Convertible Counterpoint

THE ENGLISH TRANSLATION OF THE MOST DISTINGUISHED
TREATISE EVER WRITTEN ON MUSICAL COMPOSITION
CONVERTIBLE COUNTERPOINT
IN THE STRICT STYLE

BY
SERGE IVANOVITCH TANEIEV
Translated by G. ACKLEY BROWER
With an Introduction by SERGE KOUSSEVITZKY
BOSTON • BRUCE HUMPHRIES . PUBLISHERS

$3000
Convertible
Counterpoint
BY SERGE IVANOVITCH TANEIEV
J Music scholars who are acquainted with the

Convertible Counterpoint of Serge Ivanovitch
Taneiev consider it the greatest work ever writ
ten in its field. As a textbook on composition it
has the high recommendation of such authori
ties as Dr. Serge Koussevitzky, Serge Rachman
inoff, Leopold Godowsky, Moritz Rosenthal,
Gregor Piatigorsky, Lazare Saminsky, Alexan
der Siloti, Nicholas Orloff.
Every history of modern music mentions the
Taneiev Counterpoint; every serious scholar
has heard of it; yet the riches opened up by
Taneiev's principles and methods have until
now been available only to those who have a
knowledge of the Russian language. The ap
pearance of the text in English is a landmark in
musical education, for by placing within the
reach of American and British composers and
students material that heretofore has been ac
cessible only to their Russian contemporaries, it
supplies a lack that has long been felt among
students of contrapuntal theory.
It is a book that will reveal possibilities for
the art of composition that up to the present
have been but vaguely realized or actually un
known. For the study of counterpoint is here
put on a basis of pure mathematics, and the
Taneiev method opens up an enormous range
of resources, hitherto untouched because of the
lack of that adequate approach which only
mathematics can provide.
In the Preface to his work, Taneiev states his
aims as follows :
"In studying the difficult and involved treat
ment of counterpoint— especially double coun
terpoint — as presented in the textbooks of an
cient and modern theorists, I encountered vari
ous obstacles that seemed to result from faulty
classifications, too many useless rules and not
enough essential rules. The system expounded
in the present work appears to me to be simpler,
more accurate and more accessible, as the result
of applying the processes of elementary algebra
(Continued on back flap)
IN THE
STRICT STYLE
Thi s O no
304E-18S-B5LB
SERGE IVANOVITCH TANEIEV
IN THE STRICT STYLE
Translated by
G. ACKLEY BROWER
(Mus. Doc, Toronto)
With an Introduction by
SERGE KOUSSEVITZKY
BOSTON • BRUCE HUMPHRIES • PUBLISHERS

© Copyright, 1962, by
BRUCE HUMPHRIES
Printed in the United States of America

Library of Congress Catalog Card No. 61-10738
TO THE MEMORY OF
HERMAN AUGUSTOVITCH LAROCHE
(1845-1904)
INTRODUCTION
This translation of my late friend's labor of twenty years opens up to the English-
speaking world one of the greatest musical treatises ever written, yet one that the
barrier of language has denied far too long to countless students and composers whose
careers might have been forwarded by it. If anyone ever raised the question as to what
author commanded the most comprehensive and ready knowledge of counterpoint through
the ages I believe none other than Taneiev could be named, because this work is the
synthesis of two centuries of study and learning in the realm of counterpoint.
Serge Ivanovitch was one of the most extraordinary intellectuals of the many to
which Russia has given birth. In addition to music he acquired a really deep knowledge
of natural science, sociology and philosophy. Russia's artists and thinkers sought him
out all through his life—even flocking to visit him in the poor dwelling of his last years.
Tschaikowsky, fifteen years his senior, would submit to criticism from this pupil of his
which he would tolerate from no one else. Rimsky-Korsakow, with all his technical
brilliance, felt like a student musician in the presence of Taneiev, and admitted it.
This great treatise was published in Moscow in 1909. Since I practised composition
in my early career I fairly devoured the book and urged on many others the advantage of
doing likewise. When I had to devote most of my time to my career as a double-bass
virtuoso, and later when my activities emphasized conducting to the detriment of com
posing, I found that Taneiev's Counterpoint was an invaluable asset on innumerable
occasions in working out interpretation of orchestral scores-especially those of Bach,
Handel and Brahms.
In Moscow we lived in the same neighborhood and frequently called on one another.
We had long, interesting talks, and he amazed me by the boldness of his ideas; often in
the field of musical interpretation he was daring to the point of radicalism. I recall when
I was preparing for the first time to conduct Beethoven's Ninth Symphony that I used to
go to Taneiev with the score. His conception was so striking and free from established
traditions that even I, revolutionarily inclined, did not dare to accept it.
I knew him, of course, as one of Russia's formost pianists—although unknown to
America. I have never heard the Fourth Concerto of Beethoven given a more brilliant
and vivid performance; free and logical at the same time.
And I knew him also as a composer of unique qualities. So much was he the master
of contrapuntal theory and not its slave, that his music concealed his immense technique
in its application. This is particularly true of his only opera, Oresteia, his chamber
music, and especially his cantata On the Reading of a Psalm.
It was my privilege to introduce this masterpiece to the public in two different ways.
Taneiev in his last years was much reduced from lifelong affluence, and was living in a
primitive dwelling with not even running water. In the beginning of 1913 he told me how
he wished to compose a cantata which would require two years of intense work, but that
he could not even dream of doing it because he must make a living— giving lessons and
so forth. I immediately offered to give him the sum he needed, and to publish
the cantata in my publishing house, Editions Russe de Musique, which, by coincidence,
had been founded the year his book appeared. The sum he mentioned was ridiculously
small even for those days. All he needed, he said, was 3000 roubles—equivalent to
$1500— for two years of life! This may well give an idea of how modestly he lived, but
he positively refused a higher offer.
7
g INTRODUCTION
In two years the cantata was completed and I placed it on my program in both Moscow
and Petrograd. It was indeed a masterpiece—a great and noble work. When I conducted
two performances in each city in April, 1915, it was declared by all competent judges
to be the finest work Taneiev had produced. I have never seen Serge Ivanovitch as
happy as at these performances of his cantata. Later in the month he caugh a severe
cold at the funeral services for Scribin, and this produced a heart complication which
caused his death on June 19, 1915.
The enlightment on musical structure, the mental stimulus, in Taneiev's book are of
far-reaching service. As counterpoint is presented by Serge Ivanovitch the reader finds
himself, like the author, making of it not so much the analysis of a process as a habit of
thought, a second thought, a second nature, which leads on to the creation of beauty-
flawless in its form and proportions.
SERGE KOUSSEVITZKY
TRANSLATOR'S PREFACE
The works of Serge Ivanovitch Taneiev (1856-1915), both musical and literary, seem
to be little known outside of his native country, though recent years have witnessed an
increased interest in the labors of one who as composer, theorist, concert pianist,
critic, and teacher, became one of the outstanding figures in the musical life of Russia
and who is now gaining the wider recognition that he deserves. Taneiev's compositions
must speak for themselves; the present purpose is neither to attempt a critical estimate
of them nor to give a biographical sketch of their composer, but to introduce to teachers
and students of composition his work on advanced counterpoint.
It is difficult to discuss this book in terms of restraint. Since the fourteenth century,
or earlier, many books on music theory have appeared. Amid a mass of indifferent writings
and others of considerable value but not outstanding there are a few that have made
history; one of the greatest was the Dodechachordon of Glearanus, another was the
Gradus ad Parnassum of Josph Fux. Not without good reason has Lazare Saminsky
referred to Convertible Counterpoint as "having the same meaning for musical science
that Newton's Principia has for cosmology." It may be said without reservation that
the student has at hand the greatest work on counterpoint ever written. It is a book that
will reveal possibilities for the art of composition that have hitherto been but vaguely
realized or actually unknown. Though applying, according to the title, only to the strict
style of the Polyphonic Period, its principles, as the author himself says, may be
extended to the free style of later times and to the modernism of today and of the future.
To study and master its contents will give the student a command over the resources of
composition that can be obtained in no other way. The variety of subjects in it, the
thoroughness and clearness with which they are presented, their logical arrangement,
the examples illustrating the text, and above all the astounding originality of the author's
thought—all this makes this a work compared to which other books on counterpoint
seem elementary.
From this it may be inferred that Convertible Counterpoint is not a beginner's text,
yet its study may be undertaken sooner than might be expected. The author says (in §175)
that the exercises he suggests should start as soon as three-voice counterpoint can be
satisfactorily managed, and that thereafter simple and convertible counterpoint should
be studied concurrently. To this I would add that there seems to be no reason why the
exercises in two voices should not be successfully attempted as soon as two-voice
mixed counterpoint (i.e. both voices in the fifth species) has been studied. The student
should then be well able to cope with the simpler of the fascinating problems set by
Taneiev in the earlier chapters of Parts One and Two, though the more difficult ones
will require skill in the manipulation of from three to six voices.
The first thing that is likely to surprise the reader who may think that this is "just
another book on counterpoint" is the proposition advanced by the author in his Preface-
that the study of counterpoint is here put on a mathematical basis—algebra in fact. Yet
this is quite in accordance with present-day tendencies, and the fact that Taneiev
thought about it as far back as 1870 shows that he is a pioneer in a field of research
that now includes several prominent names in the musical world. But the student may
be assured that he is not expected to know more than the fundamentals of algebra; of
this more will be said presently. Taneiev's method opens up an enormous extent of
untried resources, heretofore inaccessible because of the lack of an adequate approach—
and only mathematics can provide it. Let no one get the idea that such an approach will
9
10 TRANSLATOR'S PREFACE
stifle the imagination and yield the unwelcome result of writing music that sounds
mathematical. The effect of Taneiev's method is quite the opposite; it releases the
imagination, pointing the way to endless possibilities that otherwise would never have
been thought of. Here a few statistics may be enlightening, as showing how inadequately
a vast subject has hitherto been treated.
Referring to eighteen standard texts that claim to teach double counterpoint , I find
that while all of them deal with double counterpoint at the octave (or two octaves), none
mentions double counterpoint at the fifth, only three at the sixth, two at the seventh, six
at the ninth, seven at the eleventh, all except two at the tenth and the twelfth, six at the
thirteenth and five at the fourteenth. Two of them deal with double counterpoint only at
the octave. Several speak disparagingly of double counterpoint at intervals other than
the octave, tenth, and twelfth. Not one mentions a use of combined themes that is found
in Bach but which can be classified as neither simple nor double counterpoint. Put
together, these texts provide for only nine ways of writing counterpoint in which the
interval-relationship could be changed; Taneiev deals exhaustively with twenty-three,
not by giving endless lists of rules and exceptions but by equations in simple algebra
that eliminate all trial-and-error methods and that give positive results. All of them
are practicable in the strict style, not to mention the free. Furthermore, none of these
texts deals with triple counterpoint at any interval other than the octave, and one of them
(Jadassohn) definitely states that such a thing is impossible. Taneiev shows how it is done
Of the authors cited in the first footnote and who could be expected to know about
Taneiev's work only one, George Conus, mentions him. Conus' book A Course in Modal
Counterpoint refers briefly to double counterpoint at the octave, tenth, and twelfth, but
he gives credit to Taneiev as having written the only complete treatise on the subject.
After all, the sources upon which Taneiev's work is based were available to many who
came before him, to his contemporaries, and to all who came after, and the fact that no
one took full advantage of them certainly justifies the remark made in § 279 about certain
theorists being exposed to "grave suspicions."
Nearly all these texts confuse the issue by treating double counterpoint at the tenth
and the twelfth and duplication in imperfect consonances as belonging to the same
category, whereas they do not. Again, the changes possible in the time-entrances of
two or more melodies, called in this book the horizontal shift, are ignored in all texts
except Taneiev's, though some authors call attention to this interesting phenomenon
in occasional quotations from Bach. But none of them throw the faintest light upon how
it is done. Finally, the principles of duplicated counterpoint and the horizontal shift
combined with other varieties of counterpoint in both two and three voices leave one
amazed at the enormous scope of the subject. An inventory of what is still untried in
counterpoint would, by the application of Taneiev's methods, run almost into astro
nomical figures.
Now as to algebra: the amount needed is very small-only a knowledge of the meanings
of positive and negative as appiled to numbers and letters; the special meanings of the
signs +, -, < and >, the principles of addition and subtraction; the transposition of terms
in equations, and the rules for the removal of parentheses. Three hours spent with any
algebra textbook should be enough for the most complicated of Taneiev's problems.
The authors consulted were: Bandini, Bridge, Cherubini, Conus. Dubois, Goetschius, Haupt,
Jadassohn, Jeppesen, Kitson, Krohn, Marquard, Morris, Prout, Richter, Riemann, Spalding, and
Stohr; by no means exhaustive of the literature on the subject but, I think, a fair cross-section.
o
In Russian; it is published by the Soviet Government.
TRANSLATOR'S PREFACE n
A few additions of my own are in footnotes or at the ends of chapters. The musical
quotations have been verified—a necessary measure as the original edition contains
many misprints. On the last page of the original is the word konyetz, which I have
omitted, as this book was not "the end" but was followed by a sequel dealing with
the canon, doing this difficult style of composition what the present work does for
convertible counterpoint— puts the study on a scientific basis.
I am indebted to several whose interest, advice, and information are in no small
measure responsible for the appearance in English of the monumental work of Taneiev.
First to be mentioned in Lazare Saminsky, who, about twenty years ago, told me of
Taneiev 's book and how radically different and superior it was to other texts. Without
his description of the book and his enthusiastic recommendation of it I might never
have known about it. Next are the obligations I owe to Alexander Siloti and Nicholas
Orloff, both pupils of Taneiev whose reminiscences of their teacher were of the greatest
interest. From Serge Rachmaninoff, Leopold Godowsky, Moritz Rosenthal, and Gregor
Piatigorski I have received encouragement in a project that I entered upon with some
doubts as to its interest to a publisher but none as to its value. Dr. Serge Koussevitzky,
whose activities in the musical life of America were too well known to need further
comment, contributed an Introduction. From books I have got valuable help from the
Memoirs of Taneiev, by Leonid Sabaniev, and from the second volume of the History of
Music in Russia, a symposium on Taneiev, published in Moscow.
Tha manager of Bruce Humphries, Mr. Edmund R. Brown, and the members of the
editorial and production staff have solved most successfully the peculiar mechanical
problems that the printing of such a complicated text involves—the first of its kind to be
done in English.
G. ACKLEY BROWER
AUTHOR'S PREFACE
In studying the difficult and involved treatment of counterpoint—especially double

counterpoint—as presented in the textbooks of ancient and modern theorists, I encountered
various obstacles that seemed to result from faulty classifications, too many useless
rules and not enough essential rules.
The system expounded in the present work appears to me to be simpler, more accurate
and more accessible, as the result of applying the processes of elementary algebra to
contrapuntal combinations, and by restating certain essential rules in terms of the con
ventional symbols of mathematics. This enabled me to take into consideration a far
greater number of relevant facts, and to bring them under control of a comparatively
small number of general principles.
For many years I have used separate parts of this theory in my classes in counterpoint
at the Moscow Conservatory, and I have tried to simplify the treatment at those places
where experience has shown that difficulties were encountered by students.
The present book is an exposition, of the utmost comprehensiveness, of convertible
counterpoint in the strict style. In using it as a textbook the teacher should select, from
amid the detailed development of the subject, what is most necessary for the student.
I have dedicated this book to the memory of H. A. Laroche, whose articles (especially
Thoughts on Musical Education in Russia) have had a profound influence on the trend
of my musical activities.
SERGE TANEIEV
Klin, July 1, 1906
12
INTRODUCTION BY DR. SERGE KOUSSEVITZKY 7
TRANSLATOR'S PREFACE 9
AUTHOR'S PREFACE !2
INTRODUCTION 17
PART ONE
VERTICAL-SHIFTING COUNTERPOINT
DIVISION A
TWO-VOICE VERTICAL-SHIFTING COUNTERPOINT
Chapter
I INTERVALS 25
The Notation of Intervals 25
Addition and Subtraction; Negative Intervals 25
Compound Intervals » • • • • • • 27
Successive Series of Intervals; Division into two Groups: int. and int. 28
Order of Intervals in the Successive Series 30
II THE VERTICAL SHIFTING OF CONTRAPUNTALLY COMBINED VOICES;
INDEX OF VERTICAL-SHIFTING COUNTERPOINT 34
Notation of the Vertical Shift: v, vv. Formulas for Original and Deriv
ative Combinations 34
The Shifts 34
Relation of Original to Derivative Intervals 35
Index of Vertical-Shifting Counterpoint (Jv) 36
The Formula m + Jv = n; Inferences Therefrom 37
Conditions of the Shifts 38
Limiting Intervals; Their Signs (<, >) 38
Signs for the Shifts (<,>,<>, <) 40
Jv Equal to Zero 40
III THE GROUPING OF INDICES ; 41
List of Indices 41
Columns of Indices 42
Octave Relationship of JJv 43
Grouping of the Columns in Pairs 44
Determination of the Value of an Original Interval and of the Index • • • 44
Allocation of the Melodies of the Derivative Combination According
to Voice 47
Intervals Determining Jv; Entrance Intervals 48
Interchange of Original and Derivative Combinations 49
The Augmented Fourth and Diminished Fifth 50
IV THE RULES OF SIMPLE COUNTERPOINT 53
Limitations Imposed upon Intervals by the Rules of Simple Counterpoint . 53
Imperfect and Perfect Consonances 53
13
14 CONTENTS
Dissonances: (a) Passing and Auxiliary; (b) Suspensions 57

Resolution of Tied Dissonances; The Dissonant Note 59
The Fourth and Ninth in Multi-Voice Counterpoint 60
The Free Note to a Tied Dissonance 62
The Preparation of Suspensions "4
Symbols Indicating the Conditions of the Use of Tied Intervals • • • • 66
V THE RULES OF VERTICAL SHIFTING COUNTERPOINT; FIXED INTERVALS
INTERVALS 70
Fixed Consonances "•
Table of Fixed Consonances 73
Explanations of Table '3
Combinations Without Suspensions, Admitting of Shifts at Two or More
Indices '°
Combinations Without Suspensions, Admitting the Duplication of Imper
fect Consonances; Three-Voice Counterpoint 82
Fixed Dissonances ■*
The Preparation of Fixed Dissonances 89
VI VARIABLE INTERVALS W
Variable Consonances ™
Variable Dissonances 92
The Preparation of Variable Intervals 95
Simplifications in Compiling Tables of Indices 97
Tables of Indices Giving the Inverse Shift 97
Tables of Indices Giving the Direct Shift 100
Distribution of Variable Dissonances 101
VII SUPPLEMENTARY RESOURCES 105
Ornamental Resolutions of Suspensions 105
The Free Note Progressing to a Consonance at the Moment of Resolution • 105
Insertion of a Consonant Note Between a Dissonance and its Resolution • 107
The Cambiata Figure 108
The Use of Notes Taking the Sign x • HI
Negative Intervals in the Original Combination Ill
VIII EXERCISES IN TWO-VOICE VERTICAL-SHIFTING COUNTERPOINT. . . -112
IX TWO-VOICE ORIGINAL COMBINATIONS GIVING MORE THAN ONE
DERIVATIVE; COMPOUND INDICES 115
Double Indices 115
Triple Indices 121
The Polymorphous Index 122
X THE DUPLICATION OF IMPERFECT CONSONANCES IN TWO-VOICE • 123
The Duplication of One Voice Using Tied Intervals; Two Derivatives:
Two-Voice and Three-Voice 123
Table of Duplications (Supplementary Conditions) 126
Uniting the Conditions of Duplications with the Conditions of Every Jv ' 126
The Duplication of One Voice Using Tied Intervals; One Derivative,
Three-Voice j29
The Simultaneous Duplication of Two Voices; One Derivative Four-Voice 1 30
Duplications of Two Voices Without Suspensions j31
Duplications of Two Voices Using Suspensions 134
The Six-Four Chord in the Strict Style 141
CONTENTS 15
Tonal Combinations Resulting from the Duplication of Suspensions;

their Importance for the Development of Harmonic Technique 145
DIVISION B
THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT
XI THEORY OF THE TECHNIQUE OF THREE-VOICE VERTICAL SHIFTING
COUNTERPOINT 157
Preliminaries '57
Notation of Voices and Shifts 159
The Indices Jv' Jv" JvS. and Their Relationships 160
Rules for Ninths 163
Rules for Fourths 164
Multi-Voice Counterpoint as Developed from Two-Voice; Verification
of Multi-Voice Counterpoint by Two-Voice 169
Cases of Multi-Voice Combinations not Verifiable by Two-Voice . • . 170
XII SPECIAL CASES OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT 174

XIII COMBINATIONS GIVING MORE THAN ONE DERIVATIVE; TRIPLE
COUNTERPOINT 183
General Formula for Triple Counterpoint 183
The Difference Between Triple Counterpoint at the Octave and Triple
Counterpoint at Other Intervals 184
Triple Counterpoint at the Octave 185
XIV COMBINATIONS GIVING MORE THAN ONE DERIVATIVE (Continued) • • • 190
Extracting Derivatives from a Given Combination 194
XV DUPLICATIONS IN THREE-VOICE COUNTERPOINT 198
The Duplication of One Voice; Four-Voice Counterpoint 198
Simultaneous Duplication of Two Voices; Five-Voice Counterpoint • • • • 201
Simultaneous Duplication of Three Voices; Six-Voice Counterpoint • • • • 202
PART TWO
DIVISION C
TWO-VOICE HORIZONTAL-SHIFTING AND DOUBLE-SHIFTING COUNTERPOINT
XVI GENERAL PRINCIPLES 207

The Basic Version 208
Basic Version and Imaginary Risposta 209
Conditions of a Difficult Problem on a Cantus Firmus 210
Basic Version for Double-Shifting Counterpoint 211
Shifting Voices from One Part of the Measure to Another 214
Functions of the Basic Version «... 215
XVII NOTATION OF HORIZONTAL-SHIFTING VOICES; POSITIVE AND NEGA
TIVE MEASURES; THE SHIFTS; INDEX FOR HORIZONTAL-SHIFTING
COUNTERPOINT; FORMULAS; METHODS OF WORKING OUT EXERCISES 219
Notation of the Horizontal Shift; h, hh; Formulas for the Derivative
Combination 220
Index for Horizontal-Shifting Counterpoint: Jh 220
The Formula a + Jh = b 221
Conditions of the Shifts • 221
16 CONTENTS
Application of h and v to the Basic Version; Formulas of the Imaginary

Combination 222
General Formula of Horizontal-Shifting Counterpoint 223
General Formula of Double-Shifting Counterpoint 224
Methods of Working Out Problems 225
XVIII INTERCHANGE OF ORIGINAL AND DERIVATIVE; PROBLEMS ON A
CANTUS FIRMUS 227
Interchange of Original and Derivative Combinations in Horizontal-
Shifting Counterpoint 227
Interchange of Original and Derivative in Double-Shifting Counterpoint- • 228
Problems on a Cantus Firmus • • • • 229
XIX ANALYSIS 232
Compositions of the Basic Version According to Original and Derivative • 232
Analysis of Combinations on a Cantus Firmus 238
XX COMBINATIONS GIVING SEVERAL DERIVATIVES 240
Combinations with Several Derivatives at a Simple Jh 240
Combinations with Several Derivatives at a Compound Jh 245
Arrangement of Contrapuntal Combinations in the Reverse Order ofTheir
Composition 251
DIVISION D
THREE-VOICE HORIZONTAL-SHIFTING AND DOUBLE-SHIFTING COUNTERPOINT

XXI THEORY OF THE TECHNIQUE OF THREE VOICE HORIZONTAL
SHIFTING COUNTERPOINT '. 265
Preliminaries 265
The Basic Version 265
Notation of Voices and Shifts 268
The Indices Jh1, Jh", Jh£ and their Reciprocal Relationships 269
Formula of the Basic Version 270
XXII THREE-VOICE COUNTERPOINT WITH HORIZONTAL SHIFTS 271
First Form of the Basic Version for Horizontal-Shifting Counterpoint:
Jh = 0 Absent 271
Second Form of the Basic Version for Horizontal-Shifting Counterpoint:
Jh = 0 Present 274
XXIII THREE-VOICE COUNTERPOINT WITH SIMULTANEOUS HORIZONTAL
AND VERTICAL SHIFTS (DOUBLE-SHIFTING) 277
First Form of the Basic Version for Double-Shifting Counterpoint;
Jh = 0 Absent 277
Second Form of the Basic Version for Double-Shifting Counterpoint:
Jh = 0 Present 279
XXIV COUNTERPOINT WITH AND WITHOUT RESTS 284
Two-Voice Counterpoint With and Without Rests 286
Three-Voice Counterpoint With and Without Rests 290
XXV CONCLUSION 299
Appendix A 302
Appendix B 329
Appendix C 339
Appendix D 349
Index 353
INTRODUCTION
The art of counterpoint has passed through two eras: that of the strict style, which at
tained its highest development in the sixteenth century (Palestrina and Orlando Lasso),
and the period of the free style, of which the crowning achievements are found in the
works of Bach and Handel. The differences between the contrapuntal writing of these
two eras are to be found both in the nature of the melodies themselves and in the char
acter of the harmonies formed by these melodies in combination.
Strict counterpoint, developed on the basis of the so-called ecclesiastical modes,
was a pre-eminently vocal style that had not been exposed to the kind of influence that
instrumental music later exerted; antedating such influence, it attained to complete self-
fulfilment. Strict counterpoint excludes everything that presents difficulty to voices
singing without instrumental accompaniment. Melodies in the strict style show evid
ences of their origin in the chants of the Catholic Church — they exhibit many char
acteristics of these early canticles. They are strictly diatonic, are written in the ec
clesiastical modes, and in them are no. progressions of intervals that are difficult of
intonation, such as sevenths, ninths, augmented or diminished intervals, etc
The basis of multi-voice counterpoint of the strict style is of course the two-voice
texture. Two-voice counterpoint is subject to the rules governing the progression of
intervals, these being employed in a way that for the normal hearing is the most simple
and natural. A knowledge of the rules of simple counterpoint in the strict style is es
sential in order to understand the present work, though this is not the place to explain
them. In strict writing the rules of two-voice counterpoint apply also to more intricate
polyphony. With a few exceptions it is observable that in a multi-voice combination
each voice together with every other forms correct two-voice counterpoint; that multi-
voice counterpoint is an association of several two-voice combinations, as a result of
which is obtained a series of varied consonant and dissonant harmonies, foreign to con
temporary harmony and often sounding strange to us. Although isolated harmonies may
be classified under the heads of certain chords, the term "harmony," in the sense in
which it is used in the music of today, is not applicable to the old contrapuntal style.
Harmony of the strict style is not subordinated to the requirements of our modern tonal
system, in which a series of chords is grouped around a central tonic chord: a system
that in the course of a composition allows the tonic to be shifted (modulation), and
groups all secondary tonalities around the principal key, besides which the tonality of
one division influences those of others, from the beginning of a piece to its conclusion.
In the harmony of the strict style there is no such dependence of some parts upon
others, or of what may be called harmonic action at a distance. Only in the perfect cad
ence, where as a result of the ascent of the leading-tone to the tonic the gravitation of
dominant to tonic harmony is temporarily brought about, can be seen the embryo of our
present tonal system. Aside from such cadences the strict style does not present a ser
ies of harmonies that are unified in this sense; key-continuity may be entirely absent,
and any chord may follow any other, on a strictly diatonic basis.
In music of the Polyphonic Period — essentially vocal — coherence was provided
for first of all by a text. But besides the text an external factor not belonging to
the domain of music — the works of the period possessed another — purely musical
17
18 CONVERTIBLE COUNTERPOINT
— resource, by which composition took on coherence and unity; a resource all the
more valuable inasmuch as harmony did not as yet possess the unifying power that it
subsequently acquired. This was imitation: the recurrence of a melody in one voice
immediately after its presentation in another. The result of this use of a single melody
that appeared in different voices was to distribute the thematic material equally among
all of them, giving to the whole a high degree of coherence. An imitating melody often
entered before the preceding melody had closed, and then did not end until still another
imitation had begun, a process that served to knit still closer the contrapuntal texture.
For two or more centuries the working-out of imitative forms in the strict style re
ceived much attention from composers. There arose many different phases of this device;
imitations on a given voice and without it, canonic imitation, imitation in contrary motion,
augmentation, diminution — forms that in the course of time culminated in the highest
contrapuntal form of all — the fugue. From the introducing of one melody in all voices
it was natural to take a further step and apply the same process to two melodies at once;
hence double imitation, double canon, double fugue. At this transference to different
voices of two melodies simultaneously the question must have come up as to the pos
sibility of changing their relationship at the successive recurrences, and thereby from
an original combination to obtain another, the derivative. Thus the origin of complex
counterpoint, i.e. the obtaining of derivative combinations, also came in the era of the
strict style.
In multi-voice music melodic and harmonic elements are subject to the influences
of the time and to the nationality and individuality of composers. But the forms of
imitation, canon and complex counterpoint — either as actualities or as possibilities
— are universally valid; they are independent of such conditions, capable of entering
into the plan of any harmonic system and adaptable to any melodic idiom. The idea is
prevalent that the old contrapuntalists of the Flemish Schools exhausted the resources
of imitation, especially as regards the canonic forms, but in reality they worked out
completely only a few of them; the rest received only incidental treatment or were not
touched upon at all. The outstanding merit of the Flemish composers was that they in
vented these forms and from them developed a flexible and efficient system of tech
nical procedure.
Arising in the era of the strict style, these forms survived without material change
until the end of the seventeenth and the beginning of the eighteenth century, when under
the powerful influence of instrumental music they were enriched by acquisitions which
up to that time had constituted mere technical virtuosity ; also by harmonic, figurative
and other elements that in the preceding era had been absent. This free counterpoint
of the time of Bach and Handel, essentially the same as our own, was sharply distin
guished from the counterpoint of former times, and its subsequent development naturally
contained its own elements. The new counterpoint was not based on the ecclesiastical
modes but upon the present major-minor tonal system. Not only in instrumental but also
in vocal melody progressions are found that are difficult of intonation for voices and
which were unconditionally forbidden in the strict style, such as leaps of sevenths and
of augmented and diminished intervals, figuration based on dissonant chords, chromatics
and other resources unknown to the older order.
The harmony of the free style is no less sharply distinguished from that of the pre
ceding era. The free style enables entire groups of harmonies to be consolidated into
one organic whole and then by means of modulation to dissect this whole into factors
that are tonally interdependent. This characteristic, absent in the former harmony, pro
vided the conditions for the development of the free forms of instrumental music that
INTRODUCTION 19
appeared at the end of the eighteenth and during the first half of the nineteenth century.
This new tonal system made possible the writing of works of large dimensions that
possessed all the qualities of effective structural style and that did not have to be re
inforced by texts or by imitative forms per se, but contained within themselv es the
necessity for the latter. By degrees this system widened and deepened and its spread
ing circle embraced newer and newer resources and laws governing the relations be
tween remote harmonies. Such were the broad horizons opened up for harmony; the
creative activity of 3eethoven then appears, and he, by a further expansion of the modu
latory plans as they stood at the end of the eighteenth century, showed how much variety
of key-relationship a composition could exhibit, both in its larger and smaller aspects.
Superseding the ecclesiastical modes, this tonal system was in turn affected by a
new one that tended to endanger key-sense by the substitution of a chromatic for a di
atonic basis; this led to a transformation of musical form. Applying the principle that
by the use of chromatic progression any chord may follow any other, and pushing it too
far, is likely to compromise key-relationship and to exclude those factors by which the
smaller units of form are grouped and amalgamated into one organic whole. Neither did
the harmony of the strict style, in which any chord could follow any other, though on a
diatonic basis, exhibit the characteristics of tonality and form as now understood. The
new harmony, as it now stands and which Fetis called "omnitonal," is inimical to the
logic of tonality and form; the chief difference between the old and the new is that the
diatonic basis is replaced by the chromatic. Omnitonal harmony, though adding to the
resources of composition, at the same time lacks the virility characteristic of the di
atonic method. To remain for a time in one key, as opposed to more or less rapid modu
lation, the contrasts afforded by passing gradually to a new key, with a return to the
principal key all this, by contributing to the clearness of long movements and en
abling the listener to comprehend their forms, has little by little disappeared from music
since the time of Beethoven and far more rapidly since the beginning of the twentieth
century.* The result has been the production of small works and a general decline in
the art of composition. Unity of construction appears with less and less frequency.
Works are written not as consistent organisms but as formless masses of mechanically
associated parts, any of which might be replaced by others.
As for the music of today, the harmony that has gradually lost its virility would be
greatly benefited by the strength that the contrapuntal forms can infuse. Beethoven, who
in his later works reverted to the technical methods of the old contrapuntalists, sets
the best example for composers of the future. The music of today is essentially contra
puntal. Not only in large orchestral works, where the abundance of independent parts
often results in obscurity, or in opera, where leitmotifs are worked out contra puntally,
but even in pieces of insignificant dimensions, can counterpoint be employed to the great
est advantage. The study of free counterpoint is therefore indispensible for the tech
nical training of composers, but because of its melodic and harmonic intricacy it cannot
be studied first. The foundation must be laid by counterpoint of the strict style, more
accessible because of its simplicity. The preliminary steps as regards shifting counter
point is the subject of the present work.
The term "complex" is used for that kind of counterpoint in which an original com
bination of melodies yields one or more derivatives. The term does not refer to the
•The original reads: ". . . has little by little disappeared from contemporary music," but as
this was written in 1906 it can hardly be considered a violation of the author's thought to bring
the statement up to date.—Tr.
complexity that results from the union of many voices, nor to the complexity of their
melodic or rhythmic features. The essential mark of complex counterpoint is the pos
sibility of obtaining from an original combination of melodies a new one, the derivative.
Complex counterpoint is divided into categories according to the methods by which
derivative combinations are obtained. The principal methods are: (1) the shifting of
voices; (2) duplications in imperfect consonances, and (3) transmutation; hence the
three aspects of complex counterpoint: (1) shifted, (2) duplicated, (3) metamorphosed.
(A) Shifting Counterpoint

A derivative is obtained by shifting the voices. The following classification ex
hausts all possible shifts:
(1) Vertical shifting upward or downward——hence vertical-shifting counterpoint:
Original Derivatives
(a)
ir3EB> XE îgM £
rJ/'rrrr r
In the third example, the upper voice is shifted underneath and the lower voice above,
a special case of the vertical shift known as "double counterpoint."
(2) Horizontal shifting, in which the time- intervals between the entries of the voices
are changed, hence horizontal-shifting counterpoint:
Derivatives
a I SE BE
(3) Vertical and horizontal shifting together, hence double-shifting counterpoint:
Derivatives
" 'i fiinam ^r+^

»i k J iiiîlk
^ ~n~ s mc
î=£ IDC mc
In this work double-shifting counterpoint is included in the divisions devoted to

horizontal-shifting and is explained parallel with it, as the methods of writing both are
similar.
INTRODUCTION 21
The subdivisions of shifting counterpoint are therefore:
(1) Vertical-shifting (upward or downward);

Shifting counterpoint (a deriva (2) Horizontal-shifting (changing relation
tive from the shifting of voices): ship between entries);
(3) Double- shifting (the combination of the
two preceding).
(B) Duplicated Counterpoint

A derivative combination is obtained by duplicating one or more voices in imperfect
consonances. Therefore the number of voices in the derivative combination is increas
ed: at the duplication of one voice, to three; at the duplication of two voices, to four:
m
m -r'.
An original three-voice combination yields derivatives of four, five, or six voices,

according to how many voices of the original are duplicated. Examples will be found
in Chapter XV.
The connection of this counterpoint with the vertical-shifting counterpoint is obvious:
each duplication is nothing but the vertical transference of a voice at an interval equal
to an imperfect consonance. Therefore the study of counterpoint admitting of duplic
ations is included in the divisions dealing with vertical-shifting counterpoint.
Of the various phases emunerated of complex counterpoint, forming the contents of
the present work, the detailed treatment of double counterpoint is of the utmost value
in music theory. It is of practical importance both in connection with counterpoint ad
mitting of duplication and with certain cases of multi-voice vertical-shifting counter
point, for instance in triple and quadruple counterpoint, especially at the octave. In
theoretical literature little reference is found to any aspects of vertical-shifting counter
point other than double, and still less to horizontal-shifting counterpoint, the study of
which, as a special department of shifting counterpoint,, is here presented for the first
time.
(C) Metamorphosed Counterpoint
The derivative is obtained by a process of transmutation. By metamorphosis is meant
such a change of the original combination as would correspond to its reflection in a
mirror . this is also known as "mirror counterpoint." To avoid ambiguous terminology
"metamorphosis" will be used only in this sense, and will not refer to the shifting of
voices in double counterpoint. Since metamorphosed counterpoint does not enter into
the plan of the present work it will not be considered further.
The statement has been made that the transference of voices is a characteristic of
shifting counterpoint. The changes that this counterpoint makes in a melody amount
only to its transference vertically to other degrees or horizontally to other measures or
parts of measures. Every other change, such as metamorphosis, augmentation, diminu
tion, made simultaneously with shifting, places the given combination beyond the scope
of convertible counterpoint, and shifting ceases to be a vital characteristic.
Another feature of complex counterpoint remains to be mentioned: the existence of
rules in relation to its various subdivisions and to simple counterpoint. The study of
the latter, in either the strict or the free style, is that of a system of rules to which every
union of voices must conform. In complex counterpoint the same rules apply to both
original and derivative combinations. The significance of the rules of complex counter
point is that if they are ignored in the original combination the derivative will show pro
gressions that violate the rules of simple counterpoint. The mutual relations of the as
pects of complex counterpoint and their relations to simple may be illustrated symboli
cally by the accompanying diagram, where the large circle represents the domain of
simple counterpoint, and the small intersecting circles, with the portions of their areas
coinciding, the various aspects of complex counterpoint.
From this diagram it is clear that the combinations used in complex counterpoint
must also belong to the domain of the simple, but not vica versa; portions of what is
permitted in simple counterpoint are found outside of the circles that represent the vari
ous aspects of complex. The intersections of the circles show that certain phases of
complex counterpoint may be combined, as was illustrated in the examples given.
This work is divided into two parts; the first deals with vertical-shifting counter
point and with counterpoint admitting of duplications; the second with horizontal-shift
ing and double-shifting counterpoint. Each part consists of two divisions, one devoted
to two-voice counterpoint, the other to three-voice. The investigations are limited to
two or three voices. More than these are found only as the result of duplications; they
are given at the end of the first and second divisions of Part One, where duplications
are found using a larger number of voices, up to six inclusive.
PART ONE
DIVISION A

