be true that 'from the supposition, &c. ', it follows that it
must be supposed that the three angles of the triangle are equal
to two right angles. For a supposition can only necessitate
another supposition.
But it may be contended that the real meaning of the hypo·
thet1ca.l 1s that one assumption necessitates the other thus :
'The assumption or supposition that this triangle is rectdmear
necessitates the assumption that its three mtenor angles are
equal to two right angles. But then we should have to explain
this relation asserted between the two assumptions by adding
'for, if the triangle 1s rectllmear, its three mternal angles are
equal to two right angles, and therefore 1£ we suppose the one
we must suppose the other'. The assumption 'A 1s B' neces-
sitates the assumption 'C 1s D ', for 1f A 1s B, C must be D.
Thus the ongmal form with IF appears to be ultimate. For,
in order to explain the relation between our assumptions, we
have to appeal to a relation between the farts to which they
relate, and for this the IF 1s necessary.
Consider again what 1t 1s that 1s necessitated in the hypo•
thct1cal statement, ' 1f A 1s B, C 1s D '. It 1s not that C 1s D.
For then the statement would imply that C 1s D 1s true, which
it does not , that is only true m the causal form, 'because A is
B, C is D '. What the statement tells us is that C's being D
would be necessitated, IF a certain cond1t1on were realized. What,
then, is involved as absolutely necessary in the hypothetical
statement ? It might. be replied : 'the following of the con-
s e g. ' If tlus triangle 1s really rectilinear 1t follows that 1ts three angles are
really equal to two nght angles ', 1t being uncertam whether the S1dea of the
tnangle are really straight.
 53Z INFERENCE
$equence (C is D) upon the realization of the condition 1s asserted
as necessary and unconditional m the hypothetical statement,
that 1s to say, 'if A 1s B, C 1s D' asserts it to be necessary that
the reabzabon of the condition 'A 1s B' should be accompanied
by the realization of the conditioned 'C 1s D ', The 'realization'
of the condition, however, here means, 'if the condition is
realized', and once more we cannot really explam the manner
in which necessity enters into the statement without using the
word '1f' and presupposmg its signification.
§ 296. To understand better what seems to be the attitude of
thought m relation to the hypothetical clause '1£ A is B ', a clause
which has a meaning only m relation to the hypothetical state•
ment of which 1t 1s a part, let us consider the process of inference,
in which the hypothetical statement has its roam use and m
which also we shall find its origm. We may m this way get
some bght upon the meanmg and use of the words ' supposed '
and ' assumed'. As already pomted out,1 we have only to con-
sider the case in which A and B arc both known to be realizable,
so that so far there 1s no unccrtamty about them. The elements
A and B bemg clements of reality we arc uncertain whether
they can be combined. Let us for s1mpbc1ty take the case where
the combmatlon corresponds lo our merely assertmg B of A ;
where, that is, A and B are combined m the unity of every
subJect which is A. Let us also take again the case where the
further process of thmkmg, after we have realized this uncer-
tainty, 1s directed to removing the uncertainty, or deciding
whether A can or cannot be B. The reasonmg may have one
of two forms, viz either: Rcgre5s through the conditions of
A's bemg B, to fmd a cond1tion known to be realized, from
which therefore A 1s B would follow, as thus: 'C's bemg D 1s
a cond1tion from which A's bemg B would follow, E's bemg F
is agam a condition from which C's bemg D would follow ; sup•
pose it is known that E 1s F, A's bemg B follows'. The argument
would have the form :
If C is D, If Eis F,
A lS B j C is D;
But, Eis F:
Therefore C is D ;
Therefore A is B.
1
§ 294,
 Hypothetical, Statement and Argument 533
Or, the reasoning may progress through the consequences of A's
being B in order to find a consequence known to be impossible
and thus decide that A is not B ; thus :
If A is B, If C JS D,
C is D, E lS Fi
But E is not F :
Therefore C is not D ;
Therefore A 1s not B.
Let us examine the second of these cases We start with the
uncertainty above described about A's and B's combmation,
whether, that is, A can be B. This uncertainty, being mere
absence of knowledge, can give us nothing to start from ; we
can get no new statement and make no advance from that.
The only poss1bihty of such advance hes m our knowledge of
the certain elements A and B, which, of course, as certain
knowledge, is expressed m non-hypothetical statements. And
thus the only advance m reasonmg that is possible is by non·
hypothetical statements mto which A and B enter. Thus,
suppose we know that C necessitates A and B necessitates D.
This enables us to take the next step in the argument, ' If A is
B, C must be D '. Here we observe the ongm of the hypothetical
form ; it depends on and is condit10ned by certainty, and 1t
does not appear until we have some certamty. But how do we
take the step m which the IF appears for the first time ? W c
arc uncertain whether A can be B, we possess the knowledge
that C is A and B 1s D. Perhaps the usual account given would
be that we proceed to suppose that A is B, or, to assume 'A 1s
B ', and then draw from tlus the inference tha.t C 1s D; .:1.nd in
ordinary language A's bcmg B would be said to be the supposed
or assumed case. This, we may note, seems especially natural m
the second instance, inasmuch as 1t turns out that A is not B.
But what is the meaning of 'suppose' ? The meaning 1t most
1:ommonly has is 'to thmk something is probable '. Thus to
'suppose' A is B would mean to think it probable that A is B.
But, as already said, this 'supposition' is not necessary for the
statement 'if A 1s B, C is D ', the val1d1ty of which is no way
affected by the fact of our thinking 'A 1s B' to be probable or
improbable. Moreover, this is so much so that, even if we do
not believe that A is B (though e« hypotkesi we don't know that
534 INFERENCE
A is not B), we can still form the hypothetical statement, and
actually do so, in order to arrive by a progress through con•
sequents at a confirmation of our belief that A is not B.
However, it may be said there is another use of 'suppose' in
which (whatever else it means) it docs not mean to think some•
thing probable, and 1t 1s this meaning which is the only kind
of meaning relevant to the present question.
We may begm the refutation of an assertion 'A is B'. by
saying 'let us suppose A 1s B ', and continue 'then 1t follows
that C 1s D, for C 1s A, and B 1s D '. That does not mean 'let
us think that A 1s probably B ', for then the inference could not
be 'C 1s D', but at most 'C 1s probably D'. Nothing indeed
seems to do short of 'let us thmk that A is really B '. But now
we may suspect here some inaccurate or at least artificial use of
language. The proposal to think that A 1s really B cannot
possibly be carried out. We are, ex hypothesi, uncertain whether
A is B. That is, we do not thmk or conceive the connexion or
umfication of them in one subJcct, and we cannot think A 1s B
without thinking this connexion. Nor will 1t do to say that we
'imagine' that A is B. We may 1mag10e A and B, 1£ this means
to present to ourselves images of md1v1dua.l realizations of A
and B, but of course we cannot imagine the connexion of which
we have, as explained, no conception and therefore no image.
Nor indeed would such 'imagination' be of any use in inference,
unless accompamcd by statement. It may perhaps be said that
we create the fiction that A 1s B. But our criticism of fiction
has shown that there can be no such tlung as a fictitious idea.
It would have to mean, if 1t meant anything, the making of
a mental unage, i.e. the 1magmabon JUSt cr1tic1zed, accompanied
m some way by the statement that A 1s B, a statement that
by hypothesis we do not make.
One way of putting this would be to say that it is a fiction
of ours that A 1s B. But what can this mean ? That we
pretend A is B? We may pretend by a verbal statement to
others ; but we cannot pretend to ourselves. We cannot pretend
to ourselves to think that A is B, inasmuch as we have proved
we do not think A is B and cannot think it with our data.
Language here appears to encourage the same illusion as in the
case of the problematic conception, which has the verbal form
 Hypothetical, Statefflefft and Argument 535
of a conception which is really thought, although in fact it is
not so thought.
But perhaps it may seem that the idea of fiction or pretence
is after all only an inaccurate way of putting something that
is true about this form of reasoning (where we say 'let us suppose
A is B '). It may be said we do not pretend' A is B' or pretend
that we have made the statement' A is B ', but we treat 'A 1s B'
as if it were a statement, viz. by inferring from it. Just so we
may treat a man as we should an honest man, without assuming
that he is, or believing he is, or pretending he 1s; or, a chemist
may give some unknown substance the treatment which would
yield a certain kind of result in a certain kind of substance,
without supposing it is of that kind, in order in fact to test it.
But these analogies are fallacious.
How do we propose to treat as a statement what is not
a statement ? And more difficult still to answer-What is it
that we propose to treat as a statement ? The word-form 'A is
B ', when we say 'treat "A 1s B " as if it were a statement',
seems to give but really does not give the 'it' some definiteness,
for 'it' is not represented by 'A is B ', which properly represents
a statement, whereas, as we have seen, 'A 1s B' does not, in the
context 'if A is B, C is D', represent a statement. The words
simply correspond to my questioning whether A is B, and to
this we cannot apply the treatment applicable to a state-
ment.
Again, let 1t be said that we treat 'A is B' as if 1t were a state•
mcnt by inferring from 1t. This taken literally means that we
infer from something which 1s not a statenlnt. But it is only
from a statement that we can infer, i c the proposed treatment
1s only possible for a statement. Just so there are ways in
which we cannot treat a man believed to be dishonest like a man
who is honest ; we cannot respect the dishonest man. To avoid
this it might perhaps be said that havmg the statement 'B is C'
we mfer as if we also really had the statement A 1s B.1 But
this could only mean that we mferred outright that A is C.
We do not do that, for the uncertainty as to whether A is really
B makes us also uncertain whether A 1s really C.
1
The form 1mpl1ed here is the following • If A is B, A 1s C because B 1s
C ', or the form given in the next section • ' If A is B, A is C, for all B 1B C •,
 INFERENCE
S297. Take the simple case, ' If A is B, A is C, for all B is
C; A is not C, therefore A is not B'. We begin with uncertainty
as to whether A is B, we have the certain non-hypothetical
statement all B is C We consider whether A is B, and we have
C as a mark of B, or a necessary, though not therefore sufficient,
condition that a thing should be B. We look to see whether
A has the mark, find it has not, and decide therefore our question
in the negative, i.e. that A is not B. For the reasoning the
following is sufficient-All B is C, A is not C : therefore A is
not B. Here the statements arc non-hypothetical; and as for
the rest of the process there is no supposition, or assumption,
that A is B, no conception of A as B, no fiction. This helps
us to see that there is no 'supposition' or 'fiction' necessary :
for the process, as above described, which precedes this reasoning
also shows none.
But, nevertheless, this reasoning does not represent the whole
mental process. Being uncertain whether A 1s B, we ask what
B necessitates m order to see whether any of its consequences
are found or not found m A ; because, e. g , the absence of such
consequences would decide that A is not B. The reason there•
fore why we look for consequences of B (a reflection therefore
which must precede such search) 1s that we reflect that IF A is
B, it must have the consequences of B. Thus this hypothetical
statement starts the whole process of search or inquiry, and
therefore instead of bemg superseded by the above statement
of the non-hypothetical argument, 1t truly conditions the appear-
ance of such argument.
§ 298 This brmg! us to an important characteristic of an hypo•
thetlcal statement, viz. that 1t 1s always an inference, and an
mference from what 1s non-hypothetical. It is an inference
which results from our knowledge of reality and its form is due
to the combmation of the knowledge about something with an
uncertainty about 1t. Knowmg all B is C, questioning whether
A is B, and therefore not possessmg the premiss all A is B, we
do not mfer A is C, but only, if A 1s B, A 1s C. Thus and thus
only do we get the hypothetical statement, which 1s essentially
inference, and inference from knowledge of reality."
[" cf § 102, and Croom Robertson's d1Scuss1on of the if of doubt and of
inference 1n Msncl, vol ii (1817), pp 264-6]
 Hypothetical Statement anrl At'gument 537
The temptation perhaps is to suppose that 'if A is B, A is C 1
is merely the equivalent of 'all B is C', but it is not. B is C
doesn't state a condition to be satisfied, but an actual quality
of all the real things which are B. In 'if A is B, &c. ', we are
considering something which we do not know to be a B, and
the possession of C is a condition, for it needs to be satisfied
if A is to be a B1 and it is only m relation to an uncertainty
that there is a condition to be satisfied at all. Or, to put it
otherwise, 'B is C' is a fact realized, and a realized fact is not
a condition to be satisfied. The condition cannot be a fact, it
is hypothetical, something to be satisfied IF &c., but there 1s
no 'If' about the fact. Hence, 1f we av01d the '1f', the con•
dition has to be expressed not by the indicative mood but just
by the conditional mood. That A should be C 1s the condition
of its being B which is the mere equivalent to if A 1s B, A is
C. The latter 1s the full and adequate statement in language
of the condition as condition.
§ 299. The hypothetical expression 'if A 1s B, A is C, because
B 1s C' when general is parallel to the non-hypothetical 'when•
ever A is B, A 1s C, because B is C '. But we cannot reduce
the former to 'every case of AB 1s a cao;e of AC'· for this is
proper to the latter, the normal meaning of which is that there
are cases of AB, whereas the normal meaning of the hypothetical
1s that this 1s uncertain and there may be no such cases. Thus,
1f we take the form 'all cases of A being B are cases of A being
C', we arc obliged for accuracy, in view of the possible addition
in argument of 'but there are no cases of AC', to add this
qualification, 'all cases of A being B, IF the!'e are such cases',
and thus the IF refuses to be eliminated. The two forms given
above are parallel though not identical-and if our explana-
tion 1s successful, it ought to make the parallelism appear
natural.
To get at the meaning of the IF, we must reconsider the whole
actual process of thinking m which 1t anses We ask the question
whether A is B : we find that B 1s C ; we then ask the question
whether A is C. This must be because we think the answer to
this second question may throw hght on that of the first, as
being connected with 1t. Where the second 1s answered in the
affirmative, we cannot decide the answer to the first, since B is
 .538 INFERENCE
not necessarily coextensive with C. 1 But in those cases where
the answer to the second form is negative, so is the answer to
the first, and this is our reason for passing to the second.
Clearly we have now to consider the relation of the answers
to the questions. Whether A is B, and whether A is C, are
both of them problems to us. Now since B is C, the problem
or question whether A is B is a case of the question whether
A is C, since B is a case of C. If we ask the question whether
A is C at all, if the species of C are, e, g , BC, XC, YC, then the
question whether A is C, wluch may be for convenience sym-
bolized by (AC?), has for its various cases (AB?), (AX?),
(AY ?). That is simply, the question (AC~) necessarily takes
the forms (AB?), (AX?), (AV~). The hypothetical expression
'If A is B, A is C, because B is C ', corresponds to the relation
between the problems (AB?) and (AC?) that is, its meaning
may be put thus. (a) 'The question whether A is B, is a case
or form of the question whether A is C, because B is C'. We
see now the reason of the parallelism in form between the two
. ( wherever
expressions 'if A is B, A is C' and ' l h A is B, A is C '.
w enever
For the meaning of the latter is (/J) 'The fact of A's being B
is a case or form of the fact of A's bemg C, because B is C.'
Now (a) and (/J) only differ m the substitution m (a) of the
expression 'the question whether A 1s B ', for the expression
'the fact of A's being B ', which appears in (/J). Observe that
when we use the form (/J), which is eqmvalent to 'wherever
A, &c.', we expressly leave open the possibility that A might
in some cases have another form of C; and m (a.) we leave
open the poss1b1l1ty that another form of the question whether
A is C might be raised. The above relates to the general form
of hypothetical reasoning. The particular, 'if this A is B, &c.',
corresponds to ' the problem whether this A is B is a case of the
problem whether this A is C ', but the corresponding or parallel
non-hypothetical 1s the particular statement 'This A's being
Bis a case of its being C'.
The above seems to represent everything there really is in the
consciousness or mental attitude which corresponds to the hypo•
thetical expression ; and, if the preceding analysis is correct,
1 Viz we can't argue' If A 1s B, A 1S C, A 1S C ·. A 18 B ',
 Hyf,othetical Statement and Argument 539
and also the criticism of the ideas of supposing, assuming,
imagining, &c., there would be no other alternative.
But now we must test its adequacy by examining the inferences
based upon an hypothetical statement, and this we do by con•
sidering the relation between the answer to a question and the
answers to its cases or forms. As to those questions which have
a negative answer, 1t 1s clear that any case or form of them has
a negative answer. 1 Thus, given that a question has a negative
answer, we infer that any case of it has a negative answer. We
analyse therefore the argument 'if A is B, A 1s C ; but A 1s
not C, therefore A is not B' as follows. The question whether
A is B is a case of the question whether A is C. But the question
whether A 1s C, has a negative answer; therefore its case, the
question whether A is B, has a negative answer. Again, obviously
when a case or form of a question has an affirmative answer,
the question itself has this same answer. Hence, '1f A is B,
A is C, but A is B, therefore A is C' is analysed as follows. The
question whether A 1s B 1s a case of the question whether A is C.
But the question whether A is B has an affirmative answer ;
therefore the question whether A 1s C has an affirmative answer.
So far, then, the hypothetical statement corresponds to the
apprehension of a connexion between questions or problems
based upon a knowledge of reality. This knowledge must not
be overlooked. For we do base the connexion of our problems
upon the nature of obJective reality, i. e on the apprehension
that B is C. And so we are not confined to mere subJective
factors when we state the connexion of our problems. For we
do not merely say that the question whether A is B 1s a case of
the question whether A is C, but that, because B 1s C, 1t is a case
of the question whether A 1s C. Here the clause introduced by
'because' definitely bases the connexion of the problems on
reality. Nothing which 1s true m the matter and necessary to
the reasoning seems to be left out.
§ 300. The foregoing analysis may not seem quite satisfactory,
because we may feel that m the expression '1£ A 1s B, A 1s C, &c. ',
we do somethmg more than assert a connexion between our
subjective problems. We may doubt, that 1s, whether the verbal
expression is intended to put just this aspect of our thought ;
• This answer 1s nol an liyt,ollietical statement.
2773•2 K
 INFERENCE
since in fact it seems to do something more. In the clause 'if A
is B' we don't in fact seem to be merely thinking of the question
whether A is B or not and of its implications, but to have our
attention somehow directed rather to the affirmative answer,
though of course we should stop short of saying that we affirmed
it. The difficulty must not be exaggerated. In a certain sense,
when we say the question {AB?) 1s a case of (AC?), because
B is C, we arc alluding to what may be called at least -the
positive side of the interrogation, ev.en if 1t is not the affirma•
tive answer to it. For it is true that we are considering pri•
marily the question (AB?) and not the question (A not B ?).
These two questions are not the same, though they involve one
another.
That they are not the same may be shown as follows. If
{AB?) and (A nol B ?) were the same question, they must admit
of the same forms of answer. There is of course a form of answer
which is the same for both ; we may answer to (AB ?), 'A is B ',
and then this 1s also the answer to (A not B ?). But there is
one form of answer which can't be the same. If the answer to
(AB?) is 'Yes', the answer to (A not B ?) must be No, that is,
the answer must be negative.
Again, 1f (AB?) were the same as (A not B ?), and therefore
{AC?) the same as (A not C ?), in the statement '{AB?) 1s
a case of (AC?), because B is C ', we ought to be able to sub•
stitute (A not C ?) for (AC?) ; whence we have (AB?} is a case
of (A not C ?), whence, 1! the second question 1s denied, so 1s
the first. That 1s to say, 1f A 1s C, A is not B. Which of course
is absurd. For though it is possible that A may be C without
A being B, there 1s nothing m the form of the given hypo•
thetical statement to show that this 1s necessary.
Nor does it even follow that (AB?) bcmg a case of (AC?),
(A not B ?), the eqmvalent of (AB?), 1s a case of (A no~ C ?),
the equivalent of (AC?) For, if it were, the negative of (A not
C ?) necessitates the negative of (A not B ?) ; in other words,
if A is C, A is B, which is obviously untrue. This may be shown
as follows. If we substitute for (AB ?) its equivalent (A not
B ?), and similarly for (AC?) its equivalent (A not C ?), we shall
get (A not B ?) is a case of (A not C ?). Now what is the con•
 Hy-/>othetical StateMMI and Argument s..1
dition that the first question should be a case of the second ?
Clearly that not B should be a case of not C ; i. e. that all not
B is not C ; i. e. that all C is B : whereas all we know is that
all Bis C.
The questions (AB?) and (A not B ?) therefore are not the
same ; but the one is the complement or necessary concomitant
of the other : and the answer to the one 1s not in every form
the same as the answer to the other, but nevertheless necessitates
the answer to the other. They are two different and inseparable
aspects of the interrogative attitude; (AB?) 1s the positive side
of it (avoid calling it 'affirmative') and, as such, it is the definite
side, which alone gives any meaning to the negative.
Thus when we say the question (AB?) is a case of the question
(AC?), it is the positive side of the interrogation which we have
before us, and we affirm an 1mphcation of it, which could not
be affirmed of the negative side. We relate the positive side of
one interrogation to the positive side of another.
§ 301. But, nevertheless, the language, m the form 'if A is B,
A is C, because B is C', seems to go further even than corre-
spondence to the positive side of the question or problem, and
in a sense to indicate a contemplation of the affirmative answer.
It is in this reference to the affirmative answer that we may
look to find the true account of the ver-bal form of the usual
hypothetical statement. It will not, however, do to say that
'if A is B, A 1s C' means 'the answer or statcmeut A is B
necessitates the answer or statement A is C ', because this in its
form, as we know, properly implies the answer really given,
i. e. the statement really made, whereas the answer is uncertain.
Therefore the above will have to become, when fully expressed:
'the answer A is B, if it were given, necessitates, &c.' ; where
we have again introduced the IF which is to be explained.
Let us return to the mental process itself as previously
described. Being uncertain whether A is B or not, we know,
of course, that the decision of our uncertainty must be in a
statement which must have one or other of the two forms A is
B, or A is not B : the latter comprising under it, if we know
{!} sufficiently, A is C, or A is D, &c. ; in other words, if we
K2
 INFERENCE
know the species C and D of not B, which are relevant. We
therefore have these two forms before us, in the sense that
we contemplate them as the alternative forms which the answer
must have, that is to say, we apprehend them. Here it is of
the greatest importance to notice that the forms symbolized by
A is B, A 1s not B, arc forms of statement, and not statements ;
neither could be a statement unless we had stated 1t ; but
before decision this is impossible (Besides-as a redumo ad
absurdum-1f they were statements, they would be stated, and
we should have made two conflicting statements.}
But now we know that B is C. Consequently, we see that
the affirmative of the two forms of answer, namely, A is B,
necessitates also the form, A is C ; so that the complete expres-
sion of the first form of answer 1s ' A 1s B and therefore A is C '.
Observe here that just as the statement that there are two
forms of statement, one of which the answer must take, is not
hypothetical but non-hypothetical, so also the statement that
one of them, viz. 'A is B' necessitates m addition the form of
statement' A is C', is strictly non-hypothetical. It is not that
the form of the answer 'A 1s B ' would in thL case necessitate
the form 'A is C ', but the form of answer 'A is B' does neces-
sitate the form 'A 1s C'. 'Would' would bring back the JF, of
course. Our attitude, then, 1s this : we know that the two
forms of statement, one of which the answer must take, are
(not would be) 'A 1s B' and 'A 1s not B '. We do not know
which form it will take. We also know that of these forms the
first, 'A is B ', necessitates m this case (not would necessitate
but does necessitate} the form A is C, because Bis C Now this
is what the verbal form 'IF A is B, A 1s C' seems intended to
express. It indicates both our uncertainty as to which form of
statement the answer will take and our certainty that the
affirmative form has in it the form A is C as consequence., Here
we notice that there is no fiction in the matter and nothing
asserted of any fiction. Everything is apprehension of reality,
together with no fiction as to what reality is, but merely the
question. This property of the form of one (1 e. the affirmative}
of the two alternative forms of answer is simply an aspect of
the fact that what we apprehend, in the hypothetical attitude,
 Hypothetical, Statement and A,gument 543
is a relation between the two problems A is B and A is C,
involves this necessarily and is involved by it.
§ 302. The apprehension of the forms of answer or forms of
statement which the answer or statement must take is subserved
by the imagination: the umversal in fact is as usual appre•
bended somehow through our apprehension of the particular.
In a certain sense, though in some cases only, 1t is allowable to
say that we imagine the fact corresponding to the affirmative
answer 'A is B ' ; that is, we imagine A being B and then
imagine A being C. But, when and so far as this is possible,
neither is the imagination, although subserving the statement,
the statement, nor is the hypothetical statement about the
imagmation. There are, moreover, cases in which the imagina-
tion of the affirmative answer is impossible, as, for example, in
the enunciation of Euclid, I. vii. We cannot possibly imagine
two triangles with their sides equal, &c. 1 For if we imagine the
sides terminated m one extremity of the base to be equal we
cannot 1magme those terminated in the other extremity also
equal. But the most important consideration is that the descrip-
tion of the imagination 1s inaccurate or misleading. If we
examme the supposed 1magmation, we shall find it is not properly
of A's being B-a kmd of obJect1ve poss1b1hty-at all, and that
it actually contams the problematic element, only appearing to
avoid it through a certam confusion of thmkmg. The imagina-
tion cannot contradict anythmg we apprehend as necessary to
the object to which it refers. We necessarily apprehend the
exterior angle of a triangle as greater than the interior and
opposite angle, and we cannot imagine the contrary. We can
imagine a greater and a lesser angle, but we cannot imagme
them as respectively mtenor and exterior opposite angles of an
imagined triangle. Consequently, if the connexion of A and C
is a problem to us it remains so in the imagination ; we may
imagine A and C but not their connexion. An illusion about
this may be caused by the fact that the imagmatlon in a given
case only refers to a part of the reality of the object. When
1 ' On the same lme and on the same s1de of 1t, there cannot be two

tnangles havmg their sides terminated at one extremity of the hne equal
to one another, and hkewlSe those termmated at ihe other extremity of the
line equal to one aDOther '
544 INFERENCE
we say we imagine a serpent breathing fire, all we imagine is
the colour and exterior appearance of a serpent and the same
for fire. We arc not 1magming the physical properties of fire,
and the biological properties of the serpent which make 'breath-
ing fire• impossible for the serpent.
§ 303. The mferencc '1£ A is B, A is C ; but A is not C,
therefore A is not B' arises m the following way. Being uncer•
tain whether the answer has the form 'A is B' which really,
since B is C, 1s equivalent to the form 'A 1s B, and therefore
A 1s C', or whether the form of the answer 1s 'A is not B ', we
afterwards find that a statement of the form A is C cannot be
made ; in other words, that A is not C. The answer therefore
cannot have the form mcludmg or nccess1tat10g the form 'A is
C'. It cannot therefore have the form 'A is B ', which (as we
have seen) mcludes or necessitates the form 'A is C' ; mother
words, it must be 'A 1s not B '.
This way of putting it explains most simply and clearly the
verbal parallelism between the hypothetical and non-hypothetical
arguments. We have simply to substitute, m fact, in the non•
hypothetical form of argument, the expree.sion 'the form of
statement A is B ', for 'the statement A is B •. Thus, e. g. :
Non-hypothetical argument. Ilypothetical argument.
the statement A is B ncces- the form of statement A is B
sitates the statement A 1s C, necessitates the form of state-
ment A 1s C,
the statement A is C ncccs- the form of statement A is C
sitates the statement A 1s D, necessitates the form of state·
mcnt A 1s D,
therefore the statement A 1s B therefore the form of statement
necessitates the statement A A is B necessitates the form
is D. of statement A is D.
The second (or hypothetical) argument 1s now rewritten, 'if
A is B, A is C ; 1f A 1s C, A is D ; therefore if A 1s B, A is D '.
We may also put the parallelism as follows:
A's being B necessitates C's bemg D. (This may be either
singular or universal in form ) The affirmative form of answer
to (AB~), viz. the form A 1s B, necessitates the form C is D.
In other words, 'if A 1s B, then C must be D '. If we put the
form '1£ A 1s B, C 1s D' m the following shape, 'the affirmative
 Hypothetical Stalemenl and Argument 545
form A is B is a case of the affirmative form C is D', we have
language analogous to that used before about one problem being
a case of another, and the inference may be put similarly. In
the case where we know C is not D, the form C is D is impossible,
and therefore any case of it, as 'A is B ', is impossible. From
this pomt of view then we may say that the hypothetical stJte•
ment asserts something of one of those forms of statement, one
or other of which is the form which the answer to our problem
must take. The distmctlon on which everything turns in the
foregomg 1s that of the form A is B from an actual statement
of that general form.
In the hypothetical statement we assert nothing of a state•
ment, but we do assert something of a form of statement.
§ 304. The 'if' clause m an hypothetical statement may,
according to a familiar 1d1om, refer to something already decided
and thus appear not to embody a question. We say, for example,
'If A had been (or were) B, as we know 1t was (or is) not, A
would have been (or would be) C '. Such statements are, of
course, of frequent occurrence, e g. : 'If he had jumped out of
the carriage a mmute later he would have been killed.' Observe
that all such cases refer to the contradictory of what really
happened, 1. e , as would be said, to somethmg which did not
happen. It docs not follow that in such cases, or indeed in any
cases, the hypothetical statement 1s about a possibility which
is objective. For, as we have seen, there is no obJectlve pos-
s1b1hty of anything which 1s not nccessary. 1 The only other
possibd1ty, 1f 1t can be called one, 1s m our subjective uncertainty.
In such examples as the above the reference 1s to something
impossible and not to somethmg obJcctively possible. Nor is it
true that something is merely affirmed about an imaginary
existence. Somethmg true of reality is affirmed, as 1s easily
seen when we reflect that the example gives the information
that the man ;umped before a certam thing, incompatible with
the hfe of a man in the carriage, happened, e. g. before the
carriage fell over a precipice. The solution of the difficulty may
be got out of the case where there is a real question. Take
a case where we start with a real uncertainty as to whether
A is B but know that B is C; and suppose that we afterwards
l § g0,
find that A is not C. We say first, ' If A is B, A is C because
B is C', i.e. (AB?) 1s a case of (AC.?} since B is C. We add
afterwards, 'But A is not C', 1. e. (AC?) is answered in the
negative. •Therefore A 1s not B 'i i. e. (AB ?) is negatived,
But, when we have got so far, (AB ?} ceases to be a question,
for our uncertainty no longer exists. The proper expression of
our thought now is : '1f A were (or had been) B, A would be
(or would have been} C; but A 1s (or was) not C, therefore
A is (or was) not B.'
Now this 1s equivalent to the question (AB?) was a case of
the question (AC?), and as (AC?) was answered in the negative,
so also (AB ?) was answered m the negative
Thus the hypothetical form we now have before us, '1f A were
(or had been) B, &c. ', expresses a relation between what were
questions (AB ') and (AC '), as questions, and conveys also the
mformat1on as to how they were answered, viz. negatively.
The carriage accident 1s qmtc parallel to the above. The
question whether the man would Jump m time or not was a real
question before he Jumped, and the question whether he would
jump m time was a case of the question whether he would be
killed.
In the case of a.n ordinary reductio ad absurdum proof, it may
be said that the question (AB?) 1s known from the first to be
negatived; but this 1s not for the inquirer or learner, but only
for the teacher when he communicates 1t So in the mouth of
the teacher 'If A 1s B, A is C' means . 'the question for you
(i. e. the person learning) whether A 1s B 1s a case of the question
for you whether A 1s C ' Of course for the teacher himself at
one time (AB?) will have been a problem.
§ 305. The disJunctive statement 'A is either B or C' often,
not always, implies an uncertainty about some given subject
A, and accordmgly 1s akm to the hypothetical. It may mark
a stage m our inqumng attitude ; we may begm with uncertainty
as to whether A 1s B, but advance sufficiently m knowledge to
have the certainty that A is either B or C, or in general, either
B or C or D, &c. Moreover, it then gives rise to hypothetical
reasonmg, A is either B or C, B is X, C is Y, .·. if A is B, it
is X, 1f A 1s C, it is Y. But the statement 'A 1s either B or C'
nevertheless does not always imply an uncertamty, and is at
 Hypothetical Statement tZnd A,gument 541
least not intended to express any. It may express the necessary
differentiation of a universal and contain nothing hypothetical
at all. For example, 'Lines are either straight or curved or
crooked'; 'number is either odd or even'. 'A triangle is (or
triangles are) either right-, obtuse-, or acute-angled.'
In reference, however, to a particular instance of the universal
such a statement has the dis1unctive form which expresses a
limited uncertainty. For of a given triangle we may know it
is a triangle because wc know 1t 1s a figure consisting of straight
lines joining three given points, each to each, without knowing
whether it is right-angled, obtuse-angled, or acute-angled. Thus
we know that the figure formed by joining any point in a cir-
cumference to the extremities of the diameter is a triangle,
before wc know its property.
But, m this case, we have the d1sJunctivc statement 'the
triangle X has either one right angle, or one obtuse angle, or
three acute angles' , which is based on the disjunctive non·
hypothetical statement, ' a triangle must have either one right
angle, or one obtuse angle, or three acute angles '. In fact
the disjunctive statement of uncertamty 'A1 1s either B or C'
(it is uncertain which) 1s always an mference from two state•
mcnts, v1z : {1) a d1s1unct1vc which is not problematic, or
expressive of uncertainty, i e. 'A (or all A) is either B or C ',
and (2) a statement neither hypothetical nor d1sjunct1ve 'A1 is
A'. In other words, m the problematic disjunctive statement
'A1 is either B or C ', A1 has the alternative attributes or pro•
perties B and C only because 1t has the characteristic A which
must differentiate mto B or C.
In this characteristic the problematic disjunctive 1s obviously
like the hypothetical, for this also must be an inference from
a non-hypothetical statement. But the non-problematic dis•
junctive statement is not necessarily an inference from another
statement. A d1s1unct1ve statement, however, of the form 'A is
either not B or not C ', suppose we start with A as subJect, can
only arise as an inference from A 1s not {BC), 1. e. from 'A is
either X or Y, and X is not B, and Y 1s not C ', for obviously
a universal A cannot be differentiated into negative classes
immediately.
§ 306. The disjunctive statement, whether problematic or not,
 INFERENCE
is, like the hypothetical statement, as categorical as the statement
to which the word categorical is usually restricted.
The statement 'A1 is either B or C' (when not the differentia•
tion of a universal} implies the problems (A1B?} (A1C?) ; it
states that the problems are such that one of them has an
affirmative answer. But its meaning cannot be more simply
and directly expressed than by the use of the words 'either' and
'or', 1.c. 'A1 1s either B or C', supplemented by 'but 1t is not
known which'; for to say that one of two problems (A1B?) and
(A1C?} has an affirmative answer is to say either (A1B?} or
(A1C?) has an affirmative answer.
Or put 1t this way. the question 'what 1s A 1 ?' is such that
the answer must take one of the two forms 'A 1 1s B ', or' A 1 is
C' ; where, once more, 'A1 1s B' and 'A1 is C' are not statements
but forms of statement
Thus the problematic disjunctive statement does not affirm
a relation between statements ; it expresses a relation between
two problems, as above described, or (another aspect of the
same thing) 1t states a relation between two forms of state•
ment 'A1 1s B' and 'A 1 1s C ', that is, the statement which
answers the question about A1 must take one or other of these
forms.
§ 307. If we have the statement A1 is either B or C, it follows
that 1f A1 is not B, A1 1s C, and 1f A1 is not C, 1t 1s B. Thus
the problematic d1sJunct1ve statement gives rise to hypothetical
statements.
We may put the problematic disjunctive reasoning as follows:
the question whether A1 is not B 1s identical with the questions
whether A1 is not B and whether A1 1s C (A being either B or
C) ; or the question whether A1 1s not B 1s a case of the question
whether A1 1s C, 1 c. the / orm of the answer 'A1 1s not B'
necessarily includes the form A1 1s C. The ongmal disJunct10n,
as stated, is, however, compatible with A being both B and C.
We ought to add the caveat, if A 1s B, it docs not follow that
A is not C ; and where tlus 1s not the case, 1t ought to be
d1stmctly stated.
§ 308. The problematic disjunctive statement may by com-
bination with assertoric premisses lead to a conclusion which is
not problematic. For instance :
 Hypothetical State,nenl antl A,gument 549
A1 is either B or C j B is D and C is D
••• if A1 is B, A1 is D
if A1 is C, A1 is D
••. A1 is D.
In other words, an attribute which belongs to A1 , if in the group
B and C, mvolves D.
Here it is not the disjunctive statement as such which leads
to a non-disjunctive conclusion. B and C must involve D in
consequence of something common to both, viz. A. We know
that A1 is B or C because A1 is A, and A is either B or C. Thus
we know already that A1 has the condition of D, viz. A, and
thus the argument accurately stated is A1 is A; A is D; .'. A1
is D. The disjunctive statement only arises from an imperfect
insight into A and D, on account of which we have not been
able to attach D directly to A. That 1s to say, the problematic
statement A1 1s either B or C follows from the non-problematic
disjunctive statements A is either B or C (where B and C are
differentiations of the universal A), and A1 1s an A.
But B (= AB) is D, and C (-AC) is D,
And AB and AC make up A (which statement implies that
AB is real and AC is real) .
. ·. all A 1s D, and A1 1s A,.·. A1 is D.
§ 309. Aristotle showed a right instinct in not proposing a
reduction of hypothetical argument to a non-hypothetical form.
He does not condemn it and probably 1t had never occurred to
him to consider its poss1b1bty. The ordmary method is to
substitute for the 1f clauses, in, for example, if A is B, C is D,
if C is D, E is F, therefore if A is B, E h, F, the apparently
non-hypothetical statements 'all cases of A being B are cases
of C being D '. But, as we have abundantly seen, the sub•
stituted and non-hypothetical form properly means that it is
not uncertain whether any A is B. In fact, to represent the
reasonmg m full, we should have to say all cases of A being B,
if there are any, &c. Thus we have not eliminated the' if' and
the supposed reduction is impossible.
Consider again, if all A is B, C is D ; but all A is B, ••. C is
D. Here, if we substituted 'all cases of A's bemg B are cases of
C's being D ', we should imply that all A is B and 'but all A is B 1
would be superfluous. In fact it 1s really unintelligible, in regard
 IN.FERENCE
to such reasoning, to say all cases of A are cases of A being B 1
for ' if A is B ' contradicts the idea that all cases of A are B. If
it be said that reality necessitates A's being B, the first clause
must be A's being B necessitates C's being D, and this implies,
in ordmary usage, that A is B. The same difficulties occur in
the ordinary reduction of the d1sjunct1vc argument. Take, for
example,
either A is B or A is C,
A 1s not C . ·. A is B.
We are asked to represent this by the following reasoning:
Every case of A not C 1s a case of AB.
Every case of A 1s a case of A not C.
. ·. Every case of A is a case of AB.
But here, in the first statement, we have implied the reality of
the cases of A not C, whereas the obJect of our original statement
was to make these cases problematic. Moreover, the precedmg
criticism apphes to the way the second (so-called mmor) premiss
1s stated. Similarly we cannot reduce the argument A is either
DB or DC, . ·. A 1s D, to a non-dtsjunctivc form.
§ 310. We may now mqmre how hypothetical reasoning is best
represented so as to bring out its true form. It 1s uncertam
whether A is B , let C be A and B be D. Then if A ts B, inas-
much as C JS A and B is D, rt follows that C is D. Further,
we know that Jf C is D, E JS F. Therefore 1f A JS B, we know
that E 1s F.
The accurate way of stating this appears to be this. An
affirmative answer to the question whether A JS B necessitates an
affirmative answer to (CD?), an affirmative answer to the latter
necessitates an affirmative answer to (EF?), therefore an affirma•
t1ve answer to (AB~) necessitates an affirmative answer to (EF?).
But here we are m d1fficult1es again, because we can say that
the quest10n (AB?) 1s the same as the question (is CA-BD?)
but not that it is the same as (CD?). That 1s, we have to add
the reason for the dependence of C's bemg D on A's being B,
in order to express the hypothetical statement in terms of the
identity of one question with another. If, then, we only state
that C 1s D, 1£ A is B, without the connecting reason, we cannot
identify something relating solely to C and D with the question
whether A can be B.
 Hypothetical State1Kefd and A,gument ssr
What is adequate, then, is to say that the form or answer
to the question (AB?) is identical with the question whether
A is B and C is D, owing to the real implication of the real
elements A, B, C, and D. This exactly covers the hypo•
thetical statement. For, if C is not D, the question whether
A is B and C is D has a negative answer and therefore the
question (AB?) has a negative answer, that is, 'if C is not D,
A is not B '. But if the answer to the question (CD?) is affirma•
tive, it does not follow that that to (whether A is B and C is
D?) is affirmative, and therefore not that the answer to (AB?)
is affirmative. This corresponds to the usual expression that if
C is D, it does not follow that A is B. Thus the hypothetical
argument, if A is B, C is D, and if C is D, E is F, &c., may be
correctly put in the form of the corresponding questions. We
may, however, ass1mrlate this to the usual form more closely if
for 'the question whether A 1s B 1s the same as the question
whether A is B and C is D' we substitute 'the question whether
A is B is the same as the question whether C is D and something
else is true', This exactly covers the usual hypothetical state•
ments, and the conclusion will be 'the question whether A is B
is true is identical with the question whether E is F and some•
thing else '. The argument, then, is a relative argument, that is,
the relation of two terms to one another by means of their
relation to a third. We can put it strictly as follows:
The problem whether A is B includes the problem whether
C is D.
Th!! problem whether C is D includes the problem whether
Eis F •
•·. The problem whether A is B includes the problem whether
Eis F.
And this form of argument cannot he reduced to a syllogism,
§ 311. We thus reach the conclusion that the assertion many
hypothetical statement is that one problem includes another or
that one question is identical with another and that the grounds
of this connexion or relation are based on non-problematic con•
siderations. These latter are implied but not stated explicitly
in the hypothetical statement. If A is B, C is D is grounded, for
example, on the non-hypothetical implication that C is A and
B is D, and means not that we imagine or suppose or assume
 INFERENCE
that A is B i not that the supposition C is D follows from the
supposition A is B, but that the question whether A is B is
identical with the question whether A, which presupposes C, is B
which necess1tates D, or whether AC is BD, inasmuch as C is A,
and B is D. That is the full expression of our thought. If the
thought is not fully expressed (is expressed, that is, without the
ground of connexion, as '1f _A is B, C is D '), the adequate ex·
pression is 'the question or problem whether A is B is identical
with that whether C is D as well as A is B '. A relation, onC'C
more, of problems, not of statements.
 VII
DIRECT AND INDIRECT ARGUMENT
THE FICTIONS OF MODERN MATHEMATICS
§ 312. As the forms of language encourage certain mistakes
about the nature of hypothetical thinking, the question arises
how far they are to be observed and whether they are capable
of modification. Ilypothet1cal thinking 1s so common in ordinary
life that there can be no idea of using a merely technical and
'artificial' set of expressions. But some of the misleading
expressions are unnecessary and the ordinary language provides
the right ones. Thus the causal form of the hypothetical which
sometimes occurs in a reductio ad absurdum proof is unnecessary
and can be discarded. We need not put '1f A is B, C is D' in
the erroneous forms 'since A 1s B, C is D' or 'because A is B,
C IS D'.
The case 1s somewhat different w1th such expressions as
'assume', 'suppose', '1magme'. We have already pointed out
the natural meamng of 'suppose'.
It clearly must have passed to a different meaning, if usable
at all, when applied to an hypothesis which we do not suppose
to be true, the consequence of which we are investigating. We
have not decided here either way. The same is true. of 'assume'.
There is a sense of 'assume' (as applied to a doubtful or hypo•
thetical matter) in which it 1mphes a dec1s1on ; and we must
be careful here to avoid a confusion. The same is true of
'suppose', but yet it is more usual with 'assume' than 'suppose'.
A man is uncertain which of two poss1b1hties for him is true
and they concern a matter in which action has to be taken.
He has to decide how to act, which means that he must act as
if one or other were true. If he decides to act m one way, he
acts and decides to act as if one or other of these alternatives
were the true one. In that sense he decides for one alternative,
and only in that sense. He does not decide that it is true ;
that is an impossible feat. He does not think that it is true ;
554 INFERENCE
that also we have seen to be an impossible feat. He may
suppose it to be the more probable alternative.
There is another case of decision which may seem rather
theoretical than practical. Of two alternatives a man may
choose one to argue from and to develop its consequences. For
example, in certain chemical or biological investigations where
a man may have to choose between conducting one set of experi•
ments and another, or between treating the living body in one
way and another. Slbch dec1S1ons, however, are clearly practical,
for theoretically he would conduct, 1f possible, alternative experi•
ments. Herc we may say that a man acts on the assumption
that one alternative is true rather than another.
Putting aside these uses of 'assume' and 'suppose' as obviously
inapplicable to the case where we say '1f A 1s B, C is D ', without
either decidmg practically in favour of the alternative that A is
B, or thinkmg 1t probable, we may ask what it is that corre•
sponds to the expression 'let us suppose (or ' suppose') A is B
and see what follows'. These words arc nothmg but a way of
proposing a problem for consideration ; they really mean 'let
us fix our attention upon the question whrther A 1s B, and ask
what follows if A is really B '. We shall then see that 'if ' mtrudes
itself and therefore can't possibly be replaced by 'suppose'.
This 'let us suppose' is, then, really a way of proposing some
definite problem ; m other words it is a kmd of decision, viz. to
consider one aspect of a problem. It is so much more convenient
to use the short than the long expression which appears to
be the legitimate one, and this m language is a scarcely
resistible consideration, and usage will be too much for us. For
ordinary purposes we may therefore retain such expressions as
'suppose', 'supposing that', 'let us suppose', 'let us assume', &c.,
remembermg of course what they really stand for ; but it will
be imperative in a philosophical investigation, or where there
is a possibility of confusion, to substitute the true equivalents.
The urgent necessity of the latter will appear m the subject we
are now approaching, the treatment of reductio ad absurdum
arguments. The confusion caused by not doing so is mainly
re!ponsible for the modern pseudo-mathematical monstrosity of
non-Euclidean space. There is no need in any scientific argu•
ment to use these somewhat misleading expressions 'assume' or
 Di,ect and Indi,ect A,gument 555
'suppose'. Their legitimate use is really for the decisions we
have mentioned.
§ 313. The true view of hypothetical statement and reasoning
is obscured by a certain misleading use of language, which 1s
found in the ordinary scientific representations of a reductio ad
absurdum proof.
The reductio ad absurdum proof, stated in the way we criti-
cized above, usually begins thus, 'assume A is B ', then follows
'because A is B and B JS C, therefore A is C ; but A is not C,
therefore A is both C and not C '. Now we could not in fact
advance to the conclusion usually suhJomed, 1f the first premiss
really corresponded to the verbal forms in which it is put. We
have to restore the hypothetical form which we have verbally
suppressed before we can make another step; we have to say
'therefore if A is B, A JS both C and not C ; which latter is
impossible, therefore A is not B '. That is, we have to remember
that there was no real premiss 'because A is B ', and that what
we really meant was '1f A is B, then because B is C, A will
be C'. Now misleading expressions hke 'assume A is B',
which put quite a false complexion on thought, inclme people
to think they have 'supposed' A to be B, and hence that they
somehow start with a 'conception' of A's bemg B. To avoid
this danger the 'if ' should be restored throughout the reductio
ad absurdum argument.
§ 314. It must not be supposed that the form of the reductio
ad absurdum is peculiar to a particular kind of matter.
We can replace any reductio ad absurdum argument with its
hypothetical premisses and hypothetical inference by an argu•
ment entirely non-hypothetical 1 We infer 'if A is B, then in
general C is D ', and 'C is D' is known to be false Then it
follows unhypothetically from C is not D that A ts not B, and
the steps of the non-hypothetical proof can all be obtained by
a sort of inversion of the reductio ad absurdum proof. This
brings out its true characteristics. Starting with our question
I §§ IOS, 259 II

[ 11 Cf note to§ 259 If the matter 1s so simple, 1t may well be asked why
the incbrect argument should be used. On this, see § 316. In the author's
du-ect proof of Euclid 1 7, observe the last step, which certainly seems to be
indirect]
11773-a L
 INFERENCE
whether A is B, we conduct an inquiry represented by a train
of hypothetical inferences till we arrive at a criterion which
decides the answer to our question. Now the reductio ad absur-
dum proof represents the process by the system of hypothetical
inference till it terminates in the hypothetical contradiction to
a known truth, which then serves as criterion. The non-
hypothetical proof consists m applying the criterion simply,
which of course is done in a system of non-hypothetical rcasonJng.
We may illustrate by a particular instance, drawn from the
Seventh Propos1t10n of Euclid's Elements, Book I. The truth
which it 1s desired to establish 1s that 'on the same base and
on the same side of 1t, there cannot be two triangles having their
sides which arc terminated at one extremity of the base equal,
and likewise those terminated at the other extremity, equal'.
The reductio ad absurdum proof given in Euclid runs as follows :
If A= B, then a+p = y.
But a+P > p.
Therefore y > {3.
Agam, 1f C = D, ,,+a= fJ.
Therefore, 1£ A = B and C = D,
fJ(= r+a) <r-
Therefore "Y > y+ a.
Thus the part would be greater than
the whole
Therefore A cannot - B, and C at the same time = D.
The corresponding non-hypothetical proof runs as follows :
Make A - B, and JOtn the cxtrem1ties of AB, by C and D,
to the other extremity of the base.
Take as before vertices external to each other's triangle.
Reqmred to prove that C cannot be equal to D.
Then, if A • B, a +{J = y.
But a+{J > {J,
Therefore "Y > {J.
But y+a > ,,, and,,> {J,
Therefore ,, +a > {J.
Therefore the sides of the triangle of which y + a and p are
base angles are unequal. Q. E. D.
The last step 1s that unequal angles at the base necessitate
 Di,ect antl Indi,ect .A,gument 551
unequal sides. It is important to see that this last step is non•
hypothetical and is mistakenly represented as resting upon a
reductio ad ab1urdum proof.
It would be proved by a reductio ad absurdum proof as follows :
If the sides were equal, then the angles would be equal. But
the angles are not equal. Therefore if the sides were equal, the
angles would be both equal and not equal. But by the principle
of contradiction this is impossible. Therefore the sides are
unequal. This is a mere illusory and unnecessary transforma•
tion of argument. It is always futile to appeal to the principle
of contradiction as if it were a &pecial premiss to be used on
a special occasion. For, so far as any premiss depends on it,
every premiss depends on it equally, and every premiss would
fail, if the prmc1ple of contradiction were not observed.
The general form of the above argument is this. Let A be
a condition of B. If B 1s absent A 1s absent, because if B were
absent and A were present, B would be present and not present,
which by the prmc1ple of contradict10n 1s impossible. Now the
fact 1s that the thing to be proved 'if a thmg is absent, its con-
dition A 1s absent', 1s absolutely sclf-ev1dcnt. It is an inseparable
aspect of the fact that A 1s the cond1t10n of B · and a false
argument always arises when we try to prove the self-evident.
We really only base 1t on itself in the negative form and so
bring in the principle of contradiction. You can indeed only
argue from the pnnc1ple of contradiction as a premiss by usmg
the prmc1ple of contradiction and so ad infinitum. In the
present mstance, smce the angles at the base of an isosceles
triangle arc equal, 1t follows immediately that when the angles
are not equal the sides are not equal. The latter follows
directly : there 1s no hypothetical reasoning m the matter.
Thus what we have done in the non-hypothetical argument is
to use precisely the same material as we used before in the
hypothetical form of proof; inasmuch as the proposition that
unequal angles at the base necessitate unequal sides is simply
a necessary aspect of the truth contamed in the statement that
equal sides necessitate equal angles.
§ 315. The general proof that a reductio ad absurdum argument
may be always expressed without a reductio ad absurdum can
L2
558 INFERENCE
be put rigorously as follows. As a preliminary it is necessary
to prove the following theorem.
Suppose that we have the argument:
(I} If A is B, C 1s D ;
(2) But C is not D, therefore A is not B.
This argument is self-evident as it stands and is an ultimate
form, not requiring a reductio ad absurdum. In the reductio ad
absurdum form it would be
(3) If C is not D, and A is B, C is both D (from (1) above)
and not D.
(4) But C is both D and not D is not true ; therefore C is
not D and A 1s B is not true.
Here (1) and (2) have the form, 1£ P 1 is true P 1 is true; but
P 1 is untrue, therefore P 1 is untrue. But this is precisely the
form we have m (3) and (4) Thus, mstcad of provmg the form
represented m (1) and (2), we have merely repeated it in (3)
and (4). Thus we can't help makmg this form ultimate and
self-evident.
In any reductio ad absurdum proof the argument depends on
a proposition or system of propositions non-hypothetical in form
and given as true Let this propos1t1on or system be represented
by P 1 • Let Q1 be our question or problem. Let 1t follow from
P 1 that if Q1 1s affirmed (or true), R1 1s true. Let R 1 necessitate
P 2 as true. Then, if Qi is true, P 1 is true Let 1t be known
that Pa 1s untrue. Hence Q1 1s untrue. This may be reversed
as follows: P 1 is untrue, therefore R 1 is untrue, therefore Q1 is
untrue. By the theorem we have just proved none of these
reverse steps will involve a reductio ad absurdum proof. It is
important to point out that Q1 may itself be a system : and it
would be better to repeat the proof for this case.
Let P 1 be true. Let our problem or question be whether
Q1, Q1, Q8, &c., are to be affirmed. Let Q1, Q1, Q8, &c , in
virtue of P 1 , necessitate R 1 , R2 , R 3 , &c, i. e that 1f Q1 be true
it necessitates R 1 , that Q2 s1mtlarly necessitates R 1 , &c. Further
let R1, R1, R8 together necessitate P.11. Let P.11 be untrue, and
it follows that if Q1, Q2 , Q3 are affirmed together P 2 is true ;
but P 1 is not true; therefore Q1, Q1, Q3 can't all be true or
affirmed together.
This may now be put non-hypothetically. Pa is untrue,
 Direct and Indirect Argument 559
therefore R 1, R 1, R 8 can't all be true: therefore Q1, Q1, Q1
can't all be true. Here as before the reverse steps don't imply
a reductio ad absurdum, '
We may apply this to Euclid's proposition, I. vii.
The angles at the base of an isosceles triangle are equal.
(Q1) therefore 1£ A - B, ,- =a.+ fJ (R1 ),
{R1) and 1£ C = D, fJ = y+ll (Q1).
Therefore, if A - Band C = D, a.+/J =rand fJ - r+a.
Therefore ,, > r + a.
Therefore 1f A= B and C .. D, r >,, +a.
But y is not greater than y +a.
Therefore A - B and C = D can't both be true.
To reverse this process we must begm by showing that from
r + 8 > y 1t follows that a.+ fJ = y and y + ll = fl can't both be true.
Now ,- +a> y, hence, 1£ y +ll - x, y cannot be equal to a
magnitude greater than x. Therefore 1t 1s impossible that
,-+a - fl and at the same time r = a.+fl. I e. from ,-+a >Y
follows 1mmed1atcly that r + 1l = fJ and ,- = a.+ fJ cannot both be
true. Therefore a condition of y + ll being equal to {J, and
a cond1t1on of y bcmg equal to a.+ fJ cannot be simultaneously
realized. Hence it follows that A - B and C - D cannot be
true together.
Now this has exactly the form y + a > y, •·• a.+ fJ = y and
f3 - y + ll arc not both true, wherefore A - B and C - D are
not both true; and thus 1t exactly reverses the steps of the
hypothetical reasomng.
§ 316. We observe Ill the first stage of this last argument
a certain art1fic1ahty, 1£ we suppose ourselves to start absolutely
with the non-hypothetical and not to have the hypothetical
argument before us. The statement that 1£ one magmtude is
greater than a second, &c., 1s of course true ; but the question
is, why should 1t occur to us and why should we know 1t to be
of any use? Secondly, even granted this statement, why should
we give x the value {J? For any magmtude whatever will do
for the purposes of the statement itself. Agam, why should we
take a.+ fJ for the magnitude greater than x?
It 1s here important to remark that this kind of difficulty as
to why we should select a particular premiss, the justification
of which 1s not seen till we have got the conclusion, is not at
 560 INFERENCE

all peculiar to such a case as that before us. If it were so

peculiar, we should be oblia-ed to infer that the reductio ad
absurdum was the natural proof in certain kinds of matter, and
so that our difficulty J1ad arisen through substituting a form not
natural to that kmd of matter. But 1f we look at any proposi-
tion with a non-hypothetical proof m Euclid, we find the same
kmd of difficulty. The constructmn is mtroduced without a word
of cxplanat10n as to wl1y it should be chosen and we don'f see
the reason until the problem 1s solved. In some simpler pro•
positions this is not so evident because we sec pretty quickly
what the construction is going to lead to. But the same remark
applies to all. For instance, the difficulty of accounting for the
construction m Elements, I. 47, 1s notorious.
Hence 1t docs not follow that the reductio ad absurdum mode
of proof belongs to any peculiar matter. The truth 1s simply
this. The step which appea.rs so difficult to account for m the
non-hypothetical argument 1s accounted for and arrived at quite
naturally m an hypothetical argument. The hypothetical argu•
ment in fact represents the process by which the construction
upon which a non-hypothetical proof depends is found. We
begm simply with our problem and we ask what an affirmative
answer would necessitate ; these necessary results we derive
from previous propos1t10ns, or from direct intuition m particular
instances, and it 1s the carrying forward of tlus chain of results
which suggests the construction.
The hypothetical process therefore combines in itself both the
method of discovery and the proof, and 1s the proper scientific
exposition. The non-hypothetical proof to wluch we are accus•
tomcd is a sort of scientific pedantry, a and 1t is consequently
a great mistake first to give what 1s called analysis, which
corresponds to the l1ypothetical process, and then to follow it
by a synthesis, which 1s the non-hypothetical part, thus putting
aside analysis as 1f 1t were a sort of accident. It 1s an error
because 1t conceals the true process of thinking.h
[• From Lotze, Logic, Ji 4, § 206 • Es 1st log1sche Pedantene ••. fiir
euien Satz, der md1rect s1ch DUt zwe1 Worten schlagend bewe1sen lii.sst, eine
dtrecte Ableitung zu suchen ••• und die Einsicht 1n den mneren Grund lhre8
Vorhandensems um nichts re1cher macht'
b Indirect demonstration m geometry has often been said to show that
a theorem 1& true rather than why it 1S ]
 Mode,n M4lhematico-Phuosophical Fictions 56z
§ 317.• The modern theory of the possibility and conceiva-
bility of two kinds of space, respectively called hyperbolic and
elliptic, different from those of which we are said to have experi•
ence, 1s connected on one side with a mere misunderstanding
of hypothetical reasoning m general and of the nature of reduclio
ad absurdum proof in particular. One form of 1t began not at
all through an attempt to find whether another kmd of space
was possible or thinkable but was a quite unforeseen result
of an attempt to prove Euclid's Twelfth Axiom. In Euclid this
axiom is the premiss from which the equality of the three
interior angles of any triangle to two nght angles is proved ;
on which the immensely greater part of geometry depends.
Possibly the need of some proof of this axiom would not have
been felt or the question at least not raised so soon, 1f for Euclid's

lJn IllIIII
A C B
rill 11 IIrrr.
A

axiom a somewhat simpler proposition had been substituted,

C B

from wluch it could have been denved and which is really

self-evident. Such a simpler proposition is this 'If two straight
Imes have a third perpendicular to both of them, the first two
are always at the same distance from one another ' ; that is, 1f
a perpendicular be drawn from any pomt m the one to the
other, this perpendicular 1s of constant length. Or, what comes
to the same thing, given a str:ught lmc AB and a stra.1ght line
DC perpendicular to 1t and of fimte length, any lme through
D, such that the ordmates from it to AB are not equal but
mcrease or decrease on both sJdes, cannot be a straJght hne but
must be bent. That this 1s so, 1s absolutely self-evident. From
it Euclid's axiom can easily be derived ; and from 1t the rcfuta•
tlon of these hyperbolic and elhpt1c systems.
[• TI11s XIIth Axiom IS called 1n Euclid's Elements the sth Postulate.
In old English books and still on the Contment it is called the XI th Axiom.
For attempts to prove th.J.s 1n antiquity see Proclus in Eucl i (Fnedlem), p, 364.
I have left the author's words stand.mg as this part 1s the cherished conclusion
to wluch h1S dlSCussion of hypothetical reasoning was directed,]
 INFERENCE
It would be denied that this was self-evident, but here mathe·
matics would be involved in a contrad1ctJon. For the funda-
mental prmc1ple of the application of differential calculus to the
treatment of curves 1s certainly far less self-evident than this,
At all events 1t may even be doubted whether it is self-evident.
The proposition that the mterior angles of a triangle are
together equal to two right angles 1s further connected with
Axiom xu m the sense that either can be denved from the
other. The attempt to prove Euclid's axiom was made by the
method of 'assuming', as it is called, that it was un.true, and
then deducing conclusions m the hope of finding some contra•
d1ct10n, 1. e. the attempt was made to prove it by a reductio ad
absurdum, The contradiction to Euclid's axiom may take the
form obviously that the three interior angles of a triangle are
either greater or less than two right angles. Each of these
contradictions was, as the phrase 1s, assumed, and a number of
theorems deduced from each (merely hypothetical theorems
of course). The result would be ordmarily described thus: the
theorems deduced from the assumption that the mterior angles
of a triangle are less than two right angles never showed any
contrad1cbon. This stnctly means that no proposition was
hypothetically arrived at, the truth of whose contradictory
seemed more self-evident than the contradictory of the ongmal
assumption Otherwise the theorems arrived at were in the
figures correspondmg to them constantly m contradiction to our
faculty of representing straight lmes-a faculty on whose
correctness the whole of geometrical thinking depends ; but
this point will not be <lwelt on for the moment.
In the case of the assumptJon that the interior angles are
greater than two nght angles, a quantity of coherent theorems
can also be deduced showing no contradiction, m the above
limited sense of contradict1on. But one of them docs lead to
a contradiction a of the proposition that any two s1des of a tri•
angle are greater than the third, a proposition accurately proved
in Euclid, without presupposmg the Twelfth Axiom, But the
failure to get a certain kmd of contrad1ctton in thefirstcasereacted
strangely on the mathematicians who first tried the method
[• Wilson notes that this was discovered by Legendre, but mdependently
by lumself.]
 Modern Mathematico-Pnilosophical Fictions 563
described and who evidently were destitute of logical knowledge
or any philosophy of the true procedure of geometry. The
thought arose that the process had led to no contradiction
because in some sense the hypothesis argued from was valid.
The hypothesis that the interior angles of a triangle are less
than two right angles may be replaced by its equivalent in terms
of two Imes and a common perpendicular. But this supposed
property of straight Imes contradicts the property on which
'Eucbdean geometry' depends. The Euclidean system, how•
ever, not only doesn't lead to any contradiction of a self-evident
truth but to no contradiction of our faculty of representing
or, more accurately, apprehending geometrical figures. And thus
1f the one geometry were vabd the Eucbdean should be so
a fortiori. There was therefore no attempt to deny the validity
of the Euclidean system, but the really inevitable conclusion
that if the latter 1s true the former is false '\\las evaded, by
supposing that the former (hyperbolic) system represented a
different kmd of space m which straight Imes m a plane had
these curious properties,
The fact is that to these nonsensical words corresponds no
thought whatever · they are mere words.
Next, in order to make this more possible, the ordinary idea
of straightness was replaced by a definition that a straight lme
was one such that 1£ two of them comc1ded in two points they
must comc1dc altogether. Then of straight Imes so defined there
were supposed to be varieties, one corresponding to Euclidean
geometry, and others corresponding to spaces of different kinds.
And mathematicians who believe m this actu~lly suppose them•
selves to be able to conceive such a space Oddly enough also,
the geometry which betrayed a contradiction was now retained,
§ 318. At the root of this there arc several fallacies of an
elementary kind. If we waive an obJcction which is in itself
quite fatal to the system, that 1s, that the correspondmg .figures
contradict our faculty of apprehension, there is first the ele-
mentary mistake of supposing that because a tram of hypo•
thetical argument {for 1t 1s purely hypothetical) has not as yet
Jed to a contradiction of a certain kind 1t never will.
It ought to have been a corrective to this to notice that the
so-called elliptic system also showed a quantity of coherent
 INFERENCE
theorems, without the aforesaid k10d of contradiction, if de-
veloped in certain ways. But of course the mistake is too
obvious to need such a corrective.
The supposed theorems arc entirely hypothetical. The start,
as we have seen, 1s hypothetical, and this of course never becomes
non-hypothetical 10 the process of deducing hypothetically from
it; 1t no more becomes non-hypothetical at the thousandth
deduction than at the first. In fact the system has pret1sely
the form of a reductzo ad absurdum argument ; that 1t has not
led yet to a certain kmd of absurdity makes not the slightest
difference to its entirely hypothetical character.
As we have seen abundantly, 1t 1s not true that we beg10 by
conceiving what we start with 10 the reductio ad absurdum (or
in any hypothetical argument), and then find that 1t 1s not true;
for 1f ever conceivable 1t would be always so. In that case the
proof would be that 1t was not true though conceivable; whereas
we have seen that the proof lies 10 showing 1t 10conce1vable and
altogether 1mposs1ble for thought. On the other hand, if the
thmg was conceivable, 1t would be true, for 1t would be seen
to be necessary (as explained) 1 and there could be no thought
of test10g 1t by arguing from 1t hypothetically
The fallacy may be made explicit thus · The question is
proposed whether a straight hnc (A) can have a property (B).
We are so far from concc1vmg A to be B, that by no feat of
the 1magmat10n can we represent the connex10n of A and B ;
so that 1t 1s absurd to suppose that the m1tial step, wb1ch may
be put 10 the deceptive form 'assume A to be B ', means that
we conceive A as B . for we cannot do 1t.
But we may waive even this form of objection. It is enough
that at first we are confessedly at least unccrta10 whether A
and B can be connected. And we have shown that this means
we do not tlunk the connexion but wonder whether 1t can exist
and what 1ts nature would be 1f it did. We thmk A and we
thmk B · we do not thmk their connex10n-it is a problem to
us and rema10s so through all the hypothetical deductions.
'Assume A is B ', then, only means that we ask whether A
can be B and propose to see what would follow if A were B.
1 viz necessary to thought 1n the sense of the appreheI1SJ.on of a necessity
in the obJect
 M ode,n Mathematico-Philosophical Fictions 565
What difference, then, does the train of argument make which
as yet has not led to a contradiction, that is, if it were really
true (as it is not) that it had led to no contradiction? Merely
that, as far as we have gone, we have not found anything which
decides our original question by showing that A cannot be con•
ceived as B ; but 1t remains equally true that nothing whatever
is found which makes an affirmative answer possible : and we
have made absolutely no progress to or in the act of conceiving
AasB.
But the supposed new conception of another kind of space
means that we can conceive a space in which A is B, which is
only a disguised form of saymg we can conceive that A is B.
Hence the view that the process of deduction described has
made this a possible conception is an obvious fallacy ; and 1t
1s a very naive mistake to suppose that the mathematicians
who have conducted these kinds of deductive operations have
discovered the concept10n of a new kmd of space. This may
be put as the mistaking of the problematic conception A? B
for a true conception, or rather the idea that the problematic
conception has become a real conception.
We have seen, then, that a new apprehension of space could
not originate by the method supposed : and, as this is the only
method by which it 1s thought to have originated, this is a
sufficient refutation.
§ 319. But there is another way of makmg this clear. Really
the dec1s1ve thmg 1s that all the proposed constructions are
impossible for our faculties. We cannot 1magme such lmes
straight, we cannot tlunk of them existing as straight. But the
advocates of the theory arc in general inaccessible to such an
argument because they are under an 1llus1on about the function
of the imagination m geometry, owmg to their want of all
critical mvest1gat1on of the relation of thmking to imagining
and of the thmkmg consciousness in general to its object. In
short, they suppose that it docs not matter that the so-called
straight lmes cannot be imagined straight, which 1s as non-
sensical as to suppose that there could be a number 'two' which
we had necessarily to imagine as 'three'.
But they may be approached on their own ground. It may
seem that if a man affirms he has a conception, however con•
 INFERENCE
vinced we may be that he has not, we can do nothing with him
because he must claim to know what he has in his own mind.
And similarly, if he affirms he hasn't a conception which we
know he has and yet 1s m some illusion about. But m both
cases a refutation 1s possible. In the latter case, by showing
that he uses the conception and so must have 1t; i.e. that
certain of his thoughts and actions presuppose 1t. In the former
case, by showing that no use whatever 1s made of the supposed
conception and supplementing this by showmg that the illusion
that there is such a use arises from the mistake about an hypo•
thetical or problematic conception
Th<' truth is that no other apprehension of space is used except
the Euclidean m these theories. As we have seen, the start in
such hypothetical reasonmg 1s from the question whether A can
be B (or A and B can be m a relation R) where A and B {and
R) are elements of which we have a certam amount of undoubted
knowledge. Advance in thought beyond is only possible through
non-hypothetical reasonmg from what we know of A and B,
from what we do and must apprehend them to be, to something
further which this necessitates. Now, ex hypothesi, we start
with no other conceptions than the Euclidean. A and B are
notions we have from Euclidean space alone, and our next step
must consist m rcasomng in the only way we have as yet-the
Euclidean-to the consequences of A and B. In fact, asking
whether a Euclidean clement A can have the Euclidean property
B, we prove that A1 mvolves A and B mvolves B1, by a purely
Euclidean method m each case, and the manner of the rcasonmg
is such that our question whether A is B 1s seen to depend on
the answer to the question whether A1 can be B1 : or, puttmg
it hypothetically, 1£ A 1s B, A1 1s B1 .
Now not only must the process begin thus but 1t is the only
way m which it can possibly be contmucd. This is evident from
what 1s said above, that the process of deduction cannot lead
to any new apprehension ; every stage is as interrogative as the
preceding. But it is also at once verifiable from the actual
procedure in the theorems m this subject, In all cases the
forward progress 1s made only by argumg positively or rather
unhypothet1cally from the Euclidean elements to the Euclidean
consequences.
 Modern Mathemalico-P1"1oso,pmcal Fictions 567
The analysis is this :
I. (i) Can A be B (A and B being Euclidean)?
(ii) Euclidean argument showing A1 involves A, and B
involves B1.
(iii) Hence, if A is B, A1 is B 1•
II. (i) Can A1 be B1 ?
{1i) Euclidean argument showing A1 involves A11 and B1
involves B1 •
(iii) Therefore if A1 is B1, A2 is B 11, and if A is B, Aa is B1,
III. Then again, can A1 be B1 ? and so on.
So that unconsciously these mathematicians use no other con•
ceptions of space than the Euclidean, none other of course being
possible for thought, while they imagine themselves to be talking
of another kind of space. The origin of the illusion is obvious.
§ 320. To corroborate the view that hyperbolic space is another
kind of space, under the generic idea of space, it is usual to
point out that the various theorems in hyperbolic geometry have
their analogues in the Euclidean system, so that the latter can
be derived from the former by certain suitable changes. This
is an elementary mistake in elementary logic. For the relation
in question may hold between a system of true theorems and
a system known to be false.
Suppose we have a system of theorems derived from a theorem
in which certain elements A, B, C are connected in a certain
way Suppose we make some change in A so that it becomes
A1 and then let the connexion which the theorem would make
between A, B, and C after this substitution be untrue. The
change of A into A1 will involve a corresponding change in all
the derived theorems and we shall have a system of false
theorems side by side with the origmal system of true theorems.
Now in the first of these, by changing A1 into A we shall recover
the true theorem about A, B, and C. Similarly, by making the
analogous change in each of the derived false theorems, we shall
recover each of the true theorems.
§ 32 I.• There is a fallacy which has helped towards making
the illusion of extra-dimensional space seem plausible. It is
supposed we have 'space of one dimension' in the line, 'of two•
in the plane surface, and 'of three' in the solid body.
r He 1s referring to an amusing little book by Edwin Abbott called • Flatland •.]
 INFERENCE.
It is falsely considered that these three dimensions are three
different spaces and that it is a mere matter of adding one
dimension to another. Thus the idea suggests itself that the
addition can be continued beyond three (ad lib.). The dimen•
sions are in fact treated like units independent of one another,
each hke the other and so capable of addition. This is a serious
fallacy. 'D1mens1ons' cannot be treated as independent units.
There is no space of one dimension and none of two ; nor dqes
space of three d1mens10ns arise by addition of dimensions. It
is not formed by addition and 1t cannot be increased by add1t1on.
The so-called two-d1mens1onal space necessarily presupposes and
is a part of space of three dimensions. The one-, two-, and three-
dimensional space arc in organic unity and cannot exist apart.
They presuppose one another and form parts of an inseparable
whole
To illustrate : the idea of a plane is quite impossible except
with ref erencc to space on both sides of it. Instead of a third
dimension being aclded by erecting a perpendicular on the plane,
the plane as surh necessitates the perpendicular, and a plane
a~ a plane necessitates three-dimensional space.
 VIII
THE MATHEMATICAL EXPRESSION OF PROBABILITY
AND ITS RELATION TO REALITY
§ 322. a WE have spoken of our belief or feeling of confidence
depending partly on what we call the strength of the evidence,
and have said, what is indeed obvious, that what we call the
probability of an occurrence does not express anything objective
but the degree of confidence which we entertam, a purely sub-
jective matter. We tend, however, to forget that strength of
evidence 1s merely somcthmg for us, and this leads to a fallacy
which 1s often illustrated m the treatment by mathematicians
of the doctrmc of probability.
Knowledge has no degrees, any more than reality. There is
nothing m reality to intervene between bcmg and not being,
nothing m knowledge between truth and falsehood. But, though
there are no degrees in knowledge, there is, m our mmds, some•
thmg between absolute ignorance, the rcframmg from pro-
nouncing a judgement, and knowledge. This is the region of
belief and here only differences of degree are possible. We
express these degrees commonly by the words probable and
possible and the cxpress10ns more or less probable. The lowest
expression of belief 1s that the thmg 1s possible, a higher degree
is denoted by probable, and degrees of this are indicated by
more and less. But though possible is applied in one way to the
lowest degree of belief, the word improbable is used commonly
of somethmg lower still ; not, however, usually of our own
beliefs. We contrast an improbable opinion with another which
we call possible or more probable, and this latter would be our
own belief. An improbable belief, then, is one that is rejected
relatively to another, it cannot be rejected absolutely, however
improbable we may estimate it to be. It must be m the category
of what is possible, and if we reject it we pronounce it impossible,
s
r, cf. I r, 92 This and the next two sections are early lectures as
delivered. §§ 325-7 are from his article m Natu,e, 12 xu oo.]
570 INFERENCE
not improbable. Now, as we said above, what influences the
mind in these degrees of belief has no corresponding influence
over facts, and the terms possible and probable are therefore
alien to the objective, to reality, and merely express a subjective
attitude.
§ 323. We may ht,re offer an outline of an analysis of this
subjective attitude Consider first the probability of a universal
statement, all S is P. This implies the necessary conditioning
of S by P. Now from the side of the experience of particulars
we may approximate to belief in this universal statement in
the manner described under the head of enumeratio simplex or
by the method of agreement. This is usually called arguing
from particular instances of a rule to the umversality of the
rule. For this purpose the quality of the instances is really
important, the quantity bemg of advantage only because it
increases the probability that a sufficient variety of instances
will be examined. The more instances we take, the more probable
we think it 1s that they should be varied, and if S and P have
remained permanent while the other elements varied, we think
that they arc necessarily connected. But now this is just the
problem that we have to examine, for 1t is obvious that we
cannot be sure that while we increased the number we have in
the same proportion, or even at all, increased the variation. It
is thus possible that the larger quantity has not in fact increased
the variation, and we may therefore be altogether mistaken in
giving more weight to the increased number of instances than
to those we had to begin with. How, then, is it that the mind,
notw1thstandmg tlus poss1b11ity, inclines to attach more value
to the larger number of instances?
The general answer to all questions like this seems to be that
the mind inclines to that side in favour of which there are at
all events the larger number of possibilities, other things being
equal, that is, when we can't or don't distinguish between the
qualities of the poss1b1hties. In the case before us, then, suppose
that one instance is observed in which S is P. Suppose next
a second in which S is also P. There 1s now a possibility that
this instance may vary m its other clements. A third instance
will give a further poss1bdity, and thus we increase with each
instance the number of possibilities, at all events, of a variation
 P,obabiUty
of the different elements. Now the second case of S which we
approach might, so far as we know beforehand, not be P. If,
then, it turns out to be P, one possibility of the contradictory
of all S is P is destroyed. The same is true off urther instances,
and as a matter of fact our mental inclination, our belief,
strengthens as the number of instances which might have been
contradictory but are not increases. Whether this is rational
or not, at least it appears to be the fact. A simple instance
will illustrate it. Suppose that we are entirely ignorant of the
conditions under which a die is thrown, but know that it has
six faces differently marked, one of which is the ace. The
question might arise whether it is more probable that the ace
or any one of the other numbers would turn up on a single
throw. Now we can't d1stmgmsh between the value of the
poss1bihties that the ace, the deuce, or any other particular
number will be thrown. Thus we have one possibility in favour
of the ace and five possibilities of something not the ace coming
uppermost, and we should m fact say it was more likely that
a number not the ace would be thrown than the ace. The
mind, we see in this instance, chooses the side on which the
larger number of possibilities are. In this there is at least
a certain appearance of rationality, smce from the point of view
of our knowledge the appearance or non-appearance of the ace,
both possible cases, are not on the same footing. If we have to
choose, then, there is a difference, so to speak, in favour of one
side, and we take that side, knowing nevertheless that the
difference may be only apparent and that the favoured event
may be the one that does not happen. It is: a decision of the
will, just as if a man should come to two roads quite unacquainted
with their difference in respect of his destination. He must
decide somehow, and he may, his mind being otherwise abso•
lutely undetermined, choose the one m which he happens to see
something consistent with the hypothesis that it is the right
road. That md1cation, however slight, serves to put an end to
his indeterminate attitude. We may, then, perhaps thus explain
our attitude when we assign more probability to the proposition
all S is P as the number of instances mcreases.
§ 324. Take next the singular statement, the answer to the
question ' is this S a P ? '. Either we know the conditions of the
M
 INFERENCE
appearance of Pin S or not. Suppose we don't, the lowest point
of knowledge is to know nothing against the appearance of P in,
S. Here we may say that the bel1ef jg for us a possible one.
Secondly, though we do noL know that the elements, say, 11, b,
of S condition P, yet from particular experience we may have
:reached a belief, more or less probable, that they do. This is
obviously only an instance of our first case of all. Suppose,
however, we know the conditions of the presence of P but do
not know whether they are present. As before we have first
the case m which we know nothing but assume that P is present
on the grounds of mere possibility. Next we may know that
one of the conditions, say a, is present in S. If, then, we have
nothing to mfluence us from the side of the absent conditions,
our belief in the probability of the presence of P will mcrease
with the number of conditions we know to be actually present.
Thus if a, b, c, d be the conditions, finding a and b we are more
ready to believe than if we found a alone, and :finding a, b, anc!
c we are still more ready to believe We have, m fact, an
application of our formula that our belief increases in degree
with the increasmg number of possibdit1es in favour of the
statement m question, an increase wluch has an important
negative side, the removal of one obstacle after another to the
truth of the statement.
It remains to consider the case where we are influenced by
the rema1ning cond1t1ons, those about whose presence we are at
present uncertain The measure of probability we have arrived
at will now be affected by the probability of the presence of
these conditions. Clearly we may here have to take into account
the probability of the existence of their conditions, if we know
them, but this process must stop at last with the question of
the probability of the association of the ultimate group, say, of
those conditions with some other facts either in the present or
the past. Here we return to the first case of all and the proba•
bdity will be estimated in that way. If, however, we don't
know the conditions upon which these conditions depend, we
shall have again the question as to the probability of the
association of these conditions with certain facts before us, and
this again is the first case of all.
So far we have cons1dered the opposition of a single fact to
 P,obability 573
its negation and not to a definite number of possible alternatives.
There are cases, however, and these are the most ordinary ones
in the mathematical theory of Probability, where to a given fact
is opposed a limited number of alternatives, as in the instance
of throwing the die. The question here arises of the comparative
probability of any one clement being realized, given that we arc
certain that one or other of the enumerated alternatives will be
realized. We require some sort of common measure of the
possibilities in favour of each element in order that we may
understand how it is that we incline to one rather than to all
the rest, or to one rather than another of a given pair, and so on.
The mathematical theory of probability is an attempt to reach
such a common measure, and though there are many complicated
calculations in this part of mathematics, the principle of them
all admits of simple statement.
§ 325. We may avoid certam confusions which often arise in
the statement and application of the mathematical theory if we
form a clear idea of what is meant exactly by the fraction which
is said to represent the probability of an event,
A good statement of the ordinary account of it is given in
Todhunter's Algebra : 'If an event may happen in a ways and
fail in b ways, and all these ways are equally likely to occur,
the probability of its happening is ~b, and the probability of
a+
its failing is _!_b. This may be regarded as a definition of the
a+
meaning of the word probability in mathematical works.' A
definition must not assume and use the notion to be defined,
Here probability is defined through cases 'equally bkely to
occur'; but 'equally likely to occur' means equally probable,
and so the definition assumes the very notion which causes
difficulty, and of which we require the explanation.
The first thing to settle is the meaning of these 'equally
likely' cases. Is the equal bkebhood a quality in things them•
selves, or is it something in our minds only? If it is a quality
in things it can only mean equal possibility of occurrence or
realization. But if a number of cases, mutually exclusive as
intended in the above definition, were in the nature of things
equally possible, not one of them could happen. If the claim
112
5-74 INFERENCE
of any one of them on reality were satisfied, so must the claim of
any other, since these claims are equal, and therefore if one
happens all must. But by hypothesis if one happens no other
can ; thus the only possible alternative is that none of them can
happen. (It 1s precisely on this principle that we decide that
the resultant at a pomt of two equal forces, whose directions
include an angle, cannot be m any other direction than the
bisector of the angle, and that there can be no resultant of two
equal forces which act in opposite directions.)
The equal likelihood then intended cannot be anything in the
nature of things, because it is assumed that one of the equally
hkely cases will happen. It 1s really only in our mmds, where
there 1s an equal balance of reasons for and against two or more
events, and 1s due solely to our ignorance, smcc 1f we knew which
was to happen there could be no such balance and indecision.
This is clear 1f we consider what is the reason why we pronounce
one event more likely or probable than another ; it is because
we thmk there 1s more evidence m favour of the one than m
favour of the other, however the 'more' may happen to be
measured Two events are equally likely for us when we know
nothing more m favour of the one than we do of the other;
when the state of our knowledge-and (1t is important to add)
of our ignorance-is the same for both contmgencies. This view
agrees with the actual procedure in mathematical examples. If
a bag contains n balls, and one 1s to be drawn 'at random',
there are said to be n equally likely cases ; that is, each of the
n balls 1s equally likely to be drawn. Clearly this only means
that, as we don't know how the hand 1s gomg mto the bag,
we have no mformahon m favour of the drawmg of any one
ball as compared with any other and no information agamst
the drawmg of any one as compared with any other.
•Equally likely• cases then bemg such that owing to our
ignorance the evidence m favour of any one is no greater or
less than the evidence m favour of any other, the meaning of
the definition of probability criticized above is evident. It is
not a definition of probability, but it is the definition of a certain
way of measurmg evidence. We are entitled to say that one
event is more probable than another when, the evidence before
us being decisive for neither, that in favour of the one seems
 P,obabiluy
to us, according to some standard of measurement, greater than
the evidence for the other.
f 326. Now what the mathematical analysis does is not to
alter the ordinary meaning of 'probability' at all, but to find
a standard for the measurement of the more and less in evidence.
The whole possibility before us in any given contingency is
divided into a number of cases, 'equally likely' or 'equally
possible', in the sense that they are equal from the point of
view of the evidence in favour of each of them; then if one
event has more of these equal possibilities in its favour than
another, it has in this sense •more' evidence m its favour, and
so, m accordance with the usual meaning (as above described)
of 'more probable', 1s more probable than the other. And here
the 'more or less' m the evidence is not a mere 'more or less',
but has a definite numerical measure. The evidence being, so
to speak, divided mto equal umts, the strength of the evidence
in favour of a contingency is measured by the number of these
units m its favour. Thus 1£ the total of equal possibilities, one
of which must happen, for the events A, B, and C is n, of which
a involve A, b involve B, and c involve C, the comparative
strength of the evidence in favour of A, B, and C respectively
is measured by the ratios a : b : c, while the strength of the
evidence for A, B, and C respectively, as compared with the
evidence for one or other of them happening (which is certain),
is, on the same prmciple, measured by the ratios a,
n n1
~ and n~- •
If, then, we symbolize the strength of the evidence for A, B,
and C by ~, b, and ~ , and similarly that for cine or other of them
n n n
happenfog by !' = 1, these quantities have to one another the
n
ratios required. We then arrive at the true meaning of the
fraction which is said in mathematics to be the 'probability' of
a contingency ; and much confusion might be av01ded 1£ we
called the fraction, not the 'proba.b1hty', but the 'modulus of
the evidence', and the so-called equally likely cases not 'equally
likely' but 'equi-evid,,ntial', or by some more convenient name
conveying the same idea.
§ 327. It must, however, be insisted that the above is only
 INFERENCE
one way of measuring the evidence, and is not applicable to all
cases. Indeed, the more important matters of daily life usually
do not admit of it, for there are qualitative differences in strength
of evidence which cannot really be measured quantitatively, and
that is why the application of mathematical probability to the
testimony of witnesses is so obviously futile.
The solution of every mathematical problem in probability is
in the last resort only the finding of a modulus of evidence, .in
the ratio of the part of the whole number of equi-evidential
cases which involve a given contingency, to the whole number
of such cases ; and w1th the finding of the modulus the strictly
mathematical work ends. Mathematics, as such, has nothing
to do with the inclination in our minds to expect the event for
which the modulus of evidence is greatest (or 'the probability'
greatest), or the inclination, when some practical step has to be
taken, to act on the hypothesis that the event will happen for
which the evidence to us seems strongest. Unfortunately, how•
ever, there i& too often a tendency to confuse the mathematical
measure of the mere state of our mmds with the measure of
somethmg in reality. This is an insidious fallacy and we are
not unlikely to fall mto it in one form when we have escaped
it in another. It is the origm of various mistakes-e. g. the
inclination to expect that the actual proportion of the occur-
rences of the event will tend to conform to the proportion
represented by the mathematical probab1hty1 in other words,
conform to a formula of our ignorance. The mistake of sup•
posing that the mathematical probab1hty could be confirmed
by actual observation belongs to this head. The attempt to
regulate betting by mathematical probab1hty is another instance
of the fallacy of confusmg the subJcctive with the objective.
The truth is that an observed average may be made the basis
of a mathematical 'probability' or modulus of evidence, by
a process which could easily be explamed ; but, though a
'probability' may be based on an average, an average can never
be based on a 'probability'.
 PART IV
SPECIAL LOGIC
 I
INDUCTION. THE GENERAL FORM OF THE
EXPERIMENTAL METHODS

§ 328. THE recognition of Induction as a method seems to

come about in two ways. It arises first with the syllogistic
logic itself, because the ultimate premisses of syllogistic proof
cannot be given by the syllogism. Thus, if they are to be got
by any process at all as opposed to direct intuition, that process
(not being inference from a universal) can, it is supposed, only
be some sort of inference from a particular. Secondly, it arises
from certain farts of experience, because m scientific inquiry
gencrahzat1ons are met with which cannot be demonstrated in
the manner of the exact sciences but are founded upon the
observation of particular instances.
The general notion common to all ideas of induction seems at
least to be this, that 1t is a process in which we arrive at
a universal statement of the form 'all A is B' by the considera-
tion of particular cases of A which are B ; and this general
definition can be fairly maintained. But the definit10n is
commonly stated m a wider form, viz. that it 1s a process in
which we start from singular statements and arrive at a universal
conclusion (of which the singulars are instances) without pre•
supposing a universal statement, that is, without deriving or
deducing from that.
There is another way of looking at induction. Deduction is
conceived as (1) mferrmg the particular from the universal, and
(2) as inferring a consequence from its reason or ground or
condition, where condition is used m 1ts general sense and
includes cause. In Anstotehan terminology, deduction is
regarded as arguing from the universal to the particular, or
from the condition or ground to that of which it is the con•
dition 1 (and according to Aristotle the two should coincide in
the perfect scientdic demonstration)
1 tta9' llA011, tra.9' ftta.ffOJ', a.fr1ov, oli a.fr,w.
 ltJdumon and the E~perimental Methods 579
Induction by contrast is considered 1 to be the inverse process
to deduction, and accordingly as either arguing from the par•
ticular to the universal (the usual form of the conception of
induction) or as argumg from the conditioned to its condition
(of which a special case 1s arguing from effect to cause); the
process, that 1s, of mf errmg from something what 1ts condition
is, whether that condition 1s cause or not. Thus it is assumed
that as such an mverse process it is d1fferent m method from
deduction. This 1s uncritical because 1t assumes without
examination that the relation of condition and conditioned is
not reciprocal but one-sided, and further that this involves in
consequence a difference in method.
Now we have seen 2 that the relation of the true condition to
what it cond1t1ons 1s reciprocal, m !:.o far as each necessitates
the existence of the other : though there arc cases in which
there 1s a pnonty of the one over the other m the order of our
thinkmg.
This being so, it docs not follow that the method or process
of inferring the 'result' from its 'cond1t1on' must be different
from that of mfemng the 'condition' from its result. And we
have seen that m geometry 1t 1s actually the case that the
process is the same, even when there 1s a pnonty of one to
the other m the order of our thought.
Where there 1s no such pnor1ty we prove the converse of the
propos1t1on without assummg the proposition itself and the
method 1s as direct in the one case as m the other. The term
mverse is not confined to one of them : each of the proofs is
the mverse of the other.3 But when there 1s tac kmd of priority
referred to, so that the converse of a proposition is proved by
assuming the proposition (whether with or without" reductio ad
absurdum •), the proof of the proposition 1s the same m method
1 As, for instance, by Jevons [See, e g, Tke Principles of Science, chs VJ,
vu, xi,and x11 ]
I § 259,
1 e. g a direct proof of Euchd, Elements, 1 6, or of the converse of Elements,

ill. 31 (the angle in a sexrucU'cle is a nght angle)

• With rerl. ad abs as in Elements, 1 7, or without, as in Elemems, i. 48
(converse of Pythagoras's theorem).
[• Euchd, Elements, 1. 40, can be proved either with (as in Eucbd) or without
retl, ad abs. He also used his dlZ'ect proof of the converse of 1. 32; cf.§ 259.J
 SPECIAL LOGIC
as the proof of its converse. Both consist of geometrical intui•
tions or constructions.
In the first case neither proof can be called the inverse proof
because, each being independent, each is the 'inverse' of the
other, m the sense that the one is the converse of the other and
the one may give the stages of the other in inverse order.
In the second case the proofs are not independent and so the
first couldn't be called the inverse of the other, in so fal' as it
is really contained in the other, nor• does the dependent proof
reverse the stages of the independent. Each is, however, still
'converse' 111 its conclusion as compared with the other. But
the essential thing is that, while the 'inverse' method does not
apply to the first case, the procedure in the second case is
exactly the same m kmd m the dependent process as in the
independent, that is, whether the inverse method applies to 1t
(is employed to demonstrate it) or not.
Thus the given definition of mduction does not seem justified,
and if the account already given of the nature of a mathematical
demonstration is true, it is evident that the search for induction
cannot properly ongmate in the unreal distinction between
axioms and demonstration. So far as the demand for mduction
has arisen m tlus way, 1t is from a complete misunderstanding
of the nature of mathematical demonstration and is part of
a theory which supposes that the formal operation of the syl-
logism advances science by synthetical propositions beyond its
ongmal data. The secret of it all, 1. e. the recognit10n of the
universal and essential connexion m the particular instances,
seems never to have occurred to these mductionists, and hence,
in part, has arisen their misunderstanding of Aristotle's position
when he speaks of the induction of principles or axioms.1
§ 329. From what has been said of mathematical demonstra-
tion 1t follows that the idea of induction ought not to be traced
to a distinction of ultimate axioms and demonstrated proposi•
tions, since that d1stinct1on1 as we saw, involves a mistake. So
far as it has had such an origin, a serious misunderstanding of
the nature of demonstration is implied. Moreover, this mistake
has had a misleading effect upon the theory of induction
itself i because the true secret of mathematical method, which
' See note to § 329.
 lntlucti.on and D,,e Expm111ental Methorls 581
is the possibility of recognizing the universal in the particulm
(i.e. a universal truth represented by a universal judgement),
has never occurred to the inductive logicians. It has affected
also the understanding of the Aristotelian account of 'epagoge ', 1
which is often taken to imply the view that axioms are dis•
covered by inductive inference from particular instances. The
truth is that with Aristotle the axioms are self-evident, and
the induction is simply the mental process in which we re-
cognize the universal apart from the particular. When this
recognition is of axioms there is added 'intelligence ' 111 as the
intuition of their universal validity. In modern induction two
different forms are really recognized, about the relation of
which to one another there is the greatest confusion. One is
enumeratio simplex," the other is often called 'induction' simply,
and comprises what are called the experimental methods. This
is supposed to be the essentially modern form and we may
therefore term it modern induction. The former is characterized
by Bacon and his followers as the loose uncertain method
practised by the ancients ; and they often profess to replace it
by the new induction wh1ch they thmk of as a modern discovery :
while they speak of the old as a fallacy to be forgotten.
But we find to our surprise that the despised enumeratio
simplex 1s not only retained with a special function of its own
which the 'new induction' cannot fulfil, but has an absolutely
necessary place as a part of the new induction itself, and cannot
be got rid of. That is, these logicians are not only dnven to
retain enumeratio simplex as their only possible account of the
axioms of mathematics, but the more modern writers (as opposed
to Bacon, who would not carry h1s analysis so far) are driven
to recogmze that the new mduction itself depends entirely upon
an axiom called the ax10m of the uniformity of nature or of
universal causation. And here agam they have to resort to
enumeratio simplex to establish the axiom.b But worse than
this, we shall find that the reasonmg of the new induction as
1 lwa'Yfl"ri, Latin induct,o
• ,,oiJr,
intelhgence [as d1Btmgu1Shed from i11•tlffl/l'I, sctentlfic understanding
and sc:ience],

[" Described shortly by Mill, Syst,m of Logic, ui. J, § 2.

b Mill, 1b., w. 21, §§ 2 and 3.]
582 SPECIAL LOGIC
such is what these logicians are bound to call deduction or
syllogism. Thus the only element which could be called in•
ductive in the vaunted 1mprovement upon the enumeratio
simplex, is just this enumeratio simplex itself. There is a further
confusion of analysis : we are told that the axiom of induction
is itself the result of mduction. Now no process can possibly
prove its own presuppositions. Thus we have either this ele-
mentary fallacy or the circulus in probando must be avoided by
supposing an ambiguity of terms. For the process called induc-
tion which supphes the axiom cannot be the same in lond as
the process which works from the axiom. This seems to be the
fair account of the matter. The induction which presupposes
the axiom is the new induction. The induction which proves the
axiom is the old enumeratio simplex. This necessitates the
question, why should two processes apparently different m kind
both be called induction? What have they m common? These
logicians, however, are so unconscious of this difficulty that they
don't even raise it.