Nissuna fiumana investigatione si po dimandare vera
scientia, s'essa non passa per le mattematiche
dimostrationi.
Leonardo da Vinci
Libro di pittura, Parte prima, § 1
PART ONE
DIVISION A
CHAPTER I
INTERVALS
The Notation of Intervals

§ 1. The subject of the study of vertical-shifting counterpoint consists of an invest
igation of those combinations from which derivatives are obtained by means of shifting
the voices upward or downward. Such alterations in the relative positions of the voices
are effected by changing the intervals that are formed by these voices in combination.
For the analysis of these changes the best method is that of mathematics, by which the
quantitative differences in the sizes of intervals are expressed in figures; mathematical
operations are derived therefrom. For this purpose it will be necessary to employ a more
accurate method of indicating intervals than that in general use. This new method, used
in the present work, consists in taking the interval between two adjacent scale-degrees,
i.e. a second, as the unit. The interval is then indicated by a figure showing the number
of these units that it contains. The unison is indicated by O, since in it this quantity
is equal to zero.
etc.
TT ^
2
Therefore each interval is represented by a figure that is 1 less than its usual num
erical designation: a third by 2, a fourth by 3, etc.
Addition and Subtraction; Negative Intervals

§ 2. By indicating intervals according to this method processes of addition and
subtraction become possible. An interval may be added to another, either up or down.
In the former case the upper voice is shifted upward, in the latter the lower voice down
ward. In both cases one voice moves away from the other. For example, a fifth added
to a fourth—
zm
w Wr 3 + 4=7
25
gives an octave, an interval equal to the sum of the terms: 3+4 = 7. Addition is also
possible both up and down at the same time; here the result is the sum of three terms:
i+3 Js
±=4 + 3 + 2 = 9
Tf?
Other combinations of the same terms yield the same result; the order in which they
are taken does not affect the total:
+3
^ _r Q a Q . «- etc.
^
+2 I +3 F^T^P f^f
§ 3. The reverse process, subtraction, causes the voices to approach, i.e. the higher
voice is shifted downward or the lower voice upward, or both. For example, subtracting
a third from an octave leaves a sixth:
M*j t-Tr 2= 7-2 = 5
If the subtracted interval equals the value of the first interval the result is 0, i.e. a
unison:
^L.l.fjbph 4-4=0
r-T1 w
If the subtracted interval is greater than the value of the first interval the result is
a negative quantity:
4 -5 = -1
^
§ 4. A negative quantity therefore refers to an interval of which the lowest tone be

longs to the upper voice and the highest tone to the lower voice. These intervals are
termed negative. The same mathematical processes may be applied to them as to pos
itive intervals.
§ 5. It is possible to regard the addition and subtraction of intervals in the algebraic
sense; i.e. to consider both processes as addition, in which the amounts concerned may
be either positive or negative quantities. Results so obtained are algebraic. The sum
of two or more positive numbers is only a special case.
§ 6. The order in which the terms are taken does not affect the total. Therefore when
the two voices shift simultaneously it will be found more convenient to add their alge
braic values all at once, not to add each item in turn to the given interval. Suppose that
the given interval is a fourth and that one of the voices shifts -9 and the other +1. The
sum of these quantities is -8. Adding - 8 to the value of the interval 3 gives
3 + (-8)= -6, i.e. a negative sixth.
INTERVALS 27
-9
-9- 3+ 1 -9=-5
PP|
If other shifts are substituted of which the algebraic sums are the same the result
remains unchanged:
I -to
^.3-3- 5= -5 * etc. 3 + 2-10:
^
Compound Intervals
§ 7. If an interval contained within the octave limits is increased by one or more
octaves an interval is obtained that is termed compound, in relation to the first.
To separate the voices forming an interval by an octave, add 7 to its absolute value:
A
S
.1L/1
-9— ^C -^
K*f
mm
"^
m= etc.
-7
?*# ^zteS
is: rz= ©*. 4«- etc.
a^ hh:
-7
H2:
¥«
To separate the voices two octaves, add 14 to the absolute value of the interval;
for three octaves add 21, etc., in multiples of 7:
14 A-2-!-,
0 7
c*
-o- -o-
15
1 8
-o" -& -©- -©-
§ 8. The following table is a list of simple and compound intervals within the limits
of four octaves:
Unison 0 7 14 21
Second 1 8 15 22
Third 2 9 16 23
Fourth 3 10 17 24
Fifth 4 11 18 25
Sixth 5 12 19 26
Seventh 6 13 20 27
§ 9. To find what interval within the octave limits corresponds to a given compound
interval, divide the latter by 7. The remainder will be the desired interval and the
quotient will indicate by how many octaves the voices are separated. Suppose that the
given interval is 30. Dividing this by 7 gives 4 as a quotient, with 2 as a remainder.
The desired interval is therefore a third, and the voices in the given interval are separ
ated by four octaves in addition to the third.
§ 10. The propositions following are based on what has been established. Consider
ing each voice separately, the vertical shift in one direction is a positive operation, in
the reverse direction a negative operation. The voice for which the upward shift is re
garded as a positive operation will be termed upper, first, and indicated by the roman
numeral I; that for which the positive operation is the downward shift will be termed
lower, second, and indicated by the roman numeral II.
The positive and negative shifts of the voices may be represented by this diagram:
I II
When the voices are arranged in the order i. the intervals formed by their union are
positive; in the order V they are negative.
If two voices forming an interval a shift by intervals of which the algrebraic sum
is ±s, then from a is obtained a + (±s) (§ 6). The same result is obtained if one voice
shifts at ±s and the other remains stationary, s being the algebraic designation for the
interval by which the voice is shifted up or down.
Successive Series of Intervals;

Division into Two Groups: lint. and 2int.
§ 11. Intervals may be put in a successive series, such that from the unison (0)
positive intervals are ranged on one side, negative on the other. In the following series
the consonances are in bold-face figures, with p. or imp. added, for perfect or imperfect.
etc.
iM^
0 q a o ij p fj ttt
u^LAAAA
r r pp r r r r
etc.
-9 -8 -7 -6 -5 -4 -3 -2 -1 0 23456789
imp. p. imp. p. imp. p- imp. p. imp. p. imp.
Positive and negative intervals are divided into two groups; (1) intervals that appear
in three forms: perfect, augmented, diminished; and (2) intervals that appear in four
forms: major, minor, augmented, diminished. The first group consists of 0, 3, 4, 7 and
the corresponding intervals beyond the octave. The second group consists of 1, 2, 5, 6
and their compounds. Intervals of the first group are indicated 1int., those of the second
group 2int.
INTERVALS 29
o
'int. ~n~ etc.
■o-
0 3 4 7 10 11 14
P- p-
3E
*int ~TT~ etc.
e- -©- -O- -o- •O- -O-
1
imp
5mp. 6 8 9 12 13
imp. imp.
The first group includes the perfect consonances, the second group the imperfect.
Obs. These groups of intervals have other characteristics. For example, each 'int.,
counted upward from the first degree of the major scale, is identical in size to the same
interval counted downward; both intervals are perfect:
-f*-
£r$r ^ ** etc.
*
On the contrary, the 'quality of each 2int. is changed, under the same conditions;
those counted upward are major, those downward, minor:
Maj.
mbj. »?■ r-
Maj. Maj. "J
tT: frS —£» ;» etc.
Mln. Mln. Mln. V.
Mln.
Also notice that the first four notes of the harmonic series include all the perfect
intervals of the first group; the fifth note forms one of the intervals of the second group
from each of the preceding notes:
:¥
S 3E
§ 12. The distance between two given intervals in the successive series is deter
mined by the interval at which one voice is shifted, the other remaining stationary, a
process required in order that from a given interval another may be derived. There inter
vals, for example:

lA are separated by a fourth; these: B
4 T -2 2
by a fifth; adjacent intervals from the series: ~ra n_::za

P ~f a second, etc.
-6 -5 1 2
§ 13. If a positive interval (termed here a) is added to a given interval, the interval
obtained will lie in the successive series to the right of the given interval at the dis
tance indicated by a. If a negative interval, -a, is added, the interval obtained will be
will be found to the left of the given interval at the distance -a. For example, adding
a positive sixth to a third gives an octave (2 + 5 = 7), lying a sixth to the right of the
*£.
third: Conversely, adding a negative sixth to a third gives a negative
fourth (2 -5= -3), lying a sixth to the left of the third: =zz k
Order of Intervals in the Successive Series
§ 14. It is desirable to dwell at some length on certain peculiarities in the order of

intervals of the successive series which will be referred to later on.
(1) The perfect and imperfect consonances alternate, in both directions.
(2) Two consonances are adjacent, 4 and 5, but not two dissonances. (Here the
possibility is not considered whereby a consonance can be changed to a dissonance
by chromatic alteration.) Thus the fifth and sixth have a dissonance on one side only;
all the other consonances have dissonances on both sides.
(3) On both sides of each dissonance is found a consonance, one of which is perfect,
the other imperfect.
(4) Of two consonances found at equal distances, right or left, from a dissonance,
one will always be perfect, the other imperfect; e.g.:
r f f ipp=p i H- "t i l
5 6 7 -7 -1 5 -2 1 4
imp. p. imp. imp. p-
§ 15. In the following the consonances only are taken from the success lve series:
Consonances
negative positive i I
etc. etc.
fwm
-9 -7-5-4-2024579
u) ga a a g g g
P
imp. p. imp. p. imp. p. imp. p. imp. p. imp.
The calculation of the distance between positive consonances only or negative only
proceeds as follows:
(1) Two consonances of the same group (i.e. both perfect or both imperfect) are sep
arated from each other by a 'int.
INTERVALS 31
(2) Consonances of different groups (i.e. one perfect, the other imperfect) are sep
arated from each other bv a 'int. For example, consonance 2 (imp.) is separated from con
sonance 5 (imp.) by 3(= 'int.); consonance —7 (p.) is separated from —11 (p.) by 4 (= lint.).
Conversely, consonance —5 (imp.) is separated from —7 (p.) by 2 (= 2int.); consonance
-4 (p.) from -9 (imp.) by 5 (= 2int.), etc.
§ 16. Exceptions to the foregoing statements are not found as long as positive con
sonances only or negative only are compared. But in comparing positive consonances
with negative the following sole exception is encountered:
Two fifths, negative and positive (thus corresponding to a compound interval) are
separated from each other by a ninth (or at an interval larger than a ninth by an octave,
or two octaves, etc.). Since the ninth is a 2int. the case represents an exception to
what was stated in § 15, (1).
8 Mint) 15Mint.)
-4
p.
15(=2int
i
-4
P.
With these exceptions the statements in the preceding section relative to the dis
tances between consonances of the same group and those of different groups apply also
to all cases where one interval is positive and the other negative.
§ 17. Proceeding to dissonances:
Dissonances
negative
A r positive
etc.
SJ
hi etc.
TTTT
-10 -8 -6 -3 3 6 8 10
First the distance is to be measured between positive dissonances only or negative

only.
(1) The second and its compounds (1, 8, 15 etc.) are separated from one another by
a 'int.:
7(= 'int.)
7(- 'int.)
1 8 -8 -1
(2) The other dissonances— the fourth, seventh and their compounds (3, 6, 10, 13,
etc.)— are also separated by a 'int.:
3(= 'int.)
4(= Mnt.)
m I -6
(3) The second and its compounds are separated from the other dissonances by a 2int.:
5(=sint.)J
2(=*int)
i
-8
§ 18. Next to be considered are the mixed cases where one dissonance is positive
and the other negative. Here the statements in § 17 regarding dissonances are present
ed in reverse order:
(1) The second and its compounds are separated from one another by a 2int.:
2(=2int). | 9t2int) |
P5
-8
(2) The other dissonances (3, 6, 10 etc.) also are separated from one another by a int
9(=*int.) J 12(=2int.) I
-3
=P
(3) The second and its compounds are separated from every other dissonance under
the same conditions (that one interval is positive, the other negative), by a 'int.:
4(='int.) I lit Int) I
§ 19. If under one successive series of intervals is placed another so that a new
interval a comes directly below O (unison) in the upper series, then each interval in
the lower series will be equal to the algebraic sum of the interval above it + a. Let m
equal any interval in the upper series and n the interval in the lower series directly
underneath m; then m + a = n. In the following—
INTERVALS 33
| A J j-J gj; ^
Jiii
m
-5-4-3-2-1 0 1 2
cirrr
3 4 5, 6
i
t[iwmjfM'r'EsM
-9 -8 -7 -6-5-4-3-2-1 0 1 2 3
under O in the upper series is placed a - ^4. Taking for example 7 in the upper series,
m = 7, and adding -4 gives below it n - 3 (7 - 4 = 3).
Such comparison of two series of intervals is necessary in working out exercises in
vertical-shifting counterpoint.
CHAPTER II
THE VERTICAL SHIFTING OF CONTRAPUNTALLY COMBINED VOICES;
THE SHIFTS; INDEX OF VERTICAL-SHIFTING COUNTERPOINT
Notation of the Vertical Shift: v, vv. Formulas

for Original and Derivative Combinations
§ 20. In the preceding chapter the shifting of the voices forming separate intervals
was investigated. The present subject is the shifting of the contrapuntal union of mel
odies. To show that two voices form correct counterpoint the Roman numerals indicating
the voices (melodies) will be used, united by the plus sign: I + II. Each voice in the
derivative combination is indicated by the same figure that it had in the original. I + II
is the formula for the original combination.
Obs.— The sign +, used as indicated, refers to addition as meaning a combination of

voices (two-voice addition, multi-voice, etc.), and is to be taken in this sense only when
employed in connection with the roman numerals for the voices.
§ 21. The letter v, for "vertical" (plural vv) refers to the vertical shift of a voice,
and is placed to the right and slightly above the roman numeral corresponding to this
voice. The number indicating the direction and interval of the shift is united to v by the
sign of equality. For example, the expression iu-5 means that the upper voice shifts
a sixth upward; I^- + II r_ indicates a shift of the upper voice a third downward,
the lower an octave upward; the sign + means that the voices so shifted form correct
counterpoint.
Such an expression, indicating shifts of the voices united by the sign +, is the formula
of the derivative combination. Formulas for derivative combinations may differ, but that
of the original can only be I + II.
§ 22. When a voice remains unshifted neither the letter v nor v = 0 need be associated
with its Roman numeral. Therefore the expression I + II " and iv~° + II" are
synonymous, both meaning that the upper voice remains in place but that the lower is
shifted a second downward.
The Shifts
§ 23. The relationship of melodies in the derivative combination may present one of
the following three cases:
(l)The Direct Shift. — In this the melodies retain their relative positions; the upper
voice stays above, the lower underneath, though they may approach each other or recede
(Ex. 1). This shift may be illustrated by the diagram:
Orig. Deriv.
1 I The symbol for the direct shift is: ; ;
II II
34
THE VERTICAL SHIFTING OF CONTRAPUNTALLY COMBINED VOICES 35
Ex. 1
Original 1st Derivative 2nd Derivative

Vr-2
&
H ^ i^* "-> II
(2) The Inverse Shift. — Here the melodies change their relative positions; the upper
Orig.
I
voice goes below, the lower above. Diagram: ,r
Symbol: Zx^ This shift is what is commonly known as "double counterpoint,",

but it is only a special case of the vertical shift.
Ex. 2
Original Derivative
11:--9
JU^^li
I
Trr
IE
r- T XI"
3E3
su f^ >u
I
-Q_
.oo-
II
The final consonance in this example is the unison (0); in the derivative combination
it can occur in either the direct or the inverse shift.
(3) The Mixed Shift. — This is partly direct, partly inverse:
Ex. 3
Original Derivative
II J.J- J.I J.J

S^
•TOfrJJj ""
"oF-V f f^r r r =^ IV-.7 •
f m
Relations of Original to Derivative Intervals
§ 24. The combination of melodies I + II forms a series of intervals:

a, b, c, d, . . . . n. If one of the voices shifts at ±s, that is, takes for the derivative
formula lv = ±s + II or I + IIV = ±S (§ 21), then a new series of intervals is obtained:
a + (±s), b + (±s), e + (±»), . . . . ■• (% 10). For example, tfie original combination of
Ex. 1 represents the series of intervals: 4, 7, 6, 5, 4, 2, 3, 4.
Original (Ex. 1)
HiII '^JJJ i^H 3r

7 6 5 4 2 3
Its first derivative is I + II v . Adding 3 to each original interval gives the inter
vals of the derivative combination.
Intervals of the original combination: 47654234
+3 +3 +3 +3 +3 +3 +3 +3
Intervals of the Derivative combination: 7 10 8
Derivative
3E r^fîig
II v^3 10 9 8 7
In the successive series (§ 11) each of these derivative intervals lies trom its own
original at a fourth to the right (§ 13). In the same way a series of intervals could be
obtained for the second derivative, Iv-_2 + II, by adding -2 to each interval of the
original.
In Ex. 2 the series of intervals for the original combination is: 7, 6, 5, 2, 3, 7, 9.
Original (Ex. 2)
Jfe^
fi
7 6 5 2 4
f=f
Derivative formula: I + II v_. Adding -9 to each interval of the original com
bination gives a series of negative intervals, each of which lies from its own original
in the successive series at a tenth to the left (§ 13), and showing that the shift is in
verse.
Intervals of the original combination: 76524579
-9 -9 -9 -9 -9 -9 -9 -9
Intervals of the derivative combination: -2 -3 4 -7 -5 -4 -2
Derivative
m T -2 -3 -4 6=
m
-4 0
I
§ 25. From the definition of the shifts it follows that at the direct shift the deriva
tive intervals take the same signs as those of the original, and at the inverse shift the
opposite signs.
Index of Vertical-Shifting Counterpoint (Jv)

§ 26. To obtain the result of the simultaneous shifting of two voices it is necessary
to add to each interval the algebraic sum of the quantities indicating the shifting of
either or both voices, i.e. the algebraic sum of their vv (§ 10). This rule is of general
application, since the idea of algebraic sums includes those cases where one of the
voices has v = 0, that is, remains stationary.
§ 27. The algebraic sum vv of two voices contrapuntally united is termed the index
of vertical-shifting counterpoint, and is indicated by Jv (plural JJv), J standing for
"index" and v for "vertical shift." In distinction to the sign v, referring to the indi
vidual voice (§ 21), the sign Jv can refer only to the combination of two voices.
§ 28. To indicate that a given shift of voices applies at a certain index, the
formula of the derivative combination will be put in parentheses and after it Jv; e.g.
(Iv + IIV ') Jv= — 9. When the formula of the original combination is presented
in a similar manner it will mean that the voices admit of a shift at the index indicated;
e.g. (I + II) Jv= 2, (I + II) Jv= -2, etc. If one and the same combination admits of shifts
at two or more indices their respective figures are placed after the sign of equality and
are separated by commas. For instance, Ex. 1 admits of two shifts: at Jv= 3 and at
Jv= —2; this is indicated: (I + II) /v=3, —2. Such a Jv, referring to a single original
but stating the conditions of two or more indices, is termed a compound index. A com
pound index may be double, triple etc., according to how many indices are united in it
(this has no reference to double or triple counterpoint). A compound index is always
printed in the singular: Jv, but if several indices are to be considered that refer to
different original combinations the sign is printed in the plural: JJv. If, for example,
it is required to list the indices that correspond to perfect consonances, the expression
will read: JJv = ±3, ±4 etc., equivalent to Jv = 3, Jv = -3, Jv = 4, Jv - -4 etc.
§ 29. Shifts of voices I + II at a given index may be replaced by other shifts that
give as a result the same series of derivative intervals if the algebraic sum of these
shifts remains without change (§ 26). A derivative combination will therefore be re
produced on other degrees. In this way one and the same index can generate various
shifts of voices and therefore can belong to different formulas of the derivative com
bination.
§ 30. It is possible to get a derivative combination in which one voice remains un-
shifted. In this case the other voice must take: v - J v. This follows from the fact that
the index is equal to the algebraic sum vv of both voices (§ 26).
The Formula m + Jv = n; Inferences Therefrom

§ 31. Adding the value of the index to an interval of the original combination gives
the corresponding interval of the derivative (§§ 26-7). Indicating the original interval
by m and its derivative by n, the relationship is expressed by the equation to + Jv - n.
§ 32. From this equation it follows that-
(1) If to = 0, then n - Jv; i.e. from the unison is obtained an interval in the derivative
equal to the value of the index.
(2) If to and Jv are equal but have opposite signs, then from m-Jvis obtained n - 0;
i.e. from an interval equal to the index but having the opposite sign is obtained a unison
in the derivative combination.
(3) Since at the inverse shift to + (~Jv) = -n (§ 25), then n + (-/t>) = -to; i.e. in
double counterpoint, from the derivative intervals are obtained intervals equal to the
original. At Jv = -7, for example, from a fourth is obtained a fifth in the derivative
(3 - 7 r -4), and from a fifth a fourth (4 - 7 = -3). At Jv - -9 from a third is obtained
an octave (2 - 9 = -7), and from an octave a third (7 - 9 = -2) etc.
§ 33. Comparing the formula to + Jv = n with the statements in §§ 12 and 30 the con
clusion is that in the successive series of intervals (§ 11) a derivative interval lies from
an original at an interval equal to Jv. If Jv is positive this distance will be to the right;
if negative, to the left (§ 13).
Conditions of the Shifts
§ 34. The index (Jv) may be of positive value, of negative value, or may equal zero.
The conditions under which Jv yields the direct, the inverse, or the mixed shift are next
to be investigated. From the equation m +/v= n and what was stated in § 25
it follows that—
(1) If m and Jv are both positive or negative the shift is direct.
(2) If one of the quantities m or Jv is positive and the other negative either the direct
or the inverse shift is possible, depending on the value of the intervals in the original
combination relative to the value of Jv, namely:
(a) At m with absolute value greater than Jv the shift is direct.
(b) At m with absolute value less than Jv the shift is inverse.
§ 35. From the fact that in the derivative combination the unison (0) may be found
in either the direct or the inverse shift, the conditions of the shifts can be
stated as follows:
(1) If m> Jv, then with like signs for m and Jv the shift is direct.
(2) If m< Jv, then with like signs for m and Jv the shift is inverse.
§ 36. These principles for the shifts apply without exception to all cases, whether
the intervals of the original combination are positive or negative. But since in practice
it is advisable not to cross the voices but to use only positive intervals in the original
combination, the rules for the shifts applying to the latter may be formulated thus:
(1) At a positive Jv the shift is always direct.
(2) At a negative Jv the shift may be direct, inverse, or mixed. Condition of the direct
shift: m > Jv (§35, [l]);of the inverse shift: m > Jv (§35, [2]). The union in the same
original of these and other conditions forms the mixed shift (§23, [3]).
Limiting Intervals; Their Signs (<, >)
§ 37. A successive series of original intervals for a positive index (giving always
the direct shift, §36,[1]), starts with 0:
Original Intervals
Jv;4.
J-J. etc.
PP
i_LAAA
4
mm etc.
§ 38. A successive series of original intervals for a negative Jv, in order to yield
the direct shift, must start with an interval equal to the absolute value of Jv. Therefore,
at the direct shift, an interval equal to the absolute value of the index is the limiting
interval for approaching voices of the original combination; it is indicated by the sign <.
At Jv=—2, for example, the voice must not approach closer than a third:
Original Intervals
Jv--2.
,i j IAA + .etc.
2< PP^
j j J J j 4 etc.
P
at ,/t> = —3, not closer than a fourth, etc.
Original Intervals
Jv=-3. etc.
I=PP£
1 J J J j
etc.
§ 39. A successive series of original intervals for a negative Jv, in order to yield
the inverse shift, must start with 0 and end with a positive interval equal to the absolute
value of Jv, showing in this case the limiting interval for receding voices; it is indicated
by the sign >. For example, at Jv= —7 (double counterpoint at the octave) the voices
must not recede from each other by more than an octave:
/ Original Intervals ^
i Jii^
rrn, r r r r
0 7>
X P3 £
HO [J [) H1
-7 0
at Jv = - 11 (double counterpoint at the twelfth), by not more than a twelfth:
f Original Intervals >
r r t r i r i r r i. r r ll>
x:
-11
Obs.— Limiting intervals for receding voices are necessary only for indices giving
the inverse shift. For others the series of original intervals can be continued to the
right as far as required.
§ 40. If in an original combination at a negative index are taken intervals some of
which are less than the absolute value of Jv and some greater, the result is the mixed
shift (§23,[3]).Cf. Ex. 3, written at Jv = —7. In the second measure, where in the original
the voices exceed the limits of an octave, the shift is direct.
Signs for the Shifts. (<, >, <>, <)
§ 41. To indicate the shift of a negative Jv the same signs are used as for the
limiting intervals; they are placed after the figure for the index. The sign of the limiting
interval for approaching voices (<) will be used for the direct shift; that for the limiting
interval for receding voices (>) for the inverse shift. Placed in succession (<>) the
signs will refer to the mixed shift. For example, Jv = - 5< means that this index gives
the direct shift, i.e. that the series of original intervals relevant to it starts with a
sixth; the limiting interval for approaching voices is also shown: 5<. The expression
/w=-5> means that the index gives the inverse shift and that the series of original
intervals, while starting with a unison, ends with a sixth; the limiting interval for re
ceding voices is indicated by 5>. A> = -5<> indicates the mixed shift, i.e. the presence
of both the preceding cases, more or less in alternation. When one sign is placed above
the other (^) they refer to two negative indices of identical value but with different
shifts. Therefore, JJv= — 45 serves for two expressions: A>=-4< and Jv= -4>.
§ 42. For positive JJv it is not necessary either to indicate or to establish distance
limits for the voices.
Jv Equal to Zero
§ 43. The positive and negative indices have been considered; the index equal to
zero (§ 20) remains to be mentioned. This Jv, like all the others, can denote different
shifts of the voices; for example (Iv= 9 t IIV= -9) Jv = 0; (Iv= "* + IIV= 3) /v= 0, etc.
The result of such shifts is to yield a series of intervals identical in value to those
of the original.
§ 44. It is possible to regard every recurrence of a two-voice combination on the
same degrees or its removal to other degrees as a shift at Jv= 0. The rules of simple
counterpoint are the rules of Jv= 0, and therefore simple counterpoint can be understood
as a special case of the vertical shift.
§ 45. The important idea implied by the use of the symbol Jv simplifies the study of
vertical-shifting counterpoint; it yields numerous possibilities of voice-shifting with a
comparatively small number of indices.
CHAPTER III
THE GROUPING OF INDICES
List of Indices
§ 46. The indices of which the conditions are presently to be examined correspond
to intervals of the successive series taken within the limits of three octaves. In the
following list they accordingly fall into three groups, with seven indices in each group.
Beginning with }v = 0 the positive indices proceed to the right up to nearly an octave;
the negative indices down to the left to two octaves. Shifts of negative indices are
indicated by their proper signs (§ 41).
JJV^
Inverse Shifts Direct Shifts
14=- -13=- -12=- -11:=- -io>- -cp- -8=-, -7 -3<-2-=-i«=, o 123456.
§ 47. The positive indices end with Jv=6 for the reason that to continue to Jv= 7
would merely shift a voice of the original combination an octave higher:
Ex. 4
Derivative
Original
TVr7
1 U^JH
1^ mkMM
TT- J3Z
r 'r '
i]
.*' r 4 3tE
np-r r f ~^r
1
It is obvious that anything conforming to the conditions of simple counterpoint ( i.e. a
combination at /t> = 0) would also be correct for Jv= 7, so this index does not require
special rules. Similarly, to separate a derivative combination at Jv = 1 gives a deriv
ative at Jv = 8:
Ex. 5
gga '^"
wm
1JS3 |ra
feB E§ m
A combination at Jv=2 will serve equally well for /t>= 9, one at Jv = 3 for /u = 10; the
same relation holds between /t/=4 and /i> = 11, etc. It is therefore unnecessary to' for
mulate rules for positive indices equal to compound intervals; those equal to the cor
responding simple intervals can be used instead.
41
§ 48. Proceeding to negative indices: those of the values —1< , —2< and —3< will
always refer to the direct shift, because those of the same values for the inverse shift
would result in limiting the movements of the voices to too narrow a range (§ 38). For
the same reason it is advisable to regard -4, -5 and -6 as indices also applying to the
direct shift, though these values also admit of the inverse shift. In the former case they
are indicated -4<, -5<, -6<; in the latter -4>, -5>, -6> (§ 41).
§ 49. The next three indices —4>, — 5>, — 6> begin the series of shifts in double
counterpoint; the fifth, sixth and seventh. The indices beyond them to the left, beginning
with }v= —7 and continuing to Jv= —13 inclusive, are regarded as inverse shifts (double
counterpoint at the octave, ninth, tenth etc.). Only exceptionally will they be treated as
direct, in which case they will take the sign< .
§ 50. One index remains: /t>=— 14, but neither this nor the ones beyond (/t?=— 15,
Jv = —16 etc.) require special study. Any combination written at Jv= — 7 can be shifted
at Jv= —14, since here the derivative is only separated an octave as compared to the
derivative at Jv= —7; similarly with a shift at Jv- —21, where the voices are separated
two octaves. The same applies to a shift at Jv = — 8, also possible at Jv = — 15 and
Jv= —22 etc. Therefore by double counterpoint at the octave will be understood not only
}v=—7 but also Jv = — 14 and Jv=—21; by double counterpoint at the tenth, JJv- — ,9,
-16, -25 etc. (cf. table, § 6).
The only difference between these cases concerns their limiting intervals, which
always are equal to the absolute value of the index (§ 38). Thus, at Jv= — 14 the voices
may be separated by two octaves, but at Jv= —7 by not more than one.
Columns of Indices
§ 51. The indices listed in § 46 presented three series of figures, seven in each.
Putting these series in numerical order, one underneath another, gives the seven columns
below. Four of these, shaded, contain indices corresponding to intervals of the second
group (nnt.). The other columns, unshaded, contain indices corresponding to intervals
of the first group (int.).
-11= -10;
-42
3 4
These columns, except the first, are numbered; each reference number corresponds
to the figure of its lowest Jv. The first column to the left is not. counted, it will be
referred to as the zero column. It contains Jv = 0, i.e. the index of simple and not one of
vertical-shifting counterpoint (§ 44). The upper index in this column, Jv- —14, represents
the same double counterpoint at the octave as does the middle index, Jv = —7, hence
it is not necessary to discuss it further as a distinct index. The only essential index in
this column is Jv=—7. Therefore the zero column will often be referred to in an in
complete form, restricted to the middle index.
§ 52. In these seven columns are placed the indices corresponding to the intervals
of the successive series, taken in a range of three octaves; the positive indices within
the limits of one octave, the negative of two. If more indices are needed to indicate an
octave extension on either side, each column will contain four indices instead of three;
for an extension of two octaves each will have five indices, etc_ Therefore the value of
THE GROUPING OF INDICES 43
any index determines its place in one of the seven columns. But, as already stated, for
practical purposes three indices in each column are enough.
§ 53. It is not difficult to remember the indices contained in each column. The lowest
indices are all positive, each of them corresponds to the number of the given column.
The middle and upper indices are all negative and may be easily learned if attention is
paid to the following relations:
(l)The sum of the absolute values of the lowest and middle indices in each column
is 7.
(2) The sum of the absolute values of the lowest and upper indices is 14.
(3) The difference between the absolute values of the upper and middle indices is 7.
Octave Relationship of JJv
§ 54. The lowest Jv of each column corresponds to a positive interval within the
limits of an octave; the middle Jv to a negative interval, also within an octave, and the
upper Jv to a negative interval beyond the octave limit. If —7 is added to the lower Jv
the middle one is obtained; if —7 is added to the middle Jv (or —14 to the lower) the
upper Jv of the same column is obtained. Therefore if in a combination I + II one of the
voices remains in place while the other shifts in conformity to each of the three JJv of a
given column, shifts of the melody will result, to the same degrees though in other
octaves. All three indices of the same column are thus in an octave relationship. This
is clearly shown in the following table, using the indices of the third column:
Original
U-UlAlii
a ri ti
r r r r 77
ri a T' V p T7
It ri a a a a—a £
-11=
TV;-n
0 m o
-4: 3g fir. -±
U-AJJ
r i-t ¥ % r r r i1
WS-lA. Aiiiiiii .i e r7
m r r f r-wm
§ 55. In each column the lower Jv has the direct shift, the upper the inverse shift. Of
the middle numbers, three: -6^, -5^ and -4^ have both the direct and the inverse
shift, and three: -3<, -2< and -1< have the direct shift only (§ 46).
§ 56. If a derivative is written at a middle Jv and one of the two voices of the
derivative is separated an octave, a derivative is obtained at an outer Jv of the same
shift as the middle Jv.
Take for example the table in § 54. Writing a combination at Jv=—4< (i.e. with the
direct shift) and separating a voice of the derivative an octave gives another derivative
at Jv=3, i.e. at the lowest Jv of the same column, also a direct shift. Taking a deriv
ative at /i>=-4>, i.e. at the Jv giving the inverse shift, and separaring a voice an
octave, gives a derivative at Jv = — 11, the highest Jv of the same column and also an
inverse shift. Similarly, a derivative at /t>=— 5<, separated an octave, yields another
at a /i>in its own column: Jv = 2; one at /t>= — 5> in the same way a derivative at Jv = —12.
The same relation holds bewteen Jv = — 6< and Jv = 1; Jv = — 6> and /v= —13, etc.
In general, every combination written at a middle Jv yields a valid derivative at each
outer Jv of the same column and with the shift indicated. But the contrary is not neces
sarily true; writing at an outer Jv may be unsuitable for shifts at the middle, as will
be seen later (§ 72).
§ 57. The statement was made in § 51 that the unshaded columns contained JJv
corresponding to intervals of the first group (înt.), and those shaded, JJv corresponding
to intervals of the second group (int.). On this fact is based a division of indices into
two groups that has great importance for the whole study of vertical-shifting counter
point. In the first group are JJv corresponding to the Mnts. (i.e. those in columns 3, 4
and the zero column). In the second group are JJv corresponding to the înts. (i.e.
columns 1,2 5 and 6). When it is necessary to refer the characteristics of a given Jv
to either of these groups, use will be made of the indications '.Jv and *Jv.
Grouping of the Columns in Pairs
§ 58. In the following table six columns (1-6) are grouped in pairs; the zero column
is isolated. In each pair of columns the sum of the lowest indices equals 7, and each
lowest index is of the same value as the middle index in the other column of the same
pair, but with the opposite sign; the lowest Jv is positive, the middle Jv negative. The
zero column is unrelated and cannot pair with any of the others.
-n=> -10> -14^

-4== -3< -7^
3 4 0
2nd pair 3rd pair

§ 59. The lJJv in the zero column correspond to perfect consonances; in the third
pair of columns some JJv correspond to perfect consonances, others to dissonances of
the first group. The lJJv in the first pair of columns correspond to dissonances of the
second pair to consonances of this group, i.e. imperfect.
§ 60. Classifying the JJv in this way into two groups, arranging them in columns
showing the octave relationship, and then combining the columns in pairs that bring the
indices in logical order—all this facilitates their study and recall.
Determination of the Value of an Original

Interval and of the Index
§ 61. Reference to the formula m + Jv = n (§ 31) shows that a derivative interval (n) is
equal to an original interval (m) to which Jv has been added; i.e. the equation constitutes
a definition of what the two other quantities m and Jv are equal to. Taking-
(a) m- n - Jv= n + (-Jv)