NoTE. Aristotle's authority could probably be cited for the

view that the axioms of the exact sciences are got by induction.
In connexion with syllogism he says, 'those are first principles
which do not depend on syllogism, they depend therefore on
induction ', 1 But he nowhere gives an account of the induction
of axioms by which they would appear, as in a modern induc-
tiomst, to be the conclusion of an mfernng process from experi-
ence. Moreover, he says more than once 2 that the first principles
of these sciences are self-evident-the exact opposite of what 1s
understood by an inductive ongm. Moreover, as we have said
br1efly above, what he describes as the induction of first prin-
ciples 8 is quite obviously the gradual process by which the mind
comes to recognize, m the abstract and apart from all particular
instances, a principle which it has been using in the particular
instances. This is not the inference of the validity of the
1 ~ ,.i,, &} IINl-ywy,} dp,cq lo-T, iro.l Toii 1<1:&6&1.011, cl Bi ITIIMrry,.o-p~ l,r Tolv iro.BoAov,

•lo-I• 4pa dp,cd It cw cl avMoy1up6r, liv o,,. In, uti'A.AvycuflDr I n ~ I.pa., E. N.

113gb 28-31
1 Top. 100b 18 , At&. PF 64b 34-6.
1
A11. Po ll, 19, especl&lly 100b 3-IS voiis a,, .r,, TO/II dpxGw,,, •• '"''"'""'
.r,x+
 Induction antl the E~pe,imental Methods 583
principles from the validity of the particular instances as such.
When he adds that the faculty of principles is intelligence (110ft}
and that this is the most exact kind of knowledge, this is just
what we should expect from a writer who grounded all knowledge
on experience and at the same time made the principles of
experience self-evident. Intelligence is the intuitive faculty of
reason by which we recognize this self-evidence, when an induc-
tion from experience has brought us to think of the universal
principles in abstraction from the particulars in which we first
thought the universal.

§ 330. It is usual, perhaps universal, to regard as ultimate

and irreducible the ordinary formulation of enumeratio simplex :
and there seems to be no idea that its presuppositions have to
be discovered or examined.
The general formula of enumeratio simplex may be fairly
described as follows. This A is X, that A is X, &c., all observed
A's are X. Whence the conclusion (as probable, not certain),
all A is X. This is what has seemed to be ultimate, and it is
here therefore that our inquiry must begm. The first thing to
notice is that the instances of A and X arc also 10stances of
a good many other thmgs. This A is X is really (e.g.) ABC
is X. The important thmg, then, is that 10 the conclusion all
A is X, we have selected one element A out of the complex
ABC rather than another as the subJect of X 10 the universal
statement all A is X The first question therefore is, What
is the ground of our selection~ Now suppose that the examples
of A were all md1stingmshable in quality and the corresponding
premisses therefore each of the form 'this AB is X' ; clearly
a thousand such will be no more use than one for the selection
of A : there would be no reason for preferring A to B. Thus
we arrive at the next step. Our mstances, to effect selection,
must be different in quality. Our premisses must be this AB
is X, this AC is X, this AD is X ; and these incl10e us to select
A as the subject of X. And here we strike upon the essential
principle of the selection and have now to ask for the pre•
suppositions of this rcJection of the elements which differ, and
selection of what is common. It will be noticed that the analysis
now resembles the method of agreement m the new induction ;
 SPECIAL LOGIC
and the truth is that that method is but a particular case of
mufflff'atio simplex, i. e. the application of it to cases of causa•
tion. The modern analysis of this method then is a step, though
an entirely unconscious step, towards the formulation of the
true process of enitmeratio simplex. We shall find, too, that as
this method of agreement is admitted by the inductionists to
depend upon an axiom, so also does the enumeratio simplex,
a result totally unexpected from the inductionist point of View
and even totaHy contradictory to their theory of the experi•
mental methods.
§ 33 I. The premisses of the enumeratio simplex argument being
this A is X, &c., the interest of the conclusion, all A is X, clearly
bes in the fact that 1t is not an enumeration of observed cases,
but goes beyond them. This universality means that A as such
is supposed to necessitate X ; or the quality of A•ness neces•
sitates X-ness. This necessitation has not been observed in any
of the premisses ; for, if it had been, a single instance would
have been enough, as it 1s in geometry, and there would be no
need of an appeal to any further instaures of what is called the
inductive process. The premisses, then, do not give the element
of necessary connexion which makes the conclusion of impor•
tance. This of itself is enough to show that the process as
usually analysed cannot be ultimate : there must be something
behind it. This we may see by putting the matter in another
way. The premisses arc not merely related (as ordinarily sup•
posed) to the conclusion as particular to universal ; for the
particular that corresponds to the real meaning of the conclusion,
which meaning is that A as such necessitates X, must be a
particular in which we see that A in that case necessitates X.
That is the only kind of particular which would correspond to
the given universal : whereas that is not the character of the
statements which constitute the premisses, for if it were we
should be able to draw the conclusion from a single instance.
The premisses, then, are not sufficient and require to be
supplemented, but still it will be said that they are evidence
in favour of the conclusion. Now how are they evidence? The
act of selection, which we have described as the important thing,
shows us what is presupposed and how the particulars can be
regarded as evidence. Before we can undertake such a selection,
 Induction antl the E~j,mfflffltal Methods S8S
which is directed towards finding what element another element
necessitates, we must already have the idea of necessary con•
nexion, as opposed to the mere coexistence of two elements
together at a given time in our consciousness. Thus we argue
that if A conditions X, A and X must be always realized together
in the particular instances of A : that follows from our idea of
necessitating. In the case before us the facts of observation so
far conform to this condition that X is always found with A as
far as we know ; and thus the mere coexistence of A and X
which we observe becomes so much evidence in favour of their
necessary connexion. But this is not enough. We argue further
that any element not always found with A cannot be conditioned
by A ; consequently in AB is X, AC is X, if the analysis is
correct, B cannot be conditioned by A, because it does not
appear m AC is X ; and similarly with C. Thus, as a result
of elimination, there 1s left only X which could be conditioned
by A. We may put this accurately thus· If in a complex ABX
and a complex ACX, A conditions anything, that cannot be
either B or C, and therefore 1f it conditions anything, that must
be either X or a part of X. Now whether we represent these
processes clearly to ourselves or not, they are obviously what
the mind must go through m an argument by enumeratio
simplex, however elementary. We do not proceed absolutely
from the more particular to the more general : the particular
premisses, instead of bemg the startmg-pomts, are only of use
when subsumed under this general premiss • of necessary con•
nexion. The process itself indeed is what the inductive logicians
would be obliged to call deductive, and purely deductive. And,
for the matter of that, enumeratio simplex argument differs from
strict demonstration not m its form but in its matter ; the form
1s accurate, but we are not sure of our analysts of the objects
of experience expressed m the individual premisses. If we were
sure, two instances might be enough to eliminate all but X from
what could be conditioned by A : 'might be', because if we were
aware that the two instances of A, which we have assumed so
far to be identical, may be instances of a species to which other
and different species correspond, we might require further in-
~tances before accepting our universal conclusion.
[" It 1s an axiom rather than a premiss Cf I 330 and note to § ~42.]
 SPECIAL LOGIC
I 332. The number of instances • is too often represented as
if by their mere multitude they contributed to the universality
of the conclusion This 1s the fallacy of supposing that any
finite number, however large, can be an approximation to the
infinite and 1s analogous to that other fallacy of treating
the universal statement as if it represented the totality of the
particulars. We have seen already that the mere number of
instances as a number is of no value, because, if the quality. of
the instances is md1stmguishable, however you mcrease them,
they are of no more use than the first for selecting X as con-
ditioned by A. To be of value they must differ in quality, and
it is because of this that there must be a plurality.
Now if our analysis ABX, ACX, were certain we should
require no further mstances ; but m practice it 1s quite impos•
sible to be sure that our analysis of particular mstances is
complete. We are not sure that X is the only element beside
A remainmg invariable in the two instances ; some we notice
vary, but we are not sure of the extent of the variation. Now
what we do in attachmg importance to a large number of
instances ts to assume that the more instances we take having
A and X in common, the greater the chance that we have
varied all the other elements except A and X. This, then, is
the true and only reason for attaching importance to the number
of instances. Thus m a way their larger quantity makes up for
their defect m quality, or, their quantity becomes a kmd of
substitute for two instances perfect in quality. This also ex-
plains how the ordinary presentation of the enumeratio simplex
argument, i. e merely m the form-this A 1s X, that A is X, &c.,
can be justified. It 1s really because the plurality of instances
tacitly implies the probable variation of all the elements
except A and X. Otherwise the f,ulure to call any attention

[• 'Now the precanousness of the method of simple enumeration 1s 1n an

inverse ratio to the largeness of the generahzabon . . As the sphere widens,
this unscientific method becomes less and less hable to mislead , and the
most universal class of truths, the law of causation, for instance, and the
pnnc1ples of number and of geometry, are duly and satisfactonly proved by
that method alone, nor are they susceptible of any other proof.' Mill, 1d ,
w. 21, § 3 But m m 2, § 1, he say'I, ' It 1s the number and the nature of
the instances, &c ' H. Spencer lays stress on this point; see Mill, id.,
u.7,§2.]
 lffeluction anel the E%f,erimental Methocls 587
to the variation of quality would vitiate the procedure in the
manner already described.
§ 333. In the foregoing inquiry we have used the idea of the
conditioning as opposed to the conditioned. The argument
turns on the assumption that A is the conditioning. As we
have seen, it really comes to this as its result, that if A con•
ditions anything it conditions X. But we cannot convert the
probable statement all A is X and say all X is A, without
some express stipulation about the reciprocity of the conditioning
and the conditioned. Now the ordinary analysis of the enumera•
tio simplex inference gives the result all A is X, without ever
inquiring why the result should not have been all Xis A. Yet
as the premisses give the mere conJunction of A and X in the
same subject, both are on the same footing and the ordinary
analysis 1s helpless before the quest1on why the conclusion should
be all A is X and not conversely. If we do select one form in
preference to the other, it must be on some grounds extraneous
to the premisses as ordinarily presented. In fact we require to
add that, in getting the conclusion all A is X, we have assumed
that A may condition something, and our conclusion is, accur•
ately, only the statement that if A conditions anything, it
(probably) conditions X. This ambiguity m the ordinary method
would certainly appear m the method of agreement, if that were
only an instance of the general form of e1tumeratio simplex with·
out anything added about the relation of A and X But, first,
there is a distmcbon between conditioning and conditioned made
and presupposed m that argument. In the instances taken,
ABC and a.fly are not on the same footing ; ABC are conceived
as prior in time and already conditioning a.fly, the method only
settles which element determines which. The question really
1s1 What does A condition as subsequent to 1tselP (We must
remember it 1s correct to ask, what does a condition as preceding
itself?) Secondly1 the ambiguity does appear to a certain extent
and shows the necessity for considering whether we suppose the
relation of conditioned and cond1tionmg to be reciprocal. For
in Mill's account it appears that an ambiguity 1s caused by his
doctrine of plurality of causes, a difficulty which would not arise
in that form at least (i. e. as he puts 1t) if he had believed the
contradictory of the 'plurality of causes', i. e. the reciprocity of
N
588 SPECIAL LOGIC
conditioning and conditioned.1 We shall as a matter of fact
see that the validity of the method is not really altered, as Mill
supposes, by the assumption of a plurality of causes.
f 334. In the followmg investigation we are not trying to find
an ideal of inductive method but first to understand the true
nature of the methods actually used in science and ordinary
life, methods which arc called the experimental methods and
usually considered to be mductive, and secondly, to discuss the
account given of these methods by the inductive logicians.
The method of difference as described, e g. in Mill, deals with
the relation of antecedents and consequents. We are supposed
to have given us at least two cases m experience, symbolized as
follows:
Antecedents, ABC, BC.
Consequents, aP,,, fJ-y.
In such a process as Mill describes we assume in the argument
first that every antecedent necessarily conditions some con•
sequent and that every consequent is conditioned by some
antecedent; and secondly, that we have been able with some
probability to isolate ABC and a/J'Y from the remainder of the
umverse, so that we have only to look for the effects of ABC
in a/Jr, and the causes of a./J-y in ABC. We assume, lastly, that
every antecedent necessarily cond1t1ons the same consequent,
1. c. the same m kmd whenever it occurs ; we assume, in other
words, that the same cause has the same effect.
From these assumptions we argue that in the second instances
BC causes /Jr; for, ex hypothesi, the effects of BC are to be
found in Pr, and the cause of /Jr in BC. If, then, there were
one effect m /Jr not conditioned by BC, we should have a con•
sequent not conditioned by any antecedent, i. e. an event with•
out a cause : similarly, if there were an element in BC which
had no effect in P,, there would be an event without an effect.
Thus fly 1s the whole effcct of BC, and BC the whole cause of
/Jr. But the same cause has always the same effect : therefore
m the first instances also BC causes P,, ; similarly BC does not
cause a., for 1£ it did 1t must have caused it in the second instances
1 Add to th1S that the argument only enables us to say that if A conditions

anytbmg it JS X or a part of X, not that 1t 1s the whole of X, hence conversion

is impossible.
 lmlumon ani the Exf>erinuntal Methods 589
from which a is absent : therefore if A has an effect it is 11 ;
but every event has an effect : therefore A causes 11.
Now whether we have a right to consider BC the same for this
purpose both when alone and in company with A, or not, that
is certainly the presupposition of the argument and the meaning
of the symbolism. The inductive logic is here describing an
ideal of reasoning, to which we endeavour to approximate, in
the sense that the premisses which cover the particular instances
are assumed to be correct. 1 Now clearly the argument is
demonstrative, and the mduct1omsts m their opposition of
induction to deduction would be obliged to call it deduction.
There is no uncertainty m the form of the argument : the
uncertainty lies only in the matter. The premisses contain
presuppositions comprised m what is called the law of the
uniformity of Nature or the law of umversal causation ; and to
these we must add the assumption of the isolation of instances.
§ 335. The argument presupposes that a given set of occur-
rences can be isolated from everything which is happening at
the same time in the universe. This assumption is clearly of
the last importance practically, and the inductionist ought to
consider how it is that he can come to make it. Inductive logic
such as Mill's can hardly be said to face the difficulty; indeed
such a system as his gives no means whatever for making the
distinction and for such a system the question is insoluble. For
with him the clements related to one another and to be con-
nected as cause and effect are treated as mere antecedent and
consequent, i. e. as events. So far, then, their only bond is
succession m time, which, properly speaking, is no bond at all.
And, if this were all, it would be difficult to understand how
any definite number of events could ever be grouped together
so that we could possibly say, 'this antecedent belongs to this
consequent', inasmuch as all events at a given time, if we think
of them only under the form of time, are equally near to or
remote from every other event, however remote in space ; an
event happening here and now is absolutely contemporaneous
with another event on the other side of the world ; and a process
beginning here and now succeeds as immediately in time to
1 1 e. we are sure of what we observe and are sure tljji Ai4J- isolate
ABC and 11/t-,, •
N2
590 SPECIAL LOGIC
a process now ending on the other side of the world, as it sue•
ceeds to the processes and events which preceded here. As
regards mere time it is quite as near to one set as to the other.
And the difficulty 1s a real one for Mill, when we reflect that,
apart from logic, in such a metaphysic as Mill's, the Real con•
sists solely of events in time. The idea of any permanent
substance or substances (which might serve to group together
or connect events as somehow belonging to 1t) is pronounced
a mere fiction of subJective thought and the result of 'the
association of ideas'. It is true that m practice the difficulty
is concealed because the events are by Mill himself (1. e. in the
concrete examples) referred to subJects of them other than
events. The truth 1s that the description of these methods in
such books as Mill's is accompanied by an irretrievable confusion
as to the meaning of such words as cause, effect, agent, patient,
condition, and the hke.
But, however that may be, when Mill comes to deal with the
definite processes of inference, m these methods, cause and effect
mean m his formulation respectively antecedent event and
sequent event between wluch there is assumed to be a necessary
connexion : for that is what Mill means m spite of what he
says. He purposely avoids 'necessary' and 'necessary con•
nexion '-and with good reason, for these terms are fatal to
him-and speaks only of 'invariable unconditional sequence'.1
We find, what he 1s indeed anxious to tell us, that he does not
mean invariable sequence merely ; and it emerges that by an
odd perversion of language by 'unconditional' he means 'con•
ditioned' sequence. He uses unconditional because he means
that the effect must follow the cause and thus cannot be pre•
vented by any condition. Thus he is committed to Necessity.
§ 336. In the method of agreement we have two instances
symbolized as before, but the second instances contain A in the
antecedent and a m the consequent Thus .
ABC ADE
a./3')1 aa,
The same assumptions as before bemg made, the argument
would be : every cause while remaining identical has the same
effect : therefore the effect of A, A bcmg the same in both sets
1 Mill, 1 C ill 21, § I.
 I ntluction and the Experimental Methods 591
of antecedents, must be found in both sets of consequents.
That which 1s common to both sets is ci ; hence the effect of
A must be found in ci. It may be shown as before that this
argument 1s entirely deductive. Further, notice that we do not
come upon any difficulty so far about the plurality of causes,
which Mill supposes to be the peculiar drawback of this method.
He supposes that smce the event a. might have different causes
at different times, it might be caused the first time by B and
the second time by D, and have notlung to do with A. This
is another example of the weakness m logical analysis which 1s
found m every part of Mill's account of mfcrence. He argues
from the general assumption that a. might have different causes
at different tunes, without not1C1ng what 1s entirely obvious,
that m the particular case before him the data make this quite
impossible for the two times represented by the two sets of
instances. He 1s prevented from scemg this because he happens
to begm with the consequents. 1f he had only started with the
antecedents, he could hardly have failed to sec that a must be
necessarily connected with A. Thus even if the doctrine of
plurality of causes were true, as m the strict sense it is not,
Mill's difficulty is unreal, and 1f we assume that the plurality
of causes 1s true, the only difference 1t would make would be
this, we should have proved that the effect of A must lie in a,
and if a 1s something which admits of no further analysis, we
should have proved that A causes all a.. If, however, we suppose
a. resolvable mto x and y, though x may be necessarily connected
with A, it might be that the remainder y was really caused the
first tune by B, and the second time by B. There must be
a residuum at least, 1dentkal in both cases, caused by A.
§ 337. In the method of concomitant variations we start with
the same presuppositions as before, both as to isolation and as
to the nature of causality. Herc, however, we are supposed to
have instances of an antecedent A, and of another element
associated with it, whether contemporaneous or consequent is
left somewhat obscure m Mill. We observe changes in A, and
we observe changes m a associated with 1t, but no change in
the other elements. We argue, then, that A causes a.:
A1 BC AsBC
41/J-y a.JJ,,
 SPECIAL LOGIC
The ground of the argument is clearly as follows: We assume
that if A conditioned some other element, a change in A would
involve a change in that other element, and consequently those
elements which do not change with A are not connected with
it. Hence everything is chminated except a.. Or otherwise
thus: the identical clements have identical effects, therefore
they have not caused the changmg effects ; and thus the
changing consequents arc caused by the changing antecedents,
Herc again the argument is deductive ; and here again we
observe serious deficiencies of analysis in Mill. Mill ought to
have considered what can be meant by the conception of change
applied to an antecedent if that antecedent is an event. How
are we to distinguish what we call changes or variations u/
something from events considered simply as new events? An
antecedent, as Mill describes it, 1s an event. Now an event,
properly speaking, cannot change , and on the other hand
change cannot be reduced to the mere denomination of event.
Clearly, then, change requires something not itself an event,
something identical m time and m the change, m reference to
which the events can be called changes, or more correctly phases
of change. If we speak of a change m an event itself, this can
only mean difference m the form which the universal takes :
the universal, that is, corresponding to the event ; so that
strictly there would not be a change in an event but different
events of the same kmd but of different species. If again we
assume with Mill that the same effect may have different causes,
the fact that an clement docs not change with the change of
A docs not eliminate it as not havmg causal connexion w1th A.
If the instances arc
A 1BC A8BC
a.1/Jr a.a/Jr
it is only consistent with the plurality of causes to say that the
element /3 which has not varied may have been caused the first
time by A1 and the second time by A1• Further we may ask,
in such instances as Mill gives, how we can d1stmguish which
is cause and which 1s effect, if the variations are concomitant
in the sense of being contemporaneously concomitant, Aa the
pendulum moves towards the mountain the period of its vibra•
tion changes : therefore, he argues, its movement towards the
 I ntlumon aHll the Exf,erimental M Bthods 593
mountain causes the change in vibration. But on such an
analysis as his it would equally follow that it was the change
in the vibration that caused it to move nearer the mountain.
The reason in fact why one of these elements is the condition,
the other the cond1t1oned, must he in something extraneous to
such data or premisses as Mill gives.
The same criticism is fatal to Bacon's Tabula graduum (which
corresponds to concomitant variations), considered as a means
of discovering the 'form'.
§ 338. It would usually be said of modern induction that it
investigates the relation of cause and effect ; i.e. not necessary
connexion in general but a particular form of 1t. It is of course
in the particular subject-matter to which these methods apply
that we must look for their real significance and explanation.
The first characteristic is that the necessity of their connexion
refers usually to time. In pure mathematics, which also deals
with necessary connexion, the connexion is as such not in time
at all. But beside this there 1s another d1stmction which is
really responsible for the difference m the methods employed m
the two cases. In mathematics we understand the necessity of
the connexion, or we know why the one clement conditions the
other. The expression 'understand the necessity' 1s the better,
because the other (' know why') rather suggests something to
mediate the connexion between the two clements and rather
1mphes that the knowing why and how the two arc connected
1s knowing something else wluch connects them, whereas in
geometry we sec the necessary connexion of the two elements
directly and therefore see how one cond1tions the other. When-
ever our knowledge of one element shows its necessary connexion
with another, then we 'understand' the connexion fully. But
there are other kmds of necessary connex1on where our know-
ledge of one element does not show the necessary connexion of
1t with another, and where we have recourse to experiment and
observation ; and our knowledge about the connexion has to
be (in an obvious sense) a mediated mference. Now the general
form of this argumentation 1s as follows. We start with the
assumption that there is a necessary connexion m a group of
elements given in experience, and our problem is to find which
is connected with which. It is just our want of intuition into
 SPECIAL LOGIC
the nature of the connexion between the elements which causes
our argument necessarily to take the threefold form corre•
sponding to the methods of agreement, difference, and con-
comitant variation. The problem accurately stated seems to be
this. Given a complex of elements ABC, and another o.{Jy, and
given that each clement 10 the first 1s necessarily connected
with some element in the second, which is connected with
which? The clements are merely given in experience with S'ome
sort of juxtapos1tlon in time ; and in the experience we do not
see the connexion. If we could, the single experience would be
enough : we should not need the methods ; we should be in
the position of the mathematical sciences. Observe also that
this general form of the problem 1s not restricted to the con•
nexion of cause and effect as this is usually understood. That
relation is not the essential thing, but solely this want of direct
intuition into the connexion. This be10g so, what means have
we of settlmg which elements arc connected with which? We
have seen· (1) that one instance 1s not enough, because we do
not observe the connexion; and (2) that the mstances must
vary in quality.
Now 10 the case of a given clement A we have this exhaustive
d1v1sion:
(1) A present m different kinds of mstanccs.
(11) A absent ,, ,, 11
(m) A changmg ,, 11 11
Consideration of these cases leads to the three following
mam d1stmct1ons of the experimental method.
In the first (pre&encc) the method 1s such as to eliminate what
cannot be cond1t1oned by A, leaving what 1s conditioned as
residue.
In the second (absence) the method is such as, by combining
the presence and absence of A, to eliminate things present when
1t 1s absent.
In the third (concomitant variations) the method 1s such as
to eliminate what does not change with A.
With nothing but experience to help us, this is an exhaustive
principal division of the possible methods : the subordinate
character of the so-called method of residues being obvious.
We must not forget that the analysis given in each method is
 I ntluction antl the Experimental Methods 595
an ideal, because Wct,_assume the elements of what is observed to
be completely known, or at least with sufficient completeness
for our purpose, an ideal which we know can never be realized.
Further, if the ideal conditions were satisfied, one method would
be enough, if applicable, and one instance of it ; and 1t is only
because we cannot be sure of our analysis of facts that we
multiply the mstances and may use all the methods on the
same problem.
We see now what the characteristics of the subject-matter are
which necessitate the use of such a method and that the general
character of the method itself in all its forms 1s ehmination.1
As we start without knowing the nature of the necessary con·
nex1on, so we end without any further insight into it i for the
methods as negative and ehmmat1ve can by themselves alone
lead us only to the conclusion that A cond1t1ons ci because A
can condition nothmg else . they do not tell us how A conditions
ci. Now this has a serious consequence ; for we must admit
that m the empmcal sciences, so far as they depend on those
eliminative methods, we never reach that understanding of the
necessary connexion which we have got m mathematics : we
have the fact but not the reason why. 2 It is true we seem at
first to have advanced from the fact to a reason in so far as we
have advanced from the particular observations of A and u
together to the belief that A necessitates a, which seems the
reason of the particular fact. Still the general prmc1ple to which
we advance from these particular mstanccs is itself but a general
fact, the reason or ground of which we do not understand.
Lastly, our conclusions can never in the nature of the case be
more than highly probable. Now of this kmd of necessary
connexion the causal connexion is a particular species. It is an
example of a connexion as to the nature of which we have, in
many cases at least, no direct intuition.
1 Bacon's per re1ect101111s.

• Anstotle's in-, and &ch-,.

[It 1s just to note here that Wilson in a letter to a former pupil, Mr. D.S.
MacColl (27 ,v1 xi), wrote: • Mill was a great man, if rather confused as
a logician and metaphysman, and the older I get the better I appreciate
auch men as he.']
 II
DEFECTS OF THE MODERN THEORY OF INDUCTION.
THE FORMAL CAUSE IN BACON

§ 339. 1 IT seems, then, that the Eliminative method only dis•

entangles out of a mass of mere conjunctions in time those
which are necessary ; finds, that is, which member of the com•
plex belongs to which other in consequence of their own nature.
But yet it does not advance beyond the fact that the first element
does condition the other, it does not get within and enable us
to see how precisely the nature of the first element conditions
that of the other. We may fairly say we know that the thmg
is so, but we don't understand it fully.
This 1s character1st1c of all the so-called Natural Sciences, but
in one department, that of organic and animated hfc and matter,
we observe a remarkable tendency to explam the parts of an
organism by a teleological idea ; so that there is something more
than the mere knowledge that an element m the organism is so
and so, through the reference of it to some function or end of
the organism. This must be recognized as an attempt, whether
successful or not, to get behmd the mere mechanical fact. The
position of it, however, m actu.il science 1s dubious and probably
vanes with the ind1v1dual thmker. In the reform of modern
science the teleological idea was regarded as a kind of super·
stition and certainly with some amount of truth. Science had
been perverted by the assumption of what ought to be instead
of patient inquiry into what ts, and it was part of Bacon's
service to protest against these teleological guesses. It is
remarkable, however, that b1olog1cal science, starting as it did
with a mere mechanical theory of causation, has come to present
the complex of a plant or ammal as all ordered and arranged
relatively to a plan or function which defines the organism, and
[•§§ 339-45, These sections were ongmally inserted at about this point.
They were dropped out to save tune, but the author was satisfied with his
argument. Cf. §§ 520 and 527.J
 Defects of the Modern Theoty of Induction 59'J
it is- important to note that while many still represent this as,
so to speak, an accident of necessity working without a plan
or end, yet the idea of the plan or design is not merely a sort
of after-thought and without essential connexion with the frame
of the science. On the contrary, it forms often the most valuable
clue to direct the mvestigations of the biological student. We
find in the writings of Darwin himself that the idea of what
ought to be, on the assumption of the preservation of species
and its reproduction, led him to important discoveries as to
what is actually the fact. We may say generally that this
notion is not a mere product of the imagination but an actual
development of these supposedly mechanical sciences. But, as
yet, the students of these sciences do not seem to have made
up their minds as to the rank they shall assign to an idea which
is so important to themselves and to the general student, an
idea which appears at all events to give deeper insight into the
meamng and workmg of Nature than the merely mechanical
view.
§ 340. We have hitherto spoken of necessary connexion in
general m connex10n with the eliminative method ; but, as
a fact, the modern Inductive Logic seems to confine induction
to the discovery of relations of causation, understanding thereby
not the. cond1tJonmg of one element in general by another but
the cond1tronmg of an order m time of events. This is not an
accident apparently, for it would seem from Mill's general intro•
duction to the theory of induction that it is the special merit
of this modern method to mtroduce the notion of cause and
that it is in this way that it has its great advantage over the
ancient enumeratio simplex. So, too, Bacon's methods are pro·
fessedly intended to discover what he calls the formal cause.
However, the formal cause m Bacon 1s not the same kind of
tlung, as far as we can see, as the cause which is presupposed
in the more modern accounts of induction. It must, however,
be obvious that there 1s no need to restrict these inductive or
eliminative methods to this one form of necessary connexion,
and that it might be applied to the discovery of necessary con•
nexions which, as such, have nothing to do with sequence in
Time.
§ 341. We have seen that with Bacon and his followers the
Sg8 SPECIAL LOGIC
conception of cause is the central conception of their so-called
inductive method, and that this method, which we have chosen
rather to call eliminative, consists first in assuming universal
axioms about cause and effect and then applying them in
an inferential or deductive process to the particular facts of
experience.
The conception of cause m Bacon stands by itself and is
peculiar to him in the Inductive philosophy. 1 The cause which
his inductive methods have to discover is called by him the
Form.a or Formal Cause. His utterances about it are such as
to have caused considerable difficulty m mterpretation. He
assumes the distmction of material, formal, efficient, and final
causes from the Anstotehan philosophy without analysis or
definition, though of course it is extremely important for him,
as the inventor (as he thmks) of a new method, to give clear
defimtions of what he exactly means by these terms, and that
more especic1lly because he criticizes earlier systems severely for
usmg abstract notions without any clear idea of their meaning.
He can himself hardly be said to attempt a defimtion of any of
them exec-pt the form.a. And tlus absence of defimtion is the
more deplorable because he makes it one of the chief merits of
his Induction and of his scientific method m general that it
inquires into the formal cause and not mto the other three.
Thus he repeats with approval the Aristotelian maxim, 'to know
truly 1~ to know by the causes ',2 and adds, also with approval, 8
the current distmction of the four causes. The final cause he
reJects here and elsewhere, as tendmg to mic;d1rcct mvesttgation.
The efficient and material causes he thinks superficial and not
able to give the real nature of the thmg 4 and therefore not the
true prmciples of action or production 6 The distmction of the
formal from the material and efficient causes is further brought
out in the next Aphorism. There he speaks of them as bemg
variable 8 and, as it were, only the vehicle of the form. This
vehicle m fact vanes and, we should gather from his words, is
1 Cf § 527. 1 Vere sc1re ease per causas scire, N O u 2.
• Etiam non male constituuntur causae quattuor, N O u. 2
• Sc1entiam veram, 1b • Scientiam activam, 1b.
• flu/llae. ' At qui effic1entem et materialem causam tantummodo novit
(quae causae fluxae sunt et mlul al1ud quam veh!.cula et causae formam
deferentes in abqu1bus) •, N O u ~
 Fo,11U11, Causs in Bacon 599
only important as in some way containing the form or bringing
it about, while the form itself is the invariable essence of the
thing. The knowledge of it, moreover, should, according to this
aphorism, not only give the true knowledge 1 of the thing but
also power over nature, 1 the freest and most unlimited by the
accidents of the material.
In all this it will be observed that Bacon gives no definitions
of these efficient and material causes, but assumes them as
sufficiently understood, although they belong to that part of
the Aristotelian philosophy of which he is a pronounced enemy.
If we follow the usage of the philosophy from which he borrowed
the terms, we should make the forma simply the definition of
the thing, the formal cause of a thmg bemg m Aristotle its
definition, its essence If we do this, a difficulty at once meets
us. How will the knowledge of the essence of a thing (its
definition) brmg that power to produce it which is, according
to Bacon, the very end and aim of the scientific method? That
is, though we may get true knowledge and true contemplation
from the form, how shall we reach the knowledge of principles
of action and the free production? For production we expect
indeed the efficient cause, and are entitled also to expect that,
because he has not defined the efficient and material causes for
his own purposes, but accepted the terms from ordinary
usage. 3
§ 342. The difficulty we here feel is only confirmed when we
examine a particular instance of the application of Bacon's
method. When he gets the form of heat,' it is interesting to
notice whether it is really a precept for action, a formula by
which the thing could be produced, or simply a definition of it,
not containing the means of its production. In the last section
but one of this aphorism, heat or the formal cause of heat is
given merely m the form of a dcfimtion : 'Heat is a motion,
expansive, restrained, and acting m its strife upon the smaller
particles of bodies But the expansion is modified : while
expanding all ways, still 1t inclines somewhat upwards. And the
struggle in the particles ts modified also ; it is not entirely
1 contemplatio vera ' Quare ex formarum inventi.one sequitur contem-
platio vera et operatio hbera,' N O 11 3
• operatio hbera. • Cf note (3) to p 598, supra • N. 0, ii. 20.
6oo SPECIAL LoGrc
sluggish but hurried and with some violence.' 1 When further
we come to the practical precept, we find what seems like trifling
with the mvestigator: 'The direction 1s as follows: if in any
natural body you can excite a dilating or expanding motion,
and if you can repress and turn that motion upon itself in such
a way that the dilatation docs not proceed equably, but holds
its own in part, in part 1s counteracted ; without doubt you
will generate heat ' 1 Thus we are presented with a definibian
of heat and then told, what no one requires a philosopher to
tell hirp, that, 1f you can produce the object so defined, you
will produce the object of the definition.
Turning back to earher parts of the N ovum Organum, we
find that the utterances of Bacon vary between those which
describe the form as a mere definition by genus and difference
and those which seem to relate 1t somehow or another to active
production or causality. His phrases are so vague and general
that we can hardly hope to discover what he means, except by
looking at the examples he works out Some trouble has been
spent m attempting to define the d1stmctton between the form
and the efficient cause, but the difficulty here is that he says
so little about the efficient cause Clearly, then, the most
important thing to do 1s to try to make out how the form differs
from that of which it is the form, viz. the simple nature.
According to Bacon the obJect of our investigation should be
the simple quality, and we have to find out the form of the
given quality or nature. This implies an important distinction
between the two
I 343. The material from which we must form our opinion is,
(a) certain general statements 3 about the relation of the form
to the thing (simple nature), and (b) the actual example of heat,
in the vindemiatio prima. 4
It will be at once obvious that, if the distinction of form and
1 ' Calor 11st motus expanswus, col11bitus et nitens per parles m1nort1s Modi•
ficatur autem expam10 ut 11:tpandt1ndo ,,. ambitum, nonnsl,u tam11n 1nd1nel
vwsus supenora Modlficatur autem et nixus 1lle per partes , ut non sd
omnsno segn1s, sed 1ncitatus et cum ,mpetu nonnul/o • N O 11 20, ad fin
• ' Nam des1gnabo est tabs S1 ,n aliquo corpore natural, pot,ris emlMI
molum ad s, ddatandum aul expand1,ndum : mmqu, mon,m Ila r,pn_.e el
•• II 11erlert ut duatat,o ilia ,ion ,Procedat aequaluer sed ,Parl1m obt1n11at, parl,m
rd""1alur • p,oculdub10 gen11rabss calsdum • 1b 11 20
• N 0. 11. 13 • 1b 11 20
 FortlUII Cause in Bacon 6or
nature is not clear, this will be a fatal difficulty in the practical
application of the Baconian methods. These methods presup•
pose fonns and natures somehow given together, and there is no
note or mark to discriminate between them except the defini•
tions of the1r being. They are not separated in the convenient
way in which, for instance, antecedents and consequents are
separated in Mill's methods. That is, we cannot distinguish
them as occupying different moments of Time. They are given,
in Bacon's methods, as entirely contemporaneous and, from
some of the things he says about the form, it 1s necessary that
they should be. The method therefore will utterly fail; unless
we can tell in the group of facts before us which are forms and
which natures.
According to the Thirteenth Aphorism, the form is absolutely
identical with that of which it is the form · 'the form of the
thing 1s the very thrng itself.' The form only differs from the
thing as the real existence from its appearance ; as the inner
from the outer ; as that which is in reference to the universe
from that which is in reference to man 1 These phrases are,
however, very vague and may mean almost anything. We
must turn therefore to the example in the Twentieth Aphorism.
The opinion we should naturally form from 1t 1s clearly this :
in the case of heat we ~ay distinguish our own feeling from
what in the atomistic theory 1s its external cause, material
particles ancl their movement. And we find that Bacon's defini•
tion of heat 1s a definition of this external cause. We might
therefore conclude that our feeling of heat was what he meant
by the thing or simple nature. Thus his distiqction of form and
nature would be a recogmtlon of the atomistic theory and his
methods would really have for their obJect to discover the
material or external causes for our sensations. One may say
in advance that this is probably the view to which Bacon, in
a confused way, is comnutted ; but we shall endeavour to show
that he was not clearly conscious of it : that he could not even
describe it and that 1t contradicts not only his words but his
general method of representing nature.
To begin with, it is clear that when Bacon represents a body
1 • in orchne ad universum,', the wp/irro• t/'v1111, and • in ordme ad hominem •,

the •,,G,rw .,.p:,, tpiis of Anstotle

~ SPECIAL LOGIC
or individual thing as a complex or combination of simple
natures (e.g. gold is a complex of such simple natures as yellow,
heavy, malleable, &c.), he docs not m the least imply that these
simple natures arc only in our consciousness He not only does
not say this, though 1t would have been of the last importance
to him to say 1t, 1f l11s idea of form and of the scope of physical
investigation were such as we have described, but his language
in general 1mphes the contrary. He clearly thinks that .tl!e
simple natures are 1~ the real 'body' or nature, not merely
something in us.
A further difficulty is this : the sensation of heat is the effect
of the aforesaid external cond1t1ons and 1s different from them j
so much so that, m this atomistic theory, the one is supposed
to exist merely in the human consciousness and the other to be
entirely independent of the human consciousness, to exist indeed
apart from all consc1ousncss.
§ 344, Now we have seen first that Bacon makes the form and
the nature 111 some way 1dent1cal and uses very strong language
to express tlus, 1 and secondly that he expressly denies that the
form of heat 1s the efficient cause. He says · 'not that heat
generates motion or that motion generates heat (though both
are true m ccrtam rases), but that Heat 1tsclf, its essence and
qu1ddity, is Motion and nothing el<ie' ; 2 and a httle further on :
'nor again must the communication of Heat or its transitive
nature, by means of which a body becomes hot when a hot
body 1s applied to 1t, be confounded with the Form of Heat.
For heat is one thing, heating another.' 3 This agrees with his
reiterated statement that the form is the precise definition of
the nature and also with the use of form in the philosophy from
which the term 1s borrowed. This difficulty cannot be removed
by assuming that he meant that the nature should be defined
by its cause, that cause being different from 1t, 4 because of his
express assertion of the 1dcnt1ty of the two.
It would seem as though the only solution possible is to
assume that Bacon, hkc ordinary people, did not recognize that
certain of the attributes, as they are called, of things exist only
1 ' 1ps1SS1ma res ' N O 11 13 • N o 11 20 • 1b
• As Aristotle, for instance, says that the liT1, or demonstrable property
of a tlung, should be defined by a 11..:,.,, wluch lS in a sense duferent from 1t.
 Formal Cause in Bacon 603
in our sensitive consciousness : that he thought, in the ordinary
way, of the heat felt, as in the hot thing ; just as we speak of
a thing as coloured, as if we held the colour to be in the thing
as much as the extension is. But even this is excluded, for he
recognizes expressly that what he calls ' hot to the sense ' is only
an effect m us : 'Hot to the sense is a relative notion and has
reference to man not to the universe; and is correctly defined
as merely the effect of heat on the animal spirit. Moreover, in
itself it is a variable thing, for the same body (as the sense is
predisposed) mduces a perception as well of cold as of hot.' 1
It seems then, that, if the form of heat is {as we know it is
with him) this external cause of our sensation of heat, the
sensation of heat cannot be the nature
What, then, can the nature be~ There seems nothing left as
an answer to this question. Tl1e confusion, as far as words and
formulae go, 1s quite mcuralJle , and 1t 1s made somewhat worse
by another statement about the 'sensibly' hot. 'Now hot to
the sense is the same thmg : only (it must be considered) with
reference, so far as 1t 1s fitted to sense.' 2 These phrases, had
we not the rest of Aphorism 20, would certainly seem to suit
the d1stmct1on, given m Aphorism 13, between the form and the
nature, for correspondmg to 'the same tlung' we have 'the very
thing itself', and to 'with rc-ferenre', we have 'm reference to
man' as distmgmshed from 'm reference to the universe' and
the apparent as d1st111gmshcd from the existent; and, further,
m Aphorism 20, while the 'hot to the sense' is so distmguishcd
from Heat proper, as the mere effect of 1t, 1t has the designation
'm reference to man, not to the universe'.
§ 345 In face of these contradictions, which are so explicit,
it is futile to attempt a reconc1hat1on ; to try to present a definite
conception of form as ex1stmg m his mind Indeed 1t is idle to
suppose that such a confusion of ideas could be of any value
either m philosophy or m sctencc, to which Bacon professes to
supply a new and certam method.
The only thmg that remams 1s to attempt to account for
these confused statements ; we must try, if we thmk the problem
' N O II 20, § 5.
1 'Cahdum autem ad sensum res eadem est sed cum analogia, quahs
competlt sensu1' N O u 20 (concluding words)
2773•2 0
 SPECIAL LOGIC
worth solving at all, to trace the various influences that were
crossing one another in his mind. An indication perhaps of the
true state of things is to be found in his curious statement about
the • hot to the sense' already quoted.1 He distinguishes it
apparently from Heat proper, as merely a sort of result or effect
of it in our sensat1011 i he seems further to distinguish it from
what is properly heat, by the remark that in itself it is a variable
thing, and indeed so far distmguished from true heat that the
same body may mdiffercntly give the sensation of cold or of
heat. ('Moreover, in itself 1t 1s a variable thing,' &c.) He
himself refers to the forty-first instance given in his Third Table, 1
m the wordmg of which we find somethmg curious and sug-
gestl ve This • hot to the sense' is called 'hot as far as regards
the sense and touch of man', and therefore 'a variable and
relative thing'. The mstance of its vanabihty given is that
•tepid water, tf the hand be cold, feels hot, but cold if the hand
be hot'. Now this 'hot to the sense' occurs m the last instance
of the Table as though it had not occurred already, as though
it were m fact something separate from the quality of heat
observed m bodies and described m the forty preceding instances.
This 1s extraordmary because there could be no other test of
the existenre of heat in the cases he has mentioned except the
'human touch' and 'sense'.
Tlus seems enough of itself to prove a certain confu:,ion about
the relation of the quality of heat or simple nature (as first
observed m bodies before we have discovered what he calls the
form) to our sensation. We may therefore provisionally account
for the difficulties we have found by supposmg some such con-
fusion as the following No doubt we ordmanly think of the
heat we feel as though it were in the object. Now Bacon has
no doubt got beyond that stage, especially through his studies
of the Atom1sts ; he recognizes the sensation of heat as in
ourselves, an effect of the obJect, not m the object. But it is
possible, even when this stage is reached, to fall back into the
most natural form of our consciousness and to represent, or
rather thmk of, what ts really in our sensation as if it were in
• See note 2, p 6o3
• • Calid.um, quatenus ad sensum et tactum humannm, res vana est et
respectiva. adeo ut aqua tepid.a, Sl manus fngore occupetur, sentiatur esse
cahda; mn manqs 1ncaluent, fngida • N O u 13, [Nsl. 41.
 De/eds of Bacon's Theory of Induction 605
the thing. Thus we go so far as to think of sound as if it were
10 the sounding instrument. And this is not uncommon in our
ordinary perception when we aren't reflecting on the distinction.
The confusion is removed, of course, by reflection ; but the
truth about Bacon seems to be that his confusion is retained
in his philosophy. For him the simple nature seems to be our
sensation, the 'hot to the sense' confusedly or inaccurately
referred to the object as its attribute. And thus this nature is
sometimes half-identified with our sensation and sometimes
distinguished from it.
Thus, for example, in those first forty Instances, he simply
thinks of the nature as m the thing, without reflecting on its
distinction from our sensation, just as any ordinary person will
do. In the forty-first instance he has become conscious of the
difference between our sensat10n and the objective fact. Again,
when distinguishing the 'hot to the sense' from the nature, 1 he
recogmzes its difference from the obJective fact and presents it
as the effect of heat 2 and not as heat itself. We see why, when
he thmks of the nature as m the thmg, he makes the form its
defimtion. On the other hand, when he 1s specially concerned
to describe the nature itself, it 1s not surprismg that he should
shp into a phrase which belongs to our sensation, the 'hot to
the sense', because the nature is nothing but that 'hot to the
sense' confusedly conceived as if it were m the object.
1 N 0. II 20
• Calo, here, from the context, 1s the simplex natu,a We have three terms,
'form of heat', 'heat' (1 q simple:, natura), and 'hot to the sense' (an
effect of the simple nature)

02
 III
THE AXIOMS OF MODERN INDUCTION
§ 346 A COMMON characteristic of all the current definitions
of Induction sccmc; to be the implication that the conclusion
must be wHIC'r than the premisses. It 1s easy to see, even from
the- adm1ss1on of such an mducttomst as M111, 1 that this is not
true of the so-called Experimental Methods. These imply certain
maJor premisses"' (clauses, so to speak, of the law of the uni-
formity of nature) far wider than the particular conclusions
drawn from tlwm Ly the help of the parllcular experiences used
in the mfcrenccc; It may seem at first sight that an inductive
clement 1s ldt 1n the cnumeratio simplex argument from which
these maJor premisses arc suppoc;cd to be derived. And if we
take the pomt of view of the mduct10nists themselves, they
would have on their own showmg to admit that the only
inductive part of the new methodc; is precisely that supposed
antiquated form of lncluct10n winch the<,e new methods are to
supplant. But \\e muc;t take away even tins last claim. It is
an 11lus1on to clcc;cribc the enumeratio simplex inference as a pro-
cess in wlnrh the conclusion 1c; an advance m generality upon
the premisses \Ve have srcn that such a statement is due to
an J.nalys1s "luch 1s not ultimate, and that ultimate analysis
involvcc; a UlllVl•r-,,11 axiom winch is parallel to the universal
axioms used for the experimental methods ; the latter axioms
being only cases of the former. But, apart from this, we can
show that the Inductive logic ic; not only mvolved m the
contradiction ,1bovc described, but also that 1t 1s involved m
more scriouc; contrad1ct1on by the manner m which the enu-
meratto simplex method 1s made to supplement the new methods.
On this pomt our obJections will not be merely formal m
character. We shall be able to show that the new methods
would have an uncertamty which the mductionists have not
suspected, and that actually they could not, on their own showing,
1 System of Logic, III. xxvn
["' Note a p 585 ]
 Failu,e of Induction 6o7
be so trustworthy as the old method which they profess to
replace.
§ 347. Such inductionists as Mill admit that their induction
depends entirely upon the Axiom of Universal Causation. The
axiom bemg found by enum.eratio simplex, we ask how a process
described 1 as 'loose and uncertam' can prove the validity of
an axiom which 1s the foundation of processes supposed to be
methodical and satisfactory It 1s really clear that it could not
give a satisfactory proof and could not account for our convic-
tion that the necessary connexion of events is universal and
knows no exception. And 1t is a sign of the truth of this that
Mill 1s unable to mamtam this part of his theory without con-
tradiction. If we ask why a process wluch he condemns as
falhble should be admitted to prove tlus most important axiom,
his first reply is that the case 1s exceptional. This enumeratio
simplex can be rehed upon, because of the immense quantity
of the premisses We should naturally suppoc,e therefore that
Mill believed both that the principle of causation was universal1y
true and that this exceptional form of the ancient mductlon was
enough to establish 1t. Indeed he says 2 that 'the law of causa-
tion, for instance, and the principles of number and of geometry,
are duly and satisfa.ctonly proved by that method alone, nor
are they susceptible of any other proof'. Mill must mean by
'duly and sat1sfactonly proved' that the conclus1on, the law of
umversal causation, 1s absolutely true as a matter of fact and
that the proof of 1t 1s as vahd as a proof m geometry. Tlus
agrees so far as 1t relates to the mathematical axioms with what
he had said earlier : 8 'These arc rigorously true, even beyond
any possible experiences of ours.' a And so of the axiom of
causation (Jaw of universal causation) he says, 'we may . . •
1 Mill, 1 C • III XXl, §2 I 1b. § 3 • 1b , II v1 [see note a].

[" Th1s quotation, the reference for which 1s given m the author's typescnpt
and m the various lecture notes as Mill, System of Logic, II v1, I have not been
able to venfy. It was sometimes cited as' even beyond the lnwts of our
possible expenence •. The difficulty is that Mill's doctrmc m th.ls chapter
1s that axioms are dcnved from cxpencnce, and he 1s expressly speaking even
in regard to number of • all ob1ccts known to our expenence ' In III. v, § I,
1t 1s true, he says, ' In the laws of number, then, and m those of space, we
recognize . . the rigorous un1versahty of wluch we are m quest • ; but tlu.s
is consistent with his arguments elsewhere as to their ongm 1n experience 1
6o8 SPECIAL LOGIC
regard the certainty of that great induction as not merely com-
parative but, for all practical purposes, complete' .1 Yet at the
end of the chapter, m the course of which the uncertainty of
the method has been gradually growing upon him, he denies
that we may rely upon the law beyond the possible range of
our experience, c;aying 'In distant parts of the stellar regions,
where the phenomena may be entirely unhke those with which
we are acquainted, 1t would be folly to affirm confidently \hat
this general Ia.w prevails ' 2 Observe the contrast to his earlier
language in the same chapter . 'It is therefore an empirical law
coextensive with all human cxpcnence; at which point the
distinction between empirical laws and laws of nature vanishes.' 8
Clearly, m the later passage, he has become aware of the
untruth of what he said earlier, and rcahzed that all human
experience 1s not therefore coextensive with all nature. Again
in the cac;c of m.i.th<'mat1cs, the axioms of which he says are
duly and hahsfactonly proved by the same proof as the principle
of causation, he has affirmed 4 that we can rely upon them JUSt
in those chstant parts of the st cllar rrg10ns of which we can
have no experience "
Finally, m a note at the end of this same chapter, he rejects
his own doctrine m crit1r1zing Tame. Agamst Taine he says
that he can't sec how m 'interpreting in general language the
testimony of experience, the limitations of the testimony itself
can be cast off' ; though tlus 1s exactly what he himself did in
the carlrnr part of the chapter The contradiction is complete.
The truth seems to be that when Mdl has h1s attention fixed
on the obJcct1vc fart of causation he can't help feeling certain
about 1t, ancl t hcrcforc 1s obliged to think that the only method
wluch his ph1lobopl11c c;ystem allows him for estabhshmg 1t must
be satisfactory. On the other hand, when his attention is fixed
upon the mfcrcnt1al process, he sees that it is unsatisfactory
1 M11l, I c , III 'Uu, § 4, para 2 • 1b , last para. of ch. xxi.
1 1b , § 3, para 2, sub fin
• ' That a straight !me IS the shortest distance between two points, we do
not doubt to be true even m the region of the fixed stars.' 1 c, III. hi,
I 3, para 3

[& Notice, however, Mill's own caveat m the note to this section ]
 Failure of Emf,iricism 6og
and so now, to save his philosophy, he has to throw doubt on
the conclusion of the process.
§ 348. To return to the exact form of the proof of the law of
causation by enumeratio simplex inference. First, what are the
premisses? In the nature of the case they must be of the form
'This A causes this a', 'This B causes this /J', &c. 1 with the
generalizations 'all events have effects'; and similarly, 'all
events have causes'. Now how can we get premisses of this
form? They can't be got by the four methods: for these methods
presuppose the axiom of all induction. They can't be given by
direct observation : for, as Hume contended and as empiricists
must and do admit, we can't observe causal connexion. Indeed
if we could observe A causmg a we should not need the methods :
for the meaning of these methods 1s that they arc to remedy
this defect of observation, and the analysis of the experimental
methods shows that the inductiomsts assume that we can observe
antecedents and consequents alone.
Where, then, do the premisses come from~ In this philosophy
there can be only one answer, from enumeratio simplex. This
is implicitly admitted 1 by Mill, though he seems 1,Lraightway
to forget the admission and its serious consequences. We notice
first, then, that this enumeratzo simplex, so far from being
founded on the facts observed, is founded upon premisses ;
premisses which are themselves not observed facts, but each of
them an inference from another enumeratio simplex. Enumeratio
simplex proof, then, of the law of the uniformity of Nature,
instead of being a peculiarly mfalhble form of enumeratio sim-
plex, is a specially fallible one. For, besides the uncertainty which
belongs to the final proof by enumeratio simplex, there is also
the uncertainty which attaches to each of its numerous apphca·
tions to prove the premisses. There is not one inference of this
nature concerned, but a great number ; and, unlike the ordmary
enumeratio simplex proof, there must be a doubt attaching to
each premiss, since 1t is got by an admittedly uncertain method.
But we see the fallibility to be still greater when we look at the
form of the various appbcations of this mode of proof to prove
the premisses. How are we to prove that A1 causes a1 ? Not
from the observation of A causing a at any time, but only from
• 1 c., III. xXJ, § a.
610 SPECIAL LOGIC
the observation of A followed by a, Thus the conclusion 'A
causes a' involves two processes, each inconclusive : first, the
inconclusive generalization from 'A preceded a on the occasions
observed• to •A is always followed by a' ; and secondly, by
the fallacious introduction of a new conception not contained
in the premisses, 1 e Lhe conception of causality. For whereas
the premisses mm,t uc 'A is followed by a', &c., the conclusion
is 'A causes a'. The falhbdity of the ordmary proof by 1nµ-
meratio simplex, 10stead of bcmg removed or mitigated, is, so
to speak, multrplred by itself.
§ 349. We can now tabulate the whole process:
(1) There is an experience of the type ' A preceded a, as often
as observed '.
(ii) There ts an enumeratio simplex 10fcrcncc from many such
10stances to the conclus1011 • A causes a ' ; and s1mdarly
'B cause,; /3 ', &c, &c
(111) The results of thu, unsr1cnttfic 10ductton give the premisses
for another 10durtron, ah,o an enumeraho simplex infer-
ence, winch has for its conclusion the axiom of the
umfornuty of 11.1ture or the bw of universal causation.
(iv) After tlus come the processes of the four experimental
methods, based upon the results of the second mduct1on ;
thcmi,elvcs, though deductive m form, ~lled 10ductions ;
and these constitute the supposed scientific lnduct10n.
Now these methods yield conclusions of exactly the same
form as those arrived at 10 the second stage by the
unscienllfic Induction. Modern Induction, then, claims
that the conclusions m tlus fourth stage are more trust-
worthy than those 10 the second i,tage. Thus 'C causes
'Y' 111 stage (1v) 1s more trustworthy than 'A causes a'
10 stage (11)
Now obviously the conclusion cannot be more certain than
its premisses ; and if the method is uncerta101 the conclusion
must be less certa10 than its premisses But the 10ference in
the experimental methods 1s not certa10, because we cannot be
certam of the facts of observation. Thus the conclusions of the
new induction are less certain than the axiom of Umformity.
This axiom again cannot be more certain than its premisses and,
the method of 1ts der1vat1on being admittedly not rigorous, must
 Faitu,e of Emp;,ricism 6II
be less certain than its premisses ; but these premisses are just
the results of the unscientific induction. Thus a fortiori the
results of the scientific mduction are less certain than these last
premisses ; are less certain, that is, than the unscientific induc•
bon. Thus the conclusions of the new induction are two degrees
less reliable than those of the old. It follows that the only
rational procedure would be to abandon the methods, go straight
to experience, and use the unscientific induction.
§ 350. The question of the plurality of causes is part of a larger
question, the relation of the cond1t1on, whether temporal or not,
to what it conditions. We tend to assume a certain pnority of
cond1t1on over consequent and to look on the former as active
and the latter as passive. The cond1tton, then, seems to have
a complete hold over what 1t conditions, so that the result must
follow, whereas the result 1s not thought of as havmg the same
hold over the condition which precedes. A fact 1s not conceived
as compelling the existence of 1ts own condition in a particular
case, though appan:ntly it is concc1vcd as necess1tatmg some
condition or other, an adm1ss1on of the last importance. The
view is erroneous and probably clue mainly to tlus idea of
superior activity m the concl1tion which specially suggests itself
when the condition 1s temporal ; for, as such, 1t has apparently
a prerogative and superiority of existence over an as yet non·
existent result. Thus, when the relation is of sequence, il seems
a kind of paradox to represent the not yet existmg result as the
cond1t1on of what 1s ex1stmg There 1s a further corroboration
of the erroneous view in the doctrine of the plurality of causes,
which means that the same effect may at different times be
produced by different causes. Moreover, what is very important,
the view thus suggested by experience gets support from a general
a prion opinion about the rcla.tlon of the cond1tio11 to what 1t
conditions ; an opinion based on the fact that the hypothetical
proposition, as such, cannot be converted. From '1f A, then B'
we cannot infer '1f B, then A' , thus A seems to control the
existence of B but the converse does not seem to be true. Even
in this abstract form the condit1omng clause seems to have some
superior force.
§ 351. We shall try to show first that the relation of the
condition to the cond1tioncd 1s not one-sided ; and that the
61z SPECIAL LOGIC
conditioned has the same hold over the condition as the condi•
tion is usually thought to have over it; and, secondly, that the
illusion of a plurahty of causes is due to imperfect observation.
In such a question we must begin with something admitted
to be necessary. We shaJl therefore start from the fact that,
while it is thought that an effect may have different causes, the
ordmary observer is sure that the same cause cannot .h;we
different cffcctc;, or that the same cond1t1on, as essence say, will
necessitate the same result, say as property. In this lies the
key to our problem.
Suppose A rond1tions a : A docs that in virtue of its own
nature, 1 and the reason why B conditions fJ lies just in the
difference of the nature of B from the nature of A. It is self-
evidt"nt that the definite nature of A must, as definite, condition
not something mdcterminate, but a perfectly definite nature such
as a. We may get the same result negatively. If A conditions
now a, now /J, then there must be a reason for the difference
and that reac;on would constitute a difference between the two
<' u.ci. of A. Thus these two rnses were wrongly supposed to be
identical. Now tlus 1s doubtless the ground for the certainty
we assign to the propos1t1on that the same cause must have the
same effect. But exactly the same argument must necessitate
that a definite R determmes for itself a defimte C. • It is admitted
that a. must have !,0me condition, i.e. its nature necessitates
some cond1t1on 1 but necessity cannot possibly be indeterminate
and thu~ the detcrmmate nature of a. must necessitate a deter-
minate comhtlon. a as identical with itself and different from
fJ determmcs from its own nature a connexion with A, as dis-
tmguishc<l from any other clement such as B: and so with /l.
Again, ncgat1Vl'ly, 1£ this were not so, we should have to suppose
that the nature of a at least determined that it must have one
or other of a certain kmd of (."Ond1tions ; and, if we assume that
it docs not determmc a particular one, there wall be nothmg at
all to decide which shall cond1t10n 1t m a particular instance.
Further, if there were anything which did so decide, that would
constitute a difference between the two cases of a. Or to put
it otherwise ; 1f the given cond1t1ons were all equally indeter-
minate, a could not come about at all. We see in fact that,
I _, ai,,-6.

[• R = result, C = cause.)
 Failure of Efflf>incism
just as the definiteness of the condition necessarily determines
a definite nature in the conditioned, so the definite nature of
the conditioned must determine a definite nature in the con•
ditioning. If the argument connecting the definite thing condi•
tioned with the definite condition is correct, so must the converse
be also, for the reasoning is exactly the same. Consequently,
when the contrary seems true, there must be some difference
in the conditioned or the difference in the condition is only
apparent. The instance of a conditioned by x cannot be the
same as the instance of a conditioned by y ; for otherwise it
would be true that the difference between x and y would have
nothmg to do with the identical part of a, and that which really
had to do with it must be identical in the apparently different
conditions x and y. Thus, either the cond1t1oned is not really
the same, or the cond1t1on was not different in the two cases,
so far as it was a cond1t1on m each. The argument is self·
evident and the difficulty would never have occurred but for
certain appearances which we must now consider.
§ 352. • Consider first the • condition ' in an hypothetical state•
ment. We have already seen that m geometrical science a pro·
perty of a figure is derived from its full essence or definition
and that, 1{ the proof has its completest form, the property can
always be shown to condition the essence. 1 In other words, the
hypothetical proposition in these cases 1s convertible. This is
important because in geometry we see into the necessity of the
connexion of the conditioning and the cond1t1oned. Thus, in
a case where we understand the connexion, there 1s no plurality
but the relation of condition and conditioned 1s reciprocal. It
1s significant that the case which causes our main difficulty
relates to that species of necessity where we do not understand
the nature of the connexion but only establish that there is a.
connexion. So too for the ' cond1t1on ' of an hypothetical state•
ment; we can easdyshowthat this 1s compatible with the doctrine
we have laid down and indeed follows from it, and that the pos•
sibility of conversion depends merely on sufficient knowledge.
Suppose A to be the true condition of C : then C conditions
A, as in the case of a reciprocal relation in geometry. Now
I § 258.
[• ' In re-wntmg this section, the sense of cause as not merely event (as in
Hume and Mill) but as comhtion must be discussed.' MS. not,.]
 SPECIAL LOGIC
suppose that we have A and B connected together in experience
and let AB - %. Then 1t would be true that if x exists, C is
existing ; since x contams A, which conditions C ; but it will
not be true thJ.t, 1£ C exists, x exists. The reason is obvious.
Conly necessitates A and not B, which 1s also in x. This is why
the statement of the connexion 1s not convertible, i.e. because
the protasis of the conditional statement 1s not the truc.apd
accurate condition of the apodosis, but something more This
1s just what happens m a mathematical proof when a proposition
1s not convertible , more than the true condition has been
admitted into the protasis. Thus, when the cond1t1onal pro-
pos1tton expresses m 1ls protas1s the true condition of the
apodosis and no more, the proposition is always convertible. 1
§ 353. It 1s clear that the forcgomg investigation explams
many cases of the apparent plurality of causes : that so far as
tlie&e vanom, causes arc causes of somcthmg 1dent1cal, so far
they them,;elves are 1dent1cal. They differ only m what is
irrelevant to the effect, and 1t 1s the busme~s of science to
discover the 1dcnllty by clumnatmg the irrelevant, just as
mathematics discovers the true cond1t1on by chmmatmg the
irrelevant pa.rt~ of the protac;1s
There arc, however, othrr way'! m which the same false
appearance 1s produced. Suppose the condition of x (state,
process or event) 1s abed, where every clement is necessary to
x. Suppose only a portion of abed 1s realized, say bed : then x
will not Le there. Suppose now a 1s added· then x wdl be
there and 1t will seem as though a was the cause of x. S1mdarly,
1£ aed 1s realized, x appears when b is added, and thus x will
seem to have a or b or c or d as its causes. In fact the condition
was really always the same, 1. e. abed.
There 1s another case wluch 1s more important, because more
illusory. Suppose x 1s destroyed, not ongmated · now suppose
the condition of x 1s abed, each clement bcmg necessary ; then,
1f any one, e. g a, 1s removed, x disappears and similarly with
bed ; thui:, the d1s.ippearance of x may seem to have a plurality
of causes. Now, obviously, v. lule the cause of the existence of
x 1s always one and the same, the cause of its disappearance
1 In tlus section the term hypothetical statement has been employed in

accordance with ordinary usage Dut the more correct term here 1s c01tddt01fal
statement Cf § 293
 Failu,e of Empi,icism 615
is one and the same ; it is always the destruction of abed as
a combination. The completion of abed the cause of .x, by
addmg a to bed or b to aed, &c., is exactly parallel to the process
of destroying the cause by taking away a or b or e or d. It
may seem, however, that, whereas m the first case the addition
of a is not the cause, because a by itself is not sufficient and
x only results because of the existence of the complete complex
of conditions abed, m the second case-the negative case-the
destruction of .x, the removal is rightly accounted the cause,
inasmuch as the removal of a is sufficient to destroy x, quite
irrespectively of the absence or presence of the other conditions
bed But tlus 1s untrue. If the other conditions, bed, were not
all present, x could not be m existence and so the destruction of
x would not take place as a result of the removal or destruction
of a. Thus the removal of a can only result m the destruction of
x, because of a'c; relation to the rcmammg cond1t1ons bed, and
therefore 1s not a sufficient cause m itself, irrespective of their
existence They must be m existence if the removal of a 1s to
result m the removal of x. The destruction of .x, when a is
removed, only takes place on the disc;olution of the relation
between a and bed, Just as the produc-tion of .x, when a is added,
only takes place through the establishment of the relation
between a and bed. Thus, as the true cause of the production
of x is always the organic complex abed and not any element
of the complex, so the dc&truct1on of .x has always for its cause
the destruction of the orgamc complex abed, as an organic
complex. This is the solution of the stock mstance of the
different causes of death. Life depends on the co-operation of
certain elements orgamcally connected, each necessary, none
sufficient. The complex bcmg necessary, the destruction of the
organic connexion causes death and is necessary as the condition
of death. The removal of any one of the clements of the complex
from its connexion with the rest is enough, obviously, to destroy
the organization of the complex. The account which assigns
a plurality of causes to death, then, is a confusion because
there is no consciousness of the distinction between the negative
and positive cases. For the kmd of mistaken consideration, due
to imperfect analysis, which gives rise to the illusion of a plurality
of causes in the negative mstance, the destruction of something,
cannot affect the positive mstance, the production of something.
 rva
FAILURE OF EMPIRICISM TO EXPLAIN
THE LAWS OF THOUGHT

§ 354. THE weakness of the inductive theory in regard to such

notions as cause, substance, &c., gave rise to the doctrine of
the Association of ideas, which dealt more especially with the
question of causation. Hume's mistake, m this doctrine, was
strange, for, according to his own theory, ideas and impressions
are equally facts of consciousness. They differ only in vividness.
If, then, the notion of causation docs not appear in the vivid
series, obviously it cannot appear m the weaker replica of that
series, the ideas Moreover, Hume, contrary to his own criticism
of Locke and his own theory, was here allowing to the mind
a power of originating an idea, which did not arise from any
impression. The self-contradictoriness of the association doc•
trinc, in its variety of forms, has been exposed sufficiently by
many writers.
The empirical theory has made a last attempt to save itself
by taking into account the modern doctrines of evolutionist
biology, and in doing so most naively assumes in the biological
doctrine from which 1t starts the very principle which it proposes
to establish. Further than this, it defeats its own purpose by
endmg in a complete scepticism. That this is so will appear
from an examination of the evolutionist theory of axioms pro-
posed by Herbert Spencer.
§ 355. The inquiry into the origin of what are called axiomatic
truths, or first principles, is, as is well known, one of those
fundamental questions which divide philosophers into opposing
schools.
I propose to consider the contribution made to it by the
Evolutionist school of philosophy.
[• Tins 1s a compOSJte chapter put together for completeness (see Chrono-
logical notes of composition) § 366 is one of the last tlungs the author wrote,
and was wnttan WJth del.Jberate re!enmce to§ 365, which JS very early.]
 F aitu,e of Empiricism
In the discussion which will follow, I use the words 'evolution'
and 'evolutionist' in reference to evolutionist philosophy, and
not to evolutionist science. Why I am anxious to make this
distinction will appear hereafter.1
The Evolutionist claims to have solved the vexed problem,
and to have reconciled all previous schools of philosophy. The
principal representative of this philosophy says : 'Already I have
pointed out that the hypothesis of Evolution "supplies a recon·
ciliation between the experience-hypothesis as commonly inter•
preted and the hypothesis which the transcendentalists oppose
to it " i 2 and here we see how complete the reconciliation is.' 8
This claim is a high one and made, as we see, with some con-
viction. The theory, though distinctly empirical, recognizes
that induction is insufficient for axioms. It is well known, but
its essential features must be recalled.
§ 356. The sufficient and the only guarantee, it is said, of
a universal proposition is that we cannot conceive its contra-
dictory. This is the ground of our belief in axioms i and thus
they are a priori for the ind1v1dual, being due to the constitution
of his mind and not to his experience. But this constitution of
his mind has been itself produced by experience, not his own,
but that of the race.
The effect of experience 1s to modify the nervous structures :
the moddicat1ons have been accumulated by inheritance from
generat10n to generation, so that the mind is at last, in the case
of axioms, completely adapted to the object. Thus, in the
well-known formula, axiomatic beliefs are said to be 'a priori
for the individual, but a posteriori for that entire series of
individuals of which he forms the last term•.«
I will read a passage II which describes how all this happens:
'Little more need be said to make it clear how certain primary
space-relations are presented to consciousness under the form
of necessary relations. If a segment of a circle be looked at,
the image of it cast on the retina is necessarily such that the
arc covers a greater number of retinal elements than the chord ;
and since each of these retinal elements yields its separate
1 § 364. 1 Spencer, Ths Pri,u;iples of Psychology•, vol i, § 208.

• ib., vol 11, § 3J!Z. • Spencer, ib., vol. 11, p. 195 (§ 332 fin ).
1 Spencer, ib , § 332.
618 SPECIAL LOGIC
impression to consciousness, the series of impressions produced
by the arc is felt as larger than the series produced by the
chord. This continues to hold however much the arc is flattened :
so long as 1t has any perceptible curvature at all, it is felt to
be longer than the chord umting its extremes. Parallel experi-
ences are derived from the ocular muscles. Carrying the eye
along the line of the curve yields to consciousness a greater
quantity of sensation than carrymg the eye along the chord
does. As the curve is flattened this difference dimmishes ; but
some of 1t contmues as long as the curve continues appreciable.
Thus the truth that a straight hne 1s the shortest hne between
two points, hes latent m the structures of the eyes and the
nervous centres which receive and co-ordmate visual impressions.
We cannot think otherwise because, during that adjustment
between the organism and the environment which evolution has
establtshed, the mner relations have been so moulded upon the
outer relations that they cannot by any effort be made not to
fit them. Just in the same way that an infant's hand, con•
structed so as to grasp by bendmg th<> fingers mward, implies
ancestral hands which have thus grasped, and 1mphes objects
m the environment to be thus grasped by this mfantme hand
when 1t 1s developed , so the various structures fittmg the mfant
for apprehensions of space-relations, imply such apprehensions
in the past by its ancestors and m the future by itself. And
just aS 1t has become 11nposs1ble for the hand to grasp by
bending the fingers outwards instead of mwarcls ; so has it
become impossible for those nervous actions by wluch we appre-
hend primary space-relations to be reversed so as to enable us
to thmk of these relations otherwise than we do.'
This theory, it may be remarked, does not necessarily imply
that the mmd 1s identical with physical organs, but 1t does
imply that the mind not only perceives by them, but can only
think and 1magme by their means, and 1s therefore restricted
by the functions of which they are capable. Divested of
technicalities unnecessary for our present purpose, it comes to
this. In its early experiences the mind was not obliged to think
the arc greater than its chord. But every arc presented to the
experience of every generation has been greater than its chord :
and the constant perception of this relation gradually modified
 Failure of Empiricism 619
the perceiving and thinking organs, so that it became more and
more difficult to think arc and chord in any other relation. The
difficulty transmitted by inheritance became stronger by accu•
mulat1on in successive generations, till it became altogether
impossible to think the contradictory of the axiom. This is
how the individual's bchcf m the axiom is formed, and the
process secures that the subjective behef shall correspond to an
ob1ective fact.
This has an appearance of clearness and simplicity which
recommends it to many.
§ 357. Before asking if it is true, we may ask whether we
should hke 1t to be true. Such a qucsllon may seem unscientific,
but in philosophy at least we should clear our reasonings a good
deal 1£ we confessed to ourselves what we wanted to believe and
what we hoped was untrue.
Now at first sight the theory may well have attractions.
Many will feel that inducllon does not account for the con-
v1ct10n they have had, as far back as they can remember, in
the same strength about axioms · and yet they will feel on the
other hand the force of Mill's obJection-' I must protest', he
says, 'agamst adducing, as evidence of the truth of a fact in
external nature, the d1spos1t10n, however strong or however
general, of the human mmd to believe it.'
The evolutionist view seems to brmg a welcome help and to
show how the desired correspondence between subjective belief
and ob1ecttve fact has been brought about by the operation of
known causes.
No doubt, then, many may think it would at least be a good
thing if it were true. The evolutionist philosopher probably
believes that 1t is a beneficent result of the great law of Evolu-
t10n ; and there may be scientific evolutionists who feel at least
a prejudice in its favour. But let us look a httlc nearer to see
whether we should have cause to congratulate ourselves 1f the
evolutionist philosopher were nght.
The theory that the mmd has become modified by its environ•
ment, and that 1ts functions have become fixed, implies that at
first 1t was not modified and its functions not thus fixed. If
'Just as it has become impossible for the hand to grasp by
bending the fingers outwards instead of inwards ; so has it
!a773 !ii p
 620 SPECIAL LOGIC
become impossible for those nervous actions by which we appre•
hend primary space-relations to be reversed so as to enable us
to think of these relations otherwise than we do', then it follows
that once it was poS!tblc that these nervous actions should be
reversed. And what does this mean? It means that the mind
and its organs were so constituted that they were capable of
thinking the contradictory of axioms, for instance, of conceiving
a chord equal to or greater than its arc, or two straight• lines
which enclosed a space, only that tlus capacity was never
developed.
To some this will appear a sufficient reductio ad absurdum,
but I do not propose to take that hne of argument. Let us
rather imagine that such a cond1t1on of the mind is possible.
It will follow that our minds have been deprived of half their
powers beyond recovery. The geometrician, for instance, is cut
off from a field of thought as large as the present science-a field
in which among many other things he would have developed
the properties of those straight Imes which are not the shortest
distance between any two pomts upon themselves.
In fact, in respect of all that is axiomatic, that 1s, of all that
appears self-evident, the whole human race is in the position of
that part of it wluch has lost or never had the use of a particular
sense-the deaf, the hlmd, or the colour-hhnd, or those who
have no ear for music. We are hke the Proteus which m dark
caves has gradually lost the eyes which would be useless to it:
except that we are worse off, for though an obJect for our lost
faculties may never come into our experience, we should
obviously have plenty of employment for them. The same
process has woven a falsehood mto our nervous tissues and so
into our minds: for the confidence we have in axioms is made
to come from our conviction that their contradictories are not
even thinkable, and yet it is shown at the same time that this
is an utter mistake, inasmuch as thought in itself, though not
as developed m us, is capable of thinking these contradictories.
The individual, then, at best has got a true belief at the expense
of entertaining a false behef, and at the expense of the loss of
half his mental powers.
Is this a result to be contemplated with satisfaction?
May one not look for sympathy 1f one confesses to the hope
 Failure of Empi,ricism 62:I
that it is not true? Surely it would be one of those things for
which we could not honestly say we were thankful. At best it
would be a matter for resignation and not for gratitude. We
may bow to it as an inevitable decree of nature ; we deceive
ourselves if we say we can see it is good and desirable. The
pessimist would find another instance of Nature's unkindness.
Even where she seems to confer a benefit, she has done us a deep
injury. She has mutilated our mmds with a refinement of
cruelty. For instead of mercifully concealing a loss which could
only be deplored and never repaired, she has evolved the philo-
sophy which has betrayed the secret.
And is there really any compensation?
If we have gamed a behcf which happens to be true, can we
be said to understand it if we ground it on a falsehood. Can
we be said to 'know' at all in the proper sense of the word ?
Can a belief be even mtclbgent which is simply caused by the
want of a power to thmk otherwise, a power which might have
been developed and 1s lost to us by a process as unconscious
as that which deprives us of taste or smell when we have a cold?
In ordinary bfe, when we are not philosopluzmg, we have a great
contempt for beliefs which are merely the result of imperfect
mental development ; we expect them among savage tribes, the
prejudiced, and the uneducated.
§ 358. An attempt has been made to remove a bias in favour
of the evolution theory. If the attempt were successful, it must
be admitted that 1t would create a bias against the theory : and
there is all the more reason to try to take a purely scientific
attitude and ask whether we have to believe whether we hke
it or not.
In the first place attention must be expressly called to the
fact that this is a theory of knowledge. It does not explain
merely how we come by a belief, but how that belief is true ;
that is, how we have knowledge.
I will read a passage to make this clear : 1 ' Hence the incon-
ceivableness of its negation is that which shows a cognition to
possess the highest rank-is the criterion by which its unsur-
passable validity is known. If the negation of a cognition is
conceivable, the discovery of this amounts to the discovery that
1 Spencer, op cit., § 426
P2
 SPECIAL LOGIC
we may or may not accept it. If its negation is inconceivable,
the discovery of this 1s the discovery that we are obliged to
accept it. And a cogmtion, which we arc thus obliged to accept,
is one which we class as having the highest possible certainty.
To assert the mconce1vableness of its negation is at the same
time to assert the psychological necessity we are under of
tlunking it, and to give our logical justification for holding it
to be unquestionable.' '
Other passages to the same effect might be quoted.
§ 359. I begin with an object10n, relating to a matter already
referred to, which perhaps concerns the form of the theory
rather than its essence. The evolutionist docs not seem suffi-
ciently to notice that, while seeming to make mconceivab1hty
an ultimate test or criter10n, he 1s really throwing entire d1scred1t
on the mind's power to conceive or not concmvc as such '
Men do indeed believe firmly when they tlunk the contra•
dictory of their belief mconcc1vable; not as bemg merely
inconceivable for them, but as berng mconrc1vablc for thought
.is SU(.h. For the moment a man is convmced th.it some one
else can conceive what he cannot, he ceases to suppose that his
own state of nund can determmc the truth m the particular
case. Thus a man v,1thout a mm,1cal car nevertheless may
believe th.i.t there 1s surh a tlung as a musical order. But the
supposed lustory of the mmd's mod1ficat10n 1mphes that, though
we are not able to conceive the contradictory of an axiom, we might
have been able, and our ancestors perhaps were able . and thus
the mmd's limitation m respect of the conceivable becomes as
such entirely md1ffcrcnt.
Any reliance placed on the inconceivableness of the contra-
dictory as a cntenon must he, and mdeed according to this
theory docs he, m the way in wluch tt has been produced, and
this 1s the essential thmg to be considered The criterion has
its value because it is the effect, as supposed, of an experience
infimtcly greater than that of one md1v1dual: and thus when
Mill obJectcd that 1£ the test got its value from expenence, we
had better appeal directly to experience, 1 it was replied that
the large experience represented by the test was inaccessible
to the md1v1dual. It is easy to see that the maximum logical
1 Mill, System of Logic, II vu, § 2.
 Failu,e of Empi,icism
worth of the test would be that of an induction drawn from all
the instances which have occurred in the experience of the race,
supposing these could be presented to one mmd which would
draw the inference. The induction has registered its conclusion
automatically m the physical organism, and, through that, in
the mind. The race has come to believe that all A must be B,
because the constant experience through its history of instances
of A which were B has so modified the organs that they cannot
be used for thmking A except as B.
§ 36o. It has been madvertently assumed that in axioms such
a belief has been produced m one way only. From the very
assumptions about biological processes which tlus theory makes,
it is clear that the sufficient and necessary cond1t10n that the
belief 'all A must be B ' should be produced, is not that every
A should be B m the nature of thmgs, but only that those
mstanccs of A which have acted on the organs of perception
should have been B. And thus the conviction that A must be
B could be perfectly well produced m the automatic way sup•
posed, in a case where A was not necessarily B. The limitation
of our experience to a certam species of A which would be
necessary for this, would correspond to known facts. For
mstance, generation after generation of a particular race had
never seen swans which were not white. Countless generations
hving m certam parts of the earth have never seen water in
a sohd form.
Thus from the very manner of its formation our behef that
the arc must be greater than its chord is compatible with the
existence of arcs which arc not greater than the1r chords. There
are ways enough in which our hmitat10n to one kmd of arc may
have come about. For instance, naturalists expect to find new
forms of flora and fauna m regions which are being explored for
the first time. How do we know-not to speak of what may
be found m other planets and stars-that the polar regions have
not an abundance of that other kind of arc and chord? And
was it not as important scientifically that the naturahsts in the
Challenger should have dredged the deep seas for them as for
those objects in which they were immediately interested?
Or agam, allowing, what could not be known if the theory
we are examining were true, that all the arcs in the world now
 SPECIAL LOGIC
are greater than their chords, how can the evolutionist know
that in early geological periods there were even any arcs at all
which were greater than their chords? For aught he can say
the arcs not greater than their chords may have been con-
temporary with the 'dragons of the prime' and have perished
with them. Why should he expect us to allow arbitrary assump-
tions here, which are allowed m no other empirical science?
It is quite remarkable to find the evolutionist confidently
makmg such statements as the following, without even askmg
himself how he could be entitled to them. 'Space-relations have
been the same not only for all ancestral men, all ancestral
primates, all ancestral orders of mammaba, but for all simpler
orders of creatures.'
But there 1s another and more serious form of the difficulty.
It is not even necessary that the experience of the race should
have been uniform. Arcs may have been seen equal to or less
than their chords, but 1f they only came seldom, or if in the
course of time their number much diminished, like that of an
expirmg race of animals, then according to the biological
prmc1ples presupposed, the infinitely greater accumula.,t:ion of
contrary experiences would m time wipe away all trace of them
from our organism.
Worse than th1s-owmg to the way m which Nature has fixed
our functions, we could not perceive arcs which were not greater
than their chords, even if there were plenty of them about us.
It is therefore useless to dredge the deep seas or to go to the
Poles, and mdeed we do not know whether in our own parts of
the world these arcs have not begun to exist. We could no
more sec them than a blmd man could. And thus the evolu-
tionist, professmg to have established the validity of the test of
inconceivableness, l1as unw1ttmgly shown it to be conSJstent
with the contradictory of what he supposes it to guarantee
absolutely. It follows 1rres1st1bly that all that seems to us
simplest, clearest, most self-evident, and more certain, according
to the evolutionist himself, than anything else we can believe,
may be an illusion.
The theory which was to hate reconciled great philosophies
has destroyed itself, and has ended m a scepticism which has
not even the merit of being self-conscious.
 Failure of Empiricism
f 361. The contradiction cannot be avoided by dogmatically
affirming that the experience which has fixed our functions does,
as a matter of fact, correspond to a universal truth which has
no exceptions : for this involves, according as we look at it,
a new mconsistency or an argument in a circle.
(1} We are told that the ultimate and only criterion for the
individual who forms the last term of the series is the test of
inconceivableness, and yet on the other hand that the criterion
is trustworthy because an invariable experience has produced
it. To vindicate this position the cvolutlomst must know the
latter proposition to be true. But 1£ he does he has a higher
knowledge than the critenon itself, because it is knowledge from
which the criterion derives its value, and thus the criterion is
not the ultimate criterion for him. The evolutionist has fallen
into this contradiction apparently because he has forgotten that
he himself 1s one of the last terms of the series.
(1i) Or the difficulty may be put thus. The inconceivableness
of the contradictory of an axiom has been produced, it is said,
by a constant experience in the race, which again corresponds
to a universal truth without exception. How does the evolu-
tionist know that the experience of the race has been such, for
he cannot have had it ; and a fortiori how can he know the
umversahty of the corresponding fact m nature? Apparently
from what has been quoted he thinks he knows. Bemg the last
term of the series he can only know the fact by the incon-
ceivability which he supposes 1t has produced m himself : and
thus the invariable experience and the corresponding fact give
the criterion its value, and they themselves -are only known by
the criterion : wh1ch 1s arguing m a circle.
This may be shortly put by saying that the evolutionist
philosopher has cut himself off from the possibility of giving
the necessary evidence m favour of his own theory.
But it may be answered that though satisfactory evidence
for the theory cannot be given, it 1s a poss1ble hypothesis, and
though it may have to surrender all cla1ms to establish the
validity of our beliefs, 1t may be true, for constant experience
might produce such an effect on our organism. This of course
would be giving up a great part of the evolutionist position.
In the first place, it must be answered that according to
 SPECIAL LOGIC
biological laws the loss of a function or of an organ may not
only be caused by want of a use for them, but also by the
operation of various processes, among which, for instance, are
diseases, and how can 1t be known that some such process has
not destroyed our capacity for thinking the contradictory of
axioms?
But the most important difficulty is this. The obJector can
hardly have reflected on the real consequence of his hypotl:ie~is.
It would throw a doubt, as we have seen, on every principle
however simple, self-evident, and certam it may appear, and
therefore it would leave nothmg even to found itself upon. It
would throw doubt also upon experience, which is supposed to
be our most direct access to reality, for owing to the hm1tation
of the nervous structures to one kmd of function, our mmds
may be m1sreprescntmg the obJect.
§ 362. How deep this unconsc10us scepticism has gone will
appear still better from another aspect of the theory.
• For logical mtuit1ons ', 1t is said, • there 1s no warrant assign·
able other than that assignable for all mtmtions accepted as
certam . namely, the imposs1bihty of thmkmg the opposite.
Unless 1t be alleged that the consc10usness of logical necessity
has a different ongm and a higher ongm, it must be admitted
that the consciousness of logical necessity 1s JUSt as much
a product of past experiences as 1s every other consciousness
of necessity.'
It follows, then, from what we have seen, that these logical
intuitions must share the uncertamty of axioms. But among
them we find included the prmc1ple of the syllog1sm, the law
of excluded middle, the principle of contrad1ct1on. They are in
fact those simple 'laws' or forms of thought to which thought
must conform to be thought at all. Thought therefore cannot
throw any doubt on them without committing smc1de. As
Aristotle has said .
6 IJ' a11a&p6111 T'GIJTr/11 ,-,}11 'lr&a'r'&II ol, 'trUIIV ,r&crro-r,pa lp,i.•
But there is one short general criticism which seems enough,
and would have explained beforehand without such an examina-
tion of details that the evolut10nist theory must end in self-
[• Now he who destroys tlus behef cannot expect lus own words to be
believed Eth Nie u7311 I ]
 F ailu,e of Empiricism
contradiction. There is an elementary prmciple on which we
should expect all philosophies to agree, which is that thought
cannot question the vahd1ty of its own presuppositions or even
try to establish them without sclf-contrad1ction. Now 1t is
evident that the evolutionist theory violates this prmciple,
whether we consider the account given in it of the supposed
cnter1on of all truth, mconce1vableness of the opposite, or the
account of those primary laws of thought which have just been
spoken of.
§ 363. The mistake is of the more elementary character when
the attempt to establish the laws of thought, or a general
cnterion, is made by help of the reasonings of a special science,
for that of course must presuppose the general laws and the
criterion. But this 1s precisely the use here made of biology.
The reasonings of biology would collapse 1f it did not assume,
for instance, the principle of contradiction ; it 1s futile therefore
to prove its objective vabd1ty by biology.
As to the criterion itself, which 1s to show that a cognition
is of the highest rank (its 'unsurpassable vahd1ty', &c.), the
principles of biology cannot depend on 1t since biology has to
establish it. Hence, either they have not the highest rank,
which involves the absurdity that the criterion of cogmtions of
the highest rank 1s shown to be a valid criterion by appealing
to cogmt1ons which arc not of the highest rank : or else the
prmc1plcs of biology have the highest rank, and then 1t turns
out that there are some absolute truths which are not derived
from the criterion, and, what is still worse, that evolution has
altogether forgotten to explain their origin, thus leaving the old
problem so far from bcmg solved that 1t is not even attempted.
For instance, among these principles so entirely unaccounted for
is the axiom of umversal causat10n . for of course b10Jogy pre·
supposes this. It 1s true that in one place where Mill 1s criticized
1t seems as if the belief m the uniformity of nature (on which
biology depends) would be accounted for by the same biological
argument as the other axioms. If s01 confusion would be
confounded.
§ 364. May I say, in conclusion, that I venture to think there
is all the difference between evolution m science, and what 1s
known as the evolutionist philosophy ? It seems reasonable to
 628 SPECIAL LOGIC
believe that there are many scientific evolutionists who will not
be led by a mistaken idea of the solidarity of the subject to
suppose themselves committed to the conclusions of the evolu-
tionist philosophers. 1 It is not the first time in history that
enthusiasts for a new and successful development of science have
tried to carry the science beyond its proper limits, and have
aspired to make it universal by making it what science never
can be-that is, philosophy. ' ·
§ 365. The futility of the inductive theory was seen by
Descartes long before the development of English empiricism,
and with his name and that of Leibniz is connected the theory
of Innate Ideas. This theory has often been misrepresented,
but the gist of it is that the truths m question are the mind's
own possession, developed somehow by its own activity and not
denved from experience--m the narrow sense-of the particular
facts.
The well-known criticisms of Locke on this theory are with
scarcely an exception an ignoratio elenchi. The answer to him
is obvious : Innate Ideas are intended to account for knowledge
where 1t is, not where it is not. It is as useful to allege against
it that plants have not these ideas as that children, idiots, and
savages have not. Locke was also under an illusion about the
presence of an idea which is implicit and not explicit. The time
at which an innate idea appears and the extent to which it is
defined are entirely irrelevant to the real issue. Locke was
answered by Leibniz m his Nouveaux Essais.
The real difficulty for the Cartesian school was to account
from their point of view for the validity of these beliefs which
the mind has on its own account. Why should this necessity
for thought be true of the objects of thinkmg? The theory of
the inconceivability of the opposite as the criterion of truth is
closely albed to that of Innate Ideas and may fairly be called
only a particular form of it. We may say that it is a test of
truth in this sense that, when the contradictory of a statement
1 The speculations of WelSSmann have made a diffenmce in the behefa of

evolutionists about the poss1b1hty of anhentmg expenence ; but neither

bwlogy nor any other special SC1ence c.:au. deal with the question of &XJ.OIIIII
or of the cntenon of truth-irrpl Iii • dpx@. Aa-,or ~ 6f,rnio,, .,.i .,,.,,.i.,.,_
; '1ff'Jld"'P'I'• [Of the first principles of his science we must not teq1lll'9 an
account from the man of science, gt,4 man of sc::umc:e. .f•. Po. 17b 5.)
 Failu,e' of Empiricism 629
is inconceivable, the statement must be true. No exception can
be found to this rule ; such exceptions as are alleged by Mill
and others are mere confusions depending upon ambiguity in
the members of the statement. For example, when it seems as
if an assertion that was inconceivable has become conceivable,
there are really two assertions, not one, concerned. These two
are confusedly represented as one through an ambiguity in the
logical subject, which appears to be the same in both statements
but is not. The reason why inconceivability cannot be an
ultimate criterion (in the limited sense in which we can allow
such an expression) is that we cannot begin with negation j what
a thing cannot be we can learn only from our apprehension of
what it must be, that is, from itself as apprehended. 1
The true difficulty with which the Cartcsians were faced
depends in the last resort upon a representation of the mind
as distinct from the object, and it is one of Kant's merits that
he seems to have been the first to realize clearly the nature of
the difficulty, even though his own solution may not be satls•
factory. If, he says in substance, instead of the mind obeying
the object, the object has to obey the mind, we could under·
stand at once our possession of universally vabd truth. This
is Kant's Copernican idea and is obvious enough when once
stated. The important addition 1s that he believes that in no
other way could we have universal knowledge such as appears
in the demonstrative sciences. The required correspondence of
object and subject is secured by the simple fact that the object
is the object of a subject and the two are in inseparable relation.
However we try to represent the obJect as independent, to be
the object of experience 1t must obey entirely the law of being
experienced, which is only the other side of the law of the
experiencing mmd. Thus if we present the mathematician with
a particular case m experience of two Imes which we affirm to
be straight and which yet enclose a space, he will without
hesitation pronounce that they are not straight, thus legislating
(in Kant's words) without hesitation for the object.
The distinction then, for Kant, of subjective and objective is
not a distinction of thought from something alien to it ; the
objective is within thought. Nor has the time-order of our
• See the remarks about 1mmechate inference,§§ 216-17
 SPECIAL LOGIC
thoughts anything to do with their content or meaning. The
laws of this order (memory, association of 1deas, &c.) arc not
laws of what our thoughts mean, nor docs the whole of what
we know stand m any time relation to anything. We do not,
for instance, suppose that the truth about tlungs did not exist
till we discovered 1t.
§ 366. This pomt of view must, however, be supplemented by
what it has been the object of the preceding lectures to develop.
Kant's standpoint is whJ.t follows 1f we start from the subjective
suJc of knowledge only. Even startmg thus, we cannot represent
reality as other than what 'agrees' with the necessity of thinking
and knowmg, but the standpoint is md1rect and incomplete.
The true account of the necessity of thought explams why there
must be the 'agreement', and this shortly is that the necessity
of thinking 1s the necessity of apprehension. Necessity of appre-
hension, then, turns out to be umntelhg1ble save as apprehension
of a necessity, that 1s, apprehension of objective necessity or
apprehension of what reality must be. Thus 1t 1s plain that
agreement 1s not a proper term to employ and that the question
why reality must correspond to the necessity of thmking only
arises by a false severance of the apprehension from what is
apprehended. Out of this false severance comes all our difficulty
as to how the two should agree or correspond. The fact is that
there 1s no necessity m the thought or apprehension save the
apprehenswn of an objective necessity. When once this is
appreciated the question of agreement or correspondence, which
will still arise even when we have satisfied ourselves by the
indirect method of the above (apparent) adaptation of Kantian
pr111c1ples that we cannot mtclhg1bly represent them as not
corresponding, disappears. It still arises because we are, so to
say, only negatively forced mto the conclusion and don't see
the positive reason. We sec that 1t must be a fact but not quite
how it must be or the reason of it. In the hght of the positive
development the question disappears as altogether unreal. There
has never been any severance and so there is no need to
explain the correspondence The necessity of thought (appre-
hension) can only be the necessity of thinkmg, as the thinking or
apprehending of the necessity m the object with which we have
been uselessly trymg to reconcile 1t.
 THE METHOD OF PHYSICS AND ITS RELATION
TO PURE MATHEMATICS