(b) Jv = n - m= n + (-m)
i.e. (a) the original interval (m) equals the derivative interval (re) plus Jv taken with the
opposite sign; (b) Jv equals the derivative interval (re) plus the original interval (m)
taken with the opposite sign.
§ 62. This will be illustrated by two examples determining the value of the original
interval (m) according to the derivative (re) and Jv (equation (a), § 61).
(1) From what interval is obtained a derivative tenth at Jv=A? Since Jv is positive
and has the direct shift, the derivative tenth (9) is also positive. Adding to it the index
with the opposite sign (equation (a), § 61) gives: 9 — 4= 5. Therefore at Jv= 4 a deriv
ative tenth is obtained from a sixth.
(2) From what interval at Jv= -11 is obtained a derivative ninth? Since Jv= -11 has
the inverse shift (§ 49) a derivative ninth is negative (-8). Adding to this quantity
(according to the same equation)the index taken with the opposite sign gives: - 8 + 11=3,
i.e. a derivative ninth is obtained from a fourth.
§ 63. Considering now the determination of Jv according to an original and a deriv
ative interval (equation (b), § 61); given the original and derivative combination shown
in Ex. 6, at what Jv is the shift?
Ex. 6
Original
Derivative
I t=: I
£
I: «>rJ4 3SI ^^ 4V J. " * u Ji?ijJ-
1W
3C
fW
It would be possible to determine Jv by indicating the shifts of both voices and taking
the algebraic sum of these shifts (§ 27), but it is shorter to make use of equation (b),
§ 61. Take one of the intervals of the derivative, for example the octave that appears
at the entry of the second voice. Since the shift is direct this octave (7) is positive. At
the corresponding place in the original is a tenth (9). Adding with the opposite sign to
the derivative octave gives: Jv= 7 —9= —2.
§ 64. To determine Jv according to the intervals of the original and the derivative
(equation (b), § 61) is best in practice. This will be further illustrated by several ex
amples from the works of Palestrina.
Given the original and derivative (Ex. 7), required to find the value of Jv. The shift
is inverse. At the entry of the first voice m= 2, re= -9. Hence fu= —9 —2= -11 (double
counterpoint at the twelfth).
Derivative
II
X
g - ■
^yî **~ JDL
X£
1 FnTf rrrr
In Ex. 8 the derivative combination is accompanied by supplementary parts, these
are printed in small notes.
Ex. 8
Original Palestrina
2)
■r/.'^f'WrllJT^".^ re
etc.
Here the shift is inverse. At the entry of the second voice m- 2, n- 5. Hence Jv= —5
—2= —7 (double counterpoint at the octave).
In Ex. 9 the derivative is transferred to another part of the measure, otherwise it is
similar to the preceding, with index of the same value, —7.
Ex. 9 Derivative Palestrina

3)
a b
W Original ,
^ttt rpf r IE
fr**F=5ff
ijnU 1duH^h-^
I Jl 3E
2
ppf
In Ex. 10 the first note of the second voice in the derivative is abbreviated to half
its original value; this is often found, especially where imitation is present. Beginning
with the second note the voice proceeds without further change.
Ex. 10
^ Original Palestrina
•
£*=
II
s /JJ'/i'/ii'i'iVJ, m etc.
etc.
Derivative
■ fiiTa) tiJ^g1
etc.
The initial intervals are: m- 7, n- 0; ]v = —7 (double counterpoint at the octave). In
the third measure of the original the voices cross, resulting in the intervals —2 and —1.
Adding —7 gives —9 and —8 at the corresponding places in the derivative. The star
indicates where shifting counterpoint passes into simple counterpoint.
In Ex. 11 the shift is inverse. Initial intervals: m- 0, n= —4; Jv = —4>.
Ex.11
Original
4iJ I I J *| J:=J?
-».' o j-
Iio etc.
Derivative
«E 5E
f=8=
etc.
Allocation of the Melodies of the Derivative Combination

According to Voice
§ 65. In the preceding examples the original combinations were worked out with two
specific voices and the derivatives with two others. But it can also happen that a single
voice-part can function in both combinations, for instance the soprano in the following
example:
Ex.12
Original
Il.Xi Palestrina
=S=
=53^
etc.
Derivative
wm &.—&■
r
p—Q g
or even both voice-parts can function in both original and derivative:
Ex. 13 Derivative
Original
rTn-^ff^rff^f^W wr
or finally, that the highest melody, though still remaining the highest, functions as a
lower voice:
Ex. 14
II
£ 3E ~rr
f= j±± .&-
Original Derivative etc.
*=-2
U-
5 rjij *i IE
m 33T
etc.
The shift in the last example must be regarded as direct, notwithstanding the fact
that at first glance it appears to be inverse. In general, by "voice" is to be understood
a melody to be sung, and this sense is meant when referring to the shifting of voices. To
avoid such an anomaly as Ex. 12 the original and derivative will be printed on separate
staves, with the highest voice above and the lowest below; the nature of the shift will
then be apparent. But the method by which both melodies are assigned to voice-parts
remains uniform.
Intervals Determining Jv; Entrance Intervals
§ 66. In determining the index, initial intervals were adopted for m and n. It is un
derstood that a progression of intervals will serve the same purpose. Sometimes an
initial interval is altogether impossible for determining Jv, as for instance in fugues
where the countersubject is delayed and a definite relation is established between it and
the subject. At the tonal answer the initial motive of neither the subject nor the counter-
subject can be exactly reproduced. In such cases m and n must be taken at the place
where all mutation in the voices has ceased.
§ 67. In fact, m and n may be taken as intervals between degrees not found simul
taneously in both voices but at some distance apart. For example, at the entries of
voices in stretto it is sometimes convenient to take for m and n the interval forming
their initial notes. Such an interval will be termed an entrance interval. An entrance
interval may be positive or negative; positive if it is above the first note in the upper
(first) voice, or below the first note in the lower (second) voice:
Lj1
^-^j r=^ &U ■JjJOsJ
f
ii etc.
ii" ■
II
ff etc.
It is negative if the situation is the reverse:
-5 I
2E
-4 H
^p etc. ^^
In vertical-shifting counterpoint, when comparing entrance intervals in the original
and derivative combinations it is necessary to consider their directions. The signs of
both intervals are the same when they are in the same direction; in different directions
their signs are opposite. Consequently if the shift is direct the direction of the deriv
ative interval is the same as that of the original; at the inverse shift it is opposite.
Original Original
ft
^—»
II ' etc. 1 II
Derivative Derivative IP"11
B y \S "
Here, according to original and derivative, the entrance intervals determine Jv by
equation (1), § 61. In illustration (a) above, Jv = -4 + 6 = 2; in (b), Jv = -1 -6 = -7.
Interchange of Original and Derivative combinations
§ 68. Ordinarily the combination that occurs first in a composition is regarded as

the original, but it is possible to treat a later combination as the original and an earlier
one as the derivative. Examining the nature of the alterations in the formula of the deriv
ative and the value of the index when the original and derivative are interchanged, two
cases are to be distinguished:
(1) Jv has the Direct Shift.—Taking the derivative for the original and vice versa it
follows that in the formula of the derivative the sign v of each voice will change to the
opposite. Therefore the sign also changes for the algebraic sum vv of both voices, i.e.
the sign of the value for Jv. For example, the derivative formula (Iv_5 + IIV~__2) Jv- 3,
when derivative and original are exchanged, reverts to the formula (Iv-~5 + IIV~ 2'
Jv - -3. This will readily be seen in the following illustration:
Original Derivative
Jv=3 Iv=5
^ln^Jz+f
5=± ffi ±==
II
¥ etc.
Original Derivative
I
1 ■ a
JiQ- A s
$ *-v. etc. ij»= 2 etc.
II
(2) Jv has the Inverse Shift.-U here the derivative is taken for the original and vice
versa, in the derivative formula v of the first voice is transferred without change to the
second voice, and v of the second voice to the first. Jv remains unchanged. The voice
that previously was first now appears as second, and vice versa. For example, the deriv
ative formula_(Iv-_1 +Hv-~6) Jv = ~7, when original and derivative are exchanged,
reverts to (Iv--6 + IIV__1) Jv = -7.
Derivative
Original
mB |)ij-iJjit
Original Derivative
(■„-., irr.l),„., ,p-l
I
^
^=F
a i j- j gg f ifeEEi
II '« *
etc.
I TW-
V=-6
The Augmented Fourth and Diminished Fifth
§ 69. One and the same Jv may yield different shifts of voices (§ 29). When these
shifts are made to various degrees of the diatonic scale the quality of certain melodic
intervals may be changed; such alterations as major to minor or minor to major involve
no irregularities.
But if instead of perfect intervals in a melodic progression augmented or diminished
intervals should result one of the principal rules of simple counterpoint would be vio
lated—such progressions are not allowed. The diatonic scale contains an augmented
fourth (the so-called "tritone") and a diminished fifth, both formed by the fourth and
seventh degrees of the Ionic or major scale. Melodic progressions of these intervals are
forbidden in the strict style, and their use is limited in other ways that are doubtless
familiar to the reader. The avoidance of these intervals in the derivative is possible, by
means now to be described.
(1) Having in mind a definite shift of voices, it is possible to calculate beforehand
which two degrees give an augmented fourth or a diminished fifth in the derivative. To
do this, the two degrees forming the objectionable interval in the original are transferred,
for the voice that shifts, at an interval equal to v of this voice, taken with the opposite
sign, i.e. these degrees are transferred in the opposite direction. Having obtained the
two degrees in this way, they must conform to all the limitations of the augmented fourth
(or diminished fifth). It is well to choose shifts that will result in one of the voices
either remaining in place or shifting an octave, in which case the precautions indicated
apply to only one voice. In Ex. 11, the derivative formula of which is(iv = _7+ Hv-3), the
upper voice shifts an octave. Therefore the avoidance of a possible augmented fourth
concerns only voice II, a contingency that may be provided for in advance. Taking the
two notes forming the augmented fourth in the given key:
fening them in the direction opposite to that of the shift Hv

4y^[j 0 ^
ii
and trans
, i.e. a fourth higher,
gives: ^The degrees obtained result in an augmented fourth in the deriv-
ative, and to them must be applied all the conditions that in the strict style surround
this interval.
(2) Another and more simple method consists in writing the original regardless as to
whether or not an augmented fourth (or diminished fifth) will appear in the derivative,
and when the problem is completed and the voices shifted, to make whatever correction
is necessary for its removal. These corrections consist in altering an augmented fourth
to a perfect fourth by raising the lower note or lowering the upper note by a chromatic
half-step. The choice depends on the melodic or harmonic conditions that precede or
follow.
3E 31 ^
(3) It is also possible to change the key signature of the derivative, getting a key
resulting in the removal of the augmented fourth, then to transpose so that both com
binations will be in tha same key.
In Ex. 15, a combination at Jv - -9, the shift I v = -9 is written on the third staff. To
get the derivative without putting it in the key specified involves these irregularities:
the augmented fourth in the melody between the first and second measures, and the
interval of the diminished fifth between the parts in the third measure.
Ex. 15
I
IE
i^^ Original
m m Derivative
A-
^b* at
(Dim. 4)
m» £
PPS IEC
Supposing the signature to be one flat the errors mentioned are corrected but a new one
appears: the leap of the diminished fifth in the lower voice between the third and fourth
measures. Assuming a signature of two flats corrects this, but again a new fault ap
pears—an augmented fourth formed by the extreme notes in the last three measures:
But a signature of three flats yields a correct combination:

II
i^3> -o-
£gp3 ..-2 p etc.
me ^
etc.
Here voice I enters with the first degree of the major scale and voice II with the third
degree. Transposing this combination so that it agrees with the key of the original gives:
Iv = -ll+nv=2.
IV
ipt rirJrf
etc.
1 etc.
It is sometimes impossible to get a key-signature that will remove all errors. In such
cases select the most convenient signature and eliminate the remaining errors by one of
the other methods.
CHAPTER IV
THE RULES OF SIMPLE COUNTERPOINT*
§ 70. In complex counterpoint both the original and derivative combinations must
conform to the rules of simple counterpoint, which in no case may be infringed. These
rules may be regarded as being of a prohibitory nature; they are not so much formulations
of what can be done in a certain situation as they are statements of what not to do.
Such, for instance, are the rules forbidding similar motion to a perfect consonance, the
resolutions of the second to the unison, the seventh to the octave, etc. An interval may
not be released from the limitations imposed on it by the rules of simple counterpoint,
but to these limitations new ones may be added. To investigate the conditions under
which a certain interval is used at a given Jv it is necessary to compare the limitations
of this interval with those of its derivative. If according to the rules of simple counter
point the derivative does not show limitations that exceed in strictness those of the
original, the conditions governing the latter remains unchanged. But if the derivative
interval, according to the rules of simple counterpoint, does show more limitations than
the original, this excess must be added to the conditions of the original interval. The
sum of these limitations constitutes the conditions under which an original interval may
be used at a given Jv. All the rules of vertical-shifting counterpoint are to be interpreted
in this sense.
Limitations Imposed on Intervals by the

Rules of Simple Counterpoint
§ 71. According to the scale of their limitations in simple counterpoint of the strict
style, intervals are classified as .imperfect consonances, perfect consonances, and
dissonances.
(1) Imperfect and Perfect Consonances
§ 72. The Imperfect consonances have the greatest freedom. They are also the most
satisfactory in effect, but because of this a certain monotony will result from their ex
cessive use in parallel motion. Hence has arisen the rule in two-voice counterpoint that
recommends the avoidance of a succession of parallel thirds or sixths that is too pro
longed. Except for this, imperfect consonances are under no restrictions.
§ 73. The next degree of limitations refers to perfect consonances; the rules gov
erning them fall into three groups: (1) limitations as to the moverments of the voices;
(2) limitations regarding the proximity of similar perfect consonances, and (3) limitations
based on the fact that perfect consonances, as compared to imperfect, sound harmon
ically empty.
§ 74. The first group of rules, limitations as to the movements of the voices, in
cludes the following:
•A knowledge of the rules of simple counterpoint is assumed. They are here summarized
for future reference.
53
(1) Similar motion to a perfect consonance is not allowed. Formulated in this way,
the rule includes both parallel and hidden progressions, but it is relaxed as the number
of voices is increased.
(2) Contrary motion from a perfect consonance to its compound, and vice versa,
is forbidden:
r> " f r ' r-

4 11 11 4 7
V
Obs. -Frequent exceptions to this rule are found in multi-voice counterpoint, espe
cially where fifths and twelfths are concerned:
Ex. 16 Palestrina
Ex. 17
Palestrina
*» "^ 3
/O „.
pHH 8 3X
AAA 5
e£
So o ax
11
-Vr
4 4 11
2
Very infrequently such progressions are found in two-voice counterpoint:

Ex. 18
. Palestrina
, P-t
3E
i 11
p' -f»-pr -jg-PJ-
3= « p p
£I
-#HC
££
THE RULES OF SIMPLE COUNTERPOINT 55
Contrary motion from a unison to an octave or vice versa is restricted to multi-voice
writion, where it is allowed between the two lowest voices:
Ex. 19
Palestrina
m iJOl
p=£ JOZ
IE
9^» ^ ff mm
" 7 0
Such progressions are usually found in seven- and eight-voice textures, especially
between the basses of the first and second choruses in a double chorus.
§ 75. The second group of rules consists of limitations regarding the proximity of
similar perfect consonances. In the second species fifths and octaves are forbidden on
the accented beats of adjacent measures:
bad
IE
iUt3C
bad
IE
U±3BZ
and in the fourth species on the unaccented beats:
bad
. h± h} .1
Obs. l.-As regards the first of these rules, its strictness, as well as that of many
other rules of two-voice counterpoint, is relaxed as the number of voices is increased. If
in two-voice counterpoint infringements of it are sometimes, though seldom, found-
Ex. 20
Orlando Lasso
1) * £=:p TZ-
w
m ' ilk I a
4
o ,.
4
£3t5
in multi-voice they are quite frequent:

Ex. 21 Palestrina
21 -e-
P zaz
tt IV
i Ê IXLI J ppJ ^^=jp jx:
B P etc.
Ex. 22 Ibid
3. g—n
33E
i Ê
^m
bm'- i
7 7 etc.
In multi-voice textures such progressions are often found, but in two-voice counterpoint
they are best avoided altogether.
Obs. 2.—In the fourth species, fifths on accented beats are not forbidden:
Ex. 23 Orlando Lasso
4)
TT
PP E
3E
K
Wm 35 3 ^ 4 etc.
BE
However, the composers of the era of the strict style did not consider it impractical, in
the fourth species, to write fifths also on the unaccented beats:
Ex. 24
Orlando Lasso
5)
11 XC m
mm gg , r PPa: #*n» air
5
rr* (9- (2^- «-#
m EZ3C
4 4
i 4 4 4
Ex. 25
Palestrina
P
Such progressions must be considered as very exceptional.
§ 76. The third and last group of conditions also refers to perfect consonances; they
consist of limitations based on the fact that these consonances, as compared with the
imperfect, sound empty. The result of this is the requirement that in two-voice counter
point perfect consonances should alternate with imperfect, whenever possible. The most
empty-sounding of all, the unison, is especially limited. With the exception of the first
interval and the last the unison on accented beats is to be avoided, and it should not
progress by similar motion to any interval following.*
(2) Dissonances; (a) Passing and Auxiliary; (b) Suspensions
§ 77. The most stringent limitations apply to dissonances. It is necessary to dwell

at some length on the distinctions presented by the conditions under which passing and
auxiliary dissonances and tied dissonances (suspensions) are used.
§ 78. These distinctions refer, first of all, to the location that the dissonances
occupy in the measure. Suspensions are placed on the accented or comparatively strong
beats; passing notes and auxiliary notes on the weak beats or parts of beats. For in-
stance, in £ time a passing note (but not an auxiliary note) may occur on the third quar
ter, which is unaccented:
#
sUM
Similarly, in 2 time passing notes of the values of quarters may occur on the second
or fourth beats (third or seventh quarters):
Ex. 26 Ex. 27
Josquin de Pres Ibid.
2)
11
Ex. 29
Palestrina Ibid.
4)
=8CTt .& etc.

Itfrrf^fr
P. 1,47.
?f^ gig g r o jCC
etc.
Passing notes should not be longer than halves, auxiliary notes longer than quarters.
Such cases as are illustrated in Exs. 30—31, where this rule is not observed, are very
exceptional.
*Cf. Bellermann, "Der Contrapunkt," 4-te Aufl., s. 159. [The reference reads: "Der Sprung
vom Einklang in ein anderen konsonierendes Intervall in gerader Bewegung ist im zweistimmigen
Satz nicht gut, und daher moglichst zu vermeiden." —Tr.]
Ex. 30 Loyset Compere Ex. 31 Josquin de Pres
IL «-
3E
etc.
IrNa ? » etc.
x
O " M: ggpgj
Bj a
etc. etc.
Ex. 32
Palestrina
§ 79. The second difference between passing and auxiliary dissonances and tied
dissonances is that the two former are not restricted as to direction; they are equally
correct when the movement of a voice is by step upward or downward. On the contrary,
prepared suspensions must resolve by moving the dissonant note one step downward-the
strict style does not allow resolutions from below upward. Therefore the movement of
suspended notes (dissonances) is restricted as to direction.
§ 80. The last and most important difference consists in the fact that the conditions
of the correct use of passing and auxiliary notes are implied by their designations. When
a dissonance satisfies the definition of either, the use of it is correct. The only ex
ception is that auxiliary notes should not be taken from the unison:
bad
Obs. 1.—Some theorists, Richter for instance, consider progressions incorrect in

which a passing note approaches too closely to a sustained note in another voice, and
forbids stepwise movement into the unison:
But such an interdiction is not in accordance with the practice of composers of the era
of the strict style. In their works such instances as shown in Ex. 33 are frequent (cf.
also Ex. 10, measure 3; Ex. 11, m. 2, two lower voices, and Ex. 28).
Ex. 33 Palestrina
uaz
fy* - r >p0J
fit 1 IIP >^

Obs. 2. -A passing note in the strict style is not to be regarded as a passing note to
a chord, but to whatever note with which it forms a dissonance. Other voices may be
free to take notes consonant with each of the other notes forming the dissonance:
2)
Pf -e-
^m
Progressions such as the following are characteristic of the strict style:
Ex. 34 Obrecht Ex. 35 Palestrina

3) 4)
^s XT"
him
T ft
^=^fi
etc.
asdi^ i -e- re
etc.
etc.
Ex. 36 Ibid. Ex. 37

5),
iN au etc.
j^JJ J f
W | , , • tt
Ibid.
. etc.
S 5=**
etc. etc.
Resolution of Tied Dissonances; The Dissonant Note
§ 81. A tied dissonance represents delayed movement. Not all the conditions of its
correct use enter into the definition of it. In the successive series of intervals (§ 11)
each dissonance has a consonance on both sides (§14, [3]) of which one is perfect,
the other imperfect (the fifth will always be regarded as a consonance, inasmuch as a
diminished fifth may be altered to a perfect fifth). If one of the voices forming a disso
nance is prepared and leads to the step below (§ 79) a resolution of the dissonance to an
adjacent consonance is effected. All such voice-leadings are included in the idea of
"resolution" but not all of these resolutions are valid. Considerations of euphony pre
clude the resolution of a dissonance to a note already present in another voice, such
as the resolutions of the second to the unison and the seventh to the octave. With these
exceptions the following list applies to both positive and negative intervals:
bad bad
3 - i -t\ II Jld ri II I O hM
O 9 ^ O
w
§ 82. In two-voice counterpoint the foregoing rule admits of no exceptions within the
limits of an octave. But the first dissonance beyond the octave, the ninth, is an ex
ception. The ninth may be a dissonance, and may resolve to a note appearing in the
good
other voice, i.e. to an octave: g=pg g When a dissonant ninth is referred to, this
©
8
particular case is to be understood. On the other hand, a ninth, as a compound interval
derived from a second, resolves the same as the latter, i.e. one degree downward, and
in this case a ninth is referred to as a second. The difference between a second and
a ninth consists in the fact that a second may be a dissonance only when below, while
a ninth may be a dissonance either above or below. All the other compound dissonances
resolve in the same way as the corresponding intervals within the octave limits. Hence
it follows that every two-voice combination may without restriction be expanded by an
octave (§ 47). But the reverse is not always true; to contract two voices by an octave,
even though they are not less than an octave apart, may eliminate a ninth if it is a dis
sonance above. At the contraction of the voice it reverts to a second, which is not
admissable as a higher dissonance.
The Fourth and Ninth in Multi-Voice Counterpoint

§ 83. There are two intervals of which the use in multi-voice counterpoint is to be
distinguished from their use in two-voice: the fourth and the ninth. The fourth, when it
occurs between an upper and an inner voice, or between two inner voices, may be re
leased from its limitations and regarded as a consonance,* and used according to the
rules for imperfect consonances, for example in the parallel movement of first inversions.
In this case it does not lose its consonant character, even when the other voices form
dissonant harmony, for example:
Ex. 38 Palestrina Ex- 39
Ibid.
^=
1)
etc.
2^
-o-
r I etc.
KJLLE-E 3 je
etc.
m E== etc.
•Of course, the strict style allows between these voices also the augmented fourth (and
diminished fifth) without preparation, provided each note forming the interval is consonant with
the bass, a condition that satisfies the requirements of the diminished triad in the first inversion.
In the following illustration the diminished fifth is worthy of notice, as with the other voices it
forms a dominant seventh:
Cristofero Morales
etc.
Ex. 40 Ibid. Ex. 41 Ibid.
10
F1^^ znz 33
£
3)
U rT F^ etc.
4^
^^ etc.
IE 33C
etc.
i^ etc.
Generally speaking, in multi-voice counterpoint the fourth retains its dissonant

character only in relation to the bass, in which case it is subject to the conditions of
preparation and resolution and is governed by the rules of two-voice counterpoint. But
even in relation to the bass the fourth may sometimes assume the functions of a con
sonant interval; more will be said about this in Chapter X.
While in multi-voice counterpoint the fourth may be released from the limitations of a
dissonance, thereby is imposed a new limitation on the ninth resolving to the octave; it
should not form a dissonance with an inner voice—only with the bass. If the number of
voices is not less than four it is exceptionally allowed as a dissonance to the tenor.
§ 84. The dissonant note in a tied dissonance, as stated before, resolves one degree
downward. Simple cases have been examined where the suspended note is followed
immediately by the note of resolution. More complicated cases are those in which a
consonant note taken by a downward leap (usually a fourth) is inserted between the
suspended note and the note of resolution:
Jq-JT-i I Hi^
cons.
PP
* rrm c 6 6
In two-voice counterpoint such resolutions sound harmonically full as applied to

intervals 1 and 6, and rather empty as applied to 5 and 8. In Fux's classic work Gradus
ad Parnassum (1725) this form of ornamental treatment is frequently illustrated. Com
posers of the era of the strict style often use it in multi-voice textures.
Ex. 42 Josquin de Pres Ex. 43 Ex. 44 H- Finck
e4
etc.
On the discretion of the teacher will depend to what extent these resolutions may
be applied to problems. In two-voice counterpoint intended to form part of multi-voice,
such cases may be left to broader considerations, and the application to 3 and 8 allowed.
Obs.—ln general it is forbidden to double a dissonant note in multi-voice counter
point, but an exception is sometimes found: a voice moving stepwise in notes of smaller
value may, at an unaccented part of the measure, run into a brief doubling of the dis
sonant note:
Palestrina Ibid.
ZEE
"A A J.. Si3^
*±
etc. mu w
_cr
etc.
3E ISZ etc.
f Ex. 50 |
etc.
Ibid.
ZEE
2
UH
etc.
i iAAlA
m etc.
The Free Note to a Tied Dissonance
§ 85. At the same time that a suspension (tied dissonance) is conforming to the
conditions of preparation (§ 86) and resolution, the other note may be free; it has these
characteristics:
(1) A free note may enter or progress, either stepwise or by leap:
Josquin de Pres Ibid.

Ex. 51 Ex. 52
a) -CL
-JJJ>
b)
1)
S W r
In Ex. 52 the voice taking the free note to the dissonance progresses simultaneously
with the resolution to another note, a complex form of treatment which because of its
importance is classified separately:
(2) The free note to a suspension may, just when the latter is resolved, progress to
another note that is consonant with the note of resolution. In two-voice counterpoint this
form of resolution is more applicable to seconds and sevenths than to fourths (especially
fourths resolving to fifths) or ninths. The following examples from Zarlino's Istitutioni
harmoniche show the application of it to 1, 6 and 3.
Ex. 53 Zarlino Ex. 54 . Ibid.

±
*^dif
VWFf?*
Ex. 55 Ibid.
-a—e
f-p _e £
3) etc.
fe *£
^ f^=2 etc.
(3) A voice may enter on the free note to a dissonance after a rest:
Ex. 56 _, . ,
Orlando Lasso
fîii TT
4)
Hf m
etc.
(4) Although the doubling of a dissonant note in multi-voice counterpoint is forbidden

(cf. Obs., § 84) the doubling of free notes is allowed if this does not involve some new
difficulty, such as the incorrect resolution of a dissonance. For example, it is bad to
double the free note to a ninth resolving to an octave:
a) bad b) bad
ir:
m r-Tj
^ aac ie:
since here as the result of such doubling the upper and middle voices show: at (a) the
bad progression of the second to the unison; at (b) the resolution, also bad, of the
seventh to the octave. In all other cases the free note may be doubled without restriction:
Ex. 57
Palestrina
t m 8=T^^Tn,f
5=*
fr6 • ' '
5) etc.
etc.
In Ex. 58 the free note is doubled at the entry of the tenor:
Ex. 58 Josquin de Pres
0WT* ZCC
r fete.
:> » p i IE
etc.
(5) There remains to be mentioned that in cadences where the resolution of a sus
pension immediately precedes the conclusion of the cadence, the voice having the free
note is often silent—it drops out at the actual moment of resolution. This is an idiom
characteristic of the strict style and occurs only at the intermediate cadences during
the course of a composition. At the final cadence, or at a cadence ending an important
part or division, all the voices must participate. This silencing of voices is illustrated
in Exs. 59—63.
Ex. 59 Orlando Lasso Ex. 60
Ibid.
£
7)
£ -mTl
w
etc.
8)
m nz -e-
m etc.
3z m m
etc.
fff^ 0M&
etc.
Ex. 61 Palestrina Ex. 62
Ibid.
i ^^
m =&
etc.
Ex. 63
Ibid.
±=± fe# 3CC oo
11)
# w? -o.
¥ etc.
^^ #
etc.
The Preparation of Suspensions
§ 86. The suspensions (1 and 6) that resolve to imperfect consonances are not sub
ject to any limitations as regards their preparation; they may be prepared by any con
sonance, provided that no rule governing the melodic conduct of the voices is broken.
But the dissonances that resolve to perfect consonances (3 to 4 and 8 to 7) are subject
the to following rule: A dissonance that resolves to a perfect consonance can not be
prepared by a consonance of the same value as that to which it resolves:
3e:
bad
-e-
^lÛ
3E
f 7
Obs. 1.—In multi-voice works progressions of fourths and fifths are often found:
Ex. 64
Palestrina
11 ffii 91
naNa IPS
06s. 2.-The complex forms of resolution mentioned in § 85, (2) allow the fifth as
preparation for the fourth, with the suspension in the lower voice: But
FTf
the ninth, with the suspension in the upper voice, can not be prepared by the octave-
impossible
, so the foregoing rule holds. Such preparation of the ninth by the octave
Vf?
is allowed in the free style, Ex. 65 illustrates this. It is from Recordare of Mozart's
Requiem, and in the ingenuity of its voice-leading is a model of noteworthy perfection.
Ex* 65 Andante.
7 8 etc.
§ 87. In the strict style the relative duration of tied notes must be considered. For
example, the note of preparation should not be shorter than the note to which it is tied.
Both tied notes may be of equal value; the second may be haff the value of the first but
not one-quarter of the value. The following values for tied notes are unconditionally
allowed:* , ,
*Cf. Note at end of chapter. —Tr.

Symbols Indicating the Conditions of the Uses of Tied Intervals

§ 88. A suspension and its resolution one degree downward, in whatever rhythmical
form is permissable in the strict style, will be indicated by a horizontal line placed
above or below the interval figure, and will be called the tie-sign (or sign of synco
pation). When placed above a figure the line will refer to voice I, when below to voice
II. For example, 3 means the suspension of a fourth in the upper voice and its resolution
one step downward:
JtJjii J'tvIJu .•,|',] n,l

B Io
f T
The expression _1 means the dissonance and downward resolution of a second. Applied
to negative intervals, —3 for instance, the line above refers to voice I, i.e. to the voice
which in fact is the lower.
Placed in parenthesis: (— ) the sign will mean that the note to which it refers can
not be used as a suspension. This enables the resolution of positive and negative
seconds to be fully expressed thus: j[_and-l^.In both cases the signs - and (-) indicate
that for seconds the resolution permitted is that to a third, but the resolution to a unison
is not allowed. Applied to the suspension of the seventh the signs are 6 .and —6:
bad bad
bad bad
it « 11
ik^r^r1 TT &
1
SF
Tl, I -T 1-6 l-«
Fourths and ninths as suspensions are indicated by 3, —3 and 8, —8, abbreviated
±3 and ±8. The complex forms of resolution do not require special signs.
~~* § 89~From the foregoing it will be seen that negative intervals have the same tie-
signs as positive intervals of equal value, but the position of the signs is reversed;
the upper sign of a positive interval and the lower sign of a negative interval are the
same, and vice versa.
§ 90. The following table, which should be thoroughly memorized", contains the signs
indicating all the conditions under which suspensions are used in simple counterpoint
of the strict style.
c— )
1 -1
(— >
+8
6 -6
C D C —3
*8
Beyond the ninth the compound intervals take the same tie-signs as the corresponding
simple intervals; e.g. 10 has the same signs as 3, 13 the same as 6, —13the same
t-z-r) _ _ t— 3 c—"3 —
as —6.
If the context refers to a single tie-sign,.either upper or lower, of a given jnterval,
the other sign should be omitted. Thus,_3 denotes a fourth resolving to a fifth; 8 a ninth
resolving to an octave, etc.
§ 91. It has been seen that the rules of complex counterpoint arise from the necessity
of subordinating the rules of simple counterpoint not only to the requirements of the
the original combination but also to those of the derivative (§ 70). Hence is shown the
need of transferring to the original those limitations of the derivative that the rules of
simple counterpoint would otherwise not impose upon it. They are—applied not only to
dissonances but also to consonances. For instance the expressionc53means that a sixth
is to be treated as if it were a real suspension; that is, is to be used only on an accented
beat, that the upper note is to be tied, and that it is to progress one degree
downward (§ 88).
■HSj.-haj
etc.
p
The expression —4 means that a negative fifth is to be used only on an accented
beat as a tied note of voice I (i.e. of the upper voice, in negative intervals found below):
II II
€»-
-4 -4
I I
It is understood that whatever limitations control intervals, they do not prevent the
use of passing notes and auxiliary notes on unaccented beats.
§ 92. In illustrating the preceding, those cases were intentionally chosen in which
the intervals concerned occupied the same positions as would a dissonance and its
resolution, and in which the progressions were to two consonances, the fifth and the
sixth. In the successive series ot intervals these are the only consonances that have
dissonances on one side only (§14, [2]). All the other consonances have dissonances on
both sides. Therefore if the progression is not to a fifth or a sixth any application of
the tie-sign to a consonance will result in a progression to a dissonance, at the point
corresponding to the resolution:
.n- ■
n
P U m
Tl —gf-
igyt W=r= ^^fe
In this case the note that sounds the dissonance must be regarded as a passing-note,
continuing in the same direction. Occurring as a quarter-note, it may be treated as an
auxiliary:
As to the fact that a dissonance formed at the sign X may in its turn function as a
suspension, more will be said in § 134.
§ 93. To indicate the dissonant note corresponding to the note of resolution—there
fore either a passing or an auxiliary-note-the supplementary sign X will be used, its
addition to a tie-sign meaning a tied note and its progression one degree downward. The
expression ^jneans that a sixth can not be taken as a tie on an accented beat on voice
I , but that it may be taken as a tie in voice II on condition that the note corresponding
to the note of resolution is either passing or auxiliary.
§ 94. Three signs are now available for indicating the conditions under which tied
intervals are used. According to the scale of limitations their order is: — , —x, (—). The
first of these, forbidding the free use of an interval on an accented beat, admits of its
use as a tie in the voice to which the sign refers. The second, —x, adds a further re
striction: the note corresponding to the note of resolution must be either passing or
auxiliary (the other voice sustained—oblique motion). The third sign, (—) forbids al
together a tie applied to the note to which it refers, nor can such note be otherwise
used as passing or auxiliary.
If a tie-sign applying to a given interval according to the rules of simple counterpoint
is replaced by either of the signs indicating more strict conditions (—x or (—)), no re
laxation of any rule of the strict style can ensue; the interval will only be placed under
more stringent conditions as regards its use. But the substitution of a sign by one of
lesser force in entirely inadmissable, for no interval may be released from the limitations
imposed upon it by the rules of simple counterpoint.
"Translator's Note to" § 87
A commentary on the last sentence of this section and the illustration following it is neces
sary, for a point is involved that may cause needless difficulty if it is not cleared up. The author
fails to make a distinction between the treatment of ties in %2 time and those in 7.
4
All three of the
illustrations at the end of § 87 are available in either \ or . provided that the second note of the
tie is not a dissonant; this is also true of the illustrations following in § 88. But if the second
of two tied notes is a quarter-note it can be a tied dissonant in \ but not in 2. tor one of the most
important rules governing the use of suspensions is that they shall resolve on a beat, not on a
half-beat unless this is the second half of an unaccented part of the measure. The author is aware
of this, for in the Observation preceding Exs. 88-90 he says that "it is very uncommon to find
suspensions on the second and fourth beats," and in these examples the resolutions all come on
a quarter-note, or second half of the second and fourth beats in %. Here the textbook rules and
the practice of sixteenth-century composers are in agreement. In many of the examples written
by the author he seems to have been in doubt as to what the time is: 2%, 4r, 4|, or 8?. All of which
are represented by the signs £ and (V in an utterly indiscriminate manner. The result is that
resolutions are often found on a half-beat if the time is <f , or on the full beat if it isC . The former
is wrong unless this half-beat is the second half of an unaccented part of the measure; the latter
right, according to the author's own statement just quoted, where, as in other places in the text,
under the necessity of expressing his meaning with precision, he uses the figures, not the signs.
The quotations from sixteenth-century composers have the resolutions on the beat, except in
Exs. 88-90, where Taneiev calls attention to the irregularity.
In discarding the signs and using the figures instead I have only done what should have
happened two hundred years ago. The sign for the "tempus perfectum,"0. was long ago replaced
by the more accurate figures .i and the modern equivalents .i and X\ while for some unknown
reason the signs for the "tempus imperfectum,"(T and £were retained and to this day are re
sponsible for much doubt as to how many beats in a measure are meant. Their continued use,
instead of the figures i< 4- or whatever is intended, is an unfortunate survival of the notation
based on the old "mood, time, prolation" principles and that lacked bar-lines. Modern notation
would gain in clarity and precision were the signs to be replaced by the figures (cf. H. Elliot
Button, "System in Musical Notation [Novello]," Chapter I, especially page 6).
CHAPTER V
THE RULES OF VERTICAL-SHIFTING COUNTERPOINT;

FIXED INTERVALS
§ 95. To prepare a table of rules for a given Jv a comparison must be made of the
series of positive intervals in the original combination with the intervals on its deriv
ative. The derivative intervals are obtained by adding to each original interval a value
corresponding to that of the index (§ 31). The conditions governing the use of an original
interval follow from a comparison of its limitations with those of its derivative at the
given Jv (§ 70).
§ 96. Such comparison yields four possibilities:
(1) From a consonance is obtained a consonance;
(2) From a dissonance is obtained a dissonance.
Such intervals will be termed fixed.
(3) From a consonance is obtained a dissonance;
(4) From a dissonance is obtained a consonance.
These intervals will be termed variable. Each of the four cases will be examined in turn,
beginning with the fixed consonances.
(1) Fixed Consonances