§ 367. THE sciences called applied mathematics are sometimes

represented as having an mtermcdiate position between the
inductive and deductive sciences. They arc not pure ma.the·
matics but yet mathematics is necessary to them. This kind
of position for them is in a way recognized by Aristotle. He
distinguishes (a) the knowledge of the fact, e g the account of
the observations of the nature of the rambow; (b) optical
science which is, as he supposes, the explanation of the rainbow
by mathematics. This is for him the science of the fact and
the reason togcther. 1 Pure mathematics, in this case geometry,
he regards as (c) the science of the reason alone. This last 1s
a strange phrase. 'The reason without the fact' means that
the reason is found m mathematics but that the fact does not
appear there. He speaks as if mathematics supplied the com·
plcte reason. But, if it did, how could the fact fail to appear
as a deduction in mathematics? Aristotle has never asked him-
self this question.
The distinctions he makes arc untenable. Mathematical

(& Tlus chapter was unt.tl 1899 followed by sections, the mam purport of
which appears elsewhere in the book Of the rest I am not &urc how far the
wnter would have mamta.ined their tenets m later years Their contents were:
§ The conception of force m physics Further development of the relation
of pure mathematics to physic& We do not understand what Force means
§ On the axioms of physics Some great generalizations (e g. the mde•
structlb1hty of matter) are self-evident and a pnon, the experiments employed
bcmg suggested by our conv1ct10n of their truth
§ On the obJectiv-e v-alld1ty of our thought. The problem onginates in
difficulties felt about the axioms of formal logic [Dictum de omni, &c ),
§ The 1nduction1st theory of the ongm of axioms , Mill's theory of parallels ;
association of ideas explanation , H. Spencer's theory of inherited expenence
§ The Cartesian answer, The mconcotvab1Jity of the opposite as a cntenon
of truth Kant's solution (the last accepted broadly 1n 1899] J
 SPECIAL LOGIC
physics depends ultimately on physical laws, which pure mathe-
matics cannot give. Thus in optics, as Aristotle knew them,
there were the laws of reflection and refraction, and it is odd
he did not see that in mathematical optics, as he knew them,
these were accepted as facts given by observation, together with
certain geometrical relations which they involved. Thus all
geometry did was to deduce certain geometrical consequences
from the geometrical part of the given premisses. •
§ 368. As may be seen in the instance of optics, mathematical
physics requires physical laws, which mathematics cannot supply.
About this we find in modern times a confusion like Aristotle's,
though not quite the same. The physical laws themselves
Involve certain mathematical relations both geometrical and of
pure quantity, and all mathematics can do is to develop the
consequences of these mathematical data. How, for example,
does mathematics deal with motion? Certain laws of movement
are given, i e. laws of the way in which a body changes its
distance from certain fixed points. Given such changes of geo-
metrical relations, geometry simply ralculates certain other
geometrical relations which follow from them. But it can say
nothmg whatever of the reason for the given laws of the changes.
To put it otherwise, if we think of force as the cause of move•
ment, it is with the effects of force, as changes in the geometrical
relations of bodies, that mathematics alone can deal. Given,
for instance, the force of attraction and given its operation or
effect, in the sense that a body revolving round the sun tends
to shorten its distance from the sun at a regular rate, 1 and given
tha.t it has at the same time a tendency to mov~ parallel to
a certain fixed hne at a certain rate, we find the body describes
an ellipse. This is only one geometrical fact deduced from
another. Given that a planet moves in an ellipse with a certain
varying velocity, we can prove that its distance from the sun
tends to shorten in the way attributed to attraction, and that
its tendency to move m the parallel direction is constant.
What is not supplied by geometry 1s assumed, i. e. the
parallelogram of forces.
§ 369. This (i e the mathematical work) is entirely misunder•
stood by Mill m his chapter on the Deductive method. He
1
The mverse square of the dJStance.
 The Melhod of Physics
supposes that if the effects of the forces, operating singly, are
given, or the law of each force operating singly is given, ~he
combined effect, when the forces act together, can be found by
ratiocination, a mere name for the syllogism rejected in Book II.
He is thinking of course of mathematical physics and especially
of the example we gave in the last section. He supposes, given
the forces acting, given the operation of the law of attraction
on a body, and the operation of the original projection, that the
result, the path of the moving body, is got simply by deduction.
This is impossible and, given the laws operating singly, mathe•
matics as such is quite unable to discover their combined effect.
It cannot move a step without assuming a physical law of
combination, that is, how these forces will act in combination.
In the given problem mathematics solves a purely mathe•
matical question : 1t does not touch the conception of Force
as such at all. It deals only with the changes of geometrical
relations, assumed as the effects of the forces ; and it assumes
what 1t is supposed to prove, namely, certain laws of the com•
bined operation of physical forces. It is the easier for Mill to
make this mistake because he has so entirely mistaken the
nature of the geometrical method itself. He thinks of it in
general as a method of deduction. If he had realized that its
method 1s construction, he might have seen that the construe•
tion, being solely of mathematical quantities or relations, could
not possibly deal with the physical conception of force, which
is not 1dent1fiablc with such relations.
§ 370. In one department of these sciences, i. e. in dynamics
and statics, motions are considered without relation to their
effects on consciousness. But there is another department in
which motions are treated as causes of what are taken to be
sensations of ours-heat, light, and sound. Here we are concerned
with another set of causal laws, i. e. the effects in feeling-con•
sciousness of some action (in the way of movement) upon our
sense organs. Now here the mathematical part is concerned
solely with the movements of the material bodies. It cannot
deal with them as causes of sensation. Violet may be the colour
corresponding to a certain wave-length and the ordinary theory
is that it is a sensation of ours, like heat or cold, caused by the
impact of ether waves of a given wave-length on our retina.
 SPECIAL LOGIC
But mathematics can only deal with the undulations : it cannot
say why violet corresponds to, or is the effect of, the undulatory
movement. That is learnt " posteriori from certain facts of
consciousness taken in connexion with the hypothesis of the
undulatory theory. Sometimes, again, it is supposed that the
concord between two musical notes 1s accounted for by the fact
that a certain number of vibrations of the air corresponding to
the one take place m the time occupied by an integral number
of the vibrations corresponding to the other. But this is an
illusion. It 1s merely an interesting fact that such commen-
surateness corresponds to our musical sense of concord. The
science fails to establish any connexion between the two. We
cannot possibly show why that numerical connexion should have
as its effect a musical concord The apprehension of the sensa·
tion, if it be such, 1s in no sense an apprehension of the vibrations
which arc its physical cause.
 SYMBOLIC LOGIC
§ 37 I. THE term Symbolic Logic is taken in the following
sections m its current meaning. It denotes, that is, a theory
which, whether rightly called logic or not, proposes to conduct
operations of purely formal inference by the aid of an algebraic
calculus. Dodgson, bin calling his book Symbolic Logic, 1 departed
from what seems to be the common usage, for he did not employ
a mathematical calculus ; and how far he was from a mathe•
matical treatment is still more evident from the unpubbshed
second part, of which he, with his usual kindness, sent me the
most important of the proof sheets. I hke to think that my
main contention would probably have pleased him a good deal.
A common objection to Symbolic Logic, as thus defined, is
that it 1s mathematics and not logic, and so is not necessary to
the log1CJan. A lmc of defence taken by symbolic logicians 1s
that the subJect is not strictly mathemat1cal, the signs of multi•
pbcatlon and divis10n have not the mathematical meaning. But
while th1s could hardly be alleged in the case of the signs +
and -, there must be something essential, common to the use
in mathematics and symbolic logic of the signs of multi-
phcation and d1v1S1on, for otherwise the employment of these
signs m the latter would be ridiculous. The truth is, that
in symbolic logic these formulae follow the same laws of develop·
ment as the correspondmg formulae (or operations, rather) in
mathematics, notably in the matter of the multiphcation of the
signs + and -. 2 Besides, the idea of an equation, which is
• Symbolic Logic', Part I, Elementary, by Lewis Carroll London (Mac-
millan), 18g6. • See p 641, § Ji'9, infra
[& Redrafted from two manuscnpt lectures of Hilary Term, 1898, and
manuscnpt connected with the same subJect I have left the •propositional•
terminology
b The Rev C L Dodgson, the author of Alice in Wonderland, was lately
dead when the lectures were delivered, There 1s a kind of mathematical
calculus suggested m Dodgson's book as an alternative to the counter system,
which makes the easy part and the bulk of the treatise.]
2773"2 Q
 SPECIAL LOGIC
fundamental to the subject, is obviously a mathematical idea..
Thus Boole entitled his earber investigation ' The mathematical
analysis of Logic ', and gave as a sub-title to his better-known
work 1 ' The mathematical theories of logic and probabilities ',
§ 372. A stronger hne of defence may be suggested. Suppose
it is mathematics ; provided only it is sound mathematics, it
is futile to object. We may not like mathematics, but that is
not an objection to the subject. And we may not like the extra
trouble ; but astronomers might as well agree to ignore the
planet Neptune, as a member of the Solar System, because it
gives them extra trouble. On the other hand 1t would be a most
interesting fact that mathematics could conduct logical pro-
cesses (as Boole's titles imply), and the logician would be bound
to study it as throwing light both on logic and mathematics.
Nor is it really enough to object that the inferences can be
conducted by ordinary methods, so that a calculus is not neces-
sary, or even that these ordinary methods are simpler, quicker,
and so better. The symbohc logicians would probably deny the
truth of both obJectlons and certainly they would deny the latter
with entire confidence. The answer should rather be that, even
if the calculus were more cumbrous, we should have to study
it or we should be cuttmg ourselves off from a most interesting
piece of knowledge. The objections clearly are only relevant to
any claims of superiority made for the calculus and not to the
questions of its right to exist and of its validity.
§ 373. The original objection, however, has a deeper basis
than mere prejudice and rests on a sound notion of logic, often
insufficiently realized, especially in Deductive Logic, but tending
to become conscious of itself, when such investigations as these
are presented seriously as logic. This sound notion is that, with
regard to inference, the methods of all the special sciences are
the object-matter of logic ; it neither makes them nor conducts
them, but studies them. It is thus, in a sense, more general
than any of them. This being so, 1t is incredible that any
particular scientific method, for example the mathematical,
should be the method of logic itself. We suspect somethmg
wrong and that the subject cannot be logic at all. That is an
1
A11 ltlflutigation of Ills La111s of TliougM, o,s Which a,e fON'#dlrl Che MtdM•
malscal Thlones of Logic anrl Probalnl1lses, George Boole, 1854.
 Sy,nl,olic Logi,c
objection which is never even dreamed of by the symbolic
logicians, but the greatest danger usually comes from the quarter
which we least suspect. Indeed it would be a shock to their
feelings to have their title to be logicians disputed, for their
language is in no sense self-depreciatory and they openly profess
to make great improvements on the common or ' ordinary '
logic,• as they call it.
§ 374. The foreboding, however, of the 'common' logician is
Justified, for he will find that the symbolic logic as such consists
of the solution of particular problems, which are on the same
plane as the solution of geometrical or algebraic problems and,
though concerned with the abstract forms of subject and pre•
d1cate, as specially scientific as these mathematical processes--
no more logic than they are, and related to logic precisely as
they are. Incidentally there is a little elementary logic involved,
but the real and serious problems of logic proper do not appear,
nor is the symbolic logic able to touch them. In comparison
with the serious business of logic proper, the occupations of the
symbolic logician are merely trivial. I do not want, however,
at present to dwell either upon the comparative importance of
the ' logical ' calculus or on its relation to true logic, but to take
it rather at its own valuation, to inquire what its problem is,
and with what success the solution is attempted.
§ 37 5. The problem stated in most general terms is to find
what various inferences can be drawn from premisses btated in
the abstract symbolic form of the traditional syllogistic logic,
with no other quantitative notions than ' all ', ' some ', and
' none '. It might seem that in this comparatively humble
sphere there would be no difficulty, as every step of the argu-
ment is either immediate inference or needs no more than the
first figure and mood of the syllogism, than which nothing could
be simpler. But 1t should be remembered that, in the solution
of a difficult and complicated mathematical problem, every
single step, when the reasoning is fully expressed, may be very
simple. The difficulty consists in finding the right steps to take
and the right order of them. And so, in formal thinking, it is
possible to devise problems of such difficulty as would never be
[• The phrase goes back to Bacon. Here Wilson defends logms commrmu
polemically though he continually elsewhere speaks dlSl'eflpectfully of 1t ]
g2
 SPECIAL LOGIC
anticipated by those who are only accustomed to such specimens
of formal argument as are given in the traditional logic.
If the premisses have compound subjects and predicates, as
all ABC is D, or DE, and if they are further perplexed by
a Judicious sprmkling of negatives (e.g. all A, not-B, C is not-D),
if further a considerable number of premisses be taken, a problem
may be invented which may take hours to solve, unless we have
some definite rule and method besides the mere possession of
the first syllogistic figure.
The followmg 1s an mstance of a formal problem devised
by Dodgson, who was clever in the construction of difficult
problems.
Given that-
(1) all active old Jews arc healthy,
(2) all mdolcnt magistrates arc unpopular,
(3) all nch snuff-takers arc unhealthy,
(4) all sarcw.t1c magistrates arc Jews,
(S) all young snuff-takers arc pale,
(6) all nch old men, who are unhealthy, arc sarcastic,
(7) all magistrates who are not poets arc studious,
(8) all rosy magistrates arc talented,
(9) all talented and popular students are rich,
(ro) all pale snuff-takers are unpopular,
(II) all unpopular magistrates arc abstainers from snuff,
(12) all talented poets who are active are rich;
to prove that no magistrates are snuff-takers.
To reach this conclusion by a series of ordmary syllogistic
arguments would mvolve a very long process. By Boole's
methods I have calculated that 2,048 factors are required, but
by an improved calculus of my own the result may be gamed
10 a very short time
A simpler example 1s the following .
Given that No AC 1s Band that no D which is not-Bis A,
To prove that all A 1s either not-C or not-D.
A third example might be one given by Boole.1
Given that-
(1) If A and Care absent, Eis found with B or D, but not
with both.
• Laws of Th014gl,I, p 146.
 Sytnl,o.lic Logic
(2) If A and D are present and E is missing, either both
B and C will be found or will both be absent.
(3) If A is present with either B or E or both, then C or
D will be present, but not both, and conversely.
What conclusions can be drawn ?
This is said to be ' the most intricate of any given by Boole '. 1
It took two and three-quarter hours to work out by ordinary
methods, but only twenty minutes by my own calculus. The
solution Venn gives could not have taken very long.
One soon discovers that some definite rule and method is
required here as much as in mathematics, and on this point
Dodgson, in the preface to his book, rightly insists •
§ 376. Method, then, being needed, the subject really being
a particular science, any method deserves to be examined, pro-
vided always the thing is worth doing at all. Hence, pnma
f acie, there is a case for symbohc logic, though it can hardly
be said to have originated in a consciousness of the difficulty
of such problems as I have been alluding to. It was probably
felt by those who originated it that formal reasoning was like
mathematics, in its a priori character, in its necessity and
absolute accuracy, and hke algebra, in the use of abstract
symbols. The science of pure quantity had received an enormous
extension by the invention of the symbolic method of algebra,
and the thought naturally arose that something similar might
be done for formal thinking
Now when the perception of the aforesaid analogies has sug•
gested symbolism, two courses are possible. We may either
study the peculiarities of the actual problems which have arisen
in the science and consider what kmd of symbolism they seem
naturally to suggest as hkely to facilitate their solution, or we
may try to apply so far as possible the already existing symbolism
of algebra, as it stands, and thus invent a calculus on strict
algebraic analogy and then see what we can get out of 1t. This
second seems really to be the way in which symbolic logic came
mto being. The first method is clearly the sound one, for,
• Symbolic Logic, by John Venn, London (Macmillan), p 351.
[• In the original there 1s here a reference to the great Joss to philosophy
by the untunely death of Mr Adamson, Fellow of St John"s College, • a good
mathematician and a promJSmg metaphysician • )
 SPECIAL LOGIC
though it is right enough to take the suggestion of symbolism
in general from algebra, it is a mistake to suppose that the
symbolism adapted to one science is likely to suit another. The
lesson which modern logic has learned and has to teach is that
the method of the particular sciences depends on their object-
matter, the various forms of the method of science are discovered
by srienttfic men themselves in the study of the special problems
peculiar to their respective sciences, and they cannot be ais-
covered in any other way.
§ 377. The presumption is therefore that if the reasonings in
a given science can be conducted by a symbolic method, the
same symbolism will not suit another. The case of analytical
(i. e. Cartesian) geometry (which is perhaps too easily misrepre-
sented) is not an exception. There geometrical relations give
rise to relations of quantity, not sham quantities, but real
quantities which can enter into equations ; they thus give rise
to true algebraic problems which algebra solves in real, not
sham, equations, with real multiplication and division. Con-
sequently, if the symbolism of algebra is transferred to the
science of formal thmkmg, we (the common logicians) cannot
help entertaimng certain anxieties. We should fear that mathe-
matical analogy, the symbolism being the same, might some-
times mislead, and a relation of logical terms be confounded
with a relation of literal quantity. Again, we should fear that
the algebraical processes would only fit partially, and that the
calculus, instead of being a potent instrument as in true algebra,
might be rather narrowed in its use ; the attempt to carry it
out consistently as a general calculus might even lead to contra-
diction and positive error. We might again expect that much
ingenuity would be expended in solving problems the difficulties
of which were due to the artificial form into which the premisses
were forced by the calculus itself. Lastly, while skilful enough
in overcoming unnecessary difficulties of its own making, the
calculus might possibly not facilitate the solution of the difficult
problems which arise naturally in formal thinking itself.
§ 378. To Judge from its contents, the main ideas of the logical
calculus seem to have been developed in some such way as this.
Formal logic from Aristotle downwards has used symbols, viz.
letters of the alphabet, like algebra, for class conceptions ;
 Symbolic Logic
for instance, all A _is B. Operations with these symbols it
recognizes in symbolic form only in the case of the association of
two class conceptions to form a species. Thus AB means that
which is both A and B, and thus a narrower class than either.
This symbol AB resembles the symbol in algebra of the multi-
plication together of two quantities. These symbols, then, form
the starting-point. Equations bemg the important instrument
of algebra and its applications, an equation is sought for between
the given symbols. 'A "" B' suggests itself, and the question
arises how this should be interpreted. It is seen that the
identity of the extension of two classes could be thus repre-
sented, and this is the interpretation given ; that is to say,
A .. B means that the whole of the individuals in the class A
constitute the whole of those in the class B.
A+ B, then, would mean the united extension of A and B, or
the class formed by uniting A and B. If A' and B', then, sym-
bolize not-A and not-B respectively, we get A+ B = AB'+ AB
+ AB + A'B. Here the interpretat10n of AB + AB is not what
it would be in algebra, because if the classes A and B overlap,
the class AB is not counted twice over.
Thus AB+ AB = AB,
Similarly, A+ A = A and not 2A.
If B is part of A, the usual symbol for subtraction may be
used, so that A - B = AB'. Here certain difficulties arise if B
is not wholly contained in A. They need not, however, be
discussed here.
§ 379. The ordinary law of algebraic multiplication is said to
hold;
viz. if a+ b = :x,
ex ... c(a+b) = ca+cb
J EB
& c: b
I
for what is common to the extension of c
and :x must consist of what is common to
c and the parts a and b of x. Supposing c has nothtng in com•
mon with b, then the class cb is non-existent and so the equation
remains true because cb = o. This last is, as we shall see,
a serious assumption. It may, then, easily be verified that
the multiplication rule obtains, if for c a polynomial be intro-
duced. Thus
(a+b) (c+d) = ac+ad+bc+bd;
 SPECIAL LOGIC
also that the algebraic rules about the signs + and - hold in
this sort of multiplication. Algebraic powers, however, cannot
enter since aa = a, obviously. In fact for Boole x2 = x is the
expression of the fundamental rule of thought. The operation
of division appears to a certam hm1ted extent, and such a symbol
as ", of which the symbolic logicians are rather proud, though
)'
it 1s really a snare spread for their feet.
§ 380. The symbol ab, if compared with the corresponding
algebraic symbol, has nothing to do with multiplication m the
sense of the repetition of a umt, but serves to md1cate an opera-
tion following some of the general laws which are followed by
the symbolism of multrplicat1on. In ab, a may be taken as the
symbol of an operation, viz multiplication, performed upon b as
a given quantity, but b may Just as well be a symbol of opera-
tion, and the mathematical theory of symbols of operation seeks
to make such symbols always represent the operation of multi-
plication. Thus the symbol a standing alone represents the
operation of multiplication performed on the umt of number,
1, a thus 1s eqmvalent to I xa and ab = (I xa) xb. a and
b then are always symbols of operation, that which is operated
on being the umt whatever 1t may be. Unity, then, 1s under-
stood always and not expressed with the symbols
In the symbolic logic, however, ab denotes an operation of
a upon b, which 1s the operation of takmg the part of b common
to 1t and a. On the analogy of mathematics we seek for a way
of representing a itself as an operation. What, then, must be
the class operated upon by a to produce a, the ultimate umt,
so to speak, operated on? That 1s to say, what 1s that class
such that when we take out of 1t what ts common to 1t and
any class a we get the class a itself? The answer 1s obviously
'reality'. As reality, then, is to be understood with every
symbol like ab, just as I is m algebra, the symbol I is used in
this calculus as the symbol of reality.
This may seem ingenious, but 1t 1s really unfortunate. It
would have been much better to have taken a symbol which
originally had no numerical significance to represent reality,
e. g. r ; then ab would be understood to mean abr. The negative
of r, that is the unreal or 1mposs1ble, would be represented by
 Symbolic Logic
r', like any other negative class. A serious mistake would
thus have been avoided ; for, as we shall see presently, the
symbol o or zero, besides its significance as a true zero, has to
do duty in the calculus as the symbol of the impossible with
grave consequences.
[For the present we are concerned only with the normal
meaning of zero, in the logical calculus. A non-existent class
is regarded as a compartment of reality which is empty and is
represented by o. This appears to follow m a mathematical
manner from what has been already laid down, for ab denotes
the operation of taking the part of the class b common to it
and a. Thus oa ( = o) will stand for the thmgs which are both
nothmg and a, that is, for any empty class.]
§ 381. On the use of I to symbolize reality is founded the
equation fundamental to the calculus x + x' = I, wluch expresses
(or is supposed to express) the fact that everything must be
x or not-x. The equation expresses this in the form that the
whole of reality ( = 1) 1s made up of the class x and not-x,
i. e. what is not m the class :x.
S1m1larly a = ax+ax', wluch we may get directly, or derive
from the above by writing (x+x') for I thus
a= ax
= a (x+x') = ax+ax'
These equations we shall call for convenience equations of
dichotomy, and shall cntic1ze them presently.
§ 382. We will now turn back to the equation A = B. It
expresses primarily a relation between the extensions, not the
intens10ns, of A and B. That is, 1t does not mean that the
nature of A necessitates that of B and conversely, but that
the two classes comc1de. This leads, for example in Venn, to
the crudest and naivest nominalism. Now, in the case of truly
universal notions, the extension 1smfinite and there 1s absolutely
no way of 1dentifymg their extensions except through the con•
ceptions. We cannot know that the class A coincides with B,
except by knowing that the conception of A necessitates that
of B, and conversely. That 1s1 we can only know it through
the two propositions, all A is B and all B 1s A. This is, then,
not so much the interpretation of a given equation, but the
knowledge we must have before we can form the equation at
 SPECIAL LOGIC
all. Similarly the proposition A= B +C implies all A is either
B or C, all B is A, and all C is A. Since, then, the relation
between the extension of two or more conceptions can only be
got at through the relation between the conceptions themselves,
as expressed m the aforesaid propositions, one may be curious
to know beforehand how far any reasoning from the equation
can be carried on without referring back to these propositions.
§ 383. It is obvious that the equation A= B will not do ior
the ordinary proposition 'a.11 A 1s B' , it only represents the
case where 'all A is B' is convertible simply, that is, it repre•
sents the two propositions 'all A is B' and 'all B is A' taken
together. Now the important problem of the calculus is to find
an equation for the ordinary proposition 'all A 1s B '. The
relation or equation between the extension of A and B 1s clearly
only A = some B. This is represented by A = v B, where v has
the special significanre of 'some' This 1s merely the pretence
of an equation, a form of equation without the reality. For an
equation something definable must be identified with something
else definable, but here a definite A is equated to an mdefinite,
which seems to contradict the idea of an equation. The symbol
v does not improve matters, especially when we find that the
calculus, for other rt>asons, 1s obliged to give it the entirely
indeterminate meaning 'some' ; it may be none, or all, or a part.
Further, A = vB does not even tell us that any A is B.
§ 384. There is, however, another form of equation for the
proposition all x 1s y, due apparently to Boole, which seems to
have the exactness necessary for an equation and to secure
another advantage. In this form of calculus the proposition all
X is Y is treated as hypothetical, not so much from a theory
that the universal affirmative categorical is necessarily hypo•
thetical (if anything is X, it is Y), but from the special needs
of the calculus itself. In algebra when we set a problem and
have to determine the value of a so-called 'unknown• x, which
suits the problem, we may be asking that certain conditions
shall be satisfied whlch cannot be, and then our calculation
shows this by representing finally the unknown x under the
form of an impossible operation. Thus we cannot say that the
variables in our equations always represent real quantities. This
logical calculus proposes 1ts problems with the same generality,
 Sy,nbolic Logic
and rightly. A subject X is given as satisfying certain con-
ditions in the form of premisses, and the combination of these
may result in the conclusion that X has contradictory attributes
(say X is Y and X is Y'). We are in such a case not to pro•
nounce the premisses false, but X impossible. The important
case will be when X is compound = AB ; then the conclusion
AB is impossible yields the judgement 'no A is B '. Consequently
in a premiss all X 1s Y, we cannot say that X is possible, it
may be impossible, and so it must be read, 'X, if possible, is
Y', or 'X, whether real or unreal, 1s Y'. On this account
propositions which arc uuiversal and affirmative are said not to
be existential, that is, not to convey definite statements about
existence. They are regarded, whether rightly or not, as in
a sense ambiguous. However, it is said, the proposition all X
is Y docs involve a definite statement about existence, viz. that
there are no really existing things which are both X and not-Y,
or XY'. This 1s taken, then, as the equivalent of the proposition
all X is Y and expressed in the form there are no members in
the class XY'. This translated into equational form becomes
XY' = o. This seems at first most ingenious, for we now have
an equation of the requisite accurate and definite mathematical
form, and we need not mscrt any hypothesis about the existence
of the terms. This equation 1s possibly the most important
and fundamental feature of the calculus. I shall try to show
that it is an dlus10n, by proving -first, that the proposition which
it symbolizes, no XY' exists, though an inference from all X is
Y, is not the true eqmvalent of it ; secondly, that given the
proposition 'No XY' exists' (or XY' is impossible) it cannot be
rightly symbolized by the equation XY' = o, and that the
symbolism involves a grave error from the point of view of
a mathematical calculus, or theory of symbols of operation in
general. Finally, that it is impossible to give a true equational
form at all to the universal proposition 'all X is Y '.
§ 385. We are led to suspect some error in the fundamental
equation, if we consider a curious theorem in the doctrine of
elimination • in the calculus which depends on this equation.
Logical elimination so called means merely this. Given pro·
[• Some of this was suggested, I think, by Venn's s,,,,.,,ol,c Lo,,e, cha :iav
and xv.]
 SPECIAL LOGIC
positions involving conceptions such as x, y, ands, a given element
is said to be eliminated when a proposition is found from the
data which docs not involve that element. Thus tn the syl-
logism the middle term is ehmmated. If all X is YZ, the
proposition All X 1s Y 1s true and 1s said to eliminate Z. In the
symbolic logic elimination 1s made very easy in the case of
particular propositions by the prmc1ple that such propositions
always imply the existence of their subJects, while universal$ do
not. Thus, wlulc All X ts Y cannot guarantee the existence of
any X, Some X 1s Y 1mphcs the real existence of some X and
therefore of some Y also. Thus in Some X 1s YZ we can eliminate
each element singly and each group of two thus .
some X exists,
some Y exists,
some XY exists, and so on
Now, at first sight, 1t seems odd that the merely indefinite
subject, some X, should ensure the existence of X, while the
definite, all X, cannot do this a And, 1£ the universal X 1s Y
must, for the reason stated, mean, If any X exists 1t is Y, then
for exadly the same reason the particular, Some X is Y, should
mean Some X, 1f 1t exists, is Y ; or, Some X, 1£ 1t were to exist,
would be Y. Formal proulcms lead to such an hypothetical some
jm,t as they do to an hypothetical all. Take, for instance, All
A 1s Y, all A 1s X: therefore, Some X, viz. AX, 1s Y. Here
the some X meant 1s clearly hypothetical, since A is hypothetical
It must mean: Some X, 1f 1t existed, would be Y. There must,
then, be some strong reason for a doctrine which presents such
obvious difficulties, and the fact 1s that the symbolists arc
committed to 1t as the necessary consequence of their view of
the equation XY' = o The argument 1s .
Some X is not Y is the contradictory of the proposition All
X JS y'
All X 1s Y is equivalent to No XY' exists.
Therefore Some X 1s not-Y is the contradictory of No XY'
exists. But the contradictory of No XY' exists 1s Some XY'
exists. Therefore Some X is not-Y 1s equivalent to Some XY'
[• In Dodgson's system All X is Y is considered to be equivalent to Some
X (at least) 1s Y, and No Xis Y' The resolution is, however assumed, never
justified]
 Symbolic Logic
exists, that is, Some really existent Xis not-Y. For the affirma•
tive we have to use the negative universal, which in this theory
is No XY' exists ( = No Xis Y') or XY' = o. The contradictory
of No XY' exists is correctly stated; thus the argument must
turn on the equivalence of No XY' exists to All X is Y; and
so if we suspect the conclusion, we must suspect this equivalence.
Indeed we may reverse the argument and say the two proposi•
tions cannot be equivalent because their contradictories are not.
For clearly, while the contradictory of No XY' exists is Some
XY' exists, i e. Some really existent X is not Y, the contra•
dictory of All X 1s Y (understood as equivalent to Any X, 1f it
exists (or existed), is (or would be) Y) is Some X, 1f it existed,
would not be Y. Or, to put 1t m another way, All X is Y is
equivalent to All X, whether real or not, is Y, and the contra•
dictory of this must be Some X, whether real or not, is not Y.
§ 386. Agam the equivalence of the two propositions is dis·
proved by the fact that the existential meanmg which it neces•
s1tates for the particular as against the universal proposition can
be shown to involve the theory m a contradiction. It may be
premised that lhe existential propos1t1on, No XY' exists, must
be understood m the wide sense, No XY' can exist at any time;
for, otherwise, 1t would be compatible with the contradictory of
the proposition, all X 1s Y, which 1t is supposed to represent.
All A 1s B = If a thing 1s A, 1t 1s B (1)
No A is B = If a thmg 1s A, 1t 1s not B . (2)
Therefore (1), 1£ true, necessitates that (2) is untrue.
But (2), according to these logicians, = No AB exists, or No
AB is possible, . (3)
Therefore (1) necessitates the untruth of (3).
But the untruth of (3) necessitates the truth of its contra-
dictory.
And the contradictory of (3) 1s 'some AB exists'.
Hence, 'all A 1s B' necessitates the proposition 'some AB
exists', 1. e. some A exists and 1s B.
Thus the universal affirmative necessarily implies that its
subject is real, which is contrary to hypothesis This is a
sufficient reductio ad absurdum The equation, then, XY' = o,
if representmg No XY' can exist, is not the true equivalent of
All Xis Y.
 SPECIAL LOGIC
f 387. We shall now inquire what is true of XY' i whether
any equation for All X is Y can be got from it ; and the meaning
and consequences of equating XY' to zero. If all X is Y, then
XY' is YY' ; that is, XY' is impossible, and has the particular
form of impossibility YY'. If we take 'XY' = o' merely to mean
'XY' is 1mposs1ble' we certainly do not take the full inference from
All X 1s Y, because we omit the particular sort of impossibility
to which 'all X is Y' leads; accordingly we cannot have 'gt>t
a true equivalent for All Xis Y. This is verified by the contra-
diction test. For, as in the case already considered, the con-
tradictory of the two propos1tlons 1s not the same. Indeed the
proposition XY' is impossible is really m the same position as
No XY can exist.
The want of equivalence 1s otherwise evident ; for, given
XY' is impossible, what follows is not that all X (whether real
or not) is Y (the true eq111valent of All X 1s Y), but the disjunctive,
Either X is unreal, or Y' is unreal, or X, 1f real, is Y.
That the full inference All XY' 1s YY', 1s a true equivalent
of AU X 1s Y, can be verified by inference, or by takmg the
contradictory Thus :
(1) Xis Y or Y', that is, is Y or XY',
but XY' is Y,
therefore all X (whether real or unreal) is Y
(2) The contradictory of 'XY' is YY" is Some XY' is not
Y'Y,
but all XY' must be Y'.
Therefore some XY' 1s not Y.
But this would be impossible 1f all X were Y, therefore some
X is not Y ; which is the contradictory of All X 1s Y.
But this does not yield an equation. We cannot put XY' ...
YY' because XY' is not coincident with every form of YY'.
The nearest we can get to an equational form is XY' = some
YY', or XY' = vYY'. Thus, mstead of improving on tlus and
gettmg a more accurate form and one which corresponds to the
determmate character of a true equation, we have only come
back to it and with no improvement, for X = vY makes
XY' = vYY' unnecessary. In view of these considerations it
1s of importance to observe that Venn, 1 after a sneering reference
1 Venn, 1. c, p. 403
 to the venerable structure of the syllogism, continues : ' the
distinction between universal and particular propositions which
to it is unimportant is to us vital.'
§ 388. So far we have examined the logic which leads to the
equation XY' = o, and found that it contains serious errors.
We will now consider the mathematical side. XY' represents,
we have seen, an impossible class, and it is as such that it is
equated to zero ; the idea is that XY' bemg an impossible class
there are no members m it, and so its extension 1s o. Thus
XY' = o. Similarly, 1f A is B, AB' = o. And clearly these
equations will be very poor equations, indeed a mere shamt if
it does not follow that XY' = AB', whatever propositions XY
and AB represent. The question therefore anses, what con•
sequences have such equations? But strangely enough these
equations, though the absolutely necessary result of an equa-
tional calculus, do not seem to have been formed or considered.
The equation of 1mposs1bles as such to zero does not arise
merely from the attempt to get an equation for the universal
affirmative All X 1s Y, but is necessary everywhere to this
calculus.
The distributive law in the association of class conceptions,
or rather the limiting of one class by another, or taking the
extension common to two classes, a (b +c) = ab +ac, corre•
sponding to the distributive law of the operation of multtplica•
tion in algebra, depends entirely on the equation of impossibles
to zero.

b C b C b C

§ 389. The common part of the class a and the class formed
by putting together b and c, supposing there 1s any common
part, will certainly include either ab or ac, and it may include
both, that is, it will have a common part with b or c, or both,
but it may have no common part with one of them. It is
possible, for instance, that there may be nothing common to
a and c and, therefore, no class ac at all. Then c is a'. Supp0&e
 SPECIAL LOGIC
c ,_ a'd, then ac = aa'd. But the formula for a (b +c), although
this is intended for a real class, gives always a (b +c) = ab +ac;
therefore a (b +c) =- ab +aa'd, even when ac is an impossible
class. Thus the equation can only be maintained by equating
the impossible class ac or aa' d to zero ; then and then only the
right result, a (b + c) = ab, is obtained. The same is true of all
the equational forms which attempt to represent dichotomy.
The fundamental one is x+x' = I (=all reality). Now, 1£ ~ 1s
real, x' (or not-x) includes not only the part of reality which
is not-x, but all unrealities or all 1mpossible classes Thus the
equation would falsely make reality include unrealities, and this
can only be prevented by equating the unreal part of x'
to zero.
Similarly the equation a = ax+ax', which may be regarded
as an application of the above, is given as always true But
if a is real (and even though x and x' arc restricted to realities),
one of the classes on the r1ght-hancl side of the equation may
be unreal For instance, if all a 1s x, then ax' is xx' and ax' 1s
11nposs1blc. The equation, then, 1:,, only saved by equating ax'
to zero, when we get a = ax, wh1c,h 1s true.
All 1rnpossibihties, then, are equated to zero. Thus, 1f A 1s
Band Xis Y,
AB' = o, XY' = o, AA' = o, BB' = o, XX' = o, YY' = o,
and so on.
1n this equating of the impossible to zero, two points are to
be noticed and distingmshed. First, all imposs1bilit1es are
equated to the same, which has the effect of equating them to
one another a result which equally follows 1£ they are not
equated to zero, provided they are equated to the same thing.
Secondly, each 1mposs1b1hty 1s equated to zero
§ 390 In algebra both these statements would be considered
fallacies and elementary fallacies. Familiar examples, which
illustrate the principle involved in the equation of all impossibles
to zero, are such as the following. Every person in the next
room can prove that the diagonal of the square is commensurable
with the side. This statement 1s supposed by some to be fully
justified because there are no persons in the next room. Agam,
every person who can marry a person who hved a thousand years
ago can prove the diagonal of a square to be commensurate
 Symbolic; Logic;
with its side. This is justified by saying that there are no
persons in either category.
The justification is a fallacy, and it is partly on this fallacy,
we may remark incidentally, that the modern form of the
doctrine of the reduction of the categorical to the hypothetical
proposition depends.1 Moreover, the most bnlliant and striking
part of modern mathematics, the treatment of so-called imagi-
nary quantities, depends on a prmciple which makes it a fallacy.
In algebra the proof that 1mposs1bil1ties can neither be equated
to one another nor to zero 1s easily given. Probably most, if
not all, of us, when we first arrived at a result of the form
x2 +I= o, therefore x2 = -1, 1f we avoided the mistake of
putting x = - I, and saw x to be impossible because the equa-
tion makes x equal to the square root of - I, rightly said that
no quantity would sc1.ttsfy the cquat10n and then wrongly trans-
lated this mto x = o. The mistake 1s corrected by observmg
that x = o will not satisfy the given equation For, put x = o
and we get the absurdity o = I.
Again the equation of impossibles to one another is easily
disproved For suppose a and b real quantities and unequal.
If ✓ - I a = ~ b, because both are impossibles, then if we
operate on each side by the mverse operation of the root, that
1s, if we square them, we get -a = -b, and therefore a and
b are equal. Thus the equation of 1mposs1bles or imaginaries
to one another involves the equation of all real quantities to
one another.
§ 391." If at first this simple opera- v
t10n of squaring seems suspic10us,
we have a better illustration m a
more concrete form in the application
of algebra to geometry. Take a point A
0 in a straight line and, at a given dis-
tance a, draw another straight hne at
right angles to the hne From the
given pomt as centre describe a circle
of radius less than a and therefore not
cutting the straight hoe AY. If we combine the equation of
l Cf § 400.
[& Cf § II6]
R.
 SPECIAL LoGtC
the circle with that of the straight line we find, for the distances
of the intersection of the two from A, two impossible quantities,
that is, reals affected with the sign ~ - If we describe
another concentric circle also not cutting AY, we get for its
intersections with AY two other impossible quantities, different
from the first in form, depending on the difference of radius.
Two different sets of impossibilities thus correspond to the
different equations of the two circles, which really means that
the 1mpo,ssibihty of one's cuttmg the hoe is different from the
impossibility of the other's cutting the line. If now we identify
the two 1mpossibilitics, we identify the equations of the two
circles and get the absurd result that all circles with the same
centre O which do not meet AY coincide with one another.
That the equation of impossibles to one another would imply
the equation of unequal realities to one another, and indeed the
equation of all realities to one another, 1s a thing we should
anticipate on quite general grounds. For an impossibility is
only the other side of a necessity ; every different necessity
gives rise to a d1fierent 1mposs1bihty And so there is nothing
mysterious m the treatment by mathematiuans of impossibilities
as definite, not as mere nothmgs merged m an mdistmgmshable
zero. Thus, if we identify different impossibilities with one
another, we shall expect to find that we have thereby identified
the corresponding necessities with one another. 1
§ 392. • Professional, and mdeed distinguished, mathemati-
cians, no mere amateurs, have invented this system of symbolic
logic, and so it may seem strange that they should have allowed
the equation of each impossible to zero ; but perhaps there were
these reasons for it. First, the equational form cannot be
maintained without it. Secondly, by the help of it, an equa-
tional system is produced which is coherent within its own
limits, Just as in algebra, if we equate the unreal roots of equa-
tions to zero, we shall find the real roots correctly enough, and
may avoid the contradiction resulting from the equation of
unreals to zero by refusing to consider unreals any further. In
this way, however, algebra would be cut down to very narrow
1
Cf §400.

[• On Symbolism m this connexion see Venn, I c, ch. 1v]

 S)'IUOffc 1Agu
limits, and we should lose the most important part of it.
Thirdly, it was probably thought that the equation of impos•
sibles to zero would have no further consequences like those in
algebra and therefore nothing would be lost and no contra-
diction caused. And that for two reasons. First, the equations
which usually occur in this calculus do not present forms in
which the equation of an impossible to zero would cause the
equation of a real to zero, as in the algebraical equation x1 + I = o.
Secondly, it probably seemed, if the experiment of equating
two impossibles to one another was tried at all, that nothing
but nugatory or identical predication could result.
§ 393. However, the experiment may be tried of constructing
an equation, according to the rules of the calculus, which is
parallel to x 2 + 1 = o and leads to the same sort of contra-
diction, if impossibility is made identical with zero. Thus we
might show that, in this calculus too, the equation of impossibles
to one another results m equating all real classes to one another.
We might also attempt to show the same contradiction by
a proof analogous to the geometrical reductio ad absurdum in
regard to the concentric circles, none of which cut a given hne.
To take first a proof of the second kind. Consider the equation
aa' = bb'. According to the rules of the calculus this neces•
sitates all aa' is b, which at first seems to yield nothing but
identities. We shall get something else out of it if we inquire
how impossibles such as aa' can get predicates at all. It must
be remembered that in this calculus they do have predicates
assigned to them-m proofs by reductio ad absurdum-and 1t is
in consequence of these predicates leading to a combination of
contradictories that we discover that they are impossibles.
In such indirect proofs in geometry, in order to prove that
A is not B, we assume that AB 1s a possible combination, AB
being really impossible, though the separate elements in it
are possible. We then argue further, not from the combmatlon
of impossibles as a combination, for that we cannot do, but by
finding (or having given) predicates of the real elements in it.
We then predicate of the problematical whole AB all the pre-
dicates which belong to its parts, and the argument ends when,
among the predicates combined, contradictories are found.
Thus the impossible combination AB only acquires predicates
R2
 SPECIAL LOGIC
through the fact that they attach to elements in AB. Assuming
this, it can be shown that if we have an impossible, as in the
present case, with the form aa' and a positive predicate b, b must
be due to a alone. Of several proofs which may be given, take
the following. The conception a' as merely negative is wholly
indeterminate ; it includes all kinds of conceptions entirely
irrelevant to one another, the infinity of what is not-a. Thus
its content cannot be unified by any positive conception, · its
only unity is the negation of a. a' therefore cannot necessitate
b, for then it would be unified by a positive conception. Similarly
a' cannot be resolved into the association of two positives
(e. g. a'= xy), nor of a positive and a negative (e. g a'= xy'),
so that it might include a positive element from which it could
be derived, or from which, m conJunction with a, b could be
derived. Thus, if aa' 1s b, all a must be b, and b be due to the
positive clement a alone Now whatever a and b may be,
aa' = o = bb' according to this calculus ; thus aa' is b, and
therefore, by the above, all a 1s b, whatever a and b may be
Thus the equation of all impossibles to one another necessitates
the equation of all reals to one another.
§ 394. The proof parallel to that derived from x 2 + I = o, that
the equat10n of an imposs1bihty to zero necessi-
tates the equat10n of a reality to zero, requires
Y
,,
the use of the symbol ~, which must now be

C
x X
z explained. - (accordmg to mathematical an-
- - - ___ .vz alogy) repre?ents the mverse operation of xy.
It 1s explained most simply thus :
Let sy = x, then s 1s a class such that when restricted by y it
reduces to x. If, then, we write z is ~, we get this mcamng for
:Y
,,
~, that it represents any class which being restricted by y reduces
to x.
Thus, if we have given that z is ~, 1t follows that zy = x, in
y
fact we may multiply as in algebra. Observe, however, that
such an equation may be satisfied by more classes than a single
 Symbolic Logic 655
given one as s, e. g. by w as in the figure. Thus from sy = x,
we cannot infer s ... ~ but that s is ~

'Y
y
,,
one of the solutions of ~ (w, e. g.,
X
bemg another). Thus, though we can z w
yw
multiply, we cannot divide. We can,
however, infer from zy = x,
that s =v ,,
~.
Now suppose the equation given
a
ex =b
Where a, b, c are real classes and x is to be determined by the
equation. Then we have bcx = a (2)
a
and X=V- (3)
bc
Suppose also the equation given
C = b' . (4)
Then a
b' X = b from (1)
and bb'x = a • from (2)
where bb' 1s an impossible class.
In tho equation x2 + I = o, or x2 = - I, though x is unreal
we have not the equation of a real quantity to an unreal, for
the square of x 1s not unreal.
( ( - 1)l)I = - I
The 1mposs1blc operation on - I, viz. ( - I )l, is removed by its
inverse ( ra,thus leaving - I where it was, that is, ac; - r. The
same is true of bb'x =a; we have not got the equation of an
impossible to a real, for (by the above)
a
X =V bb'

so that the equation is bb' v b~' = a

where the impossible operation b~' is cancelled by its inverse

bb', leaving a on the left-hand side of the equation. Now there•
fore the equation bb' x = a arises necessarily out of the operation
 SPECIAL LOGIC
of this calculus just as xi+ 1 = o arises out of the operation of
algebra.
xis an impossible, equate it to zero by the rule of this calculus,
••. o = a. A real class a is part of unreality. And we have the
required contradiction.
§ 395. Again, with the help of this calculus, we can prove that
all classes are predicable of one another. Suppose A and C to
be two classes outside B. ' ·
Then A 1s B' and C is B'.
The combmation of A with B produces the impossible class
ABB'.
Conversely also, the class which, by combmation with B, pro-
duces the impossible class ABB' must be AB' (since A 1s not
contained in B).
But ABB' = o = CBB'.
Thus, whatever 1s necessary to produce CBB' by combination
with B is necessary to produce ABB'.
But CB' is necessary to produce CBB' in combination with
B, therefore AB' and CB' must be the 1,ame class, but AB' = A
and CB'= C.
Therefore A and C must be the same class.
Thus all A is C and all C is A.
Hence, all classes outside B coincide with one another, and
as B may be any class whatever, 1t follows that all classes are
predicable of one another.
§ 396. Now that the symbol x has been explained, one can
y
show why the simple proof m mathematics, that the identifica-
tion of 1mposs1b1hties is an idcnt1ficat1on of realities, from the
equation r-1 a = -1-=-i b, cannot be repeated Simply in the
calculus.
The analogue would be aa'b = aa'c = o.
If we could perform the inverse operation (aa')-1 on both
sides of the equation, as we do m the above algebraic equation,
we could prove that a = b.
But we can't divide by aa', all we can say is b is aa',c. That
aa
is to say, b is one solution of aa'~, and c is obviously another
aa
 Sy,,,1,olic Logic
solution, but we can't say that they coincide and therefore
that b = c.
This is perhaps a further explanation of the fact that it was
not seen that the equation of impossibles to one another would
have simJlar effects to those produced in algebra. But of course
the ddliculty will suggest itself, that the calculus must have,
nevertheless, led to right results; for, 1f it only led to contra-
dictions, it could not have made its way at all, or have imposed
on any logician or mathematician. Those who have studied it,
all know that it does get something right.
It may be replied that one department of 1t at least, which
is put forward as of undoubted truth, has been shown, if the
above arguments are correct, to be entirely wrong, i. e. the
theory of elimination from particular propositions and all that
depends upon that theory. What, then, of the mistake in the
operations analogous to those on what mathematicians call
imaginary quantities? In algebra a certain amount of truth
would be got, even Jf we treated all such quantities as zero. So
generally, m reasomng, we may have right conclusions from
wrong premisses. One case of the kmd arises where a premiss
includes a truth and a falsehood, and we argue only from the
truth, e. g. A is BC and C is D. Suppose it to be true that
A is C but not BC, then the argument all C 1s D1 A 1s BC, there-
fore A JS D, JS sound. I shall try to show that the same holds
in this equational logic.
§ 397. Consider the treatment of the ordinary syllogism all
% JS B, all A lS x,
therefore, all A is B.
AU A is x, or Ax'= o ... (1) i all x is B, or xB' = o ... (1i),
therefore, AB'x' = o and AB'x = o,
therefore, AB'x' +AB'x = o
AB' (x' +x) = o,
therefore, AB'= o , therefore, all A is B.
As compared with the syllog1sm, the process is recommended
neither by shortness nor s1mphc1ty. But 1t 1s also erroneous,
only the mistakes happen not to influence the conclusion. We
must see how this 1s. First of all, the reasoning, which corre-
sponds to the above process and is incorrectly symbolized by the
equation, is as follows :
658 SPECIAL LOGIC
Since all A is x,
Ax' is impossible.
Similarly xB' is impossible i
Since AJ&' is impossible,
AB' x' is impossible ;
Since B'x is impossible,
AB'x 1s impossible;
Now everything is either x or x',
therefore, AB' is either AB'x or AB'x'.
But, by the above, in either case AB' 1s impossible. Therefore
AB' is impossible, therefore A (1f possible) is B Even in the
reasoning as thus put there 1s a defect (with which, however,
we arc not so much concerned at this stage), for the full con-
clusion is 'A, whether possible or not, 1s B '.
Take the steps of the calculus which correspond to this reason•
ing. The equation Ax' = o 1s really founded on Ax' is impossible,
a proposition mvolvmg predication. If now Ax'= o and xB' =O
are taken accurately, as they should be m the calculus, then from
the other operatmns and equat10ns we get AB'x+AB'x' =
AB' = o. From these the only true mference 1s that either
A or B', or both, are zero ; a~. ~o m the original equations
Ax' = o and xB' = o, A or x'1-iust be zero. Thus A or B'
would be neither an impossible !:lass nor any real class (unless
o 1s formally real). The only possible escape 1s to take zero,
not as it should be in an accurate calculus but as a symbol of
impossib1bty. Then the equation Ax' = o mvolves first the
true statement, AJ&' is 1mposs1ble, a mere predication, together
with the false statement Ax' comc1dcs with every form of
impossibility. Of these only the first, the true one, 1s use~
the argument ; the second, the untrue one, 1s not used and~
causes no error.
§ 398. Next consider the step mvolved in the addit10n of
the equation AB'x+AB'x' = o and the all-important equation
AB'x+AB'x' = AB' associated with it. The equation of dicho-
tomy, a = ax+ax', mvolves, m general, the equation of the
impossible to zero, to justify it, and 1s so far incorrect. The
truth which corresponds to it is the disJunctJve statement,
a is either x or x', and there are cases where only the true part of
the equation affects the conclusion and so the conclusion is correct.
 Sy,nbolic Logic
Suppose that, whether ax and ax' are imaginary (impossible) or
not, we have given {or can prove) that they have a common
predicate, ax is y and ax' is y.
The disjunctive statement above gives at once the conclusion
that all a is y. The equation gives
a ... ax' +ax,
therefore, a = (ax' +ax)y,
and, if we take only the result, a = vy (i. e. a = some y),
which is included in and is a part of the above equation, we
get a right conclusion.
That is as much as to say that we use only the hypothetical
or disjunctive statement which the equation includes and neglect
the erroneous addition to the statement which is necessary to
make the equation. This is exactly what is done in the present
case. The equation AB' = AB'x+AB'x' is mcorrect, but as
we can attach a common predicate to AB'x+AB'x' and only
infer this of AB', the result 1s correct. The common predicate
is ' impossible ', and corresponds to y in the above.
§ 399. The result, then, 1s this :
All the equations, both those which represent the premisses
and the equation of dichotomy, involve a simple predicational
form. To this they add another element through which they
become equations. The argument we have JUSt been through
is sound and depends solely upon the predicational element.
Thus the equat1onal clement is not employed at all ; that
element 1s not only surplusage but incorrect and, if developed,
leads to contradiction and to the destruction of the cakulus
altogether The pred1cat10nal element is correct, and the argu•
ment is sound because it depends upon this alone and the
erroneous part, the equat10nal, is in no way involved. The
equatlonal form, then, is a mere sham ; to get at the truth we
must strip it away and fall back on the simple form of predica-
tion, and what we strip away is not only useless but erroneous.
§ 400. Two remarks may be made by way of conclusion.
Modem logicians (in the true sense) often suppose that, because
in the categorical universal statement, all A is B, we may not
be sure of A's existence, all A is B may be better expressed by
its supposed equivalent, if anythmg 1s A, it is B. 1 But this
1 Cf § 3go. [Part II, ch. 11 ; Part III, ch 6 ]
66o SPECIAL LOGIC
latter form does not include the statement, if anything unreal
is A. it is B, whereas all A is B does include this. In mathe-
matics the importance of the ddference between the categorical
and hypothetical statements is essential and is obvious, and the
identification of the two is seen to be merely an error.
Finally, the statement made above, 1 that the equation of
impossibilities is connected with the equation of the necessities,
of which they are but the other side, may have led to an expecta-
tion that I should have proved that from the reduction of all
x is y to xy' = o, and all a is b to ab' = o, 1t could be shown
that the equation of the propositions x 1s y and a is b to one
another is involved. This 1s true no doubt so far, that it involves
the result that, 1f x 1s y, then a 1s b ; and, 1f a is b, then x is y ;
that is, generally, that every true statement would involve every
other. But for reasons obvious to every metaphysician this
would not supply an incontrovertible objection to the theory,
or at least not one to which no one could object.
(NoTE. Wilson left eight manuscript volumes of close
examination of the symbolic logics, viz. those of Boole and
Venn, which he had studied. Naturally a part of these volumes
is devoted to a cntic1sm of Boole's mqu1ry mto the logic of
Probab1bty. The mam pomts which are supplemental to the
preceding sections are these :
The treatment of the prmcipal equation of symbolic logic,
z = x+y
is even more useless than that of the equation of dichotomy ;
for comphc:ated processes are applied to 1t which give as their
result the propositions s 1s either x or y, x 1s s, y is s. But these
must be known before we can form the equation m question
and are the sole means by which m actual thinking we form
the equation at all.
Again, certain famous elimination a.nd expansion theorems of
Boole I involve quite fatal mistakes from either a mathematical
or a symbolical standpoint. Such a theorem is, that if any
class, however formed, be represented by /(x) 1
then /(x) = /(1)x+/(o) (1 -x). 3
• § 391. • See Laws of Tllouglll, ch s
• In Boole's system (-i-•) m equivalent to the symbol It, used above
 S,mholic Logic 661
Thus this formula gives, if applied to s = ~,
y
I o, I o,, I
--xy+-xy+-xy +-xy.
I I O 0

But if ,,
!J&
z = -, z = x+vy', z = x(y) +vy'(x').

Thus ~ = v, if the calculus is right. This important discovery

was supposed by a strange confusion of ideas to be confirmed
by the mathematical meaning of the symbol. The principle
adopted is that these coeffidents as they stand represent logical

,, ,,
classes, just as ~ 1s supposed to do. If ~ represents a class = s,
we have zy = x,
therefore, if z = 6,z · o = o.
s, then, 1s such a class that, if operated on by zero, it yields
zero. Now any class will do this Thus s may be any class.
i
The mference Venn drew from this 1s, ' stands for any logical
class whatever; that 1s, it is perfectly indefimte •. To this
Wilson of course replied that it doei, not stand doubtfully for
some or other, we do not know which, 1t stands for every one
indifferently and so for all. So he made great fun of Venn's
symbol
6,which attempts symbolically to represent a class such
that, when we take none of 1t, we yet obtain all. As usual he
tried to penetrate to the true cause of these contradictions.
They anse not because the coefficients are wrong but from
a wrong view of the coefficients themselves. The confusion
arises, he believed, from forgetting a principle in the substitution
of definite values for variables in a given function. In ,t, (xy)
we can substitute at will, 1£ x and y are independent ; but if
there is a law connecting x and y, or limiting the range of their
independence, we cannot so substitute and still retain the

,,
meaning of the function. Thus m ~, the form of the function
prescribes that all :J& is y. Then 4' (or), or q, (10), is an impossible
substitution, if 01 or 10 1s to be a class. Thus, if such a meaning
66z SPECIAL LOGIC
is to be given to the coefficient, the expansion formula cannot
be applied.
Much care was devoted by Wilson to showing that the calculus
is not in the least adapted to the solution of such difficult
formal problems as are mentioned above. On the contrary it
is painfully clumsy and long, even in the solution of fairly
simple problems. It solves problems of its own creation, but
even those can be done quicker and better without any calcu\u·s
at ali, by a method mvolvmg only Barbara and immediate
infert.nce. Wilson spent much time on the mvent1on of such
a simple method, as well as on a plulosophic toy of his own,
as he c'l.lled 1t, a fractional method , and it was characteristic
of him that he should have spent so much time trying to beat
the symuob<1ts at their own game. Hts httle book On the
Traversing of Geometrical Figures arose out of notions suggested
by these ~tudies, and so <ltd lus arduous attempts to prove
a theorem, which is connected with cltsJunctivc reasoning, the
famou,. four-colour theorem ]
 PART V
TENTATIVE INVESTIGATIONS
AND PHILOSOPHICAL CORRESPONDENCE
 I
THE NATURE OF A 'THING'

§ 401. TnE 'thing' we treat as an identity in the multiplicity

of its attributes and m the variety of the attributes which belong
to it at different times. We say that 1t is hard, heavy, round,
1mplymg that the same thing 1s these attributes ; we say that
it 1s now something and at another time something different,
and treat 1t as the same 10 the differences. The events in time
wluch thus belong to one thing arc umfied because one identical
thing 1s pre1ient 111 l hem all They do not seem to belong to
the thmg as a mere unity of themselves Nor can we say that
they belong to thc-1r umty with one another , they cannot be
thought save as belonging to somethmg other than themselves
Th,s 1& obvwus m the case of the various kmds o! bemg which
a thmg has ,1,t different limes But 1s 1t true also of the thmg
m relation to the attributes it has at one time, or with regard
to those wluch 1t always has, 1ls permanent as opposed to its
variable bemg> Is the thmg merely the unity of its permanent
attributes, and is that the umty to which the variable attributes
arc to be assigned? If tlus body 1s hard and heavy, we cannot
say that the unity of eAtcns1on, hardness and heaviness (the
body bemg extender}, hard, heavy), 1s the hardness or has the
hardness , th.1t would be absurd In our statements the thing
is said to be something that 1s, to have the kind of being to
which the attribute refers. The thmg 1s hard, not hardness.
But we can't say that this hardness and extension .a.nd heaviness
[§§ 401-59 These 'lettmM are a very imperfect attempt of mme to present
vanous important mvestigat1ons of Wilson's as fairly and as clearly as I could
on the basis of manuscript notes §§ 406-:zx and 451--9 are part of an essay
wluch he was working upon for the Bnttsh Academy, called 'A Grammar
of Universals' §§ 412 32 a.re cntlcal of Mr, Bertrand Russell's logical views
The cntic1sm 1s given otherwise, 1n some respects, m §§ 477-518 §§ 433-7 are
a cntic1sm of a chapter 1n F H Bradley's Appearance and Reality. §§ 438-50
are condensed from W1lson'11 ma.nuscnpt and give his views QD an mterestmg
and vexed quei.t1on m the history of logic ]
 Tlte Ntdtwe of a ' T/ti,ig'
ia hard, nor that the unity of them is hard, nor that they in their
unity are hard. Further, we say the hardness is hardness of
the thing, the extension is extension of the thing. But hardness
is not hardness of weight and extension and hardness, nor the
hardness of the umty of these three.
It may be objected that if we say the thing is hard, &c., the
words mean simply that hardness is an element in a unity of
itself and other elements, or that the unity of (this) extension
and (this) weight is also a umty of them with hardness. This
has been disproved above. Moreover, the language we use is
not naturally interpreted so. It is chosen just to express that
there 1s something definite which is extended and heavy and
also hard ; not that there are three abstract terms, or something
which has those terms (hardness, &c.), unified. Nor is it a mere
matter of language. The thmg is not a unity of the three terms
as general abstract terms ; so far as it is a unity of them it is
of this hardness, &c Now this hardness is hardness of some·
thmg ; thus each attribute necessarily implies something of
which it is attribute, and this cannot be another attribute nor
the mere umty of itself with other attributes. The attributes
arc not bemgs, like the parts of a machine, which by their
systematic umty form another being. They have no being apart
from the thing to which they are attributed. Hardness only
exists as this thmg's bemg hard. On the other hand they are
ways in which the thmg exists and the thing only is in these
ways. If we exclude its temporal variations for simplicity,
everything we can say that 1t is implies it as somehow different
from what that, which is attributed to it, is. What, then, is
that identity in the thmg which is said to have these forms of
bemg, hardness, &c, and which is the same m their difference?
If we call the identity x, any defimte answer of the form x is a
is excluded ; for 'a' will be an attribute, somethmg not the
identity which nevertheless the identity is said to be. We thus
have not merely a problem we don't know how to solve, but
the attempt to solve it necessarily involves a contradiction. We
seem in fact to ask for an answer in a certain form, whereas
the form cannot be a form of such answer.
I 402. Observe that the view which we arrived at about the
identity of the thing was not the result of any logical or meta•
666 TENTATIVE INVESTIGATIONS
physical theory, it was the necessary result of reflection on the
presuppositions of our ordinary language and thought about
the thing. That 1s why the identity (sometimes called sub-
stance} appears so early m philosophy, in Aristotle in fa.ct. We
may get help by observing what those, who suppose substance
to be unintelhg1ble, find to be intelligible and not at all difficult,
in contra'!lt with 1t No difficulty is felt about the attributes.
So much so that an answer given to the present difficulty 1s
that there 1s no such thmg as a substance apart from attributes
and that the so-called substance is nothing but the aggregate
of its attributes But the attribute is quite unintelbg1ble save
as rcfe!'1cd to the thing or identity or substance, which has it.
The reference 1s essential to 1ts nature. Hardness is hardness
of something. Thus the mtclhg1b1hty of the attribute pre-
supposes that of the substance ; though 1t 1s equally true that
subF-tancc 1s only intelligible m relation to the attributes, as
that which has them
§ 403. So far, then, 1t seems that we cannot speak of the being
of substance as apart from that of the attributes, nor conversely.
Ilemg mu._t mclude these two sides m their umty, 1s only mtel-
hg1blc as mdudmg them both We ran't s,1y without contra-
d1ction that substanc.c or attribute t!> or has being simpliciter,
we must say substance 1s something, the something being the
attribute. Bemg, then, 1s not a simple notion, substance and
attribute arc inseparable clements of it and neither of them 1s
a bemg ; Just as surface and volume are neither of them a solid,
though they are clements wlurh constitute a solid Now this
general view of bemg may seem to suit a self or mdiv1dual
mmd, but presents d1fficult1es m regard to material things. The
latter arc what seem to ordinary reflection the most obv10us
instances of thmgs, as ind1v1dual and mdcpendent unities.
Take for clearness a thmg A which 1s a continuous solid. Its
extension would have, according to the above view, to belong
to some ind1vis1ble, identical unity and could not be separated
from 1t, or belong to a different umty, any more than one
person's feeling can belong to another person Now suppose
the sohd broken mto two pieces, B and C. Each is as much
a thing as the original unbroken solid. Thus the extensions of
B and C would belong to two different umt1es, to two different
 Ths N atwe of a ' Thing'
substances, whereas those extensions belonged at first to the one
substance A, to its unity and identity. This seems impossible.
Are we, then, to suppose B and C to be still extensions of one
and the same substance? If so, all extended bodies would be
extensions of one and the same substance. Now there is a sense
in which all extended bodies are unified by some common reality,
as appears in considering the presupposition of their mutual
relations, action, interaction, &c. Yet each has its separate
umty, and the separate umty of B would be that to which the
extension (extended matter) of B belongs, and similarly for C.
The difficulty might be avoided if we could mamtain that the
really continuous solid could not be broken, any more than one
self (notwithstanding the nonsense talked about multiple per-
sonality) can be divided into two or more selves. Breaking,
then, would mean separation of the contiguous, not the con•
tmuous, and all fracture would imply the atomic constitution
of bodies, the atoms themselves being unbreakable. {Notice
that 1f we say a thing 1s extended, it must have parts outside
one another ; umty therefore cannot be arrived at by mere
climmution of extension ) Suppose now that there are perfectly
continuous sohds which can be broken. Such solids, as unbroken,
are one by their continmty throughout, and such absolute con-
tmmty, not mere conttgmty, 1s presupposed m any account of
matter. Those theories which reduce an apparently contmuous
solid to a system of atoms assume the absolute contmuity of
the small volume of the atom, and the idea of extended matter
itself necessitates this assumpt10n. For, if possible, suppose
a piece of matter continually divided until pieces are arrived at
which are not contmuous ; this would mean that no part of
matter has a finite extension though it has position. That
would be the exploded "absurdity of makmg matter consist of
pomts and givmg a point a separate existence. A pomt can
only exist as a boundary of a continuous line, a line of a surface,
and a surface of a contmuous solid or of continuous matter.
A contmuous piece of matter, then, is one by its absolute con-
tinuity ; as Aristotle would express it, it is one thing m actuality
and, as capable of being broken, potentially several things.
Consider now a part of the thmg which might be broken off
but is not. Is this as little a unity with the remainder as if it
2773 2 s
668 TENTATIVE INVESTIGATIONS
were broken off actually? The answer must be, No I because
it exists only as a part of a continuous whole. It is unified
with any other part because each part exists, not separately,
but only as a part of the continuous whole of which the other
is a part. Imagine a geometrical plane within such an absolutely
continuous solid, d1V1dmg it mto two parts geometrically. The
matter on either side of the plane is not contiguous or adJacent
to, but continuous with the matter on the other side Thus the
unity of the one thing, which is one piece of solid matter, is
the umty of the continuity of its parts, the parts having no
separate ex1c:;tence Thus the extended thing is not mere exten-
sion , somcthmg 1s extended, and that something we call
matter ; the continuously extended piece of matter 1s one thing.
The question, then, what this 1s which is extended and not
mere extension, wh1d1 we call matter, is one more of those
unreal questions winch, m philosophy, have no answer Just
becaui;e they are not real questions That which we name
matter is sui generis, and the question asked about it is false
because 1t presuppol'les, probably unconsr1ously, that an explana-
t10n is wantt:d of 1t m terms of somethmg else But of course
1t 1s not anything else ; we might Just as profitably ask about
sensation what it is that is felt as sensation I
§ 404 G1v!.'n now one contmuous sohd A, consider how as one
it 1s said to be so and so, to have attributes Of A we certainly
say 1t is heavy , the pronoun 1t referrmg to a unity. Arc we
Justified m saying that as one 1t 1s heavy? The only possible
objection would seem to be that 1t has not one weight as one
but that that wluch we call its weight 1s the sum of the weights
of its parts, each of which has weight. But the same would
apply to each of its parts, and so the weight of any one part
would not be one of the ultimate constituents of the aggregate
Thus no fimtc extended matter, however small, could have
a weight and the ultimate units would have to be the weight
of pomts. That the weight of the various parts of the con-
tinuous solid are unified in one weight, as the parts of a divided
aggregate are not, is therefore obvious. The weight of a part
can never act independently of the weight of any other part.
Its effect can never appear except as a part of the effect of the
whole weight, owing to the continuous connexion of the parts.
 The N atuYe of a ' Thing' 66g
Thus, what makes the piece of matter one thing makes it have
one weight. On the other hand, it is not true that the weight
of a piece of matter, which is part of an aggregate or heap of
pieces, can only have an effect which is a part of the whole
weight of the aggregate. The one solid, one as continuous, has
one weight in the truest sense, because the weights of the parts
only exist as weights of the parts of the whole and therefore
only as parts of the weight of the whole. The same argument
holds for other attributes of a continuous solid.
§ 405. The case of a flmd causes considerable difficulties. In
mathematical physics a fluid 1s treated as a system of particles
or atoms. If these atoms are considered as solid the argument
of the previous section would apply to them. But the theory of
fluids is not clear and defimtely thought out. The continuity
of the flmd 1s repeatedly spoken of, yet on such an atomistic
theory continuity could at most mean contiguity, and not even
that literally, because the atoms are not supposed to be of
a shape to fill space continuously. Thus nothmg can be made
of the mathematical theory as 1t stands, because the equations
of hydro-mechamcs are not compatible with any atomistic
hypothesis but presuppose that the fluid 1s absolutely con·
tmuous Now such contmuity must either mean that the fluid
has no parts actually separate and merely contiguous, or that
it has parts separate and actually contiguous, these parts them-
selves having parts not separate, so that these parts are really
extended continuously in three dimensions, or thirdly that there
1s only contiguity of separate parts, all parts being really separate
from all others. The hypothesis of a perfect fluid can be shown
to imply necessarily that no force 1s necessary to separate the
parts. The third alternative 1s therefore the only possible one,
and that 1s absurd. Thus a perfect fluid as continuous is impos-
sible, and yet the mathematical equations of fluid motion are
compatible only with perfect continuity.