§ 97. Within the octave limits are four consonances (0, 2, 4, 5) and three dissonances
(1, 3, 6). Therefore not all the consonances of a given Jv can coincide in the derivative
combination with dissonances; every Jv has at least one fixed consonance. The fol
lowing indices have only one fixed consonance (cf. table, § 108): ]Jv= —10, —3<, 1, and
all J}v of the sixth column. On the contrary, the consonances are all fixed in Jv = —2 and
in Jv = —4 >. Between these extremes are found a number of fixed consonances from
other ] Jv.
§ 98. Two tables will be prepared (§ 95): one for a Vv (§ 57), for example, Jv= -11;
another for a '.Jv: Jv= —9. For this purpose is written out a series of original (positive)
intervals from 0 to an interval equal to Jv inclusive (§ 38). Under the original intervals
are placed their derivatives and under these a horizontal line. Below this line are written
the original fixed consonances.
[>■ imp. p- imp. p- imp. P-

Original intervals: 0 1 2 4 5 6 7 8 9 10 II
3
Derivative intervals imp imp. imp.
p- p- p- P-
at Jv = — n: —II — 10 —9 -8 -7 -6 -5 —4 —3 —2 —T 0
Fixed consonances: 0 2 4 7 9 II
■ Jv = — 9 =-
p- imp p- imp. p- imp.
Original intervals: 0 i 2 4 5 6 7 8 9
3
Derivative intervals imp. p- imp. p- imp. p-
at '2 Jv = — 9: —9 — 8 —7 -6 5 —4 —3 —2 —1 0
Fixed consonances: 0 2 4 5 7 9
70
THE RULES OF VERTICAL-SHIFTING COUNTERPOINT 71
§ 99. Comparing these two tables it is seen that at an index of the first group ( Jv)
each consonance that is fixed yields a similar consonance in the derivative, i.e. perfect
yields perfect and imperfect yields imperfect. On the contrary, at an index of the second
group ( Jv) from each fixed consonance is obtained another with the opposite quality:
perfect yields imperfect and vice versa.
§ 100. Such a relation holds good between the fixed consonances of all indices of the
first and second groups with the exception of Jv= —8. This can be verified by the General
Table of Indices but the demonstration follows. In the successive series of intervals
(§ 11) a int. represents the distance between any two consonances of the same group,
i.e. between two perfect or two imperfect consonances (§15, [1]), and a int. the distance
between any two consonances of different groups (§15, [2]). There is only one exception
to this, the case of 4 and —4< (§ 16), in which two consonances of the same group are
separated by 8, a int. Since Jv is that interval in the successive series by which a
derivative is separated from its original (§ 33), then at an index Jv, i.e. Jv= 'int. , each
fixed consonance and its derivative can only be consonances from one and the same
group; at an index Jv, i.e. Jv= int. (except Jv = —8) they can only be consonances
from different groups.
§ 101. The attribute of fixed consonance? to which attention has been called allows
the following broad generalization to be made regarding the entire series of rules for ver-
tical-shifting counterpoint: The imperfect consonances yielded by JJv in the derivative
must conform to the limitations of perfect consonances (§§ 73-6).
§ 102. It is important, first of all, that an interval to be shifted at any index of the
second group ( Jv) shall not contain an instance of similar motion to an imperfect
consonance, because in the derivative this will show up as similar motion to a perfect
consonance. And since the latter progression is forbidden in the original by the rules
of the strict style, it follows that similar motion at ^JJv is altogether out of the question.
However, the strictness of this condition as it applies to hidden progressions is relaxed
as the number of voices is increased.
§ 103. Therefore, indices of the first group (JJv), as presenting the first cases oi
coinciding consonances (§ 100) are not subject to restrictions as to motion. But those of
the second group (JJv), presenting the second, preclude the use of similar motion.
Obs.—It has been mentioned that ^Jv = —8 is an exception to the indices of the second
group. While the fixed consonances of the other ^JJv give intervals of opposite quality
in the derivative, the only fixed interval of ~ft,_ _8, a fifth, gives also a fifth in the
derivative (4 - 8 = -4, § 31).
p-
0 1 2 3 4 5 6 7 8
—8 7 —6 -5 —4 —4 —2 —1 0
But this fact does not admit of Jv = —8 being included in the indices that allow
similar motion. Since the fifth is a perfect consonance, and the only fixed consonance
that this Jv has, similar motion is rendered impossible. Therefore the rule stated—that
"//i> do not admit of similar motion—holds good also for ^Jv- —8, so the rule has no
exception. But it should be remembered that this rule is rigidly applied only to two—voice
counterpoint. In multi-voice counterpoint its strictness is considerably relaxed, as a
result of the use of hidden progressions.
§ 104. The first of the limitations imposed by the rules of the strict style on perfect
consonances, forbidding similar motion to them, has been applied to the imperfect
consonances of the second group. The next step is to apply to the perfect consonances
of this group the remaining limitations of perfect consonances (§§ 74—6).
Applying limitation (2), § 74, gives this rule: Contrary motion from a simple fixed
consonance (either perfect or imperfect) to its compound, and vice versa, is forbidden.
The limitations in § 75 regarding the proximity of similar perfect consonances must be
extended to include all fixed consonances of the indices of the second group. Finally,
from the limitations in § 76 arises the requirement that in the original combination
perfect and imperfect consonances should if possible be used more or less-in alternation.
A preponderance of imperfect consonances in the original results in the derivative
sounding empty, but inasmuch as supplementary voices are always available this re
quirement may be discounted somewhat.
§ 105. The only remaining limitations refer in particular to the unison (§ 76). Since
at a positive Jv derivative unisons from positive intervals are not possible, these
limitations can apply only to negative J Jv, and refer to the intervals corresponding to
the index. From this interval, as has been previously stated, is obtained a derivative
unison (§32, [2]). Therefore, with the exception of the first and last sounds, an interval
equal to a negative Jv can not be placed on an accented beat, though it may be used on
unaccented beats, and when so used the voices must not progress to the next interval
by similar motion. This rule applies also to all negative indices, both of the first
and second groups.
Obs.—If the rule mentioned in Obs. 1 to § 80 is observed it will be necessary to
forbid stepwise movement to an interval equal in value to a negative index.
§ 106. The student who has assimilated these rules can now undertake exercises
with indices of both the first and second groups. On accented beats only fixed con
sonances should be used, and dissonances and variable consonances employed only
as passing, or less often, auxiliary-notes.
§ 107. The table in § 108 consists of a list of the fixed consonances of all indices.
In working out exercises those indices should be chosen that have enough fixed con
sonances to make study profitable—not less than three. With negative JJv the limiting
interval of each index, always equal to it, must be taken into consideration. For negative
JJv with the direct shift the limiting interval shows the distance to which the voices may
approach (§ 38), at JJv with the inverse shift, the distance to which, they may recede
(§ 39). As to the order in which the exercises should be written, instructions are given
in Chapter VIII, but before starting this work the present chapter must be
thoroughly mastered.
§ 108. Table of Fixed Consonances

2 4 9 11
2 4 5
. 11
. 11 5
1st pair
0 . 5 7 . 12 0 2 4 5 7 9
0 . . 5 2 4.79
5 7 9 12
0 2 .579 12 0 2 4.79
2nd pair
0 2 4 7 9 11 -11 > -10> . . . 5 .

(0 2 4
i 4 . 9 11 -42? -3< . 5 7
. 2 4 . 9 11 3 4 0 . . 5 7
3rd pair
-7 0 2.57
Explanation of Table
To the right of each right-hand column of each pair and to the left of each left-hand
column is placed the series of fixed consonances of its Jv. A dot occupying the place
of a figure in the row of consonances means that this consonance of the corresponding
Jv is variable. If a simple consonance is fixed its compound is also fixed. It would
therefore be possible to restrict the list of fixed consonances to the octave limits, but
for greater convenience in using the table the fixed consonances beyond the octave are
also given. Consonances for JJv with the inverse shift are given from 0 to the interval
equal in value to the index. For negative JJv with the direct shifts the consonances
start from an interval equal to Jv and continue to the same limit as the upper row of
intervals for the same column. Two series of figures are given for the middle Jv of each
left-hand column, the upper series for the inverse shift, the lower for the direct shift.
§ 109. In the following examples showing the application of fixed consonances to
JJv of both groups, those that have few fixed consonances are omitted. Thus, from
V/t> is omitted Jv= -10; from 2JJv of the first pair of columns only Jv= -13 is retained.
In studying the examples the table in § 108 should be consulted. For limiting in
tervals v. §5 38-9.
(A) Examples at xJJv
§ 110. Omitting Jv= -10 (§ 109) the remaining 1JJv are taken in this order:
-11>
-4< -3< -7>
3 4
(1) fv= -11, double counterpoint at the twelfth (cf. table in § 98).
Ex. 66
a)JV=-ll j nv = -4 Orlando Lasso
g SEE PPP
w£.=mi
~rr- -»- TT
TT_ •O-
~n~
etc.
Original Derivative
X
s 7—rr
II
n: IE
nz
TV = —7 etc.
Ex. 67
Ibid.
b)Jv=-ll J II v=-4
TT
1t
X
Original Derivative
#=P
^^ zce
II ^
I=3e ~n~
ZCE
* ZCE
etc.
• p j U j. J J -N-p^Si
^ etc.
In each of these examples the first three measures' present the original combination,
the next three measures the derivative. In Ex.67 the voices pass into simple counterpoint
at the seventh measure. Cf. also Ex. 7 (Jv = —11).
(2) l/v= —4>, the middle Jv of the same column.
Ex. 68
» 1Jv=-/t>Original
x: Palestrina
II Derivative .M
i 1 $ W &3J • ■ a * aa
etc.
B je i J IE ^^ etc.
II
ÊfrW
v=-4
ft" If
etc.
Cf. also Ex. 11 (/i> = -4>).

If, in a combination written at the middle index, a voice of the derivative is sep
arated an octave, another derivative is obtained at an outer Jv of the same column (§ 56).
Thus, Ex. 68 at Jv= -4> may be shifted at Jv = -11.
The distance between the two clefs for I shows the Jv.
Ex. 69
3) l/t> = -4<
(The distance between the two clefs for I shows the Jv).
Derivative J
i£ #
a *=B
BsP
ii IE
P
^:
IP BCg
f
II
According to § 56 it is possible in this example to separate a voice of the derivative

an octave, thus obtaining a derivative at the lower Jv of the same column, i.e. at Jv= 3.
Because of the small number of fixed consonances that Jv=—4< has, while the one
under it Jv = 3, has only two, also under the necessity of relinquishing animation and
interest in voice-leading until suspensions can be used, combinations result that are
musically rather meager.
Ex. 70
4) 1/v=3 Original
There is no need to cite separate examples for the middle and lower indices of the
fourth column. If in Exs. 69 and 70, at Jv = —4< and Jv= 3, original and derivative are
exchanged (§ 68, [1]), combinations result at /t>=4 and /v=-3, the lower and middle
indices of the fourth column. In this way exercises in all essentials can be changed
from the indices of any column to those of the other column of the same pair.
(5) Jv= —7, double counterpoint at the octave. There remains to give an example at
this index from the zero column. In the free style double counterpoint at this interval is
the most frequently used, but in the preceding era /u= —11 (double counterpoint at the
twelfth) was considered of greater importance.
Ex. 71
5. 'Jv=-7 IP''7
m S*Ê i
Derivative
I, Jfef^b ^m-
Original
m ¥ÎT 'A -
m
j=i -©-
#=:q?=F
£
■i g
jfe j r r r P iP * » ■ )O n —&-
mm
\w SE g pq» XE
(B) Examples at ^JJv

§ 111. Examples at *//« (§ 108) are taken in this order:
(1) 2/i>= -9>. At this index all the consonances are fixed (cf. table, § 98). This is
the so-called double counterpoint at the tenth, and together with Jv=—7 and /i>=-ll
forms a group of double counterpoints that in most textbooks exhausts the study of
Ex. 72
L]v = -9 Palestrina
m Original
3SH
m
^l
II
m
Derivative
Si
Ex. 73 ,v=-9
*
eo? XE
-©-*-
Ibid.
m jCE
2/t>=-9
_o_ _o_
SEE**- XC
fe= W^ jBI
Derivative
v = -9
*
For further remarks on both examples vide § 119; also cf. Ex. 2. Ex. 74, in double
counterpoint at the tenth, requires a shift at Jv = -16 (§ 50). In this example the voices
are separated from each other by an interval greater than a twelfth. The limiting in
terval is 16> (§ 39).
Ex. 74
Zarlino
2Jv = -\6
fia 2 SCO XXI o gqg
i 331
?
t.-i-rrirriTii'rfrn'irn,i'f'irJ^ji^J
ipp r^ri''jjj|rfrp m m
n n
a
iRrrrri" = i"JjitJ^Jt tJrirr° wi'r
o-
w
-©-
add r^ffnn «=
I ~f^ J j J ^J i [Tr rf ,i iJ.ipJJjijf p »--#-# " a o ? ~n~

Derivative
/5 nv=7
I B O -n w&
rrrn'irrrrrirrJJjji^
3Iv=-16 X
^ s gg » F^
r =-9 ¥
rg i O . . it
- J J r if J jjjpi^ ^Jfipr" i-rr f

»-» «~ *#
** j r r =e rrJr riffM ir rg
?
.J* »'s - 1 n I■• . -m m ri ri j
4i» — P [f p n—|—1||| —f—i—h—1—.^Pif9 i | | iu

.r . r. r. R ; J r.. . & K rJ » ^ 6 M W
2)2JVr-g .
Palestrina *
Ex. 75 'Derivative
(a)
•Taken from a four-voice combination.

Separating the derivative an octave (§ 56) gives a combination at the lower index
of the same column (Jv= 5).
*Jv=B
-1v=5
Derivative
£? etc.
P
£ £z
1 30C
m ^^
IE
HP
• wrto
i\ . 3 P "6 1
irppr^ 'r^ p l9 'if pro 1 ~t=\—
\Eh r I
,-io . -w 1 =W=
Vlirt p-f-' O ■ ►-■ a
In * "
*t
s
=5S*
es=? r g
Exchanging original and derivative in the last two examples (Jv = -2 and /u= 5) gives
combinations at Jv = 2 and Jv = -5<, indices from the other column of the same pair:
Ex. 77
(5) There remains a single index from the first pair of columns: z/v= -13>.
Ex. 78
m ^
%
f#=#
f ^
m i nap
ICE rr r i rrrr J^r id
is—r-<a
9s- i PE
u:
1
^ gr •
P ? »
Sp J'rrr i r _j- r^r 3E

^
*
§ 112. In concluding the discussion of fixed consonances, some special properties

of two-voice vertical-shifting counterpoint must be mentioned; cases in which, sus
pensions being absent, there appear on accented beats only fixed consonances.
These include:
(1) Combinations that may be shifted at two or more indices and—
(2) Combinations in which imperfect consonances may be duplicated.
Chapter X will be devoted to a detailed study of combinations in which all intervals,
without exception, participate.
(1) Combinations Without Suspensions, Admitting of Shifts at

Two or More Indices
§ 113. At ^/v= 9 all the consonances are fixed (§§ 98, 100). If a combination is
written without suspensions, with consonances on the accented beats, that meets all
the requirements of JJv (non-use of similar motion etc.), and that does not exceed the
limiting interval 9>, then whatever the consonances used and whatever their number, the
combination will satisfy all the conditions of Jv=—9. Hence it follows that every com
bination from which suspensions are absent, and written at any index of the second
group, can be shifted at Jv = — 9. These shifts at two indices are called compound
(double) indices (§ 28). For instance, in Exs. 75 and 76 the indices could be designated
^/v = — 2, —9 and ^Jv= 5, —9, respectively. The only obstacle to shifting at Jv = — 9 would
be that the separation of the voices might exceed the limiting interval —9. But this is
easily overcome by separating a voice of the derivative an octave, taking Jv = —16
instead of /f= — 9. If in practice such shifts prove inconvenient because of the wide
distance between the voices, they are still important from a theoretical standpoint, as
confirming the fact that every combination (without suspensions) at an index of the
second group will also yield a shift in double counterpoint at the tenth. Exs. 77 and 78
illustrate such cases of double indices: *-Jv= — 12, —16 and *Jb = —13, —16.
§ 114. The absence of suspensions has been stated as a characteristic of all JJv,
which includes Jv = —9, but this is provisional and must not be understood as preventing
the use of every combination that contains a suspension. Certain shifts are possible
with suspensions, but not all; some dissonances may render them impracticable, a
matter that will be discussed in detail in Chapter IX.
§ 115. The attribute of JJv under discussion is the result of a more general principle
which may be formulated thus: In a combination written at conditions that are more
rigid than those of the given index the possibility of yielding a shift at this index is not
relinquished. In other words, the totality of rules for a given index must be understood
as the minima limitations by which a combination is surrounded, in order that a shift
may be made at this index. Whatever limitations are imposed in excess of the required
index do not prevent a shift at this index. On the contrary, a combination that is released
from whatever limitations the given index may have ceases to yield a valid shift.
§ 116. Guided by these considerations, it is not difficult to find a few cases where
the given combination yields shifts at compound indices. For example, the fixed con
sonances of Jv = —7 and /t>= 2 are identical (cf. table, § 108). Since a combination
written at Jv= 2 is subject to conditions more strict than those of Jv = 7 (because Jv= 2
belongs to the second group of indices, restricted as to movement), then any combination
at Jv-2 (without suspensions) can be freely shifted also at Jv = —7, if only the voices
are not separated by more than the limiting interval, an octave. Even if they do exceed
this limit a combination at Jv = 2 may still yield a valid shift in double counterpoint
at the octave, not at /t>= —7 but at Jv= —14. At the same time Jv= —2 gives a shift at
Jv= —9 (§ 113). Hence a combination at Jv- 2 may be designated by a triple index:
Jv= 2, —7, —9 or /t>= 2, —14, —16 etc. It is clear that the opposite cannot be done, such
as a combination written at Jv = —7 shifting at /t»= —2, since the admission of similar
motion in the first of these gives an impossible shift at Jv= 2.
§ 117. Two rules ensue from the foregoing:

(1) A combination without suspensions, and containing similar motion, admits of
shifts at those JJv of which the fixed consonances include whatever consonances
appear in the given combination. For example this combination-
Original 1
containing similar motion, uses two consonances: 2 and 4. The V/v that include 2 and 4
as fixed consonances are those from the third column: Jv = 3, Jv=4> and Jv - —11 (cf.
table, § 108). Therefore the combinations can be shifted at each of these *//p:
Derivatives:
11
(2) If to a similar combination (without suspensions) is applied the limitations of

^JJv it will admit of shifts at every index, both of this and of the first group, of which
the fixed consonances include those in the given combination. For example this—
Original x
includes the same two consonances as the preceding, but differs from it by the absence
of similar motion. Therefore it can not only be shifted, like the preceding combination,
at the V Jv of the third column, but also at all the other ^JJv that include, as fixed con
sonances, the intervals 2 and 4. This concerns all JJv from the fifth column and two
from the first column: —13 and —6>. In this way the combination referred to yields shifts
at a compound index that unites the conditions of eight JJv, namely: JJv = —2< , 3, —4<,
5, -6>, -9>, -11>, -13>.
Derivatives:
§ 118. If a combination is written at a middle index and a voice of the derivative is

separated an octave there results a derivative at an extreme index of the same column
from which the middle shift was made (§ 56). This was shown in the preceding illus
tration, , where Jv= —11 was derived from Jv= —4< ; Jv = —5 from Jv = —2; Jv= —13 from
Jv= -6>. Discarding the shifts that are only reproductions of others at the octave, there
remain five different combinations. But this number of shifts is obtained at the expense
of great restraint. By relinquishing similar motion the original consonances are cut
to two at the most. In general, the number of derivative combinations that it is possible
to obtain from an original is in inverse proportion to the musical resources employed in
its composition. The more numerous the derivatives, the more monotonous the original
from which they are obtained.
(2) Combinations Without Suspensions, Admitting the Duplication of

Imperfect Consonances; Three-Voice Counterpoint
§ 119. Mention is still to be made as to a characteristic possessed by combinations
(without suspensions) that belong to JJv corresponding to imperfect consonances, i.e. to
the indices of the second pair of columns {^J]v= ±2, ±5, —9, —12).
At each ~/i> from the second pair of columns a derivative may be obtained that will
be the execution of the original plus an additional shift of one of the voices. The result
will be three-voice counterpoint in which two voices represent the duplication of one of
the original voices in parallel imperfect consonances. Since the additional shift is
possible in either voice, every such three-voice combination may appear in two forms.
For example, }v = —9 gives two three-voice combinations: I + I v=-9 + II and I + II +
IIV . For brevity the added voice, doubling another, will be indicated by the letter d.
Thus, Id= ~9 will replace I + Id= ~9; IId~ ~9 instead of II + IIV= -9, etc.
Such three-voice counterpoint is most often the result of using Jv = —9, an index that
yields parallel tenths in the outer voices. In parallel motion these consonances, rather
than others, are the most euphonious. As has been often observed, all the consonances
in /i>=— 9 are fixed. Whatever consonances are taken, at the duplication I " + II or
I + II , all three voices form a consonant union, as will be seen in this table:
Id = -9+II
7 I + IId = -9
4 5
,S „v=-9
Si
Ivv-
— y9 II0 2 4 5
In exactly the same way, at d- Jv the fixed consonances of all the other JJv cor
responding to imperfect consonances give a consonant union in three voices. Exs. 72
and 73, at Jv= —9, give two three-voice combinations:
!*•> t b) IIV=-S
M
I J J , _
"' v - i A •~ ^ ^J_
The same process may be applied to Exs. 75 and 76:
4).;I*JT = I
B I
nu etc.!
j 1
Lo. 1^
II
§ 120. If Jv refers to an index that has both the direct and the inverse shift, use will
be made of the signs already known for indicating the necessary situations. For example,
the expression I + II _5< means that the limit for the voices of the original combi
nation is a sixth, consequently the direct shift; I + II **> means that a sixth is the
separation limit, therefore the inverse shift. If in the last illustration the original and
derivative are exchanged, the duplication would be indicated: (a) I + II •*< , (b)
,d= -5< + 11.
§ 121. Furthermore, every three-voice combination without suspensions that is formed
by the duplication of one of the voices is possible also, in three voices, by duplicating
the other voice. In analysing the examples it may be useful to play them on the piano,
duplicating first one voice and then the other.
§ 122. In accordance with the statements in § 113, any combination, at whatever 2/t>
(without suspensions and having fixed consonances on the accented beats), gives a shift
at Jv = —9 (or Jv = —16). Also, a combination at Jv = —9, as an index of the second group,
can be executed in three voices with the shifts I^ — y or IIV y (§ 119). It follows
that any combination at a ^Jv from which suspensions are absent may be executed also
at the duplication I y + II and I + IId . And it is possible that a two-voice com-
ination originally written at an index of the second group—corresponding to a dissonant
interval, ^Jv= —13 ior example— may be executed in three voices with either of these
duplications. Two aspects of the original and derivative are here shown (cf. Ex.78).
Derivative
Jv=-13
Jvr-i3 etc. f etc.

Original (
Derivative r
Playing the foregoing is inconvenient, due to the fact that the duplications in the
originals are at Jv = -16 instead of at Jv= -9.
Obs. -Duplications have been dealt with only as far as they relate to the study of
fixed consonances. Many more questions will arise as to duplications applied to tied
intervals, dissonances and variable consonances; they will be discussed in Chapter X.
(2) Fixed Dissonances
§ 123. A dissonance in the original combination coinciding with a dissonance in the

derivative is called fixed (§ 96). Their number varies with the different indices. For
instance, all the dissonances in Jv= —9 and Jv= —4> are fixed; all except one are fixed
in //t>=—11, -7, -5< , +2 and 5. The indices of the sixth column and also these: -10,
—5> , — 6<, 1, have no fixed dissonances at all.
§ 124. To express the conditions under which a fixed dissonance may be used at a
given index it is necessary to compare the original dissonance (m) with its derivative
(n), the sign of one or the other indicating its use in simple counterpoint (§ 90). Next,
the upper and lower signs must be compared with each other. If they are alike, or if the
sign of the original dissonance indicates a higher degree in the scale of limitations
than that of the derivative, then the sign of the original stands without change. If, with
different signs, the highest degree in the scale of limitations is shown by the sign of the
derivative dissonance, then the sign of the latter must be transferred to the original.
In other words, the sign (— ) at an original interval is always retained, but the sign — must
be replaced by (— ) in cases where the latter applies to the derivative interval. In this way
the original takes over the surplus limitation which the derivative has as compared to it
(§ 70), and thereby assumes the conditions necessary for its use in simple counterpoint.
§ 125. For example, at Jv= —11 the sign (—) must be substituted for the lower sign
of the fixed dissonance 10, and must therefore appear as the lower sign of the derivative
interval:
/v=-ll
m- 10
according to the rules
of simple counterpoint
ra= -1
Conditions of using original

at A>--11:
H
Hence, in double counterpoint at the twelfth the lowest note of the interval con
sidered cannot be used as a suspension, since in the derivative it would give the in
correct resolution of a second to a unison:
. ,(k°)
§ 126. In comparing the dissonance of the original and derivative it may happen that
the original, already having the sign (-) for one voice, acquires the same sign also for
the other voice. For instance at Jv = 5 a second, giving a seventh in the derivative
(1 +5 = 6), takes the signs 1 , the result of comparing the two intervals 1 and j>.
An interval having the sign (—) both above and below can not be used as a suspension,
but only as a passing or an auxiliary-note.
§ 127. With the help of the table of dissonances (§ 90) it is easy to decide questions
that relate to fixed dissonances; for example, at what indices is it impossible for a
given dissonance to function as a suspension?
Suppose that this question arises concerning the second. A second ( _1_) has the
sign (— ) above. Its use in the capacity of a suspension would therefore be impossible
in those cases where a dissonance in the derivative would result that has the same sign
underneath. Taking these intervals from the table, and indicating by m an interval of the
original combination, by n an interval from the derivative:
m=t TO- 1
simple simple
ctrp. ctrp.
Conditions C->
Conditions "=£
for original:
'X, for original:
It remains now to find the index according to original and derivative intervals. Using
formula (2), § 61, the original interval is added to the derivative, taking the opposite
sign, .the result is the required index: Jv = —2 for case (a) above, and /i>= 5 for case (b).
It is easy to find, in the same way, that, for example, a seventh can not be used as a
suspension at Jv = —5 or at Jv- —12, etc.
§ 128. A table of intervals will be compiled for ^/t> = —9. The dissonances in the
two upper rows are provided with their appropriate signs, according to the rules of simple
counterpoint. Underneath the horizontal lines are the fixed dissonances (at this index
all dissonances are fixed) and their signs indicated as explained in § 124.
2/v=-9>
Original intervals: 0 V A 8 9
Derivative intervals: _9 -8_ -7 -6_ -5 -4 -J_ -2 -J_ 0
Fixed dissonances: 1 r 6
f— >
8
(—)
Of the intervals in the bottom row 1 and (> retain their former signs, therefore the
conditions of their use at Jv=—9 is the same as in simple counterpoint. But 3 and 8
acquire new limitations, expressed by the sign (— ).
In any statement for the rules of a given index there is no need of enumerating those
dissonances of which the signs remain the same as in simple counterpoint. Therefore
the rule for suspensions at Jv = — 9 is expressed by the figures 3 and 8., Suspensions
are used in Ex. 79. Before working out exercises one should get acquainted with the
rules for the preparation of fixed dissonances (§§ 131—2).
Ex. 79 b
Original I
-m—e-
u I i j j-H^j
I
as 31
1
2==P
3 ^W ^^
*
Jy — _q
Derivative I i
-3
in
FS
B| i r^ Fr '^'r r F ^^
^
1
§ 129. The dissonances are also fixed in 1/v= -4>:
o 2 2 a4
_4 _T -2 J 0
The sign for the second is the same as in simple counterpoint. The fourth acquires
a new sign for the lower voice: TJT. Ex. 80 uses fixed dissonances at Jv= -4>.
Ex. 80
Original
oo
Jv=-4>
4
Derivative r
The remaining examples in this chapter are given without interval tables.
§ 130.(1) l/v=-ll>. Here the only fixed dissonance using conditions other than
those applying in simple counterpoint is 10. Dissonance 6 is still excluded
as it is variable.
Ex. 81
l)%Jv=-ll
Original IP ^
■i_ 1
^rfr-3.1. ..i. jj^vm.jj /!>
£^
ffi ?P* ^= «? f^PF—Hf PF
Derivative
^-io. -10 1-8
—flrzrriX!
fe
.
'Û M s
©
«=
Ex. 82
Zarlino
b) Jv=-n
at* ■ an wf— —-.^±am.\ ti
WVP^T* i -o- •o
p
Original
&&
jfigg ~» fs^-p grcc g 'j p r f»
II
(V
— ga P »
-»- Ê
pS3 Ep s=s:
-"—0- xsz
a
SK^rprTT 6 " d
P i 9—0
10
m^& e¥
JPsf ^=F ~TT- «- rv a 3E p B -o-
ff^ g a, p J&
-e»- rn 0-0 gg| ^ PS
V^-7
wg| ii"• —p s r; a -©-

do
£ sa fJ
.©- 3CE
Derivative
X
git 15= iFP -o- _"j*Jo_ a i j»p i*31?2 W*yarza:
^
r=-4 -8 -8 -8
in;
ii*A Prr- rf i-e p
f s J9- Ppp ,~q r 0 ll—
1S-7S-2
0^—■ -
*
-10
_p—p p—h- (—
J5 g f- --H -—7^-F1
'. r< 1*5
. tl3 r
fiL. I 0 ' M*\
-8
(2) Jv=— 7>. The use of the fixed dissonances (1, 6) is the same as in simple
counterpoint. The variable dissonance 3 is for the present excluded from the avail
able suspensions.
£x. 83 £x g4
Original , ' Original Palestrina
1^
.'JtJ J i j J _E£p -ee- m lo. JiJbJ U^ri6
no f
Derivative
/•—7
II
Derivative
/«—7 X , -1
IT etc.
jjjlpl 43-
-6
m ^
az
jv=—7
HSB -6
Ex. 85
Original 6 Berardi
II
Derivative
-1 , -I
yy.-toJjJjUp^ 3E
Ex. 86 shows a shift in double counterpoint at the octave, but at Jv = -14> instead
of at /t>= —7>, a case, therefore, of where the voices in the original are separated from
each other by more than an octave (§ 48). The limiting interval is 14>.
Ex. 86
Original Palestrina
BS y-j'J.i-H^
I
Derivative
m& ur11^! rlf-r jr

©
II
II A nRft
>m
jE :z etc.
,v = -l4 etc.
(3) /«= 2. The variable dissonance 3 is temporarily excluded. The conditions using
the fixed dissonances are the same as in simple counterpoint.
Ex. 87
Original J $ J.
JY=8
Exchanging original and derivative results in a shift at Jv= —2. At this index the
variable dissonance 6 is temporarily excluded. The signs for the fixed dissonances are:
J; 8 10.
The Preparation of Fixed Dissonances
§ 131. According to the rules of simple counterpoint in the strict style a suspension
that resolves to a perfect consonance cannot be prepared by an interval of the same
value as that to which it resolves (§ 86). The suspensions resolving to perfect con
sonances are 3 and 8. Therefore the following preparations for these intervals are
forbidden:
1
iOJ II IIoj
=*
SHE it:
RPf
The rule forbidding such preparations in simple counterpoint may be formulated
thus: In a suspension resolving to a perfect consonance the note against which the
suspension is prepared can not progress diatonically downward to the note against
which the suspension is sounded.
§ 132. At indices of the first group <}]]v) the rules for the preparation of fixed
consonances undergo no change from the rules of simple counterpoint, except the single
requirementthat the preparing consonance must be fixed. But for those of the second
group (z//t>) the restrictions of simple counterpoint as regards the use of perfect con
sonances must be applied also to the imperfect consonances. Therefore for ^JJv the
rule for the preparation of fixed dissonances may be stated as follows: At a ^Jv the
preparing fixed consonance can not be of the same value as the consonance to which
the suspension resolves (the disengaged voice remaining stationary at the moment of
resolution), furthermore, at a ^Jv the note against which a suspension sounds can not
be prepared by a note on the next higher degree.
CHAPTER VI
VARIABLE INTERVALS
(3) Variable Consonances

§ 133. Variable consonances are those that in the original combination coincide
with dissonances in the derivative (§ 96). Since within the octave limits there is one
more consonance than there are dissonances, the consonances of a given Jv can not all
coincide with dissonances, i.e. can not all be variable; every Jv has at least one fixed
consonance (§ 97). This is true of the indices of the sixth column and also of JJv- —10,
—6<, 1; in all of these the consonances, with the exception of one, are variable (cf.
table, § 108). The JJv that have no variable consonances at all are JJv= —9 and —4>.
A number of variable consonances from other JJv are found between these extremes.
§ 134. A variable consonance must take over the limitations of the dissonance
which it gives in the derivative combination (§ 70). If a note with the sign — progresses
to a dissonance at the point where a resolution would normally occur, this dissonance
must be treated as passing or auxiliary, and takes the supplementary sign x (§ 92).
For example. /u= —11 has one variable consonance, 5. Adding to it the sign of the
derivative gives_5_. But to the lower sign must be added the supplementary sign, since
the progression of the lower voice one degree downward gives the dissonance of a
I— I O
f?>
seventh
PP -The conditions of using a sixth are expressed, therefore, by 5
Similarly, at Jv= 2 the conditions of using the only variable consonance, the fifth, are
expressed which in the derivative gives:

3E
Original Original
Original
^•-J-fi-
A. S- B. C.
s
JJv= 2x 3 JJv=
-10
'-2
X A- JE3 JJv=
-11 fc^ff
rrr MH^
X r fr
i
FJT-pSF fet
f^f
=s ii
m
gjyw r f
pz-im pj=m y^
90
VARIABLE INTERVALS 91
The sign x also has another meaning as shown above: If a consonance taking the
sign —x is imperfect a dissonance is formed, to which the sign refers, at the moment
of resolution; which in turn functions as a tied interval. As the beat on which it appears
is comparatively strong, the dissonance resolves as though it were tied.
In § 161 will be shown the application of the sign x to ornamental forms of resolution
(cf. also § 167).
Obs.-In compositions in the strict style written in j, time it is very uncommon to
find suspensions on the second and fourth beats:
Palestrina
Ex. 88 S. Dietrich Ex# 89
tRR?rt | g: etc. <b - o MyimJk

-3 « etc.
Ex. 90 Costanzo Festa
J=± etc.
r . -l'i Li i
§ 135. The supplementary sign is not needed when the resolution of a variable
consonance is to a fifth or a sixth. This happens when (a) the sign - is underneath
the fifth ( 4 >. and (b) when it is above the sixth (
In all other cases the sign — for a variable consonance must always be accompanied
by the supplementary sign.
§ 136. In the table to follow, the figures 4 and 5 in the two upper rows will take a
short dotted line on whichever side the interval can not take the supplementary sign: 4,
5. In the second row of numbers this same sign will indicate the progression of a fifth
to a sixth or a sixth to a fifth. Applied to negative intervals the sign is placed on the
"
'ZtL
_ _n
opposite side: -5 and -4 will refer to — and -» respectively (§ 88).
*#
The same sign will also apply to the compounds of the fifth and sixth, both pos
itive: 11, 12 and negative: -12, -li. Examples:
a) b) c) d)
Jv = — 11 Jv=l J\ Jv=--12
m= 5 m=5 m= 4 m= 4
c—>
w=-6 n=6 «=-3 n = -8
c- >
—X
5 B 4 4
X <—i
The sixth at /v=-ll (a) and the fifth at Jv=-7 (c) are the only variable consonances
at these indices.
§ 137. With the exception of the cases mentioned, the fifths and sixths and their
compounds, all other consonances that acquire the sign — from a derivative dissonance
take also the supplementary sign *.
JV = 4 Jv = — 8 Jv=— 10 Jv = — 4
m=2 m= 7 m= 2 w=7
n=6 »=,-l w=-8 n=3
C 3 C— 3 — —
-x -x -x -x
2 7 2 7
C—3 C—3 -X -X
(4) Variable Dissonances
§ 138. The last cases of coinciding intervals in the original and derivative are the
variable dissonances, i.e. dissonances that give consonances in the derivative (§ 96).
The number of variable dissonances changes according to the JJv concerned. All are
variable in the Jfv of the sixth column, while in /u=-9 and /v=-ll every disso
nance is fixed.
§ 139. At indices with the inverse shift the variable dissonances are equal in number
to the variable consonances (this follows from § 32, [3]).
§ 140. When a derivative consonance is perfect it will be indicated by a small
p. placed above the original dissonance. For example, at Jv=3 the second will be
o P
indicated: 1; at Jv= -10 the seventh: (6^ These and other intervals give a perfect fifth
as a derivative consonance.
§ 141. Since no interval may be released from the limitations imposed on it by the
rules of simple counterpoint (§ 94), a variable dissonance must retain its own signs.
But in some cases the supplementary sign x must be added to the sign — . In fact, with
the exception of the fifth and the sixth, the two consonances in the successive series
having a dissonance on one side only, all the other consonances in the series have a
dissonance on both sides (§ 14, [2]). Therefore, when a variable dissonance coincides
with a consonance that has a dissonance on each side, the progression in the derivative,
corresponding to the note of resolution, will be to one of these two dissonances (i.e.
the original dissonance resolves to a variable consonance), and the sign x
is indispensible.
(a) (b) (c) (d) (e)

Jv= 1 Jv = -1 Jv = -10 Jv = -13 Jv = 6
m=l m=3 m= 3 m= (I) m= 8

p p p
n=2 n=2 71 =-7 n = -7 n = 14
-X
2 3,
A 8—X
When a variable consonance progresses to a fixed consonance at the moment of

resolution the sign x is not needed. This is possible only in those cases where a dis
sonance and its resolution come on a fifth or a sixth, or on their compounds (4 and 5,
-5 and -4, 11 and 12,-12 and -11; cf. § 156).
JV = 3 JV = 2 Jv = --10 JV = 4
c—>
m=l m=3 m= 6 M2= 8
C— 3
p P
n =4 n =5 n =-4 n = 12
p p
C— 3
1 8 6 8
X C—3 —X
Obs.-It follows from the foregoing that all dissonances taking the sign -x are variable.
§ 142. Among the variable dissonances are some giving an imperfect consonance in
the derivative, others perfect consonances (the latter indicated by p., § 140). A derivative
imperfect consonance can not impose any new limitations on the conditions under which
the original dissonance is used. On the contrary, a derivative perfect consonance may
add its own limitations to these conditions (§ 73). Hence the following rule applies to
dissonances taking the letter p.:
§ 143. Similar motion to a variable dissonance taking the letter p. is not allowed, for
in the derivative it results in hidden consecutives or forbidden parallels. E.g. similar
motion to a seventh:
Original
is inadmissable in two-voice counterpoint at any index where 6 takes the letter p.,
namely: at Jfv of the first column and at JJv= 5, —2, 10
Jv=il Jv=-2 jvz-13 ,ir-7
Derivatives
n ^zr pr^ff
The parallel progression of dissonances taking the letter p. is unconditionally bad-
SITE
at JJv= 1, -2 and -10.
v=l Jv = -2 Jv = -10 I
Jv=l
X
Derivatives
1 4 4
II
Obs.-The foregoing progression of parallel dissonances are found in simple counter
point of the strict style, though seldom. Exs. 91 and 92 show the use of two fourths and
two sevenths in parallel motion at the same time. In Ex. 92 notice the leap in the alto
to a note consonant with the other voices.
Ex- 91 Palestrina Ex. 92 Ibid.
3E 3E
m an
r
M
mi
gg r 33r " mrs-
Parallel seconds are occasionally tolerable, but their use is not to be recommended:
Ex. 93 Ibid.
1 1
etc.
§ 144. It follows that the rule given in § 74, (2), applying to perfect consonances,
is irrelevant as applied to variable dissonances, since dissonances in contary motion
are not allowed, either within the octave limits or beyond.
§ 145. The restriction of perfect consonances concerning the proximity of intervals
of the same size (§ 75) applies also to dissonances taking the letter p. However, this
refers not so much to the case of such a variable dissonance used as a tie on an ac
cented beat as it does to a dissonance taking the letter p. coming on an unaccented
part of the measure as resolution of a variable consonance.
Concerning the first situation mentioned, in syncopated counterpoint perfect con
sonances of the same size may be found on the accented beats of successive measures,
therefore the following cases are entirely admissable:
Original m
Jv = -2
fnfr
-2 (->
Derivative
Here the seventh in the original, giving a perfect fifth in the derivative, can be freely
used as a suspension on the accented beats of adjacent measures. But if the seventh
appears on the weak parts of the measure a faulty succession of fifths will result in the
derivative. For example, at Jv- 2:
-&gm**mp*
These cases are closely related to the preparation of variable intervals, to be
dealt with next.
The Preparation of Variable Intervals
§ 146. The object of the rules for the preparation of intervals in complex counterpoint
is to prevent the appearance in the derivative of progressions that are forbidden by the
rules of simple counterpoint (§ 70). These rules refer to the dissonances that resolve to
perfect consonances (3 and 8) and consist of statements to the effect that these dis
sonances cannot be prepared by perfect consonances equal in value to those to which
they resolve (§ 86). In vertical-shifting counterpoint corresponding rules must be estab
lished that refer to all tied intervals. One rule for the preparation of fixed consonances
was given in § 132. There are two more, referring to the preparation of variable in
tervals; they are:
Rule 2. -A variable consonance, of which the free note is approached by stepwise
movement from above, can not be prepared by a dissonance taking the letter p. (i.e.
n- perfect consonance). This rule eliminates such situations as these:
Originals
at Jvz-2 at J\T-3 at Jvz-4
at Jv = l
bad i nr
-T P- 3E12 ftIS " hri i - f-T^Sf

3 _x • 6 -r 6 t 1
iV^fp
8 —x »
at Jvp-6i- , a i Jv=-7
at ..!»'-.-/ i
%P 2X * lp
It) -x
T p X P
—X
All these originals give the same bad derivative:
Originals .
at Jvzl at Jv=4 at JV--8
bad -
The following combinations give a bad preparation of the ninth in the derivative:
Derivative bad
5teE±
^rf
These originals also give the same faulty derivative:
Derivative
Rule 2.—At JJv a variable dissonance can not be prepared by a fixed consonance
equal in size to that to which it resolves:
a) Original b) Original
Impossible
at Jv=-1
a) Derivative
: Jv=-6=
This rule does not apply to those cases where the preparation of a consonance gives
a fourth; such fourths may become consonances by the addition of a supplementary voice
below. In this way original (a) at Jv= 1 and original (b) at Jv- 2 are entirely admissable,
as they give the following correct derivatives:
a) Derivative b) Derivative
: ;.Tv-i : Jvz-2
Sup. voice
W=f
Sup. Voice
Obs.—These rules for the preparation of variable intervals must be strictly observed
in all cases where, in the derivative, an octave is prepared by a ninth (8). Likewise,
where these rules prevent the appearance of 3, prepared by a fifth, a measure of their
severity arises from the fact that in general such preparations are not allowed in simple
counterpoint. It cannot be denied that the masters of the strict style indulged in pro
gressions that later theorists were inclined to regard unfavorably (Ex. 64). In the future,
when discussing complex forms of resolution, the rules for the preparation of variable
intervals will to a certain extent lose their force (§ 162).
§ 147. As has been seen, cases breaking the rules of preparation occur only among
those progressions where the free note to a tied interval is approached by stepwise
movement from above. Therefore, no irregularities are encountered—
(1) When the tied interval is prepared by a fixed consonance that approaches the free
note by a leap, as is seen in the following preparations of variable intervals:
Original —g -, » H .. + ... -t
Derivative =^^j
If iv-.>
Exchanging original and derivative (§ 68) illustrates the preparation of var
iable dissonances.
(2) When the tied interval is prepared by ascending stepwise movement to its free note:
Derivative
Exchanging original and derivative again illustrates the preparation of var

iable dissonances.
Simplifications in Compiling Tables of Indices
§ 148. Before citing examples showing the application of variable intervals, some
tables will be given that will result in important simplifications.
(a) Tables of Indices Giving the Inverse Shift
§ 149. A table for any index having the inverse shift consists of a series of original
intervals beginning with 0 and ending with an interval equal to. Jv. Such a series is
divisible into two equal parts. If the index is an even number the division falls on the
interval equal to one-half of the index value; this is called the central in
terval. For example:
i
/v=-8:012345678 (central interval 4)
/u=-10:012 34 5 678910 (central interval 5)

i
If the index is an odd number the division occurs in the space between those two
intervals of which the sum is equal to the absolute value of the index; e.g. for Jv = —7
these two intervals are 3 and 4:
/v=—7: 0 12 3 4 5 6 7
for Jv = —9 they are 4 and 5:
Jv = -9: 0 1 2 3 4 5 6 7 8 9 etc.
§ 150. According to the foregoing, a central interval is an original interval equal

to one-half of an even-number index for the inverse shift. Since m + Jv=n (§ 31) the
central interval gives in the derivative a negative interval of the same value. For ex
ample, at /t>= —8, from the central interval 4 is obtained n- —4; at Jv= —10 the central
interval 5 gives n- —5 etc. Therefore, whether the central interval is a consonance or a
dissonance, it will appear in every case as a fixed interval.
If this interval is a consonance then it is a fixed consonance, free from the tie-sign.
If a dissonance, the upper and lower signs are always the same, i.e. it appears either
as m or m. This results from the fact that the derivative interval will be a negative
interval of the same value and with the same signs but in reversed positions, the upper
sign below and the lower sign above (§ 89). Consequently if one sign of the central
interval in the original series, according to the rules of simple counterpoint, is (—), its
derivative interval has the same sign on the opposite side, the final result being that
the original interval will take this sign on both sides: 'm. If the central interval takes
the sign — above and below the derivative will be the same, therefore the original retains
its own signs: n7.*
§ 151. The sum of the two intervals lying next to the central interval and on either
side of it is always equal to the absolute value of the index, since the division falls
either on the central interval or in the space between the two nearest the center. This
same quantity is also equal to the sum of every two intervals that lie at the same
distance on either side of the center, i.e. the sum of the second interval to the right
and the second to the left, of the third interval to the right and the third to the left, etc.
In the following two series of original intervals (one for an even-numbered index, one
for an odd) every two intervals of which the sum is equal to the absolute value of the
index are connected by bracketed lines.
Jv = -8: 0 1 2 3 4 5 6 7 8
■ ^— i ! ■
}v=-7: 0 1 2 3 ! 4f 5 6 7
uj=^= ! ' .
Either of the two intervals connected by brackets (here referred to as a and b) may
represent the derivative if taken with the minus sign: if m = a, then n- —b; iim-b,
then n- —a (§ 32, 3). And since a negative interval has the same tie-signs as the positive
interval of the same value but with the positions of the signs reversed (§ 89), the fol
lowing relation holds between the tie-signs of the original intervals:
At a negative index giving the inverse shift, any two original intervals of which the
sum is equal to the absolute value of the index have the same tie-signs but in reversed
positions: the upper sign of one interval is the same as the lower sign of the other, and
vice versa. For example:
'The foregoing applies only to the central interval of an even-numbered index. Evidently the
author did not consider it necessary to explain what happens in the case of an odd-numbered
index for the inverse shift, but this works out in such an interesting way that the following is
submitted as supplementary to the text:
A series of intervals for an odd-numbered index with the inverse shift gives two central
intervals, of which the sum is equal to the absolute value of the index (§ 149). (1) If one central
interval is a consonance and the other a dissonance they will always be a variable consonance
and a variable dissonance. If the signs for one are different, the other will have the same two
signs but in reversed positions. If the signs for one are the same the other will have the same
two signs, though this occurs in only one index: jv=-5<. (2) If both central intervals are con
sonances both will be fixed; they will have the same signs but in reversed positions. This also
occurs but once, in Jv = —9>. —Tr.
VARIABLE INTERVALS
99
/t>=-ir
(1)
Original: 5 &
f__) of simple counterpoint
Derivative: -6 -5
Original at Jv= —11:
(2)
Original:
. of simple counterpoint
Derivative: -10 -1
— c->
r—> —
Original at Jv= —11: 1 10 etc.
— t—)
No reference is here made to the sign p., but it must be placed above every dis
sonance that is bracketed with a perfect consonance (§ 140).
§ 152. This relationship of signs makes possible the compilation of tables for
indices with the inverse shift. The first row in such a table consists of the left half
of the original series (§ 151), including the central interval if the index is an even
number. Under this, as second row, is placed the right half of the original series, but
in reversed order and with all signs reversed. Below the horizontal line is the third
row, beginning with 0 and ending with a number equal to the absolute value of the
index. The derivation of the signs for the left half of this row has already' been explained;
those for the right half are derived from the left half, using the reciprocal relationship
of signs previously referred .to. The following tables for Jv = —8 and /w= —12 illustrate
the process:
1) 2JV,
1' C—J V
0 1 2 3 4
p p
-8 -7 -6 -5 -4
p p
—X C—3 c—> —X —X —X -x
0 1 2 3 4 5 6 7 8
—X —X —x —X c—y —X
11 L_ i 1 1
J
•It is of interest to note the fact that the use of variable intervals was unknown to Zarlino. In
Chapter 56 of the "Istitutioni armoniche," speaking of double counterpoint at the twelfth, he
gives a rule not to write sixths and sevenths: "Osservaremo di non porre la Desta nel principale;
imperoche nella Replica non puo far consonanza." And further on: "Non porremo anco la Sincopa,
nella quale se contenghi la Settima: perioche Replica non torne bene." In Zarlino's terminology
"il Principale" refers to the original combination, "la Replica" to the derivative.
In this table the signs of each interval in the third row, counting from the center
to the right, are the reverse of those at equal distances counting from the center to the
left. The sum of any two intervals bracketed equals 8. Intervals 1 and 8 take the sign p.,
they are bracketed with the perfect consonances 0 and 7.
Ex. 94
Jv=~8 W+
2) 2Jv = — 12.
p c—> — p
0 1 2 8 4 5 6
c—>
p
C—3
-12 -11 -10 -9 -8 -7 -6
C—3 —X —X —X C—) — —v —x y.
1 2 3 4 5 6 8 9 10 11 12
—X —X —X C— > —X —X X c .>
L ' ' J I |
Ex. 95
Jvr-12
Original
Derivative
k .. P A.<~
'w. yi. -'.". "I-,
-fr, h . l"2'l L,
><; r=-'»
-5* ''1-8" jgj i r .%*
For these examples //v were selected that have the largest number of variable
intervals. But in the exercises to follow, start with those that have the fewest, such
as fv=-ll,Jv= —7 etc.
(b) Tables of Indices Giving the Direct Shift

§ 153. In compiling tables for indices with the direct shift it is impracticable to
present any table complete-a portion from the beginning is enough. In these tables a
series of original intervals begins at 0 for a positive Jv, and for a negative Jv at an
interval equal to the index value. Either series may be continued as far as desired.
§ 154. Among the tables of indices with the direct shift, any two that have the same
value, one positive and one negative, are closely related, and a study of their properties
results in simplifying the conditions of a negative Jv as compared to those of a positive
Jv, and vice versa. This will be illustrated by two tables, one for a positive Jv, the
other for a negative Jv of the same value.
VARIABLE INTERVALS
101
p C—)
Jv= l: 0 1 2 3 4 5 6 7 8 etc.
C—O
1 28456789
C—3 c > —X p
—X — p
-!-x —X —X
0 12 3 4 5 6 7 8
—X -M —X — c—5 c—> —X —X
»-» — P — p _
Jv- 123456789 etc.
— — C—3 —
P «=5 — P — p _
012 345678
— C—3 —
1
c-> C-J -X -X - -X P
-Lx -X
123456789
-x -x —x e-a c-> —x —x
Comparing these tables, it will be seen that the two upper rows in one table are
reproduced in the other, and that as a result of this coincidence the tie-signs of the
intervals in the third row are the same for both tables; therefore, from the third row
of one table can be formed the third row of the other, if, reversing all tie-signs, the
figures for the derivative intervals (second row) of one table are used for the third row of
the other. A combination written at one of these indices will serve equally well for the
other if the original and derivative are exchanged (§ 63, 1).
Ex. 96
Derivative
PC*
etc.
06s. -In working out exercises, use can be made of the tables accompanying the
examples in two-voice vertical-shifting counterpoint .The original intervals are arranged
in three rows. The fixed consonances are in the middle row, with the dissonances and
variable consonances in the top and bottom rows, the upper row containing only intervals
with the signs - and x above, the lower row with the same signs below. An interval
having either sign above and below appears twice, both in the upper and lower rows.
Intervals having (-) as both upper and lower signs are omitted altogether. Presented
in this way the tables are of special importance in the first exercises. They show at a
glance what intervals are available, and the proper use, in each voice, of ties.
Distribution of Variable Dissonances
§ 155. Some variable dissonances give perfect consonances in the derivative, others
imperfect. Sometimes met with in the same Jv, their distribution may have appeared
quite fortuitous, but a strict order underlies the apparent disorder. The distribution of
variable dissonances from any Jv can be expressed in one of two tables, called System A
and System B. Those for A>= -1 and Jv= 1 will serve as illustrations. The variable
dissonances are listed in numerical order
System A System B
f>- -1) (*- 1)
Dissonances Consonances Dissonances Consonances
of the of the of the of the
original derivative original derivative
combination combination combination combination
P-
1 per/.
3 .... J
6 . . . . .rmP-
p-
8 perf.
10 i
13 hmp-
The same table in notations:
Dissonances of
original combination
Consonances of
derivative combination
at Jv=-1
at Jv=l
(System B)
porf. imp.
One system or the other includes the following general features: Intervals 1 and 8
(i.e. the second and its compound) give the consonances of one group; in System A
perfect, in System B imperfect. Intervals 3, 6 and their compounds 10 and 13 also give
in each system consonances of one group, but the opposite group of that to which
1 and 8 belong. In System A three intervals (3, 6, 10, 13) give imperfect consonances
in the derivative, in System B, perfect consonances.
In the preceding table the indices took the direct shift. The next shows those for
the inverse shift:
Dissonances of
original combination
Consonances of
derivative combination
at Jv=-8
(System A)
at jv=-10
(System B)
Those indices have been intentionally chosen in which all the dissonances are
variable. But no matter how many variable dissonances a given index may have— if they
are not less than two—they may always be arranged according to one of the two systems.
§ 156. It is not difficult to explain the reason for arranging the variable dissonances
in this way. It has already been seen (§17) that the second and its compounds (1, 8, 15
etc.) are separated from one another by a int. (referring here to the positive intervals
of the original combination). Whichever of these dissonances appear as variable will
infallibly give, in the derivative, consonances that belong all to one group, since only
such consonances are separated from each other by a Mnt. (§15, [1]). It is exactly the
same with the other dissonances: 3, 6 and their compounds, that are separated from
each other by a int. (§17, [2]). Therefore, since they are variable, they must likewise
give derivative consonances also of one group.
Moreover, the dissonances of the series 1, 8, 15, etc. are separated from the dis
sonances of the other series (3,6, etc.) by a înt. (§17, [3]), consequently two dissonances
from different groups must give consonances from different groups in the derivative,
since only such consonances are separated from each other by a înt. (§15, [2]). In this
way the distribution of variable intervals according to one of the two systems is shown
to be the necessary result of the distance between the intervals in the successive series.
Because of this systematic order in the arrangement of variable dissonances it is
sufficient to know the relation of one of them to an interval of the derivative—whether
it gives a perfect or an imperfect consonance— in order to determine the group to which
the derivative consonances of all other variable dissonances of a given index belong; in
other words, to determine which of the two systems of distribution apply in a given case.
§ 157. The variable dissonances in ]v= —1 and Ju= 1 are those of two indices of
equal value but of opposite signs; this is explained by the fact that in the successive
series of intervals each dissonance has a consonance on each side, one of which is
perfect, the other imperfect (§14, [3]). At Jv= 1 each dissonance gives in the derivative
the next consonance to the right, and at Jv=* —1 the next consonance to the left. There
fore the dissonance which at Jv= — 1 gives a perfect consonance gives an imperfect
consonance at Jv= 1 and vice versa; i.e. Systems A and B are interchanged.
If attention is directed to the fact that of the two consonances found at equal dis
tances to the right and left of a dissonance, one is always perfect and the other imperfect
(§ 14, [4]) it will be seen that everything referring to the distribution of variable dis
sonances at Jv= —1 and Jv= 1 applies also to all other indices of the same value
but with different signs. Positive and negative indices of the same values always
have their variable dissonances arranged according to either System A or
System B.
§ 158. The distribution of variable dissonances according to the different indices
is shown in the following table:
Order of JJv with Variable Dissonances

System A System B
Jv = — 1 JV = + 1
jjH~5 JJH±5
jjv=K jjv=!"-?
•Jv= 7 has no variable dissonances.
The tables could be continued, alternatihg positive and negative indices in the same
way. }v- 9, like Jv= 7, has no variable dissonances.
If from the foregoing lists the indices are selected that correspond to dissonant
intervals the following table is obtained:
Order of JJv with Variable Dissonances

Corresponding to Dissonant Intervals
System A System B
Jv= — 1 Jv= + 1
JJV_|+ 6 JJv-(_ 6
Jv= - 8 Jv= + 8*)
JJv_f+10*) t -10
Comparing this table with that in § 155 it will be seen that the arrangement of
positive and negative indices is in exact agreement with the order of variable dis
sonances giving n= perfect consonance and those of which rc= imperfect.
§ 159. With the investigation of variable intervals is ended the most essential part
of the study of two-voice vertical-shifting counterpoint.
•Unusable.
CHAPTER VII
SUPPLEMENTARY RESOURCES
Ornamental Resolutions of Suspensions
§ 160. Ornamental resolutions of suspensions were mentioned in § 84 and in § 85, (2).

In simple counterpoint they are applied only to suspensions; in vertical-shifting counter
point both to suspensions and to tied variable consonances. At the sign (—) in the cor
responding voice it is not possible to apply the ornamental forms of resolution. First
to be considered are the forms involving the free note to a tied interval; these are used
the most, especially in multi-voice counterpoint.
(A) The Free Note Progressing to a Consonance

at the Moment of Resolution
§ 161. Applied to vertical-shifting counterpoint, the rule given in § 85, (2) may be
restated thus: The free note to a tied interval may, at the moment the latter is resolved,
progress to another note which forms a fixed consonance with the note of resolution. For
indices of the second group is added: At a Jv the voice taking the free note can only
continue upward one step. This prevents similar motion, forbidden at a ~/v in two-voice
counterpoint. Since the resolution of dissonant notes is the result of a voice moving one
degree downward, the other voice, in order to avoid hidden consecutives, must progress
in contrary motion.
At the sign —x the note of resolution can also take a fixed consonance. The note
of resolution, to which the supplementary sign x refers, is in this case part of a con
sonant interval, and need not continue by stepwise movement in the same direction, as
it would have to do were it a dissonance (§ 92).
In the following illustrations are shown these forms of resolution applied to the
original dissonance. Above each is a list of the JJv at which the given dissonance
takes the sign — without the supplementary sign. Below each illustration are listed the
JJv at which the same dissonance takes the sign —x. The dissonances with this sign
are all variable (cf. § 141).
Examples
JJv of the 3rd column (among them Jv = 3*),K>.

'(b) JJVr 2,-7,-9
(a)
i JJv=-6£,-9, -13*'.
I
1 JJVr.1,-8.
rSFfm
JJv= 4, 6,-l,-3>, -5>, -10, -12.
105
JVr-9 JJv=2,-6*),-7,-9. Jv=-7»> JVz-7*)
(c) 6 (d) 6 (e) 3 (f) 5
ifefe igi d
? fW m
*,
JJv of the 1st & 3rd columns & JJv of the 3rd column & JJVz 4,-10.
JJv = 5. -2, -8. JJVr 5,-2,-13.
JJv of the 2nd column (amoung them Jv = -12*); JJv of the 5th column &
JVr-9
(g) 3 (h) § UV= 4',,-8**-fc JV=-11.
(i)
p—p:
^T=
2,
r^
8* 8"
JJv of the 2nd & 4th columns & JJv= 6,-1,-10. JJv of the 1st column &
JJV- 6,-1,-7*). JJv= 3,-4,-8.
It is not necessary to give illustrations for variable consonances, as the foregoing

using dissonances will serve the same purpose. It is well to select, from the derivative
combinations, an index that includes the required consonance, then to use this com
bination as the original and its original as derivative (§ 68). As to the application of
these forms of resolution to multi-voice counterpoint, cf. § 213.
§ 162. In Obs. 2, § 86, it was stated that this form of resolution allows the preparation
of a fourth by a fifth: It constitutes an exception to the rule of prepa

Fff
ration previously given in § 86. Correspondingly, this -exception is also allowed in
vertical-shifting counterpoint. For example, this preparation of the fixed dissonance
Original Jv 9
Derivative
6 at Jv* -9: -±H A illustrating an exception to the rule given
*
5 6
,-9
in § 132, would be impossible if the free note did not return to the previous note at the
moment of resolution. Exactly the same thing happens under the conditions by which
the variable consonance 2 is prepared: Jv=l
,v=l
Original Derivative
enva
This is an exception to Rulel,§ 146.

'At this Jv the dissonance is variable.
SUPPLEMENTARY RESOURCES
107
(b) The Insertion of a Consonant Note Between a
Dissonance and its Resolution
§ 163. Another complex form of resolution, one of far greater importance in counter
point than the preceding, consists in this: Between a suspension and its resolution is
inserted, where the voice leaps downward, a note consonant with the free note (§ 84). In
extending the application of this form to all tied intervals in vertical-shifting counter
point, one requirement is necessary, referring to the voice in which the signs - or x
occur: at the leap to the intermediate note, the latter must be a fixed consonance.
(1) Dissonances taking the sign -:
(a) Fixed
JJv = -9, -7, (2) Jv = -9 (b) Variable
Jv = -7. 2 Jv = -6>, 3
J3^T'
-9—0° rs—
m to
a^N 9
' 9 ri*i
Lrj r
o
(2) Dissonances taking the sign -x. In a voice taking the sign -x this form of re
solution can be applied in one of three ways:
(a) The note of resolution, to which the sign x refers, can change to a passing-note
in the derivative. This is possible only when the leap from the tied note is a de
scending third:
JJv of the 1st & 3rd column & JJv = -8, -2, 5.
(b) If the note of resolution, to which the sign x refers, is a fixed consonance, both
forms of complex resolution may be combined:
JJv of the 3rd column
w
P
i
(c) The note of resolution, to which the sign x refers, can function as the free note
to a tied interval (cf. § 134):
JJv of the 3rd column & Jv = —8.

nepB.f JJv of the 3rd column JJv of the 4th column
E
f m
Illustrations showing the application of these forms of resolution to variable con
sonances may be obtained by the method described in § 161.
The Cambiata Figure
§ 164. The descending stepwise movement of a voice may be interrupted at a dis

sonance on the weak part of a beat, from which a leap of a third downward is made,
to a consonance:
U^x±±±n
*E£
XT
iES?
*32
m -P-
Such a dissonance, often found in the strict style, is called the nota cambiata.
Following the leap the voice usually progresses one degree upward to a note of the same
value, as shown in the preceding illustrations. However, cases are found where the
notes concerned are of greater value than the nota cambiata, and even where, after the
leap, there is no return at all to the absent degree.
Ex. 97
Palestrina
o " etc.
mm xc mm 3E
C- o]
etc.
Ex. 98 Ibid.
P miH g » P etc.
*e±*
W1 lBp Ex. 100
P i . etc.
Ex. 99
Josquin Ibid.
JSl 0—=-0
mi P 0 etc. etc.
f
miW
m £p*? VSl
TtT-
WE - r.
P
Ex. 101
Ibid.
3ESE rr. etc.
1
etc.
SUPPLEMENTARY RESOURCES 109
Use of the nota cambiata in vertical-shifting counterpoint is usually on condition
that the note to which the leap is made must be a fixed consonance:
JJv of the 3rd column JJv of the 5th column:
1
t
mm â
A variable consonance may be used as a nota cambiata, in which case the preceding
interval must be either a fixed consonance—
JVr 2
.^JJ.J
or a variable dissonance. The foregoing illustration shows the note to which the sup
plementary sign x refers. Examples of such cases can be formed from the illustrations
in § 163, exchanging original and derivative.
The interval preceding the nota cambiata may be a prepared variable consonance,
the prepared voice moving one degree downward to the nota cambiata:
JJv of the 3rd column JJv of the 1st column

0—X
.i a a I rU^r?
F^
II var.
cons.
II var.
cons.
m
In a variable dissonance the note of resolution, indicated by the sign x, may be
treated as a nota cambiata, making a leap of a third downward to a fixed consonance.
Examples may be formed from the foregoing illustrations, exchanging orig
inal and derivative.
Obs.—The nota cambiata is often found in the works of Russian composers:
Ex. 102 Glinka
Si
J:
B i ^^
f= r
Tschaikowsky
Ex. 103
£S
m MP
/
HO CONVERTIBLE COUNTERPOINT
§ 165. These complex forms of resolution and the nota cambiata will add freedom
and variety to voice-leading; this applies especially to those indices that have few fixed
consonances, and therefore present the greatest difficulties in the use of intervals.
§ 166. In the examples applying to Division A (Appendix A) are found all the complex
forms of resolution—the nota cambiata etc. In writing exercises, study the examples
according to whatever index is being worked at the time.
The Use of Notes Taking the Sign x
§ 167. In the preceding chapters the various applications of the supplementary

sign x were explained, but this information was necessarily scattered among different
sections of the text. It is herewith summarized.
(1) The supplementary sign x is placed after the tie-sign, and means that the note
following the one tied is a second lower, the voice descending one degree (in the original
or in the derivative) to a note dissonant to the free note.
(2) It may indicate a dissonant passing or auxiliary note (§ 93).
(3) The note taking the sign may in turn act as the free note to a new
tied interval (§ 134).
(4) The note indicated by the sign may take a fixed consonance, if no rule of voice-
leading is thereby infringed (§ 161).
(5) The sign may indicate the nota cambiata (§ 164).
Negative Intervals in the Original Combination
§ 168. In stating the rules of vertical-shifting counterpoint it was assumed that the
original combination formed a series of positive intervals, i.e. no crossing of voices
was to be found in it. In this way needless complications were avoided that would have
been inevitable had it been necessary to study combinations in which positive and
negative intervals were intermingled. Moreover, for combinations intended for vertical
shifting, the absence of voice-crossing is, generally speaking, advantageous. This
condition enables the parts to be clearly distinguished by their melodic contours and
they are easily recognized when shifts are made, especially in double counterpoint.
Nevertheless these considerations are not sufficient to justify the complete removal of
crossings from the original combination, as is insisted upon by some theorists. No
composer of either the strict or the free style has felt obliged to observe such a condition
(cf. The Hell-Tempered Clavichord, Vol. I, Fugue 11, and Vol. II, Fugues 17 and 23,
where the subject and counter-subject cross). Even in some cases, for instance in more
or less extended imitation at close intervals such as the unison or second, crossing
certainly contributes to freedom in voice-leading, and can be recommended for this purpose.
§ 169. In the tables of indices only positive intervals of the original combination
were given, as there was no necessity for including negative intervals. To know the
conditions in which a negative interval is used at a given index, regard the negative
interval as positive, change the index to the opposite sign, and then use the interval
according to the conditions of the index obtained. If the latter is positive and equal
to a compound interval, substitute the index equal to the corresponding simple
interval (§ 47).
SUPPLEMENTARY RESOURCES 111
etc.
etc.
4 3 -5-6-5-4-3-2 -1 -2-3 8
Derivative
etc.
±=d
Jvr-6 r#
[HS^h-pf^S xr etc.
In this example at ] v- —6 all the intervals except the first two and the last are
negative. Taking the positive index (/t>= 6) instead of the negative, and regarding the
negative intervals as positive, the conditions of their use will be found in the table
for this index.
In this example at }v- —11:
Ex. 105 Original
«_
i &m~ ^m
-1-2 .3-2 2 3 5 6
-o-
ii ES=i em f-
Derivative
JVz-ll.
i ^=f m
^m *
treat the negative intervals as positive and write according to Jv= 4, the substitute
for Jv= -11 (§47).
§ 170. It follows that negative intervals in double counterpoint at the octave (/v=— 7)
are written according to the rules of simple counterpoint. In fact, ]v= 7 must be sub
stituted for Jv - —7, for at this index the derivative is nothing but the separation at the
octave of the voices of the original (§ 47).
Ex. 1-06
^d
I fcTFg
Iff UP Jl^ JsHg
-2 -3
etc.
Jv=-7
^^^^ etc.
>." i i
rr'rrr
CHAPTER VIII
EXERCISES IN TWO-VOICE VERTICAL-SHIFTING COUNTERPOINT
§ 171. The study of convertible counterpoint should run parallel to that of simple
counterpoint. In entering upon a course of exercises in two-voice simple counterpoint it
is well to apply, from the start, the method of figuring intervals used in this book; this
can be easily learned in two or three lessons. Passing to syncopated counterpoint (fourth
species) learn what is said about suspensions in Chapter IV, beginning with § 81, and
memorize the table in § 90.
§ 172. In two-voice problems on a cantus firmus written in simple counterpoint with
out any calculation for shifts, places are often found where a shift in double counterpoint
at the octave is possible. It will be useful to pay special attention to such cases, and
study the conditions under which such a shift may or may not be made.
§ 173. The first exercises in vertical-shifting counterpoint may begin after three-
voice simple counterpoint has been studied, and before four-voice, if at this time the
student has mastered the contents of the first four chapters of this book.
A good preparation for the study of vertical-shifting counterpoint is to work out exer
cises in two-voice mixed counterpoint,* under conditions in which suspensions and
similar motion are absent. The dissonances in such problems appear only as passing
or as auxiliary notes. It is very useful to be able to write counterpoint using only oblique
and contrary motion, its importance consisting in the fact that the indices of the second
group ( JJv) do not admit of similar motion. Such exercises are unusual, but it is a study
that should be mastered for its own sake in simple counterpoint, while for convertible
counterpoint it is absolutely essential.
After this, when the student gets acquainted with the fixed consonances (Chapter V
up to § 108 inclusive) the problems in double counterpoint at the tenth (Jv = —9) should
be taken up, avoiding tied dissonances (Exs. 72, 73, 74). Similar motion still absent,
these forms impose new restrictions (the limiting interval 9>) and the conditions given
in §§ 104—5). Breaking with the long-established tradition—that the study of complex
counterpoint should begin with double counterpoint at the octave—has some justification.
From the very first exercises the student is trained to the strict and unaccustomed re
quirement that similar motion must be absent, and passing on to JJv, where this motion
is allowed, his work is facilitated and his technical resources increased. The overcoming
of the next difficulty— not to use a variable consonance except as a passing note—is
comparatively easy.
§ 174. After a sufficient number of exercises at Jv=—9 (without suspensions) next
pass to the indices of the first group, especially to double counterpoint at the twelfth
(A> = —11) and continue the exercises in the order given in § 110. The indices with the
direct shift may be passed over briefly, limited to some problems on each two indices
of the same value but with opposite signs (§ 110). The division of indices of the first
group begins with double counterpoint at the octave. To take double counterpoint at the
twelfth before that at the octave is not customary, but in many ways this procedure is
the more expedient. Aside from the fact that counterpoint at the twelfth was of primary
♦The author refers here to two-voice florid counterpoint (fifth species) without a cantus
firmus. —Tr.
112
EXERCISES IN TWO-VOICE VERTICAL-SHIFTING COUNTERPOINT U3
importance in the era of the strict style (Zarlino, for example, in the chapter on double
counterpoint in his Institutioni armoniche, passes on to double counterpoint at the tenth
and does not mention counterpoint at the octave at all), to begin exercises at indices of
the first group, with counterpoint at the twelfth, is better, because with the limiting
interval 11> it offers greater freedom of voice-leading than does double counterpoint
at the octave.
§ 175. In working out exercises first write the original combination. Time can be
saved by using three staves, so that one voice remains in place on one staff, the other
voices of the original and derivative combinations appearing on the other two. Next shift
whichever voice will cause the fewest difficulties as regards augmented fourths and
diminished fifths. To do this, previously read through mentally both voices concerned
in the shift. Every place in the derivative where a rule of simple counterpoint is broken
means that in the same place a rule of complex counterpoint is also broken. It is then
necessary to be certain as to what the infringament is, and to make the necessary cor
rection. Verification is simple, and the method is easily applied by the student who, in
his approach to the study of complex counterpoint, must make free use of the resources
of simple counterpoint.
§ 176. Passing to the indices of the second group (except Jv=—9, which hasalready
been given), work out exercises in the order of indices given in § 111, Exs. 75 to 78.
Exercises in shifts at several indices (§§ 113—18) and those in duplication of imperfect
consonances (§§ 119—22) may best be deferred until later, taking up the former just
before Chapter IX, the latter before Chapter X.
§ 177. Concurrent with the foregoing, study the uses of the fixed dissonances (§ 123
to the end of the Chapter) and the rules of preparation (§ 131). Returning to the first
exercises at Jv = —9 (§ 128), work now with all the intervals of this index (in which all
are fixed). After this, continue with exercises showing the application of the fixed
dissonances to the other indices of this group, guided by the text as far as § 130 and
Ex. 87, inclusive.
§ 178. Without interruption, take up Chapter VI on variable intervals and work ex
ercises using all intervals without exception. In studying the examples in this chapter
(review §§ 152 and 154), take those indices in which are found the greatest number of
variable intervals and which therefore involve the most difficulties. But in writing the
exercises begin with those indices that have the fewest variable intervals (/v=—11,
Jv= —7 etc.) and work up gradually to those that have the most.
§ 179. As regards the difficulty of these indices, especially of those having only
one fixed consonance, i.e. JJv= 1, —1, —6, —8, —10, exercises using them should alter
nate with those at the easier indices. The abundance of conditions that surround the
the- indices mentioned involves the disadvantage that the student is often uncritical of
the melodic quality of his work, provided that it is done according to rule. The persistent
use of some of these indices is bad in the long run, as they tend to militate against the
esthetic side of the work. By alternating the difficult and easy indices the former will
be mastered by imperceptible degrees, and instead of reacting unfavorably on the tech
nique of the student will contribute to its development. Exercises at these indices may
be played on the piano, though not without some inconvenience, due to the unusual
fingering. Used in moderation they will assist in developing a virtuoso technic, but
their excessive use will fatigue and injure the hand.
§ 180. In short, in a course of study in two-voice vertical-shifting counterpoint some
indices may be omitted altogether or limited to only a small number of exercises. The
student who is familiar with the applications to be found in the tables of signs will
experience no difficulty in writing combinations also at those indices previously omitted.
In class, teaching time can be saved by working out problems at the board, questioning
the students at the same time. This will lighten the home work, especially as regards
the more difficult problems, and will not retard the learning of the simpler ones. Exer
cises in complex counterpoint should run parallel with those in simple, and application
should be made to the contrapuntal forms using imitation, such as the invention, figured
choral, and fugue. However, this subject requires more detailed explanation than can
be given here.
§ 181. In working out exercises at a given index study the example of the same index
in Appendix A. In Chapter VII will be found the explanation of all uses of the complex
forms of resolution and of the cambiata figure. As one gets familiar with the material
it can be applied.
§ 182. In the first exercises the student will find it helpful to write out
the tables for himself, as they are needed. For the exercises use can be
made of the separate tables given with each example in Appendix A.
§ 183. Problems at easy indices having a comparatively large number of fixed con
sonances can be written on a cantus firmus. This is of advantage in studying the effect
of tied intervals above and below a voice, according to which voice takes the cantus.
But the same circumstances would lead to excessive difficulties with indices having
few fixed consonances, so in such cases the use of a cantus is inadvisable.
§ 184. It is best to write the problems in two-voice mixed counterpoint, and in some
to use imitation, the latter as a preparation for canonic imitation. Strict imitation through
out is not needed in order to illustrate the application of complex counterpoint. Canonic
imitation at any index only requires that the original combination shall be given at a
certain index.
An excellent study in indices with the inverse shift is to write two-voice infinite
canons consisting of two divisions of equal length.* In such canons (except the infinite
canon at the unison) the application on vertical-shifting counterpoint is a conditio sine
qua non, the necessary result of the form itself. Exercises in two-voice canonic imitation
are also valuable as a transition to the more difficult canonic forms. But they should
not be attempted until considerable skill has been acquired in writing mixed counterpoint
at all indices. Otherwise the restraint imposed might exert a harmful influence on voice-
leading.
The practice acquired in writing two-voice vertical-shifting counterpoint gives the
student the technique necessary for the various forms of multi-voice canons, and opens
up abundant material for the application of counterpoint in general.
•Cf. "The Technique of Canon." Chapter II. -Tr.