82
 II
THE UNIVERSn.L AND ITS DIFFERENTIATIONS
§ 406. AccORDING to the natural use of words, universal would
mean that wluch 1s 1dent1cal m a number of particulars and is
thereby d1stmgmshed from the d1vers1ty m which 1t 1s identical.
Nor 1s it merely identical but something defimte which is
1dent1cal m them I have called this distmctive character the
quality or the characteristic bcmg of the umversal. The uni•
vers.11 so understood as genus takes of its own nature specific
forms or species ; lmcanty is either rectilmcanty or curvi-
hneanty, colour 1s blue, red, &c. It 1s identical m the universals
which arc· its !>pec1cs, they ,ue nothmg outside 1t but belong to
its own nature, c1ddncss and evenness arc included m the nature
of number The aggregate of the specH's belongs to the nature
of the geuu,; and might seem to constitute its full nature, so
Aristotle sclys the genus 1c; nothmg but the species " This 1s,
however, 1mposs1blc, for the genus 1s identical m its forms and
their aggregate cannot be somct hmg 1dent1cal m each of them.
We cannot say the gcnu'l 1'l each of its specific forms, or any or
all of them together, becc1use zs m such a statement naturally
means '1s 1dcnt1cd.l with'
Agam the umvC'rsal c,mnot be except as an identity in the
particulars Of 1t& own nature 1t necessitates the particulars as
that in which 1t must be identical The particulars belong to
its own nature and arc nothmg apart from 1t. Yet here again
we cannot, any more than m the case of the umversals which
are species, identify any one of them or their aggregate with
the universal. The complete reality of the system of particulars
has d1stmgu1shablc within it the umversal and its particulariza-
tion. The complete system of particulars may be mfimte and
[•Tlus seems to refer to a tentative remark at Met 998b 28 For Anstotle's
real view see Physics, 210• 17-19, and ' Further the members of the definition
which 1s explanatory of the md1v1dual are parts of a whole Thus the genus
18 part of the speCles and m another sense the species of the genus • Mel.
10.a3b 23]
 The Unive,saJ and its Differentiations 671
may not be realized all at one time, but in any particular
whatever the universal and its particularization are united.
§ 407. Thus the class is not accurately called the complete
reality of the universality, but rather the complete reality which
the nature of the universal necessitates. The saying that there
is nothing in the species but the genus is not accurate in form.
It is only true so far as the species 1s necessitated by the genus's
own nature A truer statement would be that there is nothing
m the specific universal (e. g evenness and oddness) except the
differentiation wluch belongs to the nature of the generic
umversal (e. g number)
Observe, however, that 1£ the nature of A necessitates B,
A and B are by 1mplicat10n different. A cannot be identified
with B, and it may be that B cannot be called a part of A,
though necess1tatmg B may be called a part of A's nature. For
example, though surfaC"cness necessitates volume, we cannot say
that there 1s nothmg m volume but surfaceness The latter,
however, does not offer an adequate analogy to the relation we
are considering. The necessitation through which the particular
appears as havmg its nature compnscd m that of the umversal
has no analogy , 1t 1~ a umque relat 10n between necessitatmg
and necessitated. We cannot say anythmg of the particular
(mdiv1dual) wluch is not universal, except the mere 'this' and
'that'. On the other hand, not to rest the matter on an appeal
to language (important though such an appeal is), what we
apprehend m the particulars 1s the umversal which is their
d1stmctive kmd of bemg and which is m md1v1duals of absolutely
the same umversal , besides that universal we apprehend only the
one and the other of the same kmd.
§ 408. But to return to the umty of the genus in the species.
The difference here between the species cannot be reduced to
a mere this and that. Straightness 1s a kmd of being, not a mere
this lineanty ; straightness is a form of lmeanty, but lmearity
1s not a form of straightness. They are mseparably united in
the same real individual, but not in the way m which, for
instance, surface and volume are united. These two are equally
determinate, linearity and straightness are not, the latter being
more determinate. Thus the umque idea of determination is
necessary to characterize what is meant. The genus is identical
 TENTATIVE INVESTIGATIONS
in its determinations, but not identical with them. We under•
stand this by examples and can say nothing truer or better of
the specific universals than that they are determinations of the
genus, for to say they are the forms it takes is to employ a mere
metaphor If, now, we include under determination differentia·
tion of species and part1culanzatlon in md1viduals1 the deter·
mmation of the umvcrsal 1s completed m individuals, which
themselves mvolve its determmation m species. ·
§ 409. The term nccess1tat10n 1s not adequate to express the
relat10n of the umversal to the manifold of species and of
individuals. For v.hile it 1s true that hneanty necessitates its
kmds, 1t is also true that straightness presupposes and so neces•
sitates hneanty, and through the latter necessitates curvi-
hnearity So a particular lme necessitates lmeanty and through
that all particular Imes. On the other hand, the whole bemg
of hncarity is not to be a kmd of surfaceness, though these two
necessitate one another Thus the complete reality relative to
any umversal 1<; nee css1tated either by that umversal or by any
one of its species or of its m<liv1cluals What has to be taken
mto account 1s the particular km<l of neccss1tat1on The whole
bemg of curv1lmcanty 1s, m one sense, comprised m that of
the umveri;al lmeanty, m the sense that its whole bemg 1s to
bt• a kmd of lmeanty; but, m another, 1s not, m the sense of
bemg a part of 1t 01 an clement m it , otherwise as lmeanty 1s
1dent 1cal m its two species, so must the latter be Part of,
element of, do not express the relation, whereas kmd of, species
of, differentiation of, <lo express 1t, for 1t 1s a relation sui genens.
So the whole bemg of a p.irticular hne 1s to be a particular, not
a part, of lmeanty. Again the term 'comprise' 1s itself too
general, for the bemg of lmeanty c..:m be said to be comprised
m that of surfac<'ness. Nor c.in we use include m its natural
sense, for though the genus seems to mclude the species and
the particulars, v.ha.t 1s mcluded 111 1ts nature seems necessarily
to be m 1t wherever 1t 1s, and so, 1f we said the genus 'includes'
them, we should agam have one species m another species and
one individual m another. The point 1s that the species have
not only got something identical m them but that whatever else
they have somehow belongs to this identical somethmg. Red-
ness, m fact, agrees with blueness in that in which it differs
 The Universal antl its Di,fferentiations 673
from blueness. It has colourness in common with it, but it
also differs from it in colour and in nothing else. Colourness
inevitably covers the whole nature of redness and of blueness.
We must therefore examine the language in which we express
th.ts relation, to see what it is which makes us feel that somehow
the nature of the generic universal {for instance, colour) covers
and includes the whole nature of the species.
§ 4rn. In ordinary unphilosoph1cal language we say a lme
must be either straight or curved. This really expresses a con-
nexion of universals. Can we put this technically in the dis-
Junctlve form, lmeanty is l'ither straightness or curvilinearity?
Now 1f we say A is either B or C, when A is not a universal,
we mean that A is B or that A is C. This agam would be true
if A is B and not C, or C and not B, or if A is B and A is C.
But we cannot substitute linearity for A and the specific uni-
versals for Band C in these statements. We cannot, for mstance,
say lmeanty is rectllmeanty and lmeanty is curvilmeanty. The
disJunctive statement, then, is either quite wrong or at least
misleading, for 1t cannot be mterpreted as the verbal formula
of which it is an example is usually interpreted. It is, however,
true that any particular lmearity, the lmearity of any given line,
is either rectihnearity or curvilinearity, where these two terms
do not stand for universals, so the sentence means that any
particulanzat10n of linearity is a particularization of rectl•
lmeanty or of curvilinearity. Here the disjunctive has a normal
signification. But we can't say that linearity as universal
{linearity m general as the phrase goes) 1s rectdmeanty. How,
then, can we accurately express that there is nothing but
colourness m blueness, blue is all of 1t colour, there is nothing
but hneanty m rect1lineanty? A particular lmeanty (this line)
is a particular rectlhneanty and conversely. This hne, then,
has nothmg m it but lmcanty m the sense that it is entirely
a part1cular1zation of lmeanty and has 1ts whole bemg com-
prised m and constituted by that, yet not every particulariza-
tion of lmeanty 1s a part1culanzation of rectdinearity. But the
relation of the particularizat10ns depends upon the relation
between. the umversals themselves, so the real question is about
this relation. It is not enough to say the particularization of
one is a particularization of the other, for this does not show
 TENTATIVE INVESTIGATIONS
which is the generic universal, and is compatible with neither
being genus, as in the case of tnlateralness and triangularity.
§ 41 I. The relation of course is that of differentiation. The
whole nature of blueness is constituted by its be10g a species
of colourness, not m bemg colourness. Colourness as such,
viz of itself, must have for its differentiations blueness, &c.
It 1s the peculiar nature of differentiation, umque and only
10telhg1ble m examples, which makes us say that there is
nothmg hut colour 10 blueness or that colour takes the forms
of bluene,c;, redness, &c. We may solve the difficulty by
reference to the analogy of a body which preserves 1ts umty,
though 1t 1s now 10 one place, now m another. The body as
existent at different places has a variable being and yet 1s
1dent1cal Bemg somethmg here and somethmg there arc not
something <lifferent from its bcmg though associated with 1t, they
belong to its bemg Remammg completely identical the body
is thus d1ficrcnt at different times Yet by an argument similar
to the above about genus and species we could prove that if
B 1s 111 n•~ bemg here and m B's bemg there, B's bcmg here
must be m H's bcmg there, or that B m bemg there must also
be here As we h.1ve Sl'cn, its complete bemg is the complete
series of what it 1s at all different times, what 1t was, 1s, and
will Le ; winch mcludes its identical permanent being In this
1dcnt1cal bemg 1s mcluded its quality of havmg the series of
phases of variable bemg , so that, m this sense, the whole of
its variable bemg 1s comprised m the identical bcmg Suppos10g
1t to rcmam, for the sake of s1mphc1ty, unchanged m volume
and shape, 1ts identical bemg, what 1t 1s m all times and places,
1s the universal volume and shape together with the law accord10g
to which 1t moves The variable bemg 1s not the identical bemg
nor a part of 1t, yet 1s a part of the bemg of the body and 1s,
wlule not a part or constituent element of the identical bemg, still
comprised m the above sense m the identical bemg. The diverse
forms of bemg are not somethmg necessitated by the thmg's
being, as line 1s by surface (where line 1s not what surface is), but
are mcluded 10 its being ; they are true parts of 1ts being though
not of its identical bemg, wholly comprised m the bemg of the
identical thmg, for they are what 1t 1s, and there 1s nothing in
them except the bcmg of B.
 The UnifJwsal and its Dijfwentiations 675
§ 412. It is important to observe that we do not identify the
thing which is identical with its complete being. For the thing
exists now but has not yet its future being, If we did this, we
should make the past and future present. Potentiality is not
able to express this relation, because that refers to the future,
and we seem to have no word for the relation of the present
actuality to the past actuality. So 1f we ask how the universal
which 1s an identity in present particulars can also be an identity
m particulars which have ceased to be and which have yet to
be, we have an exact parallel in the existence of the individual
thmg. A movmg body 1s now somethmg which, rcmammg the
same, will be hereafter m a position not yet reached, a state
which has not yet come into existence. Similarly it 1s the same
universal which 1s m these present particulars and will be in
others not yet existent.
§ 413. How, then, this umty m d1vers1ty 1s possible we under-
stand by the 1llustratlon, and we can only understand it in this
way. Usmg language parallel to that about body, we can say
that the genenc universal (colourness, lmeanty) 1s an identity in
diverse forms of bcmg (blueness, &c, straightness, &c.), so that
we arc nght to say lmcanty is rectdmcanty or curvilinearity
but not simply lmeanty 1s rcctilmeanty. The former expression
1s perm1ss1ble because the or shows what is meant We cannot
say ' lmeanty 1s rectilmeanty ', for that would express either
identity or that rectilmearity was wider than lmeanty. ' Recti-
lineanty is lmeanty ' is also unnatural. ' Rectilmearity is a form
of lmeanty ' is too metaphorical, but we may say ' is a kmd of ',
for this expresses the true relation and, though kmd may have
been in its origm metaphorical, it has ceased to be so. The
species, then, are parts of the whole bemg of the genus but not
parts or elements of its 1dentlcal being,-that 1s, what it is in
the species 1dentlcally. They are wholly comprised m the being
of the genus, for they are what 1t 1s and must be ; they are
also comprised m its identical being in the special sense that
the identical being 1s such as to necessitate these kinds of being,
these differentiations, for itself; not comprised, however, as
constituent elements of that identical being Further, the genus
universal is not to be identified with the totality of its bemg as
the species, for it is that which 1s identical m all of them ; nor
 TENTATIVE INVESTIGATIONS
is it the potentiality of its realized species but only of those not
yet realized. It is also true that there 1s nothing m the specific
universals except the being of the genus Colourness covers the
whole nature of blueness and redness , blueness differs from
redness only m colourness
§ 414. A s1m1lar ac1.-vunt 1s to be given of the mdividual. Its
bcmg 1s entirely comprised m the bemg of the umversal ; it is
a part of that whole bemg, not a part of its identical being, a11d
the sum of the md1v1duals, with the universal which 1s identical
m them, 1s the complete bemg of the umversal but not to be
idcTJ.tifteu with the identical bemg of that universal. In any
part of the whole set of particulars the bemg of the universal
1s incomplete but 1ts 1dcnt1cal bemg 1s not incomplete. This we
can only understand from mstanccs, Just as 1t was only by
inc;tances that we understood the 1dcnt1cal bemg of body in 1ts
va.rymg positions m space. It 1s important to observe what the
idea of this 1denhcal being leads to Tlus bemg, which it 1s,
is not any temporal bemg and cannot be identified with any
temporal being AE> we saw in the casr of the broken body,
tht: parts as ex1E.tcnt at various times m the past and as existent
now must have the same non-temporal character And 1t 1s the
same with the identity of the umvcrsal.
 III
MODERN FALLACIES ABOUT THE UNIVERSAL
§ 415. WE will now examine certain modern fallacies which
have arisen from a failure to understand the true nature of the
umversal. And first, that particularity is itself a universal.
Universals arc d1stmgu1shed from one another by their d1s-
tmct1ve characteristics or characteristic bemg. But this, which
is their quality, also d1stmgmshes their particulars from the
particulars of other universals Besides bemg this or that
particular, besides mere particularity, two particulars of the
same umversal have no other distmction, unless the umversal 1s
diversely differentiated m them, so that they are of different
species Universals, then, cannot differ from one another in
the manner m which particulars do, 1. e. as particulars. If two
universals, Mness and Nness, be particulars of Lness, their dis-
tmchve quality is Lness and Lness only. Now 1f they are not
related as different kinds of Lness, they are 1dent1cal m quality,
viz. Lncss, and only d1ffer as this or that Lness. They are
identical m cverythmg universal m them, differ m nothmg which
1s universal. And they are also presumed to be universals. But
two universals which differ m nothmg which 1s uruvcrsal do not
differ at all They arc umversals not d1ffermg m characteristic
being and so not d1ffermg m the only way m which umversals
can differ. If we particularize Mness and Nness we shall simply
get mdividual cases of Lness, For Mness is this Lness and
Nness 1s that Lness, and the only umvcrsal to be particularized
is Lness. The class of M's could only appear as so many L's,
and the same would be true of the class of N's. Thus the
d1stmct1ve character of a universal cannot be the distinctive
character or quahty of another universal, unless that other
umversal 1s a species of itself (or the genus which mcludes
itself). If, then, Mness and Nness do not so differ but are
different only as thts Lness and that Lness, they differ in their
particularity, that is, by something which cannot be identical
 TENTATIVE INVESTIGATIONS
in different things and therefore cannot be a universal. We
may express this obvious truth by the formula ' particularity is
not a universal '.
f 416. The fallacy may be illustrated very simply by applying
it to md1v1duals Let A1, A1, A3, &c., be each an A. Then
Aness is common to Lhcm and each 1s a particular of Aness.
Now since each 1s a particular of Aness, bemg particulars of
Anes~ would appear to be a new universal of which they are
md1v1dual instances But that this new universal 1s common
to th~m is only mtelhgible as meaning that Aness is common
to them .i.s the universal of which they arc particulars. What
1s common to them, then, so far from bemg 'particularity of
Aness' reduces to Aness, and the false umversal, which 1s merely
verbal, disappears Moreover, 1f the process by which the false
umversal was obtained were continued we should clearly have
an infimtc series of such universals In fact, when we say
particularity 1s common to this particular and to certain others,
we mean that havmg a umversal common to 1t and others 1s
common to ccrtam others and to 1t
However, the umversal common to all particulars might be
said to be their bcmg particulars of some universal or other,
not necessarily the same. Now 1f we say every A 1s either
B or C, we might infer that 1t 1s common to every A to be
either B or C and so 'bemg either B or C' might be represented
verbally as a umversal common to every A. But this 1s absurd.
Each of these men is either a good man or a criminal gives
a common verbal statement for each, but 1t does not assign
anythmg m common to them, nor 1s 1t intended to do so. A is
either B or C only gives something in common to them, 1f D
1s the umverScJ.l of B and C, and then the proper statement is
that every A 1s D. Consequently the statement that each
particular 1s a particular of some universal or other and there-
fore shares in particularity does not state anything common to
particulars If, however, 1t is said that every umversal is
a species of some common universal, and we then mfer from
this that all particulars are particulars of one umversal, this 1s
exposed to the previous criticism of the representation of the
particularity of a given universal as itself a umversal.
§ 417. A similar mquiry will show that universality is not
 Modem Fallacies about the Universal 679
a true universal. A universal (redness) is a definite kind of
being identical in particulars. A universal is a particular kind
of universal being, and this is what is meant by saying that
redness is a universal. It is a species of universal being, which
is the genus of all other universals, unifying them as its species,
not as a universal in particulars. A false universal can be
constructed as before and be framed in the words ' being a species
of universal bemgness '. 'Redness is a universal' 1s not equivalent
indeed to ' redness 1s a universal being ' ; redness 1s a kmd of
universal bemg, a differentiation, not a particularization of
universal being. This last sentence 1s m fact a concise refutation
of the whole fallacy
§ 418. S1mdarly speciesncss is a false universal. This would
be to say that because AXness, BXness, &c , are species of
Xness, the quality of being a species of Xness, viz (species of
Xness)-ness, is the universal of each of the species. Here the
distinctive character of the universal bemg AXness, there is
added to this the relation to Xness, viz. bemg a species of
Xness. This, then, is treated as part of the whole quality, which
is now AXness and bemg a species of Xness But the addition
is superfluous smce it is already contained in the form A added
to Xness, which precisely means a differentiation of Xness.
Blue colour means the blue species of colourness. AXness 1s
not merely a differentiation of Xness, 1t is the special differentia-
tion which is AXness. Thus its bemg as specification of Xness
is equivalent to its being as the A specification of Xness, and
this is exactly AXness In ' AXness is a species of Xness ', the
indefinite article 1s fallac1ously treated as indicating particu-
larization whereas it indicates the differentiation of a universal.
We may see this in another way. Bemg a species of Xness is
common to AXness, BXness, &c. But what 1s common to them
is Xness, as their genus. The fallacy therefore arises by sub-
stituting for the genus which is common to them, their having
the genus common to them. Xness is common to AXness, &c.,
therefore having Xness as genus common to them 1s common
to them. The form 1s artificial and merely verbal, and only
intelligible as meaning Xness 1s their common genus.
§ 419. The general nature of the fallacies which have been
examined is now evident. They arise from the assumption that
68o TENTATIVE INVESTIGATIONS
the following statement is true, whatever A and X may be : ' If
it can be truly said that A is an X, then A is a particular of
the universal Xness.' This, however, is not true for every
meaning of A and X. In every universal, on the contrary, what
may be called the character of the universal must be distinguished
from its universality vr its community (of a special kind) with
what 1c; subordinated to 1t. The fallacies in question arise from
an unguarded generalization. In regard to ordinary universals
with a definite character common to their particulars which are
mdJVJduals, 1t can truly be said that if A 1s an X, A 1s a particular
of Xness. If now we substitute for X not the character of the
umvers.~1 but its universality (or what may be called, in a wide
sense of the word relation, the relation of the character to the
particulars), or agam the fact of its part1culanzatton or some
substitution of the rcl..1.llon of d1ffercntiat1on (specification) for
the rhar..1.cter d1fferent1ated, we get the false universals already
examined. Thus A is an X, therefore A 1s a 'particular of
Xncss '.
Now the second statement 1s itself of the same form as the
first ; 1t follows, then, that A 1s a particular of · bemg a particular
of Xness ', whence there results an mfimte regress, for we can
obv10usly continue the unreal process of substitution Agam,
smce A 1s a part1rular of Xness and A 1s a particular of ' bemg
a particular of Xncss ', we have either that 'bcmg a particular
of Xness ' 1s a species of Xness, wluch 1s unposs1ble, or Xness is
identified with 'bemg a particular of Xness ', which 1s nonsense.
Indeed Xness 1c; identified with each of an mfimte series of
universals
§ 420 It 1s important, m view of what has been said, to
formulate the different wayc; m which one universal may neces-
sitate another, for otherwise the phrase 'necessitate' may be
used as 1f it had one and only one meaning
In the first place, the relation properly understood 1s always
reciprocal One universal may so necessitate another that all
the bemg of one not only mvolves but is the being of the other.
Thus every trilateral dosed figure 1s triangular and every tri-
angular closed figure 1s trilateral. If Xness thus mvolves Yness,
every X 1s a Y and every Yan X. Here it is not quite correct to
say that every form of the one 1s a form of the other. Equilateral
 ModMn Fallacies about the Universal 681
trilateralness (X1) is not identical with a form of triangularity
(Yness), but any particularization of the form trilateralness
(e.g. X 1, equiangular trilateralness) is a particularization of
the form tnangularity (e.g. Y1 , equiangular triangularity).
Thus it is not enough to say that the eXJstence of one umversal
necessitates that of the other (as the eXJstence of surface neces•
sitates that of line), but we must say that the existence of one
umversal so necessitates the existence of the other that every
particularization of the one must be a particularization of the
other. The necessitation bcmg of thts nature, all the X's are
particulars of Yness, and the members of Yness are mdividuals
and not classes such as the X class. The umversals are co-
extensive, but one 1s not the umversal of something of which
the particular of the other must be an attribute ; the universals
are not related hke hncanty and surfaceness. This is the second
kmd of necess1tat1on which we are now to consider Here Aness
necessitates Bness m such a way that the identical being of
Aness has Bness mseparable from it, in its identical being. For
example, surfaceness and linearity necessitate one another in
their identical nature, but every part1culanzat1on of surface is
not a part of lmearity m the sense above cxplamed. Aness and
Bness reciprocally condition one another, but each 1s not on
that account an 1dent1ty m all that m which the other is an
identity. All bemg of surface mvolves bemg of lme, but all
being of surface 1s not all bemg of hI\c, If we consider a particular
lme, this line is an attribute of this surface (as its boundary).
In this way the particulars x, &c (given Imes), of one, Xness
(linearity), are attributes or elements m the particulars y, &c.
(given surfaces), of the other, Yness (surfaceness), and m this
way X 1s attribute to Y.
Here the total of the particulars of Xness do not form one
particular or class of Yness, but are elements or attributes in
the particulars of Yness, each m each Thus Yness cannot have
the class of X's as its particular.
Thirdly, two umversals may necessitate one another as attri-
butes of the same subJect, for example, the weight and solidity
of a body, but neither weight nor solidity 1s an attribute of the
other. Here the particulars of neither are particulars of the
other, nor are the total of the particulars of either a particular
682 TENTATIVE INVESTIGATIONS
of the other. The universals, then, of such subjects with attri•
butes (the subjects being heavy solid bodies) have the sams
particulars. Thus Yness {bodiness) cannot have the class of
heavinesses or the class of solidities as its particulars.
(Fourthly, of two universals one may necessitate the other as
the genus umversal necessitates the species universal. Here the
class of X's are as one a member of the class of which Y is the
quality Thus any particular X has its nature determined by
the nature of Yness, and as Yness is not identical with Xness,
thr nature of Yness must necessitate that of Xncss m the well•
understood relation of genus to species. Thus the x's are all
particulars of Yncss, and therefore Yness cannot have for a
member the class of Xncss itself Xness m fact is not a particular
member of the genus class as its own members arc members of
it, but Xnt-ss ic; a part of the genus class ; its members, not
itself, bemg members of the given class Yness] •
§ 421. Lastly we may rons1cler umversals which are presumed
to have no common gcnuc;, that is, umvcrsals wluch are them-
selves summa genera An attempt might be made to umfy them
under their common predicate umversaltty. Now we have seen
that any umvl'rc,al to be umfied must be a umversal whose
particulars are md1v1duals. If, then, umversahty be supposed
to umfy these summa genera, the particulars of umversahty must
be individuals, smcc we have seen that the particulars of the
species are particulars of the genus But the only particulars
umvcrsahty can have must be universals. We therefore reach
the rontrad1ct10n that the particulars of umversahty, which are
by hypothesis universals, arc m fact the rndividuals which com-
pose the whole of the summa genera in question. Thus uni-
versality cannot be rd.ited to universals as their genus.
Thest' considerations seem to pomt to the conclus10ns that no
universal can be a particular of another umversal, that the
quality of bemg a species (spec1esness) 1s not a true universal,
and that particularity or particularness is also not a true
umvcrsal.
la This paragraph 1s supplied for completeness, cf § 423]
 IV
CLASSIFICATION OF CLASSES

§ 422. THESE considerations may now be applied to the

questions and puzzles about the classification of classes to which
they provide the key. We shall show that in the proper accepta-
tion of the word class, there is no class of which the members
are themselves classes This statement corresponds to the
formula that there is no universal of which the particulars are
themselves umversals.
In ordinary practice, as opposed to the artificialities which
we have before us, umversals are never umfied under any
umversal except their genus. As we have seen, the ordmary
practice is correct. So also classes are ordmanly only unified
as species of a genus , that 1s, each class 1s not made a member
of any class as its own members are members of it but only
as a kmd of the generic class, its members not itself bemg members
of the generic cla<,c:; This we shall .1lso find to be the only correct,
as 1t 1s the natural, method.
§ 423 We must first define a class. The primary sense of
a class 1s that it consists of the totality of the particulars of a
universal Particulars of a umversal, we have mamtained, can
only be individuals These mdividuals are urufied and form one
class m consequence of the identity of the universal in them.
We shall endeavour to show that such a class cannot as a whole
be a member (1. e. one member) of another class, m the same
sense, cannot, that is, be a particular of a universal.
It is clear that a class in this primary sense is nothing but
the complete bemg of the corresponding universal. Suppose
two classes, that of AX's and that of BX's, species of the class
X. Smee these two are respectively the complete reality of
AXness and BXness, we know from the foregoing argument
that they can only be unified by their genus. Xness is identical
m them as 1t is in every member of them, and there is nothing
in them which 1s not comprised m the nature of Xness. Xness,
'.r
684 TENTATIVE INVESTIGATIONS
then, is not a universal of which each of them as a whole is one
member but a universal of which their members are members.
Each is a species but not a member of the genus class. A class
cannot, then, be a member of any other class. Suppose that
the class of W's could be as one a member of the class of which
Y is the quahty or characteristic. Smee the full being of the
universal Wncss 1s a particular of Yness, Wness must be neces-
sitated by the nature of Yness; or smce the manifold of Wness
as its manifold 1s to be a particular of Yness, Yness must
necessitate the nature of Wness Then any individual W has
its nature due to that of Yness. But 1t 1s also due to Wness.
But Yness 1s not identical with Wness. Therefore either the
nature of Yness necessitates that of Wness as genus and species
rcspect1vely, m which case the W's are all particulars of Yness
and Yness has not the class of Wness itself for a member, or
Yncss and Wncss rcc1procally necessitate one another. Of this
there are three cases. The first where Y and \V are related not
as solidity and surface but as triangularity 1s related to three•
s1dednec;s Each necessitates the other in such a way that
a particular of the one 1s a particular of the other.
The second 1s the case where a p,ut1rular of the one
(a given surface, for instance) 1s an clement m a particular of
the other (this sohd1ty of which this surface 1s a boundary).
If W 1s m this way attributive to Y, a particular of Wness (this
surface) 1s an attribute of or an clement 111 the reality of a
particular of Yncss. Hence the total of W's docs not form one
particular of Ynes!>, but the W's are clements m the particulars
of Yncss, each m CcLch, or attributes of the particulars of Yness,
each of each.
The tlurd case is where Yncss and Wncss necessitate one
another as attributes of one and the same subject Zness, yet
neither is an attribute of the other Weight and solidity, for
example, are attributes of corporeality (bemg of body) but
neither an attribute of the other. The particulars, then, of
neither arc particulars of the other, nor 1s the total of W's
a particular of Yness. The universals of such subjects with
attributes have the same particulars. Thus m no case can Yness
have the class or total of W's as its particular.
§ 424. But it may be objected that individuals are sometimes
 Cltnsifictmoit of Classes
classed by something identical in them which does not cover
their whole nature. Suppose, then, that a set of X's have an
attribute Y, which is an element in their nature, as weight is
in bodies. The X's which are Y might be said to form a class
unified by the fact that each member of it is Y. If Yness, then,
were the corresponding umvcrsal, its particulars would be not
the particular X's but their particular Ynesses, and these would
so far form a class in the sense already considered. The fact
is after all that the umversal which unifies particulars in them•
selves members of a class must have them for its own particulars,
and the X's which are Y arc themselves, and not only their
attributes, members of a class through their Yness, because they
are unified by the universal of being subjects of Yness, not as
bemg Ynesses but as havmg Ynesses. This universal is (subject
of Yness)-ness This, instead of excluding the rest of the nature
of an X besides its Yness, expressly relates to it, though not
expressing what that other nature consists m beyond being
subJect of attributes. In this sense it covers the whole of each
of the given X's, and that 1s the necessary condition of the
whole of each X being rallcd a member of the class. If, for
mstance, Wl' cla&s1fy a man with Tones, on account of lus
political opm10ns, we don't say a part of him 1s a member of
that class, we say 'he is a Tory'. He 1s classed therefore as
a man who 1s a Tory, that 1s1 under a universal which covers
his whole nature
§ 425. If, however, it be objected that the class is not taken
in its unity, whereas 1t 1s as one that it 1s a member of the
class umfied by the umvcrsal of being a species class of Xness
in general, the answer 1s that the class (i. e. all its members)
can only be one as the manifoldness of the unity of its universal ;
it 1s one as the completed being of its universal. It is, then, as
equivalent to the being of the universal that its members can
form a 'one '. If, then, this class is one member of a class, it
is a member as equivalent to the being of the universal AXness.
The universal of the latter class must be identical m the beings
of the universals AXness and BXness and must comprise the
whole being of each. But this universal is Xncss, and as it is
the genus universal it is not the universal of a class of which the
whole bemg of AXness or of BXness 1s a member but of a class
T2
686 TENTATIVE INVESTIGATIONS
of which each particular in the whole being of AXness and
BXness is a member This is the old fallacy of taking X-species•
ness as a universal, as 1f the way in which Xness is present in
its species were some other quality common to them. Moreover,
if we accept the falla,.1ous conclus1on, we shall be committed
agam to an mfimte regress
§ 426. The f.i.llac1es we are examining arise from neglecti.ng
the special character of the umty of the mamfold in the case
of the umvcrsal and its particulars. This special character is
the reason why the mamfold of md1v1duals forming a true class
cannot a~ one be one member of another class. This 1s best
illustrated by taking a manifold not so umfied which can be
one member of a class
An army 1s composed of md1vidual men, who m ordinary
m,age may be called its members An army 1s an organism and
each member of 1t 1s unified with the rest by the character
which defines the organism. But tlus umfymg element is not
a umversal of which he 1s a p.i.rticular. Definmg the army by
the umversal, a p.1rt1cular kmd of organism, the md1v1dual
soldier h.i.s not the quality of the umvcrsal, he 1s not such and
such an orga.msm As a member of an army he 1s defined only
through his relation to the whole and to the other members,
and the attribute assigned to him m that defimt1on 1s not
something which he has irrespective of the others. It 1s not
a umversal 1dentical in him and in those others, which he has
whether they have 1t or not, whereas the blue colour 1s blue m
itself, as we say, and not by relation to the whole of blue colour
or to any other colour An orgamzed whole is not a class.
Verbally we may make· 1t'l members members of a class by
stating of each that he 1s a member of this particular army;
but the fallacy 1s transparent, for what we state of each assigns
as their umty something which 1s not a universal of which each
is a particular. The orgamsm, as one, 1s a member of a class
whose umversal 1s the umversal of which the given orgamsm 1s
a particular
§ 427. We have shown that a mamfold of md1viduals forming
a species class cannot as one class be one member of another
class. We have also seen man example that a mamfold which
forms a umty and 1s one member of a class does not itself form
 Classificdion of Classes
a class. It remains to show when a set of individuals can form
such a unity that this unity can be a member of a class. This
will make the nature of the fallacy here investigated clearer by
showing at the same time why a manifold which does form
a class cannot as one class be a member of another class.
If a set of individuals can form such a unity, the quality of
the class umversal must be able to be attributed to them as
constituting th1s umty but not to each of them separately. Now
this happens when the individuals form an organic whole and
can happen only then. The attribute which they have as unified
is one which they have only as a whole and not each separately.
If, however, the set is only umfied as havmg an attribute
identical m each, as bemg true particulars of a umversal, they
are not an orgamc whole. The attribute which they have as
belonging to the whole of the particulars, so far as it can be
called a whole, is one which they have each separately and not
as a set or manifold. We may mdeed give them as a set a verbal
attribute and say they together make up the class of X's. But
here is no umty to which a quality is assigned ; on the contrary
the only quality assigned is Xness, which each of the given set
has separately It 1s said that we say not of one of them but
of all of them together that they make the class of all X's and
so the attempt is made to make 'constltutmg the class of all
X's' a sort of quality attributed to the whole, as somehow one
whole. But consider what the expression exactly means. How
does it differ from the statement 'each of these is X'? II it
does not differ, a quality X 1s merely assigned to each separately
and no qualtty is assigned to the whole set as constituting
a unity of wh1ch the quality is an attribute But it may be
said that the difference is this, that we do not only say that
each is X but that all of them together make up the class of
X's. But now the class of X's accurately means all the indi-
viduals that arc each of them X. Hence this last form means
that all of the set together are all of the mdiv1duals, each of
which is X But now we must be able to define the subject
of which we say it constitutes all the mdividuals. We can't
determine it as merely this set of md1viduals, we must say what
set, and obviously it can only be defined as all the individuals
which are X. Thus the statement which is supposed to assign
688 TENTATIVE INVESTIGATIONS
a quality to a whole as one is the mere tautology, all the
individuals which are X constitute all the individuals which are
X. Individuals or things therefore associated only as particulars
of the same umversal do not form a whole of which as one there
is one universal appl}mg to it and not to the individuals. This
obviously destroys the fallacy about the class of classes.
§ 428. The fallacy consists m classifying classes by their pro·
perty of being clas'3es To classify a g.roup we must be able to
state something of it as a unity, as a 'one', the attribute stated
forming the classtfymg umversal. Classness can never be such
an attril,ute, because to state of a group that it is a class is only
to state that the individuals form a certain kind of umty but
not to state anything of them as forming that umty. Even if,
m the c.ise of c:1.n organism, I state of these pieces of metal that
they are a umty because they belong to a clock, I have stated
nothing of them as a umty I state something of them as
a umty only when I_ state !>omething of the clock. But I cannot
smularly go on, after saymg that this set 1s a class, to state
anything of t lus class as one, because 1t is not the kmd of unity
which makes it a one, whereof something can be stated in
contrast with stating something of each member individually.
§ 429 We may put this generally as follows. If possible, let
the class of X'!> as one be a member of the class of Y's, of which
clas'3 Y is the um vcrsal.
Now the class of X's, as the total particularization of Xness,
is only what 1t 1s as a part1cularization of Xness, as having
Xness identical m 1t. It 1s only a member, as a unity, of the
class Y, 1. e. it 1s only one particular of Yncss, because its
particulars have Xncss 1dent1cal m them. That 1s, the particulars
of the universal Xness, conceived as forming a whole which is
a part1culariza· 10n of Y, a system in which Y 1s identical with
what it is m othc-r systems, have Y m them only because Xness
1s m each of them 1dent1cally. This necessitates that Y is
identical m each of them Y, then, must either be convertible
with X, which is contrary to hypothesis ; or Y is in X as an
clement m X's identical being, which 1s also contrary to hypo•
thesis , or, finally, Y is in X as genus m species, but this again
is contrary to hypothesis. Therefore the original hypothesis is
impoa,jbJe.
 PREDICATION PREDICATED OF ITSELF
§ 430. THE fallacy that predication can be predicated of itself
is easily disposed of. Let A is B be a given statement. It is
said that this 1s equivalent to Bness 1s predicated of A. From
this is sunilarly derived the statement that predication of A 1s
predicated of Bness, and from this that predication 1s predicated,
which means that prcd1cat1on 1s predicated of pred1cat10n, that
1s, of itself.
Observe that A is B means something objective, not the act
of pred1catmg B of A. B 1s predicated of A is a new statement
relating to the subJcct1ve act of stating B of A. I may reflect
after statmg that A 1s B that I have stated this, or (as 1t 1s said)
that I have predicated B. Thus the statement Bncss is pre-
dicated of A does not mean that predication 1s predicated of
Bnesc;, 1t 1s a new statement about the statement from which
1t is derived
But allowing, for the sake of argument, the derivation pro-
posed, the interest of the statement 'predication 1s predicated of
itself' 1s th.it 1t makes pred1cation a case of itself.
We mm,t therefore determine more precisely what the form
A 1s B means As we have seen, in our previous mquines,
B represents normally an adJect1ve, so that when A is B 1s said
to predicate Buess of A, Buess 1s the universal of an attribute
of A, named B. Thus m 'this orange is round', we predicate
roundness of this orange.
If, however, B 1s a noun form preceded by the indefinite
article, as m A 1s a B (This 1s an orange), Bness 1s the universal
of which the whole subJcct A 1s the 1nd1v1dual. This 1s true
even of such statements as this orange is a round obJect. Thus
to say that Bness is predicated of any subJect A means either
that Bness 1s the universal of an attribute of A or that it is the
universal of the whole of A.
Hence, if anything is predicated of itself, it 1s either the
6go TENTATIVE INVESTIGATIONS
universal of an attribute of itself or the universal of itself, and,
as we have seen before, neither of these is possible.
§ 431. The fallacy may, then, be exposed as follows.
(1) From A is B we get Bness is predicated of A (by hypo•
thesis). Let predicated of A be here an attributive. Then
(predicated of A)-ness is the umversal of which an attribute of
Bness, viz predicated of A, 1s the particular This may be put
generally as
(predicated of A)-ness 1s a umversal of an attribute,
and this 1s of the form X 1s a Y, which means that Yness is
a umversal of which X is a particular or case.
Thus (predicated of A)-ness is a particular of (umversal of an
attribute) •ncss.
Thus a universal is a particular, which 1s absurd.
(2) From A 1s B we get Bness 1s predicated of A (by hypo•
thesis). Let 'predicated of A' be here equivalent to 'a predicate
of A'.
This is of the form X JS a Y, which means that Yness is
a universal of which X is a particular.
Hence, as before, Bness 1s a particular of bemg a predicate
of A.
But Bncss is a universal, which is absurd.
The fact 1s, that the process we are cnt1c1zmg which interprets
Bncss 1s predicated of A to mean ' predication 1s predicated of
pred1cat10n ' does not yield the statement that predication 10
general is predicated but, at best, only that pred1cat1on of A
(a subject) 1s predicated of Bness (an attribute regarded m
abstraction) or that bemg a predicate of a subject 1s a predicate
of B, an attributive.
§ 432 The truth 1s that predication should be taken m Jts
strict and accurate sense. Then A JS B does not mean that
D or Bness JS predicated of A. Its meanmg 1s objective and it
means nothing about statement or predication. But if Jones
states that A 1s B, he may be said correctly to predicate Bness
of A and to predicate that A is B Now this 1s not the A 1s B
from which we started but a new A is B, nor does B 1s predicated
by Jones mean that predication is predicated of B, but that
B 1s predicated or stated of sometlung. We require, then, a. new
statement by Brown, which is that B 1s stated of A by Jones.
 Ptedication P,edtcated of itself 6g1
We thus get the series' A is B ', stated by Jones; 'Bis stated of
A by Jones', stated by Brown; 'being stated of A by Jones is
stated of B by Brown ', stated by Robinson, and so on. Thus
A is Bis the predication by Jones; B is predicated by Jones is
the predication by Brown of Jones; predication by Brown of pre-
dication by Jones is the predication by Robinson of Brown. In
other words, Robinson asserts that Brown says that Jones states
something. Here of course nothing is predicated of itself,
because the three d1fferent uses of predication belong to three
d1fferent persons.
 VI
RELATION AND QUALITY

§ 433. THE verbal fallacy JUSt exposed is one more of a kind

which forms the staple of certain modern writers. They arise
from the habit of using abstract terms without asking what
they would mean in a given context or bringmg them to the
touch by examples. There is a serious conf us10n of this kind
in a more distinguished writer, m the chapter entitled 'Relation
and Quality' in Mr. F". II. Bradley's Appearance and Reality.
Throughout this chapt<-r there is not one single 11lustrat1on,
though it is of the last importance that there i,hould be The
writer mdeed takes the fallacy so seriously as to regard 1t as
evidence of the necessarily illusive t.har.icttr of our experience.
It is in fact a nustake of his own of a very me:>..cusablc character
which he thus naively thrusts upon the 'appearance' of the
obJect and treats as evidence of the unsoundness of phenomena.
The fallacy is a suppm,ed mfimte regress wluch results from
relating the relat10n of two terms to the terms of the relation
themselves The puzzle may be put briefly thus · let A and B
be the terms of a relation, r the relation between A and B.
A and r are d1ffl•rent and, bcmg different, s~and in a relation,
say r1, to one another Simil,uly let r 2 be the relation between
B and r. Thus we have two new relations, r1 and r 2, and r has
itself become one of thr terms of a relation. Now r 1 bemg
different from A and r \\e similarly get two new relations, and
the process is infimtr, producmg an mfimtc series with terms
all different from one another
The presupposition of this fallacy is that 1f two somethings
are different from one another they must stand to one another
in a relation which is different from either, not identical with
nor included in the separate nature of either. In other words,
r 1s not identical with A or B nor a part of what is already
understood m A or B. So far from being always true, this
 RelatiOft rMll Quality
presupposition can be shown to be never true where A and B
are in fact properly described as related.
Take first the case in which A and B, being different from
each other, there really is a relation r between them diO:erent
from either. For instance, let r be equality in 'A equals B'.
We have to ask whether there is indeed a relation between
A and r and what it is ; for the fallacy clearly lies m the first
step, and 1t is presumed by the writer that we do not go beyond
the original A and B and r by any new process of apprehension.
Now, 1f there really be such a thmg as r1, it 1s necessary that
it should be new and that the statement 'A stands in the rela·
tton r 1 tor' should be new and not a part of the original state-
ment which gave r as the relation between A and B. The
question, then, what r 1 1s or what is the relation of A tor must
be real and mtelhg1ble. We shall find that there is, m fact, no
new statement and that the question 1s an unreal one, because
the idea of a relation cannot be applied, as 1s proposed, to
a relation and a term of that relation itself.
§ 434. The ordinary statement of a relation does not contain
the word relation itself. A is B's father So we say ordinarily
A is equal to B (where the relation between A and B 1s one of
equality of magnitude), not A stands to B ma relation of equality.
The latter form, however, gives a convement general formula.
A stands to B m a relation R, and 1£ we define the kmd of
relation, r, this is the particular of which R is the umversal.
Our question, then, is what r 1 1s which 1s the presumed relation
of A to r, where r is the relation of equality. Results can be
obtained only by considering particular instances, not by a
symbolic method. Two replies, differing in completeness, a.re
open to us. We may say that the relation of equality to A is
that it is the relation of A to B, or more accurately the kind
of relation which A has to B. Here equality is not the equality
of A to B, v1z. r, but R the universal, that 1s, equality in general.
The answer, then, is simply a statement of what kmd (viz. R)
the relation r IS. Thus we have not gone outside the nature
of r itself nor reached a new relation r 1 • If, then, we are to
give the full answer and to use all the mformation which the
question supplies, we may reply : the relation of A to B is not
R but the particular instance r, which 1s the equality of A to B.
 TENTATIVE INVESTIGATIONS
If, then, with this in view we go on to ask what is the relation
to A of ,, we can only reply that the relation to A of A's equality
to B (r) is that it is A's equality to B. Here, instead of giving
the genus R for our answer, we have r, the full being of R itself,
yet we have not got beyond the original relation r, norm any
sense beyond the ongmal statement that A stands to B m
a relation r. , .
§ 435 The question, then, what the relation of A is to A's
standmg m the relation r to B is obviously unreal, for it con•
tams everythmg necessary to its own answer ; it puts as a
question what is no question to the person askmg 1t. There is
a contrad1ct1on in the very askmg, a contradiction between
verbal form and the matter to which it 1s applied. This is
enough to show that to suppose a relation (r1) of Ator is absurd;
or (smce the very question what that relation may be involves
an absurdity) tt must be an absurdity to suppose such a relation.
§ 436 If now we try to avoid this by saymg that r 1 1s the
difference betwt--en A and r, we can get the solution simply by
askmg what the difference 1s between A and r. To this question
the only pm,sible answer 1s that A is A and not the relation of
A to B, and that the relation of A to B 1s the relation of A to
Band not A. Thus the so-called difference is a way of restating
the same facts as before and therefore a mere verbal difference.
Moreover, it would be easy to show the futility of inqumng
what the difference 1s between A and the last difference, as well
as to show that the only possible answer gives nothmg not
already contamed m A and r. The puzzle here 1s equally foolish
with that puzzle which produces an mfimte series by takmg the
difference between A and B as somethmg different from each of
them and then takmg the difference between that difference
and either A or B, and so on.
§ 437. The formula, then, which the apparent puzzle pre-
supposes bemg false m this particular case, we may next consider
cases where 1t is equally false for other thmgs than a relation
and one of its terms. Here again we must proceed by particular
concrete instances, not by a symbolic method. Consider, for
example, the surface of a volume and the volume. These are
clearly d1stmguishable, but what is the definition of a surface?
It is the boundary of a volume. What, then, 1s the relation of
 Relation anl Quality
the surface to the volume? We can only reply that it is that
the surface is its boundary. If the word relation applies at all,
clearly this must be the accurate account of the relation. Hence
in the case of the two terms A, the surface, and B, the solid,
the relation r (1f there is any relation at all of A to B), instead
of being, as the presupposition of the fallacy demands, neither
identical with A nor mcluded in what is understood in A, is
either identical with A, masmuch as the description of r (being
a boundary of the sohd) and the definition of A's being (being a
boundary of the solid) are the same, or at least something
always mcluded necessarily m what is already understood by A.
This must always be the case when A has no bemg except as
an element in the bemg of B. The same may easily be seen to
be true of the so-called relation of a subJect to its attributes,
the relation, for instance, between a body and its weight or its
motion. Thus, besides the particular case where one something
is a relation and the other a term of the relation, there are other
cases where the general formula 1s not true. There may be two
somethings d1fferent from one another, not bemg a relation and
one of its terms, which yet do not stand m a relation different
from either
 VII
CATEGORIES IN ARISTOTLE AND IN KANT

§ 438 IF our reasoning 1s correct, the universalities of the

various species of a genus are not particulars of the umversality
of the genus but kmds (or parts) of 1t. Suppose now an abstrac-
t10n be€'inning from md1vidual things till we come to so-called
summa genera, or, as they would be more correctly termed,
sumn&tM species. The universalities of all other universals would
be comprised m the universaht1es of these, not as their particulars
but as kmdc;, forms, parts, or aspects of them If, thhn, we
cons1derecl t ht>se summae species as kmds which Bemg or Reality
must tJ.ke, wh<'rc Being 1s more accurately the Bemg of Things
(that 1s, the um'versal of which Thmgs are the particulars), the
Bung of Things 1s m the pos1t10n of genus to these summae
specus an<l their several univ<'rsaht1es are all comprised in the
univers:i.hty of tlus genus In regard to tlus we must avoid the
fallacy of creating a universal of umversahties, with an infinite
regress That fallacy has been already sufficiently exposed.
There is no universal of umversahtyness of which the umver-
s.i.ht1es of universals \\Oulu be particulars, for the umversality
of one genus universal as d1st111guishcd from that of another hes
not m its !Jcmg a unity which 1s particularized but m bcmg
parllculanted m these precise md1v1duals, m bemg the par-
tlrulanzc\t 10n of its own peculiar quality or characteristic ; but
this 1s precisely itself, and its universality 1s md1stmguishable
from itself m the fullness of its bemg. Thus the class1ficat1on
of univcrsahties can only be, 1f possible at all, the classifica.t1on
of umvcrsdls, there is no universal of universalityness of which
the univcrsaht1es of universals would be particulars
§ 439 This genus, then, of the summae species 1s not mere
Being hut the umversal of the bemg of Thmgs or Substances,
and 1s therefore the generalization of Substance, Substanceness,
or Substance•m•general. Now in this system every universal
has thmgs for its particulars, and thus the universals of attribute
 Categories in Aristotle a,ul Kam 697
and relation will not appear 10 the system. Nevertheless, every
variety of being, attribute, and relation as well as substance,
must be comprtsed in the system, because all these are com•
pnsed in the existence of things. The nature of substance
mvolves in itself attributes and relations ; its universal there•
fore involves the umversal of possession of attributes and the
universal of relatedness. We have now to consider the relation
of such a umversal as attr1buteness m general or the universal of
a particular attribute such as colour(edness) to the umversal
of substance in general and to the universals of substances. We
may note m passing that we can understand why Aristotle said
that Bemg was not a genus 1f we remember that the genera
par excellence (m the Categories, for example) are m fact uni-
versals whose particulars are substances. 1
Attributeness means being an attribute of a substance, hence
the umversal of substance mvolves m its own being the bemg of
attr1buteness Nevertheless, the latter 1s not an element m the
umversal substanccness, the corrcspondmg element ts 'having·
attributes '-ncss Moreover, the latter 1s not a differentiation of
substanceness The question, then, 1s whether these umversals are
capable of bemg unified by any umty beside the unity that one in-
volves the existence of the other. They cannot be diffcrent1at1ons
of one and the same umversal because d1fferent1ations of the same
umversal must have 1dent1cal particulars with that universal.
Nor can they be particulars of the same universal, for if two
universals are particulars of the same umversal, they must be
d1fferent1at1ons of a common umversal though not of that
umversal Thus, though we can state of attrtbuteness and
substanceness that they are, Bemg 1s not a umversal of which
they are d1ffcrentlatlons, nor 1s Bcmg a universal of which they
are particulars
§ 440. Thus the form 'S 1s', which as opposed to 'S is P'
does not occur m ordmary lmgmstic usage whether in ancient
Greek or m modern languages, does not represent in its subject
particulars of a umversal which 1S Bemg or Beingness. If, then,
Aristotle had carried out fully the thought which appears to be
implicit in Chapter 5 of the Categories, with the d1stmction there
made between 'said of a particular subJect' and 'existing in
1 Cmegones, 2 11 11 seq.
6g8 TENTATIVE INVESTIGATIONS
a particular subject', the result would have been a system of
universals classified as in the section above and based on his
view that the only true independent reality is the individual
tlung (or person). The summum genus would be Thingness, and
for this the only word m his terminology that appears to be
suitable is Being (ovcrla), a word which both in the Categorus
and m Metaphysics, Book Z, 1s sometimes the name for Tlti~g
as such, or First (Primary) Being. This, however, 1s nowhere
unmistakably stated, he appears mdeed to have virtually
stopped m lus classification at certam highest genera and his
own exprcso;1on, 'the lughest of the genera', 1 favours this
interpretation.
§ 441 For his view that Bemg is not a genus, the important
pM!sages are three in the- Metaphysics with which one m the
Topic-s agrees, a passage m the De Jnte,pretatione, and one in
the Posterior Analytics.OJ
The first of the passages from the M ctaphysics does not turn
on lus view that the genera pa, excellence are secondary essences
and 1s a htllc difficult to mtcrpret It rnns as follows 'But
it is not possible for One or Being to be a genus For each
difference of a genus must have one and bcmg said of 1t (sc. when
we speak <Jf 1t), whereas 1t 1s 1mposs1ble that either the species
of the genus or the genus without the species should be said of
the differences. Thus, 1£ One or Bemg were a genus, none of
their differences could have one or bemg stated of 1t ' By
difference he here means not the differentiated universal (e.g.
rectthne,mty) but the differentia (straight, &c ). 3 This suggests
that he had not realized that, accordmg to hts own theory,
Being would be the proper name for the highest genus. The
passage from the Posterior Analytics, •to exist 1s not the being
(essence) of anythmg, since the bemg existent (that which exists)
ts not a genus', docs seem to be connected with his doctrine
that the md1v1dua.l 1s the only real existent and that the true
genera have •bemgs' for their particulars. Finally, the passage
m the De lnterpretatione runs as follows: • for not even "to
be" or "not to be" is a sign of a thmg, not even if you mention
1 Met. 998b 18
1
Mal. 998b 22, 105 3b 23, 1059b 30, Top 121• 16, De Inl 16b 22; An.
Po 92b 13
1
Cf his language tn the Topacs, 144• 37 seq
 Categories in Aristotle antl Kant 6gg
"being" by itself merely. In itself it is nothing, but it signifies
in addition a kind of compounding which cannot be understood
without the things so compounded.' This is most important
for the lmguistic point involved in our problem.
It is singular that Aristotle, who in the Metaphysics is attacking
those who made Being and Unity the essence of things, does
not adopt the seemmgly clear and dec1SJve argument based upon
his own theory of true being as the mdividual and the true
genus as that of which the individuals are the particulars. His
argument, however, and this is important, is directed to showing
that Bcmg as a universal cannot be d1ffercntiated, though he
has not realized what seems to be the direct proof required (he
gives instead a single reductio ad absurdum), viz. that if being
were the genus or class universal of the universals which we say
'are', those umversals, as we have shown, would necessarily have
the sc1.mc kmd of particulars as it and consequently the same
kind as one another. But, obviously, the universals of a sub-
stance, an attribute, a rel.itlon, cannot have the same particulars.
To reach this positive point of view Aristotle would have
required to have had before lum the pomt of view from which
a universal could be represented as a particul.ir1zat10n, not a
ddierentiat10n, of another umversal
§ 442 We miss, then, two tlungs in Aristotle's discussion.
The hne of thought wluch led lum lo say 'bcmg 1s not a genus',
at least when it took the shape it seems to have in the Posterior
Analytics, might have made him recognize that though the
umversal 'being' was not a lughest genus, yet m his own
terminology such a genus was exactly 'essence' as the being of
things or, more accurately, 'the bcmg a thmg'. How far he
was from this may be seen when he says, 1 ' Umty cannot be
a genus for the same reasons that neither being (existent) nor
essence can be.' Secondly, he might have been led to see that
though Being is not a classifying genus which umfied everything
in that way, yet there is a umfymg principle in all reality.
Instead he rested content with the negative statement that
Bemg is not such a principle, where being is that 'is' which is
universally predicable. He may have been prejudiced by his
justifiable criticism of Plato who had sought this unifying
1 Met 1053b 23.
u
100 TENTATIVE INVESTIGATIONS
principle in the Idea of goodness. Had he followed up his
thought he might have reflected that just as a genus demands
its own different1ation mto species and individuals, so by the
same inward necessity the unity of reality demands the kinds
and thing'> into which reality actually is differentiated.
f 443. Takmg his own categories, they are obviously unified
in the reality of (primary) Bemg or Substance, and he dqes
elsewhere recognize that the other categories depend on the
first. But he never put this as the unity and the real unity
winch corresponds to what is common to the statement of
'being' In fact he never cleared up his mind about 1t1 or he
could not, so to speak, have so degraded 'bemg', as he does
in the pac;sage translated from the De lnterpretatione, and have
merely left it at that What he has said there 1s true and
important, but 1t 1s m1slcadmg, as 1t stands.
Moreover, the formula 'Bemg 1s not a genus', although it
shows from one point of view an accurate ms1ght mto the nature
of class1ftcation, 1s e:i..traor<linary and m1sleadmg when considered
ifl relation to his own terminology. He must speak of the
categories ,ts categories of bcmg , this bcmg cannot be merely
the common prerl,cate of everything, if we are to take categones
literally ai. predicates. For we rannot state of this 'being•in-
gcneral' that it 1s a substance. On the other hand, if it does
mean 'being-in-general', categones would surely have to mean
species or kinds, so that bcmg would indeed be a genus and the
formula be contradicted. It 1s most natural to leave category
its proper meaning a11d then 'being' will stand for 'that which
is', not for being m general. This again, if all the categories
are asserted of 1t, a~ the formula 'categories of that which 1s'
naturally 1mphcs, could only be complete being, that which is in
the fullest sense. Now that with Aristotle is the individual
thmg Thts agrees with the fact that the categories are given
not as abstract universals in the noun form (whiteness, as an
example of quality) but in the adJectival form (white, double,
&c ). If so, what 1s meant by the categories of being (that
which is) is that each of them 1s an attribute of the true and
real being of the individual thing, the primary bemg of this
treatise. It is the thing m fact of which we can properly state
that it is a substance, tv.·o cubits long, in the market-place, &c.
 Categories in A,istotle antl Kant ,ox
The only word form which causes any difficulty in this inter•
pretation is that of the adverb of time (e.g. yesterday). This,
however, is again not put as an abstract universal but as it
would occur in a statement, and it is true (and the truth) that
as every happening belongs to a substance, so the temporal
qualification ultimately also belongs to a substance (e.g. this
substance was in the market-place yesterday). This appears,
then, to be the meaning, for it is difficult to see how anything
else could be meant.
§ 444. But now, if this is so, in the categories of being, of
' that which is ', the latter expression 1s used in the general
sense, and 'that which is' represents the universal of all particular
'beings' and so is the universal of substanceness. It is just in
relation to the distinction of the categories that the meaning
of 'being is not a genus' becomes of great importance. On the
above view 'that which is' is equivalent to substance {being)
when the categories are termed categories of that which is.
Thus 'that which is' is not only a genus of which real complete
things are the particulars (complete, that is, in the Aristotelian
sense as equivalent to things) but the highest and most compre•
hensive genus, though not the genus of the categories, including
all reality whatever. This discussion seems to confirm the
hypothesis that Aristotle did not pursue the train of thought
which his view that 'bemg' taken m one sense is not a genus
might have suggested to him.
§ 445. What, then, is the fallacy in Aristotle's proof that being
is not a genus? He contemplates two ways of stating the genus
X in regard to a given subject S. Either a species AX 1s stated
in S is an AX and therefore mediately the genus S is an X, or
the genus is stated ' without its species' 1 and we say immediately
Sis an X. In neither way, he says, can we state the genus of
the difference A. He seems to mean that odd being the dif-
ferentia of odd number, we can't say that odd is an odd number
and so a number, nor directly that odd is a number.
But is the argument free from verbal fallacy? When we state
that S is a species, we mean that S is an AX, and when we say
that the genus belongs to the species, we state X, the genus
universal of a member of the species AX, we mean an AX is
an X. If, then, we state the species of the differentia we ought
U2
 TENTATIVE INVESTIGATIONS
to mean that we state that an A is an AX. Now if the differentia
necessarily presupposes X the genus, the statement must be
true. We do not say that odd numberness is numberness, but
that an odd number is a number. So if we state odd number
of odd, we mean that an odd t hmg is an odd number, which
is necec;sanly true It looks as though Aristotle meant that
linearity cannot be predicated of straightness because straight•
ness jc; not a hm: , but neither is rcctlhneanty a lme, and thus
he would appear to have fallen mto a verbal fallacy.
§ 446. There is a danger that m appreciatmg the insight shown
by AristC\tlc 111 hi<; d1Ptum 'bemg is not a genus', we may our-
sclvl's fall mto the <;J.me one-s1dcdnrss. There are certam
charactcrn,t1cs of Being which, 1£ not identical with those of
a true genus, arc par.tlld to them m a remarkable way and
mu-,t t herl'forc not be neglected These characteristics are not
destroyc·cl l1y the discovery l hat bcmg 1s not a true genus ; only
a ccrt.1111 WJ.Y of rc-gardmg them, i c. as parttculanzattons or
d1ffcrcnt1at10n-, of a universal, 1s destroyed. An attribute or
a. nlatton may rightly bt· said to havt• bcmg as well as a sub-
st ancc, and l lus bcmg 1s not idc-nttcal with that of substance.
We can study the ,tttnbute m abstraction from the subject, as
we do 111 the ma.thcmat1cal sciences, and l lus proves that 1t is
d1stmgmsh,tble m be111g '-o of relation, 1t 1-, csc;ential to relation
that 1t should be' between' tlungs wluch have some other nature
than that of stantlmg 111 ,t given rclJ.t1on. Hence 1t 1s natural
at fir~t :.1ght to say that Hung 1s a genus with its species 'bemg
<lf attribute', &r Again we t hmk thal llcmg m its own nature
necei,s1tatcs lhc!>c forms; of 1tsdf, and t lus 1s parallel to the self•
dctcrnunatton of a um ,·cr,,,11 m its d1ffcrcnt1ations. If, then,
the forms of Being arc not species nor kmds as ordinarily
understood, nor of cou1sc particulars of a universal, how should
they be descnbcd ,tccuralcly? \Ve have to revise one of our
usual conceptions ,md either to rcfw,c to ccill these universals
or to admit umversals wluch haYe no particulars and no dif-
ferentiation, the umversahty of one quality (attribute) being
identical in kmd \\11th the umversaltly of every other. In this
d1fficu\ty we may prov1s1onally term them common principles.
They are live universals because they arc a unity in a manifold ;
universality m each universal, particularity in each particular,
 Categories in Aristotle and Kant ;03
and so on, but different from true universals because the mani-
fold is not a particularization or differentiation of the common
prmctple. This will meet the difficulty we found in regard to
classes. We say of every class that it 1s a class, yet we saw that
classness 1s not a true universal of which classes are particulars.
Class has no differentiation and no particulars We have a unity
in a manifold, but the manifold 1s not a particularization of the
umty. Instead, then, of our ordinary view we have to recognize
that there are some umversals which admit of differentiation
and parttculanzation but not of md1v1duahzatton, and that
some (which we have called common principles) <lo not admit
of either particularization or d1fferenbahon
§ 447. In Bemg, then, called provisionally a common principle,
we recogmze a umty m a manifold which is sui generis, just as
much as the umty of a true universal m 1ts particulars and its
species is sui genens. Being, hkc a true umven,al, also deter·
mines 1ts own manifoldness but in a different manner. If we
are to seek an illustration or analogy we may refer to the self,
which is an absolute unity m its different thoughts, or a body,
which ts 1dcnttcal 1n its different pos1t10ns and aspects, but
neither the self nor the body is the universal of those ddfcrcnces
nor they its particulars
§ 448. If we now consider the forms of Bcmg, the manifold
which it must assume and wluch simulates the differcntiatton
of a universal, we may perhaps find the true significance of the
phtlosoph1cal classifications called systems of categories. These
categories are obviously of a very special kmd and are philo•
sophical and not scientific. They arc, that is, though compre-
hensive, not a class1ficatton and could not be reached by
abstracting successively from the whole field of mdividual sub-
stances. In this way we should att •.un to the classifications of
the natural sciences but not to categories. Now Kant's criticism
of Aristotle's categories has shown that while there may be
agreement about some of the main c.i.tcgorics, there may be the
greatest disagreement about the real meanmg and object of
them as a whole. Kant believes that Aristotle was led to his
grouping of the categories without reahzmg its true character,
and that in consequence he did not carry it out consistently.
Whether, then, Kant's own view of the categories was right or
 TENTATIVE INVESTIGATIONS
wrong (and surely it was not right) this suggests that there may
be a certain instinctive impulse to search for categories, without
full con.;ciousness of its nature. The impulse will lead to an
arrangement of a quite peculiar kind, and reflection must then
supervene in order to understand the impulse and to correct its
imperfect work. Kant naturally makes the categories forms of
tho uni{ymg understanding, because of his dominant confusion
of the apprchens10n and the apprehended. Aristotle's tendency
s far soundt.r, for necessity of apprehension can, after all, only
mean apprehension of a necessity m the obJect.
The cxpla.nat1on, then, of systems of categories may well be
that we come in time, by reflection on the use of the verb
'to Le', lo recognize a corresponding unity of being, that the
totality of particulars m all their variety 1s a unity. Long
before we have recognized this unity in particular sets or varieties
of bemg. Then comes the plulosoph1c impulse to determine the
forms which bemg in gmeral must take, suggested by the
analogous dctcrnunation of such a umty as the section of a cone,
and so to cover exhaustively the differentiations of being in
general m the whole of existence Tl11S impulse need not be
fully aware of itself. Aristotle, for example, doesn't consciously
go to work m tlus way or, 1f he did so begm, he probably gave
it up when through his clear idea of d1fferent1ation and class1•
fication he realized that Bemg 1s not a genus Still 1t may
have been the fact that we state of everythmg that it is, which
suggested the idea of an absolutely unified system of bemg.
1449, Thus the charactern,tic of Aristotle's system 1s that it
is a sort of exham,tive attempt to cover the whole variety of
reality. 'Evcryllung 1s either substance or an affection of
substance' is its 1mphut mcanmg These are in fact his two
main categories. But he gets mto difficulties about substance,
so that his thought m the end is that everything is either
substance, or quality of substance, or quant.Jty of substance (or
of what belongs to substance), or relation of substance (or of
what belongs to it}, and so on. Now it 1s clear from our analysis
above that the impulse to determme a sort of differentiation of
Being must produce an altogether unique system. For the
general forms of Being are not true d1ffcrentiations of it, nor
are the individuals, which' arc', particulars of Being but of other
 Catego,i,s in Aristotle and Kant 705
universals. This explains the fact that the classification (as it
first seems to be) of all being could not be attained by any
abstraction from individuals or any classification of them ;
because this, the natural form which classification first takes,
proceeds to universals of which the original particulars are those
very individuals, and the successive universals are differentia-
tions of one another in succession. This systematization is by
means of differentiation and part1culanzation1 the other (the
system of the categories) is achieved by neither. Hence the
latter acqutres its uniqueness; for the ordinary scientific method,
even when carried to the utmost unity, will not bring to light
one single category except substance itself. If contmued ideally
upwards, it cannot even brmg to hght the umversal 'mere being'
of wluch the categories are the unfoldmg The same is of course
true of any classification proper which starts from the mdiv1dual,
whether thing or individual attribute Attnbutcness, for in-
stance, if regarded as the universal of individual attributes, is
neither a differentJatlon nor a particularization of being; the most
general universal which could have individual attributes for its
particulars is just attnbutencss Now substancencss, attnbute-
ness, relattonness, &c , cannot be treated as summa genera
because they would then be members of the same classification j
this they can never be, for such genera must have the same
sort of particulars, and here the particulars of the one would
be substances, the particulars of the other attnbutes.
§ 450. If, then, this is the true account of the philosophic
impulse under discussion, we see that it is a most senous and
fatal mistake to regard 1t as a classification, when once we
understand what a classification truly 1s. (That leads to the
further error of vamly attemptmg to adapt the system of cate•
gories to the ordinary classifications and to make them depart·
ments of it.) Properly understood, it is simply the just endeavour
to determine the manifoldness in which Be10g in general must
unfold itself, and Aristotle proceeded correctly when he exhibited
as categories such general forms as those of Substance, Attribute,
Relation, &c. Contenting himself, however, with pointing out
that Bemg is not a genus and therefore could not constitute the
essence of things, he seems by his merely negative attitude to
have m1ssed the true significance of his own list of categories.
 TENTATIVE lNV'ESTIGA TIONS
§§ 438-50. Translation of principal passages
referred to in the above investigation.
[Arist., Cat 2• 11. 'Substance most properly and pri~arily
and especially so called is that which is neither said of a particular
subject nor 1s m a particular subject, for instance, a particular
human bemg or a particular horse. Secondary substances are
what the species arc called, m wluch the substances primarily
so called cx1<it and be,;1des them the genera of these species; for
mstarcc, w!ui'e a particular human bcmg exists in the species
human bcmg, the genus of the species 1s ammal These, then,
arc c..illecl secondary substances, for example, both human bemg
and ammal'
JJe Int 16b 22 • For not even" to be" or" not to be" is a sign of
a t hmg, anrl not even 1f you mcnt10n "bemg" by itself merely. In
1t'lclf tt 1'l not hmg but 1t s1gmhcs, m addition, a kmd of compound-
mg wluch c..rnnot be undcr!>tood Mt houl the thmgs compounded.'
Post An yzb 13 'Agam we <;ay that it 1s necessary that
ever} tlung be· proved to cx1c;t by demonstration, unless it be
essence But to c:xu,t 1s not tlw essence of anything, smce the
bcmg existent (th.it 'l.l11rh cx,..,t~) 1s not a genus. That a thmg
exists tlwrcfon. will he ('luuJLCt of) demonstration. 'flus is what
the sciences rn fact do The gcoml.lncmn J.<1&umes what triangle
mean!>, but prow., that 1t l'Xlbl., '
Met 9981i 22 ' But 1t 1s not possible for either bemg one or
being cx1st.-nt to be a genus of e:>..1stents. For wlule 1t is
nel·c~&..i.ry that the ch!Tcrcntuc of c,1.ch geuuc; should each both
cx1"t ,tnd he one, 1t I"' 1mpo..,!:-1l>lc either for the species of a genus
to be said of the appropriate d1fTercnti..i.e or the gcnuc; (to be
said of them) \\ 1thout 1ti. "pu·1c,;; Therefore 1£ we assume bcmg
one or bcmg cx1sknt to be a genus, no differentla will either
be or be one ' 1
The same :1rgu1rn nt 1s used 111 Met 1059b 30 and Topics, 1211.
16. In Met 1053b 23 he ,;<'lys, 'being one cannot be a genus
for the s,1me n·al!ons th..1t neither being existent nor essence can be.'
Met. 1017• ::?.? '\\.hatcvcr the forms of prc<l1cat10n signify
arc said to be Cl>l>l ntl,1lly. For to be Ii.ts as many significations
as there .-uc form~ Inasmuch, then, as the predicates signify
what the subJect is, others its quality, &c., . . to be has a
&gnification cqmvalcnt to each of these (for there is no difference
betv.een "the man 1s \\alk1ng" and "the man walks").']
1 [J C W translated ' It 1s not possible for One or Bcmg to be a genus.
For each d1fierent1a of a genus must have one and bemg said of 1t (m state•
ments), whereas 1t 1s 1mposs1ble that either the species of the genus should
be IIAld of 1ts d1fferent1ae or the genus without the species So if One or
Bemg were- a gc-nus, none of thr1r d1fferc11ba(' could have one, or being, stated
of 1t ']
 VIII
UNIVERSALS. LAST WORDS

§ 45 I. We will now return to a difficulty presented by the view

we have taken of organisms m relation to their universals. We
have seen that a whole organic system is one particular of a
universal and that its members are not particulars of this
universal. We have also seen reason to think that 1£ a class is
classed, that is made a member of another class, the totality of
the members of the class must form an organism or system.
It is essential to a system to have as a whole a quality which
the members of it taken smgly have not. Now a work of art
may seem to cause a difficulty. As a whole 1t 1s beautiful but
each part of it is also beautiful. Thus each part appears to be
a particular of the same umversal as the whole system. There
must then either be some amb1gmty m the assignment of the
common quality, or elc;e the quality assigned is not ,.a. universal
of which the system as a system is the true particular. We shall
find that beauty 1s an attribute of the whole wluch is of the same
general kmd as one attribute of each member of 1t, and it would
be difficult, perhaps impossible, to find a system in which this
was not true. But the attribute as attribute of the member and
of the system as one is not the same m species and in this sense
the universals arc not the same. Moreover, the universal of the
attribute, beauty, is as such the universal not of the whole
system but of an clement in it. This system then is not its
particular. Thus the supposed exception will prove to be no
exception but confirmatory of the view already adopted.
§ 452. Consider, for example, a musical melody. It has a
beauty as one, as a system of its notes. The smgle notes may be
beautiful either considered apart from the whole or as in it. A clear
note quite by itself is beautiful and that same note in a melody
may naturally be called remarkably beautiful commg where it
does. The single note by itself is not merely beautiful, it baa
a definite quality as a sound and it is to that that we attribute
 TENTATIVE INVESTIGATIONS
the beauty. Thus beauty is a universal not of the sound but
of an element in the nature of the sound. It is the universal of
an attribute of the sound. So of the melody. Its universal is
a definite number of notes in a certain order, relation of pitch,
and with fixed relations of duration. It is an attribute of this
that it is beautiful, and the universal of that is a universal of
which only a part of the whole nature of the melody is a particu4i,r1
whereas the umvcrsal of the system is its whole nature of which
the individual note cannot be a particular. Finally, the single
notes as parts of the melody are beautiful because of the quality,
duration, &c., of the sound as related to the quality, &c., of each
of the other notes and as havmg a certain place in the whole.
The umversal of the smgle note as thus regarded is clearly an
element in the umversal of the system and not the universal of
the system itself Agam, though this is not necessary for our
present contention, we cannot accurately treat the note so
considered and the melody as two beautiful things and then
compare their bc.i.uties as bemg homogeneous. For the beauty
of the note in the system involves the beauty of the whole
system. More precisely the description of it is the description
of the beauty of the whole system as contributed to by its
relations to all the notes, not as contributed to by the beauty
of the note per se, and this involves exactly the same thing for
every other note of the whole. In fact we only call the note
beautiful as being an element in a system which is beautiful as
a system. It is therefore equivocal to speak of a beauty of its
own which can be rhstmgu1shed from that of the whole. So if
a clock keeps time, we may say equivocally that any part of it
keeps time too because it contributes by its quality and relations
to the other parts and the whole keeping time. And here the
artdiciality is obvious. The timekeeping of the clock is not
a system composed of the timekeepings of its several members,
nor is the beauty of a whole work of art composed of the beauties
of its parts as members. The beauty is not a system but the
beauty of a system.
§ 453. The same may be said of a difficulty raised about ' the
good'. The difficulty is not quite clear, but appears to be that
the universal good consists of the individual goods which are
members of the system together with the system itself as one
 Uniwrsllh-Lasl W o,ds
good. Before considering whether we have here an exception
to the above theory of universals, we may question the validity
of the idea itself, First let the good as universal be one which has
a limited number of differentiations each existing in but one
particular. The context of the argument we are criticizing
suggests that it is assumed that such a universal is contemplated
and that it must have a finite number of members. Yet we see
no limit to the differentiations of good which may be evolved
in history, and therefore, if the differentiations are assumed to
be finite, the hypothesis seems idle. As the good again is
actualized in infinite time, we can see no limit to the particulars
which are good. We must therefore assume that the hypothesis
applies to particulars supposed to be, possibly at least, infinite
and yet somehow composing a system. Are then the particulars
to be called good only as members of a system? This seems
clearly untrue ; for there are goods, one would think, which are
good without any such reference. For instance, the love of truth
is such a good. The love of truth and hatred of falsehood are
unaffected by relation to any other kind of good. Suppose it to
be said that the good meant is real and that we, if our knowledge
were complete, would see that every instance of the love of truth
is good in relation to all other particulars of good. That would
be another kind of good wluch it has and not the good, the really
impo1 tant practical good, of which we have spoken. This good
is not determined to have its quality by any such relation, The
particulars then would be good in themselves, though they would
also be good as members of the system. This leads to the other
genel'al alternative that the particular goods are each good apart
from the system but members also of a good system. The treat-
ment of this will obviously be the same m kind as that already
given in the case of a work of art and its relation to beauty as
a universal. It would follow that the good system could not
possibly be the true particular of the same universal as the goods
which are members of the system. Goodness cannot be assigned
to any particular except as attaching to a quality in that par•
ticular other than good. Thus the good particular cannot have
its nature- covered by mere goodness. As Aristotle would say,
' being something else it is good '. Indeed Aristotle in the sixth
chapter of the First Book of the Ni&omacheon Ethi&s reaches in his
710 TENTATIVE INVESTIGATIONS
criticism of Plato a point of view which is correct and which he
does not succeed m sustaming everywhere else.
Goodness is not the true universal of any good thing, it is the
universal not of the thmg but of that thing's particular goodness.
The true umversal of ruch a particular will mclude its goodness
as somctlung neccs'litated by the quality expressed in the
univen,al. Similarly truth is not, accurately speaking, tl\e
umversal of particular truths
§ 454 The idea that the members of a class may be limited in
number i:uggests the question whether there may not be a limitmg
case, to use mathematical phraseology, of a class with one member
or a umversal with one particular. Those who speak of such
a case as possible. or actual probably have not considered
whether the account they have given or implied is consistent
with anythmg less th,1.n a plurality of members m a class or of
particul,1rs of a universal. Certamly one ordinary way of
thmkmg of a um versa! as an i<lentity m a manifold is not con·
sistent with thl' 1de,1
a an IDSt,rnce could certainly be given of such a universal, the
nature of the umven,al must show m, a priori, as \\e say, that
1t could haw only one particular What umversal is there about
wlurh \\e have tlus knowledge ~ One may mdeed doubt whether
a defimtwn of Go<l coul<l b(• given m \\Inch a umversal element
should necc%1tate there bung hut om· parttcul,u, but for the
present purpose we nce<l not ask whether God, or the totality
of reality, 1" one particular of a umvers..11 which has no other
partieular, \\e must r,ithl·r com,1der whether the nature of a
umvers,1.I adnuts or c:,,,.clu<lcs rc!>tnctton to one particular. If
wr: allow, as seems re.1son,1hle, that the universal can be confined
to a lumted number, wh.\t n•a-:on is there for say1Dg the number
cannot he less tlun l wu? For no um versa I, not even the universal
of a rcl<1twn, is ,\ rclatwn bl't\\een particulars wluch could
necessitate at least t\\ o of them. The formula that causes our
d1fticulty about the umHr!>al hav1Dg one particular only is that
the umversal is an identity m its various particulars, so that but
for a mamfold-at least t\, o-it would appear as if the universal
could not be. Y ct 1f it 1s an 1dent1ty ID a stnct sense, 1t must
be in any given particular taken by itself and yet not identified
with it If then 1t were ID one particular only, it would still be
 Umve,sals-Last Words 7II
•universal' as at least distinguishable from the particularity.
It is in this particular what it is and is not constituted what it is
by being in any other particular besides the given one. On the
contrary it is because it is not identified with particularity that
it can be in more than one particular, where it does realize itself
in a manifold. We could not indeed recognize it as the same
universal in several particulars and identical in them unless we
could in a singular instance already so distinguish it. Hence,
if it were only in one particular, its universal nature would not
be affected thereby. We may illustrate this from the identity,
of a body in its several positions or from our recognition a
second time of a person whom we have seen but once before.
In short, what we call the universal is something of such a nature
that because of this nature it can be in a variety of particulars.
It is not necessary that 1t should be m more than one, and its
nature is not constituted by bemg in a plurality. It is what it is
not because 1t is in a plurality ; 1t can be m a plurality because
it is what it 1s.
§ 455. Universal then is not the most adequate term for what
we have endeavoured to elucidate, for universal suggests identity
in plurality and plurality is not involved in its essential nature.
The term we inherit from Aristotle 1s seriously misleading i in
his own language 1t 1s a definition of a thmg not by its essence
but by an attribute wluch 1s not a property of a given universal.
If we admit universals with only one particular, the designation
'universal' does not even apply to all universals. In Plato the
terminology used docs not imply this presence m a mamfold.
Plato's attitude is often that of one who cannot understand how
this unity can be in a manifold and who supposes that this
apparent immanence is an illusion. 1 This negative characteristic
of his terminology has advantages, but the positive side is only
conveyed by help of the metaphorical terms form and Idea.
Again the Greek term 'word', like such modern terms as con•
ception and Bcgnff, is useless, because 1t is subjective in
suggestion and therefore m1sleadmg. Nor is the analogy drawn
from the individual thmg or person as distinguishable from being
• In the Pal'tnenides he indicates vaguely a procedure for the solution of
the difficulty which is so far right that, had he clearlyreabzed 1t for what at
was, be would have dU1COvered the aolub.on
 TENTATIVE INVESTIGATIONS
here or there to be transferred literally to the character of the
universal. This latter is not merely to be distinguished from
particularity as though particularity were merely spatial, for
then its identity would not be distinguishable from the identity
of the individual It is distinguished from all particularity.
To treat particularity as merely spatial or temporal is indeed
a fallacy to which primitive reflection seems to incline.
§ 456. What then is a proper designation ? It must be
technical and not explanatory, for the distinction being sui
gtntris we require 1t always to understand or explain anything
whatever. This is too often forgotten in discussions of the
matter now before us. Just so the terms 'particular' and
'individual' are none the less technical because familiar. They
are adequate and useful because we know what is meant, we
forget their etymology and are not misled by it. But as soon as
we treat them as explanatory designations, we are obliged to
recur to their etymology and they become misleading. Aristotle
has recourse sometimes to the 'this' and 'that' of everyday
language, but the practice though convenient is not adequate for
several reasons. 'This' and •that', though they often refer to
a particular thmg, do not mean particularity ; moreover, they
are apphed to the so-called universals, although incorrectly, and
by a transference from their proper use, this colour standing for
this kind of colour. Now a universal can always be designated
otherwise than by this or that, whereas any particular cannot ;
even if we appear to do otherwise, we find, on scrutiny, that the
particulr1r in question 1s really designated by a relation to some-
thing only designated by • this' or •that'. Again this use is so far
really subJective because 1t means the individual which I am
pointing to now
In this difficulty the only important thing is to be able to
distinguish something, the so-called universal, from its particu-
larization, that is from its being this thing. Whatever we call
it we can explain that it 1s everythmg that the individual can be
said to be besides being merely a this or a that. If an individual
is this X, then X 1s an instance of what we mean by the technical
term we are in search of. Or, to put it obJectively, Xis all that
the individual is besides being merely particular.
§ 457. It is repugnant to create a technical term out of all
 UniTJe,sals-Last Words
relation to ordinary language. If we took the term 'a this' or
•being a this', as Aristotle does so often, we might take for the
universal 'being somewhat' or •being such', a use also found in
Aristotle. This again corresponds to the Latin term 'qualitas ',
but it is difficult in English to make ' quality ' a technical
term for the universal, since then it would have to embrace
quantity also. Yet, with this proviso, quality seems the most con-
venient of terms already in use. In distinguishing, however,
the particularity from the quality we must guard against making
the particularity something which has the quality, for then the
quality will inevitably become the particular quality and appear
to be the attribute of the particular. Moreover, attributes are
themselves particular, the wluteness of a surface being a par-
ticular whiteness. Thus our distmcti<>n would not give us the
true universal, which is 'whiteness'. In this way Aristotle
represents the particularity as matter which has the form. That
form is the particular quality of the thmg and not the umversal,
it is the particular defimteness of the thing. Yet in the Meta-
physics the form 1s often treated as 1£ it were the umversal.
Now the particular 1s not somethmg wluch has the quality,
1t 1s the particularized quality This animal is particularized
animality. Particularization belongs, as much as differentiation,
to the nature of the quality.
§ 458. We might substitute for ind1v1dual, or particular, the
term existence without danger of bemg misunderstood. Exist-
ence docs not suggest an explanation as ' Dasem ' tends to do, thus
carrying with 1t a wrong 1mphcation. Every existence combines
being as this and being as such, and the inseparableness of the
this from the such might be represented by 'a this such', or in
better English 'a this such and such', which brings us back to
Aristotle's language. This avoids the confusions caused by
matter and form, where matter is inevitably regarded as having
again a form of its own ; and 1£ Aristotle had made 'this such'
his formula of the existent md1vidual (1 e of his composite
being), he might, with the help of the conclusion of the Par-
menides, have given the true solution of the difficulties into
which thought had got m his day, as it still does m modern
discussions of the universal. The advantage of this formula is
that 1t makes no pretence to be an explanation of what can be
 TENTATIVE INVESTIGATIONS
recognized but not explained and that it does not employ
a metaphor, which 1s sure to mislead It indicates directly what
is meant, bemg simply a generahzat1on of the language we use
in any particular case , 1t contams both sides of the existence
in inseparable umon and has separate words for each side. It
has finally the advantage of not definmg the quality, the 'such',
through its presence m a diversity of particulars. , .
§ 459 Of late years 1t has become customary to speak of
a concrete and an abstract umversal. These are terms to be
avoided Concrete was ongmally merely opposed to abstract
aud should mean a particular existence Nothmg 1s gained by
calling an existence concrete, and the term has the danger of
seemmg to g,vc an explanation. Concrete umversal would be
a veritable contrad1ct10n m terms Agam we hear the words
' more concrete ' or 'less concrete ' used loosely and erroneously
Notlung 1s more or lc,;s md1v1dual or particular This language
arises from ,\ con[m,1011 of the d1ffcrcntmt1on of a universal with
its part1c.ul . mzat1011
 IX
TRICHOTOMY

LETTERS TO PROFESSOR I I. GOUDY

§ 460. (From Dr. JI. Goudy, Regius Professor of Civil Law in

Oxford)
All Souls College, Oxford.
7.xu 04.
My dear Cook W1l'lon,
May I venture to bother you, as a great Aristotelian, with
a question? Am,totle \\-,ts, I lhmk, ..1. ucbcvcr m the symbolic
value of the No 3. Can you say that this behef has affected
his class1ficahons of Ins suhJect-mattcr, so as to make them
stramed and 11log1cal? Or 1f not in •11 maestro', perhaps you
may have found the mfiucncc m some of his followers This
notion of symbolmn seems to me to ha\'C eur10usly affected
some of the St01cLil writers and, inter alias, the Roman jurists.
Perhaps you can put me on a good track for my inquiry,
and oblige
Yours very truly,
HENRY Gouov.

[OJI.ford, 9th Dec. 1904.]

§ 461. I should be gbd to know m what form and from what
source the view that Aristotle was a believer m the symbolic
value of the No. 3 has reached you. In general there 1s scarcely
anythmg one would thmk more alien to the sound and sober
sense of Aristotle. He had such superstitions before him m the
traditions of Pythagoreamsm and the vagaries of the later
Platonists, and m his cntic1sm of them he shows his usual
samty. He clearly has a contempt for mysterious symbolic
meanmgs attached to number · and in all of his principal works
there is no trace of his takmg any of them seriously.
If he had attributed any value to a threefold d1vis1on as such,
2773 2 X
7I6 TENTATIVE INVESTIGATIONS
one would expect it in his account of Division, or in his remarks
upon it in his logical treatises. There is nothing of the kind
and no advocacy of trichotomy. On the contrary he gives the
sound doctrine that while dichotomy was always possible, it was
useless in so far as the negative side not-B admitted no further
division from its own notion. True division, he rightly says, 1
must proceed mt~ tlVT&nr[lpT/JJ.Eva,~ ~,aq,opat~ {e. g. as when 1/'C
divide triangle mto equilateral, isosceles, and scalene). Not only
1s there nothmg m Aristotle's theory of Division about the value
of any trichotomy, but l11s own divisions, m his writings, do
not, as far as I am aware, show any special mclmation to it.
§ 462. On the other hand there 1s m the De Caelo, a treatise
accounted genume, a smgle passage, which as far as I know
stands alone 111 the works of Anstotle about the number Three 1
lie is t,Llkmg about the three d11nens10ns and IS very naturally
led to spl'ak of the completeness of threefoldness here 'Magni-
tude in one <l1mens1on (i,1,' fr) 1s the line, 111 two {f'll'l Mo) the
plane, m thn•t• (iut r,.iu) the !>ol1<l body, and there is no other
mar,nitu<lc (1. l' J.s the fourth cl11nens1on, c. g ), because 'the
three d1mcrn,ions arc all ancl the threefold d1v1s10111s complete '. 3
For as the Pythagon'anc; say, the All and all thmgs arc clefined
or ddermincd by the number three. For Begmmng, Middle,
and End comprise the Number of the All and thJ.t 1s the number
of the triad (let three comprise the number of the triad) There-
fore accepting tlus as a law of nature we make use of the number
in rl'hg1ous ritual ' lie rontmucs (c;howmg the real use he makes
[ at all) of the 1de,i) • Solid body is the only complete magnitude,
because 1t alone 1c; drfined by three (d1mcns1ons), and that is
all that 1s po~s1blc (m the way of d1mens1on) And bemg d1v1ded
in a threefold w,1y 1t 1s d1v1s1ble m all the ways (possible for
magnitude).' The last 1s further explained m what follows.
The one d1mcns1on 1s only d1v1s1blc m one way (1 c mto length) :
the two dimensions m two ways . . but the sohd only 1s d1v1s1ble
in all ways, for 1t is div1c;1blc in the three, which are all, viz.
length, breadth, and thickness.
§ 463. I have written out the whole thmg (with an exception
remedied lower down about 'all three') m order that you may
1 [Arl&t, Top 143• 36 seq] 1 De Caelo, 1 I, 1D1t 268• 7
• &cl Tel T4 Tp<a ,run 1Tva1 •al n\ TP:s """1,.
 T,icholon,y
see how the land really lies. It was the bad luck of Aristotle
that certain isolated passages in his works, because they showed
some apparent affinity to the fancies of the decadent succeeding
epochs in philosophy, were made great texts and their meaning
perverted or exaggerated. An instance is the name atvTe,,cu
ova-Ca, (secondary essences) in the Categories. This in medieval
times was associated (I am told, I haven't verified it) with the
view that Universals were Substances, than which nothing was
more remote from Aristotle's meaning and beliefs. So also it
is possible that the passage before us was misunderstood and
made too much of.
§ 464. It is worth while looking attentively into the passage.
In the first place we notice that there is no attempt to assign
a symbolic value in serious philosophical thought to the number
Three. Aristotle mentions the sober fact that in space three
dimensions comprise all possible dimensions and constitute com•
pleteness, and that (m time) a. process is distinguished (defined)
exhaustively and completely by its beginning, middle, and end.
This also is a mere fact and neither statement has any symbolism
about it.
The completeness of the threefold d1stinct1on in the case of
spatial dimension naturally reminds him of the Pythagorean
view that the number three defined the All (the whole proeese
of the world) because it was the number of the exhaustive
distmction into beginning, middle, and end. And it is noticeable
that even here he does not quote any of the fantasies of Pytha•
gorean symbolism (e.g. the mysteries of the TfTpatcris, but
something which, as far as it went, was right. One might para•
phrase him thus : ' there is a sort of completeness about the
number three in regard to space, for three is the whole of all
possible dimensions. So also the Pythagoreans noticed a similar
completeness about Three m respect of the process of time, for
that is all comprised in beginning, middle, and end.'
I 465. But Aristotle does not argue from a general idea of the
perfectness or completeness of three-he neither says ' because
it is the complete number, it is the number of the dimensions ',
&c., nor does he propose to argue from its completeness in
general to conclusions in other spheres. And there is no trace
of a view that every division should be trichotomic. For
X2
718 TENTATIVE INVESTIGATIONS
symbolism he only refers to religion and clearly only in the way
of showing how the observed completeness of the number three
in time and space d1c;tmctions naturally suggested a mysterious
use of it m rd1gion-wl11ch any modern philosopher might
have said
So also m the context immediately following what he says
about the rchg1ous use as depending upon the observing of
certain stnkmg cases m Nature, where every possible alternative
was comprm.d m three, he goes on (less happily) to connect it
with an or<lmary use or idiom of language (Greek, (so) it is in
English but not (m) German or French) whereby we do not
.i,ttach all to two and say all two, whereas we do say all three.
I dare say the passage was purchase (sic) enough for medieval
fancy and may be the whole source of what I believe to be
a men· 1llus1on-thl' view that Anc;totle assigned symbolic value
to the number threl', and I should be glad to know where you
find tlus view
§ 466 I may repeat that I believe tlus passage to be the only
one upon "luch c;uch a view could be founded, and I may
i,hortly put the gist of 1t once more shortly thus: 'The solid is
magnitude, m the complete or perfect sense, for 1t alone has
every kmd of dimension, for 1t has three dimensions and three
here compri'-<''- all Tlus reminds us of the importance attached
by the Pytlugorcans to the number three, for they emphasized
the fact that three 1s 1 the complete number of the distinctions
of temporal proccr;;s (cor;;m1r) It 1s this observed completeness
of the thrcefoltl d1v1s1011, 111 certam great natural facts, that has
made the number three a nJ.tural religious symbol : and 1t is
perhaps due to the same tlung that \\ e find it natural to say
'all tlm."t'', whcrc-as wc nevc-r say 'all two'.
I may repeat that I know of no trace m Aristotle's philosophy
of an .i,fkct10n for a threefold division as such.
§ 467. It 1s significant that m S1mphcms' comments on the
place (S1mphcms ment1onc; Ptolemy, vn 33 (Kunster's reference),
and 1s asc;igned to the sixth century A n ) two misunderstandings
(occur), JUSt of thl' km<l one n11ght expect m a late writer. He
mterprets a,a TU Ta Tpla wdvra Eli·a, «al TO Tpl.s '11'4VTTI as if it meant
complete magnitude had three dimensions ' because three in
1 M~ 'was'
 Trichotomy
itself was a compute number' 1 (the meaning is really what
I have given above), and so gets himself into a difficulty and
asks whether a man ought not to have three fingers if three
was a perfect number I 2 He erroneously supposes Aristotle is
arguing from Pythagorean principles: whereas, as I have shown,
he does nothing of the kind. Here again Simplicius gets himself
into a difficulty, since he knew it was not Aristotle's way to
argue from Pythagorean principles, and so he says' it is worthy
of notice that Aristotle here, contrary to his wont, 3 is using
a Pythagorean proof I ' The interpretation of the words mis-
understood by S1mplicius 1s put beyond a. doubt by a comparison
of 268• 8 above• and 268• 24 seq. below. 5 As he explains in the
latter place, linear magnitude in one dimension is d1vis1ble only
in one way (1. e. into lengths) and the surface, on the other
hand, is d1v1s1ble only m two ways (1. c so many umts in its
length and so many in its breadth), whereas body is d1v1s1ble
in three ways (rpixfi) and so 1s the only one d1v1s1ble in all ways,
since three ways is all possible ways. Thus in 268 1 IO 11:al ro
Tpls- waVTr, has exactly the same mearung, the d1v1s1on of magni-
tude in respect of three different wn.ys (length, breadth, and
thickness) is a. division in all possible ways

[Undated.]
8
§ 468. Dichotomy seems everywhere possible, because we
can divide a class mto the members which have a certam quality
and those which have not (I shall raise the question of the
value of this presently). There seems no such general ground
for trichotomy, so as to make it applicable everywhere. But
can we find any general reason for it or principle of it where
it is applicable, and so determme certain general characteristics
which the subject-matter to which 1t applies must have ?
Suppose two species of a genus are related as 'contraries' and
are the species of a given genus furthest removed from one
another according to the same principle of order. 7 If there is
1 6b 35. • TlJ,,.or dp,9µ6t, 7• 19 seq
1 •a,A 7',) mqfn. 1 11iiJp.a.1Ji 'Tel WTII a,a.ipu&v.
1
'Tl"Xii Ha. llia.&PffW, ·•"111 &a.ipmw.
• [Dr H Goudy had sent Wilson h!S article ' Trichotomy m Roman Law ',
with a request for cnticism]
• Anstotle defines iva.vTia. as ri •A1urrw dAA~Aoi" a.,,,T.,iroTa. ,., 'Tqi a.uT+; .,..,,,,
 TENTATIVE INVESTIGATIONS
one class 'between' them and with affinities to both, we get
a tripartite divtsion as of triangles into
those which have all their sides equal (equilateral},
those which have none equal (scalene), and
those which have two sides only equal (isosceles)-the latter
species bemg intermediate between the other two. But this
does not necessarily apply when there are two 'extreme' species,
for it is not necessary that there should be only one species
intervening. E. g. we may apparently divide four-sided figures
into three:
(I) those with all sides equal,
(2) those with no sides equal, and
(3) those with some sides equal.
But the third species divides into those with
(1) 2 sides equal and the remammg two unequal,
(2) 2 s1d1..-s equal and the remammg two equal to one another
but not equal to the first two,
(3) 3 sides equal.
And the co-ordinate d1v1sion 1s therefore mto 5, there being
three species intermediate to the extreme species.
Such tripartite d1v1s1011 therefore, wluch 1s based on there
being two 'extreme' species, depends entirely on the sub1ect-
matter and is not of universal apphcat1on even to the cases
which present 'extreme' species.
§ 469. Another trichotomy which has the look of generality
11 that of which masculine, femmme, and neuter, (in our) 1 class1-
fication of species of certain animals (e. g. bees and ants), is an
example. Here the members of a genus are classified according
to the presence or absence of a qua.lity (c. g. sex}, and the case
of presence shows two species (e. g male and female) only,
which are opposed to one another. Clearly this 1s not of general
application to the cases where a class can be divided according
to the presence or absence of a quality, for it is not necessary
that the presence of the quality should give rise to two species
only, like male and female. E. g. conics are divided into those
which have a centre and those which have not. If we confine
ourselves to real curves (excluding, i. e., straight lines as limits
of conics) there is only one species without a centre, the parabola.
• MS 'or'
 Trichotomy
But there are three which have a centre, the circle, the ellipse
and the hyperbola.
§ 470. Thus the possibility of trichotomy depends on the
subject-matter entirely and {I} there is no general principle of
trichotomy applicable to all subjects, like dichotomy, (2) there
1s no general reason or principle of it for all cases where it is
applicable : no general characteristics which all the subject·
matter to which 1t 1s applicable must have and which always
determine that the division must be tnpartite.
§ 471. But we may go further. In dichotomy there is
apparently a general prmCJple of d1v1sion mto two : and one
applicable to all matter. This is an illusion. If we divide
X into A and not-A, the negative (mere) designation 1s mdeter-
minate, cannot determine a class. There 's nothing to prevent
the negative not-A from having sometimes several true classes
corresponding to 1t, which arc co-ordinate with A, so that the
d1v1s1on of X mto A and not-A should really be, e.g., into A
and Band C.
The negative may have one class correspondmg, Thus number
ts either even or not-even, and not-even = odd. But this
depends agam on the sul>Jcct-mattcr The mere negative not·
even cannot cktermme a dass, that 1s due to the positive
character of number d1v1dmg itself mto even and odd, both
pos1t1ve.
The mere ncgat10n, v,1thout thus correspondmg to a pos1t1ve
designation, may seem to detcrmme a class m such an example
as the d1v1s10n of cats mto those which have tails and those
which have not. The general account of this 1s that X being
divided mto A and not-A ; XA really = B + A and X
not-A= B. Thus X not-A 1s really defined by the quality B
and X is divided mto that which 1s B and A, and that which is
B only.
Thus there is not even a general prmc1ple which would yield
a true d1chotomy or d1v1S1on of a genus into two co-ordinate
species. And there is no general principle which would yield
a dlv1s1on into any defimte number of species, whether 2 or 3 or
more. The number must depend solely on the nature of the
special sub1ect-matter.
§ 472. This brings me to Hegel, for Hegel's principle does
seem to lead to a threefold d1v1s1on, and that would appear to
722 TENTATIVE INVESTIGATIONS
conflict with my statement above. But the truth is that Hegel's
threefold division 1s not a d1vJS1on of a genus into species at all.
The Hegelian triads, m fact, arc not an example of trichotomy.
He supposes that 1f you take a conception and then its contra-
diction, there arises a tl11rd conception m which they are united
or reconciled E. g. umty and plurality are thus combined or
reconciled m totality. Now tlus 1s not a classification of members
of a g('nu,; mto 3 mutually exclusive species One and the same
thmg &hows the umty of a plurality of parts or elements in
a totality. There 1s here no question of the d1stmct1on of
a genus mto species. So also \\1th the combmatton of Being
and Not-Being m Becoming Not only 1s Hegel not properly
then a tnchotom1st, but I doubt whether 1t 1s at all fair to
Hegel to say that he ,..,as aficcted by any merely numerical
influence or superst1t10n His method 1s not founded on any
respect for the number three or for a threefold d1v1s1on. But
his principle of the umty of eontrad1rtories, as above explained,
necessarily produces tltc&c triads of roncept1ons I am not here
defending !us prmc1plc IJut only pomting out that 1t 1s founded
on no cons1dernt10n of number as such, though 1t gives nse to
these group,; of three
§ 473 Kant would be under more suspicion than Hegel, for
each of the 4 classes mto wluch he d1v1dcs Judgements (as to
quantity, quality, relation, and modality) contains three mem-
bers. In three of these he nught be said to find a threefold
d1v1s1on (Qua11t1ty mto Umvers,tl, Particular, and Smgular;
Relation mto Categorical, Hypothetical, and D1sJunctive; and
Modality mto Problematical, Ac:;sertorical, and Apode1cttcal).
But in Quality he will only find Affirm.itive and Negative. So
one suspects his addition of the Infuut<.', whtl h 1s clearly artificial,
was due to the wish to m.ike a thr<.'cfold d1v1s1on of the class,
correspondmg to the threefold d1v1s10ns m the other three
classes. But even tlus would not be due to a respect for the
number 3 but to a respect for symmetry. When Kant had got,
very arttfic1,1lly, his 12 Categories out of his logical d1stmction
of propos1t1ons, he discovered that, m some of the 4 classes,
the 3rd Category was the union of the 1st and 2nd. He agam,
art1fic1ally and affected by symmetry, supposed this true of the
3 categories m all the four groups, though 1t clearly is not the
 Trichotomy
case. It was this unity of two in a third which suggested
Hegel's principle. But while Kant does not arrive at his cate•
gones from 1t, but observes the peculiarity in them when
found, Hegel finds his system of categories by means of the
principle.
§ 47 4. 1 You think there 1s here 'the most patent disregard of
logic'. I must here join issue with you on the logical point.
Your statement as it stands would condemn a procedure which
is perfectly logical and may only appear not to be so through
a difficulty which is mamly verbal.
Suppose we have a class X, which is divided into two, A and
B, and the class B 1s divided mto two, e. g. B1 and B2

B appears as a genus in relation to its d1v1s1ons B1 and B1 as

species, and A, being co-ordmated m the first mstance with B,
may seem also a genus in contrast with the species B1 and B2
and therefore improperly co-ordmated with them as species of
X. But there 1s a case m which this is but a verbal fallacy and
the co-ordination 1s qmtc nght , and it looks, too, as if this were
the case before you Suppose whzlc B zs thus capable of further
d1fferentiatton, A is not. A, then, is as much differentlated as
B1 and B1 ; or, 1£ you like, X is as much differentiated in the
direction of A, m A itself, as X 1s differentiated in the
direction of B, m B1 and B11• And, 1£ you ask the question
what are the infimae species of X, the right answer 1s A, B1,
and B1•
It is in fact the case where there is only one species of A,
which coincides therefore with A itself In the above figure
A1, the species of A, quite co-ordinate with B1 and B1 , is equi-
1 [A fl.ot11 to Dr H Goudy on an extract from the ' Digest ' ' Convent1onum
autem tres sunt species. aut crum ex publlca causa fiunt, aut ex pnvata
(pnvata) aut leg1tima aut 1uns gentlum ' Ulpian, hbro 1v, ad Eriiclum.
Dt1 II XIV s]
 TENTATIVE INVESTIGATIONS
valent to A. There are plenty of instances in the sciences where
A1 ( = A), B 1, and B1 are thus co-ordmated, e. g. :
Curvilinear conics

~,,,/
Central
_,,,...- ------ ::---.
Non-central
/---------------------
asymptot.il non-asymptotal
I / ,

hyperbola L(:isymp.) circle elh~se parabola ( -non-central)

Here the poir1t is illustrated twice over. In the division of
central comes into two, asymptotal and non-asymptotal, the
latter being genus to the species, circle and ellipse, asymptotal
would (as you put it) seem like a genus m contrast with them
because co-ordm.ited with then genus But its one species is
the hyperbola, which therefore comcides with it, and the hyper-
bob 1s alv.ays co-ordmated in a div1S1on of comes with the
ellipse and the c1rclc1 and nghtly. Further, non-central curv1•
hnear come has only one ,;pcc1es, the parabola, and this agam,
though non-central c1s c.o-ordmc1te with central (wluch is genus
to the species cmJe and clbpsc), seems hke a genus m comparison
with these latter But having but one species, with which it
comc1des, it 1s nghtly co-ordmated with hyperbola, circle, and
ellipse These forms arc called the come sect10ns (curv1linear),
and quite rightly, because each represents the functional dif-
ferent1at1on of the genus come section which 1s relevant to
geometry
The consequl•ncc 1s that m the d1v1s1on of conventions into
ez publua causa and er: privata causa, if there arc no ddierentia-
tions further of the former parallel to those of the latter (and
from the silen<'e m your text one would certainly infer there
were not), then the former is rightly co-ordmated with ez privata
legitima and ex privata iuris gentium, and the logic ( of Ulpian)
1s quite flawless
§ 475 The example of the come sections is a true example
of the class1fication in which one member is co-ordinated
with a true alternative member of division and yet co-
 Trichotomy
ordinated as. inftma species with the species of that alternative
member.
There is an incorrect way of producing the same thing in
appearance. We may take a species X 1 , which is co-ordinated
with others, X 1, &c., in an immediate ddferentiation of one
element, say X.
Then divide this X into Xi and not-Xi, and not-X1 into X 1,
X 8, &c. So that X 1 is co-ordinated first with the genus not-X1
and then with the species of it, X1, X 8, &c.
But this is obviously a merely verbal device and the class
not-Xi, defined by the mere negative of X 1, is not a true class
at all.
The fallacy here would not lie in co-ordinating X 1 as a genus
with X1 and X 8 as species, but in co-ordinating X 1 with not-X1
as both genera. An instance would be dividing triangle into
right-angled and oblique-angled ( = not having a right angle)
and then dividing oblique-angled into obtuse-angled and acute•
angled, so as to co-ordinate right-angled first with oblique-angled
and then with the species of the latter.
§ 476. (The following is a full account of the true case
illustrated by the come sections i
A universal A has, e g , two clements, X and Y. A = XY.
The ddierentJahon of A according to X produces A1 and At,
which are of the form X 1 and X1 and are truly co-ordinated,
because the d1fferent1at1on of X can go no further in either of
them.
But the connexion of X and Y in A is such that the differentia-
tion of A mto X 1 involves necessarily the differentiation of Y at
the same time into Y1, so that A1 = X 1Y1•
Here A1 is necessarily an infima species, both elements in the
universal X being d1fferentiated.
But the connexion of X and Y in A may also be such that
the differentiation of X into X 1 does not carry with it the
differentiation of Y, so that Aa = X1Y.
Then A1 is not an inftma species and 1s divided into such by
the differentiation of Y. Thus
• [Thts was added later, I think, by the author to his copy of lus letter to
Goudy He has marked 1t as an addition to the logic lectures (Part II,
ch 16) and noted that they are, where necessary, to be corrected accordmfly]
726 TENTATIVE INVESTIGATIONS
A("" XY)

, - ~ ~
A1 -X 1Y1 A9 •X8Y
L ~=
A21 = X1Y1 A22 X8Y8
Thuc; A1 =- X 1 Y1 is nghtly co-ordmated m the differentiation
of X with A8 = X 2Y and nghtly co-ordmated as an ultimate
d1ffcrentiatton of A, or as an inftma species with A81 and Au-
The same kind of thmg happens if X 1 mvolves the d1fferentia-
t10n of some clement or elements m Y, but not of all.
Let Y = a/J, and let A2 = X 8a 2/J, because X 1 involves the
differentiation of a mto aa
Then A2 = X 2a2/J
/ '
_.,,/ '-..._
/~ ''-..._
A21 = X2a2f3J A22 = Xaaa/3a
We may get more steps, 1£ there arc more elements in A.
Thus 1£ A = XYZ, "e may get steps corresponding to the
above classificat10n of come sect10ns For suppose that X 1
necessitates Y1 Z1 , and X 2 leaves Y undifferentiated, whde Y 2
necessitates 2 2, and Y3 lcav<..'b Z undifferentiated :
A =XYZ
-- i\l - x;vz (central conic)
A21 ~ X 2Y2Z2 (h)'pcrbola) A21 ----= X 2Y8Z {closed)

A, • X,Y,Z, (pambola) , ,,- \

A 211 = X,.Y9 Za (circle) A222 = X,.Y8Z,. (ellipse)

So that A1, A11, A11111 , and A111 arc co-ordinate as infimae species.
Note also that, m strict accuracy, the language used above
(pp. 723-4) is not admissible. A ought not to be represented
 Trichotomy 727
as a genus with only one species (A1 which coincides with it)
and similarly for the further concrete instances. In the fu])er
symbolism A1 ( = X1Y1Z1} is a species of A ( = XVZ), just as
much as A111 ( = X 1Y 8ZaJ. It looks like a genus, only because
A,., with which it 1s co-ordinated through the division of X1, is
a genus with regard to the infimae species A11, A 1111, A 1u,
Strictly, a genus with one species is a contradiction ID terms
and it 1s only allowable 1£ it means a genus which now has only
one existing species, though 1t had others once For examp1e,
if all species of a kind of plant have died out except one. But
this case can't occur in the mathematical examples)
Note to § 462 [ Ka8&.7rtp yap (/,arn Kal ol IlvOayopflOI, TO 'll"GJI Kal nl
'11'41/Ta TO°i§' TpllTLII i':Jp,crTal. TEh(VT~ yo.p ,cal µ.lcrov ,cal cipx~ TOIi cip16p.o11
lxu TOIi Toil 'll'a11nh·, TaVTa ae n'lv Tij5' Tp1&.ao§'. 311'1 'll'apa. Tijs (/,VITECIIS
i,>..71q,,fru: i'!Jrr'll'fp voµ.ovs EKt(v71s ,cal 'll'pOS Ttlf b:yl(TTf4ar XP~p.E9a TWJI
9EWJI T,ii lip,8µ.ip TOVT'f' • • • TO crwµ.a µ.ovov Av Ef1J TWII /J,EYEB&iv Tlhnov•
µ.ovov yup ~plUTat TOi5' Tp1rr£1•. (Aristotle, de Caelo, 268" IO seq.)
'For as the followers of Pythagoras also say, the whole
and the all is determined by "threes". For, they argue, end and
middle and beginning embrace the number of the whole, and
that is the number " triad ". Accordingly we receive from
Nature, as ID a sense her laws, this number and use 1t for the
worship of the gods . . Body alone of magmtudes would
appear to be complete, for 1t alone 1s determined by three.'
St. Thomas Aq. remarks on this passage: 'Quod ergo
Aristoteles d1C1t per hunc numerum adlubmmus nos 1psos magni•
ficare Deum unum, emmentem propnetat1bus eorum quae sunt
creata, non est sic mtclhgendum quod 1pse poneret ternarium
numcrum m dmm1s Sed uult d1cerc quod ant1qm utebantur
tcrnano numero in sacrificus et orat1ombus propter quandam
ternarit numeri perfcct1onem.' S. T. I, qu xxxu, art. I, ad
pnmum. This shows that the passage had been used in con·
troversy in the thirteenth century. There 1s a curious reference
to Aristotle's supposed doctrine m the Port Royal Logic, Part III,
ch. xix, § 3 (Non causa pro causa)
(It was usual m Greece to mvoke the gods in threes, not in
pairs; thus ID the Clouds, Socrates swears by Void, Clouds, and
Tongue (424), and agam by Breath, Void, and Air (627), but
the conscientious usurer by Zeus, Hermes, and Poseidon (1234).)
Cicero cnt1c1zes Epicurus for a mistaken trichotomy, De Fin.
1i. 9. Hegel ascribes to Kant the credit of drawmg attention to
the predominance of trichotomy m the sphere of the mind.
Logic, (Encyc.) § 230. Cf. Kant, K. der Urth., Introduction, § 9
(note).]
 X
CORRESPONDENCE WITH B. BOSANQUET
§ 477.
I2 Fyfield Road,
7th July, 1903.
My dear Bosanquet,
I am returning your I st six pages and am sending also some
new notes on the subJect, as I do not think you quite realized
my points. If I have got wrong, I trust you will be as frank
as you wished me to be Error m these childish and verbal
(for that 's all they are) subtleties is exasperating, for it tends
to make these abject trifles seem important-as doubtless they
seem to the writer you wot of so I shall only be too glad
to have any mistake I have made put as clcar]y and poinudJ.y
as possible. Though I don't pretend to thmk I have made any
senous error, I know one 's never sc1fc
What I feel is that these arc simply difficulties for minds that
do not realize .1.ccurately the nature of grammatical distinctions,
and 1t makes me very angry that they should masquerade as
philosophy.
There are d1fficu]t1es coming from a want of realization of the
fact, on the part of the logician, that he 1s dealing mainly with
a linguistic (grammatical) question with which one has more
sympathy-e g the whole confusion of the stock distinction
(and its treatment) of the connotation and denotation of terms.
But this stuff makes one feel why PJato and Aristotle hated
the quibbling eristic with such a holy hatred. They must have
felt it so cheap and been so vexed that it could attract attention
and be mistaken for clever thinking.
As to the papers on Royce (by the way I daresay you would
like them back. How long may I keep them? You needn't
return what I am sending you this time as I have taken a copy
of the essentials-just as you like) I would rather write later on
1h41, when you have had time to consider what I am now
sending.
 Corresf,onunce with Be,na,d Bosanquet 1•9
I will just say that though not quite ready yet to take the
field upon the whole • ketten' business on which R. 's puzzle
seems to depend, I have just a little to say on the fallacies to
which distinguished mathematicians are committing themselves
in our day, which may interest you.
I feel sure we shall get on, because I am attracted by what
I venture to think the grasp of such matters as' infinite number',
&c., shown in your Logic.
Will you however m the meantime when you next write tell
me what you yourself understand by a true and a false infinity,
and what you suppose Royce to mean by them. I observe that
reflective mathematicians are begmnmg to speak of true and
false infinity: but I don't know where they get 1t from nor am
(1) altogether sure I know what they mean by 1t. It can hardly
be Hegelian (can 1V), for would Hegel's idea of the 'mfimte', as
the non-finite and ac; therefore somethmg whole and unhm1ted
by anything else and therefore properly only the whole, be
applicable at all w1thm quantity (mere) ?-and of course 1t 's only
of quantity that these mathematicians are speaking (except, as
Ca1rd points out to me, that the formula which covers the whole
infinite series might be allowed by Hegel as the truest approach
to mfimte in quantity).
May I add I like controversy with a single person? In
a general talk of several I find I am too inclined to be afraid
of bemg shown to be m any way wrong, and to fight for my
hand. But that sort of feeling hardly exists, 1£ at all, when
discussing with no 3rd person to hear, and I like the clearest
and most unsparing statement
Yours truly,
J. CooK WILSON.

§ 478. The first three pages 1 are a little difficult to deal with
because the obJection does not seem clear. It mentions some-
thing about the difference between the class of classes and
ordinary classes which 1s on the whole true but seems irrelevant
as an ob1ection. Besides you interpose a cribc1sm on it yourself
in which you abandon an important statement, so that at is not
perhaps easy to see what really remains.
1 [Note1 appended to letter of 7 vh 03]
730 TENTATIVE INVESTIGATIONS
I venture to imagine that l see what the matter is because
something similar happened when I read a paper like that sent
to you to our small society Isn't it that you began a little
too hastily-you observed I was talking of a class the 'members'
of which were not, or not necessarily, themselves groups or
classes; whereas it is distinctive of the class of classes to have
groups unified by a common element (or notion) as \ts
members.
When I read my paper and had completed the account of the
fallacy applied to any class whatever, and its refutation, one of
my hearers at once ob1ected that the 'class of classes' had the
above d1fference from an ordinary class. I simply begged him
to observe that I of course knew the supposed characteristic of
the class of classes, and asked Jum to be content at present to
judge whether I had proved my first pomt or not, viz. that the
paradox enunciated of the class of classe~, that it was a member
of itself, c.ould be estabhshed, on the line of the paradox, of any
clasc; whatever and was not a privilege of the class of classes-
and also that I had stated its refutation
Then he admitted this (as every one else chd), and then I asked
lum to hear me out for I had provided for the class of classes.
He admitted when he had heard the whole thing to the end
that I was nght. And I really think you began JUSt as he did,
whether you ,vill end as he did is another matter
I had really written a longish piece upon membership of
a class and the special character of membership of class of
classes, wluch I spJ.red lhem and you, because it seemed to me
one could shortly and neatly refute the fallacy m the way
I proposed. If I had sent that, you would have seen I was
quite awake to your runsiderataon.
I had better state exactly what my method is in order to
show the irrelevance of the objection (or half ob1ection).
§ 479. The paradox about the class of classes 1s that being
like the classes subordinate to it, m so far as it 1s a class and
they are each a dass, it h::i.s a common element with them (being
a class) m virtue of which 1t is classed with them, i e -a member
of a class of wluch they are also members
Now I undertake to show that it 1s a mistake to suppose such
a procedure applies only to the class of classes, and therefore
 Co,respondence with Bwnard Bosanquet 731
the paradox or supposed bit of metaphysics (l 1) was not due
to the wonderful subtlety of making a class of classes, but
belonged to any class, however ordinary, common, or un•
clean.
I. I proved I think rigorously that, by the method of the
paradox, [that] the paradox could be shown true of any class-
taking the quite general symbol X for any class of things which
are X.
Now on page 4 you say apparently that you doubt whether
without niuch forcing I could apply the paradox to any class.
This puzzles me extremely-unless I may suppose that in hasti•
ness you merely thought of the aforesaid difference of the class
of classes and my X class, without first going through my
demonstration. It seems to me 1£ you would look at that, you
wouldn't find the slightest forcmg about 1t. It seems easy and
simple and I certamly thmk I have shown that to any class the
paradox could be applied and you haven't anywhere mdicated
any flaw m the demonstration.
§ 480. 2. (a) I showed that, without makmg any special pro•
vision for the 'class of classes', what I said of a class in general
could be apphed immediately to the class of classes-syllo•
gistically. For that it was only necessary to make a certain
substitution. For Xness substitute simply 'being member of
a class'. I didn't begm this way, but showed independently how
the paradox was developed m the class of classes and what was
its refutation But 1t struck me, after I found the general
application of the paradox. that it was quite easy to represent
the 'class of classes' case as a particular case of it : and it was
important to do so as the full nature of the fallacy 1s the better
elucidated. Besides it helps to show that 1t is a mistake to
suppose the whole thmg a unique characteristic of the class of
classes.
(b) I gave, what I put rather shortly at the end of the paper
sent to you, a treatment of the class of classes as such. and
without bringmg 1t under the general theory as m (a).
In this way I satisfied my obJector (who is one of the acutest
people here) because I not only treated the class of classes on
it, ~wn merits-with special attention to its nature as having
,lasses for its members, which was what he wanted, but l also
1773-a Y
732 TENTATIVE INVESTIGATIONS
gave him a new view of it whlch hadn't occurred to him (or to
any of my hearers) by showing the analogous thing in any class
whatever. I attach importance to the last, because clearly the
people who advertise the paradox imagine it due to their clever
•class of classes '.
(By the way it 1s a mistake to suppose-if anybody does-
that the class of classes is the only class which has classes for
its 'members'. I shall come to that presently.)
§ 481. Let me now return to the first part of my argument.
Perhaps I put it too shortly (a habit coming from my smattering
of mathematics) but I stated essentials, and thought it could
be easily worked out.
I will now detail 1t fully-or as fully as necessary. First we
may show that, 1f W is the class of classes, [that] it is nghtly
represented as 'the totality of things which are members of
classes', and that the general character1Stic of the class W is
rightly put as 'being member of a class'. (Of course I do not
deny that we should naturally at first put 'classness' in this
general notion, and indeed that 1s right enough, but it doesn't
suit my present purpose And the notion I have chosen 1s also
right from another pomt of view, and is the pomt of view which
suits my purpose.)
W bemg the 'class of classes'.
Let the classes subordinate to 1t be A, B, C, &c.
Now each of these classes has its own members unified by
a universal, A by 11 (say), B by fJ, C by y, &c -so that A consists
of all thmgs wluch are 11; B of all that are fJ, &c, &c.; but
for the purpose of the 'class of classes' we abstract 11, {J, y, &c.,
and regard A, B, and C, &c , &c , only as groups umfied by
some prmc1ple or another ( = dasses)
Thus the members of each group A, B, C, &c., &c., have the
common designation of being 'members of a class'.
W is the totality of the groups A, B, C, &c
. ·. W is also the totality of the members of these groups.
. ·. W is the totality of thmgs which are members of classes
and the generic conception corresponding to W is 'being a
member of a class '.
The relation of the general case to the class of claues may
QOW be shown in parallel columns :
 Cor,ssf,ondence 'll1ith Bemarl Bosanquet 733
Let X be any class whatever, Let W be the 'class of clasaes '.
of which the class conception Then, as above, the class
is~- conception is 'being a mem•
ber of a class '.
X then consists of things each W then consists of things each
of which is an x. of which 1s a 'member of
a class'.
Let A, B, C, &c., be classes Let A, B, C, &c , be classes
subordinate to X. subordinate to W.
A 1s a group of things each of A is a group of things each of
'which is an x (as well as which 1s a 'member of a
being an ci-u being charac- class' (as well as being an u,
teristic of the sub-class A). u characterizmg the class A).
So B is a group of things each So B is a group of things each
of which 1s an x, &c., &c. of which is a 'member of a
class', &c., &c.
&c., &c. &c., &c.
X is the totality of the groups W is the totality of the groups
A, B, C, &c. A, B, C, &c.
. ·. X is the totality of the ·. W 1s the totality of the
members of A, B, C, &c. members of A, B, C, &c.
X then is the totality of thmgs W then 1s the totality of things
which arc each x which arc 'members of
classes'.
This latter statement is then The latter is erroneously stated
erroneously put m the form · thus.
X is a totality of things "luch W 1s a totality of things which
are x. are 'members of classes'.
The mistake of substituting the indefinite for the definite
article being made, the fallacy proceeds thus .
X being 'a group of tlungs W being a totality (or group)
which are x' 1s exactly like, of things which arc •mem-
in this way, each of the bers of classes', is thus
groups A, B, C, &c , for each exactly hke each of the
of them 1s 'a group of thmgs groups A, B, C, &c, for each
which are x '. of these 1s 'a group of things
which are members of
classes'.
It is therefore hke them a W is therefore hke them a
member of the class defined member of the class defined
by the general quality com- by the general quality com-
Y 2
134 TENTATIVE INVESTIGATIONS
mon to A, B, C, &c., of mon to A, B, C, &c., of
•being a group of things •being a group of things
which are % S '.
1
which are members of a
class'.
But the class so denned, being But the class so denned, being
the totality of groups which the totality of groups which
are 'groups of thmgs which are •groups of things which
are x's ', 1s exactly the class are members of a class', is
X itself. exactly the class W itself.
. ·. X 1s a member of itself. . ·. W, the class of classes, is
a member of itself.

This I think proves the exactness of the parallelism. I suppose

I need not add the refutation m parallel columns ; it may be
easily shown that the refutation I gave for the X class applies
with the same exact parallelism to the W class.
I will come to the treatment of the 'class of classes' by itself-
taking •classness' as its general conception instead of 'being
member of a class '-presently.
§ 482. Now I want to add the proof that any class whatever
may be treated as a class of classes, 1. e. as a class of which
the members are themselves classes. This I thmk has the
additional interest of making clearer the meaning of membership
of a t'lass.
Consider the class of rectzlmear polygons. Take as its general
characteristic •polygonal' The species are triangle, quadri-
lateral, pentagon, hexagon, &c An individual triangle, as
triangular, is a member of the class triangle.
As polygonal 1t 1s a member of the class 'polygonal', then
the members of the species •triangle' as 'triangular' are the
same as the members of the genus 'polygonal•, and the members
of the genus are the same as the members of the various species.
The class •triangular' is not a member of the genus 'polygonal•
simply because tt must be a definite mdnmlual figure which is
polygonal, The class triangle cannot as a totality have the
predicate 'polygonal'. Thus the species •tnangular ', 'quadri-
lateral', &c, are ,wt members of the class polygonal.
Consider the class triangle as the total of all triangles, 1t is
thus th, class consisting of all triangular ngures and at is also
" class consisting of polygonal figures. Si111ilarly the class
 Cor,es-pouence flJHh Bmwtl Bosflnquet 135
pentagon is ta class consisting of polygonal figures. Taken in
this way each of the species, triangle, quadl., pentagon, &c.,
is " class consisting of polygonal figures.
As having the general predicate 'a class consisting, &c. ', they
come under the genus 'class consisting of polygonal figures• of
which they are true members. Thus if we look at the class
of rectilinear polygons, as \\e may, as the class whose cla,s
predicau is '11 class consisting of polygonal figures' and . ·. as
the class consisting (of) the various classes of polygonal figures,
J. e. as the totality of classes consisting of polygonal figures, it
has for its members now not individual figures (for no individual
figure is a class, i.e. can have the predicate 'a class of polygonal
figure'} but the species or classes •consisting of triangles',
'consJsting of quadls. ', &c.
Class X with generic predicate 'a class consisting of polygonal
figures'.

class triangle. class quadrilateral. class pentagon.

specific predicate, specific predicate, specific predicate,
• the class consJstmg ' the class cons1stmg 'the class consJstmg
of triangles ' , of quadrilaterals'; of pentagons' ;
generic predicate, generic predicate, generic predicate,
•a class consisting of 'a class consisting of 'a class consisting of
polygonal figures' polygonal figures' polygonal figures'
&c. &c. &c.
Thus any class wluJ.tever may be exhibited as a class of classes,
i. e. as a class of "hJch the members are classes and not indi•
viduals (and you can easily repeat the fallacy for this class,
i.e. Ike class cons1stmg of polygonal figures).
§ 483. I venture to hope you wdl find that the above com•
pletely ans\\ers the difficulty you have raised on page 4, not
that the addition 0£ detail I am now sending is necessary i for
it i& contamed m what I sent you at first. You say m short
you don't see how, in an ordinary class, groups can be members
of the class-and how . ·. the paradox could be appbed to any
class. Well, if you had looked carefully mto my demonstration
about the X class-which I have repeated in parallel columns
above-you would have seen I had proved that the paradox
could be applied to any class, and so that your objection was
 TENTATIVE INVESTIGATIONS
either irrelevant or incorrect in itself. But if you had then
examined the implication of the argument you would have seen
that it does show any class whatever can be so represented that
itl apecies appear as its members (and not the members of the
species themselves). For the classes A, B, and C appear there
as each 'a group of things which are x', which makes them true
members, and the only possible members, of the general class
determined by the generic predicate 'a group of things which
are x'. And the general class of which they are the true
members 1s obviously coincident with X.
What I have added now is simply the illustration of the
polygons to make the thing clear, and exhibit the general
principle apart from its appearance m my other argument.
I thought after all I needn't take the separate case of the
general class of all classes over again by itself, because you seem
to agree with me-for you say your criticism (middle of your
third page) comes to the same thmg.
§ 484. But now frankly-you "ish me to be frank-I do not
tl11nk your treatment 1s clear or satu,factory (merely, I think,
because you ,,ere too hasty over 1t) (1) It 1s encumbered by
\\ hat seems to me untenable, viz that the species cannot be
represented as members of the genus, i c as classes members
of a class I have tried to show how this is really true after
all. (ii) I tlunk 1t important to keep these questions about
classes and their members quite apart from the relation of
a 'concrete' whole hke a rhomboid to its parts or of a square
to its component squares Such comparisons are hable to lead
to much confu,;10n, .i.nd so I thmk 1t a pity you should work
your cnt1c1sm (p 2 1 bottom, 3, top), which you thmk comes to
mine, through these dangerous analogies More especially as it
seems to m\'oh·e you m something untenable (p 2, bottom) :
for the whole class 1s m no sense co-ordmatc ,,1th 1ts members:
and this leads to my third difficulty m this part of your paper.
(iii) Page 3, middle, you seem to me, frankly agam, to shp back
into the fallacy yourself. 'The class of classes could be looked
at, if it interested one to do so, as "a class" and so co-ordinate
with all other classes.' On the contrary I should say this can't
be done, and indeed that the simple account of the fallacy is
that it consists m reprcsentmg tlie class as a class-a mere
 Co"espondence wlf Bmuwd Bostmpet 737
fallacy of confusion of the two articles. The essence of the
argument of mine with which you profess agreement is that we
cannot regard the class of classes as a class : ' a class' is an
adjectival predicate, and as I have said, we have to see what
such adjectival predication exactly means.
§ 485. In these linguistic puzzles I think we must not only
hold strictly to the normal meaning of the language, but sub-
stitute symbols for the words likely to deceive and then interpret
our result back again.
The class of classes is a class, is of the form ' Y is a Z '.
'Y is a Z' strictly means that Y is one of the things in which
Z is realized or realizable ; and 1t does not mean that Y is the
only thing m which Z 1s realized, i. c. does not mean that Y is
the total realization of Z. It means, then, that Z is realizable
m things other than Y, in the strict sense that they are not
included in Y.
These other reahzat1ons of Z are said to be co-ordinate with
Y, and this is the proper meanmg of co-ordmation : they are
members of the class as each 1s one realization am.ong others,
and that is the meanmg of membership of a class, which must
not be considered as merely idmtical with the possession of
a predicate (this answer to one of your questions I will return
to further on).
Now to translate back : make Z = class, then Y 1s a Z = Y
1s a class.
By the above this properly can only mean that Y is not the
total realization of 'classness' : that classness 1s realizable 10
classes other than Y which are not included, m Y.
Thus in the proposition 'Y 1s a class•, Y cannot possibly be
the 'class of classes' ; for Y as the totality of classes is the
total realization of 'classness ', and 'classness' 1s not reabzable
in anything not included m Y.
Thus the expression the 'class of classes is a class' is simply
false. If it were true 1t would necessitate the co-ordination (as
the above shows) of the ' class of classes' (Y) with classes not
included in it.
Thus the paradox 1s shown to be a mere verbal paralogism,
for it all depends on the false proposition, 'The class of classes
1s a class'.
 TENTATIVE INVESTIGATIONS
The above while directed to a remark of yours is also the
promised separate treatment of the class of classes making
X .., ' classness '.
§ 486. You ask (p. 1), after a correcting sentence, 'What is
membership, as distinct from possessing the predicate?' To all
questions of the form 'what JS A except B ? ', or to statements
of the form 'A only means B ', I should answer that their
insufficiency 1s shown by taking them stnctly. If A only means
B, then we need never mention A at all again : substitute B for
it everywhere, and strike out A. In any concrete irtstance the
result of the substitution 1s enough to convince, e. g. 'A straight
line is only (means only) the shortest distance between two
points'. Very well-never use the word 'straight' again and
see what happens How, for instance, should we feel when we
came to the word curved? In fact the statement• A only means
B ', 1s only of any interest because A is somehow distinct from
B, not as a mere word, but as a meanmg A and B are different,
but yet they are inseparably connected, and this latter 1s con-
fused with their 1dcntificatlon.
Thus m the present instance • possession of a predicate'
(adJect1val predicate) necessitates m the way above shown
membership of a certain class, but 1s not to be merely ide,itiftd,
with 1t.
Obviously, when in experience we observe that a particular
Y has a property Z, in general we only think of the particular
Y having the particular property-we do not thmk of it as
a member of a class, or think of other obJects as possessing the
same property. Of course our Judgement necessitates that Y is
so classed, and on reflection "e may realize this.
But the answer to your question doesn't affect the other
issues
§ 487. The puzzle you gh,c on p 6-wh1ch was apparently
taken seriously on the authority of your I C S. man-is only
a corollary of the general fallacy explained above, and may be
e-asJ!y refuted by help of the foregomg consideration. The
statement of alternatives 'everything 1s either a member of
itself or JS not a member of itself' 1s altogether incorrect, and
in it bes the fallacy. The correct statement JS .
Every class ia either.-
 Co"est,tmunce rJJith Bmwtl Bouffl/Ud 139
(1) • class which is a member of itself,
or (2) the totality of all such classes,
or (3) a class which is not a member of itself,
or (4) the totality of all such classes. 1
The first and third alternatives may be destroyed as in the
argument which you reproduce on p. 6. The second and fourth
alternatives are untouched by that argument.
The second alternative is destroyed because the class in
question 1s the totahty of classes not members of themselves.
The fourth alternative remains and is the true one. The truest
destruction of the first alternc:1tive is that no class is a member
of itself : and the truest destruction of the third is by the truth
of the fourth. (It's the fallacy of the indefinite article again.)
By the way, I don't understand your solution of the fallacy.
You have put 1t too shortly for me to understand. Perhaps
you will expand it for me.
I am afraid I am obliged to thmk that a man 1s conceited
as well as silly to think such puenhties are worthy to be put
in print : and it 's simply exasperating to thmk that he finds
a publisher (where ,vas the publisher's reader?), and that in
this way such contemptible stuff can even find 1ts way into
examinations.
J.C. W.
(PS.) 2 Perhaps formally one should add a fifth alternative
class, consisting partly of classes members of themselves and
partly of classes not members of themselves : 1f so 1t would be
easily disposed of, for the class cons1stmg of total of non•
members of themselves couldn't belong to 1t. But on a proper
division this fifth mixed member would not appear.
The division bemg .
Class of all classes

class of all classes members of class of all classes non-members

self of self
A, B, C, &c. 111 fl, y, &c.

J. Coox WILSON.
' [See PS. iDlra J 1 [On a po5tcard, 9th July 15103 ]
 TENTATIVE INVESTIGATIONS
18 July 1903.
§488. I2 Fyfield Road.
My dear Bpsanquet,
In reply to your courteous expression about sending such
decided objections to my views with 'trembling, &c. ', let me
assure you that, in controversy with one other man at least,
I do like the points urged against me put as clearly as possible·
and with all their proper force. If a man from motives of con•
sideration doesn't quite express his full convictions, there is
a danger one may not fully appreciate the ob3ection, and either,
if the thing is nght, one remains in error, or more controversy
is needed to get to a right mind. It is true that a strong
expression of difference, or unqualified accusation of error, does
in the first instance rouse the fighting instinct {in its irrational
form) and that obstructs clear vision, but this has this kind of
drawback rather m conversation, where one is tempted to hit
quickly back, but even so, it is the least of the two dangers :
and in a written objection the danger 1s almost nil, and the
thing is to be absolutely clear. This 1s no affectatlon, at my
age one wants to save time wherever controversy is necessary,
and one is more aware of the limitations of one's own mind and
more anxious for help from others.
I would say this to any one who was good enough to discuss
with me. But to you I would say especially, you cannot possibly
offend me by the most downright and uncompromising state•
ment that I am m error m some particular. Because I am quite
sure that even if you detected me in a pretty bad' mistake, you
do not and would not think I was rather hopeless. I should
not feel abased, for I doubt whether any one has given such
generous commendation to some of my theories as you have,
and it has, I assure you, been a great encouragement to me
that you, especially, have so much belief in me.
I have indeed very flattering 'testimonials' from time to time
from younger men here; but then they have been my pupils,
or attended my lectures, and have got into my ways of speech
and thinking ; and besides are under obligations for mere trouble
taken with them.
But you are perhaps a trifle my senior, and are a 'Master in
Israel', and so I have been greatly delighted and helped by your
 Co",sj,onlence with Bem"1'tl Bosllff(/Wt 1.P
very unstinted e:xpre•ions of agreement on certain important
matters.
I am sending you a discussion of one of your obje.ctions1 and
without any tremors. I am only making clear a point I believe
of importance and you needn't answer at all, and if you wish
to, seeing something obviously wrong, you can delay as long as
you please. We are off actually to-morrow {Sunday): perhaps1
however, I may send in another paper on another of your points.
My address will be Kurhotel, Schonwald, bei Triberg, Baden.
Your very kind letter (July 17) about Dedekind and Royce
came whde I was wnting the paper. I am so glad that what
I said was of some use.
§ 489. There seem to be two main faults in my argument from
your point of view. 1
(i) That whereas I suppose that arguing on the same principle
as 1s shown in the argument by which the paradox that
'the class of classes' 1s a member of itself is proved, it
could be shown in the same way that any class whatever
may be a member of itself . the truth 1s that my argu-
ment 1s not on the same principle.
{u) That m the argument itself there is a non sequitur, viz. the
last statement but one in the left hand parallel column,
'the class so defined as being the totality of groups
which are" groups of things which are x's" 1s exactly the
class X itself'. You say it 1s not this class at all.
It seems to me really on examination that there ,s no serious
difference after all: 1t is almost a mere question of statement.
§ 490. Let me take the first point first. Then of course we
hypothetically consider the argument in itself sound and not
affected by the second objection. We are only asking whether
it falls {if sound) under the principle of the paradox. I think
you only differ because I mean only the general and essential
principle of the argument m the paradox, while you mean
something more detailed, some closer agreement 10 detail.
perhaps.
The argument of the paradox is this :
The class of classes is a 'class',
:. it is a member of the class 'class',
• Notes appended to letter of 18.v1103.
142 TENTATIVE INVESTIGATIONS
••. it is a member of itseli.
The first two propositions here being of the form
YisaZ,
. ·. Y is a member of the class Z.
As to the first proposition : I suppose myself to have shown
that the way 1t 1s got 1s really this :
The class of classes 1s the complex or total of all the classes;
which are themselves totals or unified manifolds. Then it is
argued really that it shares with its subordinate classes the
property of being a totality, or a unified manifold, and so
comes 1 under the general idea. That's how the first proposi-
tion 'class of classes 1s a class' 1s got, whether rightly or not
doesn't for the present purpose concern us. In the case of the
class X, assuming for the moment of course that it 1s rightly
represented as the totality of groups of y things which are x,
arguing in exactly the same way as above we should get that
X shared with its subordinate classes the property of being
a group of things which are x, and thus get by a parallel method
an exact parallel to the 1st proposition m the paradox of the
class of classes. Then the argument continues with the same
parallehsm.
The class of classes is a class, X is a group of things which
are x,
. ·. 1t 1s a member of the class . ·. X 1s a member of the class
class ' group of things which are x'
. ·. 1t 1s a member of itself for . ·. X is a member of itself, for
the class class is the class of 1t 1s the whole group of
classes. things which are x.
This 1s what I mean by saymg that I 'prove' the paradox
about the X class on the essential principle by which the para-
dox is proved for the class of classes-or, 1f you like, on the
principle which really underlies the proof of the paradox, how-
ever it may be stated. I think the above way of putting the
parallelism may very bkely satisfy you on the first point, for
I have avoided what you object to, viz. that resolution of the
class of classes (which by no means 'destroys' its character as
1 MS, • some come, •,
 Cor1esf,ontk,u;e with Bernard Bosanquet 743
you think) in which one is abJe to represent its 'class conception'
as 'being a member of a class' instead of classness.
§ 491. But supposing this difficulty (No. i) to disappear, there
remains (ii), and this is the important one. Consider the class
X of which the class predicate is x, i. e. every member of the
class X is an x and so from this point of view its 'members'
are certainly not classes. I say the same class X can be repre•
sented so that its members are classes. You reply this is [is]
not true. It seems to me there is no more difference between
us than between t.wo people one of whom should say a curve
was convex, and the other should object 'no it is concave'.
Only the parallel is not accurate, for I suppose my position to
be that of the man who says it is both concave and convex.
I will make a suggestion which I think will conduce to your
appreciating the lines upon which I was thinking. You say the
class which is the totality of groups of things which are x, is
Mt X, i. e. it is not the class which (is) the totality of things
which are ,c,
Now I should suggest you should not stop at the mere nega•
tion, but ask yourself the question, if the first class (call it X')
is ,wt X, what is the positive relation between the ISt class and
X? 'X' is not X, how is X' related to X?' I imagine that if
you had asked yourself the question, you would have revised
your criticism and possibly have quite agreed with me.
Now my pomt JS this · 'The Class X as the class X means
a certain part of reality, or say a certain total of reality-it
consists of all the things which are x.
'But the very same class X, as this total of these realities,
is divided into groups A, B, and C, each of which is (a) group
of x's (a.x's, {Jx's, yx's), and so it is a total of such groups-
a total of groups each of which JS an :r group (in contrast with
which it may be called the :r-group).'
Or to put it otherwise :
'One and the same part of reality-or total of realities if you
like-is (i) a total of individuals (say) all of which are z, and
so is a manifold unified by x-ness, the members or elements of
the manifold not being classes, and (ii) it also is divided into
groups of individuals: groups A, B, C. These groups are each
unified by a kind of x-ness (az-ness, fJ:r-ness, rx-ness) within
744 TENTATIVE INVESTIGATIONS
themselves, which make them groups and distinct groups. Now
these groups, taken as units, themselves constitute a manifold :
and they are unified by the common property of 'being a group
of z's •. Thus the totality now appears as the totality of these
groups-its members from this point of view are classes, and
the class predicate is no longer x but 'being a group of x's '.
Surely this is all right, it is but an instance of the conca\r~
and the convex-one reality with two inseparably united aspects.
Supposing you agree to this, the parallel argument [to] which
you object to can be eastly put m a form (for it 's I feel a mere
matter of statement) which would satisfy you.
I might say 'any unified manifold constituting as such a
'"class", even if it, the manifold, is presented as individuals,
• not classes, is also or consists also of a manifold of classes
'unified, &c., &c.' And then I should continue, 'If any manifold
'of individuals constituting an ordinary class is considered from
'this other point of view as the unification of a manifold of
'classes, the argument of the paradox can be applied to it in
'exact parallelism.'
§ 492. Now in truth I had taken to myself some credit for
observing two things which hadn't been observed, and hoped
(may I still hope?} to be patted on the back by you at least
for it.
(I) I think I have shown how a class of individuals is also
(not merely 'may be represented as') a class of classes
(Agamst this the 'class of classes' is simply the class of
classes, but that doesn't matter even if you don't like
that way of putting it).
(2) I have also shown how 'the class of classes' is also the
class of mdividuals defined as 'members of classes' so
that, as in the case of an ordinary class the class con•
ception is altered, when represented as a class of classes,
so in the case of the class of classes, when its members
are represented as mdividuals, its class conception
becomes as I showed m my previous paper 'being
member' of a class.
The parallel argument I first sent you, you wdl see if you
look at it again, combines this double aspect of each class. It
isn't necessar')' it should, for we can simply express the X class
 .
Co"sspatknce 'llJilh Bema,d Bost11UJuel
.
1.JS
in the form corresponding to the class of classes, without any
reduction of the class of classes, and represent the parallelism
as I have done it on p. 3 of the paper. It interested me, how•
ever, to do it that way too in order to make the dass of class11
appear as a case.
§ 493. Now in conclusion I do venture to think that the
theory of the class of classes, &c., is not complete until the
double aspect of an ordinary class and of the class of classes
is recognized: and it 1s not till this is done that one realizes
the full likeness and difference between the 'class of classes' and
what we call an ordinary class. (You will note that what you
called destruction of the characteristic of the class of classes, is
to me the recognition of something essential 1 to it.)

The Heath Cottage,

Oxshott, Surrey.
§ 494. Aug. 2, 03.
My dear Wilson,
Many thanks for your kind letter of July 18th. I noticed
what you point out in --'s review of Adam, but did not know
the inner history. Is he not also wrong about the shadows in
the cave (Adam, ii. 95)?
I almost wish I had gone in for the Logic Profship, it seems
to me that all the real battles arc on Logical ground, and these
discussions attract me tremendously. I will say a word or two,
to state how I think we stand. I do not think what I say worthy
of the discussion you have initiated. It must be taken as
conversational.
I don't think we are far apart as to Y is a Z. You say, if
I understand right, that it means Y is one among Z's, and so,
if applied to The Class of classes is a class, becomes false. I say
it doesn't merely mean 'one among Z's ', and . ·. if applied to
prove 'class of classes is one among other classes', Jt doesn't
prove it. But I am still interested in the point that Class of
classes shares definition of a class. I think 1t a peculiarity
which necessarily suggests a possible abstraction from the
character of the Class of classes as a totality. The abstraction
would not give a truth, but it would be inevitable if we were
' [' essential ' 1s lllegible ID the original 1
746 TENTATIVE INVESTIGATIONS
discussing the nature of the Class of classes as a class, or as
'class' if you like. I am sure it remains an exceptional case
that the class-predicate can be attached to the class as such,
and I know of no natural form for this assertion but 'Y is a Z'.
§ 495. In your analysis of an ordinary class X, you construct,
as it seems to me, a new and artificial class GX, G equalling
definition of 'group'. Of this, the sub-classes gx's are fair ahd
regular members. But of this class GX, x's are not members,
and GX is not a gx, You cannot escape from the law of member•
ship expressed in the class-predicate. X is a gx, but then gx's
are not members of X, so this does not help to show X as a
member of itself. This can only be done if X = GX. But
surely there is no ground for this. Just because x's were not
groups, you have introduced the reflective notion of a group
explicitly into the class predicate m order to be able to make
. .
groups its members. It is your whole instrument for putting
the class apparently on a level with its members, and you cannot
possibly identify a class determined by this elaborate notion
with one in which it does not appear. It is like saying that
you and I are members of the class of municipalities, because
particular municipalities contain us. Of course, X is a class;
but this does not mean that the idea of class determines its
connotation, which in GX it does. Is it not 'a material use of
a formal principle'?
Well; I must break off. I shall try not to write again, and
I know you will excuse me. This got between me and my
Inaugural, and I had to try and ease my mind. I must try to
give up these things to Stout now; they will be his business I
I should rather bke to wnte something about tlus problem
some time. I think there is more behind m the question of
membership.
I hope you are enjoying Germany, and that Mrs. Wilson
is well.
Yours very truly,
B. BosANQU&:T.
 Co",sf>ontlfflu with Bmaa,tl Bosanquet 741
§ 496. Sunday. [August 1903.]
My dear Bosanquet,
We leave in 2 hours so that will excuse the extra bad writing
of accompanying paper. But I thought I had better get this
off my mind before I started. For same reason it isn't as
detailed as I should like it to be for clearness, but, if you've
time, I really think you will make it all out especially with my
previous paper before you.
Yours truly,
j. COOK WILSON.

PS. I hadn't time to notice other details in your paper. Will

look at 1t again when I return in September, but these are the
essentials.

§ 497. On the ntl?am.ng of the general form Y is Z.

Though it is necessary, for the reason I gave in the previous
note, sometimes to substitute a symbol for a significant word,
a symbolic statement like the above is inadequate and may
mislead because the only word with significant grammatical
form in it is the verb 'is'. The other symbols have nothing to
represent the grammatical terminations, which are absolutely
essential for determining the meaning of a bngu1stic form. There
would be less chance of error 1f one could by some barbarism
give a kind of adjectival form to the predicate symbol, (thus.)
Y is Z-ical, Y is Z-y, Y 1s Z-ible,
for the meaning of the predication 1s quite indeterminate with•
out the precise grammatical form which follows the word 'is'.
(From this point of view the form Y 1s a Z has advantages-we
will return to it presently.)
Of course it is true that language is elastic and that not only
are certain artificial forms introduced for technical purposes like
Aristotle's A j1r&pxn T'lfl B, and 'whiteness is inherent in snow'
or 'whiteness is in snow'-but also linguistic forms get away
from their original meaning otherwise.
But now I must in this reference hope to gain assent to a point
which is to me of the last importance. It is the business of the
student of Logic to determine the normal use of an idiom or
z
 TENTATIVE INVESTIGATIONS
linguistic expression. Everything depends on that. If this is
once done, one gets a hold over the 'implications', of which you
speak, and derived and extended meanings. It not only gives
the key to fallacies which come from the confusion of different
meanings under one form of expression : but it prevents certain
mistakes m the logician himself. This is not sufficiently appre-,
ciated by very able logicians · they do not humble themselves
sufficiently to the banausia of grammar, and pay for it. For
instance, in Bradley's Logic there are very serious consequences
of this kind · much that he has to say on some topics is merely
worthless in consequence.
§ 498. I. The form Y 1s Z may represent what I call equa-
tional predication, where Y is the total realization of the universal
in Z. But whether it 1s so or not can only be learnt from the
word-form of the particular judgement itself. The symbolic
representation tells nothing.
2. The form may also represent what I call adjectival pre-
dication, when Z has the form of an adjective, or some gram-
matical phrase equivalent to an adJectivc. 'This blotting-paper
is red'; 'this letter 1s heavy' About the normal meaning of
the adjectival predication, I should have thought there could
be no manner of doubt Language has developed certain
different forms for special and different purposes, and that is
their normal use as opposed to shifting and change. For instance,
(it) is not for nothing that language has developed the two
forms 'all AB is C', and '1f A 1s Bit is C'.
Now I venture to say that the adjectival form of predication
has been developed for the precise purpose of expressing the
case in which the universal referred to by the predicate is either
known not to have the subject Y for its complete realization,
i. e. only realization, or, where this is a matter of uncertainty,
to leave 1t an open question, and not state more than that Y is
at least a realization of Z.
Thus when we say Y is Z-ical (or Z-ible or Z-ial, &c.) the
reason for choosing this form of expression rather than any
other-the reason why language has developed it-is that we
do not mean to state more than that Z is realized in Y, that
Y 1s at least a realization of it, whether we are sure there are
more or simply don't know whether there are.
 C o " e ~ with Bmuwd B0$tffifuet 749
I hope that th.is way of putting it may satisfy you that it is
quite an irrelevant objection to say that often in such a judge-
ment as Y is Z we don't know that Z has more realizations than
in Y. Just because we don't know we choose a form of expres•
sion which commits us to no more than we know to be true or
think true, viz. that Y is at least on, of the realizations of Z.
And I think it a conclusive answer also to the objection, that
if we do intend to affirm that Y is the whole and sole realization
of Z, we should never naturally choose the adjectival form of
predication. Not only so, but I suppose 1t literally true that
we never do, in ordinary speech (of course artificial or technical
expressions invented arbitrarily to support a theory do not
concern us).
§ 499. The above applies without change to the form Y is
a Z. There 1s a reason why we say 'Socrates 1s ci man' on the
one hand and 'Socrates is white', without the mdefmite article,
on the other, which docs not here concern us, and I won't dwell
upon 1t ; but its normal use always indicates that we do not
state more than that Y 1s one of the realizations of Z, and the
normal implication is that we at least leave open the possibility
that there are more. This form again draws attention through
the indefinite article to the fact that the subject Y is being
represented not only as having Z realized in 1t but as a member
of a group of similar tlungs in each of which Z is realized.
I venture to think that the soundness of this view is brought
out by the fact that 1t 1s upon the well-understood normal
meaning and normal implication of a given phrase that pre•
varication often depends. It is this that makes certain kinds
of this form of deceit possible. Suppose a man described to
a foreigner really every part of the British Empire, and after
finishing the list said, 'All that is a territory now in possession
of the British'. The foreigner would of course suppose it was
not the only British territory and would expect "to hear about
some other. The speaker might take advantage of this by then
enumerating certain parts, not before mentioned by name, which
fell within the whole he had mentioned and calling them British
territory. Some practical end might turn on this. When it
was secured, and the deceived man rounded on the prevaricator,
the latter might claim he told the truth : that his first statement
Z2
750 TENTATIVE INVESTIGATIONS
'all that is o British territory' was true. Whereas the fact is
he lied : for he knew his statement would be taken in its normal
sense. If he appealed to Logic to defend him, his opponent
might well conceive that- Logic was some contrivance for justi•
fying knavery.
§ 500. This, I must maintain, must be applied without any.
abatement to the statement that the class of classes is a class.
It is only because the idea of the class of classes is new to us
or unfamiliar that it seems strange not to be able to say that
(it) is a class. We don't realize in so speaking what the 'class
of classes' means. To say that it is a class is as mistaken as
to say that Space 1s a space. Just because Space is thought of
as a totality we never speak of it as a space.
Just as you call Space 'Space' without any article, so you may
1£ you please say 'Class' as equivalent to the class of classes or
the totality of classes, but you cannot without doing violence
to the normal and understood meaning of language call it a class.
But now people do violate the normal meaning of language
in order to take advantage of it-and henrc, as I have said,
prevarication or paradox. And this statement that the class
of classes is a member of itself because it 's a ~lass is a mere
instance of unconscious prevarication. The statement is only
of any interest because 1t is paradoxical. If membership was
so construed as to mean the whole of the members of the class,
then that the class of classes 1s a member of itself only means
that it is identical with itself or with the totality of its members
in the proper sense of members. But then it would be a silly
tautology and no one would take (the) slightest interest in it.
It is interesting because a member is understood naturally to
be a part of a whole group, and so the sentence implies that
somehow the whole is identical with one of its parts. That 's
the whole story. Now such a result comes from the prevarica•
tion: to call a thing a Z normally implies that it is one among
other Z's, or at least that we don't state it 's the only Z, and
normally leaves open the possibility that there are other Z's.
But if you do more, and say that Y is a Z and therefore that
Y is a member of the class Z, you go further than merely leaving
the possibility of more members open, or of other realizations :
the normal implication is that there are other members.
 Co,r~ondence vnth Bernard Bosecnguet 751

Thus (i) if we say the Cl yof cl is a Classz and go no

ass• • apes
further, we normally imply that there is at least a possibility
that there are classes outside the class of classes, but even the
possibility is false 1 and so our statement is really false.
(ii} H we add 'because the class of classes is a class, it is
a member of itself ' we commit ourselves to a slightly worse
falsehood, for the normal implication of this is not only that
there may possibly be classes outside the class of classes, but
(that) there are actually such classes.
Similarly I thmk there is no difficulty whatever in showing
the trivial fallacy of language involved in making any class
a member of 1tself.
J. c. w.
 XI
1
ON THE NOTIONS OF A CLASS AND OF CLASSES

[Oxford, 23 September 1903.]

§ 501. (The distinction of the 'class as one' from tlie 'class as
many'.
The accurate account of 'a class' 1s that 1t is a unified mani
fold; but 1t is not any umfied manifold that is a class. For
instance, an animal organism is a unified manifold of organic
parts, but it is not a class.
A class 1s a particular kmd of unified manifold and its d1s•
tinctivc character hes m this that its umty consists m the
realization of one and the same universal m each of its con•
stituent members. This is the same as saying that 'the unity
is constituted by the truth of a certam assertion concernmg
each of its members severally' ; though I prefer the objective
way of puttmg it
§ 502. We can therefore distinguish between the umty and
the mamfoldness of the class, and this is a valid d1stmction, for
the unity of the class 1s not I ts manifoldness, though the two
are inseparable from each other.
Now, of coun,e, the class as one (or as a unity) is not identical
with the umty of the class, and the class as manifold is not
identical with the mamfoldness of the class. We therefore
cannot argue from the above true statement 'the unity of the
class 1s not its mamfoldness' to 'the class as one (or as a umty)
is not the class as a mamfold '. Indeed the latter statement is
not true ; on the contrary, 'the class as one' is •the class as
manifold•.
It must be remembered that we have to do with the distinction
of the class as one from the class as many, as 1t appears in the
1
Letter to Dr. G F Stout, F BA, Professor of Logic in St Andrews
Un1vers1ty.
 On the Notions of a Class and of Classes 753
statement that the class as one may be a member of (stands
therefore in a certain relation to) the same class as many.
§ 503. Suppose A is X and A is Y ; suppose also that X and
Y necessitate one another and are inseparable.
Then A as X is Y and A as Y 1s X ;
Suppose A as X is B, then A as Y is X, therefore A as Y is
Bi and A as X and A as Y have exactly the same predicate
and can't be distinguished by difference of predicate.
Further, let A as X stand in a relation R to C : then A as Y
must as X stand m this relation R to C.
Let C ""A as Y.
Then, by the above, if A as X stands in a relation R to A
as Y, A as Y stands 10 the relation R to A as Y, which is
nugatory.
And 1f 1t has been asserted that in consequence of the distinc-
tion of A as X from A as Y, the relation R 1s possible betwee12
A as X and A as Y, this statement will be nullified, for A as
Y will stand in the relation R to A as Y, where there is no
distinction of the terms related.
This will be still clearer 1f we suppose a relation R, which is
not convertible. If we suppose, that 1s, that M 1s in a relatiOll
R to N but that 1t doesn't follow that N is in the relation
R to M, but ma relation r to M. For instance, 1f Mis greater
than N, N is not greater than M.
Suppose then R to be such a relation that if M is in the
relation R to N, N ib m the relation r to M, R and r not being
coincident.) 1
Now 'A as X stands m a relation R to A as Y' means that
A, because it is AX, stands 10 the relation R to A, because A is
AY ; but, smce Y necessitates X, AY is AX ; or A, because it
is AY, is AX.
And since X necessitates Y, A 1s A Y, because A 1s AX.
Therefore it follows from 'A because it is AX stands in the
relation R to A, because it is AY' that A because 1t is Y (and
is therefore AX) stands in the relation R to A because it is X
(and is therefore AY), i. e. A as Y is m the relation R to A as X.
Now it has been shown that A as Y is in the relation r to
1 The part in brackets is lost from the ongmal letter and JS supplied from
Wilson's manuscnpt notes.
154 TENTATIVE INVESTIGATIONS
A as X, but this is impossible on account of the relation between
R andr.
If we substitute in the above general theorems for A, 'the
class' i for X, •a unity' ; and for Y, 'a manifold' i and for R,
•the relation of member of', the above refutation will immediately
apply to the statement that a class as one can be a member ,of
itself as many.
For the unity and the manifoldness of the class involve one
another in the way necessary for X and Yin the above theorems.
The unity of the class only exists as a umty OF the manifold
in it, and thus the class itself-not the unity of the class, but
the class-is 'one' only as a unified manifold. That is, the unity
of the class is such that the class as one is a unified manifold.
Also 1t is only a manifold as the pluralization or manifoldness
of a unity. That 1s, the manifoldness of the class is such that
the class as manifold 1s a manifold unity. Thus as one it is
manifold, as manifold it is one ; 1. e. the nature of its unity
necessitates its manifoldness, the nature of its manifoldness
necessitates 1ts umty. This corresponds to the inseparable rela•
t1on of X and Y, and 1L ,,ould then follow that 1£ the class
'as one' was a member of itself 'as many' (rather 'as a mani•
fold'), that the class as one was a member of itself as one and
that the class as 1na11y was a member o{ itself as one.
§ 504 Or agam we may put 1t otherwise thus:
The expression 'A as B' 1mphcs that the statement 'A is B'
is true. The expression therefore 'the class as many' should
imply that the statement •the class 1s many' 1s true. But this
1s not a proper use of language. The class as class is not many,
it is one : it is the members of the class that are many, and it
is of them that 'many' is properly predicable, and in such
a statement. The best ch,mce for makmg possible the d1st1nc•
tJon of the class as many from the class as one is to substitute
for ' the class as many', •the class as manifold'.
The class has something to do with • manyness ', with 'many ',
with 'a many': but what is it exactly? The class has only to
do with ' manyness ' or •a many' as being the unification of
a many. The expression then 'the class as many' if allowable
at all should mean 'the class as having the aforesaid relation
to many-ness ', that 1s, 'the class as unification of a Qlany'.
 On the N anons of a Class and of Classes 755
Thus the expression •clus as many' must be capable of having
substituted for it • class as unification of a many', and conversely~
But the class is one as unification of a many: therefore the class
as one (a) is the class as unification of a many (b). Substituting,
in accordance with the above, for the second phrase (b) in this
last sentence, we get 'the class as one is the class as many'.
Whence to say that the class as one is a member of itself as
many is to say that the class as many is a member of the class
as n1any ; and we should also easily derive the other statements
already given above (§ 503 fin.).
§ 505. But however this may be, I should have thought the
idea of a class as a unified manifold made it easy to disprove,
shortly and simply, that any class can be a member of itself
We have but to consider what membership of a class means.
A class being a umfied manifold, 1t 1s the elements of the
manifold wruch arc the members of the class. That is the whole
story. Consequently it 1s a contradiction to say that a class
can be a member of itself, for that would mean that the umfied
manifold of elements was one of the elements urufied. Hence
it would follow that any belief that a class could be a member
of itself must depend on some fallacy of language. It doesn't
matter what class 1t is-the 'class of classes' included. This
would be qmte certain a priori, before examining any particular
statement : and I have tried to show, in previous papers, what
the lmguistic fallacies exactly arc. I should have thought the
above refutation, short as 1t 1s1 sufficient and decisive, and that
the only matter of dispute could be, not whether there was
a verbal fallacy, but what the exact nature of lhe fallacy was.
§ 506. It follov.s that I am (in) entire disagreement with your
two papers, so far as these presuppose (I) the distinction of the
class as one from the class as many, in what seems to me an
untenable form, and especially (2) that a class as one can be
a member of itself as many.
Nevertheless, as criticisms are asked for (by you) on these
papers, I may venture to discuss certain parts of them in detail.
§ 507. Distinction of
(a) class taken simplicittr,
(b) class quA class1
.(c) class taken as many.
756 TENTATIVE INVESTIGATIONS
Conaider what would be the natural interpretation of a dis-
tinction so stated, before considering the interpretation given of
it and the use made of it.
If the words are to bl.' taken strictly, the distinction is of the
general form
(a) A taken simplmter,
(b) A qua. A,
(c) A taken as B =- A qua B.
A term like 'simpliciter' does not seem very suitable, when
clear analysis 1s aimed at, for it may mean various things and
itself requires mterpretation. We ask how can A simpliciter
differ from A qu! A~ 'A simpliciter' seems naturally to mean
'A pure and simple' (&w>.&i11) 1 as distmct, e.g., from any analogous
or metaphorical uses of the term A ; but that is just what is
meant by A qua. A, and so (a) and (b) would be indistinguishable.
But perhaps the formula is not to be taken strictly, and we
should read each time for ' class', 'a class ' : and then the general
form will be
(a) an A simpliciter = a particular A simpliciter,
(b) an A qua. A = a particular A qui A,
(c) an A qua. B = a particular A qui B.
And this turns out to be the case : for the application of the
formula shows that 'class qua class' means a particular class
considered as class only in general, and not as the particular
class which 1t really is ; and from the same context one gathers
that by 'class simpliciter' is actually meant 'a class as bemg
the particular class which it is', or 'a class qui the class it is',
or 'qua itself'. The distinction, then, between (a) and (b) is
merely the distinction of the particular or concrete (the particular
class) from the universal (class m general).
§ 508. This being so I cannot understand what is gained by
describing the distmction so artificially, not only by the odd
use of 'simpliczter', but in the formal explanation given of 1t.
In this latter explanation, instead of the simple statement that
by 'a class simpliczter' is meant a particular class as particular,
we find an account which seems to me to be indirect and con•
fusing. One doesn't gather what it means till further on. But,
besides this, the form of this explanation seems to me open to
criticism. The procedure is to explam •class simplicit8r' by
 On the N onons of a Class aHd of Classes 157
appeal to certain linguistic forms : the meaning of words in
certain statements. This may have seemed simple in itself and
clear, but I confess it does not seem to me to be so. Is there
any particular reason for taking such an example as m,n? No
reason is given and yet there is one, for it would not do to
substitute, e. g. 'the class "classes'" for 'the class "men 11 ' .
What is exactly meant by the 'complete meaning' of the word
'men' ma certain statement? The phrase may have seemed so
clear that the question may be held captious, but I cannot think
it clear. To what kmd of meanmg of the word 'men' is the
'complete meaning' opposed. The next sentence, 'it is equi•
valcnt, &c. ', suggests thJ.t equivalence is the test. But in such
an appeal to verbal forms, the equivalent should be capable of
bemg substituted for that of which it 1s the equivalent. The
context requires, then, that the 'class man simpliciter' should
be equivalent to 'every man' m such a propos1t1on as 'every
man 1s mortal' ( = every man 1s a mortal), or to 'all men' in such
a proposition as 'all men are mortal'. But this seems to fail,
for we can't substitute 'the class men' for 'every man' in the
above propos1t1on ; for 'the class men is a mortal' would not
be true, nor mdecd 1s 'every man' equivalent to 'all men', as
the context implies : e. g \\-e cannot in the sentence 'all men
should love one another' substitute 'every man' for 'all men',
nor can we in '.all men form a commumty '.
§ 509. Now compare your page 2, where a. similar procedure
1s adopted to define (a) 'class taken as many'. It will be found
that this, if taken literally, can be made to coincide with the
account given of (a) just discussed-'we refer to the class
"men" as many, \\ henever we speak of a man, or of "some
men" or of "any man" '. But as 'a man' 1s here meant, the
statement 'a man 1s mortal' is equivalent to 'every man is
mortal'. So agam 'any man must die' 1s equivalent to 'every
man must die' or to 'every man is mortal'. So that the above
amounts really to explaining 'the class "men II as many' by the
way •we speak of' every man, m 'every man is mortal' ; whereas
•the class " men" simplicite,' has been explained in the self•
same way.
It seems to me that this sort of thing is always liable to
happen to us, when we are trying to define something quite
 TENtATIVB INVESTIGATIONS
simple and well understood. Because it is perfectly well known
what object we are defining, what we really mean, we may
overlook the fact that our formulae are not adequate, or are
indirect. We must ask whether any one who didn't know what
was meant beforehand would discover 1t from our formulae,
For instance, in the case before us, would a foreigner who did.
not know the meaning of the English word 'class', or of its
cognates m other languages, really find out from the description
what 'class simpliciter' meant? And would he, from page 2,
understand there was any difference between 'class simpliciter'
and 'class as many'?
I am obliged, then, to think that the method adopted for
explaining the distmction really intended between (a) and {b)-
not to speak of (c), 'the class as many'-is curiously artificial
and indirect, and that it does not really explain the distinction.
Even if it wdl not be admitted that it could be misunderstood,
it is at least assailable, whereas on the one hand such defimtions
are bound to be accurate and must submit to the test of being
taken literally and strictly, and on the other hand the distinction
can be stated simply and without possibility of misunderstanding.
§ 5IO. Further, it seems to me that whereas the account which
I offered was rejected, the present account reproduces the
essentials of a distinction which I gave-as something merely
obvious to anybody-between the 'class of classes' and the
classes subordinate to it (or of wluch it is composed), but asso-
ciates this d1stmction with an .i.rtlficial formulation which tends
to obscure it.
I may be allo\\cd to recall, summarily, the distmct1on as
I gave it. A class is a unified manifold, m which the clements
of the mamfold arc unified by the appearance in each of a.
common characterh,llc, This account, as common to every
class, constitutes their 'classncss '. On the other hand, any
particular class is umficd by some special common charactenstlc,
and not by a. common characteristic m general. And it is this
special or particular nature of the common characteristic which
makes any given class the special or particular class that it
is, e. g. such special characteristics as 'humanity', 'redneas '1
'triangulanty '. What is called the 'class of classes' is obviously
therefore the totality of such special unifications of manifolds.
 On the Notions of a Class and of Classes 759
Juat as triangularity is the common characteristic of each
member of the class triangle, so unification of a manifold by
a common characteristic is the common characteristic of the
members (themselves classes) of the class of classes.
§ 51 I. Not only so, but an important conclusion which I
arrived at reappears, though in a somewhat different dress,
viz. that classness cannot be predicated of the class of classes.
This was one of my principal points, and one I laboured to
elucidate. I do not suppose this is a conscious acceptance of,
or coming round to, my view, because you are so ready always
and so frank in acknowledgement, and it would be difficult to
account for your opening expressions of regret at disagreement.
If this is only carried out consistently, it results, as I think
I have shown, in the refutation of the fallacy that the class of
classes can be a member of itself. But though there is this
important agreement in the conclusion, is the method of arriving
at the conclusion successful? Would it really convince a
doubter? Is it not open to an objection for which no provision
is made? • If you call "the class •classes'", the class •classes',
how can you refuse to predicate classness of it?' Would it not
also be inevitably said that m your statement there is a mere
assertion of what has to be proved. that the question is begged,
and no answer given to the above objection? For my own part
I am convinced the objection cannot be answered except by an
analysis of the meaning of idioms of predication, •A is B
(ad1ective) ', 'A is a B ', &c:., and their relation to class member-
ship. It is a matter of linguistic analysis almost solely. I
adopted this method in the accounts submitted, and the account
now under discussion only confirms my conviction that this is
the only method by which these paradoxes can be dealt with
(inasmuch as they arise entirely out of a misuse of language).
And I venture to thmk the use of this method makes the whole
matter transparent ; shows clearly that the idea of any class
being a member of itself is a mere verbal fallacy. It seems to
me also to settle the supposed contradiction in two ways,
(1) because the' contradiction' is grounded on the mistake about
a class being (a) member of itself; and (2), in the way which
I have already detailed before.
§ 512. With respect to certain developments which I added
 TENTATIVE INVESTIGATIONS
to the criticism of the fallacies-though not necessary to it-
these depend in part at least on a characteristic of classes of
which account had not been taken, which may be put shortly
as follows : Consider a mamfold of elements unified as a class.
The elements are members of the class. This manifold of
elements may be divided into sets, each set itself a manifo41.
unified by some common characteristic. These unities (i. e. the
sets} constitute a manifold, and the unity of the whole may be
presented as the unification of this manifold, the members of it
being the sets, and themselves unified manifolds.
[One section omitted ]
§ 513. 'A class 1s essentially a distributive unity.'
'A class collection is essentially a collective unity.'
This distinction seems to me (to be) put in an untenable form.
I think I understand what is aimed at, but the formulae do not
seem to me to express it ; the distmction being not really due
to the difference between two kmds of unity between the
members of a class.
{i} What would be naturally understood by a 'distributive
unity' in such a connexion? That a certain unity
{universal) manifested itself m a plurality of particulars.
This agrees with the definition given of 1t-'constituted
by the truth of a certain assertion concerning each of
the members severally'. The common predicate repre•
sents the unity manifested in each particular. Shortly
then the 'distributive unity' means a pluralized unity.
{ii) Secondly, what would be naturally understood by a • col•
lective umty' of a class? That a certain number-or
an indefinite number-of particulars, though many, were
collected into a unity, or unified, by something common
to them : that is to say, a unified plurality. Thus the
distributive unity would be a pluralized unity (manifold
unity) and the collective unity would be a unified
plurality (unified manifold). But these are merely two
verbal methods of describing the same thing.
§ 5I 4. You say, however, that for the collective unity 'it is
essential that its constituents should be connected by some other
form of connexion than that which simply consists in the truth
of a certain assertiQn concerning each Qf them severally'. (One
 On the N otiOffs of II Class tm4 o/ Classes 761
observes here that both unit~s are conceived as having con•
stituents; that is, both as pluralized unities or unified pluralities.)
Prima facie one must wonder what the •other fonn of combina•
tion' could possibly be ; for one would suppose the unity in
question must be manifested in each particular, and inevitably
give rise to a common predication. Unfortunately no illustra•
tion is given : and the general formula, 'any collective unity
having for its constituents all members of a class', is (I should
have thought) exactly apphcable to a class and to what has
been called a distributive unity.
§ 515. But one may judge, by help of the other paper and
also by apparent implications in this one, that the collective
unity has somehow corresponding to it a predicate not predicable
of each constituent severally and that the real intention is to
distinguish between the two forms of predication illustrated in
these two examples :
(a) Every man is a vertebrate animal.
(b) The class man is a species of the class vertebrate, or, is
a member of the class of vertebrate classes.
For we canno( predicate the predicate of (a) of the subject of
(b}, the class man is not a vertebrate animal; nor can we
predicate the predicate of (b) of the subject of (a). The di'ltinc•
tion between (a) and (b) seems to me certainly not due to a distinc•
tion between two different kinds of unity of constituents at all.
§ 5 r6. But before coming to this, I think 1t may be shown
that, if the distinction of (a} and (b) is, as I imagine, the fact on
which the distinction of the 'distributive unity' and 'collective
unity' is founded-(then) the characteristics above assigned as
distinguishing these two unities can hardly be maintained ·
viz. that in the distributive unity the constituents have severally
the same predicate, but that in the collective unity there is
some other bond of connexion. For 1f the collective unity
is that represented by the class •man' in (b), subject of a singular
verb and with one predicate, and similarly that represented by
the class 'vertebrate' in the statement • the class vertebrate has
the class man for a species', then such 'collective unity' owes
its collective unity precisely to the bond of union expressed by
the predication of the same thing of its constituents, and to
nothing else.
 TENTATIVE INVESTIGATIONS
The 'collective unity' or •class collection' represented by the
•class vertebrate' in the above statement, with its singular verb,
is unified solely by the common principle which gives each of
its members the common predicate ' vertebrate '. That is to say,
its unity is made in exactly the same way as that assigned to
the distributive unity, and 1t has not 'some other bond ,o!
connexion' besides this. So also ' the class man' representing
a 'coJJective umty' or 'class co11cctJon' m (b) owes its unity
precisely and only to the fact represented by the predication of
one and the same universal (vertebrateness and whatever else
1s common to men as men) of its members. Supposing, then,
that the right kind of example has been given of what is intended
by a collective unity, the prima f acie impression above stated
that the 'collective unity' could have no other bond but that
which makes the 'distributive unity', would be confirmed.
§ 517. Now as to the difference between (a) and (b).
The class being a unified manifold, as described (by you) is, as
a unified manifold, a unity, and so can have a predicate attached
to it as a unity, a predicate which cannot be attached to any
of its members individually. This corresponds to the instances
of the form (b). But more than this, there is no other form of
predication of the class as class possible. As a unified plurality
or pluralized uruty, it is a unity and all its predicates as its
predicates are of it as a unity. Thus (b) 1s the only real form
of predication about a class as a class. But we may also pre•
dicate something not of the class-the unified manifold as such-
but of one of its manifold members, or of some severally, or of
each severally. Even in the latter case there is no predication
of the class as class, the predicate is attached only to each
member as individual, e. g. 'each triangle is a figure with interior
angles, &c.', but we can't say 'the class triangle is a figure, &c.t.
This predication of each of the members, and not of the class
as such, is represented by such instances as (a). Apparently,
however, (a) is regarded as a case where something is predicated
of a class 'as many'. As I have contended before, the class is
not many, it is its members which are many, and to speak of
the class as many seems to me a loose way of speaking, which
leads to serious confusions when we introduce it into strict
analysis.
 On the Notions of a Class and of Classes 763
§ S18. However-and this is what probably leads to what
I have ventured to deprecate as a confusion-from such pre-
dication as (a), which is not a predication of the class as class,
a true predication of the class can be derived : e. g. 'the class
triangle is such as to have each of its members a figure whose
interior angles', &c., &c. ; 'the class man is such that each of
its members is a vertebrate animal.' And there is a further
relation which may simtlarly promote confusion. Though the
predication (a) is one in which the predicate can only attach
to each member of the class, and not to the class as class, that
is, not to the class as the unity of such members, nevertheless
the fact that this very predicate, which attaches to each severally
and cannot attach to the class as class, is present in all, makes
a unified manifold of them, that is, makes a class as class.
E g. the property predicated of each triangle umfies them all.
The above seems to me the true analysis ; the explanation
not bemg by what I venture to thmk an untenable distinction
between a distributive unity and a collective unity. And the
distinction between (a) and (b), which I imagine a simple matter
notwithstandmg the length at which I have treated it for the
sake of clearness, seems to me merely this :
In (a) we have a predication in which the predicate can only
attach to each member of the class as a separate member, and
not to the class as class ; m (b) a predication in which the
predicate attaches to the class as class
To avoid misunderstanding of the use in the above of the
expression 'class as class', I may say there is no mystery about
it; it simply stands for class, but is put for clearness as against
the representation of a class which seems to violate the idea of
a class as such. And so by ' class as class' I only mean 'a class1
if it is to be a class at all'.
J.C. W.