CHAPTER IX
TWO-VOICE ORIGINAL COMBINATIONS GIVING MORE

THAN ONE DERIVATIVE; COMPOUND INDICES
§ 185. In order for the original combination to yield two derivatives, each having
its own Jv, the conditions of both JJv must be united. An index combining the conditions
of two or more JJv referring to one original is called compound; it may be double, triple
etc., according to how many indices are united in it (§ 28).
§ 186. In uniting indices into one compound index, the rule must be invariably fol
lowed that the fewest limitations are supersded by the most.
Double Indices
§ 187. Before stating the conditions under which the intervals of a compound index
are used, it is necessary to define the groups to which they belong, and thereby to
determine whether or not similar motion is allowed. In conforming to what was said in
§ 186 a compound index is a 2/t> if one of the components belongs to the second group.
§ 188. Next is the question regarding the limiting distance between voices. Here no
control is necessary if the combined JJv are both positive (§ 42). In all other cases one
must be guided by the following rules:
(1) If of two JJv one is negative, its limiting interval is valid also for the combined
(double) Jv.
(2) If both JJv are negative, and both give the direct shift, the limiting interval for
approaching voices (§ 37) equals that of the index which has the greatest absolute value
(§ 186). For example, in uniting the conditions of Jv = -2 and Jv - -A< the limit for
approaching voices is 4<, that is, a fifth. This is indicated: 2Jv - -2, -4<C
(3) If both JJv are negative, and both give the inverse shift, the limiting interval for
receding voices equals that of the index which has the lowest absolute value (§ 186).
For example, in uniting the conditions of Jv - -7 and Jv - -11 the limiting interval is
7>; for Jv = -9 and Jv = -12 it is 9>. The former is indicated: lJv - -7>, -11; the
latter 2/v = -9>, -12.

(4) If one of the JJv gives the direct shift and the other the inverse, the former
defines the limit for approaching voices, the latter for receding voices. For example, in
uniting the conditions of Jv = -2 and Jv = -11 the limiting intervals 2< and 11> are
obtained; this is indicated: 2Jv = -2<, -11>.
§ 189. After deciding to what groups of indices the combined JJv belong, and having
found the limiting distances between the voices, the conditions of using the intervals
must now be expressed in signs; this is easily done (§ 124). Taking from the General
Table of Indices the two series of intervals pertaining to the combined index, next
compare each interval having the tie-sign with the same interval of the other Jv. If the
latter interval also has the tie-sign it is necessary to compare separately the upper and
lower signs of the two intervals, and to retain the sign that shows the greatest degree
of limitation. In the absence of any signs for one of the intervals compared the signs
for the other must be kept. If one of the intervals takes the sign p. this must also be
retained in the final result.
1 ID
Having found the rule for the individual Jv and made the comaprison between the
various intervals of the original and derivative, resulting in the same rule for the com
bined index, next compare one and the same interval as functioning in the two different
indices. First are united the limitations imposed on intervals in simple counterpoint of
the strict style, then those by complex counterpoint.
§ 190. Suppose that the situation unites the conditions of Jv = —7 and Jv = -11.Then
(§188, [3]) the corresponding double index will be indicated: V« = —7>, -11. Similar
motion is allowed at a ^Jv; the limiting interval is 7>. Now pass to the conditions of
using the intervals, and from the General Table of Indices write in successive order
the figures of the intervals applying to Jv = —7. Under them, from the same table, write
the figures applying to Jv = —11, within the limit 7>, so that the figures of both series
coincide in position. Then below the second row a horizontal line, and under this the
sum of the conditions of the combined JJv.
i)1Jv= — 7, — II.
Jv = -- 7: 3 4
—X
C—O -X
Jv = --11: 3 5 6
C— J
—X
p
X C—3 —X
Jv = -- 7*-,--li: 3 4 5 6
—< —X
Ex. 107
'JVr- 7^,-11.
Original P P
6* 3 3
*
• " 'oo
II *
1st Derivative
.ative
•*» Lr -trrrrir j irMfmrrrfi-i

2nd Derivative
JqUjuJ- AAJ u. :»: i

z*^,ri'1jTfJr^rrir r FrJrrcJr^r nc
The next two examples, similar to the preceding, show combinations at two JJv
with the inverse shift:
a)'iJv = " ft — II
C—J
2Jv = 9: 3 8
c—>
c—> —X
1JV = — li: 5 6 10
—x. C— 3 C 3
C—3 —X — —
-Jv = — 9-, — il: 3 5 6 8
—X c—> ^
TWO-VOICE ORIGINAL COMBINATIONS GIVING MORE THAN ONE DERIVATIVE m
Original
ex.108^-9^-11, r
6X
o _ --== i ., r~^~*î i" =—e 3=C
Jvz-9
XII
1st Derivative :tfcfc "
JSes °
iv>9
s y J J -J J J
l^l *
m 331
^
JV--11
X"
2nd Derivative 1 13 9 it*
" _ I _ _ m* P °
1
H
jjP^SBM
P !*=eetT M AA
M 1
3)2Jv--ll, -13
\Jv«-ll: V cLS
SJv«-13: tf
— 5-* -
JL y r
(-)
Ex. 109 »*~U>.-1* £ t A t 5 a
2jv = _n>,_i3
Original
Sa=
*frfcSF J^
2nd Derivative
Jv = -13
;» °pof JV Al
I °" io "=pg ^
r© S9-
,v=-13
In the next three examples one index has the direct shift, the other the inverse shift.
4) 2Jv=2, —7
2Jv = 2: 3
—X A
_x
ijv = -7: 3 4
Lx —X
2Jv = 2, --7>: 3 4
—X (-)
Ex. 110
2Jv = 2, —7>
s jjjjCj Û
Original
s it if M-rrrrrH-fj
ii F *
a* ^...•'J.UJ jjj
ft
J
* J o #
fp ?¥PP W
Jv = 2 . . Tv = 2
» i *
1st Derivative \} J J I ,J , .hnM-J^
5
B
rrr r^r rrr r
Jv = -7 II
h , ip1 rTr r t
2nd Derivative jj i j J J JTj
&BI f r r f?
frfj,4^^j^?j ^ J J i ^
* rn rrr1 r " FrMf rr
TWO-VOICE ORIGINAL COMBINATIONS GIVING MORE THAN ONE DERIVATIVE 119
5) 2Jv=-2,-ll
(_) _ -x _
SJv=-2: _3 §. A A
\Jv=-ll: §« A H
2Jv=_2 <,_!!>: £ t & |, g
The interval 1 is not included in this table, since it lies beyond the limit for ap-
<— 1
proaching voices, 2<.
Ex. Ill
2jv = _2<, -
'^jd^MUS^MkMj
Original
'îffliiV rrpt q\W.W.'r

-©•
^£=
Jvr- 2
|T=-«
.^ J J JL m ga_ g —t.
a=&
1st Derivative ?##
JJ
II
^ w&
XL?
si H^r r fjt r f ^
JVz-11. I IV
x jir
2nd Derivative in |ftft
i ,\ 1 J- t±± i +++±t
w f- r r for
=DP
m ZZZ
6)^ = 3,-8.
—x — —x —x —X
1 Jv = 3: 0 3 5 6 7 8
—X c—3 -X c—3 —X —X
P P
—X c—3 c—j —X — — -X —X
2Jv = — 8: 01235678
—x —x —x e—9 —x c- 3 c—3 —x
P P
—x c—3 c— 3 —x — —x —x
.2JV = 3, — 8^: 0 12 3 5 6 7 8
—X —X —X C—3 —X C—3 C—3 —X
Ex.112 Jv=8'*T: 7X
Jvr 3
JVr , "
1st Derivative
r r err [f
Jvz-8 .
><. ."^ a
2nd Derivative ^^(jff E
v;-8 ^+ fc5^
In the next example both indices have the direct shift. Since the limiting interval
is 2< the preceding intervals 0 and 1 are omitted.
7) 2Jv = -2<, 3
2Jv=-2: ¥ 5
—X
<6-)
1<
—X —X
1Jv= 3: <3> 5
—X
(6) 7
—X
8
—X Eix
Ex. 113 2jv=_2<, 3:
X 5
—X
A 7
—X
A
Original 1st Derivative E-x —X —x
Jv = -2 6
i3£ P2^£ 1 PE 1 P n
?P
d
S ? PB W2 i P P
Jv = 3 IIV = 2
2nd Derivative
I + IIV = 3
IP pg Hrfirr^Jiîr ^ O-
3X
Ofcs. -Stating all these situations in the form of mathematical symbols instead of
words has the incontestible advantage that the study of complex counterpoint is thereby
greatly simplified. The way in which almost insurmountable difficulties can be set up
in the composition of such examples as the last four is well illustrated in Habert's
book,* written according to the system of Simon Sechter.** He lists fifty-six separate
cases of combinations in "double counterpoint," and for each of them detailed rules for
the use of each interval; all this together with the examples occupying 155 pages in a
book of 260 pages, not including the appendixes.
•Habert, "Die Lehre von dem doppelten und.mehrfachen Kontrapunkt," Leipzig, 1899.
••Simon Sechter (1788-1867), theorist and composer, with whom Schubert had arranged to take
lessons in counterpoint shortly before his death. —Tr.
TWO-VOICE ORIGINAL COMBINATIONS GIVING MORE THAN ONE DERIVATIVE 121
Triple Indices
§ 191. The method of indicating the union of two indices has been shown. In the
same manner can also be expressed the conditions of a greater number. To compile a
table for a triple index it is sufficient to take a table for a double index and indicate
the process of adding the third index, for example:
2Jv = — 7, — 9, — n.
C—3 C—J X
2Jv = — 9,— ll: 3 5 6 8 10
X c—3 C—3 C )
Px X
^ = — 7: 3 4
p
C—3 X C— 5 —X
2JV = — 7=-, — 9, — 11: 3 4 5 6
—X c— )
Ex. 114 Original

-W
JVz-7,-
X
1st Derivative -4-Ar
JVr-7
- i—i ' i i ' ' ii rr i i ut f
2nd Derivative -
JVz- 9 tt"T»rMr rnr nr rr'r rcr'r %fr IE
3rd Derivative jg^^ J-J,j^|J3j^JjJ-7Jjj^jj^

JVz-ll EB 1 k- Mr rr'r- r'f ffrff XT"
§ 192. Another way to unite the conditions of three indices is as follows: Write the
table of conditions for each index, and comparing separately the upper and lower signs
for the same interval in the three rows, leave to the final result the signs that show the
highest degree of limitations. For example:
2Jv = i, 3. 5-
c—> C—3 X
t
—X — i>
-X
p
X
—X X
Jv=l: 0 1 2 —3 0 6 7 8 10
X -X X C—3 C—3 —X -X —
X _ _ X —X "*
Jv=3: 0 8 i) 6 7 8
C—3 —X c—y X
P
Jv = 5: 1 5 6 8
c— > —X C —3 C - 5
c—> C—3 -x P
- X
V
X -X
P
—X —X
2Jv=i, 3, 5: 0 1 2 8 5 6 7 8 10
Original
Ex.115 Jvrl,3,6. p P 2,
I 6X y .6 3
U. A P
5^
4 Si$
I], _Q_
^£ w 3d;
jggl §?? m 2
^ e:
jYrl n *x —x
1st Derivative ; ; II v:l
■+ir. bî^ rir rr w IT
JVz3 ;
2ndd Derivative
Derivative ; HV: 3
F* v^
±2 TTi-j
JVrB
3rd Derivative ; J£5
#-•
wr=. '^.»-nrrrrr ^
To the conditions of a triple index may be added those of a fourth index, or a fifth,
etc. But added constraint results as the number of indices is increased, and writing
becdmes more and more difficult.
The Polymorphous Index

§ 193. Here is encountered a problem that used to be of considerable interest to the
old theorists—one even shrouded in secrecy—that of writing a two-voice combination that
would yield all shifts without exception. This concerns polymorphous counterpoint, and
the matter can easily be settled by applying the principles already learned. Uniting the
conditions of all indices gives this table:
(n> (r> W (7> (7> (?> (T> W (o>

A A A A A A A A A
Obviously, to write counterpoint under such conditions is impossible, for there is not
a single interval that can be placed on an accented beat. Therefore the question as to
the possibility of polymorphous counterpoint in the strict style must be decided
in the negative.
CHAPTER X
THE DUPLICATION OF IMPERFECT CONSONANCES

IN TWO-VOICE COMBINATIONS
§ 194. In concluding the treatment of fixed consonances in Chapter VI reference was

made to a special characteristic of indices corresponding to imperfect consonances (in
dices of the second pair of columns: JJv = ±2, ±5, -9), consisting in the fact that they
admit of the simultaneous execution of the original plus a shift of one of the voices at
the given index, thus obtaining three-voice counterpoint in which two voices move in
parallel imperfect consonances (§ 119). It remains to examine to what extent all tied
intervals—dissonances and variable consonances—may be applied to duplicated voices.
First of all will be studied the duplication of one voice (§§ 195-202 and §§ 209-12,
three-voice counterpoint), then the simultaneous duplication of both voices (§§ 213-27,
four-voice counterpoint).
I (A) The Duplication of One Voice Using Tied Intervals;

Two Derivatives; Two-Voice and Three-Voice
§ 195. In three-voice counterpoint resulting from the duplication of one of the voices
in imperfect consonances (§110) three combinations are to be distinguished: (a) a two-
voice original; (b) its derivative, also two-voice, consisting of a shift at an index cor
responding to an imperfect consonance; and (c) a three-voice derivative, the result of
the simultaneous execution of the original and a shift of one of the voices. As long as
the accented beats are limited to certain fixed consonances the correctness of these
two-voice combinations—original and derivative— guarantees the correctness also of the
three-voice combinations obtained by executing them simultaneously. But when tied
intervals are used they are changed by duplication. In this case, though each two-voice
combination might be individually valid, showing correct two-voice counterpoint, this is
insufficient in order that their simultaneous combination shall be equally correct. To
secure this result it is necessary to be sure that the dissonance of a ninth (8 or —8) is
not found between the upper and middle voices. According to a rule of simple counter
point of the strict style, 8 may form a dissonance only with the lowest voice; as a sus
pension it constitutes the only superfluous limitation that multi-voice counterpoint
offers as compared to two-voice (§ 83). For example, this original at Jv = —9:
/ /i * "\" o\d *H gives this derivative:^ |»op| p",, 1 g g but to execute
II 8 6 I
II 8
them simultaneously is impracticable, for the forbidden dissonance 8 appears, involving
8
the middle voice:
impossible
To prevent the possibility of 8 appearing between the upper and middle voices is the
object of the rules presently to be given.
123
§ 196. Before passing to the statement of these rules, those cases will be considered
in which the ninth will not be found at all between the upper and middle voices, con
sequently offering no obstacle for execution in three voices.
(1) When the upper and middle voices progress in parallel motion 8 obviously cannot
appear, therefore such a three-voice combination is always correct. This refers to four
cases of duplication:
Ic 2 » ■,d=-2 9 d=5 f ,d=-5<
* •
Indicating each voice in detail gives the following expressions:

2 + I + II, I + 1^ ~2 + II, Iv~ 5+ I + II, I + 1^ ~5 + II.
r
All the examples in Appendix A at these indices can be executed in three voices.
Obs.—Since combinations at these indices always give correct three-voice counter
point when one voice is duplicated, the problem can be solved in reverse order, i.e.
start with a three-voice combination, writing it in simple counterpoint with the highest
voice duplicated in thirds or sixths. From this three-voice combination extract two of
two voices, using one as the original and the other as the derivative.
(2) Every combination at Jv= —5> can also be freely executed in three voices, since
5> shows the limiting interval for receding voices, whereby 8 is automatically eliminated.
Such correct duplications are Id= _5> + II and I + IId~ ~5>.
Ex. 13 in Appendix A (p. 312) can therefore be executed in three voices, as follows:
Pw^5 :um
i
w iei j'liV-TTrrrf W-m\U
n rr p etc,
*4—e± 9 a
^&
r- :-5
§ 197. Ending the duplications of which the correctness is provided for by the cor
rectness of each of the two components, the next cases of duplication are those where
the correctness of the two-voice combinations does not guarantee the absence of mis
takes in the three-voice combination. The error, as stated before, is the appearance of 8
between the upper and middle voices, the prevention of which is the purpose of these
rules. The two voices concerned may belong either to the original or to the two-voice
derivative. In the first instance the rule refers to the original ninth, in the second to
that m from which is obtained the derivative ninth. Hence the duplications about to be
examined fall into two groups.
§ 198. 1st Group of Duplications.—Here the upper and middle voices form the original
combination, the lower voice duplicates one or the other. This group includes (a) the
direct shift of voice II at a positive Jv, and (b) the inverse shift of voice I. Aside from
the cases of duplication mentioned in § 196 where d- 5>, the remaining ones are: (a) the
direct shifts II^ 2, IId= 5; (b) the inverse shifts Id= -*, Id=-12. This group of dupli
cations can be illustrated diagramatically:
(a) Direct Shifts (b) Inverse Shifts
I Original * I Original
II 1 Combination II | Combination
| jV^-a
a=2 at a = 5- a = — 9 at a = — 12.
DUPLICATION OF IMPERFECT CONSONANCES IN TWO-VOICE COMBINATIONS 125
The arrow goes from a voice of the original to the duplication of the same voice in
the derivative.
Since in these diagrams the original combinations are formed by the upper and middle
voices, in them the resolution of the ninth to the octave must be forbidden, consequently
*B*
8 appears as a supplementary condition for the Jv of the original combination. It is
understood, of course, that the lower sign belonging to the ninth of this Jv is retained,
therefore in situations such as II ' and Id , where this sign is (—), the ninth must
be excluded altogether from the available suspensions
Based on this, Ex. 23 at Jv = -9 in Appendix_A (p. 320) can be executed at I,d=-9
only as far as the seventh measure, in which 8 is found, making an incorrect three
voice combination.
■ I I , ■ i I , , o.impossic
8 impossible
i .k% M &-.
- *—TLyqê; etc.
f?fQf - rr'r TTpFff
1 ^ HE a? -EL
¥
&—-P-
P ^m ***f *--& etc.
A similar obstacle is encountered in the eleventh measure of No. 29 (A>= —12),

Appendix A.
§ 199. 2nd Group of Duplication.—Here the upper and middle voices are those of a
two-voice combination. This group includes both the direct and inverse shifts of voice II
at a negative Jv. The duplications referring to this are: (a) with the direct shift, II ,
II -S . (Cf. what was said about Jv= -5> in § 196). Expressed in diagrams:
(a) Direct Shifts (b) Inverse Shifts
♦ II
Original Original
combination combination
II
1 II . II
a = —2 or a = —5 o = -9or a = -12
Here the upper and middle voices appear as the voices of the derivative, therefore
to prevent the appearance between them of a ninth resolving to an octave the interval in
the original (a) from which the ninth was obtained must be reversed. To find m giving
n- 8 (or with the inverse shift n= -8), add the value of the index, taken with the opposite
sign, to the interval of the derivative (n) (§ 61).
For an original interval found in this way the sign (—) shows the conditions that must
be added to those of the given index. Furthermore:
(1) If the derivative ninth is positive (i.e. if Jv has the direct shift, (a) in the pre
ceding diagram), the sign (—) must be placed above: m.
(2) If the derivative ninth is negative (i.e. if Jv has the inverse shift, (b) in the pre
ceding diagram), the sign is placed below: m..
The lower sign of m with the direct shift and the upper with the inverse
remains unchanged.
§ 200. The original interval from which is obtained a derivative ninth, together with
the changes in its tie-signs, will be given for each index belonging to the second group
of duplications.
(1) For indices with the direct shift n- 8 is obtained: at Jv = —2 from m- 10; at/v = — 5<
from m- 13. For these intervals the sign (—) above is obligatory, cf. § 199, (a). Since
at Jv = —5< the sign is already found below at m- 13 this interval is excluded from the
available suspensions
<8>
(2) For indices with the inverse shift n- -8 is obtained: at Jv= -9 from m= 1; at
Jv= _12 from m- 4. When these intervals are duplicated they take the sign - below,
cf. § 199, (b). Since the second, according to the rules of simple counterpoint, has
already the sign (-) above (1 ), at Jv = -9 it is excluded as a suspension:^ .
§ 201. In the following table for the first group of duplications (§ 198) is indicated
the conditions under which as original ninth (m= 8) can be used; in the table for the
second group (§ 199), the conditions using that m which gives a derivative ninth (§ 200).
-12 /.
-a
:8») (8>
Obs. -The duplications Id = 2, Id = "2, Id = 5, Id = -5<, Id=-5>, IId = -5>, do not

require any change in the conditions of their corresponding indices (§ 196).
§ 202. In order for a combination at one of the upper indices of this pair of columns
(at Jv = —9 or at Jv = —12) to get a reciprocal duplication both above and below a voice,
it is necessary, in the conditions of either index, to insert the additional changes re
quired by one or the other of the duplications. For example, if in the table of intervals
for Jv- —9 are inserted 8 and l' this combination will give both duplication: 1^ + II
andI+IIv=-9 ~ C-,
Uniting the Conditions of Duplications with

the Conditions of Every Jv
§ 203. By the methods given (Chapter IX) the conditions of any index may be united
with the conditions of duplication at any imperfect consonance. The absence of similar
motion in the original is a necessary requirement, since the indices corresponding to
imperfect consonances are all those of the second group ( ] Jv, § 59). If a duplication
is required of an index of the second group, if, above all, it has few fixed consonances
and is therefore hemmed in by many limitations, then the number of changes inserted into
the conditions using the intervals may be very small. Probably the greatest advantage
of this is that the limitations belonging to the final process in a duplication are already
present in the initial conditions. This explains the fact, strange at first glance, that a
combination written at one of the difficult indices, Jv = ±1, ±6 etc., sometimes admits
of a duplication of either voice at any imperfect consonance.
•In the first group of duplications the upper and middle voices form the original combination.
DUPLICATION OF IMPERFECT CONSONANCES IN TWO-VOICE COMBINATIONS l27
§ 204. It will be found that Ex. 1 at Jv=—l in Appendix A, from beginning to end, may be
executed in three voices at the duplications Id_ ^ or Hd= 2 The beginning of the second
of these duplications is herewith quoted, as it presents more inconvenience as to the
augmented fourth than the first.
Jv=-i Eg
Ot.
g f ffrrir etc.
ipP* U (b)
What is the reason for such coincidence? Why does a two-voice combination, written
fe£ etc.
with no thought of parallel thirds, nevertheless admits of them? The answer to these
questions is obtained if the conditions of the duplication at h>= 2 (II ) are united to
those of Ju= —1, and their differences made clear. Add 8, the condition of the duplication
II d=2 to the intervals of Jv=2 (§ 201), comparing them with the intervals of /i>=—1.
C— 3
Jv = 2
—x -8 >:
II"
Uniting these and X c-3 I'

c—>
other conditions: 0 •2 8
C--3 c—3
Comparison of the third row of figures with the first shows that few changes had to be
added to Jv= -1 as the result of annexing to it the conditions duplicating at Jv= 2. The
absence of ties below the fifth (4) and above the ninth (8) is sufficient in order that a
combination at Jv= -1 shall admit of duplication in thirds: Hd~ 2. The lack of these ties
in the previous illustration was what made the duplication practicable.
§ 205. Next will be shown to what extent the conditions of Jv= -13 must be supple
mented in order to obtain a duplication at Jv= -9, namely: Id= """ and Hd= ~^. According
to the table in § 201, to the conditions of fv= -9 must be addedând^
r—•> c—j
c-o "-X C 3
-Jv — 13: 0
X
5
—M
6 7
- X
8
—X
d = -9. C—J <—3 C —3
;jv <• = — ». 1 3 8
II C—> C— 3
C—3 C—> C—3 C—3 c—3

0 1 3 5 6 7 8
-X X C—3 —x C--3
Thus, 13 8 are the changes that have to be incorporated into the conditions of
Jv= —13 in order for a combination written at this index to admit of the duplication
Id= ."' and IId ', In Ex. 31 at Jv= -13, Appendix A, the first three measures may be
executed in three voices by duplication at the tenth, i.e. II d=-9. but the second used
as a suspension in the fourth measure, 1, prevents its continuance in three voices (the
illustration is transposed an octave lower):
iSL
etc.
~o~
impossible
t^=^£1
~i Aim UJ4 etc.
f rs: P
The derivative of this example (v. Ex. 32, p. 329) allows the duplication [™ * up
to the eighth measure, and II 4 up to the fifth measure. The example at Jv= —6
(Ex. 15, p. 314) allows the duplication I * up to the sixth measure.
§ 206. It is a very simple matter to unite the conditions of duplication at Jv= —9 with
Jv= —7, i.e. with double counterpoint at the octave. Since the limiting interval is 7>, 8
is excluded from the table of duplication at Jv= —9.
Jv = -7: 3— X
rid=-9:
2Jv = -9 \lld=-9:
s
%
f —X
4 2Jv, 7>
&*<&)
To the conditions o( Jv= — 7> is added: JL 3x and the absence of similar motion.
§ 207. Because of the greater practical importance of double counterpoint at the
twelfth, a table will also be given uniting its conditions with those of a duplication
at A; =-9.
—X
Jv 11 —X
r ld=-9
2JV = - : (i) t X
—X
? —X A x ^Jv, 9>
It will have been observed from the last two tables that few changes were needed
in the conditions of Jv= -7 or at Jv= -11 to enable them to take duplications at the
tenth. Likewise, similar motion is forbidden, as it alters the conditions of these indices
to such an extent that a combination written at Jv= -7 or at Jv= -11 is seldom found
that would allow of imperfect consonances; at every point similar motion is frustrated.
§ 208. AH possible cases of duplications may be united not only with any of the
indices but among themselves in any permutations desired. For instance, the conditions
of two-voice counterpoint can be deduced that will admit of shifts at two indices and a
three-voice presentation besides; the rule of counterpoint can be found that will satisfy
the conditions for duplicating various imperfect consonances-the sixth, tenth etc. The
number of permutations may be so increased that nearly every three-voice combination
will admit an interchange of voices and the transposition at the octave of one voice or
another. To write counterpoint satisfying such requirements is less difficult with the
use of fixed consonances; more difficult with dissonances and variable consonances
from which tied intervals are excluded.
(B) The Duplication of One Voice Using Tied Intervals;

One Derivative, Three-Voice
§ 209. In the preceding sections the combinations were discussed that have two
derivatives: (a) in two voices at a given index; and (b) in three voices, resulting from
the duplication of one of the voices in imperfect consonances. The present discussion
will be limited to three-voice derivatives.
The necessity for avoiding 8 (or —8) in relation to the upper and middle voices still
applies (§§ 195-202). The three-voice derivatives now considered differ from the pre
ceding cases only in the use of derivative fourths between the upper and middle voices.
Appearing before in two-voice derivatives, fourths were subject to the limitations of
dissonances; they are now released (3 and 10 instead of 3 and 10), and may be used
according to the rules for perfect consonances (§ 83). An original consonance giving a
derivative fourth, now consonant, is released from its former limitations and become a
fixed consonance. Therefore the original dissonance, of which this consonance is the
resolution, is released from the supplementary sign x. Such are the situations that allow
consonant fourths to be added to the conditions of a given index. In a three-voice com
bination resulting from the duplication of imperfect consonances, successive first in
versions may be freely used: f\ | | | [ I , ,
r i
and therefore similar motion may be added to combinations at JJv.
Obs.— It should be understood that this releasing of fourths from their limitations is
in no sense a requirement; that also in counterpoint of three or more voices the pre
paration and resolution of a dissonant fourth according to the rules of two-voice counter
point is still valid. But, having obtained the status of a consonant interval, the fourth
need not be prepared and may be used freely.
§ 210. In duplicating one of the two voices of the original combination at imperfect
PS
4
consonances, a consonant fourth is found only in first inversions: it r»
?? L8 o
—
LS
since any attempt to include it in a triad in fundamental form would result in the omission
of the third, with no imperfect consonance present: A " — > a structure that will
admit of no duplications.
§ 211. The results of legitimate duplications may be represented by this chord:
. Here only the third or the sixth above the lowest voicemay be duplicated.
A chord with such an arrangement of intervals admits of the following:
(1) At IId= ~2 from m= 5 (n= 5-2=3, § 31).
ii
(2) At IId = -5 from m = 2 (n = 2 -5 = -3).
3£«±t
m
II
In conformity with this, at II , 5 is released from its limitations, and at II
2 Is released. Since each of these consonances reverts to a fixed consonance, the dis
sonances that resolve to them are released from the supplementary sign x. At II ^ the
seventh is released from this sign, as the interval resolves to 5 C6 instead of 6 ), and
at II the fourth reverts from 3 to 3.
§ 212. It would also be possible to examine the conditions presented by two other
*►
o s*
aspects of first inversions: M_
rt -,those found where compound consonances
1
fm
^V n -fi ,
c °
are used: II,d=- ind II d=-12 . But this is unnecessary, because all instances of where
fourths are released from their limitations are included in the following rule:
A derivative fourth changes to a consonant interval whenever the lower voice of the
original combination is duplicated at the upper third or upper sixth, or their compounds.
A duplication at the third is released from the limitations of the original sixth; at the
sixth from those of the original third.
Both these intervals revert to fixed consonances, whereby the dissonances of which
they are the resolutions are released from the supplementary sign above. It is such
changes that must be inserted into the conditions of a given index. In the following
example, where the duplication is II d=-2.
Ex. 116 Josquin de Pres

-I
1atea: «-© 3E
ICC
^m
xxzzztzna.
-#«-
?T~n .. JlSi^jg 1 » 8 "J» H 8 W S EL^ ^S

w nfc-2
II
the original combination shows a two-voice canonic imitation in the outer parts. In
measure 6, where it is interrupted, is found an augmented second (between the lower
voices), the use of which is uncommon in the strict style. It is an example of the free
way in which dissonances were treated in the works of the Flemish composers who
came before Palestrina.
II The Simultaneous Duplication of Two Voices;

One Derivative, Four-Voice
§ 213. In concluding Division A is presented a study of the conditions under which

a two-voice original will yield a four-voice derivative, involving therefore the simul
taneous duplication at imperfect consonances of both voices. First will be considered
the duplications in which suspensions are excluded from both the original and the deriv
ative, later the duplications that use them.
(A) Duplications of Two Voices Without Suspensions
§ 214. When one of the voices is duplicated at the tenth (d- -9), from each original
consonance is obtained a consonant three-voice harmony (§ 119). But when both voices
are duplicated (I 9 + Hd *) not all the consonances will yield a consonant four-
voice harmony. One of them, the sixth, gives the dissonance of a seventh (13) in the
outer voices:
v = -9
Pippf m
I
vtr=y
Therefore, if the sixth is excluded from the fixed consonances and the others from 0
to 9 are used, and counterpoint without suspensions written in conformity with the
conditions of Jv= —9, such counterpoint will admit of the simultaneous duplication
of both voices:
Ex. 117
(l*-* +Ui:a)
m mm
Original
m*— .,
I
^
s^sp
IS wm
o-
LM tMi =U £ i
W f^rrm ^ -OOi
II
3S
SE m ^^ m £
§ 215. Returning to the dissonant harmony, it results from using the sixth. The
outer voices form a compound seventh, 13. How does this dissonance originate? There
are four voices, of which two constitute ihe original combination I + II. Every other union
of voice I with voice II is a derivative; there are three of them:
(a) (Iv = -9MI)/U=-9;
(b) (I + IIv = -9)/V= -9;
(c) (Iv = -VlIV, ')/«=-18.
The first two combinations have Jv= —9, the third has Jv= -18, i.e. double counterpoint
at the twelfth. The index of the third combination is always equal to the sum of the
other two indices. The original combination must, therefore, satisfy the condition of the
double index Jv= — 9 >, —18 (§ 190). All the consonances of this index except 5 are fixed.
Clearly, then, the dissonant interval 13 in the outer voices is the result of the shift
m- 5 at Jv= —18, since 5 -18= —13.
§ 216. Indicating the imperfect consonances that duplicate voice I by the letter a,
and those that duplicate voice II by b, gives the following list of derivative combinations
and their indices:
(a) (Iv = a + II)/v=a;
(b) (I + IIv = b)/t>=b;
(c) (Iv = a + IIv = b)/t>=a + b.
Hence it appears that if vv of voices I and II are unequal, all three indices are different
in value, therefore the original I + II must satisfy the conditions of the triple index
,/t>= a, b, a + b.
Furthermore, the supplementary conditions must be inserted into the conditions of
this compound index, in view of the fact that a two-voice derivative is not independent
two-voice counterpoint, but is only a constituent part of a multi-voice texture.
§ 217. In Ex. 118, a duplication at Id~ *• + IId~ ^, one of the consonances, the fifth,
*6^
gives dissonant harmony at the duplication: ' o' * The other consonances
ii ** ir *
give consonant harmony:
§ s
IXX
312
tfi» " iS CS C2"
Excluding the fifth from the fixed consonances and using the rest, it is possible, by
observing the limitations of ^JJv, to write a two-voice combination that admits the
above-mentioned conditions. The derivatives are:
(a) dv: -5 + II) Jv-= 5;
(b) d + IIV = 2)Jv-= 2;

(c) (IV == 5 + iiv = 2; Jv
The index of the third combination, Jv= 7, is rejected, since a derivative at this index
would be nothing but the separation at an octave of the original (§ 47). The latter will
have, therefore, the double index Jv= 5, 2. Since both indices are positive a limiting
interval is not needed (§ 42).
§ 218. In a table for these indices must be inserted the following supplementary
conditions, inasmuch as the two-voice combination to which the indices apply is not
independent two-voice counterpoint, but enters into the structure of four-voice
counterpoint.
When a derivative fourth (±3 or ±10) between the upper and middle voices or between
the two middle voices is obtained from an original dissonance, it is released from the
limitations acquired and becomes fixed. It is therefore allowable to include 5 among the
fixed consonances of the double index Jv= 5, 2, listed in the preceding section, as it
yields a consonant fourth.
DUPLICATION OF IMPERFECT CONSONANCES IN TWO-VOICE COMBINATIONS
Ex. 118 133
1 ?E£
t Jv = 5.2
^ S
10
£
10
i JJ: ±=M £ M:
IS II ~n-
r5 r r r r r-rr=P^
* ^^ ?^ *=
ss f r r r r r be r r g r r r rr
§ 219. The only chord resulting from the duplication of two voices in imperfect
consonances and that also contains a consonant fourth is the first inversion with doubled
third. It occurs in the following forms:
EE ii, ru,i gL
\>o' *=* ^ £ etc.
Since two first inversions with doubled third must not show parallel octaves—as
would happen if all the voices moved parallel— successive parallel fourths, possible in
three-voice counterpoint, are not allowed in four-voice. This is what distinguishes four-
voice duplications from three-voice; therefore, at the simultaneous duplication of two
voices the limitations of JJv remain in force, and similar motion is impracticable.
§ 220. Since parallel fourths can not be used in the three upper voices (§ 219) these
voices may appear in other octaves, i.e. may be shifted in double counterpoint at the
octave (it must not be forgotten that this refers only to those cases of duplication where
the accented beats take consonant harmony). Such shifts can be freely made as long as
the lower voice remains the lower, i e. as long as a combination of the lower with any
of the upper voices forms the direct shift. Otherwise, at the inverse shift of the bass and
one of the other voices, the free use of a fifth would result in an unprepared fourth.
As long as the bass continues to remain the lowest voice, shifting the other voices
at the octave can be freely done if an occasional crossing of the parts is not considered
a disadvantage. The use of these other voices as bass would be possible only if they
did not imply second inversions. For instance, the alto in Ex. 118 (voice I of the original
combination) could not function as bass, since the notes indicated (***) are unprepared
chord-fifths, which, if appearing in the lower voice, would give unprepared fourths:
I
*E5E XE
wm m
The same applies to the tenor (voice II of the original)— it cannot appear in the role of
bass since this is prevented by the triads in the second and third measures. On the
contrary, no such obstacle is encountered in the upper voice, so it may function as bass.
Therefore the following shifts are possible for Ex. 118:
(1) Bass remaining as lowest voice:
b)
^^
d)
i.Ji^J ^J
(2) The upper voice of the original functioning as bass:
a) b.
iÔi i
f i
rmmf etc.
rr r rr r-?
ast ■ i- J J J J J J-^L
y^ v=$
^¥ f
(B) Duplication of Two Voices Using Suspensions
§ 221. To find the conditions under which both voices of the original may be dupli
cated simultaneously it is necessary (1) to compile a table of compound indices that
unites the conditions of three derivatives, and (2) to insert in this table the supplementary
conditions by which the use of fourths and ninths in multi-voice counterpoint differs from
their use in two-voice. The method of compiling tables of compound indices was given
in Chapter IX. Fourths and ninths require further explanation.
§ 222. The appearance of 8 (or —8) should be prevented in all those two-voice com
binations that do not include the bass voice. If the original 0 + II) is such a combination
8 will appear as a supplementary condition in the table of compound indices. As to the
derivative combinations in which the bass is not concerned, it is necessary to find what
equals m, given n- 8 if the shift is direct, or n- -8 if the shift is inverse. In the former
instance the supplementary conditions show at, in the latter, 'm' (§§ 199-201). This
supplementary condition is also inserted in the table of compound indices.
§ 223. A consonant fourth (±3 or ±10) may occur only in those two-voice combinations
that do not include the bass voice. First of all, from such combinations can be eliminated
any original that forms independent two-voice counterpoint, for such a combination can
not contain unprepared fourths. If in a combination where the bass voice does not par
ticipate n equals ±3 (or ±10), derived from m, and if at the same time m is a fixed con
sonance in each of the other derivatives, then m is released from the limitations imposed
on it by fourths. The dissonance resolving to m loses the supplementary sign x, owing
to the fact that m becomes a fixed consonance (cf. § 209). If one of the indices is/t>= 0 a
derivative fourth of this combination is not released from its limitations, otherwise an
original fourth would have to be regarded as consonant, an impossibility inasmuch as the
original is a two-voice combination.
§ 224. Because of the absence of parallel fourths (§ 219) and of 8 between the three
upper voices (in conformity with the rules of simple counterpoint), these three voices can
be transferred to different octaves on condition that none of them crosses the bass. In
the absence of suspensions, no obstacle is encountered in such shifting of the three
upper voices (§ 220). But if suspensions are used these shifts are governed by the
following rule: A voice forming 8 (or —8) in relation to the bass cannot approach the
latter any closer, since doing so would contract the ninth to a second— resolving to a
unison— a progression forbidden in the strict style. Aside from this, all shifts of the three
upper voices in double counterpoint at the octave are possible.
The conditions at which any of the three upper voices— suspensions being absent—
could function as lower voices were explained in § 220. But if suspensions are applied
another condition must be added to those already in force: A voice that contains a ninth
resolving to an octave cannot appear as lowest voice.
§ 225. A table will be compiled of the conditions for Id= ~9 + IId= ~9. Here the
original combination is for the alto and tenor, the soprano and bass take the duplications.
S. „v= -9
"a. I
Original
T.
B. I jv= -9
Corresponding to this diagram are these duplications:
dr=-*)
(r-9)
The original combination I + II (A.+ T.) yields the following derivatives:

(a) (Iv = _9+ II) /w= -9B. +T.;
(b)-(I +IIv = -9)/t;= -9: A. + S.J
(c) aV = -9+ IIV = _9)^=-18: B. + S.
Since (a) and (b) have the same index, Jv= -9, that of the original is not triple but
double: Jv= -9>, -18, equivalent to Jv= -9>, -11 (cf. table in § 190).
Next is to determine the changes that must be inserted into the table for this compound
index. There are two combinations in which the bass does not participate: the original
(A. + T.) and derivative (b): (A. + S.)/t> = — 9. To prevent the appearance of a ninth re
solving to an octave, the former required the supplementary condition 8 ; the latter 1
(§ 222). As for consonant fourths, they cannot appear at all in the original, only in a
derivative, and then only when the lowest voice is not involved (§ 223). But since com
bination (b), alone satisfying this condition, represents a shift at }v= —9, wherein m,
giving a fourth, is dissonant (m= -3 + 9= 6, § 61), a derivative fourth can not appear as
a consonance, and therefore must be inserted as a change in the table for this compound
index. Writing in the supplementary conditions in the table for the double index Jv= —9,
—11, the conditions are obtained for the above-mentioned duplications: I ' + II" "
C3' V 6 8 (v. § 190, [2])

^=-9,-11:
— — X C— ) c— >
1 C—3
Supplementary conditions: I 8
< -O
Conditions for I + II 13 5 6 8
C—5 — —X C—3 C—3
These conditions are met in the following combination:
Ex. 119
ld=-9+nd = -9
II v = -9
-p-
m
Original [
Derivative
LJ
I
3C
u m .or J=
II w r^
,v = -9
B
gj^S m
m
±t^J2A A J X*.
W ^PP w f -©-
m -9-
§ 226. Table for the duplications Id~ -9 + Hd- ~2.
rS. r I
. IIV= -2
A.
Original
-T. L ii
B. jv= -9
I Ciiv~-2)
corresponding to: O
\
m. \
Original: I +11, S. + T.
Derivatives:
(a) (lv = -9+II)/t, = -9, B. + T.;
(b) a + Hv = -2)/f=-2. S. + A.;
(c) (Iv = -9+IIv = -2)/v=-ll, B. + A.
Triple index for the original: (I +11) ^Jv= — 2<, — 9>, —11. Conditions not involving
the bass, original: (S. + T.) and derivative (b): (S.+ A.).
Supplementary conditions: (1) the ninth, resolving to the octave (§ 222), excluded from
the original (hence 8) and from derivative (b) (Jv= —2, supplementary condition 10).
Because of the limiting interval 9>the latter need not appear in the table.
(2) The fourth can not be used as a consonant interval. The only derivative in which
it could be found is (b) (S. + A.) and here it is obtained from a dissonance (m= 3 —2= 1).
It is therefore not inserted as a change in the table.
C— 3 C— 3 -X —
2JV = — 2-«,-9>,- n: 3 5 6 8
Supplementary condition: H
Conditions for I +11 =~ 3 5 6 8

—X c > c >
XE3
I
^m m
g FF m ^fefeg
f¥ r-
-*-in
^feffi
ft *
£
feti
fF
^t f i
IE
# XJ *
3
IfejbCfl IE
a ^g 5E
p. fPTt^T
^^_
^ eqE£ ^^ BEEi
Ots.-Comparing the preceding table with that in § 225 (Id= -9 + IId = _9) it will be
seen that they are identical except that the table last considered has the limiting interval
2<, absent in the table in § 225. The similarity is explained by the fact that both tables
have the indices Jv= — 9 and Jv= —11 and both the supplementary condition 8. In the
latter table the addition of Jv= -2 to these JJv inserted no changes except the limiting
interval 2c Ex. 120, in which the voices of the original do not approach closer than a third, also
satisfies the conditions of I'd=-9 + nd=-9.
Jd--9+ Jjd=-9
Ex. 121 ir
£m. -»-, *-J J J J J £ ±
S
rA
m p5
i Wr?
B 3^
lA A Uli s Ji
zed
gp?I ^
TV:-9 «
febi ^ ^y= jû

^
jii-j i j JaU-i ii
giggi f= i 33C i mc
ee
j4^ m
Ssê;i i£
*
m f 3 I
ffi
■^ —;:
i #:
f^
o#
f p
St
§ 227. Table for the duplications Id~2 + IId —2.
, ,
Original
r A. 1
L I
T. 4 II
a:
B
corresponding to: ^ EC
if
~~& ^S^
Original: I + II, A. f B.;

Derivatives:
(a) (Iv = 2 + H)/t;=2f S. ♦ B.;
(b) a+ Uv = -2)Jv=-2, A. + T.;
(c) (Iv = 2 + IIV = ~2) /t; = 0, S. + T.
Double index: 2/t;= 2, -2<.
Combinations in which the bass does not participate: derivatives (b) (A. +T.)and
(c) (S. + T.)-
Supplementary conditions: (1) in accordance with § 222 the ninth resolving to the
octave (8, —8) is excluded from derivative (b) (Jv = -2, hence the supplementary condition
10) and from derivative (c) (Jv= 0, supplementary condition 8).
(2) In combination (b) the fourth can be consonant. It is obtained from an original
sixth (m= 3 + 2= 5), which appears as a fixed consonance at the other two indices
(Jv= —2, Jv- 0), i.e. it enters into the structure of consonant harmony:
I (Iv
tv = -2
)
Z3S.£
II C« (IIV = -Z)
Therefore 5 is a fixed consonance in combination (b). As regards combination (c),

the fourth in it is dissonant. This is a combination at fv- 0, consequently a fourth in it
would be derived from a fourth in the original, an impossibility, because in a two-voice
original a fourth can not be consonant (§ 223).
The supplementary conditions 8, 10, and the use of the fixed consonance 5 are there
fore obtained, all of which are inserted in the table for the double index: Jv- 2, — 2<.
C—3 —< C 3
-,JJv = 2, — 2-=: 3 4 5 6 8 10
-X c--o —X e—> e—i> X
C—3 C 3
Supplementary conditions: 8 10
P
c—> C 3
Conditions for I (1 = 2 IId 3 4
- X
6
G—3
8 10
X c—> C—3 C—3 X
Consonance 5 is absent from the lower row, since it is fixed. Also, because the limiting
interval is 2<, interval 1, which is outside of this limit, is not included in the table.
Ex. 122
ea J i J, -± k+
I 1bP
ZEE
u w
JP mil (^=-^=t
— Y
4"
I
I
^
ft£
IPS if m =s^ fU»—iL
i m
P^^
^
i ui i * £=fl
P» vw a 4 V> ' '
F*
&:
jtsft » 4+ wp £W s=
Ots.-What was said in §§203-8 about uniting the conditions of duplications with
the conditions of every index is valid also for the simultaneous duplication of two voices,
and for all cases of duplication in general.
§ 228. With the simultaneous duplication of two voices in imperfect consonances is
completed the study of two-voice vertical-shifting counterpoint. The remainder of Division
A consists of additional matter referring to questions previously brought up.
The Six-Four Chord in the Strict Style
§ 229. In connection with the use of fourths there arises the question as to the
application in the strict style of triads in the second inversion. Although this is within
the province of simple counterpoint it is necessary to refer to it here because of the
frequent appearance of opinions that are unintelligible in view of certain facts. Such, for
instance, is the common assertion that six-four chords are forbidden in the strict style,
whereas in reality they are often found. In order to define the conditions under which
six-four chords are in keeping with the character of the strict style, they are brought
under the following general rule, which refers to all dissonant chords:
Every union of sounds, dissonant intervals included, by whatever term it is known
in the study of harmony, may be found in the strict style under circumstances in which
the dissonance may be analysed according to the rules of two-voice counterpoint as a
suspension, a passing-note, or an auxiliary-note. This is based on the proposition that
every combination in the strict style is correct if each voice forms correct two-voice
counterpoint with every other voice, taking into consideration the changes in the rules
of two-voice counterpoint when this passes into multi-voice counterpoint. This involves,
as previously observed, greater freedom as to hidden consecutives (§74, [1]) and the use
of fourths as true consonances between the upper and middle voices or both middle
voices (§ 83).
Take, for example, this "seventh-chord Here there is a dissonant
-*
interval between the outer voices: 13; this dissonance must be both prepared and resolved
correctly; besides which the free note may progress to another note at the moment
of resolution:
The remaining voices do not participate in the dissonance but form consonant intervals
with each other, therefore they are free to move either stepwise or by leap, on condition
that they do not form dissonances at the moment the dissonance in the outer voices is
resolved. Hence is obtained a large number of "resolutions" of this chord:
*0 is beyond the limiting interval 2<. therefore at this place the alto and tenor cross.
§ 230. Six-four chords are also found in the preceding illustrations (at the places
indicated*); they result from the movement of the voices, according to the general
principle mentioned. One dissonance, 3, is a component of the six-four chord, and if
this dissonance is prepared and resolved according to the rules of two-voice counterpoint
of the strict style, then the six-four chord that results is correct. But here a reservation
must be made: it has been seen that in two-voice counterpoint the resolution of fourths
(3 and 3) to fifths sounds very empty. This is because the inherent meagerness of the
dissonant fourth is further accentuated by its association with an empty-sounding fifth.
In multi-voice counterpoint this use of 3^ is always aided by the sharper dissonance of
the second
M
EjSf *
rhis does not include those instances where the fourth
is the only dissonance in the given harmony: though such use of the
six-four chord is by no means unknown in the strict style.

Passing to specific cases, the six-four chord may be used in the following situations:
(1) On accented beats, as preparation and resolution of a dissonant fourth:
Ex. 123 Palestrina

Ex. 124 10
kdhidi
SS »i-e-
=i M M
T» S3? Wf fTTP rr -Ist-
etc.
-o-
Ex. 125 Ibid. Ex. 126

m Ibid.
-fo$ a Egz _„ ,) ,1. J 3 J j i J

S *
s
etc.
gS ql-JcE go P s^ £ m i
Ex. 127 Ibid. Ex. 129 Orlando Lasso

Ibid. Ex. 128 Jo
3ES "3£§£ i mg tf B a
^ rj
^3~
trrr -©-
a1 -©-
-.IfJ
ir
ff
Ex. 130 Idid. Ex. 131 Palestrina
10 10
^■gff d=£ ti -e-=- i
ffTr -e— jCC
rrmrrrrr
m -©•- ii id i -©♦ m
^
rr
(2) As the resolution of a suspended seventh (6), which as passing-note forms a
fourth with the bass:

I 6
I £ ^3 5r
ii
*
±zztp "—P"
r=f
(3) Simultaneously with the resolution of a suspended seventh:
Ex. 133
Orlando Lasso
1 1 rj|>J tr
™lW
i
S» IE -©-
(4) As the result of passing-notes:

Ex 134 Orlando Lasso Ex.135 Ibid.

_, L_
^
so fi
8= xc
(5) Formed by an anticipation:
Ex. 136 • Palestr
etc.
Many other examples might be cited.
§ 231. It is difficult to assume that these tonal combinations, known in modern

harmony as six-four chords, were regarded as such by the masters of the strict style.
There also appeared a number of other effects that are now analysed as secondary
seventh-chords, ninth-chords, etc., but which in the period considered were not designated
by any terms, nor understood as situations that could be classified according to the
rules of any system or theory. In only one case can a "six-four"chord—or more accurately,
a dissonant fourth— be regarded as a departure from the normal usage of the strict style.
This is when it is used as a preparing chord. Here the dual nature of the fourth is evident,
as it functions as a consonance though really a dissonance. Entering as a syncopation,
usually by conjunct movement, the fourth on an unaccented beat plays the part of a
preparing consonance.* On the accented beat following it is understood as a dissonance,
and accordingly resolves one degree downward. Such use of fourths, whereby a "six-
four" chord is obtained at the moment of resolution, is a special effect intended by the
composer, and is exceptional inasmuch as it ignores the conditions under which sus
pensions are ordinarily employed. Following are some examples of this use of
the "six-four" chord:
•Cf. Bellermann, "Der Kontrapunkt, " 4-te Auf., s. 200.

c-x j37 Orlando Lasso
Ex. 139 Ibid. Ex. 140

Palestrina
t a gg3
rr^* he *^» »=S= **
r*r
i=Lr .w-
»
iiii
*
£
^f-5
a a o'
ff
fj
r»-
fffff ICTT 10 '
s FfPF .-
J4 «-5 -o o-
II J J J JJJJ
«=
?q^ bo. IT-
£? IBF
f
Tonal Combination Resulting from the Duplication of Suspensions;
Their Importance for the Development of Harmonic Technique
§ 232. In two-voice counterpoint there is of course no difficulty about holding the

free note during the resolution of a suspension. Such is the normal treatment of the free
note; the more exceptional treatment is when it progresses to another note at the moment
of resolution. But when duplicated in imperfect consonances situations are encountered
where the note to which the dissonance resolves forms a dissonance with another voice.
When this happens, either the free note progresses to a note consonant with the note of
resolution, or else the latter continues in conjunct movement. These and other cases
result in various harmonic combinations.
a— o
Taking, for example, the dissonance of the seventh at the duplication I +
Ha
IId" ~9 (§ 225):
I the seventh resolves to 5, and 5 at Jv= —11 is a
variable consonance; therefore at the moment of resolution it yields dissonant harmony.

Consequently either the free note must progress at the moment of resolution to a note
consonant with the other voices:
wm mi m
or the free note is held and the other voices continue one degree downward from the
note of resolution:
or the note of resolution progresses one degree upward, so that the progression of the
free note forms the resolution of an auxiliary dissonance:
or finally, a new dissonance, is formed by the free note at the moment of resolution and
in turn resolves according to rule:
Hence a considerable variety of harmonic combinations is possible when dissonant

chords are formed by the methods described. Chords interconnect-one progresses to
another-and a valid, durable harmonic texture results thereby; it is the logical result
of applying the rules of two-voice counterpoint that refer to suspensions.
The introduction of these dissonances-also variable consonances-in exercises with
duplications provides, therefore, the means of developing not only contrapuntal skill, but
to an important extent harmonic technique also, as it enriches harmony to an extent
inaccessible by other methods.
There are other harmonic effects where dissonant chords are obtained not only at the
place where the suspension sounds, but also at the resolution; they result from the
parallel motion of the voices in imperfect consonances:
Ex. 142 Palesttina ^x* "3 Orlando Lasso
etc.
These are best mastered by the continued study of exercises in duplication, using
variable consonances and suspensions.
§ 233. There is a very important problem as to the use of suspensions and imperfect
consonances that hitherto has escaped the attention of theorists. It is customary in
teaching counterpoint to pass in silence the possibility of applying these intervals in
cases where they might be used and to limit the accented beats to certain fixed con
sonances. In the examples in duplication cited by Zarlino tied intervals—dissonances
and variable consonances—are entirely absent. Zarlino throws no light on the subject, as
he fails to isolate the problem, confusing it with the use of variable sixths and sevenths
in double counterpoint at the twelfth.* In their treatment of duplications Fux, in his
Gradus ad Parnassum, and Marpurg, in his Abhandlung von der Fuge, allow a fixed con
sonance on accented beats and exclude the sixth and all suspensions. Most later
theorists were too concerned in supporting Marpurg— repeating what he said and quoting
the same examples. Marpurg's authority was thereby confirmed, as though the absence
of suspensions were a necessary condition when imperfect consonances are duplicated.
The fallacy of this attitude must be shown, for his treatment of duplication is not only
meaningless but he closes the way of approach to those combinations which in a high
degree are useful for the development of harmonic technique.
§ 234. Exercises in two-voice florid counterpoint contribute greatly to the development
of this technique in the direction stated. In them the pupil for the first time encounters
the situations described, where suspensions are resolved to notes that are dissonant
in relation to the implied harmony. In this connection it may be of interest to consult
Bellermann, Der Contrapunkt, 4th edition, pp. 227-37 and 266-8.
§ 235. Mention has been made of the significance for the development of harmonic
technique of the application in multi-voice counterpoint of those suspensions wherein a
note, dissonant in relation to one voice, forms at the resolution a new dissonance with
the other voice, forcing the latter to pass into consonant harmony, with the result that
various resolutions of dissonant harmony are obtained. Theorists, in excluding dis
sonances and variable consonances from these two-voice combinations that admit of
duplication, thereby eliminate the very combinations that are of most value in improving
harmonic technique. Not only this, but some writers make the assertion that these forms
of resolution are entirely forbidden in the strict style. Fetis, for instance, in his Traite
tin contrepoint et de la Fugue (2nd edition, part I, pp. 51-2) says that consonant chords
only are the basis of the strict style; he therefore allows only those suspensions that
resolve to a note consonant with the other voices, for example:
•Cf. footnote. § 151.

cons.
e .. —
as if only such suspensions as these could at the moment of resolution be followed by an

other chord:
Those cases where the changing of the chord at the moment of resolution appears
obligatory, owing to the fact that had the voices remained stationary the resolution
would have formed dissonant harmony, he rejects as wrong. According to this the fol
lowing harmonies: (a) the seventh with the third and fifth, (b) the fifth and sixth with
the third, and (c) the second with the fourth and sixth:
are rejected as "inadmissable dans le contrepoint simple." He also states that such
instances are not found in the literature of the period and quotes a solitary example from
Palestrina, calling it a mistake. This is what he says:
"All these circumstances have probably been noticed by the composers who preceded
the invention of modern tonality and the use of natural dissonances, for one cannot
find examples of these harmonies in their works.* The following, an excerpt from the
Hymn of the Martyrs of Palestrina, is the only one (!) in which I have found the harmony
of the third, fifth and sixth, but his resolution has the defect that I shall indicate,* and
one must regard this example only as a mistake on the part of the great master. The
passage follows: "
These statements of the French theorist can with entire justice be flatly contradicted.
The composers of the era preceding our tonal system did not hold such views, and the
progression criticized by Fetis was not considered a mistake. The assertion that such
progressions cannot be found in their compositions, and that the one cited from Palestrina
is unique cannot possibly be substantiated— they are often found. In addition to several
already quoted in other connections that illustrate the progression forbidden by Fetis, a
•The italics here and in later quotations are mine. -S. T.
"Cf. the following from Palestrina: Exs. 39, 45, 46, 47, 131. 136; from de Pres: Ex. 43; from
Orlando Lasso, Ex. 56.
series of several more are herewith quoted from Vol. 1 of Ambros' History of Music.
These examples illustrate in particular the harmony of the third, fifth and sixth that
Fe'tis calls a mistake of Palestrina.
Ex. 144 _ Ex. 145 Ex. 146 Pierre, de la Rue

Obrecht I Josqui
Josquin de Pres
11
fî £=3EI
^
m Eyj do
<D etc. scp.

i)
SS ss o Q
S
Ex.147
r
^
W
, „
de Orto
Ex.148
_
„ ,
Carpentras
Ex.149
,
r , . _ .
Benedict Ducis
il
rrr ietc.
IPHW B -&- 3
JJa.i tin e tc.
.
s 351
s Id
Ex. 150 Ex. 151 Ex. 152

Henricus Isaac /\ Ludwig SehfL . a Adrian willaert
If?
Ex. 153 Ex. 154 Ex- 155
Ibid. A Escobedp a Cristofero Morales
MS ^ a I 1 = £S_
O"
i •T
8* « ~e n
eJ
fff etc. f etc. etc
*j»j'.so"j"iy 1,8 i fywfppf

rf ^^
It is freely admitted that considering the extensive literature of the period Fetis may
not have come across the works from which these examples are taken. Turning now to a
well-known composition, Palestrina's Missa Papae Marcelli, a statement in regard to it
is available. In his Biographie universelle des musiciens, article on Palestrina,

Fetis says:
"Few historic monuments of art present so much interest for study as this mass
dedicated to Pope Marcellus, for it marks one of those rare epochs when genius, over
leaping the obstacles by which the spirit of his times was surrounded, suddenly opens
up an undiscovered vista which it takes with giant strides . . . With regard to the tech
nique—the purity of the harmony, the simple and natural voice-leading within the range
of each, the melodic conduct of six parts with all the combinations known to scientific
composition, in the narrow range of two and one—half octaves—all this, I say, is above
praise; it is the despair of whoever has seriously studied the mechanics and the diffi
culties of the art of writing. "
And in this mass, the merit of which, according to Fetis himself, is "above praise,"
on the very first page, extending twelve measures from measure 7, are found the same
progressions which in his opinion must be called mistakes of Palestrina— namely these
"erreurs de se grand maitre:"
E-x- 156 Palestrina
Such, according to Fetis, are the "inadmissable" cases that are of frequent occur
rence in the strict style, and which are found not only in those compositions of Palestrina
cited by Fetis in Part I of his book as exemplary models,* but even in his own examples,
for instance in measure 2 of the first example on page 54:
etc.
and in measure 11 of the five-voice example on page 84:
•P. 123 (last measure), p. 127 (m. 9). p. 132 (m. 6). p. 135 (last m.) p. 136 (m. 3). p. 139 (m. 18).
It is not hard to find a more or less plausible explanation of how the distinguished
theorist came to fall into such contradiction. The study of strict counterpoint begins, of
course, with two-voice writing in the first species (note against note), from which dis
sonant intervals are excluded altogether. Therefore Fetis concluded that the beginning
of the strict style established a basic principle. This idea, expanded to include consonant
chords, offered a simple explanation of the difference between the strict and free styles:
the former, based on consonant harmony as opposed to contemporary harmony, which
allows the independent use of dissonant chords. This generalization, concerning only
the first species of strict counterpoint, was adopted as a guiding principle for the entire
domain of the strict style. From this basic assumption Fetis drew the furthest inferences,
not noticing that the laws he deduced were in opposition to the musical literature of the
period he was dealing with. It is not surprising that the first attempt to verify these
situations by reference to the classic composers ended unsuccessfully; Fetis found in
Palestrina a harmony that contradicted his theoretical assumptions. But this failure did
not raise any doubts as to the correctness of his theory, nor prevent him from continuing to
verify it to his own satisfaction. He preferred to attribute this one instance to a lapse
on the part of Palestrina, and presumably abandoned further investigation. Nothing else
can explain how he came to regard this one example as unique when, as has been shown,
such examples are to be found in abundance.
There is a further significance about the basic principles of the strict style as under
stood by Fetis. In a reference on page 81 he says, a propos the obscure matter of sus
pensions: "these considerations throw a bright light upon the origin and classification
of secondary seventh-chords."*
An inconsistent judgment can do much harm when sanctioned by authority. Much
space has therefore been given to a refutation of Fetis' statements, though in other
regards his treatment of the subject is a valuable contribution to theoretical literature.
§ 236. In approaching vertical-shifting counterpoint from these various angles the
first result is the possibility that double counterpoint at the twelfth may become more
familiar. Counterpoint at this interval, so important in this period of both Palestrina
and Bach, in now passing into disuse. This is due both to lack of cooperation on the
part of teachers of composition and to the superficiality which ignores the technique of
the preceding era. For example Marx, in the fourth edition of his Kompositionlehre,
presents double counterpoint only at the octave. The previous editions were no better in
this regard, and the author attempts to justify his lack of thoroughness by referring to
counterpoint at other intervals as a weakness to which he is determined not to yield.*
But such an opinion as to complex counterpoint by one of the most authoritative theorists
of the nineteenth century cannot be regarded as binding on his successors; his lack of
cooperation has only impoverished contemporary contrapuntal technique.
To raise their technique to the level of the preceding era means that composers of the
present day should face the problems involved and act accordingly. Here the right course
to follow is indicated by Beethoven, whose later works show a striking novelty and
originality, due to a return to the older contrapuntal methods. The way in which he used
•E.g., cf.
**'He [the author] will not deny that the appearance of completeness has been carefully
avoided. To retain worthless details against the reproach of incompleteness is only a weakness
to which the author will not yield." (PP. 595-6)
these resources is illustrated in the first movement of his Sonata Op. 106, where he has
a two-voice imitation in double counterpoint at the twelfth:
Ex. 157 Beethoven
etc.
P
Later it occurs with both voices duplicated, first in tenths, then in thirds. Counter
point at the same interval is used in the second section of the fugue from the Sonata
Op. 110. The following examples are from the Finale to the Ca minor Quartet Op. 131,
showing double counterpoint at Jv= —4, —14, —18:
Derivative Beethoven
Original JV=-4
Ex. 158
n
T=-4 ii^-_&-
JV=-14 II £1
Later uses of these indices:
Derivative
| Original Jv=-4 V-~\
Ex. 159 Ibid.
^^J7JU 7>i J7^J7^J74lJ
ffa= srffccrtfcfr^
n
JV=-21 ip*-
Modern music offers plenty of opportunity for the use of counterpoint at those indices
which in the strict style are only found exceptionally, i.e., Ju= ±1, ±6, —8, —13. Counter
point at these indices, troublesome in the strict style because they easily give augmented
fourths in the melody and also because the abundance of variable consonances requires
rigid observance of numerous conditions, loses its difficulty in contemporary music owing
to the unrestricted use of chromatics and free dissonances. This opens up a vast expanse
of unexplored territory, large enough to satisfy any modern composer in his striving for
originality. Here again Beethoven points the way, with a use of vertical-shifting counter
point that differs but little from the usual. In the following passage from the Missa
Solemnis, Op. 123 is found the application to free writing of counterpoint at the seventh:
Jv= 6 (if one of the voices of the derivative is brought an octave lower, Jv= 1):
Ex. 160 v=3

Presto Beethoven
Jv = 6
??F#? e£
Original
PP ^
Derivative
II v=3
gggfr s p PS FJPPp
(The original combination admits of the duplications Id lf> + II, I 2 + II, and
jd= 2 + nd= -2}
It may be said that the same intervals which were most carefully avoided in the era
of the strict style—the augmented fourth and diminished fifth—are the ones that in con
temporary counterpoint are often the chief cause of strange and unusual harmonic com
bination. Reference can be made, for instance, to a shift of the lower voice at a dimin
ished fifth, to be found in Glinka's Kamarinskoi:
Ex. 161 I Glinka
fi >
In Ex. 162 the augmented fourth (and its enharmonic equivalent the diminished fifth)
is used as the motive for the two voices that shift, with the upper voice as soprano
ostinato; a style of writing often found in the works of the Russian composers.
Rimsky-Korsakow
Ex. 162
f5— r -»^~^*r—
An interesting example of a ground-motive shifting a second above while the upper

voices are repeated without change is found in the Finale of Tschaikowsky's
Second Symphony:
Ex. 163 Tschaikowsky

(Allegro vivace. Jv=-i
mIfT f m u
r: &
!
#
&f w
3 i
ÊE3
v
III
z»—y 3E £E ^s 5**
nr--1
The last two examples are in three-voice vertical-shifting counterpoint, the detailed
study of which, applied to the strict style, will be the subject of Division B.
DIVISION B

DIVISION B
CHAPTER XI
THEORY OF THE TECHNIQUE

OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT
Preliminaries
§ 237. The next problem is the study of those three-voice combinations that yield
derivatives as a result of shifting the voices vertically. Such counterpoint will be termed
three-voice vertical-shifting. A necessary condition is the presence in the derivative of
all three melodies of the original. If, therefore, one melody appears that was not in the
original, even though the other two shift, the whole texture will be termed two-voice
vertical-shifting counterpoint, not three-voice. Various relations are possible, involving
either the direct or the inverse shift, or both. The derivative may show the result of
shifting one voice:
Ex. 164
Original Derivative
I ft V-3
4m m P
1
wdm
Ex. 165
in
Original
mMM III
Derivative
&-
I u
miFF HI jQ_
m
m p *E
3U
3&
J3L
as rV=-ll
s=F
•Voice II Is shifted an octave lower; this shift is indicated n^ ±7 (cf. § 241).
157
15g CONVERTIBLE COUNTERPOINT
or of two at the same time:
Ex. 166 Original

I
Derivative
u
fr m TSl
P
Every combination that represents a shift of all three voices may be so transferred
to other degrees that one voice does not appear to be shifted. Conversely, every deriv
ative in which one or two voices have not shifted, as in the preceding examples, may
be transferred to other degrees so that all three voices show a vertical shift. The various
relationships in which a derivative may stand to its original, due to the possibility of
transferring the derivative to other degrees, involves the uniting of indices, a matter
requiring detailed treatment.
§ 238. An original combination may have one or several derivatives. An example of
counterpoint giving several derivatives may illustrate the so-called triple counterpoint
at the octave. Its derivatives show the result of shifting the voices to other octaves, and
any voice may function as upper, middle, or lower part. Most textbooks, in applying to
three-voice counterpoint the process called in this book the vertical shift, do so only to
the extent of giving triple counterpoint at the octave, a knowledge of which, together
with quadruple counterpoint at the octave, is generally regarded as necessary for a
basic technique in counterpoint. The present work goes further, and applies the shift
at Jv- —11, i.e. double counterpoint at the twelfth, as also necessary. In general, double
counterpoint at the twelfth, as stated at the end of Division A, claims consideration on
the ground that earlier usages, significant for the technique of composition, may
thereby be restored. Several phases of its application will be dealt with sepa
rately. If to the shifts in double counterpoint at the octave and twelfth is added
the shift at Jv=~9, making possible the duplication of voices, together with
what has already been given in Chapter X, the resources of three-voice vertical-
shifting counterpoint are almost exhausted. What remain are some combinations
that have not been fully studied and others not investigated at all and still
untried. Three-voice vertical-shifting counterpoint, within the limits indicated, is
one of the most difficult phases of the whole subject of complex counterpoint. It
has reference more to the virtuoso aspect of contrapuntal technique, and the skill
necessary for its mastery means the meeting of almost innumerable requirements.
The approach to learning it can only be by way of a firm grasp of two-voice
TECHNIQUE OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT 159
§ 239. Also, the rules of simple counterpoint for more than two voices are
only a further development of those governing two voices, in the same way that
vertical-shifting counterpoint for two voices is the basis for the vertical shift
in some parts. An original multi-voice combination must be dissected into two-
voice combinations, implying that each voice must combine with any other. In
finding the index for each of the resulting combinations the possibility of ex
plaining the conditions that must satisfy the original multi-voice combination
will be obtained. To increase the number of voices in simple counterpoint means
that changes must be inserted into the rules covering two parts, and in the same
way the rules for two-voice vertical-shifting counterpoint undergo analogous
changes when these two-voice combinations are parts of a multi-voice texture.
Therefore the problems to be investigated are: the ways in which three-voice
combinations differ from those in two; the relations between the indices of these
two-voice combinations, and the nature of the changes that have to be inserted in
the rules for these indices, owing to the fact that a two-voice combination is
part of a larger one. But first it is necessary to explain how the voices and their
shifts are indicated.
Notation of Voices and Shifts
§ 240. A voice of a three-voice combination will be indicated according to

its position as highest, inner, or lowest voice, figure I, II, III respectively. The
sign + unites these figures, showing that the voices represented form a con
trapuntal union. Therefore the formula for the original combination is: I + II + III.
§ 241. The shifting of the outer voices (highest I, lowest III) will be indicated
as in two-voice counterpoint, i.e. the shift of the highest voice upward and of
the lowest downward are regarded as positive operations and indicated by the
plus sign; the shift of the same voices in the opposite direction are negative
operations, indicated by the minus sign. As regards the middle voice (II) there
is a double interpretation, according to its relation to the other voices. Therefore
a shift of the middle voice will be indicated by the use of both the plus and
minus signs, placed one under the other, the upward shift by the sign +, the
downward shift by ±. The upper part of the sign refers to the union of voice II
with voice I, the lower to its union with voice III. For example, the expression
jV= _9 + nv=±3 + rav= -5 meanS tnat voice I shifts a tenth below, II a fourth
below, and III a sixth above. Iv_ " + IIv=+^+ mv= " means that the middle voice
shifts a ninth above, and I and III remain unshifted. If separate combinations of
voice II with either I or III are taken (I + II or II + III), one sign for voice II is
sufficient, referring to its union with whichever outer voice is given. For example,
the shift of a fourth below for voice II would be indicated either I + IIV or
jjv = — J + ill. The shifting of the voices may be represented by this diagram:
+ +
tI III I
III
1 1 ± +
§ 242. In solving multi-voice problems in vertical-shifting counterpoint the

combination must be analysed into its components, each of two voices, so that
any voice may appear in relationship with each of the other voices (§ 239). In
three-voice counterpoint there are three such pairs of combinations: I + II, II + III
and I + III. For each of these combinations it is necessary to find the index
which, according to § 27, equals the algebraic sum of the shifts of both voices.
The shift of the middle voice will take a double designation (§ 241), of which
the upper sign indicates the union of II with I, the lower sign the union of
II with III.
The Indices Jv' Jv" /v£ and Their Relationships
§ 243. The index for the combination I + II will be indicated by Jv'; for
II + III, Jv". The third index, retering to the union of the outer voices, I + III, is
always equal to the sum of the first two indices.* It will be indicated: JvH, the
Greek sigma standing for "sum". As an example of the relationship described,
take the derivative: Iv= a + IIV= ±b + IIIV = C. In the first two indices (JV + Jv")
the quantity b enters into the equation with both positive and negative signs;
these cancel each other and the result is o + c, the index for the combination
I + III.
(IV=a + „v= b) Jv' =a +b
(IIV= _b + IIIV= c) Jv" = -b + c
(Iv=a f mv= cj /w2_ . +c
Problem: To find the indices for the derivative Iv= ~10 + IIV= ±5 + IIIV= 3:
(Iv=-10+ nv=5)/t;, =_5
(IIV= ~5 + IIIV= 3) Jv" =-2
(Iv=-10 + MIv=3)/t;v=_7
§ 244. The correlation of indices is expressed by the equation:
Jv' + fv"= JvZ.
Determining the value of each term of the equation:
JV = JvZ-Jv"
Jv" = Jvl-Jv'
]vt= Jv' +Jv"
Hence it follows that: (1) a derivative can take only two indices, the third is not
obtained independently; (2) one of the indices may equal 0; (3) all three indices
may equal 0; (4) two indices can not equal 0 except on condition that the third
index also equals 0.
•It should be observed that this refers to the index of a single derivative, not to an index
that unites the conditions of several derivatives. In the latter case it is not necessary to indicate
the correlation between indices (cf. Obs. to § 281).
§ 245. In finding the three indices for a given combination it is possible,

without changing their values, to substitute other numbers for the vv of the
separate voices. The series of intervals for each derivative will remain the same,
though a voice may be transferred to other degrees. For example, for Iv * +
jjv = ±9 + inv = -2 may be substituted Iv=0+ IIV = ±7 + IIIV = -4, since in both
cases the values of the indices are identical; that is, the algebraic sum of each
pair of voices remains unchanged: Jvx = 7, Jv" = —11, Jvl =—4.
§ 246. Whatever indices there are, any voices of the derivative may remain
unshifted, that is, can be taken at v= 0. In the preceding section, for instance,
the highest voice remains, and the vv of the other voices would be indicated:
Iv=0+IIv=±7 + IIIv=-4>
If the middle voice remains unshifted the derivative combination will be:
v= 7 + IIV U + IIIV . Here as in other cases JJv remains without change,
the derivative shifting to other degrees.
§ 247. If all three indices are numerically equal but cancel each other by
opposite signs so that the result is 0, various shifts may be assumed; for in
stance Iv= ~3 + IIV= ±3 + IIIV= 3; Iv= 5 + nv= +5 + IIIV= -5 etc. The result of
such shifts is a series of combinations identical with the original. In this way
every recurrence of an original combination or its transference to other degrees,
the intervals between the voices remaining constant, may represent a shift at
indices equal to zero. This, and simple counterpoint in general, is only a special
case of the vertical shift.
§ 248. As in the two-voice examples, the crossing of voices will be avoided
in both the original and derivative combinations, therefore each index can show
either the direct or the inverse shift. This enables the voices of the derivative
to be disposed in any one of six ways; these permutations may be represented
as follows:
(1) (2) (3) (4) (5) (6)