1773.2 Aa
 XII
PRIMARY AND SECONDARY QUALITIES
Lette, in criticism of a pape, on P,ima,y and.
Seconda,y Qualities
July, 1904.
§ S19. Some preliminary points.
According to the formulation of Locke {e. g.) the secondary
qualities, which are qualities of the external object, are strictly
speaking not the sensations, as heat and colour, produced in us
by the object, for these are not qualities of the object, but the
powers in the obJect to produce the sensations. Still the sensa-
tions themselves-inaccurately I think-get called secondary
qualities I observe this m the present article. It doesn't
matter, but 1t is well to remember it or it tends to confusion
in thinking.
For clearness m a discussion bke the present it is well to
confine attention to an example hke heat and to exclude colour.
For m the case of heat every one easily recognizes, without
philosophy, that he himself has a definite sensation of heat-
we know this to be a sensation inasmuch as we say we feel hot.
This is not so with colour People do not easily recognize that
colour is a sensation of their own On the contrary, the belief
that 1t is a sensation is an inference from a theory scientific or
philosophic. We all of us, both the plain and the philosophic,
think of colour as in the thing and put it there in a way we
never put our sensation of heat. Thus, too, ordinary people say
they feel hot, but never say they feel coloured, e g. feel red.
Moreover, whereas they speak of a sensation of heat, they never
speak of a 'sensation of yellow'. It is only the scientific or
philosophic who speak thuq in consequence of a theory which
they hold, and that only because they think they have proved
to themselves it is a sensation, not because they recognize it
directly as such : for it 1s not natural to do so. A connexion
with sensation becomes apparent when the colour dazzles us :
but even then we think of the colour as in the thing and of the
dazzling as its effect upon us. That colours are only seen in
 On Prima,y and Secondary Qualities 76-'
the light makes no difference (and here Locke's argument is
wrong), for we simply think of their colour as a property ehey
have in the light, just as we think of floating as a property cork
has in the water.
We may therefore for the present purpose when speaking of
secondary qualities confine ourselves to the sensation of heat and
its correlative in the thing.
§ 520. Consciousness of the plain man.
I don't think you give a quite accurate representation of the
theory of the plain man's consciousness implied in the view you
are attacking. It needn't be assumed that he definitely thinks
his sensation as in the hot iron, so as, e. g., to be liable to what
is called a 'flagrant absurdity'. Rather 1t would be supposed
that in feeling a hot surface he tends to think that besides
having a sensation of burning or warmth, he is feeling something
in the thing which causes it, just as when he feels a rough
surface he not only has a special kind of sensation but recognizes
something different from 1t, viz. an uneven surface. Only that 10
the case of heat the only pos1t1ve nature he assigns to the
supposed quality felt m the thing 1s derived from his own sensa-
tion. So it might be said that by a confusion he assimilates
the character of the quality attributed to the thing to the
character of his own sensation-a confusion in which he 1s not
fully conscious of what he is really doing.
(Oddly enough, this seems to be exactly what your own theory
comes to, as I shall try to show presently, so that you don't
really differ from the theory you are attacking.)
I shouldn't mind calling this an unconscious transference of
the quality of his own sensation to the thing ; but he certainly
distinguishes from it his own feeling of being burned (and pained
possibly) or of being warmed (and pleased possibly). The quality
of his sensation thus {does) double duty; and I am prepared
to show that this is done in Bacon, and that it 1s the key to
the standing puzzle about the causa formalis or forma.
I should be quite prepared to find that some people who hold
the theory that in heat there is really nothing in the body but
a state of its primary qualities {dynamical)-especially scientific
people who think crudely on these things-have said the plain
man simply thinks the heat he feels (his sensation) is in the hot
Aa2
 TENTATIVE INVESTIGATIONS
object ; but that is merely an awkward, perhaps unintelligent,
presentation of the consequences of their theory, and the philo-
sophic view would be, I think, what I have represented.
I think it is impo1tant to notice that we have not to deal
with the plain man here as opposed to the philosopher. We are
all 'plain men' m this sense. It represents the operation of.oJlr
ordinary unreflective consciousness : and it is because we who
are reflective and philosophic have this consciousness that we
are able to deal with it. Thus it is true that the only course
we can take is to ask the plam man (m this sense) what he
means. As philosophers we mterrogate ourselves as plain men
§ 52 I. On the 'secondary quality' as a 'mere possibility'.
Unless I have mistaken your dnft, you seem to treat the
theory you are cnt1c1zing as if it implied (that) the Secondary
quality m the thing was a mere power to produce a sensation in
us, and you seem to be controvertmg that and trymg at least
to show that our 'plam consc10usncss '-' ordmary unreflectmg
thought '-considers the quality as havmg some actual nature
beyond bemg a mere power ( = mere possibility). But the
moderns who hold Locke's doctrme {and I should say, with
doubt, Locke himself) do not hold this at all. They suppose
that correspondmg to our sensat1011 of heat there is an actual,
not potential, activity of the particles themselves, a vibratory
movement-a dynamical state which is entirely comprised within
what Primary qualities mean for them. This exists whether
there 1s any sensation or not and it 1s a potentiality only with
regard to sensation. If there 1s a sensitive surface upon which
this vibratory movement can act, then it produces a sensation
of heat. Your argument certainly gives the impression that you
suppose yourself correcting a pomt m the Lockean view, and
ins1stmg on some actual or positive quality m the body said to
be hot, other than the mere possibility of producing a sensation
in us. If this is what you really mean, it is a misconception.
The thing you are trymg to establish against the theory is an
integral part of the theory itself, i.e. so far as you are merely
trymg to prove that the quality m the thmg is not merely
potentiality of sensation
It seems to me that there is not only this false issue in the
paragraph before us, but that it has helped to put you off
 On P,imary and Secondary QuaUties 76'/
the scent, and to prevent you from realizing the true significance
of the 'fact' to which you appeal. I shall return to the last
point presently.
Perhaps you do not intend your contention as to the actual
and positive nature of the quality in the body to be a criticism
of the Lockean theory ; but I feel sure every reader would
suppose you did, and therefore I should suggest in re-writing to
make your meaning plain. But m any case I venture to think
that it JS your attention to this partJcular point as following
from (' shown by'} the fact to which you appeal, that has obscured
for you the main significance of that very fact.
§ 522. Sensation without representative value.
There 1s another point I should hke to clear off before coming
to grips with the main issue. You suppose that the gradual
diminution of heat sensations as the sensitive organism is with-
drawn from the hot object 1s without representative value.
Here you have overlooked the fact that, according to your own
meaning of representative value, this change has a very definite
and very important representative value. It obviously corre-
,;ponds to the objective physical fact that the heat in a body
has effects at a distance from itself, and that these effects
diminish in intensity as the distance increases: this is as
objective as the dm11nut10n of attractive force with the distance.
It shows itself not only m our sensations but m the fact that
a body gets warmer from the heat of the warming body as it
gets nearer to it
This suggests another point which you don't notice and should,
I thmk, be kept m mind m the discussion of secondary qualities.
Sometimes we feel the warmth or the cold of a body by touching
it and when we touch it, then it is easier to confuse the quality
which belongs to our sensation with a quality in the body. Bttt
how about the case where we feel warmth 'in the hand', when
the hand is advanced to a certain distance from the object?
Here we do not naturally say we feel anything in the object,
but I suppose we thmk of ourselves as feeling heat conceived
as something in the space between us and the object, and 'coming
out' from it (think, e. g., of warming the hands at a good hot
blazing fire). And in this case we clearly and definitely locate
a sensation of warmth in the hand. And here I would call your
 TENTATIVE INVESTIGATIONS
attention as a psychologist to a very striking difference between
heat as a sensation and colour (1f it is to be called a sensation) ;
though the perceiving organ is always at a distance from the
perceived coloured object, we always locate the colour in the
perceived obJect and never in the perceiving organ. In the case
of heat felt at a distance from a hot object, on the other hans),
we locate the sensation of heat in the hand (e. g) and the 'heat'
(as obJective) which we suppose ourselves to feel {through the
sensation) we do not locate in the object, though we may
connect 1t with the object.
§ 523. The theory criticized and the theory opposed to it.
The theory criticized seems properly to be this : Heat in the
hot body is only a vibratory movement of its material particles ·
and this 1s confined entirely to its' primary qualities'. Symbolize
this heat by Hm. (Heat as a mode of motion.) There is nothmg
else in the body except the po\\er of tlus Hm to produce a sensa-
tion in us. Call the sensation Hs. In the case of heat felt by
us, then, the whole reality consists of Hm, the power of Hm to
produce Hs, and of Hs.
This 1s the theory of the physicists and philosophers referred
to. (By the way, the theory began with the Greek Atom1sts,
who, though philosophers by nature, had surely rather the
scientific 1 than the phtlosoph1c spirit ) They would usually
admit that we ordinarily thmk of heat as something in the
body like our sensation, and not as a mechanical condition of
its particles (which mdecd we only know by scientific investiga-
tion). And this latter tendency of ours they regard as an illusion
though they may differ in the way in which they describe 1t.
The theory you oppose to tlus I cannot think clear, and
I doubt whether you have got yourself clear as to your own
meaning. The one certain thmg by which one has to try to
interpret you is that you deny (that) the plain man's view is
an illusion This would seem to necessitate that he is right in
thinking there is some quality in the body which, whatever else
it is, 1s neither his sensation of heat (though apparently somehow
near akin to 1t) nor the vibratory movement of the particles.
It must be different from the latter or you would agree with
the theory you criticize. This then is neither Hs nor Hm ; let
1 rn the MS •scientific• and' philosophic' are inverted.
 On PrimM)' and s~t;()Mfr)I Qualities 769
us call it Hx. The distinction, however, is obscured, from your
tendency to represent the 'primary quality' theory of heat as
though it only meant a power to produce a sensation. Thus
where one would expect Hx (the secondary quality as ascribed
by the 'plain man' to the thing) to be distinguished from the
primary quality as a vibratory movement of particles, it is only
distinguished from 'their power to produce a certain sensation
in us'. One would expect you to say 'he does not merely mean
the secondary qualities as v1brat1ons of particles together with
the power of this dynamic state to produce sensations in us'.
Thus you seem committed to recognizmg Hs and Hx and Hn,
which latter 1s not Hm.
The question arises : do you really mean to recognize Hm at
all? And, 1f so, (m) what relation does it stand to Hx, since
both Hx and Hm are supposed to be in the body? This last
question is a serious one, and the fact that you haven't asked
1t seems to me to indicate that you have not fully thought out
your position. Nor on the other hand do you deny Hm.
The first criticism I should therefore pass 1s that you have
not made clear what it 1s exactly that you do hold. But,
assuming that what you don't deny you retain, the reasonable
supposition seems to be that you recognize as real and as distinct
from one another Hs, Hm, and Hx.
§ 524. Sensations as representative of secondary qualities and the
general idea of suck representation.
You begm an important section of your argument by assuming
the idea of sensations being representative.
( represent}
(They)~ express something other than themselves
Lstand for
without any cntical elucidation of the idea.
Now, I venture to thmk that the idea of such representation
in philosophy, or psychology rather, is very loose and treacherous
and, if used at all, should be preceded by a 'critique' of such
representative character, and an explanation of the exact sense
in whicll the word representative is used.
If you had undertaken such a 'critique' I think you might
have written very differently in the sequel. It seems to me
that in what follows you not only assume this idea of repre•
770 TENTATIVE INVESTIGATIONS
sentation, but assume its application to Hs and Hx, thus
assuming the existence of Hx, the very thing which you have
to establish.
Let us consider the idea of 'represent', which you make
equivalent to 'express ', and 'stand for'. I am convinced that
we are all hable to fall mto serious confusion by an uncritii;al
use of such words. For example, I am convinced that Bradley's
theory of Judgement, and with it therefore a great deal of his
Logic, is hopelessly shipwrecked in consequence of a great con-
fusion m his mmd as to the meaning of 'significance'-' meamng'
-•sign'-' symbol', which have to do with representation. (I do
not publish my criticism of Bradley ; but this criticism I have
lately been encouraged to circulate in MS. among Oxford
students of philosophy. You are welcome to see it if you like :
though I shouldn't care for you to talk much about it-except
perhaps to me-especially as among its consequences (not
developed m my circulation copy), there is involved an uncom-
promising rejection of the general account of inference which
Bosanquet follows. For I have an uncommon regard for
Bosanquet, begotten greatly by his own generous treatment of
myself.)
In the strict meanmg of 'represent' and 'stand for' . nothmg
in the world in itself ever 'represents' or 'stands for' anytlung
else. In itself it can only be said to 'represent' itself and 'stand
for' itself ; though even this 1s an abuse of terms. Representa-
tion is our subJective act. Nothing has a meaning in itself. It
is we who 'mean'. It is we who make something represent
something else, somethmg other than itself. (This pomt I have
worked out in detail m the aforesaid cnbc1sm) Thus 1t 1s we
who decide that x shall represent the problematic quantity, and
a, b, c, &c., determmate quantities. It is we who make a weeping
willow a symbol of sorrow. There may of course be something
m the object which prompts us to give it a meanmg, e. g. the
resemblance of the weeping willow to a human figure bowed
over m the attitude of grief. But the willow m itself can neither
'mean' grief, nor 'represent' nor 'stand for' nor 'express' grief.
We do all that. The fatal thmg in Bradley is that he thinks
an 'idea' can m itself have 'meaning', be 'significant', or
'symbolical•.
 On P,ima,y and Secondary Qualities 771
The thing we make representative, as opposed to mere symbol,
is, in the strict use of representative, always somehow like
the thing it represents. But there is an extended use. If the
existence of A necessitates that of B, we may, taking the
existence of A as evidence of the existence of B, speak of A as
a sign of B-i. e. a sign to us because we infer B from A; but
when we are fully conscious that this is the relation between
A and B, I thmk we stick at saymg A is 'representative' of B,
though by a confusion we may perhaps do so, when we are not
dearly aware of the true relation between A and B, or not
accurately remembering it But we should never say A
'expresses' B.
Now, in the present case, 1£ I understand you rightly, your
text seems necessarily to mean that Hs is a sensation which
'mediates our knowledge of' Hx, because Hs •represents' Hx,
or 'stands for' Hx or 'expresses' Hx. Now I would venture to
cut out at once the word •express '-a sensation cannot in any
proper sense of the word 'express' anythmg different from itself.
Let us then merely retam 'represent' and 'stand for'. How is
it we come to speak of a sensation as 'representative', for
I suppose we all tend to do 1t, not you only ? Before cnttcizmg
the application to secondary quantities, which themselves (in
the Hx form) are sub Judice, I should like to consider the case
where the existence of the quality (the primary quality) 1s
undoubted, and where we are hable to speak of 'representative'
sensations.
I think the best case to take 1s that of felt extension, for that
gives the most plausible instance, and besides I do not agree
with what you say of visual extension. Here (m felt extension)
we seem to recognize in feeling round the surfaces and edges
of a die (e. g) that we have certain feelings quite different from
the extension, and yet that it is somehow through these feelings
that we become aware of the extension. Thus we may get to
call these sensations 'representative', without any distinct
envisagemcnt of what we mean-and the word 'representative•,
instead of being an explanation, is itself a problem, and indeed
is itself due to a confusion m our ordinary thinking. Now the
very description of these sensations (tactual and the like), as
mediating our knowledge of the extension, implies that they are
772 TENTATIVE INVESTIGATIONS
different from extension and have no extension in themselves.
They are not therefore 'representative' by reason of likeness.
How, then, are they representative? As already contended, m
no kind of way can they of themselves 'represent' anything
other than themselves. If we consider them as representative
it must be either on the ground of hkeness (excluded in th_is
instance) or because of their associatjon (m their qualitative
character and their given temporal order) with extension. We
must therefore be already acquainted with extension either to
recogmze the likeness of the representative sensations to it or
to know of the association. But this 1s 1mposs1ble 1f the media-
tion of our knowledge of extension by the sensations (together
with their order) means, as 1t usually does, that we depend
entirely on the sensations for our knowledge of extension at all.
But suppose, for the sake of argument, we know already that
a certam order and quality m our tactual sensations and (say)
'sensations of resistance' 1s caused by the fact that our fingers
are moving in contact with the surfaces of an extended obJect.
If we know this in general already, we can apply it to a particular
instance, and we might say that this order of tactual and other
sensations is 'representative' of extens10n here, the phrase being
really inaccurate (and m1sleadmg) ; our true meaning, then,
being that the so-called 'representative' sensations are known
to be associated with such and such an extension : and that
they therefore can be taken as evidence of the existence of the
extension. This waives the question as to how we could have
got our knowledge of the correlation of the sensations and
extension, and 1s merely given to show how little can be made
out of representation, m even what 1s perhaps the most favour•
able case.
As to this knowledge, 1t seems clear that we could not get it
at all if our datum was the mere sensat10ns, g1ven without any
relation to extension. If we really get both together we might
as well caJJ the extension representative of the sensatJOns as the
sensations representative of the extension.
The idea of representation, then, in fine, seems to me not
only useless in philosophy but misleading as tending to obscure
the solution of a difficult problem. In any case observe that
here the 'represented' idea (extension) is as clear to us as the
 On Primary and Seconda,y Qualities 773
supposed 'representative' ideas, indeed clearer and clearly dis-
tinguished from them. Thus there is a science of the so-called
represented idea--geometry, which takes no account whatever
of the ' representative' ideas.
§ 525, Sensation (Hs) as 'representative' of a quality (Hx) in the
body.
To apply the above general considerations to the case before
us: ls Hs representative of Hx as bemg like 1t? No. for then
Hx as like a sensation in quality would itself have to be a sensa•
tion : and that, according to your own view, would be a 'flagrant
absurdity'. It would seem, then, by the above, to remain that
Hs would be something unlike Hx, but correlated with it, and
that we could only make Hs 'representative' of Hx because we
had found both associated (1. e. 1£ we allow the word 'repre-
sentative' at all). Hx, then, ought for us to have a positive
content ddiercnt-qurte different-from Hs, and clearly dlS•
tingu1shable from it : and according to your theory clearly
distinct from Hm. Clearly also there ought to be a science of
Hx. These seem to be necessary presuppositions of the only
mtelhg1ble sense in which Hs could be said to represent Hx.
§ 526. Evidence of the existence of Hx.
If Hx exists it ought, according to what seem the implications
of your representative theory, to satisfy the above conditions.
But (1) 1t seems evident that there 1s nothing which does satisfy
them. As d1stmct from our sensat10n Hs, there is nothing in
the body with any definite quality clearly d1stmguished from
Hs and from Hm, except only the power of Hm, or, 1£ you like, of
the body m general, to produce Hs : precisely the contention
of the theory you attack. If there were, there should be a science
of it, and there 1s none. In fact there ought to be a science
(empirical at least) of the relation of HJC to Hs and to Hm,
but there ts no science of its relation to either, the only science
of such relation is the science of the relation of Hm to Hs.
But (2) your own argument undes1gnedly leads straight to
the same conclusion : and here I come to the main significance
of the fact to which you appeal. 'The fact that m ascribing
secondary qualities to thmgs we normally think of the things as
if they were actually producing the sensations in an hypothetical
percipient.' This proves that if we try to find any positive
774 TENTATIVE INVESTIGATIONS
quality of heat in the body which we may suppose ourselves to
feel (just as we feel extension m 1t), we are always thrown back
on the sensation itself for positive content, and can only give
a derivative positiveness to the property the body has of causing
it, by thinking of 1t as producing this definite quality of effect.
This means that we are able to say nothing of the body's quality
so far, except that 1t is the cause of this definite kind of sensa-
tion , though we may, by a confusion, be misled mto trans-
ferring what can only be a quality of the sensation into the
body itself.
All that you say in this connexion confirms the theory you
are attacking, and 1s indeed precisely the kind of argument which
the advocates of that theory would and do employ. They would
say that 1£ any one thmks there is a positive quality perceived
m the body (allowmg the case most favourable to you, when the
hot body 1s actually touched) other than the sensation and the
mere extension of the body, he may be challenged to assign any
positive predicate to this quality in itself ( = Hx), and it will be
found he can give none--and you have given none whatever-,
and a further proof that he JS under an JllusJon and can really
give none 1s that 1£ he tnes, he can only describe its positive
character by refcrenre to the positive character of the sensation
which 1t causes, and can never do anything else ; and this either
means that there 1s no quality assignable to the thing except
the power of producing the sensation m us, discoverable or
1mphed m our perception (and he therefore agrees with them), or
that he is confusedly transferring the quality of his sensation
to the body. Now this JS Just true of your own account, i.e. you
really say that no positive character can be found in the thmg
except by this reference to the quality of the sensation. Now
1t follows, from the above analysis of such meaning as 'repre·
sentative' can have, that the quality of a representative sensa-
t10n cannot possibly enter 'm consequence of its representative
function' into the ' essential constitution of the corresponding
secondary attributes of matter'. Just because of 'its repre-
sentative function' 1t can do nothing of the kmd, and the
statement is a mere contradiction m terms. This, I think, you
will see 1f you ask yourself now what could be intelligibly meant
by 'representative function'. In fact I hold that in this sentence
 On Prima,y and Secow:la,y Qualities 775
you have exactly illustrated the unconscious transference of the
quality of the sensation into the object. I may here add that
I think this explains why some quite acute people among your
audience at Oxford were puzzled by your paper. I think they
felt they were not quite sure what you were driving at : and
some of them were inclined to suspect some confusion, if not in
your own thoughts, yet in your presentation of them.
You will see, then, that I cannot help thinking you have
proved the contradictory of what you originally intended, or at
least that your arguments are on that side and confirm the
position you meant to shake. The right formulation of the
theory seems to me what I have given above (in§ 520).
§ 527. On the relation of the supposed Hx to Hm.
As I have said, I consider it a sign that you haven't fully
cleared yourself up that you do not ever raise the important
question of the relation of Hx to Hm If Hx 1s really the
quality of the body itself which 1s correlative to our sensat10n
of heat (Hs) why 1s Hm needed ? One might expect you would
have to say that Hm is itself an illusion arising from people's
not recogmzing the existence of I Ix. If, however, m deference
to science you rctam Hm, how is 1t to be related to Hx ? Is
it the cause of Hx and Hx agam the cause of Hs ? If so, the
important thmg for science would be a theory of the relation
of Hm and Hx m the body. Thus, in any case, you would have
to propose a new science. If you reject Hm, the new science
would recognize Hx and study its causal relation to Hs. If you
do not reject Hm, you would introduce a new science to study
the relation of Hm to Hx.
But, of course, you know I should mamtam there could be no
such science, because no positive content at all can be given to
Hx. I will here add, what I have already indicated m conversa-
tion, that I am fairly sure the secret of the whole puzzle about
Bacon's 'forma' lies in the fact that he himself got into complete
confusion between Hs 1 Hx, and Hm. The real question to ask,
I am sure, is not the usual one, 'how does the "formal " cause
differ from the "efficient" cause m Bacon?' which no one has
been able to answer satisfactorily (or even to give a coherent
account of what Jonna itself means)-but we should ask what
is the relation of the forma to the simplex natura of which it
 TENTATIVE INVESTIGATIONS
is the /ormtJ. No one seems to have undertaken this inquiry.
If it is done, the result is a mass of hopeless contradictions of
statement. Some time ago I thought I detected the source of it
all in the fact that the simplex natura as conceived in the body
was precisely Hx, and the f orma of this is Hm. Bacon some-
times seems to recognize that heat in the body is only Hm, •and
besides this there is only the effect in our sensory (organism)-
a sensation-and this accounts for a certain class of passage.
But sometimes he unwittingly transfers the sensation into the
thing; thus appears the simplex natura of heat as in the thing,
which, nevertheless, derives all its positive character from the
sensation with which it is confused This accounts for another
set of passages, and the two sets are thus m irreconcilable
contradiction to one another.
As the Baconian difficulty is a recognized problem of philo-
sophy, and I believe it capable of being quite cleared up, I had
thought of givmg a public lecture on it next term. Your paper
has itself made me undertake the rethinkm~•over process which
I should have had to undertake, and so I may as well execute
my project. The lecture itself was written long ago, I should
say fourteen years ago, and I have occasionally delivered it in
my ordmary logic course. There 1s perhaps nothing in it I want
to alter, at least as far as I can see. I dare say it will interest
you to know that I have given my course on 'hypothetical
thinking' again this year, and I think I have improved it by
developing 1t and makmg 1t quite defimte m certain d1rect1ons.
§ 528. The case of the Primary Qualities.
(A) For the mam issue, which must be settled one way or
another by the preceding, it would not be necessary to go on
to what you say upon the primary qualities.
With regard to your words 'This analysis of the secondary
attributes of matter holds good in all essential respects for the
primary also', it doesn't follow, because, as I think, the analysis
is not valid for the secondary, that something like it may not be
valid for the primary ; and so I propose to consider the primary.
Moreover, I don't find I can agree with your account of sensible
extension, &c., and want to talk about it because I think your
account of the secondary has infected your view of the primary
qualities. The facts in the case of the secondary quahties seem
 · On Primary and Secondary Qualities 777
to be {taking heat} that we perceive only our own sensation
{Hs) and possibly the extension of the hot body, if we touch it :
we infer a special power in the body to cause it, and we arrive
at our idea of the state of the body which has the power (Hm)
by inference and there is nothing else in reality. Hs, then,
is perceived, Hm is not perceived but inferred by a scientific
theory. Add that there is the illusion Hx, letting Hx now stand
for our tendency to transfer confusedly the quality of our
sensation (Hs) to the body.
What is there to correspond in our knowledge of extension?
Take felt extension first. There are certain tactual and muscular
sensations : call these Es ; qmte distinct from these is the exten-
sion felt and which we say we feel in the body : call this Eb.
Eb is conceived as distinct and different in its nature from Es.
Hm JI JI JI JI JI Hs.
So far the relation of Eb to Es 1s like that of Hm to Hs. But
there is a fundamental difference.
Hm is the mere result of scientific inference, it is not the
object of perception, we neither perceive it actually, nor suppose
we perceive it. Eb we certamly suppose we percei~. We say,
not we feel the tactual feelings, but we feel the extension itself.
A corroboration of this is that we arc able to describe the
extension we say we feel, as it appears to us in perception or
in what we call ' feeling ' it, we cannot possibly describe Hm
from the experience in which we have Hs.
If it is objected-we don't really perceive Eb but infer it,
and are under an dlus1on in thinking we perceive it, it may
be answered : (I) If so, how is it we do come to think we perceive
Eb, and never come to think we perceive Hm? Why can we
from mere perception describe the one and not the other ?
(2) There would still remain the fundamental difference that we
at least suppose we perceive (feel) the one and do not suppose
we feel or perceive the other. (3) Eb, or extension in a body
in general, cannot possibly be something originally unperceived
and only inferred from what is perceived : for, as shown in the
discussion of the idea of representation, if what we perceived
were only tactual, &c., sensations, without the perception in
them of any relations of extension, extension could never be
inferred from them.
 TENTATIVE INVESTIGATIONS
Thus again, while there is this correspondence that Hm is
conceived as the cause of Hs, and Eb also as the cause of our
having certain sensations (Es) in a particular order, there is this
fundamental ddferenre that our idea of Hm is only acquired by
our searching for a cause of the given Hs, but our idea of Eb
is not acquired merely from a search for a cause for Es. .
Is there anything to correspond to Hx, not as you conceive
Hx, but as an tllus1on, as I conceive it? There probably usually
1s. We not only thmk of the hardness of the extended substance
in associat10n with our feeling when we endeavour to compress
it, but confusedly transfer the quality which belongs to our
feelmg to the thmg itself.
Therefore in the case of the sensations relative to felt extension
(not seen), I should say that these sensations and their order,
though of course correlated to the extension in the body, most
certainly do not 'enter mto the constitution of it', and do not
'represent' somethmg other than themselves, i. e the extension.
As to the former phrase, the context shows that you hardly
mean it seriously yourself, for 'only so far as', &c, practically
cancels anything naturally meant by such a strong phrase as
'enter into the constitution of it', and as to the second phrase
I needn't repeat what I have said about 'representat10n '.
§ 529. To come now to visual extension.
It 1s here I specially find the 'infection' of your view of
secondary qualities. You hold here that we perceive a visual
(coloured) extension, apparently, which is as much our subJective
state as the sensation of heat, and is merely representative of the
real extension which 1s other than itself. The words m italics
seem absolutely necessary to your v1ew and to the analogy you
institute with the secondary qualities.
I confess I cannot understand how you could have persuaded
yourself that the distinctions you make in this part of your
argument were tenable, if it had not been that you were affected
by the v1ew you had developed about the secondary qualities.
Let us take first your instance of the extension (visual) perceived
with closed eyes.
This is nothing but an mstance of the following. Suppose we
are looking at a flat surface--say the wall of a room. We see
the rectangular shape of the wall when the eye is at a. sufficient
 On P,iMa,Y affll Secondary Qualities 779
distance. This surface is a part of ' real ' space, of the space
'in which attraction varies', &c. As we move the eye nearer
the wall, its limits which constitute its shape get out of the
range of distinct VJsion. When the eye is near enough to
the wall we lose sight of them altogether ; when it is close to the
wall we see but a comparatively small part of it : the centre
of this distinctly, and parts at an increasing radius less dis-
tinctly, fading off mto nothing visible at all by imperceptible
gradations. And so, as in your case, we cannot say what shape
it has, but nevertheless what we see 1s a part of actual extension
and has a position in actual space, for it is on the wall Now
m the case of the closed eye-Jf you really mean we perceive
or see 'visual extension', as opposed to merely imagining-the
only ddlerence 1s that the surface has come into actual contact
with the eye, for it's nothmg but the eyelid itself which is seen
by transmitted hght, and so what we see 1s actual extension, it
certamly has a position in real space. As the case of the wall
shows, the fact that we can't tell what shape limits what we
see is no proof whatever that 1t 1s merely some subjective
extension that we see. I feel I cannot be too emphatic about
this, and I feel you have been seriously misled. You say again
it is an 'mtrms1c impossibility to express its magnitude m feet
and mchcs '. The magnitude of any part of it, on the contrary,
can just as much be expressed m feet and mches as the magnitude
of any part of the visible part of the wall, whose limits we can't
perceive on account of the nearness of the eye. The only way
m which it is impossible is irrelevant, i. e. it may be impossible
for me to use a measure so as to measure a given bJt of the
extension, but nevertheless it has a dimension expressible in
inches.
§ 530. (B) Leaving this subject for the moment, let us con-
sider the perception of a piece of coloured surface, as of a blot
of colour on a piece of white paper. This is with you 'visual
extension'; you also speak of it totulem verbis as' sensation' and
'extended sensation•. This 'visual extension' again with you
is not m the thmg (and if seriously meant to be extended
sensation, could mdeed not be} but has a 'representative func-
tion', i.e. represents something m the thing.
With you 'representative' means 'representing something
2773•2 B b
780 TENTATIVE INVESTIGATIONS
other than itself' ; it would be important to say what ' other
than itself' exactly means, whether 'other' as one smarting pain
is other than another such pain or different as the smart is
different from the thorn which causes it-' different in kind'. You
don't expressly say, but, as I have shown above, your theory
necessitates the latter kind of difference, for 1£ Hs, which ~e{>re-
sents Hx in the body, were like Hx, what you call the flagrant
absurdity would be committed of supposing the body had a
sensation.
In several ways it may be shown that this view of the relation
is quite untenable.
For simplicity, suppose the blot to be bounded by rectilinear
edges, and to be a parallelogram.
(I) The quality of the colour-supposed to be a sensation-
is irrelevant as, e. g. green, or red, or blue : it is the extension
only that matters. This extension I describe as a parallelogram,
and say it is bounded by straight lines. By hypothesis this is
to represent something in the body different in kind from itself.
This extension in the body 1s different in kmd from anything of
the nature of a parallelogram and cannot be bounded by straight
bnes. What is that? Clearly we could say nothing whatever of
it, and you would be committed to something worse even than
'a thing m itseH', viz. to 'an attnbute m itself'. (In our con-
versation in the quadrangle, in order to reply to a remark of
mine, you committed yourself to the view that the real extension
of the thing was for aught we knew as ddierent from the visual
as 1s Hm from Hs. I build no argument on a statement m
a hasty conversation ; but 1t confirmed me in thmkmg you
had not worked out your position clearly and realized its
implications.)
But (2) we always assume that the extension in the thing is
exactly the same in kind as the visual extension, and the science
of geometry depends entirely upon this. (This 1s one of the
reasons why your criticism of Kant is invalid.} There could be
no science of the 'extension in itself' any more than of the
thing in itself. Moreover, science assumes not that the extension
of the blot that I see is like the extension in the thing, or
represents the extension, but that it is the actual extension in
the body. It is visual extension simply as real extension seen.
 On PriMMY and Suontla,y Qualities 781
(I mean of course science as scfonce. If the scientific ffltJ7I was
asked, he would very likely give a ridiculous answer. They
usually do, for they are more incapable of reflecting on their
own presuppositions than the ordinary 'plain man'.)
(3) Let us analyse this 'visual extension' and inquire what
parallel it presents to the secondary qualities. If we attend to
the colour alone and distinguish Cs, Cx, Cm, then Cm as the
movement of particles of ether actuated somehow by the body
corresponds to Hm, for it is not perceived and is solely an
inference in scientific theory. But the relation of Cs to Cx is
very different from that of Hs to Hx. As already indicated,
whereas we can and do locate Hs in our sensory organ, even
though at the same time we may erroneously locate something
like it (Hx) in the body, colour is clearly and distinctly located
in the body perceived and we no more locate it in the per-
ceiving organ (the eye) than we locate there the extension of
the body itself. We have no choice: we must locate the colour
in the surface: we cannot imagine yellow save as on a surface.
Nor will all our scientific theory, that 1t is not there but our
sensation, enable us to 'locate' it in the eye or represent it
otherwise than as at a distance from the eye and on the surface.
The colour cannot be made many sense 'representative' of the
extension, simply because we can't perceive it except as
'extended', i.e. on an extended surface, nor even imagine it
without extension. Thus tf we 'perceive• colour, we must at the
same time 'perceive' extension.
§ 531. (4) But the important thing now is the extension element
in this; because it 1s as extended sensation, or visual extension,
that you make it 'YepYesentative • of something m the body-
apparent and not real, and, as it would seem, something merely
subjective.
As remarked in the discussion of the foregoing point (the
comparison of secondary qualities with extension), the percep-
tion of colour 1s impossible without the perception of extension,
not at all as an inference but as a necessary part or aspect of
the same perception, and such extension again is absolutely
unintelligible save as the extension of a SU1'fac~i. e. a real
surface. lf two colours are juxtaposed which do not shade off
imperceptibly into one another, i. e. two colours homogeneous
Bb2
 TENTATIVE INVESTIGATIONS
in themselves--as the coloured blot on white paper, we neces•
sarily perceive their common boundary as linear, that is really
the common boundary of the surfaces to which we refer the
colour. (N.B. If two sensations different from one another, other
than colour (granting 1t to be a sensation), succeed, no such idea
of a boundary arises, e. g. a taste and a smell, one toothacpe
and another pain, one sound and another.)
We 'perceive' the bm1t or boundary just as directly as we do
the colours.
The aim of this detailed analysis (in which for clearness I have
put the same thmg m different ways at the nsk of wearymg by
repetition) (1s to show) that in no sense does the perception of
what you call visual extension (as a kind of what you call
sensible extension) involve any extcns10n m sensations as sensa-
tions. When we analyse it, we find that though the extension
(is) inseparably associated with the colours, it is not assigned
to sensations but to the surfaces, a.nd to these alone, to which
the colour (1f this is to be sensation) 1s referred. It is the
coloured surfaces which are extended, the coloured sur-
faces which have a boundary. It is either a confusion or
a mere fafon de parler to speak of the colour sensation as
extended, and as a / afon de parler 1s only excusable as meaning
that the surface to which the colour is referred is extended.
Therefore, strictly speaking, in a philosophic discussion such
language should be avoided and the accurate expression sub-
stituted Accurately, then, the sensat10ns, whether tactual or of
colour, are not extended, and the idea of extension is absolutely
inapplicable to them
Now the view that visual extension (and sensible extension
in general) is merely subjective, is nothing in the thing, is an
appearance as opposed to reality; an appearance which some-
how represents a reality, which as unreal and merely subjective
has no position in real space, depends wholly upon the view
that sensations can be in themselves extended, and 1s therefore,
I am obliged to think, entirely untenable. This seems in itself
enough to me, but I will consider some of your principal state-
ments in support of your view m detail.
§ 532. (5) I should say it was impossible to develop such
a theory into details without contradiction or some verbal con•
 On Primary and Secondary Qualities 783
fusion. I seem to :find contradiction in what you say of the
apparent size as given by the visual extension in contrast with
the real size.
The visual extension as consisting, according to you, of repre•
sentative sensations must, I have contended, according to you,
be different m kmd from that which it represents. You never say
this quite clearly, though it is a point on which you ought to
have pronounced defimtely, and I can't help thinking that it
was a sort of instinct which prevented you. You use, for
example, the very strong expression 'extension as a charac-
teristic of visual sensation is quite distinct from the extension
of things in space', of which the natural interpretation is that
the two things differ m kmd. But your remarks on real and
apparent size entirely contradict this The contrast of the
apparent size of a thing with its real size 'as measured m feet
and inches' means simply, e g., that it looks z inches long
whereas 1t 1s 3. This means again that the apparent length
is a line 2 inches long and the real length a Jine J inches long.
(With this kind of difference between a real and apparent
distance one is very fam1Jiar m the exercise of' judgmg distance'.)
This is only the comparison of one line with another, of two
things (two extensions) exactly the same m kmd.
Verbal confusion I seem often to find in what you say of this
curious 'sensible extension' as opposed to real extension. First
you say sensations arc extended, then you seek to take away
the paradox, by addmg an explanation of 'extension', in the
case of sensation, which takes all that 'extension' means out of
it ; (thus you say} 'they are not extended m the same sense as
corporeal things', ' bodies are extended in space', &c. But what
other sense could there be if it was extension at all? This to
my mind is as bad as saying 'our sensations are numbered and
counted, but then it is not the same kind of numbering or
counting as applies to corporeal things'. ' Bodies are extended
in space', and it is therefore implied that sensations are extended
but not in space. This to my mind is sheerest contradiction.
What is extended must be in space or it is not extended. For
instance, if it is extended as a surface (and it is really this case
we have before us), that necessarily presupposes space on both
sides of it. It will not do to meet the difficulty by allowing
 784 TENTATIVE INVESTIGATIONS
the 'extended sensation' to be in a space of its own, quite
different from and not included in 'objective space', or the
'single homogeneous infi01te space', as you describe it, for this
would involve a terrible mess which you will realize of course
well enough.
I seem to find a good deal of verbal confusion. This 'sensible
extension', notwithstanding the difference you make between it
and the ' real ' extension, you speak of as a ,nanifestation or
expression of spatial extension. Here I note the expression
'spatial extension' which would be paralleled by 'numerical
number' and implies a non-~patial extension ( = a non-extended
extension) just as the other would imply a non-numerical
number. But anyhow, what could these somewhat loose and
popular phrases 'manifestation' and 'expression' mean philo-
sophically? How could the sensible extension be the •expres•
sion' of the spatial extension which is quite different from it?
It seems connected with the fallacious use of 'representation'-
' expression' here is only to be tolerated as a metaphor, as such
it is of no use whatever as an explanation ; one has to explain
what the metaphor means. 'Mamfestabon' is used legitimately
enough m philosophy for the particular as a 'manifestation' of
the universal, but then you do not mean to relate 'spatial
extension' to 'sensible extension' as umversal to particular.
Possibly, as the context with the word 'appearance' in it sug-
gests, 'manifestation' means the manifestation of spatial exten-
sion to us in perception. But then obviously 1t must be spatial
extension 1ts very self which is before us-spatial extension
must as manifested be the object of our perception : and not
an extension 'quite different' from it. If the latter presented
itself as a 'mamfestation' of spatial extension it would be a mere
fraud : it would be only a 'manifestation' of the 'sensible
extension' and not of 'spatial', for clearly what was really mani-
fested would be simply the 'sensible extension' before us.
If you want to introduce Jones to me, 1t won't do to introduce
to me somebody 'quite different' from him ; that wouldn't be
'manifesting' ']ones to me, but that other person.
§ 533. (6) Apparent antl real sue.
The supposed difference of apparent size is what seems to
have affected you greatly in your view of the difference between
 On Pritnl.W)' •ntl Sscontlary Q#tllities 185
sensible and real or spatial extension. Here again I think you
have assumed what is little better than a popular distinction
without criticism. Just as 'representative and represented' so
'real and apparent' seem to me regular trap worde, which cannot
be safely used in any particular case without a most patient
examination of their meaning and justification in the particular
instance. The distinction of real and apparent size seems to
me very slippery and to require more than usual care-all the
more so because at first it may seem familiar and not to require
any discussion at all. There are two ways of treating the theory
of 'sensible visual extension' versus 'spatial extension'. The
first and the easier for mere negative purposes is to show the
untenable consequences of adopting it. The second and harder,
but the one required to complete one's estimate of the true
facts, is to analyse what is really meant by and involved in the
conception of apparent versus real size.
(a) Assume the theory to be true. Suppose I see a chair
before me upon the lawn. The strict meaning of the theory
absolutely necessitates the view that the 'visual extension' or
the extens10n which J sec is my own 'extended sensation' (your
phrase) ; that 1t 1s not the real extension of anything, for such
extension is in space, while sensible extension (e~ hypothesi)
cannot be. The same holds of the lawn itself and of all tho
obJects (extended) which I see about the chair. These 'appear'
to me to be 'in space', tke,e seems to be empty space between
them-indeed this is absolutely necessary to and involved in my
thinking that the chair has any extension at all. (This, I trunk,
you have hardly realized.) Now the space, which indubitably
thus 'appears' to me and in which the apparent objects have
their extension, is non-real; for, if it were real, the objects
I appear to see would have their extension in real space, whereaa
by hypothesis they have not. (We thu11 come in a different
way to a previous thought.) Is it necessary to develop the
consequences of this further? May I not assume that it lea.di
to an uncompromising subjective solips1stic Idealism? (N,B.
that this space (apparent) couldn't possibly be called 'extended
sensation', and so has no place in your non-real extension.)
{b) No doubt we are accustomed to say that as an object- is
removed farther from the eye it 'looks smalls,', We believe it
 TENTATIVE INVESTIGATIONS
has not become smaller, and so people come to speak of the
variation as 'mere appearance' and so to call the size itself
which seems to vary 'apparent size' as against 'real size'.
Again the apparent variation of size is said to correspond to
a real variation of the distance from the eye, and is often said
to be 'representative' of this real variation. • •
All these phrases are most misleading and tend to pernicious
consequences m ordinary psychology In your own mvestiga-
tion I think you omitted to ask yourself at the start what could
be meant by real size. You conceal the difficulty from yourself
by describing the real size as the size measured by feet and inches.
'Thus we contrast the apparent size of a thmg ... with its real
size as measured in feet and inches.' Tlus of course suggests
two important questions :
(a) How is the apparent size measured if not by feet and
inches?-a question which doesn't seem to have occurred to you.
It must be measurable or it wouldn't be a 'size' at all, and we
couldn't speak of its becoming smaller or larger. Moreover
(part of the same question), smce the apparent size is a size
and the real size a size they must stand m a relation of measure.
So any apparent size must stand in a definite relation of magni-
tude to the real size Moreover, we can mcrcase and diminish
the apparent size ad indefinitum by removing the object far
enough or bringing it near enough to the eye. For this purpose
the magnifying glass is merely an instrument which enables us
to see the object, when brought much nearer the eye than is
consistent with clear vision without it Consequently some one
of the apparent sizes must coincide or be equal to the real size
(and be measurea in feet ana inches). There must therefore be
some distance from my eye at which the apparent size of the
object is its real s1ze. What is this distance? and how can it
be determined? Clearly this result is absurd, for there could be
no reason why one distance should have the prerogative over
another. I expect you will readily agree in the absurdity of
this necessary consequence. But it shows the untenablencss of
the ordinary distinction of apparent and real size. You come
partly into this train of thought, but you have not followed it
to its consequences, nor realized that it should have made you
revise all that you had stated previously. E. g. you say •still
 On Primary and Secondary Qualities ']87
less can we select this or that apparent ex.tension and identify
it with the real. For each of them has in principle just as much
and just as little logical title to be so regarded as any of the
others.' This makes me say you will agree with me about the
foregoing absurdity; but you have not seen that the ordinary
distinction of real and apparent size, which you yourself use,
makes it necessary that there must be one of the apparent sizes
or extensions which is equal to the real.
(b) Virtually defining real size as that measured by feet and
inches, you have, as I said, disguised the difficulty. For now
the second question arises, 'What are feet and mches? ' To say
the real size of our obJect is two mches is to say its extension
is twice a given extension which 1s called an inch. But what is
the real size of this extension? What 1s the real size of the
extension which you call an inch? The question shows that we
are no 'forrarder' in the definition of real size by this reference
to feet and mchcs
The d1stmction of real size and apparent size, as above
described, I must then hold to result m hopeless contradiction.
Your theory depends very largely upon the d1stinctlon, so con-
ceived, and so for this reason agam I must think your theory
untenable.
§ 534. We must therefore undertake anew the investigation
of the facts upon which the misleading language employed above
[§ 533 (6) (b)] 1s based
We may introduce this by asking the question, 'How should
we go about to show that the real extension of a given lme was
two inches, in reply to the inquiry as to what it was, or as
agamst a conJecture that 1t was, say, an mch and a half?'
Suppose we are given a six-mch rule. We should define an
mch as the sixth part of the extension of this rule, which we
may again simply define as the rule we are holding in our hand.
We want to compare the real extension of this rule with the
real extension of the given line, e. g. by applymg the one close
to the other.
But how can this be done, since, by hypothesis, in experience
only the sensible (take here the visual) and merely apparent
extension of the given line and the sensible or apparent extension
of the ruler are given ?
788 TRNTATlVE INVESTIGATIONS
The answer would doubtless be given readily enough. 'We
simply apply the visual extension of the ruler to that of the
given line or edge, and 1f the line's visual extension coincides
with two-sixths of tbe visual extension of the ruler the real
length of the line is two-sixths of the real extension of the ruler,
i. e. two inches. If we ask 'Why?' the answer wdl be 'because
then the real extensions are in the same ratio to one another f.&S tJu
apparent extensions'. These answers, probably given with some
confidence, mvolve a pretty cod; even 1£ we omit to criticize
the application of one visual extension, as equal to one extended
sensation, to another extended sensation.
(1) This implies that the apparent extensions stand in a ratio
to one another, and so are measurable. Nor can the implication
be avoided that the apparent stand m a ratio to the real. And
from this of course 1t follows, as before, that some apparent
extension must coincide with the real extension, with the con•
sequent absurdity
(2) How can we possibly know that the real extensions
{'spatial extensions' you call them) are in the same propor•
tion as the apparent? For the former arc never objects of
our experience-by hypothesis-but only the latter, i. e. all
extension we see is visual extension This 1s an unanswerable
difficulty.
(3) The idea of the application of the visible extension of the
ruler to that of the given hne has also serious implications.
In mere visible extension we can make the visual extension
of the given hne appear to comcide with any fraction of the
ruler. It will be readily replied, 'But that is because the ruler
and object are not at the same distance from the eye, and though
we may see no space between them, they are not in contact ;
the true measure is got when they are at the same distance
from the eye.' But this refers to actual real (spatial) distance
from the eye. That we can never get at, our perception being
only of the 'sensible extension', and in such a case we can only
compare what we can perceive. Consequently we could not
know when the real distance from the eye was the same for the
ruler as for the given object.
(There is a point about the use of 'representative' which
I ought to have mentioned explicitly. I have continually been
 On Pnmary and Secouary Qualities 189
presupposing it. If two things different from one another are
correlated, why, in this use of 'representative', do we make the
one (A) •representative' of the other (B) and not the latter (B)
representative of the former (A)? The one (B) is said to be
•represented ' because it is supposed not to be itself present to
our consciousness ; the representative (A) is that which is sup-
posed present to consciousness. This is vital to the distinction
of 'representative' and 'represented', and perhaps tends to be
forgotten at important points.)
Or, to put it otherwise : the apparent siies which correspond
to the real sizes, when the latter are at the same distance from
the eye, are to be compared, a.s these are in the ratio of the
real sizes. But how are we to know which apparent sizes corre-
spond to the same distance of the objects? For it is only the
apparent sizes which are present to consciousness, neither the
real sizes nor the real distances from the eye are, by hypothesis,
present to consciousness, nor indeed could they be. The question
is quite unanswerable by the theory before us. For instance,
we couldn't say they (the apparent sizes) had the same apparent
distance from the eye, for that would not only put them in
a space in general of their own (of which we have already seen
the difficulty), but smce 1t would be a distance from the eye,
this subJective space would, at the eye, have a common point
with real space-an add1tlonal absurdity-but perhaps 1t can
be dodged by the doubtful substitute of an apparent eye for
the real I Or suppose, to avoid such a difficulty about the
position of the eye, we say the visual or sensible extensions of
the ruler and the obJect, which (1. e. object and ruler) are at the
same real distance from the eye, are observed to be conterminous
(at least the tactual extensions, as there are more obvious
difficulties about supposmg we observe conterminousness in
visual extensions by eye only). This conterminousness inevitably
means prox1m1ty m a space including both the conterminous
extensions. (Besides, though this is not necessary for our
immediate purpose, there could be no test for such conter-
minousness even in the case of touch-an interval of time
between the apprehension of the sensations constituting one
tactual extension corresponding to one surface, and the appre•
henaion of those correspondmg to another, does not necessarily
79o TENTATIVE INVESTIGATIONS
correspond to their discontinuity or non-conterminousness. This
can be proved, but we needn't stop for that now perhaps.) If
we try to do away with such space by reducing the sensible
extension, whether visual or tactual, to a sort of analogue to
a surface with no space on either side, then sensible extensions
representing surfaces of objects at very different distances from
the eye would, m such 'extension', be conterminous with one
another.
§ 535. This may serve to introduce the positive analysis of
the facts : for the explanations which are inevitably given to
account for the variation of 'apparent' magnitude, and which
cannot themselves be accounted for by the theory now criticized
or reconciled with it, are the result of a very different view of
our perception of extension and show that this view is the
ground of all our thought and perception of the spatial.
I shall try to show that it is precisely because we reason as
1f we saw the real (spatial) extension of the real obJect that we
are able to explam what are (as I hope to show) misleadingly
called 'apparent variations in its seen size', as due to a variation
of distance from the eye. That 1s1 our mental attitude pre-
supposes that we see the real extension m real space and not
some simulacrum of 1t wluch 1s not many space at all.
The explanation which would always be given of variations
of apparent magnitude while the real magnitude remains the
same is of the following kmd .

--
A...-=------------C:_

& -

B 0

Let A B D C be the tops of four vertical posts rismg to the

same height above the same horizontal plane. Let E be the
position of the eye of the observer and let this be m the same
horizontal plane as A B D C, outside the rectangular quadri-
 On Primary a.ntl Secondary Qualities 791
lateral A B D C, as shown in the figure. Let the straight line
AE cut CD in a and the straight line BE cut CD in b. Let
AB = CD and be parallel to it. Then the observer will see
the points A and B between the points C and D, which more
accurately expressed means that he will see the point A imme•
diately behind, or in alignment with a point a between C and
D upon CD, and B behind, or in alignment with, a point b upon
CD between C and D. The two points a and b are 'between'
C and D. Similarly every point between A and B will be seen
behind or in alignment with a pomt in ab. The observer then
will see A, B, C, and D in this kmd of relative position C a b D,
measured along CD.
This is then described (erroneously) by saying that AB 'looks
smaller' than CD · and the above figure is supposed to explain
why AB being equal to CD 'looks smaller' than CD to the
observer at E.
We are only concerned to show what tlus theory-umversally
accepted-presupposes
It presupposes a fact about the manner of our visual per•
cepbon as given and known-1t does not attempt to explam
this fact.
This fact is that the observer sees a point in the real extension
of a real ob1ect in the direction (real) of the line drawn from
his eye in real space to the given point in the real object.
It would doubtless be said-as another way of describing this
fact-that the pomt seen 'appears' to the observer m the
direction of the lme from it to his eye. The expression may
seem harmless, but 1t 1s likely to lead to a fallacy, for 'appear-
ance' tends to be opposed to reahty. So it may seem as 1£ the
point seen merely appeared in the direction from it to the eye ;
but 1t really is in this direction.
Now the important thmg in this presupposition is that the
observer 1s supposed to be lookmg at the real extension of
the object itself, and also that what he sees m so looking is
definitely a part of the real extension. He is looking at the
point A as the end of AB in the figure, and what he sees is
the end of this line AB. It is most important here to keep out
,of one's mmd all physiological and psychological theories to
account for the way we come to do this ; our attention must
•
 TENTATIVE INVESTIGATIONS
be fixed solely on the conception of the fact as implied and
presupposed in this theory with its figure, i.e. on the conception
of what we do apart from any theory of how we come to do it.
Since the lines EC and EA are apart, or in different directions, and
the points C and A are in these different directions and are seen
in the direction in which they are, C and A wm be seen apart, .
i. e. A will not be behind C. Similarly B wdl not be behind D.
This is a result of geometry and of the geometry of the space
as the real space in which A, B, C, D, and E are. The geometry
further shows that A and B will be seen 'behind' points ia and
b between C and D, and that m general the whole hne AB will
be seen as behind a line ab less than CD But this relation of
AB to ab is no illusion, no mere 'appearance' : 1t is an objective
fact. When we say we see A behmd a, this means that we see
A in the same direction as a from E.
This means that we see A upon the same straight line from
E as that upon which a is. But A is actually on the same
straight (hne). So, what 1s really true, we see a true relation
between A and a, and so in general a true relation between
B and b ; and not a mere 1llus1on or a mere appearance. More-
over, and th1s is of the last importance, it is just because we
are seeing the real extension of AB and seemg the real A and
the real B, that we have this particular visual perception (which
we call perspective). Or, to put 1t otherwise, because we are
actually seemg the real extension of AB and not some apparent
extension different from 1t, [that] we see AB and ab in this rela-
tion, for it is the real AB and no other length that stands in
this relabon to ab and E.
In short, instead of perspective being an illusion or at best
a mere subjective appearance, the whole geometrical theory of
rt implies that we see the actual extension of things m actual
space and the 'perspective' itself is nothmg but the recognition
of true objective spatial relations, determined by the real rela-
tions, of the real extensions seen, to the position of the eye in
the same real space. Now to see a real extension or size is not
the same as being able to measure it and say what size it is.
Measuring size means assigning the proportion of the extension
to be measured to a given extension In this accurate sense we
do not in seemg a real extension see also its size, simply because
 On Primary and Secondary Qualities 793
we do not see its relation to the given measuring extension.
We only do it exceptionally if the length of the measuring
extension and of that to be measured are in actual con•
tact, and then only in one of two cases, either (1) when the
two extensions are exactly equal and their extremities are seen
to coincide, or (2) when the measuring extension is graduated
like a ruler and the extremities of the extension exactly coincide
with (two of the) graduations marked. The general case, then,
m perception, whether visual or tactual, is that we certainly
do not see sise or perceive sise.
§ 536. This consideration enables us to see the misleading
tendency of the customary expression that the 'apparent size'
diminishes as the ob1ect recedes. What really diminishes
is the part ab of CD 'behind' which AB 1s seen, and the
diminution 1s no mere appearance, 1t 1s really an objective
fact-it appears but 1t 1s also real-a reality here, as reality,
appears.
To say the apparent size dimm1shes is a mere fafon de parler
and, as often with such fafans, a very treacherous one. Now
that we realize what size means, it 1s obvious that in 'seemg'
as such we do not see size at all, and so a fortiori we cannot
be said to see an apparent size, and so cannot be said to be
seeing an apparent size getting smaller.
We do mdeed m practice speak of the apparent size of an
obJect, but we do not mean the above by 1t. We see an object
m the distance and we try to guess its size-an operation which
presupposes that we do not see size. We guess that a distant
ob1ect (the side of a building) 1s ten yards long. This is our
estimate of its real size, but we call 1t ' apparent', meaning
simply and solely that 1t 1s our guess, a guess depending often on
a guessed distance. So we record our guess, 1£ necessary, as the
real sise guessed. If now we see a horse walk up close to the
buildmg we know the real size of the average horse and take
its length as a measuring unit, and then can see that the side
of the house 1s more like twenty yards than ten. We ma.y
express this by saying that the apparent size was 10-i. e. we
guessed the real size to be IO-whereas the real size was nearer
20 : the latter is also merely the result of a guess, but a guess
with better data.
794 TENTATIVE INVESTIGATIONS
Now if we move, say, roo yards farther off, the horse
remaining still, we don't say the apparent size diminishes-in
the discussion of size-on the contrary the apparent size (our
guess) depends on a comparison of the extensions of the house
and the horse, the result of which may seem to us exactly what
it was before .
It is true that the further objects are jl.Way, the more difficult
docs 1t become to Judge their real size. In the case, e. g., of
the sun, it is impossible to form any approximately plausible
conjecture. But it is not always true that our estimate of the
size of the object is less for a greater distance than a nearer,
m practice it is sometimes the opposite. And m any case this
is nol the meaning of the expression that the apparent size of
the object dim1mshes as its distance increases.
We have now attempted to explain the true meaning which
corresponds to this mi'!leadmg language, and tried to show
how sorely such language needs analysis and how impor•
tant it is that it should be undertaken In fine, then,
the theory appealed lo, and umversally admitted, to explain
what is called the dimmution of apparent size with increase
of distance implies that we see actual extension, and not
any subjective simulacrum of 1t , the dmunution or varia•
bon to which 1t refers and which it explams is not a merely
'apparent' vanat10n, but a true objective variation which
•appears' to us, and 1s not on that account a mere appear-
ance.
Further, we do not usually sec size at all, and certainly not
in the experience called vanat1on of the apparent size with the
distance.
§ 537. Refraction and reflection.
The theory and its figure imply normal conditions of vision.
We have to consider the visual •image' m refraction and reflec-
tion. For there we have to modify the presupposition to the
form (which includes the previous) that the object, or point
in it, 1s seen m the direction m which the ray of light from it
enters the eye. In the case of reflection 1t would be said
ordinarily that the position in which the obJect appears is not
its real position, it 1s an illusion and therefore a mere appearance.
Further, the reflection 1s called an •image' of the object, and
 On Primary and Secondary Qualities 795
thus again distinguished from it as mere appearance. But now
notice again the absolutely necessary presupposition of the
explanatory theory itself. The very distinction of the 'image'
from the object implies that what we see 'directly' is the real
object and not an 'appearance'. But if this is so, clearly what
we see as image is also the real object-it's the same object 'seen
round a corner', so to say. If it is assumed that the real object
is seen by means of the rays of hght proceeding from it (no
matter how they mediate the result), it follows strictly, just as
much in the case of the reflected image, that what we are looking
at is the real object and what we see is the real object. And of
course the geometrical optics of the explanatory theory expressly
assumes this, though still speaking of the 'image'.
As to the apparent pos1t10n of the reflected object, note again
that the explanatory theory implies not that 1t is some mere
appearance that we look at and see, but that we locate what
we see, i e. the real object, in a wrong pos1t1on in real space.
That 1s the important pomt, that ' the appearance• is our error
in locating the object not in any subJective space or any' apparent'
space, but m a wrong positwn m real space : 1mplymg therefore
that the space we suppose ourselves to be perceiving, or to be
aware of, 1s real space.
In short, according to the explanatory optical theory, what
we see is not a mere appearance, nor is the space in which we
see it an apparent or subjective space we see the real parts
of the real object and locate them in the real space though m
a wrong part of 1t The theory would entirely fail if these
assumptions were not made.
Finally, the theory enables us to substitute for the wrong
judgement of position of the object the recogmtion even in this
of an ob1ectwe fact, viz. the real position in real space of the
rays of light from the real object which enter our eye.
It might be said that, owmg to the hypothetical form of the
argument at the beginning of this section, the alternatives really
are that either both reflection and object are real objects seen
or that they are both images. But the latter alternative is
excluded by the geometric optics of the explanatory theory,
which (as shown) demands that the directly seen object should
be real.
cc
 TENTATIVE INVESTIGATIONS
All this applies, mutatis mutantlis, to refraction. If we see an
object through an uneven pane of glass, the theory implies that
what we see are parts of the real object, just as much as if the
pane were not there, &c.
There is much more to be considered in relation to these last
subjects. I have confined myself to what seemed relevant as
against the view that the extension we perceive in seeing is
something merely subjective and an extension merely in and of
our sensations in no space at all. My argument is that the very
theory by which the statement that •apparent size' varies with
the distance is justified, and from which it is therefore inferred
that 'apparent size' so varying proves what we see to be an
extension not of any object at all but of our sensations, actually
implies as a necessary presupposition that we see the real
extension in the real space.
§ 538 Note upon a certain confusion to which we are liable in
regard to the conception of appearance.
If we perceive some property of an object, there is presupposed
on the one hand the property of the object as existing on its
own account whether we perceive 1t or not ; and as distinct
from this, our act of perceiving or recogmzing the nature of
this property.
This latter, the subjective act of ours, is sometimes spoken
of from the side of the obJect as the appearance of the object
1
to us This 'appearance then gets distinguished from the
object, and that in itself is justified in so far as our subjective
act of recognit10n of the object's nature is not the same as that
nature. But next the appearance, though properly the appearing
of the obJect, gets to be looked on as itself an object and the
immediate object of our consciousness, and being already, as we
have seen, distinguished from the object and related to our
subjectivity, becomes, so to say, a merely subjective •object•-
' appearance' in that sense. And so, as appearance of the object,
it has now to be represented not as the object but as some pheno-
menon caused in oqr consciousness by the object. Thus for the
true appearance ( ... appearing) to us of the object is substituted,
through the 'objectification' of the appearing as appearance, the
appearing to us of an appearance, the appearing of a phenomenon
caused in us by the object. (The thing to emphasize on the
 0,,, Prima,y anrl Seconda,y Qualities 79'1
contrary is that the so-called appearance is the appearing of the
object, that is, we have the nature of the object before us
and not only some affection of our consciousness produced
by it.)
It must be observed that the result of this is that there could
be no direct perception or consciousness of Reality under any
circumstances or any condition of knowing or perceiving : for
the whole view is developed entirely from the fact that the
object is distinct from our act of knowing it or recognizing it,
which distinction must exist in any kind of knowing it or per-
ceiving it. From this error would necessardy result a mere
subJective idealism. Reality would become an absolutely
unknowable 'Thing in Itself', and finally disappear altogether
(as with Berkeley) as an hypothesis which we couldn't possibly
justify.
I venture to thmk that this insidious and scarcely 'con-
scious' dialectic has done much mischief in modern meta-
physics and theories of perception, and that people, who
would disclaim subjective idealism entirely, unconsciously
fall into it.
§ 539. Modern accounts of the theory of vision.
It is extremely important to distinguish the nature of our
acts of consciousness from a theory of the manner in which we
come to perform these acts. But the two seem to me to be
often confused. The theory which psychologists give of the
origin of our acts of perception unconsciously affects and corrupts
the account of these latter. The facts assumed are partly at
least deductions from the theory, and do not represent the
actual facts of our perceiving consciousness.
Thus, since Berkeley's time, there seems to be current an
account of visual perception which is absolutely false to experi-
ence. The psychologist's account of perception commits him to
a view as to what the data are which are produced in the con-
sciousness of the perceiving subject, from which the subject is
somehow to advance to the perception of the object. These
data are really presented as something which the subject has
to interpret. But the account given is entirely false, for on
examination it will be found that the subject, in order to effect
the interpretation, must be already equipped with all the ideas
CC2
 TENTATIVE INVESTIGATIONS
about the nature of reality which he is supposed to acquire
through perception and from the data.
It is among other things generally assumed that we do not
•see distance' and that somehow the original visual data are all
in space of two d1mens1ons.
This is an absolute falsification of the facts. The object
cannot appear to us as in 'space of two dimensions' strictly
understood, because there is no such thmg as 'space' of merely
two dimensions. It is absolutely mconceivable to us and there-
fore to us objects could not be presented in such space. The
only intelligible sense of 'space of two dime11S1ons' would be
a space of two dimensions, as a surface, and, strictly speaking,
a plane surface. If we had any such perception, we could not
represent the surface as a surface at all save as having a position
m space, and havmg space on both sides of 1t. (We cannot
think of any extension at all, except as m space of three dimen-
sions.) And clearly no position could be assigned to such
a surface, to which our perceptions are supposed (to be) confined.
Further, we most certamly do not refer our 'visual data' to any
such surface.
The only chance for such reference would be if there could
be present to our perceiving consciousness space which consisted
only of a surface. But this, as we have seen, is absolutely
impossible and a contradiction to the idea of space itself. Agam,
as the apprehension of the supposed surface necessarily involves
that of 3-d1mensional space m which it is, the presupposition of
the theory, put in the form that we do not see distance, is
falsified. In fact such an account of visual perception-what-
ever form it takes-and 1t perhaps is generally put confusedly
without any full consciousness of what it involves m the way
of mere surface space, &c.-is a mere untrue figment of the
psychologist to which he is compelled by his theory, instead of
bemg a representation of the facts which his theory is to explain.
His theory therefore itself is untenable, and there will be no
progress till the nature of the facts of consciousness to be
explained is accurately recognized.
§ 540. You would perhaps reply to the argument in §§ 528--g,
that the fJisible extension of the wall was what was here con-
cerned, and the visible extension was not in space but in our
 On P1ima-,y and Secontla,y Qualities 799
sensation. I have delayed therefore coming to this point till
I should have discussed this view of 'visible extension'. But
the purpose of the argument in §§ 528--9 is to show (that) the
issue is somewhat confused by appealing to the visual extension
before the closed eye, inasmuch as, if you had taken the example
of the wall, every one would say that in what could be seen of
it there was measureableness in feet and inches and that it was
in real space. But we also know that under normal conditions
the actual limit of the VlSlblc part of the wall surface, when
from the nearness of the eye the whole is not visible, is a circle,
though we may not be able to determine its radius, because of
the insensible gradation of the vividness of the perception of
colour down to zero. Just so in a surface whose limits we do
see there may be a colour mscnsibly getting fainter down to
the white of the rest of the surface. There ism reality a boundary
of zero points for the colour, w1thm which, however near, there
1s colour increasmg msensibly and continuously from zero. We
cannot distinguish the boundary, but we do not think that the
extended coloured surface we arc lookmg at (contained in the
surface whose definite limits we sec) is, on this account, not
a real cxtens10n but a mere appearance to us. The subjective
element is merely our mab1hty to say where in the general
surface the colour ends.
Again you don't seem to me to have asked yourself what the
exact presuppositions of your line of argument here are. You
seem to argue that a real extended surface or a real extension
must have hm1ts, and (that) therefore a perceived extension
which appears to have no hm1ts 1s not real extension.
But how can 1t be known that real extended surface must
have limits except from the notion of extension itself, for of
course it is erroneous to abstract extension from reahty and
then say it has particular mod1ficat10ns when 1t 1s the extension
of something real? But 1£ 1t 's true of the notion of extension
in itself, then it must be true of 'sensible extension', and, if not,
the latter is no extension at all. And as already stated the
'sensible extension', which you think has no shape or limits, has
shape and limits, though we are unable to give an accurate
estimate of them. Compare what I have said about the visual
estimate of size.
800 TENTATIVE INVESTIGATIONS
PS.-1 don't want to go into your criticism of Kant, which
I cannot think very happy. People like ourselves who try to
think out everything for ourselves, not fearing any authority,
are not likely to acquiesce either in Kant's account of reality,
nor in his theory of knowledge ; though we learn so much by
the way from him. But this criticism of yours depends on your
making his 'sensibility' equivalent to your 'sensible extension',
which I do not think will do. And there are other things to
be said.
 XIII
LETTERS TO MR. H. A. PRICHARD, FELLOW 014'
TRINITY COLLEGE, OXFORD
53 Cambridge Terrace,
Hyde Park, W.
§ 541, 6 Jan. 1904.
Mv DEAR PRICHARD,
Your letter is very welcome, for though I am not at all in
a fit cond1t1on to answer it, 1t accords with a strong feeling
I have had for some little time that I ought to get mto shape
something on the pure ' theory of knowledge' m d1stmction from
my Logic lecture. There are places in the lecture itself which
are expressed m a particular way, rather than in another, because
of certain views I have on the nature of knowledge and reality
which are not developed m the lecture itself. And I have of
late been feclmg that I had got my own views and doubts and
difficulties mto a much clearer shape. I have prepared the way
by a determined attempt to analyse hypothetical thinking and
especially the exact meaning of the thought we express by the
attachment of an apodos1s to an hypothetical protas1s In the
lectures I gave last summer I arrived at a definite opinion, at
least, on the problems m it which most perplexed me. . . . It
is not a complete solution (or attempt at one) of the proper
presentment of the whole theory of knowledge, but is an indis-
pensable preliminary, in my opinion : and I do not think that
any one can be supposed to have anything really clear to say
about it, or even to know his own mind, unless he has got
himself clear on the precise character of hypothetical thinking,
where our activity seems most to distinguish itself from reality
and where, even, we assert a necessary nexus, while in a sense
there may be no reality to correspond.•.. We are both there-
fore rather wrecks. But tho' so unfitted, I will say something
on one question, and that may be some help for the moment,
802 TENTATIVE INVESTIGATIONS
though I hope when more free to address myself to it m the
precise form m which you put it.
I think it is probably an instance of what I suppose to be
a certain illus10n we are bable (to) in all investigation of knowing
and willmg. Whereas we have to do with the relation of subject
and object, we try to express and explain various aspects of
this relation in our ordmary categories which are all of the
relation of object and object, easily seen m crude matenahsm
where thoughts arc treated bke sobd obJccts inside the mind
like another such object-a box, but not so easily detected in
more modern forms. The only remedy is to look into the nature
of the thing before us where we are certain of it, and see if it
really admits of such categories If we do that, we shall find
these functions or activities of the tlunking subject often cannot
admit of such categories. If we think of knowmg as an activity,
as doing somethmg, then, as 1f we had to do with relations of
objects, we require a somethmg to which something 1s done and
a somethmg m 1t winch 1s done somethmg (to)-m fact, as one
object m causal act1v1ty produces a change m another object,
we think that the kno\\mg subject must, in knowing, do some-
thmg to the obJcct 1t knows and (that) that object must suffer
something. Or 1£ we don't envisage this to ourselves as clearly
as you seem to do-for to put 1t so clearly 1s, I suppose, half
the answer, we tend to thmk on this prmc1plc. Now we must
know something about knowledge, and we know when we reflect
that the very idea of it 1s incompatible with any such action
upon, or suffering in, the obJect known. You can no more act
upon the object by knowing 1t than you can 'please the Dean
and Chapter by strokmg the dome of St Paul's'. The man
who first discovered that equable curvature meant equ1d1stance
from a pomt didn't suppose he had 'produced' the truth-that
absolutely contradicts the idea of truth-nor that he had changed
the nature of the circle or curvature, or of the straight line, or of
anything spatial. Nor docs any one else suppose so. Obviously
if we 'do anythmg to' anythmg in knowmg, it 1s not done to
the obJeCt known, to what we know, for that simply contradicts
the presupposit10ns of the act of knowledge itself. If we persist
in trying to find something done to the object, we are simply
using a category applicable to the relation of object to object
 Correspondence with Mr. Harold A. Prichard 803
and not applicable to the relation of sub3ect to object, and must
fall into all manner of fallacies.
Now representation 1s only another form of the same fallacy.
We want to explain knowing an object and we explain 1t solely
in terms of the object known, and that by giving the mind not
the object but some idea of it which is said to be like it-an
image (however the fact may be d1sgu1sed). The chief fallacy
of this is not so much the 1mposs1bdity of knowing such image
1s hke the obJect, or that there 1s any object at all, but that it
assumes the very thing 1t 1s intended to explain. The image
itself has still to be apprehended and the difficulty is only
repeated. We still distmgu1sh the 1mage and the knowing, or
perceiving, or apprehending it. The theory which is to explain
sub;ect1ve apprehens10n of the object cannot, as one could pre•
dict, do anything but presuppose the absolute ultimate fact of
apprehens10n of an obJect, and so explain apprehension of the
object (unconsciously) as apprehending another obJeCt hkc it.
Obviously neither can apprehcns10n be explained m terms of
the object apprehended, nor the object m terms of apprehension.
In a way the distinction 1s not only ultimate but of extreme
s1mphc1ty-nothmg can make 1t clearer than 1tself. It 1s
'simple' because we absolutely must always presuppose 1t to
know anything, or doubt anything, or think about our knowing
anythmg. Perhaps most fallac1cs m the theory of knowledge
are reduced to the primary one of trying to explain the nature
of knowmg or apprehcndmg. We cannot construct knowing-
the act of apprehendmg-out of any elements. I remember
quite early m my phdosopluc reflection having an mstinctive
a version to the very expression 'theory of knowledge'. I felt
the words themselves suggested a fallacy-an utterly fallacious
inquiry, though I was not anx10us to proclaim (1t). I felt that
if we don't (know) what knowledge 1s, we know nothing; and
there could be no help for us. I feel sure many most respectable
theories commit the fallacy of supposing the presupposition of
all explanation can be explamed. What on earth is gamed by
'construction' or 'reconstruction' over ' representation'? When
you have made your construction you still have to apprehend
it I It's no good-knowledge and apprehension can only be
described m terms which already mean knowledge and appre-
 TENTATIVE INVESTIGATIONS
hension. Is it not true that just as those who consciously or
disguisedly hold a representative theory are leaving out appre-
hension altogether and substituting another object for it, so the
xap,lvns, the idealist constructors or reconstructors, are either
leaving out the obJect and substituting for it the activity of
perceiving 1t-th1s I thmk is their general tendency, or merely
like the others constructing something which is an object but
still requires apprehension : object on the one hand without
apprehension, apprehension on the other hand without object?
§ 542. As you know it already, I won't repeat what I have
said about the fallacy of explammg judgement as any kind of
'putting together'. As to your present special difficulty, it
apparently does not take you in, but you see it has to be dealt
with. 'Doing to' is a very general phrase. If we ask what it
specially means m the case of knowing or apprehending the
object, i. e. what 1t is we are doing to it, the only sane answer
must be we are knowing 1t, apprehending it. Knowing can't
be explained by domg to, doing to has to be explained by
knowing, and the knowmg as I have said is itself utterly mcom·
patible with any 'domg to' m the sense of effecting any change
m the thmg known. You want to know what is changed; the
answer 1s ourselves, our subJect1ve attitude. It is quite right to
say our conception of the object is changed in the advance of
knowledge, and I should always stick to that phrase as the
simplest. I admit it requires elucidation because of the mistakes
people make about what 'our conception of an object' means.
It 1s, however, the nght and natural expression and m accordance
with the use of ordinary speech ; it reqmres vmd1cabon, not
the subst1tut1on of something else for it. I will take your own
case, probably chosen advisedly, where the progress of know-
ledge about a geometrical figure is due to my own thinking.
Without believing it right1 you don't see how the following
reasomng can be met : 'm knowing, it is the thing which must
be directly before me, as my content, and, in coming to know,
more and more of the being of the thmg must come before
me . . . but it is my thinking which brmgs this about. This
must be taken to mean that I have somehow produced what is
before me and ex hyp. what is before me is the {thing or)
facts.'
 Co,,espontlence with M,. Harol,tl, A. Prichard, 805
The last sentence seems an almost artificial non sequitur. Why
must it be taken to mean, &c.? If I wonder what is the use of
a certain part of a machine about which I know of course
something already and, by examining the other parts and
thinking and mferring, finally arrive at a knowledge of what
the part is for, what its function is, must that be taken to mean
that I produced that function of the machine? And so if
I wonder whether there 1s any rule about the magnitude of the
angles of the triangle and then, by thinking about the triangle
and its relation to a lme parallel to one of its sides, discover
the equality to two right angles, why have I any more produced
the add1t1on to the content than I have produced the function
m the machine?
I wonderwhether the words I have underlined(' as my content')
have betrayed you. If you have unconsciously identified the
object as my content with some subjective state of mine (which
contradicts the very idea of 1t), then the fallacy may be caused
{in that way) : for my sub1ect(1ve) state has changed ; I appre•
hend more of the obJcct than I did before ; my conception of
it has changed, and by the activity of my thinking. Identify
then my subjective state with the object and we have your
conclusion. Any idealism winch led to such a conclusion would
have to maintain I made the function of the part of the machine.
When I apprehend the proposition about the triangle, I appre-
hend something universal, which can't therefore be any state of
my sub3ective consc10usness. Suppose you repeat to yourself
the argument I have quoted from your letter, substituting in it
the umversal proposition about the triangle for the general
'fact' or 'tlung', and see how it sounds to you. (Just let me
say, while I thmk of it, you seem to have overlooked a serious
defect in a distinction you quote-as between
(a) the thing about which we think,
(b) idea as = what is before the mind when we think (Locke),
(c) the idea as = the activity of thinking.
There seems to me here to be much confusion. Idea can never
be the activity of thinking ; a more confusing use of it I can
hardly imagine. Idea is surely always • idea of', always implies
something apprehended, and not merely apprehension. And
when people use it in distinction from reality, they always mean
806 TENTATIVE INVESTIGATIONS
some mental image together with the Judgement that the reality
is more or less hkc it.
And agam, how distinguish (a) and (b) m the case of this
universal proposition about the triangle? 'The thing about
which we tlunk' 1s the equality of the angles of every triangle
to two right angles, and that is precisely 'what is before the
mmd when we think' The only distinction that can be made
here 1s between the part of the thing which 1s before us when
we thmk [as opposed to] (and) the parts of it which are not before
us. E. g. the equality to two right angles involves many other
properties which are not 'before us' when we have completed
the proof of this property itself.)
You must infer, as you apparently do, that according to me
Judgmg is 'putting thoughts together'. I repudiate all manner
of 'puttmg together', and I don't see how 'putting together' 1s
ever an mtelhgible express10n , at best 1t would be a clumsy
fafon de parler requirmg a good deal of explanation. Suppose
M m the syllogism of mme to wluch you refer is tr,angularity.
I may apprehend 'tnangulanty ', 1. e. understand a closed figure
can be tnangular. I may apprehend that three angles m a more
or less complicated figure before (me) are the three angles of
one and the same tnangle. I may not yet apprehend the pro-
perty of triangulanty as involvmg equality of three mtenor
angles to two nght angles, and so haven't yet got the premisses
necessary to complete the syllogism and prove that the three
given angles = two nght angles. When I do, M has suffered
a change, that is, my conception of tnangularity has suffered
a change: which means sm1ply tlus, that, whereas I knew some-
thmg about triangularity, I now know something more about
1t, which implies no change m the thmg but a change m me.
I not only know M 1s a, as I did before, but that M 1s also b.
This doesn't mean puttmg thoughts together, but that I have
passed from thmkmg M is a (and also contains an unknown
margin x) to thmkmg or apprehending M as a and as b. All
language about puttmg thoughts together or things together is
simply misleading I haven't here put any thoughts together
or any thmgs together ; I pass from the apprehension of M as
a, and the apprehension of M as both a and b, to the apprehension
that a and b are together m M. This is the simple account and
 Correspondence with Mr. Harold A. Prichard 8o7
nothing can make it clearer than 1t is. I think S 1s M, appre•
bending only the a in M ; afterwards I apprehend b in this M,
which I recognize as the same M as I attributed a to, and so
my thought that S is Ma is replaced by the thought that S is
Mab, and, if b = P, I have the conclusion of the syllogism.
I have the same M before me each time, the second time I appre•
hend more in 1t than I did before. In some ways it would be
more accurate to say not that my conception of M has changed
but that I have changed from one conception of M to a different
one. That 1s, I have changed from one judgement about M to
another Judgement about M. If you like to use the language
about my content, you may say that more of M has become
my content, that 1s simply, more of it is revealed to me or
apprehended by me ; but M is in no way itself changed by this,
because, as I have contended, my apprehension produces no
change m the obJect, in the nature of the object as 1t is known:
such categories do not apply at all.
§ 543. Now let us consider what we mean by our 'idea of
a thing', or 'conception of a thmg'. Suppose an object of per-
ception-Cologne cathedral. If asked what was my idea of 1t,
I should at once state certam Judgements of mme that the
church (a reality) had certain real attributes when I saw it.
Such Judgements are accompanied by a mental image, but that
IS not my concept1011 or idea of the church, nor do I say It 1s.
On the contrary I think the image 1s more or less inadequate,
and I say the church was something hke the image I am forming.
The Judgement 1s obviously not the image and can be reduced
to no kmd of terms of images. The image 1s as distinct from
the judgement as the obJect (the church) of which I say I have
an idea. In the Judgement about an obJect not present to
perception, what is the fact before me? It is not the church
but the fact that the church looked so and so when I saw (it) ;
it is a fact now that this was so m the past. It is the fact which
is before me, the fact, i. e. which I apprehend in the judgement
now. There is nothing between me and it, no intervening image.
I may forget details but I remember a general character, such
as two spires, very much crocketed. And this which I do not
forget, viz. that I apprehended then the general character, is what
I now apprehend. Besides this there is an uncertainty about
8o8 TENTATIVE INVESTIGATIONS
details not certainly remembered ; this is not an image, it is
a state of my thinking consciousness which can be expressed
in no terms except its unique self, uncertainty. If now we take
a universal reality hke triangularity, my conception of this, my
idea of this, when analysed, is not my image, but a direct appre•
hension of a part of the reality, viz. the enclosure of space and
the three angles as necessitated by the three sides. I may, ·
mdeed must, have either 'perceived triangles' before me, to think
it at all, or mental images, but they are not what I think;
I think the reality which I directly apprehend. It seems to me
that even the most advanced people one reads are under the
dominion of the view that judgement is somehow idea as =these
representative simulacra.
§ 544. The judgement of knowledge is apprehension of reality
or fact, it is not the fact, it is not the reality ; but neither is
it any image of the reality nor the apprehension of any such
image. Nor is it mere apprehension as a subjective state-mere
apprehension is impossible-it Js (1) apprehension (2) of the
reality.
The ideal element we are lookmg for, and always (all of us)
tend to misrepresent as an 1mage of the reality, is the appre-
hending side as our act: the fact that we apprehend the reality.
It is not here the least necessary to say how we come to be able
to apprehend reality ; the whole point is that the very concep-
tion of knowledge as such must necessitate this and nothing
short of that wJll do. . . .
You have really done me a great service. It is one of the
most useful things possible to me that I have the good fortune
to know a number of earnest thinkers ... who will put difficult
questions to me and not let me off. The cleverest man in the
world gets into his own grooves of thinkmg, and that can only
be cured by the independent activity of other people's minds,
and I consider myself specially fortunate in having so much
access to it. . . .
I went to the Watts Exhibition on Saturday (this letter has
flowed over to the Sunday). It is splendid and probably a unique
opportunity.
 Correspondence with Mr. Harold A. Prichard Bog

New College, Oxford.

23 April 1906.
§ 545, Confidential.
l\fy DEAR PRICHARD,
Do you see any answer to the short criticism of Joachim's
book overleaf? 1 If there is none, it is quite enough to nullify
his whole book. If you think there is, do not trouble to write,
for it might be a long business, but just say you think there
is an answer and we can talk about it some time. It seems to
me shortly that his book turns entirely on two fallacies : (i) the
identification of what is apprehended with the apprehension of
it, which produces the statement I criticize ; (ii) the fallacy of
making truth some sort of correspondence between. He can't
shake himself free of this. It reappears in the form of' coherence
between' and leads to the fatal nonsense I refer to (below).
§ IO, p. 281 'I answer "Euclidean space" is itself a system of
such judgements.' But a judgement is a judgement aboul some-
thing else. That something else is not the judgement itself
No judgement is a judgement about itself. The judgements
Joachim refers to are of the type 'the internal angles of a triangle
are equal to two right angles' (p. 27, bottom).
(i) If the system of such judgements is space, then, by the
above, since what they are about is not themselves, what they
are about is not space : quod est absurdum.
(11) To (meet a reply based upon) the difference between a
single judgement and the system :
(a) A system of judgements, as a system, is about something
else not themselves. The system of judgements is not
about itself nor about any one judgement in the system.
Therefore the system of Judgements which is space is
about something other than space. Same absurdity as
before.
(b) To (avoid) the absurdity in (i), reply that a single one of
these judgements is about space. Then it is about the
system of judgements which = space (which includes
itself). But it is too obviously not about any such
system and, if it were, it would be about itself.
1 Tiu Nattw• of Truth, by H. H. Joachim (Oxford), 19()6. Cf f 240.
810 TENTATIVE INVESTIGATIONS
'The truth of each of them (i.e. the judgements)', says
Joachim, 'is "constituted " by its coherence with all the others.' 1
This I consider absolute nonsense into which he is driven by his
theory or rather by the idea of correspondence between two some-
things which he can't get rid of. I think I demonstrated in
informal instruct10n this year that the attempt to represent
'truth' as any kmd of 'correspondence or agreement between•"
1s suicidal.
Yours,
J.C. W.

New College, Oxford.

Thursday(? 1906).
§ 546. As to truth of single Judgements (I 3ust sec the title of
your paper), I have been accustomed to contend, either as
undictated comment to what I say of syllogism m lecture (or
in informal), that the syllogistic form itself gives the directest
contradiction to the view that sometlung cannot be known as
true m itself without its coherence with everything else. On
the contrary, the syllog1sm implies that we can know a smgle
truth and that the truth of it cannot possibly be altered by any
new knowledge It means that when we say S is M (all A is M)
this cannot possibly be altered by any new mformation about P.
This is what makes the syllogism SM, MP, SP vahd, otherwise
when we got MP we should have to reconsider the vahd1ty of
the ma1or. 2 Of course S 1s M means just S is Mand whatever,
known or unknown, M agam necessitates. Also I maintain (m
informal mstructlon) that, if the truth of a smgle 3udgement
depended on coherence with others, its truth could never be
known, since (e. g geometry) the number of possible Judgements
m the subject is mfimte. It 's only because we know it is true
that we can know it coheres with (the) infinite system. In fact
it is a theory of pure Scepticism, 3 not of knowledge. Agam it 's
absurdly contradictory to the facts of science. That three
straight Imes are enough to enclose a space and that they, as
so enclosmg, necessitate three angles is known by itself, and can't
be modified by or depend upon any other proposition for its
validity. Agam, if you have a set A of judgements all 'cohering'
1
Cf , however, 1b § 37 • A slip for • Zlllnor •. • Cl 1b pp. 199-8<>
 Corres,Ponaence with Mr. Harold A. Prichard 8II
with one another and a set B of judgements all cohering, if
A and B don't cohere how the devil are you going to say which
set is true? Scepticism again and of a naive character (other
considerations in this line too). Agam, ask any one who holds
such a doctrine whether he thin](s this 1s true :
If A is B
and if Bis C
then A is C
and, if so, if he knows and how he knows 1t coheres with every
other truth. Of course one sees that idealism of a certain kind
is driven to this theory of truth, which 1s enough to refute it.
]. C. W.

By the way I have to go to my doctor (Osler) at 4 30-not

well yet I

Bad Kreuth, Bavaria

13 July [1906].
§ 547.
As regards your mterestmg Jetter 1t 1s very welcome mdeed
because I have been rcflectmg that m preparing my lecture for
publication I must probably expand or put in a different form
the introductory port10n. It contains m the para• on the def•
of Logic the essentials from wluch started my whole critic1c;m of
the use of 'idea'. I have only been developing more clearly
what I committed myself to in the para. about the 'content'
of thought bemg the ob1ect's own nature and there bemg nothmg
else m the thought except the apprehension (or as I have more
imperfectly said sometimes in the lecture 'presentation' of the
object's naturc-prcscntat1on won't do, 1t 's too metaphorical
and causes difficulty when we come to perception). I have been
thmkmg e, g. that this important position ought perhaps not
to come m as incidental to the discussion of the defimt1on of
Logic; but perhaps I ought to start with 1t. It comes m agam
in a clear manner m Judgement when I mamtam that, when
I am thinking of Jones (or Williams, was 1t)) m the next room,
the object of my thought 1s Just Jones and not a picture of
him, &c., &c.... You have only expressed m your own language
(and I think not in the right kmd of language) exactly what
a77ra Dd
Sn TENTATIVE INVESTIGATIONS
I meant.•.. I think I can convince you of this in the present
case.
(1) I did not mean to say .•. that 'idealism {of any kind)
upset the view that logic studied thought and science studied
things'. What I said, and I feel sure I said it accurately, was
that we couldn't retain the statement simpliciter now that forms
of idealism had developed themselves which in their respective
ways equate thought and thing-whose very business it is just
to deny somehow very emphatically the opposition of thought
and thing. I pointed out that in consequence of the apparent
emph0 denial of the very distinction upon which the above
differentiation of Logic from Science depends we were obliged
to take notice of these idealisms and we might naturally too
suppose we had to decide for or against them But then that it was
obvious on examination that idealism in whatever form involved
the direct presence of the obJect to consciousness and so merely
the distinction of the presentn or apprehension of the object
from the object.
§ 548. Similarly that the materialism opposed to it made
ultimately (in its theory of perception) the ohJect with which
we are concerned something directly present to consciousness
or thought. Thus we after all hadn't to decide for the purpose
before us between these opposing theories, for they agreed in
this important point that the object of thought was not some-
thing •outside' thought but 'inside' it. Now this of course
implies that subjective idealism is for this purpose m exactly
the same position as the materialism which d1stmgmshes between
thought and thing (I am sure I made this clear) : for both of
them there is something d1stingmshed from the thinker as object
and in both of them this 1s supposed to be directly apprehended.
This I said was enough for Logic, i. e. the distinction of the
apprehension (subjective side) from what 1s apprehended (objective
side). This of course exactly means that idealism does net
destroy the distinction between the subjective and ob7ective side,
when we come to examine it, but leaves it as the distinction
upon which the difference between logic and science is grounded
just as much as the ordinary materialism leaves it.
§ 549. Now this seems to coincide with what is the real truth
jn your supposed objection, but I think you do not express
 C011'espondence with M,. Hr,old, A. P,icna,d 813
it in a form which [it] is at all advisable, and J think I can
make you see why I avoided that expression.
If you like to call the object Thing and the apprehension of it
Thought, then the above doctrine may of course be put thus :
in subjective idealism there is a distinction between the object
thought and the apprehending of it. Therefore, if we call the
object 'thing' and the apprehension of it 'thought', then subj.
1d• distinguishes thought and thing ; and as I go on to make
logic of the apprehension and science of the object (shortly), so
I make logic of the thought and science of the thing, even if I
adopt subjective ideahsm-which may be put in your language
that 'subj. idm does not upset the statement that science studies
things and logic studies thought'. But in choosing this sort of
language you seem to me to be obscurmg again what I had
taken some trouble to get clear. It does not do to retain sim-
pliciter the statement that Logic studies thought and that Science
studies things, for the very vital reason that the formula usually
conveys the implication that things are something outside
thought altogether. It most certainly did, and does, in the
common sense realism mean that. The thought not bemg at
all supposed to be the apprehension of the thing, but something
self-contained, something entirely mental and only apprehension
of the thing as being an apprehension of some sort of a replica
of the thmg (i. e. when you push them to say what knowledge
of or thought about the thing 1s). In fact by using this sort of
language, which I carefully avoided, you obscure the very point
on which you agree with me, and thmk, as I do, of the last
importance to keep clear. Remember the sort of locus classicus 1
for this wrong way of putting the object of logic, De Interp. 1
ch. i, where language 1s said to be a sign of thought and thought
a sign of the thmg. Clearly the writer makes the distinction in
the interest of what we call logic (but I won't say any more
about the passage). I am getting a bit tired and so will stop-
but if I haven't made it clear even yet please write to say so.
I KAl ll,,r,r1p ou3i "ff)4Jl,/Jl1.TO.,,a,,,
Td drd, wlli ~r,al 'al_a.vral· @I µlrmM ftffl
,,.,,.,,.. ,rpJn-on, TIIUT« ,,a,,, ~/Jl1.TO. riir t/Nxiis, so.I @I TaiiTa. 1i,...,,r1,,_TO., "flG"fpa:rG ff"'1
nlrra (tl, l•t 1611 5-8) Whereas all men do not use the same letters nor
utter the same vocal sounds, the a:ffect10ns of the mind, which these are
signs of, are the same for all and, further, the thmgs of which these aflections
are sunula&f'a are the same.
od2
 TENTATIVE INVESTIGATIONS
I will only add now that I am sure there is no difference but
a verbal one, and I think this time your verbal form not good
but positively misleading....
The very essence of what is contended in thos'e passages is
that thought about the thing is inseparable from the thing: that
th13 thing is an 'clement' in the complete act of thought, the
significance of the latter commg out when we put it negatively-!
viz. that tf we take the thing's nature away, the thought is gone
too, for without the thmg it has no 'content'.
§ 550. (As to my lecture on complex and simple ideas) On
consideratton of thmgs akm to this I thought it important for
polemical purposes to express the same thing two ways, 1•t in
the ordinary (which is really conceptualist) language, and then
in what would be the accurate language acc. to my own view
of rein of thought and thing, and I have done that in one
important place 1 m my revised hyp 1 lectures as you will
discover.
Oxford.
18 Dec. 1906.
§ 551. •.. What you have practically asked me to do is to
bring what I say about the relation of Thought to its obJect in
the Introduction to my Logic lectures ... into clearer and more
definite connexion with what I have communicated to you on
my views about the mistakes m all forms of ideahsm. Or, m
short, you ask me to adjust my Logic to my Metaphysics....
So m introducmg my new and more detailed consideration of
the subject to my present class ... I said it was due to a colleague
who had made me feel the necessity of putting this part of my
logic into clearer and more satisfactory relation to my views on
the theory of knowledge-metaphysics mdeed.
For some years past I always felt a certain uneasiness about
this part of my lecture, and have each time thought about it
again. But I always found myself pulled up by a passage in
the § entitled 'On the object as within consciousness '. 2 After
contendmg that what is supposed to be the very 'content' of
thought, i. e. what gives it actuality and life as opposed to
a mere potentiality, is the nature of the Thing (directly and
immediately-nothing between), that (what) we think is the
l 1 C §§ 289-90
 Co"espondence with M,. Harold A. Prichard 815
Thing's own nature, not some thought different from the Thing,
I go on to the very objection which is the real ground of your
present difficulty and the ground of whatever difficulty I have
felt myself about the preceding part of the §-a difficulty there-
fore which I have had before me from the beginning. I find
it e.g. in Brade's notes taken in 1896; but I find a better
version of it much earlier in W. A. Craigie's notes, which are of
hoar antiquity. 'Again we may try to separate the object from
our thought, by saying that our thought is not the nature of
the thing but the recognition of it (i. e. of that nature). But
we shall find again that the content of the act of recognition,
which makes it more than a mere form, hes in that very nature
which we suppose thought to recognize. Or, otherwise, 1f we
abstract from the act of recognition what 1s recognized, we shall
find 1t empty' (Brade's version adds 'and meaningless'). Every
time I came back to this I felt I was not going to give it up,
that rt was the essential-and that the mam drift of these§§ was
therefore right whatever it might lead to.
Now I always was careful to put this by itself, on its own
merits, without assignmg rt to, e. g, any form of idealism (for
from the first I would not commit myself even to the most
attractive form of idealism, tho' greatly attracted by it). It 1s
here that you misunderstood my attitude.... Certainly I was
clear enough as to what I meant. I say m effect 'I am not
going to decide about these realtsms and idealisms, I don't need
to, for such and such 1s my view about the relation of "what
is thought" to thmkmg, and these theories themselves involve
what m that view 1s essential to my purpose.... •
It 1s the inseparableness in the act of apprehension of the
thmking from the ob1ect thought, which is not the least got
rid of because the oh1ect can exist as unapprchended, which
forces itself on one as the most important fact ....
However partly your criticism and partly my resolve to begin
to prepare these things for publication made me decide to put
the wholethingmore plainly, and as I have of late years become
much clearer as to my attitude to any form of idealism 1t was
easier to do so. I devoted myself therefore to trying to find
the exact character of the 'inseparableness' and what could be
fixed as its necessary presupposition and consequences. And,
816 TENTATIVE INVESTIGATIONS
keeping the essential statement exactly as it was, I have recast
and greatly added to my previous discussion. . • • And when
I had finished • . . I found it advisable-again with a view to
publication-to develop more fully a good deal more of the
lectures beginningwith the various senses of theword 'subject',
and this kept me at work all the term. Among other things
I added, after the negative criticism of Bradley, a positive'
attempt to represent the real genesis of his views, so as to show
how he gets into the statements which appear so queer and
eccentric.... I feel not only that what I have written in the
new lectures is the best discussion I could give of the points
you raise, but it anticipates some of your points in the very
form in which you have put them, e.g.... I find that I have
added a passage making still clearer than before how I use
'within•. But this was only natural because your difficulty ...
was that this representation of the inseparableness of Thing
from Thought in the act of apprehension (with therefore somehow
the necessary inclusion of the being of either within that of the
other) might be supposed to really identify the Thmg with state
of consciousness, especially when one uses the word 'within'.
I was therefore led as naturally as you were to make the 'within'
clear, though I contend it was clear enough in my somewhat
carefully phrased statements already. . ..
I may remark that I think I have made an improvement in
termmology. I used to employ the words recognize, recognized,
recognition, but for a year or two (perhaps more) I have steadily
used the words apprehend, apprehended, apprehension, as being
the simplest and truest expression of what is meant. It 1s
partly the feeling of (the) necessity of some general word which
introduced the barbarous cogniu and cognition, which nothing
would induce me to use. Such technical terms always tend
to make a sort of meaning for themselves-as if they were
some new metaphysical discovery-and so produce no end of
confusion. . . . I have re-written my account of Judgement
accordingly.
 Co,,esj,ondence with M,. Ha,-old A. P,icha,-d 817
[? Oxford.]
§ 55z. Monday, 9 Nov. 19()8.
MY DEAR PRICHARD,
I am greatly pleased by your very liberal acknowledgements
and am touched by the kind words which accompany them....
It isn't quite enough to make the distinction between 'of'
and 'between'; because, jf you add 'to' after 'of', you get the
same result as if you wrote 'between'. Thus it must be 'relation
of universal and particular' not 'relation of universal to parti-
cular'. You might say 'marked in language by the use of the
preposit10ns "between" or "to"-the relation between A and B
and the relation of A to B in the one case-and by the use of
(the word) of only in the other, "the relation of universal and
particular'". . . . Observe also that one uses this OF idiom of
thmgs related which have a nature other than what is con-
stituted by the relation, when we are only characterizing them
as related-thus 'the relation of father and son' .•..
I don't agree with this, for I thmk 1t misleading. You
don't give it as merely Kant's <view) but approve it as valid.
I have no doubt you would agree that our judgements in
Geometry are in a sense (you say no sense} based on experience
because we can't possibly have them without experience. We
must experience particular space relations before we can get the
universal geometrical judgement. The method of Geometry is
nothmg but apprehendmg the conjunction of universals in the
particular, and, though we don't base them mductively a pos-
teriori on the particulars, we must have the experience of the
particular, &c., &c. lf you meant that 1t was the best argument
Kant had or that 1t was valid with a change of phraseology
(1. e, om1ttmg the statement (that) they were in no sense based
on experience) you should say so...•
I am overjoyed at the completion of your difficuJt task, and
much honoured by the handsome acknowledgement to myself.
Both are a great comfort to me at a time when my own power
of work is so unfortunately limited, and just when I need it
most. With very kind regards and very hearty thanks,
Yrs. truly,
J. C. WILSON.
 XIV
LATER CORRESPONDENCE WITH B BOSANQUET
The Heath Cottage,
Oxshott.
§ 553. March 2 1 1909.
MY DEAR WILSON,,
The pomt on which I differ from Joseph's account of Induc-
tion is the relation of disproof of hypotheses to the establishment
of an hypothesis. I understand him to hold that an hypothesis
can only be established by the disproof of competing hypotheses.
The view is formally very convenient, and not easy to refute
in detail, which makes me all the more regret that a man of
great ab1hty should not have assisted in statmg and illustrating
the more difficult doctrine, which I discussed under the heading
of 'Fus10n of data and Hypotheses'.
His view {practically Jevons' also) seems to me a piece of
what I call 'the negative approach' which 1s fash10nable to-day
-a sort of sh1rkmg the establishment of rational connexions,
deeply akin to many philosophic phenomena of the day. It is,
at root, I suggest, the same as saymg that A is A is secondary
to and rests on A 1s not not A. One recalls Bradley's account
which takes the opposite view.
Of course, we have been warned that cause need not be
'similar' to effect. But none the less 1t seems clear that there
must be a positive continmty of nature ; and that, where this
can really be shown, no operation wtth suggested other con-
nexions is necessary, and where it is not shown, no such operation
can supply its place.
Stout, I think, agrees with me, but only on my ex parte
statement, about J. 's view. I have no real right to cite him
in my support, therefore.
It seems to me that the disproof of alternative hypotheses
may, strictly speakmg, tell you nothmg at all about the hypo-
 Correspondence with Bernard Bosanquet 819
thesis which is not disproved. And it is very difficult to see
how the truth of any conclusion can be established without
analysis of its positive continmty with its premisses. Somebody
-it may have been Hoernle but les absents ont toujours tort-
gave me the impression that you were with Joseph on this
point. I hope it is a wrong impression.
Yours truly,
B. BosANCJUET.

The Heath Cottage,

Oxshott.
§ 554. March 21, 1909.
DEAR WIL-SON,
I will answer quite straight without trymg to guard my-
self ; it is best among experts, I thmk.
I do not know the mathematical proofs to which your first
question refers But I thmk that in strict logical parlance the
proof of a 0T1 without the iufr, is a contradiction m terms.
Proof I understand to mean the establishment of such a con-
nexion between a propos1t10n and the whole of experience that
the two stand or fall together. Explanation I understand to
be the establishment of a connexion between an alleged fact
and the whole of experience. I have no doubt I could deal
with any case the detail of which was w1thm my comprehension.
But it would be waste of time to deal with cases hypothetically.
Besides, you can predict exactly what I should say. I attribute
weight, of course, to what you say about a mathematical pro-
cedure , I attribute none to a mathematician's logical description
of his procedure.
As to the second question, I used often to talk it over with
old Faulkner at Univ. He said that those proofs could only be
drawn directly from the consideration of the nature of whole
and part, or, as I should say, from the nature of quantitative
comparison in which a plus or mmus can be established without
bemg determinative. As to the case you give, I am wholly
unfamiliar with the language and technique ; but what it looks
like to me 1s th1s. Within the formula the diminution of 0
without limit is sufficient to explain the conclusion that the
magnitude•is L. If you were applying tt to any fact, say, in
 TENTATIVE INVESTIGATIONS
statistics or actual magnitude, I do not think the proof would
hold without an explanat10n of what you mean by the diminu-
tion of 8 without ltmit, apart from considerations showing L to
be the value. In many cases of this type the mathematical
proof so-called is simply invalid when applied. Piazzi Smyth
applied a method of eliminating error (' least squares' I think)
to different measurements of a rough box of stone found id
Egypt, and brought out a result which he applied to determine
the umt of measurement used in constructing the box. The box
had. no dimension ; it was rough granite, and you could not say
where the surface was intended to be. The whole calculation
was a sheer :figment.
This is a distant illustration, and perhaps irrelevant ; yet
I thmk 1t goes to (prove) this pomt ; I will never accept any
mathematical proof of any concrete fact, unless it is explained,
or shown to be exphcable, how the proof connects the charac-
teristics of that fact with the whole of experience.
It would be waste of time to write more, till one knows what
further pomt you take.
Yours v. truly,
B. BOSANQUET.

[7 April 1909.
[To B. Bosanquet.] Oxford.]
§ 555. . . . The case I referred to is this.
Suppose an attribute a 1s proved of a subject S whose whole
definition 1s A1B1C1, where A1 is a d1fferentiat10n of a umversal
A. Suppose A 1s enough to necessitate a, bemg what mathe-
maticians call the 'sufficient and necessary condition' of a.
A then is the true a&efr,. If we prove a of S by deriving it
from A1 we haven't then got the true auSn. As e.g. if we
prove of an isosceles triangle that its interior angles ... two right
angles, arguing partly from the equahty of the sides. This
corresponds to Aristotle's dictum, of course, that the property
must be proved of the ,,,~ro11 ,ca8Aov e. g. of triangle (A) and
not isosceles triangle (A1). (Correct as this is, he gives no
criterion for knowing when the proof is from the wp. ,ca.(J. ; nor
would his syllogistic theory of scientific proof allow him to.)
If then we prove a. from A1, we haven't got th, true a,&n,
 Cor,es,Pondfflce unth Bernard Bosanquet 821:
if we prove of the isosceles triangle (using equality of sides as
we may) that it has the given property, we don't know the real
reason, which is not that it 's an isosceles triangle, but that it 's
a triangle.
Still further are we from the a,Jn if our proof should be from
the whole definition A1B1C1•
Now it might seem that in the demonstrative sciences where
a is really proved of S that generally (i) we should prove a. from
its true cond1tion A. And as a matter of fact this is so nearly
always in Euclid. We always, with some rather unimportant
excepbons, do argue from the 11piirro11 ,ca86>.ov. But (il) supposing
it happened in some case this was not so and we derived e.g.
from A1, we should expect that we should recognize A1 as
a differentiation of A, and recognizing A we might either
(1) observe m our proof that nothing really depended upon its
different1ation mto A1, and we had only used A; or (2) could
observe that we had proved a. in the special form o.i, and see
that a depended on A, or (3), at least, recognizing A, we might
test our proof by omitting its d1fferentiabon, and try whether
we couldn't denve a. from A.
But suppose, what at first might seem an impossible hypo-
thesis, that we denved a property a of S from the whole definition
of S, not from the element A alone in S on which 1t depended ;
not knowing that it depended upon A alone ; not being able
to see in our proof that the connexion was with A alone ; and
not even recognizing A as A at all-would it not be accurate
to say, 1f such a thing were possible, that we bad proved the
fact, the 6T1, that S has a, but had not shown its reason-
..-a ~,on-we had not proved 1t from its true reason ; for indeed
we do not even know its true reason or condit1on-we neither
know its condition as its cond1tJon, nor do we know of the
existence of its condition, for its condition has not even appeared
in our proof ? Consequently though sure of the existence of the
fact we can't be said to understand it.
§ 556. I imagine that when this statement is made the first
answer in general would be that this would undoubtedly be the
state of our knowledge: i.e.
(1} that we had proved the ST& not from the &,dn ;
(2) that we don't even know the a,oT,;
 TENTATIVE INVESTIGATIONS
(3) that therefore while being sure of the fact we don't
properly understand 1t : i/ the state of things described
could occur.
But it would probably be said such a state of things cannot occur
in a true demonstration issuing in certain knowledge of
the fact.
It might seem the only thing at all corresponding would be ·
cmpmcal-experimental, inductive argument (1. e. eliminative)
where we have not eluninated the true condition, but perhaps
more than the true condition, e. g. have eliminated AB as true
condition where B 1s superfluous; or we may even suppose
a substance X to be the condition, or part of the condition,
and yet not know an element A in 1ts nature to exist, which
element is the true condition (And here too the kn0ttJledge of
the fact so far as we have it is not derived from the condition
supposed ) But now the thing does happen in exact demonstra-
tion, as we find when we advance beyond Euchd to Come
sect10ns. 1 Among the properties common to all comes as comes,
there are some, notably those connected with tangents, which
it is usual in the elementary books to prove of each come. Such
proofs start from the full defimtion of each conic and use the
full definition. The properties in question are not connected
with an element in the definition common to all conics, but
depend upon the 1(()hole definition : nor even does the element
common to all comes appear either m the defimtion itself or in
the proof.
For instance, the harmonic property of pole and polar is
proved in the case of the circle by a proof which depends upon
and uses the whole defimtion of the circle as the defimtion is
given in Euclid (I can send you references to books 1£ you hke).
Nevertheless the property proved 1s not pecuhar to the circle
and does not belong to it as circle but as 'conic section'.
But though circle is a differentiation of a universal 'conic
section', this does not appear m the aforesaid definition of
a circle at all.
. We might, out of the definition, get the genus 'plane figure
included by one line', but this will not do. It is too wide and
is not the definition of come section Thus, you observe, the
1
Cf § 104.
 Correspondence with Bernard Bosanquet 823
universal on which the property depends cannot be abstracted
from the definition used : and the definition has to be used as
an inseparable unity. Moreover, this universal cannot be dis•
covered in the proof, nor can a generic character of the proof
be abstracted from the specific proof before us, which would
fit any conic section.
The student first reading the proof sees in it no reason what-
ever for supposing the property attaches to anything but the
circle. He is not the least conscious of any difference from the
ordmary procedure of Euclid which (as I have said) correctly
deduces a property from its true condition and so is not applicable
beyond the subject of which it is proved.
This, by the way, raises the question how it is we know that
the Euclid proofs are thus correct. To which I think the
answer is that in general we cannot see it from the mere form
of the proof any more than we can see that the aforesaid proof
of polar properties 1s not' correct' (m the given sense). We can
only test it by seemg whether the converse is true, for if the
property is derived from its true condition { =sufficient and
necessary condition) the converse 1s always true. Now this can
always be done m the ordmary case in Euchd and is done
explicitly, e. g. propos1t1ons 5 and 6 of Book I, propositions 47
and 48 of Book I, unless the converse is felt too obvious to
need proof. However, Euclid is a httle deficient here because
the converse of I. 32 ought to be proved. The fact that
e g. the converse 1 has to be proved m Euchd, I. 4812 shows
that the proof of I. 47 does not in itself reveal the fact that the
property is proved from its sufficient and necessary condition.
(The kmd of unimportant exception to reciprocating proofs is
Euchd, III. 2 1 which is true of the ellipse and parabola as well
as the circle.)
§ 557. In the instance then which I have quoted the student
in proving the property (1) does not prove it from its true a,orc,
(2) he does not know the true aufr,, (S) he does not know even
of the existence of the true auST, : he doesn't know the circle
is a 'conic section', (4) even an examination of the proof will
not at all reveal the generic element.
1
Viz theconvenieo/theenunc1abonof Euclid, I 47 (Pythagoras' theorem).
1
MSI47
 TENTATIVE INVESTIGATIONS
On this account it seems right to say such a proof is a proof
of the Sn-a real proof-so that we are certain of the ST,, but
it is emphatically not a proof by the &,on ; and it is right,
I think, to say that the fact is not properly 'understood' and
not properly 'explained'.
In order to bring out this better, I may draw attention to
the fact that even when the properties in question were seetl.
to be common to the vanous conic,;, 1t was some time before
a definition of the element common to the conics was obtained
from which the given tangential properties could be proved in
one proof.
'Conic section' is no help, because the cone has itself to be
defined as the surface formed as the locus of a hne which passes
through a fixed pomt, i. c the vertex of the cone and through
every pomt on a circle (an clhpse also, &c., would do). We can,
1t 1s true, then, save 1 separate proofs for each conic section, by
provmg for the circle only, and extending by 'projection' (1. c.
perspective) to the other conic sections. But we begin with the
proof from the definition of the circle and so arc no 'forrarder',
for that proof uses the whole definition of the circle.
The genenc element which has been discovered at last, wluch
forms a general dcfimtion of which every come is a species, is
this : 'A come section 1s the locus of a point which moves so
that the anharmonic rat10 of the pencil of Imes drawn from it
to four fixed pomts is constant' [figure drawn]. And from this
defimtion the pole and polar properties (of tangents) common
to all conics have been proved. You will observe that this late
product of geometry 1s not to be found as an element in the
defi.mtlon of the circle, and certamly 1s not discoverable in the
proof-such as I am referrmg to-of the tangential properties
of the circle common to 1t and other 'conic sections'.
§ 558. I may add that the case 1s different m analytical
geometry. The defimtion of the circle gives us its equation
x3 +y8 = r2 where r 1s the radius. If we form the equations
of the tangents from a point, we easily establish the properties
in question which belong to other conics also. But in this case
we can discern the generic element, for we see that the variables
have certain definite coefficients, viz. + I for each of them. The
• Vu: (apparently) escape the necessity of , .
 Cor,espontknce with Be,na,tl Bosanquet 825
equation then is seen at once to be a species of the genus
a,r;I +byB • cl, where a and b may have any value and any sign
(c doesn't matter-I needn't explain why}. This is the equation
to a curve of the second degree.
This may suggest to us that it is as a curve of the second
degree that the circle has the given properties. We then see
that we can similarly generalize the equations to the tangents i
and then we find we can demonstrate the properties universally.
It was then really comparatively easy to find the generic
clement in Analytical geometry, and it is true that the special
proof suggests the generic one. But m pure geometry this was
not so-' conic section' has a verbal generality parallel to 'curve
of the second degree', but, as you have seen, isn't at all in the
same position.
§ 559. Mathematicians are, of course, aware of the necessity
of getting generic proofs of properties which belong to every
species of a genus ; but they are not, as far as I know, at all
alive to the sigmficance of the facts to which I have drawn
attention, and the kmd of curious paradox they involve which
we may put as either (1) proving the fact not from its 'reason'
and not knowing its 'reason', or (2) as somehow using the
' reason' m our proof without knowmg we are domg so 1 or even
knowing what the reason is.
Nor do I know that any logician has called attention to this,
much less discussed it. (But I must grumble that logicians in
general seem to me to have too little studied actual processes of
demonstration m the sciences ) I should say also that tho'
mathematicians know in general they should get generic proofs,
owing to the difficulty in non-analytic geometry of formulating
the common element sometimes, they will present a proof which
should be generic adulterated with specific proofs. I have before
me (Drew's Conic Sections, Appendix, prop. VI} a proof of
tangent properties which professes to be generic, and yet in one
place uses a proposition which has only been got by proving it
of each conic separately I
Of course, I am aware that there's a certain possibility of
misunderstanding if one speaks of 0.1.e reason ; but if C is the
sufficient and necessary condition of E, everything such that it
and C mutually necessitate each other is in the same position
 TENTATIVE INVESTIGATIONS
as C as a10T1, and we don't know how many such equivalents
there may be: the point remains that in the proofs I mean we
don't prove by a.ny form of true a,on.
Yours truly,
j. COOK WILSON.

The Heath Cottage, •

Oxshott.
§ 560. April IO, 1909.
Mv DEAR WILSON,
I am sorry indeed to hear of all your worries. Anything
affecting your wife's health must be so much worse than any-
thing that could happen to yourself. I hope the whole thing
will settle down before term begins : it is awful to have to work
1f you are not sure that the household is running smoothly.
It is extraordinarily kmd of you to find time to write all this
to me. I must put your letters together so that when your
biographer's time comes, not m my day, they will be accessible
to him. Your correspondence will be a large and valuable
philosophical asset.
I had better recapitulate the points as I understand them
from your letter. Of course I cannot here, as perhaps I could
in some matters, hope to give an ongmal analysis and solution
of the difficulty. So I won't ask for books or attempt to
criticize the proofs m quest10n. I must take them in the light
of what you say about them
I certainly seem to myself to find a contradiction in the con-
ception of the proofs of which you speak. I thmk it can be
accounted for only in one way, wluch is that suggested in your
(2) of p. 12. 1 (I return your letter, as probably you have no
copy ; please let me have 1t back, whether you write further
or not ; 1t 1s my property I)
In the case you put I understand it to be admitted that :
The whole definition of S is A1B1C1, while the property to be
proved a depends on A only
If, therefore, we prove a. true of S from a.1, 1 or a fortiori from
A1B1C1, we haven't got the true a,0T1. (I remark on this that
I hold 1t is a matter of degree, and that we may have a ~10T1.)
1
§ S59 • A shp of the pen for A,.
 Co,,esf,ondence with Bernard Bosanquet 827
I understand then that II is not dependent on the difference
between A and A1, still less on B1 or C1•
But in the case put, I understand that you do prove II of S
from A1B1C1, and moreover you do this for each conic separately
and independently, i. e. you prove a. of the circle from A1B1C1,
of the ellipse from A1 B1 C., or for all I know from A1 DE, and
of other comes so on in order. And we are assured of two things :
(1) In accordance with the mitial statement about dependence
of a. on A solely, that the property really depends on the nature
of the come section, but not of the circle, ellipse, &c., as such
(N.B. words 'the universal on which the property depends',
p. 6, upper half 1).
(2) This universal nature of the come section, on which the
property depends, appears nowhere m the several definitions or
m the proofs derived from them , note sentence, upper half
of 5. 2 'The properties m question are not connected with ...
but depend upon the whole defimtion.' This must refer to the
actual procedure of the proof? If not, it surely is in fiat
contrad1ct1on with the assurance which I summarized in (I).
I take it, then, to mean that in reaching the proof the student
takes 1t to be so, viz. 'the properties m question to be not con-
nected' (though, surely, it is not so). It seems to me that a proof
of which (1) and (2) are true, cannot be accepted as a scientific
demonstration of the 6n.
§ 56r. (a) We have it totidem verlns that the conclusion is
drawn from a whole from which are inseparable elements on
which the property concluded to is not logically dependent, and
that the element on which 1t is logically dependent does not
appear m the premisses or in the proof. (Or would one say
that the property a is logically dependent on the whole of the
elements in the definition as a whole, though in some other way,
e. g causally, not dependent on them but on A only? I don't
think you mean this.) I understand, in a word, that a is taken
to be proved from properties with which 1t has really no logical
connexion.
(b) This is supported 1£ we look at the arguments all together,
when each of them appears analogous to the case of imperfect
ebmination which you suggested. You demonstrate II by, say,
1 I §56, p 823, hne 1 • § 556, para 2, • Among the properties', &c.
2773•2 Ee
 828 TENTATIVE INVESTIGATIONS
four (is 1t?} independent arguments, presenting no feature of
generic connexion, of four independent S's, using in each case
the whole defi.mtion of the S in question. It follows at once
that the nexus of a and S is not in any one of the cases reciprocal
(cf. your test of a 'correct' proof), and therefore that there is
irrelevance m the grounds of proof ; and irrelevance in a ground
of proof 1s a false affirmation of nexus, and this seems to mt ,
to be the case you put. It is a proof conducted according to
the convention which fails to demand reciprocity. The true
reason is no doubt operative somewhere within the argument ;
but there 1 1s hidden m it an irrelevancy from which we have
failed to disentangle it.
§ 562. No doubt it would commonly be held that (your p. 4)
such an argument gives you the iST1 without the &10T1. But
I do not think a logician should hold this.
As to the iSn.
If part of your ground of proof is, certainly, not logically
connected with the point to be proved, and, as 1s the gist of
the statement, you do not know which part ; I do not see that
you can be assured that the proof establishes a, even in the
current isolated sense in which you would describe a for index
purposes
And it seems to me irrefragable that the conclusion sets a in
wholly a wrong light and leaves the student m a state of
ignorance about 1t, which 1s not compatible with the establish•
ment of' it'.
As to the a,oTt,
But if you msist that this imperfect proof establishes its
existence, then I thmk you must say that it establishes a relative
fi,on. Take your description of the student's state of knowledge
after he has mastered the single proof of a about S from the
whole defimt10n of the circle and nothing more. It is obvious
that the Sn is not established for him as being what 1t really
is-only a part of the fact and not the whole is before him
It 1s a property of a conic as such, and this principal detail of
it he does not know. On the other hand, so far as it 1s established
for him at all, it is established as somehow derivable from the
nature of a circle , he does not, I gather, know exactly how,
1 MS, 1t.
 C<>tTesponflence with Btrnara Bosanquet 829
because the true nexus has never been before him at all. But
in so far as he henceforth considers himself entitled to anticipate
a; as a property of S, so far he can only anticipate it in virtue
of some connexion hidden within the nature of a circle ; and
this is so far a aufr,.
§ 563. This is really adopting your suggestion (2) of p. z2 ; 1
and there is nothing in what I have said that does not come
withm suggestions that you have made, although in the hostility
to non-reciprocal connexions I am recurring to the views of my
own Logic.
I will speak more popularly in concluding. I feel strongly
that your description of the student's state of mmd makes it
plain that the proof which you allege as dispensing with a aun-,
is a smgularly bad type of proof. Our argument to begm with
was about contmuity, I really can't think that any one would
agree that these cases make against the need of continuity for
a true logical proof I thmk any one would say, 'Why, what the
alleged proof without continuity fails to establish is the centre
of the whole fact ; without it-the nexus with the nature of
the come section-the fact is Hamlet without the Prince.'
However, I must stop
Yours truly,
B. BosANQUET.
1
§ S59

Ee2
 xv
DIVINE AND HUMAN CONSCIOUSNESS

[Oxford.]
IAug. 1904.
§ 564. 'There may be, &c.' 1 The proper reply 1s not that
there may be such things m themselves but that the statement
that there arc is unmtelhg1ble. At all events, if 'out of reference
to a thmkmg mmd ' means a thmg not bound by the laws of
our reason, as 1t often does, then if any one maintains the
existence of what 1s unthinkable he is professing to thmk about
the unthmkable, and so contradicts h1msclf. His statement 1s
merely verbal and no thought corresponds to it.
'The great presuppos1t1on' 1s neither a venture nor an act of
faith. A principle or theory can only be a venture or an act
of faith 1£ the contradictory of it 1s thmkable. The great pre-
supposition, however, 1s the necessary presupposition of any
attempt whatever to think the world ; and so 1t 1s merely
artificial to represent 1t as in any sense a venture or act of faith
as opposed to knowledge for 1t 1s the condition of knowledge
and thought, and therefore indisputable Now the md1sputable
cannot be a venture or an act of faith ; cf. Aristotle, cniic ltrr,
a' wo8fl1',s olia' af7"Y/µ,a & IUIO)'K)j Elva, ai' awo ,cal 601Cf&J/ IUl4)'IC'11 K7'A.1
The statement that the whole universe is a rationally coherent
unity 1s not 1dcnttcal with the statement that it exists only as
the obJcct of consc10usness. If the former does necessitate the
latter, the connexion ought to be shown and, 1£ 1t could be
successfully done, 1t would be an 1rrefragable argument m favour
of idealism.
Relations and things related are correlative and inseparable.
The popular faUacy is to suppose them separable, so that the
thing 1s something apart from all relations. The opposite fallacy
1 Notes for a letter in cr1t1C1sm of a phtlosoph1cal paper
1
that which must be and be thought to be because of its own nature
'
1s neither hypothestS nor postulate' An Po 76b 23
 Dimne and Human Consciousness 83x
is to mistake the inseparableness for identity, an exaggeration
not unusual in metaphysics. I am afraid Green falls into this
fallacy. We may satisfy ourselves that any argument from the
fact that relations belong to the reality of things to the con•
clusion that a thing was only relations would inevitably apply
to each man's self, and this would reduce to a 'congeries of
relations', and that no advocate of this idealism could admit.
There is in such a theory as Green's a far more serious
difficulty m the relation of the md1vidual's self and hrs con-
sciousness to the eternal self and the eternal consciousness,
which tends to be disguised by ambiguous phrases in which the
transition from one to the other is concealed. This seems to
me the weak point of the relation of morality to the ind1v1dual
self, and also of the account of the freedom of the will in Green's
philosophy
... [he now turns to the paper under cntlc1sm.] To derive
huma.i consciousness by evolution from physical elements 1s
(said to be) a i5unpo11 r.p6TEpov [a ' hysteron proteron '] because
the material elements presuppose consc10usness as something for
which they exist This 1s only ' hysteron proteron ' 1f the human
conscioui,ness to be derived from the material elements is that
consciousness which the material elements presuppose and that
for which they exist Whence one would suppose that it is the
human consciousness for which matter exists. None but a sub-
jective 1deahst believes this. Scientific knowledge presupposes
that things, objects of my consciousness, existed before they
became objects of my consciousness. Now the miseh1ef 1s that,
though this 1s not mtendcd, the logic of the 'hysteron proteron'
argument demands 1t. Thus the consciousness which is the
condition of the material world must be the human, and according
therefore to one statement made must be eternal and infinite,
and according to another must be umversal. This certainly is
the conclusion a man readmg such an argument for the first
time would draw. He would be surprised when he came to the
protest: 'I never said my consciousness made the world.' He
would feel rather taken in and might reply : 'No, mdeed you
didn't, but everything you dul say naturally bore this inter-
pretation and, if it didn't, you should have said so at once.'
Clearly, if a d1stmction is to be made between human conscious•
832 TENTATIVE INVESTIGATIONS
ness and the eternal universal consciousness, it ought to be most
plainly shown that the argument about human consciousness
can be grounded on that other consciousness The fact that
this 1s not done seems to me to show that . . the difficulty is
really concealed by the identity of the word consciousness in
both cases That 1s why the thing seems to me intrinsically
unsound and to profit by an ambiguity. How could we refute
a man who chose to answer in this way? 'I grant that the
material clements exist only m and for an eternal consciousness.
But your protest means they do not exist for and in human
consc10usness solely, and that they exist before a given human
consc10usness. Very well, the material evolution of human con-
sciom,ness would now mean that tlus consc10usness is developed
out of clements which at all events existed before it and do not
exist merely as obJccts for it, m it, and relative to it. Moreover,
the bram and other organs, wluch I maintain are necessary to
human consciousness, never are obJects of the human conscious-
ness which they condition. Thus my theory of the evolution
of human consciousness 1s m no wlut altered. Relative to human
consciousness these material clements rcmam as material as
anything need be. They are eternally existent matter and it
makes no ddferencc 1f they are said to have their eternal existence
m an eternal consciousness. It 1s hardly more than a verbal
change since the eternal consciousness 1s not human conscious-
ness And, m fact, 1s not the eternal consciousness so completely
other than a human consciousness that it makes precious little
difference what we do call 1t? And at least there is no
' hysteron proteron ' These difficulties are rather concealed
than answered by such representations as that ' consciousness
awakens m me', \\ here the consciousness 1s apparently the
umversal eternal consc10usness which has been pronounced like
my consciousnesc; Tlus sort of identification 1s of course
important to help out the' hysteron proteron' argument. And yet
for some purposes, as the protest (' I never said my consciousness')
shows, it is absolutely necessary not to identify the two con-
sciousnesses But consider the above phrase, 'consc10usness
awakens in me'. 1£ we here use consciousness as a universal in
the ordinary sense of a universal, this (1s) quite simple ;nd
intelbg1ble, and only means that my consciousness is a particular
 Divine and Human COffsciousness 833
case of consciousness in general-such a general consciousness
only exists in each particular consciousness.
But this is not the kind of universal consciousness in which
material objects exist. For, if they are in consciousness at all,
it clearly must be in an existent consc10usness. If such a really
existent consciousness is meant by the eternal universal con-
sc10usness, such could not 'awake in me' Actual consciousness
belongs to an actual subject of consciousness, must be in it, be
its consciousness, and cannot be the consciousness of any other
subject whatever Therefore the existent eternal universal con•
sciousncss, as belonging to the eternal sub3ect, cannot be in me
or 'awake' m me or m any other person ; that would be possible
only if (I, as) sub1ect, were identical with the eternal subject.
Indeed the consciousness of no subject can be identical with
that of another. My learnmg and apprehending a theorem
cannot be the same as your learning and apprehending it.
(You yourself say) that the consc10usness which does make
the world is like my consciousncss Thus it can't be mere con-
sciousness in general, but particular and real as mme is, and
have a definite subject for it, as mine has. One might mdeed
say shortly · a consciousness like mine cannot possibly be
identical with mme, as it certainly must be 1f 1t awakes m me.
By me is meant an actual ind1v1dual subject with an actual
ind1v1dual consciousness. If this 1s thought of as a subject
existent but not yet conscious, it is impossible to represent the
consciousness of another already conscious subJect as transferred
to this subject
There tends to be an equ1vocat10n between the subject of
consciousness and consc10usness itself Consciousness is said to
'awake', which makes it not independent of time. In the same
passage 1t is made independent of time as something present to
both past and present. Now 1t 1s the subJecl which is present
to the elements of the temporal series and so independent of
time, but the subject's consciousness is an activity the phases
of which take place in time. In relation to this independence
of time there is a great difficulty for any theory. Each indi-
vidual's life has a temporal beginning and we think of the
subject himself as having a definite beginning in time. The
idealism before us is specially bound to meet the difficulty
834 TENTATIVE INVESTIGATIONS
because of what 1t says of the eternal consciousness and the
tendency in it to seem to equivocate between that and the
human consciousness.
The organic theory of society does not depend on this idealism
and can be stated w1thouL 1t. It depends solely on the meta-
physic which rccogmzes that no individual has a reality com-
plete, separate, and apart from that of others, and that the only
complete reality is the complex of individuals, &c.
Morahty (1s) not properly affected by any theory of how
consc10usncss originates, material(1st1c) or otherwise. The
intrinsic value of the moral consciousness is mdependent of any
theory of how consciousness m general arises, as the value of
a geometrical judgement is independent of any theory of the
brain processes which make thought possible. On the other
hand, the derivation of moral elements from mental elements
which contain no morality 1s dangerous to moral ideas and
affects tl1em. This should therefore be refuted.
 XVI
RATIONAL GROUNDS OF BELIEF IN GOD
§ 565. THERE are 1 subjects which do not appear at first in
scientific form, art, morality, relig10n. In approaching such
subjects for the first time 1t 1s natural to begm by trying to
make proofs, or by cnt1c1zmg given proofs, of moral, religious,
or aesthetic principles, without any preliminary consideration
of the meanmg of proof or of the possible hm1tat10n of its
province. But such a process may be endless. We may con-
struct and reJect proof after proof and yet make no real progress,
because we may have presupposed, without d1stmctly recogmzmg
1t, that proof must have a certain form, a form which 1t happens
cannot fit the matter to which we are applymg it. The defect
therefore will not be in the particular proofs as such, but m the
form of proof in general, and till we recogmzc th1s we shall
labour m vain.
Or we may have somethmg in our mmds which would, whether
right or not, prevent our acceptmg any proof of whatever form.
Or 1t may and does happen that bemg already convinced of
the conclusion, moral or aesthetic, we admit a faulty proof of
it: our interest m the conclusion blmdmg us to the flaw. This
happens even m purely speculative proofs, e. g. even m mathe-
matics.
Or a proof may seem correct and yet it does not quite satisfy
us because 1t does not touch our feelmgs. Perhaps we are not
aware, in our search for scientific completeness of demonstration,
that this is the fact and, if we are at all (aware), may be afraid to
own it, because 1t seems unscientific. But the demand of our
nature remains, and we shall not be satisfied unless we raise the
question of the relations of feelings or emotion to truth.
§ 566. The attempt to vmd1cate our religious beliefs and, it
Rough notes for an address read at Canon W Sanday's Discussion Society
1

Chnst Church JJl Michaelmas Term, 1897, subsequently slightly revlSed

JJl
and expanded. The title announced was • The Ontological Proof for God's
Ex11tence •.
 TENTATIVE INVESTIGATIONS
may be, our moral and aesthetic decisions is often conceived as
the attempt to find a 'rational basts' for them.
Now the inquiry after a rational basts for a belief has in its
form certain important (1f unconscious) presuppositions. We
do not seek for something for a thing which it has got already, if
we know 1t : and so also to undertake such a search we must
be able to thmk of the thing, and must have thought of it, as
existing without what we seek for it.
Thu:;; it S(,f'J'1S as if our question to have any meamng-unless
it 1s to be entirely idle words-must imply that the belief exists
somehow already, without the so-called 'rational' basts, and
that the scarrh for such a basis is an after-thought or later
development. This 1s a serious implication It might be con•
sidercd a concession which imperilled the worth of the belief,
and so perhaps be pressed by one side and reJected by the other.
On the other hand, 1t might be thought (to be) m the interest
of the dignity of the rehgiouo; consc10usness to mamtain 1t : to
insist indeed that the belief has a lughcr basis than reason, or
what 1s usually meant by 'rational', though (1t would be said)
it cannot disagree with reason, which must confirm it.
We do not ask for a 'rat10nal basis' for a proposition m the
exatt scICnces-e g for the Pythagorean d1o;covery of the rela-
tion between the sides of a right-angled triangle-because such
a propos1tton has no meanmg for us except as <restmg) on a
iat10nal basis. Even when a theorem m the sciences 1s uncertam,
e. g. a guess about some cause, 1t 1s not natural to put our
inquiry mto its validity m the form of askmg for a 'rational
basis'. On the contrary, so far as 1t exists al all, 1t has some
sort of rational basis , 1t has been formed somehow on evidence,
though the cv1den(,c is <recogmzcd to be) msufficient and the proof
but probable. In the case of facts, agam, m nature which we
know m some way, but of which we do not know the explana•
tion, when we ask about them, it would be quite unnatural to
put our question m the form of asking for a rational basis ; we
assume as self-evident and without reflection that they have
such (basis) and are members of a rational system.
When we do come to speak in this way, 1t 1s because some
explanation of the fact has been given which would take it out
of a rational system
 Rational Grounds of Belief in God 837
For instance, we do not ask for a •rational' explanation of
the cure of diseases by ordinary medicines or surgery, we do
not seek to 'rationalize' these. But m the case of faith-healing
(in modern times), 1f scientific people suggest as a true explana-
tion of them that many diseases are really due to hysteria, and
are in a certain sense apparent rather than real, depend therefore
on certain states of the patient's mmd, which again can be
mfluenced and changed by the will : this may well be called
rationalizing or findmg a rational basis for the phenomena. And
this would be done m consc10us oppos1tlon to the 'miraculous'
explanation.
It is clear then that (when) we come to think about the
religious consciousness and its beliefs, we have some different
attitude from that which we have towards nature and m the
sciences, m respect of rationality. But it is not here alone, for
m tlus particular the moral consciousness and, to some
extent, the aesthetic or artistic consciousness are m the same
position.
From one point of view we unhesitatingly speak of all these
as rational and even par excellence rational We thmk that 1t
is only a rational (thinking) consciousness which can be moral,
and, more than that, we tlnnk that we show our rationality
especially m morality. So also we think that it is only a rational
consciousness which can be art1st1c. But m all of them we
discern a certam distmction from reason or rationahty and even
a certam opposit10n to it. For instance, 1t 1s said that Greek
philosophers, especially Socrates and Plato, endeavoured to put
morality on a rational basis This implies that morality with
its conceptions, judgements, rules, and practices, had been
existing independently of such basis.
We must not make this exactly hke the relation of our
ordinary experience of Nature to the development of the same
principle in scientific thmkmg. For instance, the principles of
counting and thinking about distances in space, methodically
carried out, become the mathematical sciences. But we don't
look at them as the putting of our old experience on a rational
basis. It is rather that we know more of the same kind of thing
which we knew before. We do not regard the preceding experi-
ence as somehow unreasoning and requiring to be just.med (if
 TENTATIVE INVESTIGATIONS
it can stand) by a rational basis ; on the contrary, we look upon
Jt as giving the true basis to science itself.
§ 567. The difference and even opposition shows itself in the
study of the sub1ect when the rat10nalizmg attempt has begun.
Philosophers assume readily enough that morality is rational
just as the ordmary consciousness does when it comes to reflect
on the sub1ect. Yet when they come to try to fix the nature
of the reason or rationality, to find the place of reason and
rational thmldng m morality, the attempt has proved to be
extremely difficult and there has been no sure unhindered pro-
gress from experience to science as (there has been) m geometry
and the natural sciences.
Am,totlc struggles vainly with the problem and ends in putting
the moral hfe actually below the speculative on the ground of
a cert..in irrational element in it i a difficulty not to be glossed
over by calbng this element imperfectly rational rather than
irrational
The Jiving actual moral consciousness, however, protests
against such a theoretical posit10n as absurd. We feel indeed
Lhat what is noble and d1gmfied and most worth having m hfe
1s, on the contrary, above the mere scientific and speculative
reason ; and so (in modern times) when this consciousness has
become more fully aware of its own value we find a ddierent
kind of reaction agamst, or resistance to, 'reason', yet still some
reaction.
Thus, in modern times, in our English philosophy, after the
reason m morality had been represented from the point of view
of the knowledge m morality, with great emphasis, as if (1t were)
exactly hke the scientific and speculative reason and, like that,
producing a system of eternal truths, there 1s a reaction of which
the characteristic is a certam repugnance to puttmg everything
m morality under the reason. So much so that the very faculty
which supplies moral judgements, 1 e. moral truth, is not called
Reason but Moral sense, with a deliberate and expressed dis-
tinction from Reason, 1 and assimilated to a corresponding
faculty m the aesthetic consciousness. And this faculty is con-
sidered more valuable to us than our reason, giving us directly
1
Hutcheson, An Inquiry, Preface, pp. xiv-xv, • The weakness of our
Reason, &c •.
 Rational G,ounrls of Belief in Gotl 839
what Reason apparently, if perfect, could discover (but only as
the result of a long train of argumentation}, while such reason
as we have, from 1ts weakness, will not make the discovery. 1
Besides Reason,2 as a principle of action, is apparently conceived
as only calculating and contriving means to promote our private
interest, while the Moral sense is directed to the good of others.
Whatever criticism might be passed on the passages which
contain such statements, whatever difficulties or inconsistencies
they may involve their author in, the important thing for the
present purpose is that the faculty supposed to give us moral
judgements is not only distinguished from reason but preferred
to it. The same thing holds of the popular associations of the
word conscience and finds expression in the philosopher of
conscience, Butler.
Without professing to identify Butler's theory with the Moral
sense (theory), it is enough to notice that, for our purpose and the
pomt of view from which we are regarding moral systems, the con-
science theory and the moral sense theory are in essential agree-
ment ; there 1s the same reluctance to make the faculty of moral
principles simply reason ; the same representation of the moral
faculty, i. e. that which gives us moral judgements, as giving
us directly what our reason, 1£ 1t could give it at all, could only
give as the result of a laborious process : with the further
belief that our reason cannot give them and that by conscience
we feel the force of certain obligations which our reason, con-
ceived as trying to argue from ideas of the good of the world
or the happmess of mankmd, is not able to recog~ze. There
is the same tendency to make Conscience the higher faculty, it
is the ' voice of God in us '. 8
Another school of thought affirms deliberately that reason
can only calculate means for ends given to it, thus admitting
that the most important thing in moral conduct, the end and
motive, is not an affair of reason at all.
About the aesthetic consciousness one or two things seem
1 Cf Butler especially, (Sfffll v, xu, and xv).
1 Hutcheson, An Inquiry, u, § 1, compared with ii, Introduction
• See Butler, Sermon v, p 49 ' Reason alone, whatever any one may
WlSh, IS not 1n reality a sufficient motive of vll'tue 1n such a creature as
man• (Oxford, 18g6, vol. 11, p 98) Cf last note to Swmon xu. (Id.,
p. zz6, note c J
 TENTATIVE INVESTIGATIONS
clear. No artist, and probably no theorist, is likely to admit
that mere thinking or reasoning without feeling or emotion
could give an aesthetic judgement. There is no science, in the
proper sense, of the fine arts A body of rules like the theory
of Music is no exception , there is reasoning there, but such as
would be admitted m a moral sense or conscience theory of
Ethics, the application, that is, of given rules ; but the rules
themselves are the important thmg.
It is worthy of note that a writer like Hutcheson while
speaking sometimes at least in a manner which, if the words
were pressed, might imply that in the case of the moral con•
sciousness an absolutely perfect reason could give the moral
rules, in the case of the aesthetic consciousness refers its judge•
ment to Taste and Feeling or Sense, and never suggests that
here a p<'rfect reason might reach the same results ; the idea
never seems even to occur to him Further, m his later treatise,
Illustrations upon the Moral Sense, he mamtains outright that
Reason does not give us moral judgements, that 1t cannot
originate the moral ends, or any cnd,1 m such a way that he
seems to deny the poss1b1hty for any kmd of reason, not merely
for ours, (this interpretation is) confirmed by this that later,
speaking of the Divmc Consciousness, he says 'Why may not
the Deity have sometlung of a superior Kmd, analogous to our
moral Sense (, essential to him)? '
§ 568. In the foregoing remarks two objects have been in view.
The first is to remove an unfair prejudice which represents the
religious consciousness as 1£ m a unique position of antagonism
to reason. The charge agamst 1t might be put in the most
unfavourable way thus: 'Our reason has a right to call every
behef in question, to ask for and criticize the evidence for them.
Religious beliefs will not really stand this test To conceal the
weakness Theology tries to avoid the attacks of reason by
taking refuge behind mysteries. It assumes a mysterious source
of religious truth, Revelation, and a mysterious faculty
• Sect 1, ' " does not every Spectator approve the Pursuit of pubhck good
more than private ? " The Answer 1s obvious, that he does but not for any
Reason or T,uth, but from a mo,,al Sense • • 'Does a Conformity to any Tndh
make us approt•e an ultimate Enri, previously to any moral Sense?• (and
Hutcheson maintains that 1t does not) [Hutcheson, Illustrations upon the
Moral Sense. See Selby B1gge, British Moralists, vol 1, §§ 4S4, 4S9)
 Rational Grounds of Belief in God 84z
of religious truth, Faith ; and, the more effectually to put
reason out of court, it insists that these are above reason,
superior to it and with an authority beyond the canons of
reason.'
I have been trying to show that if we look at the thing fairly,
we find a very parallel tendency in the moral and aesthetic
consciousness. Shortly one may say that in the case of the
moral consciousness also we find a faculty assumed to give
truths, which is distinguished from reason and considered as
even superior to reason. Nor can we say this is a merely popular
and unscientific view, the deliverance of the untrained moral
consciousness On the contrary, we find it in the very philosophic
treatment of these subJects itself . undertaken by those who
may fairly claim to try to take a thoroughly philosophic view.
Nor can it be said that such doctrines belong only to unimportant
people not really able thmkers. It 1s true that philosophy suffers
from the fact that it is m a sense every man's subject, and one
has to read some queer stuff under that title, but the thinkers
m question belong to the strongest and ablest who have written
on morals in this or any other country.
The students of philosophy therefore can hardly presume to
take the superior attitude towards theology implied in the charge
above formulated. We must rather first see that we set our
own house in order.
If it be replied that, after all, these faculties of Conscience and
Moral sense are aspects of Reason and misunderstood by their
advocates and that the kind of reason to which they were
opposed was really a kind irrelevant to moral principles, then
prima facie the same kind of excuse may be made for the
religious consciousness
The second object is to show that the opposition implied in
the phrase •a rational basis for belief m God' was no mere verbal
accident, but corresponded to a real opposition, which has made
itself felt in other departments of consciousness besides the
religious, and that, too, an opposition so far from being destroyed
or eliminated by a scientific investigation undertaken by reason
in the interests of pure knowledge, that it actually sometimes
appears as the result of such an investigation.
§ 569. To return now to the opposition itself.
 TENTATIVE INVESTIGATIONS
It seems that where we can really ask for a rational basis for
a thing and the language is really appropriate, we assume that
the thing has already a basis which somehow is not rational,
or is thought to be something not rational.
Then comes a twofold d1v1s1on.
We may suppose that in truth the thing has only a rational
basis, that the other basis given to it 1s an illusion, and (that)
in seeking for the rational basis we are trying to remove this
illusion. The Greeks sought, for example, a rational basis for
the phenomenon of thunder and hghtmng, which had been
assigned to Jupiter, and supposed they found it m the
quenching of fire, and this does away with the non-rational
basis
If the thing which we treat m this way is a belief itself, we
essentially alter the character of it, we are explaining the belief
to be an illusion and ass1gmng the basis of fact from which it
sprang. Thus a rationalistic explanation of the belief in ghosts
has been offered that men in a pnm1t1ve state of half-developed
consciousness sometimes confused appearances m dreams with
real visions. It must be allowed, however, that in this case we
use the phrase 'rational basis' for the phenomenon or fact which
1s misunderstood rather than for the belief.
The second case would be that m which the thing 1s supposed
to have a basis already ; but then 1t becomes a serious question
whether it could have another basis, viz. the rational, at all ;
especially as we observe that where in moral theory a belief is
supposed to have a basis which is not rational m some sense,
this view 1s held on grounds which would make it impossible
to have a rational one.
And anyhow, our fir&t business must be to examine into the
nature of the basis which the thing or behef 1s supposed to ha.ve
already and before our mvestigation · for that it can have such
a basis must be essential to 1t and, therefore, essential to our
understanding of it In fact the possibdity of a 'non-rational'
basis will be exactly characteristic of such a belief as compared
with those which have a rational basis only ; and it 1s obviously
to the distinguishing characteristic that scientific or philosophic
investigation must be directed.
§ 570. What would be the state of mind of one who was really
 Raonal Gf'outuls of Belief in God 843
unbiased and looked on it really as an open question, a thing
to be proved or disproved ?
At all events he will not have any immediate knowledge or
conviction, such as his belief that he is warm or cold or uncom-
fortable, or the question would not be an open one to him. He
will regard the opinion so far as it exists as based on some sort
of evidence, or on some desire, or (on) both together; perhaps
in some such way as this :
In the earlier stage of the attempt to 'explain' natural
phenomena, men naturally prefer what seems most intelligible
to themselves and, in a first crude reflection, they think they
understand best their own actions ; or, rather, this is so natural
to them that they don't ever reflect on it, and so their explana•
t1on of natural phenomena naturally takes the form of attributing
them to a personal agent hke themselves.
lnclmation as well as reflection contributed to this result,
because m any case men would prefer the superior force in
nature to be like themselves and especially because, if it were
so, they would have a better hope of a comfortable life, and
that for two reasons · Such an agent as being a person might
be sympathetic with them and mclined to help them, or, 1f not
such but selfish or even malevolent (as they knew persons could
be), yet at least accessible and amenable to influence from them
1f they could contrive to please him, and so better than mere
unconscious nature.
Later, even when scientific knowledge has advanced very far
and mere superstition (become) recognized as such, the same
influences both m thought and m feeling will operate still, though
in a higher form When the ordinary phenomena of nature have
been reduced, with more or less success, to mechanical or non-
personal causes, the laws of this mechanical nexus will appear
as ultimate facts. We do not understand how these ultimate
facts came to be there, and the mmd tends always to be dis-
satisfied, whether rightly or not, with the recognition of a mere
fact. It will tend therefore to be dissatisfied with mechanical
causes.
There will also be a tendency to assign to the highest cause
the highest quality ; and the more our intelligence grows, the
more the mind becomes aware that reason and consciousness
a773•a Ff
 TENTATIVE INVESTIGATIONS
have a paramount worth as compared with the unconscious :
and so we have a bias towards representing the highest cause
as spiritual (especially as it 1s to be the cause of our own
existence as well}.
With increasing knowledge men begm to reflect on the adapta•
tion of their own organism to obtam certain ends, which are
desirable to them, and the more they know of organic life m
general, the more marvellous appears to them the elaborate
apparatus by which the ends of orgamc hfe are attamed ; and
even to a very superficial knowledge this is seen to be far beyond
what men could have thought out for themselves. Such adapta•
tion of means to ends is most easily understood, by the analogy
of human action, as the work of a self-conscious spirit conce1vmg
ends and designing and plannmg accordmgly, just as m the
more primitive stage the so-called activities of nature are most
easily understood on the analogy of human action.
But this time the argument has come to stay, for not only
is this the easier explanation but it 1s difficult to beheve in any
other and the difficulty seems to mcrea,;e and not to diminish
with increasmg knowledge We have the celebrated argument
from Design which has mfluenced every per10d of thought.
Once conceive the highest Cause as spmtual and it 1s natural
to consider him as the perfection or ideal of Spirit, and hence
perfect knowledge and goodness are assigned to him. He must
then Le conceived as author of the moral as well as of the
natural order and thus the belief m God has now enlisted on
its side all the mterest of the moral consciousness.
This might be taken as a sketch of the prmc1pal arguments
and mfluences which could and have brought about such a belief
in mankind m general apart from a special revelation, whether
a fact or not. There are special philosophical and metaphysical
arguments as well, but the important fact for us is that the
behef is held by the masses of mankmd, and 1t 1s to this that
we have above all to direct our attention: and the mass of
mankind 1s not metaphysical.
The unbiased Judge, whom we are imaginmg, 1s bound to
put the case strongly against the belief, to be very critical about
the evidence, and we may suppose he would treat it in some
such way as this.
 Rational Gt-ountls of BeUef in God &ts
Two different forces contribute to produce the belief. First,
there is a rational element in the form of reasoning from evidence,
the evidence being from acknowledged fact. Secondly, there is
an element which some at least would, when viewed as the
origin of a belief, call an irrational element ; and, if not so to
be called, it is not reasoning from evidence at any rate, it is
our desire, and this appears as much in the later stage of human
culture as in the earlier. We are dissatisfied with mechanism
as an ultimate fact because we want the spiritual and conscious
to be supreme. The interest of the moral consciousness biases
us also, for we are too glad to give 1t the highest possible
sanction. As to the 'rational part' :
It has not the completeness of demonstration. There is some
probab1hty, perhaps much, in the argument that the existence
of spirits depends on a spiritual cause. The argument from
design is also questioned
(1) as not more than a high probability; (2) the existence
of evil and imperfection m general 1s so much evidence
the other way ; (3) m modern times the theory of
development and evolution explains the apparent
adaptation of means to the ends of orga01c life, without
any need of a designing conscious inte]bgence : at least
some scientific men hold this view, and 1t 1s the popular
conception of evolution.
(The impartial Judge) might ask therefore whether the great hold
the belief has on men's mmds is not really due to the strong
interest they have in its bemg true ; Wish and Desire, not
Reasoning in the sense of reasoning from evidence. He might
add that strong desire and hope that a. thmg should exist,
however good 1t might be in 1ts intent10n, was certainly no
guarantee.
§ 571. One may venture to think there may be something
beyond and more satisfactory than the defence of or attack
upon these arguments, but I want to say a word, before going
further, on the argument from Design.
I suspect we don't do it justice, that it has more hold on
the mind of those who reject 1t than they are willing to confess
or are perhaps aware of. Probably, when we are first introduced
to it, it impresses us very powerfully. When we are more
FfZ
 TENTATIVE JNVESTIGATIONS
educated and get some idea of science and exact knowledge, it
doesn't satisfy the standard of such demonstration as we are
familiar with. Besides 1t has been subjected to severe criticism
from the scientific (point of) view, and so we tend to be afraid of
it and regard 1t rather as something natural to immature thinking.
Perhaps our d1fficult1t. s are summed up m this, that the very
effort to know Nature sc1enttfically presupposes the necessity of
everything m nature, and this seems, at first at least, the con-
tradictory of design. But it makes aU the difference whether
we thmk m this general abstract way about the argument from
design, or whether we are considermg some of the actual facts
on wh1c-h the argument itself 1s founded. We may imagine we
have got rid of it as a prejudice or 1llus1on ; but 1f, forgetting
our general theories, we are absorbed m the study of the adapta•
tion and harmony of the various parts of a plant or animal to
the maintenance of its hfe and funct10n, 1s there any one so
adamantme that he can maintain an unemotional contemplative
or 'scientific' attitude ~ Can he help feeling admiration and
wonder ) Whether 1t be an 1llus10n or not, the idea of plan and
design and c1101c<' of means comes on us with irresistible force ;
we cannot shut 1t out.
This impression is strengthened and not weakened by the new
knowledge gamed by modern biology, the gradual development
of an orgamsm m successive generations We see the gradual
'perfecting' of the organism for its functions ; it isn't now
as though we saw the marvellous adaptation m a given orgamsm
and inferred, or could not help thinking of, a designer , 1t is as
1f we saw the design or plan m process of being worked out.
But tlus 1s not merely the impression of the outsider. This
idea of design 1s actually the clue by wh1ch sc1ent1fic men are
constantly working, not only 10 dcscnbmg the parts of an
animal, but especially the evolution of the organs ; for the
meaning of the elements in the lower stages 1s found by them
in the higher. Sometimes this comes home to them and they
admit its s1gmficance 1
Secondly, (as to) the objection derived from evil and imper-
fect10n.
First notice that the impression above described is not one
• See Aubrey Moore, Setenre and 11111 Fadll.
 RatiDflal Grounds of Belief in God 847
which we have before the thought of evil and imperfection
occurred to us ; it comes over us strongly although we are aware
of these things. They are not evidence against design in general,
The want of adaptation and fitness in one department does not
affect the evidence for it in another, so far as it is evidence, but
leaves it untouched. The inference we should rather draw if
we confined ourselves to the evidence is one to which our modern
habit of thought dismclines us. Helmholtz • m his account of
the eye takes note of certain optical defects in the crystalline
portion, and adds significantly that if such an instrument had
been sent to us by an optician we should return it with the
request that such defects should be remedied One understands
the innuendo. It 1s directed agamst the theory of design. But,
if so, the inference was not the one which an accurate scientific
man should draw. Hts own description of the structure of the
eye gives an instance of that wonderful adaptation, m choice
of material and form of organism, which eventually makes us
think of a des1gnmg intelligence. If we are to argue strictly
from such evidence alone with nothing else to go upon, we should
infer the plan itself, as much as ever and as clearly and forcibly
as ever, but we should infer something which hindered its
execution, destroyed 1t altogether, or made 1t imperfect. But
we have a scruple in drawing such a conclusion, because we
think of the designer as also God, and mean an all-wise and
all-powerful Spirit who could not make a mistake m plan or be
hampered by the necessity m Hts material in carrying out Hts
plan. But 1£ it is mere design we are thmking of, the natural
tendency would be to a duahsm, and the tendency is realized
in ancient authors hke Plato and Aristotle.
(Note on the argument from Design. b In the foregoing I have
not been concerned to offer an opimon on the value of the
argument from Design because I wanted to keep clear what
I have represented as the mam issue. But I may here add
a development of what has been suggested. If we put aside
such scruples as I have referred to, the evidence is in favour
[" Helmholtz, Pop14lar Lectures on Sc1etd1fic S1'b7ects (Eng translation),
London (Longmans), 1873, Lecture VI, 1, • The Eye as an Optical Instru-
ment' Wilson's cntic1Sm 1s hasty, see especi.ally p 228
b Added after rece1vmg a copy from Professor C C J Webb of his book,
Problem&, 6-c., in 19u. See f 582 J
 848 TENTATIVE INVESTIGATIONS
of supposing that there is an agency m Nature working with
design, but neither omniscient nor omnipotent, and learning
gradually how to fashion to its ends a material which it has
not created.
But such agency would no more need to be identified with
God than human agency needs to be, even from the theologian's
point of view. The theologian has to admit the existence of
human beings, imperfect m will and knowledge and power. -He
admits 1t 1s God's will that such beings should exist and that
they should endeavour to work out their own progress both in
their spmtual hfe and m fashioning the material world to their
comfort ,md good. If so, there can be no difficulty in assuming a
very similar agency m Nature which God also has willed to exist.)
As to the objection derived from Evolution.
I doubt whether Evolution as such has added to the objections
aga.mst Design ; on the contrary, as I have mdicated, if the
facts as imperfectly known suggested it, as more perfectly known,
they suggest it more strongly. But 1t tends to suggest the
picture of a designing mmd workmg with methods curiously
like (those of) the human mmd It is as if nature made experi-
ments and reJected the less satisfactory, retained the better and
even advanced m knowledge and skill. (Many instances, e. g.
Chrysalis. The outer casing of the chrysalis at one period
shows the outlines of hmbs and a definite organism, but the
interior has nothmg to correspond. The case 1s :filled merely
with a fluid substance apparently homogeneous The inference
is that a stage between the caterpillar and butterfly has been
omitted)
But the idea of necessity 1s inherent m Evolution as m all
science; the material must be supposed to go through its
changes according to fixed laws. What difference does this
make? Assuming evidence of plan at all, the difference it would
make would be to represent the plan and the gradual working
out of it as mhcrmg m the necessity of nature itself and con-
stituting its essence. If we were told that it is a law of all
conscious beings to desire to mamtain their life, a law m their
nature to fear death, we should not think of this as destroying
our planning and designing activity in endeavouring to compass
such ends.
 Rational Grounds of Belief in Gorl 849
§ 572. I have tried to imagine what might be an impartial
and unbiased view. But there are many who are not in this
condition ; what is their attitude to the problem ?
Some firmly believe in a divme Spirit and in such a way that
the success and failure of what I may call a mere external
argument would not really affect the1r belief, if they would only
come to terms with themselves and recognize that this is so.
We cannot in a ;c1entifie or philosophic interest simply put such
out of consideration. They are not men whom we think other-
wise irrational, they are to be found among the very wisest.
With this kind of belief often goes a digmty and nobleness of
character, and, m other spheres of thought, a finer perception
of what we admit to be higher , and so there is at least a prima
facie presumption that their judgement is not wrong in a sphere
for which the highest qualities seem necessary. Now 1f these
(thinkers) engage in a mere speculative argument, it is with little
profit either to themselves or others, because they are not really
talking of the conception of God as it is for themselves ; of the
living reality of 1t m their own minds In its true reality for
them it cannot be seriously considered as a matter of argument.
What they ought to do 1s to try to come to as clear a conscious-
ness as possible of the nature of their conviction · that 1s the
important thmg and vital to science and philosophy ; this 1s
what they ought to do 1f they would help m a discuss10n like
this. Perhaps many are deterred from this by a fear, which
they have not fully confessed to themselves, that they may
appear irrational. Well, they know well enough that they are
not otherwise accounted 1rrat1onal, perhaps indeed the very
opposite. They know there 1s nothing petty or contemptible in
the belief, but that 1t is confessedly associated at least with what
is highest m human nature, and they should not fear to look
into their own souls here.
Now some of those who seriously engage m argument are in
this case ; they are not aware of their true position and that it
1s their conviction which produces the argument and not the
argument their conviction. (One observes that sometimes
practical theologians will gladly take up some supposed estab•
lishment of theism by a philosoph1cal argument. But where do
they suppose they themselves were, before they heard of the
850 TENTATIVE INVESTIGATIONS
theory, and what was the meaning of their lives? They possessed
something far stronger to them than any such theory and, if
they should find a flaw in the philosophical argument after all,
they wouldn't admit they now had no ground for the belief.
It is this, whatever it may be, which precedes all such mere
theory, which 1s the valuable thing for them and has been the
movmg prmc1ple of their hves. This 1s the thing for them and
for us to try to understand Till this 1s done everything else
1s lost labour.)• One sees this sometimes in the greatest phdo•
sophers. The arguments by which they seek to recommend some
of their most important prmc1ples to others are clearly not the
process by which they arrived at them themselves, for they are
sometimes not only unsatisfactory but fallac1ous, and 1f they
had not been so convinced of the truth of their conclusion they
would have seen that themselves. (This 1s true where it is
perhaps little suspected m theories which concern our conviction
that we are perceiving 'real objects'. It is a fact demonstrable
by the exammat1on of theories of Percept10n, even the latest
and supposedly most advanced, that there 1s no such thing as
a theory of perception, m the sense of a theory able to account
by argument for the conv1ct1on m question-an absolute con•
v1ct1on as 1s shown m the most convmcmg way by our actions.
None of the theories extant from Aristotle till now succeed in
the least m accounting for the conv1ct10n, and they do not
merely fail in proving that we do perceive external objects, but
they entirely fall to prove how we get this (firmest} conviction
that we do perceive them, or even how we could get the idea
of such objects at all. Some of them have actually resulted JD
explammg away the fact that they can't explam and reducmg
perception to an 1llw,1on. A certam kmd of psychology of
perception must indeed end JD subjective idealism. The mterest
and sigmficance of this for the subject before us will be evident
m the sequel. We have m our ordinary perception a conviction
antecedent to everythmg offered m the way of proof.)b
Thus Plato and especially Aristotle try to present us with
proofs that the rational hfe is the highest and best : mistaking
the nature of proof and not perceiving that what they would
prove is the absolute presupposition of any argumentation what-
[• Much later addition by author. b Much later addition by author.]
 RR.tiontll Grounlls of Belief in Gotl 851
ever about the good. The argument as presented in the Ethics
should take no modern student in. One's first thought when
trying to prove anything about God or morality should be-did
I really get this conviction myself in this way? Then and then
alone do we find the meanmg and value of our own thoughts.
§ 573. I have ventured to suggest above 1 that there might
be something better than the attack or defence of the arguments
I referred to. It might be mferred that I was prepared at once
to suggest some other arguments, perhaps metaphysical ones.
But I may repeat that I am concerned with the arguments and
considerations that may be supposed to have affected the masses
of mankind ; for here must Ire the secret, and it is here we must
look for the inmost reality of the conception or belief. The true
business of philosophy seems to be to bring the behef to a con-
sciousness of itself.
It is no use producmg philosophic arguments external to this ;
we should never satisfy ourselves, for we should shirk the thing
absolutely necessary for understanding and 1t would return
again upon us as the problem That would be to commit the
error so often repeated m so-called ' Moral Philosophy ' ; to
suppose that its function 1s to find a criterion of goodness, as
a guide to moral action. If this 1s taken seriously, men couldn't
have known what was good before the cntenon had been dis-
covered i they couldn't have hved moral hves. Besides, there
is the further difficulty that the 'philosophers' have disagreed
about the cntenon.
When we do get a phdosoph1c account, then, we must show
that it harmonizes with the workings of the unphilosophic
consciousness.
§ 574. About the popular arguments (which I don't believe
to be a true expression of the popular consciousness) ,-we see
something can be said for and against, which 1s enough to show
they are not cogent. A simple proof m geometry is cogent and
it can't be a matter of debate at all m the proper sense of the
term.
Now I want to propose two questions connected with the
preliminary considerations about proof with which I began.
Could any proof of such a form be perfect and cogent?
1
§ S?I snd.
 TENTATIVE INVESTIGATIONS
If we had a perfect speculative proof, should we really be
satisfied (remembermg what the object of the proof 1s), and
would there be nothmg wanting ?
§ 575. First, on the form of proof. The form is 'empirical',
starting from facts supposed to have the best evidence, viz.
direct experience. The special form 1s 'analogical' It is mani·
fest that m such (a proof) we can only have probab1hty and not
necessity To 'empirical' we oppose a proof in which we -see
the absolute necessity both of the conceptions and (of) the chain
of argument mto which they enter.
It may be asked' why, m this account of the popular reasoning,
(shoultl we) assume it hasn't argued m the necessary and non·
empmcal form?'
Perhaps the truth 1s that such a question will not be seriously
asked. If there had been reasonmg of that kind we should
hardly have had such stress laid on the 'design argument' 1
considering the difficulties with which that argument 1s encum•
bered. We should certainly have heard much more of the other;
and who has heard of it at all? For the moment I will only say
that it has not been consciously present to the popular reasomng.
As to the other form of empmcal proof, the matured form
of 1t in empmcal science ; this moves only under certain pre•
suppositions, and 1t 1s really the nature of the behef m God to
try to get behind these It 1s no accident that this method has
not been apphed ; 1t couldn't be. Could any proof satisfy us,
as proof 1s usually understood-the type of geometrical proof
and the proofs m natural science? Perhaps some one may think,
IF they could get a proof of the empmcal form found m natural
science, they would be sattsfi.ed But would they? It 1s charac•
ter1stic of the natural sciences, however careful the experiments
and observations, that they never see mto the nature of the
necessity, and this 1s why they never attam mathematical
clearness and certamty.1
But what of a proof m which we see the necessity? One must
question whether a demonstration of such quasi-mathematical
form could satisfy us, if we try to realize what it is we want
to prove. In some subjects demonstration of a fact is more
valuable knowledge than the direct experience of 1t. (E. g.
1
See for this Part IV, ch. 2, on Induction and Lotze.
 Rational Grounds of BeUef in- God 853
eclipse of moon at a certain date ; the property of a geometrical
figure.) In the case of the geometrical figure we know even
more certainly than we could by experience. But in some
subjects we have an impulse1 which cannot be stilled, for what
is called direct knowledge.
If we think of the existence of our friends ; it is the ' dirett
knowledge' which we want : merely inferential knowledge seems
a poor affa1r. 1 To most men 1t would be as surprising as un-
welcome to hear it could not be directly known whether there
were such existences as their friends, and that it was only
a matter of (probable) empirical argument and inference from
facts which are directly known And even if we convince our-
selves on reflection that this is really the case, our actions prove
that we have a confidence m the existence of our friends which
can't be derived from an empmcal argument (which can never
be certam) for a man will risk his hfe for his friend. We don't
want merely mferred friends. Could we possibly be satisfied
with an inferred God ?
It may be objected : You are only, if right, stating what we
earnestly desire. It doesn't follow we shall get it. It doesn't
even follow the desire is a rational one.
It may be answered . The demand for this kmd of knowledge
is not somethmg of ours outside the conception of God, 1t hes
in the concept10n itself• We can't think the content of the
conception otherwise ; tf we do, then we haven't really the con-
ception before us. To substantiate this (reply) we must think as
fully as we can what the content is for the religious conscious-
ness. We can only realize it m certain actual experiences of
our own, which we may call religious experiences. Times of
great emotion, when, as the saying is, we seem lifted beyond
ourselves. Times, 1t may be, of great trouble ; or of great JOY,
This can't be described in abstract language, any more than we
can describe our feelings and attitude m hearing, for instance,
a moving piece of music. We have to appeal to the experience
of those who have had such feelings. So it is here. In order
to understand what the conception of God is for us we must
1
On this question, however, see Part III, H 263-8. [Cf Butler, Sermon
XIV]
[• This 1s where the author touches upon the Ontological proof.]
 TENTATIVE INVESTIGATIONS
recall such experiences and they must be experiences of our
own, not any account of the experience of others.
It might here be obJected: The sum is shortly this: (I) You
have not rehabilitated after all the argument from design.
(2) You admit that an empirical argument such as is used in
the natural sciences 1s not applicable to such matter and, even
if 1t were, 1t couldn't be complete in a direction in which we
demand completeness. (3) As to a 'demonstrative argument',
in which the necessity ts understood, you rather imply it doesn't
exist yet, and you don't say whether we could have it. But
you say that probably even this, if we could have it, wouldn't
satisfy us (4) You indicate that the conception itself demands
some direct knowledge--' experience' or something like it.
Now (it may be said} (a) this direct knowledge of God, some-
how directly in our consciousness, most certainly cannot be had.
People know at least what 1s m their consciousness, and they
are not conscious of this ; and as you have said of the demon•
strative argument, 1f it existed we should have heard about 1t ,
-1£ this direct consciousness were a fact, the same would be true.
Besides, (b) people wouldn't look for a demonstrat10n if they
had the direct knowledge, they don't look for a proof of what
they are immediately conscious of.
So (1t might be urged} the conclusion of thew hole matter seems
to be this : If the suggestions were nght about the different forms
of argument, &c., they pomt to one avenue of knowledge as alone
satisfactory, and this avenue 1s closed. The belief in God, then,
would be reduced to a mere hypothesis and an hypothesis which
on the one hand your argument cuts off even from the poss1biltty
of verdication, while, m fact, 1t has serious difficulties to con-
tend with m our experience. It may so far be possibly a true
hypothesis, though the demand for direct knowledge is not
realizable
But can even tlus be allowed? If you put the demand in the
conception of God itself, and don't treat 1t as a mere subjective
attitude, 1s not the case stdl worse? If the conception demands
an imposs1b1bty, 1s it even valid as a conception? If not, it is
no use considering whether a real ob3ect corresponds to it.
The above objections depend upon the two propositions
(a) and (b).
 Ratlonal Grounds of Belief in God 855
§ 576. Let us first notice some facts in the religious conscious-
ness, whether connected with a definite form of religion or not.
We find in the religious consciousness as expressed in its
highest and purest form-i. e. not mere rhetoric, for instance-
that the direct knowledge of God (or ' ') • is regarded not
only as a consummation to be desired but certainly to be
expected. However, this is not looked on (in such a mood) as
a present fact-that we now have this directly in consciousness,
but rather as if necessarily in the future. Compare 'but then
face to face '. b
Definitely as this seems to give up any direct knowledge at
present, yet there 1s another tone qmte as real. Men speak of
God as being everywhere, all-pervading and so necessarily in
themselves. They speak accordingly of being in the 'presence
of God', and clearly not as a metaphor.
In the religious consciousness, when not so definitely associated
with rehg1ous forms and expressions, we find the same antithesis.
In one frame of mmd even to those who have the belief, God
seems a long way off-an mfimte way, as they would say. He
obviously can't be among the objects of the material world.
We speak of him as 'behmd' it-but as behind an impenetrable
veil. In a way it would be a comfort 1f he only were a long
way off m space, but he 1s 'behmd' the mfinity of space itself.
Even 1f we say he 1s in the obJects which are near enough to us,
he seems concealed in them and, necessarily, even more con-
cealed (it seems) than the forces of nature, which m a way are
themselves not obJects of a direct experience either.
On the other hand we find expressions which imply a con-
sciousness of the nearness of God, of his most absolute closeness.
Every one will recall the lines of Tennyson.c Nor can we object
(that) it is poetry and mere poetry-for we feel these lines of his
[• This lacuna was illegible in the first draft and IS left as a lacuna by the
author. ? • Ult1D1ate reality • as 1n the Stoics
b • For now we see through a glass, darkly , but then face to face . now
I know in part, but then shall I know even as also I am known '-St Paul,
I Cor Xlll 12
c • Speak to lum thou, for he hears, and spmt with spirit can meet,
Closer is he than breathing, and nearer than hands and feet '
TIie H1gller Pantheism.
The lines are quoted by T H. Green (WorAs, vol.111, p 273) 1n a passage worth
companng)
 TENTATIVE INVESTIGATIONS
express the strongest personal conviction-not a hope, not an
inference.
These things are so far no proof-indeed they are antithetical
-but they are most important and real elements in experience
and may possibly help to a solution. It seems as if our solution
ought to find a place for them. They at any rate make us
hesitate about accepting the above propositions 1 as so cogent
as they claim to be
§ 577 Let us now try to face thcsc,proposit10ns
Is it true that we must be aware of what 1s in our conscious-
ness and actively operatmg in it~ No. Even m the acts of
knowmg and perce1vmg there may be something really existing
and operatmg m our mmds of which we may not be explicitly
aware
(r) Take first a humble example-the syllogism-' Barbara'.
This 1s a form under which we necessarily thmk, if we thmk at
all, and all must have 1t; but. 1t 1s long before we recognize it.
Yet it is real m us because a law according to which our real
thought operat.es But yet for a time we arc unconscious of 1t.
Mill's good village matron, with her daughter Lucy, even believes
she hasn't got 1t and has unfortunately persuaded the philosopher
that she hasn't • A httle conversation with her, however, would
easily brmg out the implicit form of the syllogism
(2) More important for our purpose Take the familiar topic
of Causality When we speak of what we know directly, of
that of which we are consc10us, we tend to thmk of objects of
sensuous experience only, feelmgs of our own, or objects per-
ceived-that is given (as the phrase ts) as sensuous perceptions.
If we think of these as all the content of consciousness, we shall
get mto immense difficulty, because there are elements necessary
to our expericnrc which can never be objects of perception in
the above sense-the ordmary sense. Now we do all tend to
make this hm1tatton at first about the nature of that of which
we are actually conscious; and 1t 1s a principle, too, m philosophy
as well as in the untutored consciousness. Thus there was no
place in Locke's philosophy for the conception of cause ; causality
never being an object of such experience as he 1s thmkmg of.
1 p 854 (a) and (b).
[• Mlll, Sysllm of Logit:, II. m, § 3 )
 Rational G1-otmds of Belief in God 857
But he fails to notice that, because he appreciates the working
of the conception and thinks really only of that. Hume's
criticism detects this fault. Hume holds the same view of the
object in experience and is driven by his logic, in effect, to
a denial of the conception of causality altogether; though really
he doesn't get this clear, because the conception is there and
can't be dented away, and consequently he gets into great con-
fusion, sometimes seeming to ehmmate the idea altogether, and
finally substituting another idea for 1t, something which it
certainly 1s not.
The fact is, the conception is there and has conditioned the
very experience in question. Here it is not merely that we
have not become aware of a necessary element in our thinking,
but we have actually dented that we have it at all.
(3) And then the most important illustration of all 1s the
notion of our own Self. Following such presuppositions as
Hume's, we can predict with certamty what must happen to it.
That which itself experiences-experiences these feelings, these
perceived ob1ccts-cannot possibly itself be one of such objects
of experience It 1s m a far worse cond1t1on than the conception
of Causality. And the Self m this ph!losophy had to go. The
very presuppos1t10n of experience 1s condemned by the test of
that experience to be non-existent, simply because we can't be
aware of ourselves as obJects of sensuous experience. Yet that
we are conscious of ourselves-though, of course, not in the way
of such expenence-1s the most absolutely certain thmg of all.
Every one will recall the words of Descartes.•
As to the second propos1tlon, 1 if people have really laboured
to disprove that which 1s most directly present to conscious-
ness, we shall feel no surprise that they should have tried to
prove what they have got already. The attempt to derive the
law of umform1ty of nature by an empirical argument from
experience is a notable mstance. We have the paradox that
a philosophy, whose metaphysic has destroyed the notion of
causality altogether, really tned to establish its validity (for
science must have it) in the world of experience by an inductive
proof. The proof is obviously an utter failure. The thing to
1
p 854 (b)
[• Cogtlo ,rgo sum, ]11 pens11, done JII su1s, I am consc1ous, therefore I am.
D, 111 Mdthod11, 4me Parbe.
 TENTATIVE INVESTIGATIONS
be proved is 'tumbling under our feet' • all the time, being the
very condition of examining scientifically that experience from
the examination of which we hope to derive it by scientific
examination.
The fact, then, that people have tried to find a proof of God's
existence is so far compatible with His direct presence in their
consciousness ; and the fact that they think, or some think,
they certainly have no direct experience or knowledge of God
is compatible with the same hypothc:s1s. And now the facts
quoted from the religious experience, oscillating between two
extremes-the representation of God as immensely far and
unspeakably near-get a new s1gmficance and we can hope to
explain both sides of the antithesis .
• •
§ 578. Is there any evidence of such a direct experience which
we thus ignore ?
Let us try to consider again some peculiarities of the religious
consciousness. As already said, we ought to try to keep it
before us m the most definite and concrete form, recalling m
ourselves definite experience and feelmgs. Though of course we
can't attempt to describe 1t exhaustively, we can call attention
to some stnkmg features of 1t The conception has a certain
settmg m consciousness and pecuhant1es which present an
analogy to the moral and aesthetic consciousness I do not
now refer to doctrmes of revelation and faith as modes of
knowledge, because they come from reflection on the conception
or behef, but to the active presence and working of the con•
ception m what we may call, all of us, our 'rehg1ous experience'.
We cannot mamtam towards the idea the mere unemotional
attitude of the scientific or speculative understandmg. We do
not think of God, to use popular language, merely as the cause
of the material world and the power which is always sustaining
it ; or m more philosophic phrase as the primum movens, or, say,
as the umty of the material world-i. e. the one reality of which
1t is all the vaned expression, and the true substance, so to say,
of which they are the attributes. Nor do we thmk of God as
[& Dai\111 </'(UllfTIJI ,rpa ,ro&il• ti"" IE dp,ris irllMll'lltfuBcu Ital wx llilp@pfll' IJ.p' ai'iT6.
Plato, Rep 432 D , cf Bacon, N O 1 110
b Lacuna marked m onginal J
 Rational Grounds of Belief in God 859
the basis of the spiritual world, merely in the same way as we
think of the abstraction of our own selves, or our own self·
identity, as the basis of our life manifested in time. But the con-
ception is inseparably connected with certain feelings or emotions.
The one to which I want to draw principal attention IS the
emotion of reverence. This is not to be confounded with the
fear of the superstitious. It 1s not confined to the lowest or
least intellectual, but reaches its completest development in the
highest and the most cultivated.
Yet, possibly some one might obJect, this is an utterly wrong
method. We want the dry hght of science and philosophy;
we know that our judgement can be obscured and prevented
by feeling , whatever notice the anthropologist must take of
such things, it 1s the very busn1ess of the scientific or philosophic
investigator to free himself resolutely from their influence. On
the contrary, 1£ we imitate scientific procedure, this 1s exactly
what we must not do. With the precept of Bacon a.nd the
example of Darwin we must omit none of the facts which we
know of, and this fact, that the conception of God can only be
realized by us with certam emotions, 1s not only a very mteresting
fact but it 1s an essential characteristic of the cancept1on, for
it distmguishes 1t m a very stnkmg way from any scientific
conception. If 1t belongs to the nature of ~he conception as
such that it should produce or be accompamed by such feelings,
we must, 1f we are to understand 1t, attend to these feelings
whether any 1llus1on 1s connected with them or not
§ 579. Let us change the argument to another sphere.
Consider such emotions as are proper to the moral conscious-
ness. Take (e. g.) the emotion of gratitude. We have feelmgs
which, as we say, are due to physical causes; they do not
or1gmate from any previous thmkmg of ours, are not excited
by conceptions. But there are others which are only possible
to a thinking rational consciousness as such. In this sense they
are rational, and 1t 1s a want of appreciation of the mseparable•
ness here of reason and feeling which constitutes the difficulty
in moral philosophy of which I spoke at the begmning.
Gratitude belongs to this class of feelmgs. And it presupposes
a good deal of rational act1v1ty in consciousness. We are
grateful to some one for some good (as we think) done to us.
:1773,.11 G g
86o TENTATIVE INVESTIGATIONS
Now what does this presuppose? (1) Not merely our enjoy•
ment of a pleasure, but reflect10n upon it, distinct thought of it
as some good to us. (2) It 1mphes also that we refer it to
another as an effect of his causality. Which means again that
we have, in some form, conceptions of ourselves as distinct from
other selves, also that we have the category of Causality and
more -:lspec1ally of the activity of a person as a cause. (3) It
implies also that the person, whom we have thought of as
a person and a cause, \\·died and intended the action ; otherwise
we are not grateful. And further (4) that he is benevolent to
us, "1shcs us well, has done 1t to please us, or certainly to give us
good The apprec1at1on of these relations between persons perhaps
may seem to require considerable rationality, but whether it
does or not 1t must be present or there could be no gratitude.
It may be asked whether this can be true of the consciousness
of children. Obviously 1£ they arc susceptible of gratitude these
elements must be present, though not abstracted from action
and reflected on, or the chtld could not be grateful ; and in
actual life we thmk they arc grateful and we certainly treat
them as if they regarded the person lo whom they are grateful
in the way described.
Now here is a feeling \\-hose whok meamng 1s relative to
conceptions and acts of reason, impossible without them ; and
for it to be possible these relations must be somehow appreciated.
But more than this, the feehng itself 1s mseparable from the act
of Judgement m which the rec1p1ent of the benefit appreciates
his true relation to his benefactor. The Judgement 1s not there
unless the feeling 1s there too. We cannot separate the Judge-
ment from the act as something m itself speculative and in
itself without the emotion. We cannot Judge here except
emot10nally. This 1s true also of all moral and aesthetic Judge•
ment Reason m them can only manifest itself emotionally.
Now the feclmg of gratitude would be admitted as a fact
given m experience, and I want now to argue from the facts of
experience In the first place the feeling is obviously only possi-
ble for us through the conceptions , 1f it exists, then, it guarantees
the poss1h11ity of the conceptions for our thought at least. We
must be able to thmk of ourselves as something more than and
different from our feelmgs and as having the feelings ; we must
 Rational Grounds of Belief in God 86.t
be able to think of another self than our own ; we must be
able to think of such self as willing, and willing us good :
and therefore to think of a certain relation between selves, So
much for the possibility of the conception.
It may be asked, What of the reality of the object corre-
sponding?
This means now for us: Arc they objects of experience? Do
we suppose we could commumcate by a priori teaching to any
one the meaning of the feeling of gratitude? This is no more
possible than to explam the meaning of music to one who has
not experienced it. He must have really experienced it. He
must have an actual self and (have) d1stingu1shed himself from
his feelings. He must have had an experience by which he re•
presented a person m such a relation to himself. But now even
suppose he was mistaken m attributing goodwill to another
person; for that, 1t may be objected, 1s but an inference ; or
that he is mistaken even in supposing another person besides
himself, for this again (1t may be contended) 1s but inference;
he hasn't direct experience of the person (1t might be said), he
really infers him from certain elements m his own experience.
Well, the idea (of the) goodwill of another person agam, even if
here mistakenly applied, can only come through some experience
of this attitude in ourselves. Goodw1ll of a person, then, must
here be a real experience.
But it may still be pressed (that) the other person is some-
body not in our experience and never can be, and that experience
has only guaranteed one person. Let us suppose that 1s so,
though it seems to me a serious over-statement. Then at least
the reality in experience of gratitude guarantees the reality in
matter of experience of every element we have said to be
presupposed except the last. But as to the last, I at least can't
have goodwill towards a person whom I suppose revealed to me
m an experience unless I certamly believe he 1s there.
§ 58o. Let us return to the feeling of reverence with its
solemnity and awe. What conception does this presuppose?
In itself it is not fear, love, adm1rat1on, respect, but something
quite sui genms ; it 1s a feeling directed to a spiritual being
and presupposes the conception of a spirit.
Can it be directed to a human being? We speak of love and
G g 2
862 TENTATIVE INVESTIGATIONS
respect and admiration of a human person, and 1t 1s true that
we speak of a person sometimes of exceptional goodness as
revered. But one must contend that this is not right and that
we shall detect, on severer analysis, that it 1S a metaphor of
ecaggeration, and strictly only implies a high degree of respect
and admiration and regard ; but we cannot really say we have
a feeling of reverence in the presence of such a person ; and, if
we add the feeling of solemnity which belongs to reverence,• it
seems still clearer that it ts not so.
The feeling, then, seems directed to one spirit and one alone,
and only possible for spirit conceived as God. 1
Agam, it 1s true that "e speak of reverence for the moral law ;
but again I believe no such feeling possible for a mere formula
and that, so far as it exists, 1t 1s only possible because we think
of the moral law as a mamfestat10n of the nature of the Eternal
Spirit. Here again the feelmg of solemnity seems to help. For
those who have such a real experience of reverence and solemnity
the existence of this emotion guarantees the possibility of the
conceptions, and "e retrace the same argument for the reality
of the experience as m the case of gratitude. \Ve must affirm,
here as there, that we cannot have the conceptions without real
experience to correspond There we admitted, verbi causa, the
possibthty that the assumption of the existence of another
person, though grounded on our own experience, might be only
an imagination grounded on an experience of ourselves. But
here the feeling presupposes the conception of another spirit,
whom we must tlnnk as nol human or like ourselves, or the
conception \\ould never give nse to the feeling. Can we resist
the conclusion that we must have had experience of the reality
of such a bemg somehow w1thm us ?
(This• argument should have a special claim on a set of thinkers
who, from their mtellectual bias, may seem to be the most
unlikely to admit that there is any sound evidence for a belief
m God. These are the adepts m the methods of the empirical
sciences and the adherents of empmcal systems of philosophy.
For 1t is Just these who maintam and emphasize the doctrine
1
See further the last paragraph of tlus section, p 864 (top of page)
[• An mterpolatum subsequent to 1903 and probably Jlltich later]
 Rational G,ountis of Belief in Goa 863
that we can make no 'idea' ourselves, that the ideas we have
~e 'ideas' of objects actually experienced, or arbitrary com-
binations of them by us. In the latter kind, only the combina•
tion is supposed to be ours; the elements combined must come
from experience.
Have we not here, m what we may call the religious experience,
an emotion which absolutely guarantees the existence in u11 of
the idea of a spmtual being transcending everything human,
(a being) which 1s no combination, arbitrary or otherwise, of the
'idea' of other experienced objects ? In the actual emotional
experience itself there can be no thought of any arbitrary com•
bmatlon by us of ideas of what is already familiar in experience.
Rather something we could not have imagined beforehand,
something spiritual which we utterly fail to describe, seems t-o
be present to us without our will or co-operation and, without
our will, to fill us with a unique emotion. Is 1t not certain
that the solemn reverent attitude 1s only possible because we arc
convinced of the presence of something entirely transcendmg
everything human and all other experience? No one will think
it worth wlulc to contend that such feelings could be awakened
m us by the contemplat1on of an hypothetical and problematic
combination of our own. When (we arc) not in the actual
experience we may have recourse to combinations of ideas from
ordinary experience and to analogies from such material. These
attempts fail and their failure confirms what has been just
maintained, for no combmation of objects of ordinary experience,
however magnified, can produce the conception of a thmg which
could affect us with such emotion.
It would seem, then, that the idea of God is one of those ideas
which, m accordance with the empmcal principle quoted, must
be the ideas of sometlung actually experienced. But the principle
itself is not confined to empiricists Rather the essentials of it
(for empmc1sts understand 'what 1s given in experience' in too
narrow a sense) and all that 1s necessary for the present applica·
tlon seem necessarily admitted m all philosophy.)
But, at the very lowest, we are entitled to argue that just
as the person who has a real experience of gratitude must believe
that there 1s another actual person to whom he is so related
that he is grateful to him, so the actual feeling of solemn
 TENTATIVE INVESTIGATIONS
reverence ill only possible because we actually do believe in God.
So, in such feelings, in the actual experience of reverence and
solemnity we are believmg (not fancying, not imagining); we
are believing in God, for it is impossible to have them at all
actually except through that belief.
(Those• \\ ho have had such an experience may, at times when
it is absent, search for a proof of God's existence and may think
they require one. But they arc putting a question which could
not possibly be a question to them when they have such an
experience-an experience m which they have been absolutely
convinced of the reahty which they now seek to establish other•
wise. The corrective is to recall such experiences if we can.
But that is not easy. There is an analogy m our aesthetic
consciousness ; for often we can recall but impcrfectly the effect
of a piece of music though "e are certain the effect was there.
Besides, these 'religious' experiences are often, perhaps oftenest,
such as have touched us so deeply that we mstmcbvely shrink
from trymg to reproduce them They may h.t.vc been at times
of unspeakable sorrow, such as the loss of a friend. Is it not
here that the great poet can help us? For he can awaken the
mood m us agam, and m his imaginary world bring home to
us the livmg experience itself. Poetry does us here a far higher
service than philosophy can ; yet 1t needs philosophy to
vindicate and complete 1t For to turn to poetry m such issues
we tend lo thmk of as .t.bandonmg the difficulty of reasonmg
to take refuge 10 the realm of fancy. This 1s a great mistake
"hich 1t 1s the privilege of philosophy to correct
But to avoid m1sunderstand10g, somethmg must be added.
It is not mt>ant that m "hat we have called the religious
experience a man must say to himself 'Lo, God 1s here I' or use
any ordinary rehg1ous formula. It 1s enough that 10 such an
emotional state he should be convmccd of something infimtely
beyond himself and all humanity, 10spirmg him with an emotion
which nothing human could inspire-reverent awe. One recalls
the fine saying of Tacitus about the religion of the Germans,
'secretum illud quod sola reverentia vident'.b)
[• Another interpolation to end of § 58<>, of same date as the last
b 'That mystery which they behold with the eye of veneration alone,'
-Tacitus, Ger111a111a, g]
 Rational Grounds of Bel,ie/ in God 865
§ 581. The feeling, then, points to the reality of the experience.
Such a view would throw light upon the antithesis we mentioned.
We can see why we can fail to recognize what is 1n us, and how
this which works within us is the original conviction which,
under the influence of science, we misunderstand ; and that it
is really this misunderstanding which made us cast about for
arguments such as that from design. But no argument from
design, illustrated by examples however wonderful-and there
are wonderful ones-can have anything like the effect of a single
experience of the emotions of reverence and awe. The one is
but an argument; the other is a sense of the reality itself and
1s the only thing that can satisfy us. This need not interfere
in the least with any one's belief m the edifying effect of religious
services. These are essentially for those who have already
a belief m God. Even the rebg1ous exhortation intended for
'conversion' (or whatever 1t may be called) as a rule assumes
that the hearer believes m God. If it 1s really a question of
convincing the hearer of the reality of God, the best chance
of the preacher hes m the poss1bihty of awakening in the hearer
what has been 1mpbc1t in some of his own experience. And does
not this depend on what is essentially a poetic faculty?

XVI (Appendix)
§ 582.
Nov. 3, 19II.
My dear Webb,
I was very surprised and pleased, and indeed touched by your
kindness m honourmg me with the present of your book. 1 What
can I do m return? Send you c1. copy of 'Traverses' 2 or of
'Interpretation of Timaeus' 3-both much at your service I My
stock 1s (alas I) so small. I can't pretend to have read yet so
far m the book as to discover your kmd and conscientious
reference to myself on p. 269. It was pointed out to me by
Williams. I rather wish I had known you thought of bringing
me in, for I should hke to have talked to you about it. It
doesn't put the thmg quite m the way or m the settmg m which
1 P,obl,ms an Ills Rslr,dions of Gotl antl Man, by C. C. J. Webb, Fellow of

St Mary lllagdalen College, Oxford (London, N1Sbet), 19n,

• On llae Traversing of Geometrical Figures (Oxford), 1905.
• On lh, lnlerj,relation of Plalo's T11naeus (London, Nutt), 1889.
866 TENTATIVE INVESTIGATIONS
I gave it. But the essential point I suggested as important is
given or implied, p. 2701 11. 8-13 1 viz. that whatever we might
suppose we had to think of God's omnipotence, we had the
solid fact that in the world there existed imperfect beings-human
beings {whose wJllmg and designing &c. we must distinguish
from God's) · and that therefore there was no d,i,fficulty in
supposing that the agency in nature might be of a similar kind
(or, as you say, no new difficulty) and no more God's di'rect
agency than our own agency was
You have put 1t a little differently. but this was my essential
contention, 1 e. 1t saved one mere a priori difficulties about
dualism, because we have to admit and do admit such a difference
from God's will and action as our own, and the suppos1tlon of
an imperfect development m nature would therefore be no new
difficulty (1. c. 1t no more commits us to dualism to suppose
God not 1mmed1ately operative m nature than our belief that
He 1s not immediately operative m us-our will not Hts, our
imperfect knowledge not His imperfect knowledge, &c) We
explain our own wrong or imperfect domg by ascr1bmg it to
our own causality and not to God's. S1m1Jarly we may explam
the 1mpcrfect1on, which so troubles the argument from design,
as due to this other agency m Nature-not God. And Imam•
tamed 1t might be a mistake therefore to do what was usually
done ... attnbutc all that happened m nature directly to God
as though 1t was His special province, forgetting that we have
already made our own action our special province and not His.
This 1s m effect JUSt what your statement 'assumption we are
nowadays too apt to make . attributed wholly ... to God'
comes to
Curiously enough, I have been led myself lately to the con•
s1derat1on of this very subJecl. The MS. of my own effort some
years ago (14 I I fear) 1s so imperfect and unsatisfactory that
I have with one exception refused to let any one see it...•
I undertook what I feared an almost hopeless task of recon•
structmg the fragments, although when I got to 1t I found it
not so hopeless as I expected. And though I am dissatisfied
at not being able to reproduce what I said and (worse I) what
I felt, I have succeeded in making the thing intelligible. . . .
I have said m it something pretty decidedly about the argument
 Rational G,ountls of Belief in God 867
from design and evolution, which leads most naturally to the
thoughts I have just been referring to. I didn't develop it
because my great object was not to vindicate the 'design'
argument, but to make something quite different the essential.
. . . I observe that you have not added what I should have
thought the quite obvious correction of Zeller, viz. that Plato
m the Timaeus commits himself to the most decidetl dualism.
I explained this point m my reviews• of Archer-Hmd, though
not m the pamphlet. It 1s really to my mind rather absurd to
make such an objection as Zeller's m (the) face of this. Once
more let me thank you very heartily for the honour of the
present of your book-a thmg I could never have expected.
Yours truly,
J CooK WILSON

[• One of the mam doctrine& of the T1mae1,s ' rel.ttes to the existence of
evil all cosmogonies "Inch attribute the world to some d1vme act1v1ty
find a difficulty here Some assume another \pmt, an evil one, though
partly subordmate to the good one , others, to avoid making an e, 11 spl.1'1tual
principle, as&ume an unmtelhgcnt matter, or m general 'iome form of Necessity
beside the Good Spmt We should suppose that Plato, 1f not monist, would
mchne to the latter and should have thought he clearly adopted 1t m the
Timaeus Thus Plato say, there .tre two cause,, i,oiis- and ,1.,,1'1"'1• each
with its own effects and therefore the ongm of the world 1s mixed , the
necessary causes are 1rrabonal and disorderly (40 e, 48 a, &c ) Again,
Plato does not represent l'oiir a'i entirely creating matter, but only as mocb•
fymg 1t God 1s never reprc-sented as makmg 1t entirely good but as only
domgh1s best with 1t (30, 47, ~3) '-Review 1n the Oxford Magazine, March 13,
1889, p 275]
 POSTSCRIPT

SOURCES AND DATES OF THE

SECTIONS OF THE LECTURES
 POSTSCRIPT
W1LSON, as JS clear from many places m his familiar corre-
spondence, made no pretence to be writmg a systematic treatise
on logic Thus he says, 'm the meantime I have been y~a:lY
improving (?) my logic lectures and have come to be pretty
clear on certam departments-that may be publishable some-
time, and yet here agam I feel no mclmation to write a general
treatise on logic but only to discuss those parts to which I think
I can make real additions'. 1 He \\as also most averse to
premature publication. The revision of his lectures begun m
lus last years, tentative and imperfect as he knew it to be, was
cut off m the middle by his death. Yet this has seemed no
valid reason for w1thholdmg from a larger audience work of
unusual, 1£ unequal, distmction, perpetually reconsidered by 1ts
author and of proved service to many generations of Oxford
youth The cdit10n no,, submitted to the pubhc will, I hope,
m spite of many 1mpcrfcct1011s, be of profit to more than the
group of Oxford tutors who desired its achievement and arc
already fam1har with its general purport The merits of the
author \\ill perhaps cover blenushes m the execution of a task
which seemed not mdeed a trust expressly committed to my
fellow executor and myself but still a debt we owed to Oxford
and to the memory of our distmguished friend
That the author hoped some day to appeal to a\\ 1der au<hence
seems certam Certamly also the lectures as they now appear
are not ,,h.i.t he would have made them or h.i.ve desired them
to be. Tlus 1s due as much almost to his own literary pro-
c-ra&tmat1011 as to lus consc1ent10us desire for truth. Painfully
industrious m his solitary reflection, reckless spendthrift of his
time upon subsidiary themes, profusely generous of his leisure
to pupils and younger colleagues, m actual composition he
resembled Dr. Johnson, he worked 'dilatorily and hastily,
unwilling to write and writing with vigour and haste'. His
many manuscript remains are not suited to publication ; they
1
Letter to 5 C. P., 13 vo1.
 POSTSCRIPT
are the dialogues of a man with himself, where the thread is
often hard to discover. This habit of liberal self-colloquy gained
upon him with advancing years, and he cared less and less to
concentrate upon and reconsider the form of what he had thus
composed.
His own apology for the delay to commit himself to the
public is contained in a preface confidentially circulated to a few
friends with his last prmted volume of Dictata, that corresponding
generally to Part I of the present book.
' In printing from a MS. which contains many corrections and
many mterpolations, representing, as they often do, gradual
changes of opimon, 1t was natural to find much that at least
required adjustment and much that had to be rewritten alto-
gether.... My first reason for printing such extracts was that,
as my views developed, I found I could not cover enough of
the subject in the time available for lecturing But I have
found the process of passing from MS. to pnnt so salutary that
I hope gradually to print the whole, adding the separate courses
on Classification (and Defimt1on) and on Hypothetical Thinking.
These extracts however are neither to be published nor to be
treated as pubhshed Those of my colleagues who were kindly
anxious that I should publish my lectures years ago ... will
be, I am certain, well content now that I did not comply with
their wishes. I really felt that there was too much which
required a more thorough treatment and this uneasiness has
had time to develop mto ( I thmk) a clearer consciousness and
to produce considerable changes m the old matter as well as
additions to it.
' There 1s a greater danger of fixing one's thought by pubhca•
tion and arresting one's own progress than is generally recog-
nized. I have often noticed that qmte able thmkers have the
greatest reluctance in retractmg anythmg to which they have
committed themselves by publication, though the mistake may
be perfectly obvious to the critic (whose work is mcomparably
the easier) and the author could only gam by admitting it. But
the (printed) letter k1lleth and it rs extraordmary how it will
prevent the acutest from exercising their wonted clearness of
vision....
• I hope, by my present method, to gain that greater clearness
 POSTSCRIPT
which is usually the result of printing . . . and at the same
time to preserve the comparative freedom one enjoys as long
as one's thoughts are only in manuscript. I hope also it will
enable me at least (for I dare not count on more) to remain
nearly as amenable to reason as 1£ I had printed nothing. Of
one thing I am fairly sure and that is that I shall want to
revise both this pamphlet and the others which I have printed;
while it is fairly probable and certainly desirable that seoveral
stages yet may precede publication, if I ever publish at all.
· The reader may observe 111 some of these lectures and in those
on inference which I hope to print later a s1m1lanty, and some•
times a close similarity, to passages 111 the work of contemporary
writers. I should hke to explain that I have been lecturing
a long time and the passages I have in view, or their equivalent,
were delivered years before the appearance of the books afore-
said. In this connexion I may be permitted to add something
which concerns a popular criticism of Oxford, as a University
where philosophic systems are taken up when they have been
worn out on the Continent. Y cars ago when, after the excellent
instruction I had received from most d1stmgu1shed philosophers
in England and Germany, I began to try to think things out
agam for myself from the beginning (a tendency character1st1c,
I thmk, of Oxford in those days), I was gradually led ma direc-
t10n very different (unless I am much mistaken) to the tendencies
of the schools in which I had been educated, Perhaps this
showed itself most m the subjects usually comprised under the
title of the theory of knowledge, subjects but httle represented
m what I am now prmtmg 1 For a long time I heard no sign
of any such tendency as my own on the Continent, but, after
the lapse of some years, one of my hearers told me a book had
just appeared m Germany which had important affinities with
my own views. This was in 1908, and the title of the book
was Das sogenannte 2 Erkentnissproblem. The main thought in
the book seemed to me that indicated by its significant title,
and that thought had been familiar m Oxford teaching and
discussions long before the appearance of the German book.
I am wondering whether I shall also m time hear that people
• Cf T1s'1mo,n11, p luiv, from Mr H A Pnchard's preface.
• 'sogenannte' 1s underhned by Wilson m his or1g1nal manuscript
 POSTSCRIPT
have begun to realize the failure of every theory of perception
which has yet been proposed and to attribute this to a fallacy
in the very idea of a theory of perception as ordinarily under•
stood.
' Lastly, I find in looking at some contemporary works that
certain passages in my lectures might well give the impression
of having been motived by a polemic against statements in
those works. , The passages however had no such origin but
arose entirely from mdepcndcnt thinking about the facts and
about ordinary current logical doctrines. Of course some lectures
are directed agamst contemporary authors, and I need not name
them as the polemic is obvious, but the appearance of polemic
in a certam number of other passages 1s quite accidental.'
To bring the work as it stood to a form more tolerable to the
critical reader was the task I proposed to myself, and I have
endeavoured to keep the mean between too great divergence
from the original and too scrupulous an observance of strict
fidelity. We must regret that Wilson allowed his work to
accumulate m pupils' note-books, m manuscript and typescript
in the disorderly way I shall shortly describe, a way fatal to
consecutiveness, conciseness or complete clarity, whether of
matter or form, in a subJect where these virtues are peculiarly
necessary and mdeed incumbent upon an author. In the midst
of maturer passages, dashed off m great haste, he left standing
immature and mcons1stcnt fragments of earlier date, themselves
of varymg degrees of elegance and prec1s10n. The only com•
pensat1on 1s the light thrown upon the difficulties we may all
encounter m treadmg the same paths. My own work, then, has
been this. I first established a definitive text, by collation or
comparison of the various sources enumerated below. This gave
a sounder basis for this edition than any that existed before.
This text I then edited for the press, endeavouring to bring out
a result which might have been acceptable to the author had
he superintended my anxious labours Disorderly sections I have
rearranged, have corrected manifest slips, and excluded minor
inaccuracies of expression while restormg m very many minute
pomts the sense and words of the genuine original. The whole
I have arranged in chapters, with a continuous series of sections,
generally shorter than those which Wilson devised. I have
 POSTSCRIPT
never consciously amended the matter nor have I destroyed, by
misapplied industry, the flavour of the spoken word. The
author, when committing his thoughts to paper, always anti-
cipated delivery, and to give his sentences another turn would
often have obscured the meaning. I have therefore generally
left colloquiahsms and characteristic abruptnesses ; asyndeta,
even anacolutha sometimes ; and preferred a racy or vehement
turn to a less strikmg alternative, where that existed. Arria.n's
Epictetus was my model, and the reader will, so far as he can,
please to 1magme, ao; he turns the pages, that he is listening to
the eager and confident, the hurried tones of the man I have
endeavoured to present m my mcmo1r. I may not have suc-
ceeded, but I believe that I have not erred too much on the
side of familiar speech. Wilson had good models in English for
a colloqmal style, and made the admirable effort to write m
plain, honest, even homely language, avoiding cant terms and
eschewing jargon, though he lacked the felicity of masters hke
Locke and Berkeley. His own view m this matter recurs
repeatedly in the lectures, but nowhere so clearly as m the
following words 1 : ' The authority of language is too often
forgotten m philosophy, with serious results. Distinctions made
or applied 111 ordinary language are more likely to be right than
wrong. Developed, as they have been, m what may be called
the natural course of thmking, under the mfluence of experience
and in the apprehension of particular truths, whether of every-
day hfe or of science, they are not due to any preconceived
theory. In this \\ay the grammatical forms themselves have
arisen ; they are not the issue of any system, they were not
invented by any one They have been developed unconsciously
in accordance with d1stmct1ons wluch we come to apprehend in
our experience.
' On the other hand, the actual fact is that a philosophical
distinction 1s prima facie more likely to be wrong than what 1s
called a popular distmct1on, because it is based on a philosophic
theory which may be wrong in its ultimate prmc1ples. This is
so far from bemg appreciated that the reverse opm1on is held
and there 1S a tendency to regard the linguistic distmction as
the less trustworthy because it 1s popular and not due to
1 Onutted from the text at f 127.
 POSTSCRIPT
reftective thought. The truth is the other way. Reflective
thought tends to be too abstract, while the experience which
has developed the popular distinctions recorded in language is
always in contact with the particular facts.
' Now it is not uncommon in philosophic criticism that some
popular term, when reflected upon, presents great difficulties to
the philosopher ; difficulties which are often due to some false
theory of his which 1s presupposed. The criticism sometimes
ends in pronouncing that really the term only means what is
intended by some other term, so that it is in a manner explained
away and any distinctive use of it is supposed to be an illusion,
or the meaning of the term may be pronounced to be altogether
an illusion. When the philosopher arrives at such a conclusion
it too often happens that he is satisfied with this negative result
and considers his work done. But the truth JS that an mvest1ga-
tion cannot be fimshed m this way. We ought under such
circumstances to mquire how 1t 1s, if the given term only means
something else, that language ever developed it and still so
obstinately holds to it, and, when ,,e suppose that we have
explained a term away or shown that 1t 1s a mere unnecessary
way of d1sgu1smg some other meaning, we ought to put our
result to the test by trying to do without the word criticized
and seeing what would happen 1f we everywhere substituted for
1t what we suppose to be the truer expression.'
His conv1ct1on of the value of the unsophisticated d1stmct1ons
which underlie popular usage led the author to those hngu1st1c
inquiries which are among the most remarkable m his work :
the patient analysis of grammatical forms and the elucidation
of the implications wluch arc contained in ordmary speech.
This may be seen in the chapter on Denotation and Connota•
tion, 1 m that on the Meaning of grammatJcal forms, 2 and often
elsewhere, as, for example, in the passage where he claims that
' ordinary language reflects faithfully a true metaphysic of
universals '. 3
I think, however, that in his use of speech as witness to logic
and in his grammatical analysis he supports his argument upon
evidence drawn from too narrow a field. Thus hi$ treatment is
coloured by an almost exclusive consideration of Graeco-Latin
1 Part II, ch. zvi11. • Jd. ch. VU]. • I 82.
877J•ll Hh
 POSTSCRIPT
forms of expression, some of which are not altogether in agree•
ment even with our own cognate tongue. This limitation may
be seen in his decision to adopt the term I nominative to the
verb ' for what we ordinarily call, in accordance with time-
honoured usage, the subject of the sentence. This narrowness of
survey has also perhaps preJudiced his treatment of the copula
and of modahty. 1 Yet even if this be true, his examination of
language in this and other connexions appears to be of gre:at
value and interest to philosophy.
His belief m the efficacy of natural expression as an unsophisti-
cated mstrument of thought further led him to write down his
reflections in the words which came most naturally to his pen.
This is responsible for a certain looseness and, sometimes,
inconsistency of expression. This fault was exaggerated by the
manner m which the work as he left it had gradually grown up.
A few details will make this plamer. The lecturer opened
with a very hurried survey of one or more sections which were to
come ; these sections were then dictated word for word. Thus
he mverted the plan followed by Hegel In Wilson's case the
commentary preceded the text and at least from an early period
did not leave the text very far. Ile wa'I diffuse, difficult to
hear, and used geometrical example'! Thr present text is
thus based on a written original, the substance of wlurh varied
but little 111 oral delivery. The notes of pupils rarely, if at all,
supply matter "h1eh was lacking m the text. The early drafts
had careful trans1t10ns as ::i rule between the sect10ns, each part
ending with a characteristic aphorism The developmrnt, too,
followmg closely upon the tC'xt-books, was approxnnately
methodical in its progrl'SS In very early days there 1s some
trace of actual improvisat1011 from manuscript drafts, and the
author would omit a pomt, sometimes add, as oral delivery
suggested, consequences which had escaped him m the study.
But as early as 1892 he had begun to dictate from notes taken
by a pupil m some earlier year One obJect was no doubt to
hm1t hunself, another to have a fair copy before him. His own
drafts were hurried and abbreviated, difficult even for himself
to read These pupils' notes he would keep by him and work
upon them by way of correction and addition. Some years
• Part JI, ch x
 POSTSCRIPT
passed and he took a later note-book to read from and to
annotate in the same way. Thus the course grew and altered i
some parts dropped out, inaccuracies crept in, and the whole
was never seriously revised until a very late date and even then
in a great hurry and without due consideration. His standpoint
altered. The change is marked by two principal dates. Between
1907 and 1909 he worked over a large part of an unknown
hearer's notes of 1904-5-6, inserting on spare leaves, on
the verso, and on additional pages many subsidiary or sub-
stituted sections, as well as making minor additions to and
verbal alterations of the original text.
Parts of this volume were copied very skilfully by a typist,
the notes of two other pupils were similarly copied, and from
these typescripts, with minor manuscript corrections and some
additions, three volumes of Dictata were printed in 1913-14.
The printing, however, went on in the reverse order, first the
middle of Part II, then its beginning, and finally Part I. In
the process Part I was very much altered and enlarged from its
original lecture form, and matter belongmg to Part II was
heaped inconsiderately into Part I.
Moreover, the original words of the author suffered in these
continual copyings, less indeed than 1f the lectures had not been
precisely and verbally dictated, but sometimes seriously. The
ends of sect10ns suffered most, omissions having been made either
by the lecturer to save time or by the hearer m his hurry. Many
minor errors crept m, sometimes ludicrous mistakes. Moreover,
the changes m verbal termmology were seldom exactly carried
through, and when they were, the alteration was ,Jften, especially
in regard to the words Judge and judgement, merely mechanical.
None of this ,vould have been of grave moment, where matter
rather than manner 1s concerned, had Wilson really corrected
his typescripts or l11s prmted proofs with the sedulous care they
deserved. There is clear evidence that he did not , obvious
blunders are left standmg , he added after-thoughts lightly,
often corrected a faulty sentence with uncommon carelessness,
not seldom accepted a pupil's variant for his own words. 1 Thus
1 Since writing tlus I have come upon a letter from lum 1n which he himself
dnc:nbee the ongin of a mistake m the pnnted text wh1eh arose very much
in the way I have 1nd1cated

Hh2
 POSTSCRIPT
the printed pamphlets or dicta.ta. tended to stereotype a text
corrupted by the author himself, by his pupils, by the typist,
and by the printer. Moreover, the matter consists partly of
entirely new sections or paragraphs, dashed off with mature
conviction but very hurriedly in 1912 or 1913, partly of ancient
paragraphs remodelled to suit the changed pomt of view. Most
strange of all, certain parts represent word for word his own
earber lectures but have, by a slight turn, been made to ap~~r
as the arguments of some imagmary disputant and are then
confuted. New patches have thus been oddly mounted on old
garments and sections moved about with some disregard for
methodical procedure.
In the circumstances an editor's task was perplexing and
a compromise -inevitable. I have endeavoured to indicate by
my tables the mam character of the changes made. By dating,
at least approximately, the various sections or parts of sections
and by some rearrangement I have tried to remedy obvious
blemishes and at the same time to satisfy the more curious or
careful reader. To have marked every change precisely and pro-
duced a critical text would have been out of proportion to the
expense and might have distracted the reader from the general
burden of the argument, itself at times sufficiently perplexed
and mtricate. I have not endeavoured, any more than Wilson
himself did, to make the termmology consistel'l.t throughout.
This will be especially clear m Part III. Where, however, late
manuscript evidence existed I have cautiously preserved the
expression nearly exactly, though the style of it is more unstudied
and less elegant than his earlier work. Lost m the matter and
having long ceased to read, except at rare mtervals, the great
English models, he neglected too much the clearness and dis-
tinctness which cost care m compos1t10n ; his sense of form
declined and his word order of ten became German rather than
English. This last fault I have endeavoured generally to remove.
In regard to the lectures as a whole, there 1s one further cause
for a certam limitation. Wilson never broke loose from the
structure dictated by their origin. He combmed, from the first,
two original courses : one a lecture in Formal Logic, where he
followed the order of the text-books, the other an historical and
cnt1cal inquiry mto the Theory of Knowledge, principally the
 POSTSCRIPT
English psychological school from Locke to Herbert Spencer.
This will explain why two different motives so constantly meet
and cross one another in this book. He is combining two
different, if cognate, inquiries, embraced in Oxford under the
traditional title Logic. A third current comes from the historical
study of Plato and Aristotle, particularly the Organon. Thus
some of the interpolated sections arose from a preoccupation at
a certain date (1908-9) with the origm and process of knowmg.
Again the first half of Part II, which was m origm an exposition
of the modern theory of Judgement, was gradually altered into
a criticism of contemporary idealistic logic, and agam Part III,
origmally a criticism, under Kant's mspirat1on, of the syllogism
as the form of mathematical demonstration, developed into
a general view of 111fcrence which was to Justify Wilson's attitude
to mathematical reasoning. Similarly the two discussions of
Conception seem hardly logical in the strict sense, turnmg
instead round a criticism of the ideology of the eighteenth
century and its modern descendants Finally, the whole discus-
sion of Induction appears m the place occupied by the treatment
of Method 111 text-books subsequent to the Port Royal logic, and
the author never had time to umfy an original criticism of Mill
with his own treatment of Inference. These arc the causes which
make the book uneven 111 texture and which have mevitably led to
some mconsistcncy. The dates will, I hope, help to explain them.
I have been asked, and the question 1s natural, whether it
can be detcrmmed what thinkers most mfluenced Wtlson's pro-
gress. In the prefatory words which I have cited above he
himself speaks of the advantages of avo1dmg publication which
may stereotype an author's views At the end, ho\\ ever, he
claims to have reached unportant opmions many years before
the date of the preface Scattered about his papers are many
notes of dates, showing his anxiety to assert mdepcndence of
the contemporary progress of philosophy. Both thmgs appear
true. From a very early date the germs of what has come to
comparative maturity in this book are present in his writings
and lectures. This 1s certamly true of his dissat1sfactton with
the idealism current m his early manhood m Oxford. It appears
also to be true that he originated h1s characteristic tenets for
himself, 1£ there can be said to be any strict origmality in the
880 POSTSCRIPT
matter. That is to say, his whole effort was directed, as we
shall see, to maintaining the objectivity of mathematics and the
mathematical sciences. He was at first content with the Kantian
solution to which he was no doubt guided by Green, and so far
as mathematical method is concerned he accepted this solution
to the end. But he was dissatisfied very early with the excessive
subjectivity-as he considered it-of Kant's general position and
his own conviction of the reality of the mathematical object was
reinforced by lus growmg conviction of the immediacy and
certamty of perceptual experience. These two things played
mto one another and V\, ere strengthened by his continuous
absorption m problems of mterpretation of Greek philosophy ; 1
he fell more and more under the mfluence of the frank objectiv1sm
of the typical Greek scientific view of the world. This also
helped to shape his doctrmc of the umversal.
If now we ask what modern mfluences, besides his reaction
from Kant, led him in tlus direction, the answer is not Lotze.
Lotze's presentatiomsm is fundamentally opposed to Wilson's
views, though with charactenstlc loyalty he did not go out of
his way to say so The mflucnce ot Lotze 1s plam m much of
his detailed work as well as 111 the fundamental conv1ct1on that
man has, m that philosopher's words, 'an immediate certainty
regarding what 1s umversally vahd, upon which all conviction
rests, a certainty which, whether called mtu1t1on or by some
other name, must be admitted to exist, though its ongm 1s and
will, 111 all likelihood, remam unexplained '. 2
The determuung mfluencc was, I thmk, that of Ueberweg,
whose Logic and /Jistory of Philosophy "ere familiar companions
of Wllson 111 lus formative penod. In the preface to the second
edition of his Logic, that philosopher recommends his work to
the reader 'as a thorough attempt to reach a relatively obJective
theory of knowledge as agamst Kant's subjective cntlc1sm '.
No doubt the tendency had long been in the air, especially m
Germany. At Gbttmgen Lotze was the successor of Herbart.
But 1t 1s Ueberweg who, m his Logic, contmually msists on
1
Apart altogether from plulosophlcal debate, we can see this, I think, 1n
hlS paper of June 1890, • On some apparent anomalies 1n the use of 1'9°.
See Transactions of O:tford, Philological Socaety, 1889-90, pp. 27-8,
1
Cf Lotze, Logic, I§ 356-7.
 POSTSCRIPT 88t
a view which is, in essentials, identical with Wilson's. Ueberweg,
too, rests his theory upon the certainty of mathematiaal truth
and upon a conviction of the objectivity of our apprehensions,
Not that Wilson foJlowed him slavishly or literally, but he read
him early and adopted his method, partly historical, partly
critical, in the treatment of logical problems. Thus his general
attitude to Kant was probably affected by Ueberweg as well as
by his own English predilection for the scientific rather than
the philosophic attitude to Nature, What, however, weighed
with him more than anything else was his firm conviction of
our immediate apprehension of the truths of Euclidean geometry,
of the pure science of quantity and of mathematical physics,
especially optics and acoustics, and of their objective validity.
With this conviction went that other, that the method of
mathematics 1s not by process of mferrmg but by direct intui-
tion. These convictions explain his comparative neglect of
Descartes, Spmoza, and Le1bmz. Spmoza's method of exposi-
tion he thought fundamentally false, and he disagreed of course
with Leibmz's views of mathematical method. Thus there
remain for predommant mfluences Plato, Aristotle, and Kant,
especially the Prolegomena of the latter, with the author's con•
viction of the certamty of our direct mtuition of the truths of
mathematics as they were conceived by the Greeks.
This 1s not the place to review his philosophy, even were
I competent for the task. Excellent, if brief, summaries of
Wilson's doctrme have been given elsewhere,1 but it is best to
allow hun to speak for himself. He certamly tloes not disguise
h1s teachmg. Two clues to his thought, however, I have found
useful m studymg him. One is stated by himself m almost the
last words wh1t..h he wrote. They will be found m § 370, which
I have placed immediately after a passage of his very early
lectures to \\h1ch he was rcferrmg. His ob1ectiv1ty 1s there
declared to be the corrective of a one-sided sub1ectiv1sm from
which, under Kant's mfluence, he had started. 2
Wilson, hke his master T H. Green, 3 spent much tod over
the ordinary problem of perceptton and the theory of knowledge
1 T11st1mon1a, p lxxv.

• See Hegel, Logir; (the Encyclopaedia), • Third attitude of thought to

obJect1vity ', ch v.
• Green's Works, Preface to vol. 1 (Nettlesh1p) and vol 1i.
882 PO!TSClUPT
connected therewith. In the endeavour to refute the scepticism
of the English school of Locke's successors his argument recurs
perpetually to the presuppositions of Locke, Berkeley, Hume,
and Kant. His own view of reality comes out most clearly in
the letters on Primary and Secondary Qualities. 1 I do not think
he ever wavered from the faith which is there freely and fully
set out. His conviction stands there m a form which cannot be
doubted and 1s nowhere else so patent. He believes that we
have direct apprehension of the physical world as it was phtlo-
:;ophically conceived by Newton and his successors. Wilson,
however, appreciated m a way m which Green, not being
a mathematician, could not, the problem of knowledge in regard
Io pure mathematics This is the second clue to his logical
mqumes His central problem was not the quest10n of the
validity of our thought m regard to tlungs-the objects of the
senses. He came to think of a theory of Percept10n as philoso-
phically preposterous. His logical inqumes were largely guided
by the attempt to Justify our knowledge of the world-the
world of three-d1mens10nal space and of time as conceived by
the physicists of the eighteenth and early nineteenth centuries.
To this end his arguments on the subordinate question of what
he calls modern mathematico-plulosophical fiction really lead.
The elaborate crit1c1sm of part of Mr. B. Russell's work 8 is an
attack upon the opposed school of thmkmg and his d1ssatisfact1on
with the current Idealism really sprang as much from these
convictions of his as from his views of the presuppositions
of everyday experience and his teaching as to Perception and
theories thereof.
This brings me to a lc1,st pomt. Wilson was m extreme earnest
m his search for truth This 1t is wluch gives his wr1tmgs a tone
of almost religious zeal m regard to questions where en-
thusiasm 1s uncommon. This made his teaching so 1mpress1ve
to his listeners. He supported his convictions m a region from
which emotion 1s supposed to be excluded with an ardour which
never flagged and which fretted at every obstacle. This aspect
of his mind and character comes out very clearly m the splendid
fragment of a reconstruction of the Ontological proof of the
existence of God. 3 There the old rational argument gets a new
1
Part V, H 519-40 • H 422-32, 471-500, 501-18, • ff 565 seqq.
 POSTSCIUPT
turn from the introduction of the emotional element as part and
parcel of man's religious experience. The same unexpected
intrusion of human emotion into what appears at first sight
a purely intellectual problem may be found in a striking passage
in the inaugural lecture with which he entered upon the duties
of his chair. ' Before asking if it is true, we may ask whether
we should like it to be true. Such a question may seem un·
scientific, but m philosophy at least we should clear our reason•
mgs a good deal if we confessed to ourselves what we wanted to
believe and what we hoped was untrue.' 1 This 1s surely a daring
saymg for a professor of Logic and rs made more striking by rts
place m the midst of a merciless exposure of a superficial
philosophy.
Wilson had a moral repugnance to Scepticism in any form.
Herc, especially, he rcmmded one of Socrates. In the Meno,
Plato puts mto Socrates' mouth the uncompromismg reply to
a sophistical dilemma that, even 1£ the eris be difficulty 1s
verbally unanswerable, he 1s sure that we shall be better men 1f
we undertake our mqumes m the faith that they can be solved.
Thus behmd the Athenian's acute dialectic lay a background of
natural bchef which gave even to his abstract speculations
a moral fervour. So was it with Wilson. Acute though his
intellect was, 1t 1s always the servant of his conviction It was
the same in his moral philosophy, as, for example, m his belief
in the retnbutory basis of punishment. He is hke a preacher
with a text to expound, he knows the answer and exhausts
himself in the search for proofs This makes his negative and
critical work so powerful, renders lus positive 11ttcranccs some•
what dogmatic. Where he traces an error or a mistaken con•
victlon to its c;ourccs, as in his bnlhant treatment of Negation
or of Symbolic Logic, he 1s unrivalled Even where he seems to
fad, he throws more hght on the question than thinkers who
rest satisfied with the mere detection of error and eontrad1ct1on.
But 1t is time to say farewell to a task m which I have been
supported by the conv1ct1on of the advantage of these lectures
at least to earnest begmners and of the value of the rest of
the book to more advanced students. There is somethmg very
bracing in the way in which Wilson attacks his problems and
1
13S7•
884 POSTSCRIPT

judges even the immortals 'with all the freedom of an equal',

I close my labours with a heightened regard for that I warmth
and determination of spirit', to translate Francis Bacon's words,
'which aged not m the pursuit of its purpose and grew not cold
in so long a space of years'.
A. S. L. F.
UNI\'ERSITY CoLLEGE,
OXF0llD.
ao September x9a5.
 NOTE
IN I f)07 Wilson began a systematic rewriting of his lectures
on what is usually called the theory of the judgement. His
objects were partly to substitute a more correct terminology for
his old phrases, especially to get rid of the erroneous use of the
terms subject and predicate, partly to eliminate language which
might suggest a subjective view of the nature of apprehension
and the meaning of statement, and partly to work out more
fully the grammatical problems which he believed to underlie
the mistakes of the current logical doctrmes. Part II, Chapters
I-9, were thus largely rehandled.
In 1911-13 he reconsidered this work and the whole of his
lectures Parts I and II as far as the end of ch. 15. He was
working on chs 13-15 up to the end.
The result of these two rev1s1ons was the issue in printed
form of his Dzctata (that is substantially the present book to the
end of§ 123) The work was very much hurried, and he intended,
as he says, to reconsider and revise it all.
These printed pamphlets are of the following dates :
§§ 72-3, §§ 89-93, §§ 1:zz-3 • December 1912
§§ 78-88 and§§ 94-121 • May 1913
§§ 37-8, §§ 50-71, §§ 74-7 . December 1913
§§ 1-13, §§ 22-36, §§ 39-42 and 49 January 1914
The remainder of the sections mcluded in this group are
§§ I 4-2 I, "luch " ere intended to be recast and were never
printed, and certain sections which have been remodelled by
the editor, as ts shown below
During the same period he made many small verbal changes
and some additions to Part Ill, Chapters 1-3, but he never
reconsidered ch. 4, and 111 chs 5-7 mclm,1vc the alterations were
purely verbal and often very msuffic1cntly weighed.
To represent exactly the date of the various clements would
have involved great expense m printing and confusion to the
reader. I have therefore shown by tables the principal dates
as accurately as I can.
886 NOTE
By 1904, I indicate sections which, except for slight editing,
had reached their present form in 1904-5.
By 19071 I indicate sections which appear to have been
written between 1907-9
By 1904 and 07 1 I indicate sections completed in 1904-5, with
additions of the later date.
By 1910, &c., I indicate sections which seem to have been
composed in those years.
By square brackets I show sections which I have reconstructed.
It must be understood that all the sections contain re-handled
matter, often of very much earlier date ; the dates show the
period of their final rev1S1011.
In the text itself I have marked §§ 14-21 and §§ 70 and 71
to show additions (( )) made much later. Those sections will
give an idea of the manner 111 wluch the work grew and was
altered. In the first group the add1t1ons arc usually mere
additions, m the second they indicate a change of pl11losoph1cal
outlook A few sections (e. g § 42), as I have explained in my
notes, are of special interest as containing views which the
author himself had held and later abandoned.

SOURCES AND DATES OF THE SECTIONS

OF THE LOGIC LECTURES
50URCES CONSULTED AND COLLATED FUR THE TEXT
1. A Wilson MS , &c ·
(1) Early lectures, summary notes.
(11) 2 vols , Logic lectures, revision of an ca1her
course m MS
(111) MS note-books and folios. 1900-4
&,
(iv) Later MS note-book'> C 1907-14
(v) MS fohos for typist or printer 19n-13
('-1) Typescript of Part I, anti Part II-§ 123, ? 1910-1.2
basctl on I. F (1) and MS of a pupil (not
available) of ~ earher date
(vu) Typescnpt of Part III (cxcludmg chs 5-8), ? 1909
from MS of a pupil (not available) taken
from I D m substance
(vu1) Typescnpt of Part IV (exclutlmg ch. 5)-
same source as (vu).
(1x) Printed Pamphlets (Dictata)
 NOTE
Pupil's note-books, representing developments, between
dates shown, of Part I, Part II, 1-15, Part Ill, 1-4,
Pa.rt IV, 1-5.
B. (i) G F. Hill, Merton (a.nnotated by J.C. W).
(ii) J. Pedder, Oriel
(ui) W. A Craigie, Onel (annotated by J. C W.).
(iv) A. S. L Farquharson, University.
(v) C C Edgar, Onel
C. (1) W. R V Brade, B.N C {annotated by J C W 1896-1902
C 1898),
(ii) T F Cattley, Hertford Inference and Special
Logic only.
(iii) G B Cobb, New College. Inference and
Special Logic only
(1v) H J Creedy, St John's (annotated by J C. W ).
inference and Special Logic only
(v) P Neville Pnor, Umvers1ty (annotated c 1903-4)
D Anon (lMge MS corrections and additions, c 1907-8) 1904-6
E (1) K Lyon, Merton 1907-10
(11) Anon (1908-9)
(m) C D Carew Robmson (lender's notes typed)
F. (1) J Flynn, Pope's Hall
(u) 0 H T Rishbeth, Merton
(111) A M Tnstram, Worcester (last. lecture taken
14 June, 1914)
II. Part. II, chs 16-17
A H R P Gamon, Exeter (annotated by J C W) 1902
D (1) G F Hill, Merton 1889--90
(11) W A Craig1e, One! 1891-2
(m) A S L F,uquharson, U111vers1t.y. 1892-3
III. Part II, ch 18 MS note-book not later than 1go1-2.
IV Part. III, chs 5-7
A Author's MS note-books §§ 305-8 were never
delivered, probably for want of time
R (1) A J Jenkinson, Fellow of B NC. 1904
(u) W D Ross, Fellow of One! 1904
V. Part III, ch 8 Mam!y from II, B (1)
Of the above sources ·-
I A and D, II A and H (11), IV A and B (1), have all
been collated.
Remamder compared and where necessary collated.
The authority for the rest of the text is shown m the
notes below The Table of Contents 1s the Editor's
work
 CHRONOLOGICAL TABLE
Part I. Cb. 1. §§ 1-5, 1904; [§ 6 to 'provisional• 1904; 1912] § 7,
1904 ; [§ 8, 1904].
Cb. 2, §§ 9-10, 1912 ; §§ II-13, 1913.
Ch. 3. Except[§ 141 and§§ 22-3, 1912, all is of 1904 and (1907).
Ch. 4 [§ 24, 1912] [§ 25, 1907 and 12] §§ 26-8, 1907, but last,para.
1913. § 29, 1907 and 13, §§ 30-l, 1913
Part II. Ch. I. [§ 34, 1907] § 35, 1904; §§ 36-7, 1907, [§ 38, 1907],
§§ 39-40, 1912 [§ 41] [§ 42, 1912] § 43, 1912.
Cb 2 [§§ 44-7, 19n and 12]
Ch. 3. [§§ 48-g, 1913] [§§ 50-4, 19IIl,
Ch. 4 § 55, 19u, § 56, 1913 [§ 57, 1913] § 58, 1913, § 59, 19u.
Ch 5. §§ 6o-3, 19u
Ch 6 § 64, 19u, § 6,; 'If we assume-at all ', 1913, remr. 1911.
§§ 66-7, 19II, [§ 68, 19IIl
Ch. 7 § 6g, 19n, §~ 70-1, 1910 and (1913), § 72, 1907, § 73, 19u.
Ch. 8 [§ 74, 1907, 1912] ~§ 75-7, 19u (and 13), § 78, 1910.
Ch 9 § 79, 1912 [' Now we often nom,•e to the verb '] § So,
1912, last para 1913, § 81, 1912, last para 1914. § 82, 1912,
§ 83, 19xz, except' Finally' to end, 1913. Note to§ 79, 1907.
Ch. IO [§§ 84-5, 19u] §§ 86-8, 1907, §§ 89-93, 1904, Note 1907,
§ 94, 1910, except ' In the above ' to end, 1912
Ch II. 1904 except §§ 102-3, 1qo7, § 104, 1901-2
Ch 12 § 106, 1907 [Hlstoncal d1gress1on, 1891] § 107, 1904, except
'What 1s charactenstic' to end, 1913 § 108, 1904, except
'It seems then '-end, 1g13 § 109, 1913, § no, 1907, § III,
1904, but pen para 1907, ult para 1912 § n2, 1907, ult.
pa.ra. 1912, §§ 1q-14, 1904, ult para 1907, § n5, 1904, except
• We have been cons1dermg '-end, 1907 § II6, 1913, § n7,
1907.
Ch 13 §§ nS-19, 1904 and 07, ult para. 1912 §§ 120-1, 1907,
§§ 122-3, 1904 §~ 124-5, MS C 1904
Ch 14 [§ 126-' meamng them'] Rem §§ 126-37 and §§ 140-1, re-
vised and largely rewritten 1914, but § 141, 1907 §§ 138-g,
MS C 1904
Ch 15 § 142 and§§ 144-7, 1907, § 143 and§§ 148--50, March 1908,
Chs. 16-17 1902, see II A and B, above
[Ch 18] c 1900, see III, above, §§ 206-7, 1906
Part III. Ch 1 §§ 208-9, 1907 [§ 210, 1907, 1913] § 2n, 1907 • But
imagination '-end, 1904 §§ 212-15, 1904 and 08, §§ 216-18,
1904, except • In the facts ... equal to B, the same bolds ',
1912. §§ 219-27, 1907
 CHRONOLOGICAL TABLE 88g
Ch, 2. § 228, 1904, § 229, 1904 except foot-note, 1912. §§ 230-3,
1904, § 234, 1907, § 235, 1904, § 236, 1904, except' If 1t were
abstracted• to end of section, 1907. [§ 237, 1904, 07, and 14]
[§ 238, 1914] § 239, 1904, § 240, 1904, except ' It would usually
, • , clearly then• and note, 1912, and' The foregoing '-end,
1914, § 241, 1904,
Ch. 3. §§ 242-3, 1904 and 07, § 244, 1907, § 245, 1904, §§ 246-7,
1907, § 248, 1908, to' confined to itself'. Rem. 1914. §§ 249-
50, 1907, §§ 251--61, 1904.
Ch. 4 § 262, 1904, § 263-end, 1907 All Part III, chs. 1-4, was
worked over very slightly m 1913-14.
Chs 5-7. 1904, see IV, above
Ch 8 1890, see V, above,§§ 325-7, from Nat11re, vol 63, No 1624,
13 Dec 1900
Part IV. Ch. 1 § 328, 1910 [last para 1899 J §§ 329-31, 1904, note
to § 329, 1899 {from I C 11). § 332, 1904, except 'Tlus also
explams '-end, 1909 § 333, 1904, foot-note 1909, § 334, 1904
and 09, §§ 335-7, 1904, ~ 338, 1904, except 'We have seen
that'-' change with A', 'and m consequence•-• connexion',
and' It is true'-' understand', 1909.
Ch 2. §§ 339-40, 1890-1, ~§ 341-5, 1890-1, collated with 1898,
I C (n).
Ch. 3 §§ 346-50, 1904, § 351, 1904, except' If we assume .. an
1llus1on ', 1909 § 352, 1904, § 353, 1904, except 'The com-
pletion of the cause abed as an orgamc complex ' and ' ancl
is necessary '-end, 1909
Ch 4 § 354, 1890-1, §§ 355-64, 1889, § 365, 1891, § 366, 1914.
Ch 5. §§ 3t>7-9, 1904, § 370, 1904, except' and the ordinary theory
, . retma ', and last para , 1909
Ch 6 §§ 371-400 MS of two pubhc lectures of 1898
Part V. [§§ 401-5, §§ 406-14, §§ 438-50, MS of relatively early date.
[§§ 415-21, §§ 422-9, §§ 430-2, MS 1912-14, intended for Bntish
Academy.]
§§ 433-7 MS c 1904, collated with two pupils' MS of l.:i.ter date.
§§ 451-g MS. C 1913-14.
§§ 460-564 MS correspondence, dated m text.
§§ 565-81. Typed notes for address, 1899, with later additions,
dated m text., collated with notes {of a. portion) by S C. P,
denved from onginal MS
§ 582. Letter, I9II
 INDEX TO PARTS I-IV
The references are, unless othel'WlSe stated, to the pages.
Abelard, 213 279, as subJect of logic, 18, 59, 78
Absolute Idealism, 6o, 63. seqq , and Judgement, 85-6, 93,
Abstract and Concrete, 170-1, 389- 278-9, 295, that on which uruty
90 of different activities of thuikmg
Abstract nouns)(ad1ecbves, 173-5, depends, 97, and statement, 196--
206--8, 399 ; )( concrete nouns, 188- 7, 218, 313-17; and knowledge, 15,
91, 206-8, 399-400 40, 97, 279, )( conception, 309-10,
Abstract1011, true nature of, 28, 42 , 317 , ) ( perception or experience,
ddlerent klDds of, bo, and de- 508, 509, 514-18, m relation to
finition, 378--g ; where 1mposs1ble, inference, 427-g, 431 , of sensa-
256 tion, 313-14, 329, of universals,
Accent, vid Strebs Accent 35, ,9, 42-,, 4S, 67, 292-3, 307,
Accident, Ill Aristotle, 159, 16o, 16 I, 316--7, 456--62, 482-~. 513-18, of
164 negation, 255, 256, 201 seqq, 271-
Activity, of consciousness, 37, 81, 2, i7 3 • .i.lways of complex, 312,
none of ' Judgement ·• 8 5-7 ,14, ,i7--9, hard to &eparate from
Adjective, u'!e ID language, 1 54, 160, its ob1cct, 67, 70, 72, does not
171-2, 174-5, 179, 181, 184-5, 187, mclude obJect m itself, 73-5, 1s not
189,190,198, 204-5, 211,219,390, exph1-able m terms of obJect, 'I 5-6 ,
396-7 , logical symbols for, 199, Kant'i, mJStake 111 treatment oI,
205 6 10
Affirmative, vid Categoneal A priori, logic 1n what sense, 49, 59 ,
Affirmation and Negation, part II, method, 29 , forms of thought,
ch I.l 49, so , sciences, 29, w , classlfica-
Agreement, method of, 584, 587 t1on, 368, 372-3. syllog1stlc logic,
(cf 594-5) 436-;
Algebra, science of pure quantity Anstotle, founder of logic, 30 ,
27, meanmg ID 289, universal Analytics (name for logic), 31, 32,
and particular m, 477-8, equa- 49; nowhere defines logic, 30-1 ,
tions, 219-20, ab applied m • sym- 'b Organon, its contents, 31 ;
bolic logic', 6~9 seqq. 651-J, Categories, I S9, ItJ6, 222, 223, 37 I ,
1magmary quantities m, 268-71, De Jnte,p,etatwne, 1 34, 177 .113,
651 1.97 , Prior Analytics, q l, 223,
All, m u111versal &tatement, 447-8. 450, 45,, Posterior Analytics, 133,
Analogy, 3b4 159, 1fio, 381, 4 n n , 471 ; ropics,
Analys1S) ( bynthes1s, 56o, m log1c, 49 IJl, 38o, Physics, 55, De Anim•,
Analytic geometry, c,40 1.97, Metaphysics, 55, 134, 159,
Analytic statement, ,q1-5 aflimty 214, 16o, Nicomachean Ethics,
to, of negative &tatement, 27 2 450 n , on Socrates, 132, view of
Analytics, v1d Aristotle statement as analysable mto simpler
Anaxagoras, on space, .i48 thoughts, 8~, .i97-8, Jl4-I5, of
Antecedent, in hypothetical state- pred1c.i.tion, 114, 129, 131, 147-8,
ments, 242, m induction, 588 159-66, 177,213, of ,lva,, 181 n ;
of sub1ect and substance, 1 52, I 56,
A.;:::l!ct1c statement, 224 seqq 157, 159 seqq , of essence, 38o-1,
A postenon )( a p,,_iori, 368, 372-3 , 471 , of property, 380-1, 471 ;
for ' Evolut10111sts ', 61 7 of accident, 159, 16o, 163, 164;
Apphed logic, 57-8 , mathematics, of definition, 377-81, 453, 471,
57-8, 631. 47 3, 6o2 n ; of genera and species,
Apprehension, in general, part II, .?22-3, 36o, 367, 371, 376, 377, of
ch. I ; ultimate and mdefinable, relations, 71 , perception 1s of uni-
I 1
 INDEX
versa!, 336 ; view of inference, 433, Met 1040• II 340 n.a
445, 450 ; of syllog1Sm, I 29 seqq , ,. 1051• 22 134 n.a, 455 n.1
4S3 &eq(J, , defimtion of deduction, 1051b 3, II 66 n a, 286 n a,
578 ; VJ.ew of induction, 581, 582 , 214 n.1
of axioms, 581-3; of hypothetical 1
,. 1055 28 364 n.a
statements, S49 ~q : ol pure and ,. 1078b 23 132 D,:Z
applied mathematics, 631 , failure EN 1130'" 12 65 n.a
to d1Sbngu1sh obJective and sub- ,. rr3gb 28-31 $82 D,I
jective, 446 n .. u47• 6, 24 25 n b, 450 n.1
Aristotle, passages cited or quoted " II73• I 626
(References to Bekker, quarto Article, mdefimte, 399-400.
ed) Assertive word-forms, 176-8, 'IS3-5.
Cal. 2b IS 100 n I Assertor1c statement, 224-5.
., 3•36 ,. .. 2 Association of ideas, not irrelevant
,. 6&17 366n I to logic, 56-7, )(impbc1tthmlung.
De Int 16& 3-4, 12-14 286, 297 n I, 303-4, 1n Hume, 61t1; ID Bradley,
314 n a 294. theory of, criticized, 112,513.
16b6-11,19 177n1 Assumption, 553-4
16b24 181n1&213n2 Attnbute, as baslS of defimbon, 377 ;
170.2,8 298n1 metaphysical, 153-61, 178-9, 200,
,. 17•2s 81n1 508 , how expressed in language,
.. I ga 33 286 n a 166, 171-2, 173, 174, 175, 177-8,
An h I 7 477 n 2 & 453 n 1 18o-1, 186-go, 394 seqq., 1n rela.•
24• 16 8J DI tion to meamng, 391 seqq
u 64b 34-6 582 D 2 ' Attribute-element ', I 53-4, I 56-7,
.. 67• 33 450 n .2 171, 172, 177, 185, 187 n.
A11 Po I 1 31 n 2 'Attributive verbs', 172-3, 185
II 8 478 n 1 & 453 n .:i Axioms, 458, 465-6, 490, 580-3, 616
71b 5 .25 n b seqq
72•7,b18 84nb
77b 5 6.?8 DI Bacon, Francis, view of mduction,
83"'1 1oon3&162n1 259, 581, 593, 597-605
83 14, 17, 18
11 16o n 7,8,9 Bemg, as subJect of metaphysics, SS;
83• 30 159 no m general, 214-15, 216, 2;q, )( not•
87b28 355n1 being, z47, 248, .263, 267,268
gob JO 377 n 1 & 31:!o n 1 Belief, wrongly mcluded under'Judge-
97b 7 378 DI ment ', 98 seqq , )( knowledge,
97b 24 ,. D 2 100, 101, 109 , )( opllllOD, 101 ;
.. 100• I 5-16 4S n a, 340 n I, does not adDUt of ' error ', 106 ,
355 DI vid Opm1on
10ub 3-15 582 n J Berkeley, his idealism, 34, 6o-2, 7 5.
Top' 100b 18 582 n 2 B1ology, refutation of attempt to
.. 103 15 8 .>,77 n 2 make 1t basis of all knowledge, 627
., 103b 7-19 38o n 2 Boeth1us, on subJect and predicate,
., 143• 36 367 n 2 114-15
Phys 185b 27 217 n I Boole, Investigation of the Laws of
.. 19211 36
De At1. I 1
55 n., Thought, 636, 644, criticism of,
31 n4 496, 66o-1
424• 5 05 na Bosanquet, B , 63 n
.. 430• 27 J 14 n a, 297 n 2 Bradley, F H, hinciples of Logac,
De Msm 449b 31 37 na 166, 325 , theory ol Judgement,
., 450a I 43 na 166-9, 2~4, on association of
De Part Anim. 642b 21 307 DI ideas, 325-6, criticized, 255 n
Mt1t 995"'28 155 na Browning, ll, Sordello quoted, 40.
., 1024•26 377 D 2
,. 1024b 8 36o D,I Calculus, logical, 635 seqq Vil.
,. 1024b 32 211 D 2 Probability
1028b 36 159n7 Carroll, LewlB, vul Dodgson
,. 102g&8 , ... 8 Cartesiamsm, innate ideas 1n, 628-9.
,. J0J7b 29 380 DI, 377 D 2 Categoncal and hypothetical state,.
., 1038• 5 36o D,I ment, 235-46, 492-3, 644, 651.
 INDEX
Categories, metaphysical, 150. Connotation and DenotatJOQ of
Causation, general nature of, ~16, Terms, part II, ch. 18; VIII Int.en•
517; presuJ>P(llled m induction, SlOn,
581, 007-11 ; "Mill's view of, 61>7-8 Consc10usneas, s11i generis, 38 ; ~
Cause, studied by metaphysics, 59 : of knOW1!11J, 37-8 ; suppoeed be-
' conception ' of, 309 ; not em- ginning of, m sensation, 329 ; ume,.
pmcally given, 515, 517; u cogn1zed forms of, 42, of objects,
found bf mduct10n, 588-95, 597 , 6a-3, 74-5.
Mill's v1ew of, 590-1 , Hume's, Consequent, in hypothetical state•
616; 'plurality" of causes,' 245, ments, 242 ; m 1nduct1on, 588
587-8, 591, 592, 611-1 5 , cause )( seqq
preml!IS, 53 • causal )( cond1tJonal Construction, mathematical, 455,
statement, 494 , v,d Formal Cause. 459, 465, 466, 474, 633.
Chance, 111d Probability. Content and Form, 63 n , vsd. Form
Charge, not absolute, 178, 592 and Matter
CU'Cle, one of the come sections, 26 Content and ObJect, 63, 64, 66, 70-r,
Class, two lnnds of, 361 , d1vJS1on of, 75, 306-7, in Bradley, .z8o, 28r,
,16.i-4, 367-8, JD ' symbohc logic', 285,287,289, 293-4
641 seqq. Contradiction, principle of, not to be
ClassJlicatlon, part 11, ch 16 , diffi- treated as cipec1al premiss, SS7;
culty of, 27 , relation to definition, not to be established by an appeal
29, 38J-3. depends on progress of to evolution, 6.26
sciences, 30. a priori, 368, ;72-3 Conversion, of hypothet1cals, 245,
Coherence, of theorems, 27, coher- 613-14, of sc1enblic propositiou,
ence theory of knowledge, cntlc1Sm 245, 384, 484
of, 451 Co-ordmate geometry, 30, 271 n.
Colours, why called by the same Copula, 114, u6, 117, 181, 184, 2u
name, 19 seqq. 223 seqq, 274. vid Verb' to
Command, logical position of, .i.:8-- be'
30 Copymg U1eory of perception, 61,
Common n.tmcs, 188-91, 19;, 202-4, and thought, 64, 75-b, 285--6, 287,
:::06-8, 2z1-.:, 389, w9-400 307
Comparison, charactenstlc of thmk- Corre~pondence, between 1deu and
ing, 35 , mvolved m percept10n reality, 286
&c, 46, 47, Cntenon, of truth, 107, 424-5, 6::i-3,
Complete enumeration, v,d E11u- u.?8--<;J
me,at,o Simplex, Cymcs, said to have made all Judge-
Complexity, nature of, 502-8 ments identical, 217.
Concept)( conception, 305, ;10
Conception, ordinary meaning of, Decision, proper meanmg of ' Judge-
100-10 ; ' theory of ', go , )( ment ', 92-; • practical, 100-3.
Judgement, 295-319, reduction of Deduction)( mduction, 578-So, 1m-
Judgement to combmation of, 274- phed m mduct1on, 58.z, 585, 589.
9, 295-9, may be true or false, Deducttve sciences, 57-8, 59.
;04 , supposed conceptions of Definition, m general, 24, 41 seqq
1mposs1ble-, 2(17, 468-9. 'pro- part II, ch 17, 4fJ7, ideal of, 29.
blematic' conr:epbon~. 267, 318, =search for umversal, 29 , not
509-rn, 5z.z-3, 525 , &1mple )( com- merely' analytic•. 234, and pro-
plex, 1ie,t11er capable of bcmg perty, .z6, 383-4, 470-3 , m geo-
made ', 31 g---20, 326, 329, 467-8, metry, 25, 40, 382, 383, 466-72,
497-524, 111rl Ideas m empirical sciences, 46g-72,
Concomitant variations, method of, errors m, 249, 257-8, of logic, 24,
591-3 (cf 594-5) 28, 3.z, 65-6, not found 1D Ans-
Concrete and Abstract, 170-1 , virl totle, 30-31, provJS1onal, 32; none
Universal (verbal expression of) of thinking and k110W1ng, 38-9 ; of
Cond1tion, expression of, 235-6, 239, subJect-predicate, u4 seqq, 127 ,
244; relation to conditioned, 611- of Judgement, 249, 274~
15. Degree, differences in, 369-71 ; 1n
Conditional statement, viii hypo• importance of practical decwon on
thebcal ; )( causal statement, 494. '?P.in1on &c • 101, 102
Come sections, analogy between, 26 Dehberatmg, a kind of tlunking, 37;
Ii2
894 INDEX
cf, opining and wondering, Ji' 314: F. H. Bradley's, 285 ; Hume's;
and judgement, 86. su; asllOCl&UOD ohdea.s, su--13;
Democntus, asserts void, 248 empmcal, 616-:118 : cartesi.all,
Demonstration, helped I>)' umty of 6:118-9 ; LeibmzlaD, 628 , Kantian,
aclence, 26 • bm1ts of. 84 ; )( 629 ; the author's, 629-30. YNI,
u10D11, 465-6 : vitl. Inference. Berkeley, Locke, &c.
Denotation and Connotation, 176, Equat1on, identity how far implied
part II, ch 18 ; denotative ex- 111, 219-20, 1n 'symbolic logic',
pressions, 176, 178: ind. 357, 643 6,p-62
Dtii!cartes, his theory of knowledge, Error, problem of, 104-1~: 1mpbes
628; and co-ordinate geometry, d1St111ct1on between thinking and
JO, obJect, 68 : does not belonc to
Dialectic, Socratic, 132. mental pictures, 276, nor to
Dichotomy, 367, 650, 658-9 opbuon, 105, 108, 109,
Dielam, De omni et nullo, 444. Essence, lJ1 Anstotle, I 59, 380--1 ,
Difference, method of, 588 (cf 594- and defin1t1on, 38o-2, 384--5, 470--3,
S), perceJ?t1on of, 255 , statement primary and .secondary essence&,
of, .zoo, 111d Negative 223, 371, 376. •
Di/ferenha, ,ml. Genus and Specres Ethics, relation to metaphysics, 55;
Different1at1on of genus, 29, 356-8, v 1tl J ustlce
361-4. Eucbd, vid Elements and Geometry
D1menS1ons, spatial, 567-8 Event, d1St meaning of thought, 51 :
Direct perception of obJects, 61-2, studied by psychology, S3
D1s1unct1ve statement, 494, S4t'.>-9, Evolutionist school of philosophy,
658-9 t116-28
D1str1bubve law regardmg t,la55 con- Existence, of geometrical figures,
ceptions, 649 467-8
D1v1S1on, 11 jwiori )( actual d1ffercn- Existential statements, 166, 181 : 111
babon, 29 , log1<,al theory of, ' symbohc logic ', 645, 647 , as.set'•
361-4, 367-8 tmns of non-ex1Stence, 2 59
Dodgson, Rev, C L., 635, 635 n, 638, Experience, not coextellSive with
639 apprehension, 513-18, 1n a sense
Doubt, presuppobition of a1:>Sert1on, not source of all 1dea.&, 512--15. as
183. means of acqu1nng knowledge, 81,
Dreams, 320 95, 512, )( inference, 413-17, 482.
Experiment, method of, 6o6 seqq ,
Element, as du,tmct from attribute, part IV, ch 1
199 , • attribute-element •, 1 53-4. Explanation, usual se11Se of, 38, 361
156--i', li'l, lj'2, 177, 185, 187 n Extension of Terms, 35;,, 643, vatl
Elements of Euclid, 25, 27, 455, 458, Denotation
465,475,485, 486-7, 556, 500, 561 •
bk V not geometrical, 27, Si' Fallacy, W, 254, 255 n, 3S4, 436,438,
Ehnunation (cf, Abstraction), general 447, 404 ; vid. Error, Question
meamng of, 645--t> ; Ill ' symbohc Feeling, 1f apprehended ID know•
logic ', 646 , the method of m- ledge, 35 , )( thmktng, 35, 42, 45 :
duction, 594-5 • ' ehmmabve argu- hard to separate from what 1s felt,
ments•, 259 67-8 , continuity of changes ID,
Empedocles, on 'lpace, 248 370
Empmcal, sciences, 29, 368-9, 456, Fictions, 497, 511, 519, 521--4, 525,
469-72, S95 , theory of reality, 62 534-6
EmpmcJSm, refutation of, 417, 513- Figures, of Syllog1sm,first, 132, 430--r,
14, 515-16, 006-11, 6to-J8 4S3 : third, 430,453 ; ID mathema-
Enumn-atio Simplex, induction by, tics, 455 seqq
570, 581-2, 583, 6o6, 6o8 Force, 111 science, 632-3
EnunCJation, bJStoncal .starting-pomt Form and Matter, 05, 66, 238, in
of logic, 81 Bacon, 6o1-5; of propos1bons, 435,
Ep1Stemology, )( Logic, 56, 6o-75. 442
varieties of, do not affect logical Forms of expression, via. Grammar
dlSbnction of the ' apprehended • and Parts of Speech,
and ' appreheDS10n ', 78 , appre· Forms of knowledge, .. forms of
hensionsand the apprehended, 277, obJects, 149 : hypothetical state•
 INDEX Sgs
meat 8lllel't8 something about 321: on ideas, 276, su: WTtreaOlllt-
form of statement, 545, view of imagination, 320 ;
Forms of logic, in a seme • priori, ment of cause, 616.
49-50 ; )( metaphysical concep- Hypothetical statement, general
t1ons, 91. nature of, 235-46, 491-8: JlOt
Forms of thought, logic not study of, relation between two statements,
• 33, 65-6. 495-6, 527-JC) : •assertion that
Formal Cause, 5g8-6o5. , one problem includes another, 537-
Formal Logic, 58, 4! 5 ; vul. Syllogism. 46, sso-2: alwar: an inference,
Funda.mm"4m DJvi,io,ns, 364. 536: )( categonca statement, 235-
46, 492-3, 644, 651 , must not be
General logic, 58-9 reduced to non-hypothetical form,
GenerabzatJ.on, in science not logical, 549-50 , and problematic con-
26: not due to search for defini- cept1ons, 526-7 ; )( fiction, 497,
tion, 26 534-6, 542, )( supposition, 529-31,
Genus, relation to species, 29, 38, 533-6; redumo ad absurdum re-
ducible to, 554-9, 56o , converti•
253, 3.lS-6, 356-61, 502-3; m bihty of, 245, 613-14, in 'sym•
Anstotle, 222-3: how expressed bohc logic •, 646 , )( conditional,
m language, 207 491-2 (cf. 530)
Geometry, science of space, 27;
Anstotle on, 631 ; )( general Idea, in Locke, 34 n , 61, 497-524 ;
theory of quantity, 27, 57. as in Hume, Z76, Su ; in Hobbes,
11 priori, 29, definitions in, 25, 40,
382, 383, 466-8, 469-70, 471-2,
276, tn Bradley, 28o-8: a.,
psychical image ) ( content or mean-
paracular statements m, 331-2: ing, 275, 28o-8, 292-4: in theory
postulates 1n, 467-8 , inference m, of knowledge, 61 seqq , 75-6, 112,
381, 439-40, 442-3, 475-6, 483, 282-8, 506 ; simple )( complex,
485-7, retluctio ail absurtlum proof, 319-20, 326, 329, 497-524; can
556-7, SS9-6o; constructions m, ideas be •made' ), 497-8, ;u-12,
455,459,465,466,474,6~3, sen114;
m which 1t involves makmg s18-24 ; are all derived from ex-
perience), 51 z-18, 616 seqq, 628-g
by as, 522 ; not concerned pn• Ideas, association of, not irrelevant to
manly with precbcabon, 1 3 I ; not logtc, 56-7, )( impllcit thmking,
syllogistic, 131, 4S4 seqq , how 303-4, 1n Hume, 616, m Bradley,
far non-metaphysical, 56 ; analy- 294, h1& theory cnticized, 513
tic geometry, 640, • 1magmary ldeabsm, relation to definition of
points• m, 271 , supposed non- logic, 24, 6o seqq , 66, 1 50 , 1denti-
Eucbdean, 561-8 , applied)( pure, ficabon of thought and object not
57-8, 632 , used to refute equation pecuhar to, 64: baslS of, 7S ,
of 1mposs1bles, 651-2 absolqte, 6o, 63 , subJective, 6o,
Grammar, d1&t. Logic, 50, S1 : rela• 63
tlon to Logic, 27, 83, 88, 400-1 , Identity, 1mphro m change, 178 ,
not merely verbal, also studies between sub1ect and predicate,
thought, 50 , in Ar1&totle's Logic, 217-18, 219-20, of universal m
1 ~4 , analys1& m, 48 • treatment
of subJect and predicate m, 119 particulars, 34r, 344
Idiom, vitl Usage
seqq, 124, I.!8, 130, 143-4, 148 i&ov, 111d Pl'Operty
Green, T H , his 1deahsm, 45 n , 6o ; Images, mental, 27 S, )27 ; )( sensa-
view of relations, 71 n tions, 320-5 , )( conceptions ·,
300, 302; in Bradley, 280-7, 290,
Habit of language, v,tl Usage. 292, 293.
Hegel, use of term ' applied logic •, Imaginary quantities, 268-71, 651.
57 n, and •content•, 63, on Imagination,)( th1nkmg, 37, ,S34;
identical Judgements, 217. )( reality, 320, 520-1 : essential to
Hobbes, on ideas, 34 n , 276: on sciences, 415 , ) ( hypothetical
Judgement, 92 n. thinking, 543-4.
Hume, sub1ective idealism of, 6o : Immediate, inference)( mediate, 417-
ep11temology of, 512 : confuses 31, 449
tfunlung with series of mental Impersonal statement, 228-30.
pictures, 276, on impreuions, Impbcation, )( meamng, 387, 393-6.
8g6 INDEX
Impouib1bty, nature of, 270; treat- the act which refers an ideal con•
ment in mathematics, 271, 652; tent to a reality beyond the act,
in • symbolic logic •, 647-59. :iSo-94: )( conception, 300-19.
lmprelllOD (sense-), 61, 321 ; • being Justice, process of defining, 28, 41,
under an impre11111on that ', 113 43, 44, 132.
Inconceivability of opposite as test of
truth, 424-5,621-3,628-9 Kant, nature of bis • ~ '
Ind1catJve mood, 176, 177. revolution, 629-30, cntic11m of,
Ind1v1dual, umty of, I 56 n 630 , h11 divwon of logic, 57 n. ;
Induction, )( deduction, 578-So, but on cnterion of truth, 107 n : on
really implies it, 582, 585, 589 , in error, u 3 n. ; tries to find forms
science, part IV, ch. r (cf 570-1) ; of thought by analysis of \.erbal
ID Aristotle, 580-3, 1n Bacon, 259, fonn of statement, 3u, on
581 seqq, 59~. 597-605 ; cntic1sm mathematical knowledge, 456.
of modern theories of (especially KaT.iAf/if,ir, 78 n ; vid. Apprehension.
Mill's), 580-2, 583--95, 6o6-I I ; and Kind (genus) and Quality, 36g :
pnnc1ple of causality, 6o6-r S • by differences m kmd ) ( differences in
ehmmat1on, 259, 585, 594-5, 1m• degree, 369--71 , Mill's doctrine of,
phc1t, 303. 373-6
Inference, meaning of term, 412-16, Knowing, species of consciousness,
429-30 ; IS 1t mam SllbJCCt of 33, 38 , sui generis, 38-9 : )( think-
logic?, 32, 81 ; process of =activity ing, part I, ch 2, 34 seqq., 79, 105-
01 knowing, 35, m relation to 6; and statement, 313-14: and
Judgement, 84 seqq , 89--9o, 426-7, conception, 31 3 : potential and
430 ; act of, 449 seqq , )( experi- actual, 42-3, knowing that we
ence, 413-17, 482, 483 , )( non- know, 107: bow far possible
inferred knowledge, 481--go; subJect of inquiry, 39-40
novelty of conclusion 1n, 416-17, Knowledge, theory of. vt4. Epistemo-
425-30, 431, 446-7, as subJect1ve, logy, )( ob1ect known, 61, 64,
S4 ; not nghtly viewed as con- 69 seqq , and ideas, part I, ch 4 ,
nexion of Judgements, 432 , hypo- )( 0p1DlOD, 90, g6, 97, 99, 100, 103 ,
thetical statement 1s an, 243, )( behef, 100, 101, 103: )( infer-
536, syllog1stic, 57-8, 419-21, ence, 35, 84 seqq., 89-90, 413-14;
430-r, part III, ch 2 , in sciences, bow related to modality, 225,
S7--9 , dependence on universals, vid Apprehension, Judgement.
43~, 414 n, 448, 479-81 • poss1-
b1hty of error m, 108 , 1mmed1ate
)( mediate, 417-;1. 449 Language. study of, 50 , normal use
Infinite regress. 444 of, vid Usage.
Innate ideas, 628-g Law, of nature, 227 : Laws of
Intension, 357,643; vid Connotation thought, m1Suse of term, 33
InterJectJon, logical sign11iance of, Leibmz, 49 , reply to Locke, 628.
229 Leucippus, asserts void, 248.
Inverse ratio, of extension and mten- Life, m biology, 41
91on, H7 Lmnaeus, on classification, 371.
Locke, real11m of, 61, 62 , use of
Judgement, as treated in modem 'ideas', 34n, 61, 112n, 276,
logic, 84 seqq • part II, ch 2, 295 296-7 , on Judgement, 296 , ' Sim•
seqq , true nature of, 92-6, 108, pie ideas', 312, 497-524: on
;01, 491 n , wrongly supposed mnate ideas, 628 ; hlS empmcism,
common activity, 86--7, 9:z-3, 98--9, 417 , compared to Ar1Stotle, 71 :
relation to opm1on and behef, 100- on substance, 152, r 56, 500 : on
2, 104 , never false, 104-8 : )( pro- modes, 499-500, 508 : on relations,
J)Olllbon, 94; )( inference, 84 seqq, 71, 499 , on opinion, g6 n
Sg--go, 426-7. Bradley's theory of Logic, ongm of term, 32 , aims at
cnticized, 166--g, 28o-g4; cate- truth, 53, implicit to man, 49;
gorical )( hypothetical Judgements, method of, so, 11 pno,s, in what
235-46, 492-3, 644,651 , erroneous sense, 49, 59 : not_ defined br.
defirutions of, (a) as ' affirmmg or Anstotle, 30 : defi.mtion of, part ,
denying ', 249 : (b) as combina- ch 1, 32, 65-6 : defined u atudy of
tion of ideas, 274-9, 295-9, (c) as thought, 24, 1 50, as study of forms
 INDEX
of &(>prehension of objects, 59 : of natute of objects, 59: concemed
)( Epistemology, 56, 6o-75 : )( with Reality, 55: ongin of name,
Metaphysics, 56 ; )( Grammar, 50 SS; relation to logic, 27, 82, 88,
(cf 124) : )( Psychology, 51-4; Sg, 1 so : to sciences, s5 , modern
)( other branches of plitl010phy, tendenciea of, 71 ; studies cause, 59,
27 : relation to cognate subJects, Method (vttl. Induction, Mill), de-
27 ; to the sciences, 24-6, 28-30, pendent on object of science, 48,
57-8, 6o, 65, 78, 79, 636; )( mathe• 59 ; of logic, 48-g
matics, 635 seqq : how far con• Middle term (v,d. Inference, Syllo-
cerned with time, 52-4 ; applied)( g11m), 135, 141-2, 144, 145; und11•
pure, S7, a false distinction, 58 ; tributed, 436.
general)( s~al, 59; 'symbohc Mill, J S , on logic, 3a , on definition,
l~c ', part IV, ch, 6 467-8 : on connotation and de-
Logical calculus, v1d Boole, Venn. notation, 386-7, 390,403: on pro-
Logical subJect, 123-6, 130, 148, 158, positions, 2;4; on real kinds, 373-
100, 167-8, 174, 184, 194-7, 221, 6 , view that inference 1s from par•
226, 229, 240-r, 26o. bculars cntic1zed, 479-81 ; on
Lotze, use of term ' apprehension ', induction, 416, 588 seqq, 6o6,
78 n , on Judgement, 296. critlc1sm of Spencer, 622 ; on
mathematics, 514, 612-3.
Major, vid. Premiss Mmd, treated hke a • thing • ID
Mansel, Prolegome11a Log1ca, 64 n , language, 1 56
view of judgement, 92 n. Mmor, premiss and term (vid Infer-
Matenal inference, 41q-21 ence, Syllogism)
Mathematics, nature of, ,o , )( logic, Modality, 223-9
635seqq , 1mportancefortheoryof Mode, m Locke, 499, 508
inference, 59,454 seqq , neglected Moods, vid Syllogism
by ArlStotle in developing his logic, Motion, laws of, 632.
131, 133, definitions m, 466-8,
proofs in, 244-5, 381, 454 seqq , Name, meaning of, 40 seqq, 131,
reductio atJ absurdum, 554-00 , 357 seqq , ,88 seqq
classification 1n, 366--7, 368, 370 , Necessary, d1sbnJullhable from possi•
mductlon m, 593, 6o8 , equations ble not obJectively, 224, but sub-
m, 219-20, mathematical calcula• 1ectively, 225; necessary con-
tion of probabihty, 57 3--6 ; mathe- nex10n, 372, 585, 593-5.
matical theory of logic, part IV, Necessity, obJective, 432, 516, 517,
ch. 6 ; supposed non-Euclidean 630
SP.ace, 561-8 , 1magmary quan• Negative statements, 93, 115, 216,
tities, 268-71 , modern mathe• part II, ch u, 424
matical theories of number, 352-3 ; v62air, 78 n
pure )( applied mathematics, 57-8, vov,, 581 n , 583.
6,1-4 Vid Algebra, Geometry, Nommalism, m ' symbolic lc,gic ',
Come section 643
Matter, vid Form Nouns, 1n language, 50, 166, 170, 172,
Meaning, of names, 40 seqq, 151, 173, 174-5, 187, 190, 191, 20I,
388 seqq , 397 scqq , logic studies, 204-5, l l 5-17 , logicalsymbols for,
53 seqq, 88, du,t event, 51 seq_q , 199, 204, proper and common,
)( hkeness, 275,286: )( expression, 202-4, 389, 404 . ab-ltract and
141 ; in Bradley, 280-9, 291-1 , m concrete, 188-91, 206-8, 399-400.
algebra, 289 Yid part II, ch, 18 Number, theory of, 352-3
Mechamcs )( geometry, 58
Mediation (and Immed1acy), vid Object, apprehenSJon of and logic,
Inference, Syllogism 59 , conception of and meta-
Medieval, logic, 88. physics, 59 , subJect and predu:ate
Memory, nature of, 321-5, 508-9 , mean obiects, not ideas of obJects,
of apprehension and proof, 427-8, 139-40 , obJect (and subJect}, 49,
428-g, 431, 433, 451 , cannot 53, 56, 61 seqq , 65, 67 seqq , 74-5,
explam error, 111. 7g-8o, 629-30
Metaphysics, dlSt. logic, S5-7, 59 , ObJective side of knowing, atated
studies obJectlve and sub1ective without subJective m verbal form
mdes of thmking, 56; conceptlODS of '111dgement ', 85, 87, 141, 1-49,
8g8 INDEX
298: studied by sciences, 56, 85; and not-being, 248 : on negation,
subJectlve by ICJS1C, 56 ; both 256 n.; doctnne of reminiscence,
objective and subjective by meta- 514: unaddible numbers, 352,
phyncs, 56 ; modal dJStJnctions Phlb. 24 D 365 n.1
not obJectlve, 224 , ' conception ' T/tl. 201 D 217 D,2
ob1ective or sub1ect1ve ?, 300, Sph. 237 D 215 D,I
306-7, 309, universals ob1ect1ve, ,. 251 B 217 D,2
344 seqq ; defimtions obJective or R.p 510 D 465 ll.I
sub1ect1ve ), 378, 38o-1 , neces- Plurality of Qi.uses, 245, 587-8, 591,
sity in inference obJectlve, 432 , 592, 611-15
hypothetical statements express, POS1tive, v,d Negative
528 (cf 526), 545 , probability not Possible, 224; 111d Necessary, •Prob-
obJective, 569 seqq , ob1ect1ve in able
Kant, 629-30. PosllhO¥ Analytics, vil Aristotle.
ObJecbve 1deabsm, 6o, 63. Postulate, in mathematics, 467-8.
Ockham, on ' Judgement •, 92 n : Potential, 36o
on connotation, 405--9 Practice, effect of on belief or opmion,
Opinion, a kind of thinking, 36, cf 100-3
wondenng, dehberatmg, 36, in- Predicate, derivation of term, Jr 4 ;
cluded under • Judgement ' ) , 87, definition of, 262 ; modal d1S•
92 n , 93, 95-7 , )( knowledge, 90, tmctlons fall w1thm, 225-6; 11,d.
g6, 97, 99, 100,104, )( behef, IOI , Subject
)( conception, 302-3, 318, not Pred1cat.ion, Anstotle's view of, 159-
reducible to combination of ideas, 66, 1 77 • not expressed by ' to be '
277 , does not admit of error, 105, more than by other verbs, 176,
108, 109 , distinct from its ob1ect, 177, 213, 214, syllogISm confined
68 to, 420, 436, 439-41 , 11,tl Sub)ect
Opposition, 364-7 , )( contradiction, Premiss, v1d Inference, Syllogwn,
365 , absolute )( relative, 366 maJor, 135, I 36, 142, 144, 447-8,
Or, either, v1d D1s1unctlve 458 seqq , mmor, 135, 142, 144,
Organism, for science, 596 448 , fOnciple of inferences not
Organcm, v1tl Aristotle a prenuss, 457 seqq
Origination (absolute), ;r9 Present tense, 182, 185, 186, 227,
Pnce, use of term 'appreheu11on ',
Parmenides, on being and not-being, 78 n
247 : on space, 247, Probabibty, 103-4, 224, part III,
Particular Statement, 330-2 , infer- ch 8 : probable opm1on, g6
ence from why impossible, 436 Problems, hypothetical statement
Particulars, knowledge of, r68, vid gives relation between, 241, 243.
Universal Problematic, conceptions, 261, 318,
Parts, ) ( ' elements • 1n whole, 65 n 509-10, 522-3, 525-7 ; statement,
Parts of speech, logical s1gruficance 224 , universals, 376.
of, 83, I 70-91 Proper Names, 202-4, 389; have
Perception, species of knowing, 35, they connotation ?, 404
482, )(tlunking,40,45,46,79,94, Property, in Anstotle, 3Bo-1, 471 ;
339, 481 : )( feeling, 35 , Judge- and definition, 26, 383-4, 470-3 :
ments of, 93-5, 225 , not source properties common to comc sec-
of all ideas, 513-15 , not irrelevant tions, 26.
to logic, 56 : realist account of, 61 ; Proportion, theory of, in Euclid's
involves universals, 336 seqq Elements, 27, Si•
Peripatetics, term ' logic • attributed Proposition, or1gmal meaning, 83-4;
to, 32 n vid Judgement, Statement.
Peld10 Pnncipu, is syllog1Sm a?, Psychical image ) ( content or mean•
447,464 ing, 275, 28o-8, 292-4.
Philosophy, relation to science, 26, 59 Psychology, dist logic, 51, 56-7: its
Phyncs, 58, 631-4 scope, 51, 53 IM;q.4·; studies tune
Plato, contribution to logic, ,30 ; on order and thinking as event, S3 ;
umty in logic, 8 3 , Parm.n1du, 348 , as phy11ologlcal, 54; inquires how
R,Publsc, 248, 328 , Soplnst, 248, thoughts occur, 54; errors 111, 317,
::.r56n : Tunuus, 248, theory of 328-9
Ideas, 146, 147, 345 ; on being Pythagoras, h1S theorem, 4SS, 459.
 INDEX 899
Oaalio/., apprehension of, 46-7. Self-evidence, 449, 490.
Otz_antity, science of pure, 476--7 : of Sensation, and perception, 46 : com-
Statements, 3J0-2, an improper parison ID, 46 , a~henmon of,
term. 333• JIJ-14, 329; )(unagmation., 330-l;
Question, logical position of, 2 30, )( memory, 321-$.
539 seqq, 550-2; fallacy of un- Sentence, as studied 1n logic, 83;
real, 39. fltll. Statement.
Queatiomng, wl. Wondering. Sign, Bradley's definition of, 280-1 ,
11,tl Symbol.
Realism, 61-2. Sigwart, cnticized, 53, n 2.
Reality, theonesof, are metaphysical, Simple Ideas, 312, 498, 500, 502 seqq ,
$5 : supposed sub1ect of all su-12, 519-21
Judgements, 166--g, or of some 1m- Suigular Statement, 166, 336--7; not
~ l Judgements, 228-30, m hypothetical, 238--g,
symbolic logic,., 642, 650 , v,tl Socrates, seeks definitions, 28, 1 32 ,
Bein, method, 41, 44, 38o; his dialectic,
Reasomn.g, v1tl. Inference: actual- 132: influence on ArtStotle, 133.
ized m sciences, 59 Sophists, general teaching of, 133
Recogru.tion, ;8 n Space, 358 : supposed non-Euclidean,
Reduction, to syllogistic form, 443-4, 242 n, 456-7, 561-8
454 seqq, 457 seq<(, , to first Special logic, 58-g
figure, Son , 453 , to impossibility Specie'!, )( genus, 356-61, 502-3 : de-
(1'i,duceio a.d 11bsurdum), 21 S, 24..1-3, pendence on universal, 29, 25~,
246, 381, 475, 553-6o, 653 335-6, 36o , may not derivo their
Reflective thmkmg, So umty from common quality, 39
Relations, apprehended by thmkmg, Spencer, H., hts cnterion, 621-2 :
35 : and things, 71 seqq , 153, 157- doctrme of evolution as explana-
8, zo4, 501 , and terms, 250, tion of axioms &c, 616-28.
imply ddlerence of related terms, Statement, general defirution of, 261 -
72-4, 158 , fallac1es about, 254, 2 ; 1ts meaning wholly obJeotive,
255 n.: verbal expression of, 172, 149 (l..f 139 seqq ), 298, 299, 3II,
174, 175, 185, when simple, 119, d.5 studied m log1c, 81 seqq, 87, 90,
of Statements, not expressed by modern logic founded on, part II,
hYJ)Othetacal form, 495-6, 527-30, ch 2 , always categorical, .143 , and
Remmascence, involves apprehension, knowledge, 313-14. v1tl, Judge-
37, Plato's doctrine of, 514 ment, Impersonal, Command.
Representative theory, 61 , v•d Copy- Stoics, treatment of logic, 32.
mg theory Stress Accent, as mark of predicate,
IZJ, l.!4-6, 196, 197,
S-P, as symbol of statement, 127-8, Subject, grammatical, II9 seqq , 124,
138--9 128,110, 143-4, 148,166,167, 19:z.
Scepbc1sm, Greek, 217. logical, 123-6, 130, 148, 15R, 160,
Scholastic, view of Judgement, 92 n 167-8, r74, r84, 194-7, 221, 226,
Sciences, or.1g11i of, 24 , sphere of, at .!29, 240-1, 200 , metaphys.tcal,
first confused, 27 , defimbons of, IH seqq, 158, 178--9, m "Bradley,
24-30 , relation to metaphysics, 168-g, how expressed in language,
55-6: as the study of tlm1gs )( 171, I 74, 179, 195, vatl Substance :
logic as study of thought, 32, part I, and Predicate, 54, 8.1, part II, ch.
ch 4 , study objective side of 4, 5, 6, 193, 196, 231 seqq , 240 :
thinking, 50, 79 , essentially differ- in particular statements, 330 ,
ent from logic, 79 , organ.1c relation relation of, confused by Ar.1Stotle
of, 26 , classification of, 26 : ancil- with metaphysical, r 5g-66 ; of
lary, 27 ; special methods of, de- attnbution, 153 seqq, 158, 193,
pendent on subJect matter, 59, 368, 195, 196, 210, 218, 259, 391 : sym-
6.Jo: definitions m, 382-4, 472: bols for, 205-6 ; of statement,
claYificatlOD in, 371-2, 385 , • con- 194-5 , attempt to reduce all
cepts• of, 305, 308, inference m, propo11ttons to sub1ect-predicate
414 (111tl. Induction) ; as deductive, form, 197 seqq : and attnbute,
57, $8: scientific proposition con- 153 seqq., I68, 178--g, :zoo; Ill
wrtible, 245, 384, 484 ; natural At.lStotle, I 59-66,
sciences, S96: pure) ( applied, 57-8. SubJective, aspect of thanking studied
goo INDEX
by logic, 54, So ; baslS of logical Thinking, defined, 33 ; species of
dl81:inct1on between subject and consc1ousness, 33 , as ong1n.i.ttve
predicate, 139-41, 148, 161 : are activity, 35 , )( sensatton &c , 32 ,
negative statements subjective ?, d1St perce1vmg, 45, 9~, 415,
265 seqq , probabibty subJective, involves imagination, 37 , ( know-
5b9 seqq , v,d Objective tng, 35-6, as opining c, 36;
SubJective Ideahbm, 6o, 63, v,tl relation to knowing, 13, part I,
Berkeley, Kant ch 2, 81, 105-6, based on know-
Substance, 154-7, 178, ~ 11 , m ing, 37, on apprehension, 97, com-
Aristotle, 159 seqq , how ex- parison a speues of, 35, appre-
pressed m language, 171, 174, zo1, hension of relations, H , apprehen-
405, logical symbols for, 199, zo1, 11on of universals, H, 39, 42t.3, 45,
204 , )( attribute, 200 , system of, 67, .292-3, 107, 336, 337, 482-3,
zo4, primary and secondary, 223, 513-15, =(1) event, (2) mea.nmg,
171, ,76 51 , obJectlve and subJective side
Substantive, v,d Nouns of, 53, m relation to error, 112-13
SuppOS1tion, 106, hypothetical Judge- 1homas of ~trasburg, 407-9
ment not relation between sup- fhought, forms of studied by Gram-
positions, 5.29-31, 533-0, 553 seqq mar and Logic, 48, 49, 50, 81 • )(
Syllogism, origin of, 8.z-3, 132 , m forms of being, 149-50, as special
Ar1Stotle, u9 seqq , 45, seqq , ob1ect of logic, 32, B, 48-54,
relation to everyday reasoning, 4 3 , 50 seqq , 6o 'leqq , 78, 94 , m
supposed basis of all demonstr.i.tl'l,C relat10n to objects, <>1 scqq, 65,
tnference, 57 , influence on logu.,al 67 !>eqq , 7 5 , relation to ' ioeas ',
theory, I .Z9 l 31 seqq, l 1<>-7, 147- 277-9 , verbal expression as re-
8, 418, criticism of, 82, 14.? 147-8 present.J.tion of 274, 311, ~con-
420-1 , deals with ob1ecbve, Il0t 'IC10usnebs (m Berkeley), 34
sub1ective, 142-3, 145 , )( maten.i.l Time, tune order m logic, 51-4, do
tnference, 4ID, 44.z, method a t11ne-d1stm<-hom, belong to copula
pnora and not analytical 416-7 , or pn.d1cate ~, Z:!6-7 , as expressed
not purely formal, 4 ,7-8, 44.i , by tense, 177, 18o, 182,18 1, 185-6,
why analysed mto tllree elcmenl.b, not empmc.i.lly given, 515-16
134 , why two premisses nece'ISary, 1 opics, 1 id Aristotle
446-8 , relation of conclus10n to 1ruth, aun of log1c, 53, and of other
prem1Sses taken togetller, 449- 52 , subJects, 57
symbolism for, 200, 'fallacies'
in, H4, 438 Umform1ty of Nature, presupposed
SylloglSbc, not a general form of pure m induction ~89
reasontng, 57, , 31, 147-8, 420-1, Umty, of forms of thmkmg 38 , m a
435-6, 438-46 , as pure log1c, 57 science, 26 , of different depart-
Symbols, Bradley s dehmtion of, 288- ments of thought, 78-9 , of a
92 , importance of 477-8 thmg, 15 5 , of complex ob1ect, 501-
Symbolic Logic, 57 n pd.rt IV d1 u 1, 507-9, different kmdb of 155-6
Symbolu.m, be'it for logic, 19b-.mo, Umven,.ih umfymg cognate science!>,
440, 639 beqq z7 pre~uppO'ltld mall knowledge
Synthes1b, Ar1&totle on, ..? q and practical action, 1 55 , relation
Synthetic )( and Analytu. ..? F-5 to abstr.tcbon, 2b , as apprehended
synthetic umverb.i.l Judgementi., by thmkmg H, W, 4i-,, 45, 67,
how possible, 449 .z9i-,,307,3,u, ,,7,4&2-3,51,-15,
System.i.tizabon, a& ideal of science, relation to particulars, 146-7,
.19, ud. Clai,s1ficJ.t1on 156n, 170-1, 189-90, 199, 207,
20b, .?li, 311-6, ,18-54, 36o, 375,
Teleology, 596-7 37b seqq , 389, 3q4-5, 401-4, 437-8,
Tense, 177, 18<.>, 182, 18J, 185-6, 227, 456-8, 503-4, obJectlve, 335 , two
2.28 kmds of, 37 5-6 , verbal expression
Term, vid Inference, Syllog1Sm, Re- of mamly by common names and
lation , definition of, begms from adJectlves, 188-91, 202-3, 206-8,
usage, 40-1 tnference dependent on, 433, 434 n ,
Thing, )( tllought, 6o seqq , nature 448, 479-81, not-Aness not a
of a, 15.?, 153, 155, 168,375, v1tl universal, 251-8
Object Umversal propositions, not always
 INDEX
hypothetical, 236-7, )( ~rticular, Verb, general function of, 166, 171-3,
330 seqq , 47g-80 , m symbolic 175-81, 183, 185-7, 219, verb
logic ', 644 seqq 'to be·. 181-4, 186-7, 190, 191,
Usage, relation to definition, 40, in 194, 201, 209, 212-19, 227-8, 262,
apeech, appealed to, 34, 40, 46, 48, 263-4
63, 67, 68, 70, 113, 119, 170-91,
196, 208, 212, 214, 216, 221, 225,
227, 237, 242, 244, 259, 263, 268, Whately, his treatment of logic, 32,
272,275,288,292,298, 300-1, 304, 119n
307--10, 324, 332, 336, 340, 344-5, Whcwell, on class1ficat1on, 372
348--9, 351, 359,361,369, 37S, 387- Wish, Jog1cal position of, 230
404, 413,422, 438-9, 445,481, 491- Wolf, his d1v1S10n of logic, S7 n
S, 497, 503, 521, 529, 532-5, 537, Wondering, species of thinking, 36,
539, 545, 548, 553-5, 569, 574-5, relation to knowing, 37, 522; tm•
5go, 593, 001, 650-r, 659, an portance of, neglected, 36
1d1om of Greek, 211, 228 Words, relation to thought, 292,
298-g
Variations, method of concomitant,
591-3 (cf 594-5)
Venn, .'>ymbolu Logic, 619, 645, u48- Zero, in logical calculus, 643 , equa-
9, 661 , nomm.thsm of, 643 tion of 1mposs1bles to, 647-59