I III
II II II-^M II^ÎII "C^.1 IIÔui II^-II
III III in III nKîi HKÎI Ilk^N iir xi
In each diagram the numbers on the left refer to the location of the voices
in the original combination, those on the right the location of the same voices
in the derivative. Each number on the right is connected by a line with the same
number on the left. A line crossing another means that the corresponding voice
has the inverse shift; if the lines do not cross the shift is direct. It must be
observed that these diagrams illustrate only the process of shifting, not the
final result of the shift itself. For example, in diagram (2) voices I and II take
the inverse shift, and this is all the diagram is intended to represent. The same
applies to the other diagrams, their purpose is only to show the different lo
cations of the voices. The diagrams are abbreviated thus:
(2) (3) (4) (5) (6)

(1)
The dots to the left refer to the positions of the voices in the original, those
on the right to the same voices in the derivative. For instance, the disposition
of voices in Ex. 164 would be represented by diagram (1): \_ 3
Ex. 165 by diagram (5): ^^^^=E3 > *n Ex. 1^6 by diagram (3):
The following table shows how the signs for the shifts (§ 23) are applied to
each index, for all six diagrams:
CD 09 (3) (4) (5) (6)
(I + II; jv ><: ><:

(II + III)Jv" x: x:
Q + DIfME x x:
y\ y\ y\ y\ S\
All JJv with Two JJv with direct, Two JJv with inverse, All JJv with
direct shift. one with inverse shift, one with direct shift < inverse shift
In working out exercises attention must first be given as to which one of

the six diagrams the disposition of voices in the derivative shall conform. Stating
the problem more definitely, the arrangement of voices in the derivative must be
decided upon, and for the shifting of the voices, those intervals not chosen that
would bring about the mixed shift, or cause a shifting voice to approach too
closely to another that does not shift, or cause them to recede too far apart.
§ 249. Three-voice vertica-1-shifting counterpoint may be regarded as a union
of three two-voice combinations: I + II, II + III and I + III, represented by the
indices Jv', Jv" and JvT.. Each pair of voices must conform to the rules of the
corresponding index, i.e. it must satisfy the conditions of two-voice vertical-
shifting counterpoint, but some changes must be inserted into these conditions,
owing to the fact that two-voice counterpoint is here a constituent of three-voice.
§ 250. These changes consist, first, in allowing the use of certain hidden
consecutives. The possibility of doing this relaxes the strict rules by which
similar motion at indices of the second group (2//tf) was forbidden. Between
two voices combined according to a Jv the hidden progressions (similar motion
to fixed perfect consonances) which are forbidden in simple counterpoint may
be freely used in the original combination.
§ 251. Secondly, into the conditions of two-voice counterpoint to which the
pairs of voices in three-voice counterpoint must conform, there is incorporated
a series of changes that are the result of the different uses of the ninth and
fourth in multi-voice counterpoint. This difference, as has already been stated,
consists in the fact that multi-voice counterpoint, as compared to two-voice,
imposes on the ninth only the limitation that this interval (8 or —8) is not allowed
except where the bass voice is concerned. Therefore in three-voice counterpoint
the necessary condition for the union of upper and middle voices is 8 and ->8..As
for the other interval, the fourth, the opposite holds good. The fourth, when
occurring between the same two voices, is released from the limitations of a
tied dissonance and may be used correctly as an imperfect consonance (3 instead
of 3). These are conditions of simple counterpoint; in complex counterpoint they
must be observed both with reference to the original and derivative combinations.
Hence for each of these two intervals there arises two rules which for conven
ience will be indicated by the first four letters of the alphabet: Rules A and B
will secure correctness in the use of the original and derivative ninth, Rules C
and D for the original and derivative fourth.
I. Rules for Ninths
§ 252. As regards the use of ninths in the original combination, there

results from the rules of simple counterpoint, forbidding 8 in I + II, the following
rule of complex counterpoint:
Rule A (for original ninths).— 8 is a condition obligatory for every Jv'.
§ 253. Correct use of ninths in the derivative is secured by the next rule:
Rule B (for derivative ninths).— For an index referring to those voices which
in the derivative function as upper and middle the sign (— ) is obligatory for that
m which gives the derivative ninth (i.e. which gives n- 1£ or n- —8). If Jv has
the direct shift the sign must be placed above: m; if the inverse shift, below: m..
Therefore in applying Rule B it is necessary to take into consideration the
shift— direct or inverse— of that Jv to which the rule refers. The following table
shows this index for each of the six diagrams, also those in Which m, giving a
derivative ninth, takes the sign (— ) above, and which below:
(s) ^ C4) (5) (a)

W (m 'm' (m
Jvrtx: Jv4>^ jv» |^< jv- jX
Hence it is seen that for diagrams (1) and (2) rule B refers to Jv'; for (5) and
(6) to Jv"; for (3) and (4) to /t>2. A positive derivative ninth appears in the odd-
numbered diagrams, in these the required condition is m; the ninth in the even-
numbered diagrams is negative, condition m.
§ 254. In the following table is shown, for each index, the original interval
that gives n- 8 or n- —8.
Table of Supplementary Conditions Inserted into Rule B
11 -10 &
-4- ri -3
3 5 4
•7- 15
0
II. Rules for Fourths
§ 255. The rules for ninths were interdictory, since they imposed new limi
tations. In three-voice simple counterpoint consonant fourths are allowed between
the upper and middle voices, though in general the treatment of fourths is similar
to that of dissonances. If in three-voice vertical-shifting counterpoint only the
rules are used that impose the limitations of two-voice, while the rules that
release fourths from these limitations are not drawn upon, no errors will result,
but to do this will deprive three-voice counterpoint of many advantages. This
releasing of original and derivative fourths from their limitations adds many
resources, especially with reference to some of the difficult indices, as will
be seen presently.
§ 256. Rule C (for original fourths).-li an original fourth of an index that
refers to a combination of the upper and middle voices (i.e. Jv') gives n equal
to_a consonance, this fourth is released from its limitations (3 and 10 instead
of_3 and 10) and may be used correctly as a consonance. This fourth, indicated
by the letter p. (i.e. giving n equal to a perfect consonance), is subject to the
limitations of the latter. The conditions are expressed by the figures 3 and 10, to
p. p.
which the letter p. is added, deleting at the same time the tie-sign (3 and 10
instead of 3 and 10).
§ 257. In the following table are given those indices at which the fourth
yields a consonance (except a consonant fourth) in the derivative. The figures
3 and 10 to the right of each column, all without tie-signs, indicate that the
corresponding intervals may be used as consonances; the sign p. means that the
interval is subject to the limitations of a perfect consonance. These figures
should be substituted for the same figure in the table for each index when this
index applies to three-voice vertical-shifting counterpoint at the Jv' concerned.
Changes Inserted into Rule C
At Jv= At Jv^ At Jv'=

p p
-10 3, 10
p' p
-3 3, 10
4 h io
p
1 "7 3
In this table the indices that have the direct shift apply in the Jv' concerned
to diagrams (1), (3) and (5); those having the inverse shift to diagrams (2),
(4) and (6).
The following examples illustrate cases of where original fourths are released
from limitations, therefore of changes in Rule C.
Ex. 167 a ♦ ir-1) jv=i
Derivative(IITz"1 + IIF=4) Jv=3

Original
m
SB
I 3.
m
M
S 2
mP
M
3S £
I4,
4
JvE = 4
M
m zaz
Ex. 168
fir* '
(r=-7ÎI;jv'=-7
Original Derivative^ + HI) jvô.
I II * JvI=-7
±m
Ty-
30I
m Ptarfff
i *-
ix:
ffi
in
3E s f III
m=
33=
Ex. 169 (IMP6) JV = 6

Original (IT-' + IIP'') JT-.-11
Derivative Jvi:=-7
3E 3& -pj—^—
9
y Ax iiJi j
S 3±LL
X_
as t-PF
in
(Iv=^7 + ^=4) Jyl =_3

aiv=-4 + mv=-7)jv"=-ii
Ex. 170 10 IHV
Derivative
dgy J J J
(Iv = -14 + nv= 7) Jv. =_15

Ex. 171 Giv~~7 + ni) Jv" .» 1
Original j , Derivative ,v=±l Jv2=-14
_ fP
PP
12 12 12 12
etc.
fc^ni
4
^
v=-14
i
^m i
v=-14 + IIv="l
) JV =-7
aiv= x + mv~ ~7)jv'
" Jv' =-i6
=-16
Ex. 172 JvS=-23
Originalj3. Derivative
In each of these examples notice the fourths in the original, and the con
sonances obtained from them in the derivative. In Exs. 167, 168, 170 and 172
these consonances are perfect; in Ex. 169 and 171 imperfect, hence the possi
bility of parallel consonant fourths in the original of Ex. 171.
§ 258. To the indices listed in the preceding section must be added three
more: /i>'= 0, }v'= -6 > and Jv'- -13. They all refer to one diagram, that of the
indices at which a fourth in the original yields a fourth also in the derivative. Since
TECHNIQUE OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT , ,n
10/
fourths in any case can be consonant only when appearing between the upper and
middle voices, the latter, the union of which each of these JJv refers to, must
remain upper and middle. A positive derivative fourth is obtained from an original
at Jv= 0, and corresponding to this, also at Jv= 7 and Jv= -7<. Since these
diagrams have the direct shift, they must, in order to satisfy the above condition,
be used with diagram (1): Jv' {I *
Ex. 173
(I +11) Jt=o
Original Derivative^ Up"*) jT'=_2
f
ii T *
V* J J J r
m
A negative fourth can be obtained in the derivative according to diagram
(2): IV II 2J , where the voices concerned still remain upper and middle. In
this case the required indices are Jv'- —6>and Jv'- —13.
Ex. 174
Original Derivative^ 8(nv=8MII) ^
Jv'=3
I 3 3 3
JtTT-3
sS ±t £ ar
^jPff F^F oIV=-3
X.
^P^T
ill
'>■*■ ?r
in
m
Ex. 175
Derivative
n
fp=,e(ir6+iir=4)jT-io
*) lot-
X
tr 10
si
III T=4
§ 259. Rule D, of which the object is the releasing of derivative fourths from the
limitations of dissonances, refers, like Rule 3, to the two voices that in the derivative
remain upper and middle (cf. table of diagrams, § 253).
Rule D (for derivative fourths).—If at an index referring to the voices which in the
derivative remain upper and middle (either '■Jv or ^/v), from m- consonance is obtained
n- ±3 or n- ±10, then m is released from the sign required by the obligations of n. It
becomes a fixed consonance, used in accordance with the rul"es of simple counterpoint.
Therefore if m is an imperfect consonance, parallel progressions of this consonance
are possible, and ipse facto, parallel motion at an index of the second group &v). A
dissonance resolving to m loses the supplementary sign x, since m becomes a fixed
consonance. t
In Ex. 167 the original third in the second measure: gives in the
I ^
JII 2
derivative a consonant fourth: d~at Jv'= 1. Therefore the rule requires that
32
nv=l
_x e* el-
a change be made in the table for this index: instead of 2 3 to substitute 2 3. In Ex. 171
the derivative fourth (10) between the upper and middle voices is obtained from a third
(9) also at fv= 1, in the table for which must be made the same substitution. Since the
original consonance is imperfect its use in a parallel progression is possible, as in the
first measure of Ex. 171, also the proximity of two of these consonances (mm. 2 and 3).
§ 260. A list is herewith given of those indices at which a derivative fourth is
obtained from a consonance, hence its release from the tie-sign, i.e. it becomes fixed.
Alongside of these consonances are the dissonances of which they are the resolutions.
Since the consonance is now fixed, the dissonance that resolves to it loses the sign x.
Changes Inserted into Rule D
JV = | 3 | 0, 7 | 8
10 0, 7 | 8
3:
p
1- 7 4 3
1' i'
C—3
1 « 2 0 1 3 10
- 8 5 6
C—3
—x
- 1 4 3
p p
C— 3
-12 2 9 1 8
P
C—3
- 5» 2 1 3
-X
_ 2 5, 12 6 13
C—3 C 3
P
5 5 6
C—3
TECHNIQUE OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT lgg
Obs.— Numerous illustrations of changes in the tables of indices will be found in the
twelve examples of Appendix B. To master thoroughly all the rules of three-voice
vertical-shifting counterpoint one should make an exhaustive analysis of these examples
and of the application to them of the tables.
§ 261. From a consideration of all these changes that two-voice counterpoint under
goes when it enters into the structure of multi-voice counterpoint, it is apparent that in
the latter a union of the bass with any higher part differs but little from independent
two-voice counterpoint. The only difference consists in the possibility of hidden con-
secutives, which in two-voice independent counterpoint are reduced to a minimum, if
not entirely excluded. As for the other two-voice combinations that can be extracted
from multi-voice counterpoint, that is, those from which the bass is absent, in addition
to the use of hidden consecutives is added the possibility of the independent use of
fourths as valid consonances.
Multi-Voice Counterpoint as Developed from Two-Voice;

Verification of Multi-voice Combinations by Two-Voice
§ 262. To regard multi-voice counterpoint as the union of various combinations of

two-voice counterpoint, and to accept this fact as a basic principle of the strict style,
will mean a clearer understanding of the numerous complexities found in this province
of composition.* From this principle may be deduced both the conditions of simple
multi-voice counterpoint, beginning with the first species, and the conditions of complex
counterpoint in all of its many phases. To say that the first species of multi-voice
counterpoint yields concords in the form of triads and first inversions is to employ the
terminology of harmony, and to assume that these harmonies are used, while their cause
remains unexplained. But it is sufficient to return to the principle mentioned, and to
observe that any two voices extracted from the whole must make correct two-voice
counterpoint. Only such chords as consist of consonant intervals (including the fourth
between the upper and middle voices) can form 'the two-voice combinations that enter
into the structure of multi-voice counterpoint. In this way they provide for the conditions
of multi-voice counterpoint in other species than the first. But every attempt to apply
the terminology borrowed from harmonic theory only obscures the issue. In all intricate
situations the principle stated is the sole one to be applied. Only its continued use
makes freedom and facility in voice-leading possible, qualities of which the acquisition
would be attempted in vain by means of the guidance provided by rules and terms taken
from the study of harmony. To understand fully the significance of the strict style it is
necessary to grasp the idea that the wide diversity of effect is the result of applying
the principle here discussed, and to assimilate the facts by a study of them in clear
and definite order.
§ 263. Therefore, two-voice counterpoint is the basis of the strict style. To know its
rules is to possess the key to the whole domain of strict counterpoint. In analysing the
conditions of this dependence of the strict style on two-voice counterpoint one may not
observe immediately that these conditions are very simple. The strict style employs
melodic progressions that are simple and above all easily sung. Everything is rejected
that makes melody difficult to sing or to understand. The intervals formed by the com
bination of two melodies are also subject to simple and natural conditions. The free use
of consonances, together with the precautions that surround the use of dissonances, is
seen not as the result of rules arbitrarily imposed but have a deep-seated origin in the
*Cf. Bellermann, "Der Contrapunkt," 4-te Auf., s. 200.

auditory mechanism of mankind. One need not be a musician in order to appreciate the fact
that the resolutions, for instance, of fourths to thirds, or sevenths to sixths, are entirely
satisfactory to the musical instinct, and on the contrary, that the forbidden resolutions
of seconds to unisons and of sevenths to octaves are unsatisfactory. This instinctive
understanding of the functions of intervals in two-voice counterpoint is retained when
the number of voices is increased, and the rich and intricate effects attained that are char
acteristic of the strict style. It is all evolved from the simple basis of two-voice coun
terpoint. Each voice in a multi-voice texture enters with notes that are easily sung. It
moves with freedom yet without confusion with the other voices. Each of the latter
contributes to the correct intonation of the first voice, supporting it and receiving sup
port. With no other voice does it enter into irreconcilable contradiction. If with another
voice it forms a more or less sharp dissonance this is soon resolved, simply and nat
urally. Silent, a voice is not abruptly cut off, but the melody quietly subsides, leaving
the other voices free. The result of this simple and natural voice-movement is a rich
harmonic texture and an imaginative interweaving of parts. The independent movement
of the voices generates various dissonant harmonies— secondary sevenths chords, ninth
chords, passing chords etc.— with a great variety of resolutions, but such chord com
binations are not an impediment to the free movement of the voices. Due to the simplicity
on which it is based, counterpoint of the strict style is an art unique with symmetry,
naturalness and logic as its foundations. All its infinite variety is but an evolution from
the basis, easily comprehensible, of two-voice counterpoint.
Cases of Multi-Voice Combinations

not Verifiable by Two-Voice
§ 264. In formulating the principle under discussion, according to which multi-voice

counterpoint appears as developed from two-voice, Bellermann puts forward this method
of verification: Three-voice counterpoint must be equivalent to three two-voi ce com
binations (I + II, II + III, I + III). If each of these combinations forms correct two-voice
counterpoint (in which may be found the exceptions mentioned, hidden motion and the
independent use of fourths) the entire combination is correct.
If verification could proceed solely by the method recommended by Bellermann, if it
were applicable to all cases without exception and no corrections were required, the
problem with reference to three-voice vertical-shifting counterpoint would be definitely
solved. It has already been shown how to find the indices for all three pairs of the
combined set, and the nature of the changes to be inserted into the conditions of these
indices because of the fact that they now refer not to independent two-voice combinations
but to those entering into the structure of multi-voice counterpoint. In three-voice coun
terpoint written at these conditions both the original and derivative confirm this method
of verification, and if separated into three two-voice combinations no errors emerge. But,
as has just been seen, this does not in all cases guarantee the correctness of all multi-
voice counterpoint. Errors may still creep into the derivative, since Bellermann 's
method of verification does not correct unconditionally, and must be taken
with certain reservations.
§ 265. There are two situations where the method of verification just described does
not hold good. It does not guard against errors caused by violating the rule that forbids
doubling of the dissonance in a suspension, and it condemns as incorrect certain progres
sions in which no errors appear, but in which the free notes to suspensions are doubled.
TECHNIQUE OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT m
§ 266. Case 1. -As already known, the dissonant note in a suspension should not be
doubled. Therefore all progressions are incorrect in which the bass, used as a suspended
fourth (3 or 10) is doubled by another voice; for example:
fa) M Cc) (d) (e) Cf)

O
$
~rr IT ~rt-
n
*=?
m gH
If any of these incorrect combinations is dissected into three of two voices, each
one shows correct two-voice counterpoint; therefore the method of verification just
mentioned can not serve as a criterion of correct three-voice combinations. Since the
fault does not extend to any of the three two-voice combinations, regarded individually,
it is clear that the insertion of any supplementary conditions into the indices controlling
them can not prevent the possibility of faults appearing in the derivative. However, this
possibility can be avoided. In approaching the solution of a given problem it is first of
all necessary to obtain the original for each of these instances of wrong doubling. Then
from these originals must be discarded (1) those that contradict the conditions of simple
counterpoint (because such cases are not found in originals), and (2) those in which the
I
distribution of the voices departs from the normal: II (since voice-crossing is avoided).
Ill
The remaining cases not excluded, i.e. originals that employ the conditions of simple
counterpoint and in which no voice-crossings occur, must be kept in mind while working
out problems; in them must be avoided whatever degrees cause errors. In the foregoing
illustrations (a), (b) and (c) have fourths between the upper and middle voices. Obvi
ously, if the original interval is a consonance that releases these fourths from their
limitations for no other reason than that they become consonant, then notwithstanding the
presence of the consonance in the original, an incorrect derivative results.
§ 267. Therefore, in order to prevent the doubling of dissonant notes at the interval
of the fourth, it is necessary to find, for each of the given combinations, the indices of
the originals that give the faulty derivatives. There are various ways of doing this, one
of which is herewith indicated. In each of the illuatrations (a) to (f)» § 266, the voices
are provided with numbers in accordance with whatever diagram shows their allocation
in the derivative. For instance, if they are arranged according to diaragm (1):
I II
the voices take the order II ; if according to diagram (2): , the order I , etc.
III % * •" -■ III
In each illustration one voice is kept stationary; this is numbered II, and voices I and
II are shifted, using for voice I v equal to Jv' taken with the opposite sign, and for voice
II v equal to Jv", also taken with the opposite sign. In this way the original is obtained
of which the derivative shows the combination of voices that were shifted.
Problem No. 10 in Appendix B may be taken as an example. The voices of the deriv
ative are arranged according to diagram (5): *^^<* . therefore their order in the
derivative is III. The six illustrations in § 266 are in this order. Since Jv'- -14, and
Jv"= 1, then for voice II to remain stationary the other two voices are shifted thus:
jv 14^ jv -1.
*; and the corresponding originals are obtained:
Derivatives:
(b) (c)j{ (d)H (e) Cf>Il
=ttHZ ft
e =tn
■e-
III III III III
in HI
■ i n F 0 r:
y r r f*
i i I I I
Obtained from Forbidden in simple
these originals: counterpoint:
cf!i....?ij.iwl1
e5
aff i j 2t
Iff
m (M±
1f-
-STfT
ffl- II t
ii
^~
m St 4ft ift ffi
According to § 266, from the six originals found must be discarded (d), (e) and (f),
because they violate the rules of simple counterpoint and double the dissonance (fourth
from the bass). Moreover, (d) and (f) are useless because the voices are not arranged
II
in normal order: I , crossing therefore results, which must be avoided. Three cases
remain: (a), (b) and (c), and these also are discarded. Correct usages are cited in No.
10, Appendix B.
§ 268. From the problems in Appendix B have been extracted all those original
combinations that yield faulty derivatives; they are indicated by the same letters used
in §§ 266-7, and refer to corresponding situations. The absence of these progressions
in an original written according to the rules of the indices given will in
sure its correctness.
§ 269. Case 2. -Verification by dissecting three-voice counterpoint into three of
two voices each also includes the following:
It is known that in suspensions in two voices a third voice may double the free note
on the second half of beats, for example:
P m ^
f FP
SE
ail t P W m
In the second and third measures the dissonance is formed on the accented beat
between the voice suspended (the dissonance) and the whole-note (free note to the
dissonance). The third voice, moving in quarter-notes, leaps to a note that doubles the
free note (second quarter of the measure). If from this correct three-voice combination
is taken one of two voices, consisting of the voice moving in quarters and that having
the ties, such a combination, regarded as two-voice counterpoint, is incorrect, since
the leap to the dissonance is not allowed, and can be justified only by the presence of
the upper voice, which takes the note earlier in the measure, with the syn
copated dissonance.
In the original combination such leaps are allowed in two instances. In the first a
leap may be made to a note that forms a fixed consonance in relation to the free note
and a variable dissonance in relation to the second note of the tie; in other words if the
free note to which the voice progresses by leap is to appear in the derivative as a
consonance to the two other voices; for example:
Original
1 1'jjJnj
Jv'=-7
JvE=-16 II
fe^
In the second, such a leap may be freely made where the voices represent only double
counterpoint at the octave. Consequently, this leap is possible at all shifts in so-called
triple counterpoint:
Original
A-±—-J-\4-l3>
fft^rffj^r ~ * 5r-^1
J I J *
Ifr rr Baa
^
The cases examined in this section are of far less importance than the preceding
ones. If in writing exercises what is here said about them is not used at all, no mistakes
will arise in the derivative. But on the contrary, to neglect the first of the situations
dealt with (§ 266) will involve the appearance in the derivative of progressions forbidden
in the strict style.
CHAPTER XII
SPECIAL CASES OF THREE-VOICE

§ 270. Up to this point only the general theory of three-voice vertical-shifting count
erpoint has been dealt with. Before going on to the study of the original combinations
that yield several derivatives, of which among others is the so-called triple counterpoint
at the octave, it is advisable to dwell for awhile on a few special cases of originals with
one derivative, used in composition and having considerable importance for the devel
opment of contrapuntal technique. The situations referred to have much in common with
the shifts in double counterpoint where one of the two voices remains stationary.
The difference consists in the fact that here two voices remain stationary and one of
the outer voices shifts. If the lower voice shifts, the voices in the derivative are ar
ranged according to diagram 4: f^C^* . if the upper shifts, according to diagram
3: *-^^!^» • But while in two-voice counterpoint it is a matter of indifference whether
the upper or the lower voice shifts at the given index, in three-voice counterpoint an
other combination of indices is necessary for shifting the lower voice than that re
quired for shifting the upper, for as just seen, the diagrams referred to represent different
derivatives. Indicating by the letter a the interval at which one of the outer voices
shifts, while the other two do not, one of the following two formulas is obtained:
(1) I +11 + IIIv=-a Indices:

Jv' - 0
Jv" = -a>
/v£=-a>
(2) Iv= ~a + II + III Indices:

Jv' =-o>
Jvn- 0
/ul= -a >
The difference between the two formulas is that in the first Jv'- 0 and in the
second Jv"= 0.
§ 271. First will be examined a shift at two of the most useful indices: Jv- —11 and
Jv- —9, then examples will be cited of some of the more difficult indices.
§ 272. The following example has /t>"= -11 and /t£=—11, therefore it admits of
the shift IIIV= -11 (cf. the first formula in § 270).
174
SPECIAL CASES OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT
175
Ex. 176
Original
^ i+ii + nr-11 Derivative
Palestrina o ur-u
m^MT T k K
ii
3(E
in i jce;
^< i
S ff
f n
The original combination is taken from the Sanctus of Palestrina 's mass 0 admirable
commercium, and the derivative as far as the asterisk from the Benedictus of the same
mass. This, a posthumous work of Palestrina's, was published in 1599. But also in
Zarlino's book Istitutioni armoniche (first edition 1558) are found the following examples
(in Chapter 62), both with the same shift as the preceding:
III vr-4
r]v=i7V
-7 i Jv=0
T•
Ex. 177 TITV--4 1 ilkli
Original 111 JvE-.-ll Zarlino
# m -u—z —a ps 3T 4s- m
- n i
s ^
p P . , ? r
£
^ y
in^
$
» d* «— « gj-îi^ j__, J_u. m
f rr
H
-si-
= j: „J3J
^Ê m^ $&
**
ffrr B s
i^rfrfr r t
^^
p*=#
31 ££Mfr ^
R
Derivative III v = -4
i -7^
f fS* ^¥^
frfff ,v = ±7
S DEC 3 zx: 3
JLl i J
In combination I + II, measure 7 and 13, the voices cross. There are suspensions
in measures 7 and 14. In Ex. 178 there are neither crossings nor suspensions. The
absence of the latter makes the execution of the problem very easy. It all amounts to
this: that for an outer voice to combine with each of the other two, the sixth can not be
used in the capacity of a consonance. Then this outer voice admits of a shift at Jv= —11.
It is understood that a shift cannot be made that is beyond the limiting interval of this
index; such cases must be taken at v- —18, not at v- 11.
>SI y--i ^\\y -1a IV *
Ex. 178
§ 273. This use of double counterpoint at the twelfth was passed on by tradition to
the composers of the era of free counterpoint, as can be seen in the following, from
The Well-Tempered Clavichord (Vol. I, Fugue 17):
SPECIAL CASES OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT 177
Ex. 179
Bach
Original j
jg^Vj &*_ Derivative

11 ,rvt="|^»^ IIP**10
,y rm
m <3i ^ ttij'to
*
.V;b''i,M-, j
."t^â^ t^fttLfCJ kr
§ 274. If the application of this counterpoint to the free style is attended by some
needless difficulties because of harmonic requirements, modulation in particular, on the
other hand the possibility of the free use of sevenths (therefore sixths) opens up an
extensive field for diversified harmonic combinations. Moerover, a detailed study of the
application of complex counterpoint to the free style does not enter into the problem of
actual composition. It may be observed that in analysing polyphonic works one is not
limited to a study of the derivative combinations actually used; analysis can be carried
further, and be made to include all possible derivatives, whether or not the composer
employed them. This attitude contributed materially to the development of what may be
called the spirit of contrapuntal combination. This may be illustrated by a few examples
from The Well-Tempered Clavichord where double counterpoint at the twelfth is ap
plicable. The derivatives given are only possibilities, they are not found in
the works quoted.
Bach
Ex. 180
Possible Derivative
m-4
te=#ffl
in v=-ll
PPJJ^ fe
mw
as
^
*9=
s
*
Such shifts can be made also in four-voice counterpoint, as is seen in the two fol
lowing examples, where the two lower voices are shifted a twelfth higher:
Ex. 181
Original
I | . . | | | 1 , . i Bach
Ex. 182
Possible Derivative
Original Bach
In conclusion is cited a four-voice example in which only one upper voice shifts in
double counterpoint at the twelfth, relative to the other three voices (Iv 1°).
179
Ex. 183
Original
Bach
UJi.ian
Possible Derivative
#^f I
§ 275. The skill to manage counterpoint at the tweltfh in all of its many aspects is
what distinguishes the technique of earlier composers from that of present-day writers.
Most recent theorists do not emphasize the importance of this counterpoint. From the
fact that Bach assimilated the processes of composition— known to Zarlino—that existed
at the time of strict style it would seem that the attention of theorists would have been
directed to such passages as Ex. 179. But in the textbooks of van Bruck, Jadassohn and
Riemann, dealing especially with the analysis of the preludes and fugues of this col
lection, is found only silence on this point. The attitude here taken toward counterpoint
can have much significance for contemporary technique. Counterpoint at the twelfth, as
compared with many other intervals, is very simple; its use is easy; multi-voice com
binations where it is applied sound full because of the freedom with which thirds and
fifths are employed. In this connection this counterpoint plays a predominant role in
canonic imitation, about which details are given in The Technique of Canon. A general
acquaintance with the various applications of counterpoint at the twelfth should be
regarded as indispensible to the technique of today. The persistent and assiduous study
of the exercises here indicated is recommended. To recover technical skill, now lost in
contemporary music, would be in the highest degree desirable.
5 276. In the following example counterpoint at the tenth is used. This example
gives the derivative: *^!S<^ IV ^ + II + III.
Indices: Jv' =-16

]vn- 0
/t>2=-16
£^S -o-1- Hj
+-+i+ if5* kU^k

m Tf TT
m^
ai ^1 O^
Regarded seperately, both the original and derivative combinations sound empty in
places. On the contrary, the simultaneous execution of_the original together with a shift
at Iv= ~~ 1° gives complete four-voice harmony. Since 8 is absent in combination I + II
and I + III there is no obstacle to this duplication.
§ 277. The most useful counterpoints are those at the twelfth and tenth. The next
two examples illustrate the use of some difficult and infrequent indices. Ex. 185 gives
the derivative: ,v=

Iv -12 + 11 + III.
Indices: Jv' = -12

/:;"= 0
/v2= -12
Ex. 185
Original
P' 0
m p^p
m
II
^
I
t j- ±a i
m
Derivative
,v = -12
^
iW
181
Z Tn Z K
F^ m m P^^=5=3
E
^5iw.
f Q si DOC
tr
^ i i * 3 a
This example, like the preceding, can be executed in four-voices, although the
movement of the outer voices in parallel sixths sounds rather indolent and far less
finished than would parallel motion in thirds.
In Ex. 186 an outer voice is shifted at the ninth (Iv **), Formula for derivative:
^* jv= —15 + n +nL*
Indices: Jv' = -15

Jv"= 0
Jvl- -15
Ex. 186
^3^ Derivative \\
y**
Be
^
ps
III i~
^S y^£ J-Lfc ^&
tJ[^ f^-Ef^-f^-fr
*Cf. also the shifts at the ninth in Nos. 7 and 8, Appendix B.
§ 278. It is doubtful if a single instance could be found in the literature of the

Polyphonic Period of where shifts at Jv = —8 or —15 are used in three-voice vertical-
shifting counterpoint. There are infrequent cases where along with Jv= -3 are also
found Jv= —13 and Jv= —21, as in the following passage from Fugue 17, Vol. I of The
Well-Tempered Clavichord (cf. Prout, Double Counterpoint, § 236), with derivative:
X * jv= -14 + nv= ±6 + mv= -7
Indices: Jv' = -8
lv"=-U
/v2=-21
Ex. 187 Bach
Original
mv00
Derivative
ft > "-'• j " . i ' " ^^ m±

S f
sqLJ»
^^ BB
jv=14
§ 279. It would be useless to look for illustrations in elementary counterpoint text

books as to how such shifts are obtained. The possibility of three-voice counterpoint
admitting of a shift in double counterpoint at the ninth is a matter that exposes certain
theorists to grave suspicions. For instance Sechter, in § 75 of Von Doppelten Contra-
punkt, says: "The impossibility of three-voice counterpoint at the ninth is obvious if
one remembers that at the doubling only the fifth is free. What then can the third voice
do?" etc.
§ 280. In Appendix B are twelve examples of three-voice vertical-shifting counter
point, each with one derivative, and arranged in the order of the diagrams. These ex
amples can be used both for analysis and as guides for the writing of exercises.
CHAPTER XIII
COMBINATIONS GIVING MORE THAN ONE DERIVATIVE;

TRIPLE COUNTERPOINT
§ 281. The methods of deducing the conditions of a compound Jv for two-voice

combinations giving more than one derivative were stated in Chapter IX. The same
methods will be applied to three-voice counterpoint, taking each of the three indices
separately. For instance, if one of the required indices in a derivative combination is
Jv'= a, and another Jv'- b, the compound index Jv'- a, b is obtained. If two derivatives
require the same index the latter remains without change. The conditions of the other
indices Jv" and /t>2 are defined in exactly the same way. To the conditions of two
derivatives can be added those of a third, a fourth, etc.
Obs.—When uniting the conditions of indices that refer to several derivatives /v£is
no longer equal to the sum of the two other indices. The general formula Jv' + Jv"~ Jvl,
(§ 244) applies only to indices referring to one derivative, not to J Jv that indicate the
united conditions of several derivatives.
§ 282. All the problems of three-voice counterpoint giving more than one derivative
can be solved by methods indicated in the preceding section. From the infinite number
of possible combinations will first be examined some highly characteristic cases, first
of all the so-called
Triple Counterpoint
§ 283. Theoretical works ordinarily deal with triple counterpoint at only one interval,
the octave. By this term is to be understood three-voice counterpoint giving derivative
combinations by shifting the voices at the octave, besides which each voice may be
upper, middle, or lower. The result of these shifts is that arrangements of the voices
are obtained according to all six diagrams (§ 248). In this way an original combination in
which the arrangement of voices corresponds to diagram (1)* • gives five deriva-
• •
tives, with the voices arranged according to the remaining diagrams. An absolute con
dition of this counterpoint is that each voice must be written in double counterpoint at
the octave with each of the other voices.
§ 284. Two questions arise in enumerating the features of triple counterpoint: (1)
why does this counterpoint, though giving several derivatives, not require a compound
index (as would follow from § 281), while each pair of voices takes only one: ]v= —7?
(2) If the voices are written in double counterpoint at any interval other than the octave,
can the same number of derivatives be obtained by shifting the voices? The answer to
these questions gives the
General Formula for Triple Counterpoint

§ 285. To deduce this formula use is made of the definition of triple counterpoint at
the octave (§ 283), substituting interval a instead of the octave for shifting the voices.
Then the sum of the five derivatives is obtained to get the required index. The inverse
shift will be assumed between the outer voices, so that each of them, shifting at —a,
exchanges place with the others. Hence two diagrams are obtained: the shift of the
183
lower voice at -a gives diagram (4):£^2^* , and the shift of the upper at -a diagram
(5):£^5*^1 • Diagrams (2): I^Xând (3): •^____-» , with the direct shift be
tween upper and lower voices do not show the result of shifting an outer voice, but only
the middle, which exchanges position with one of the outer voices. To obtain any of
these diagrams it is sufficient to shift only one voice. On the contrary the remaining
diagram: (6) *^><C3 retluires tne simultaneous shifting of both outer voices, the
middle voice retaining its position.

These are the shifts considered in the following table, where for each derivative is
given its indices according to the method indicated in § 281.
Derivatives:
^ (I -+- IIv = =Fa-r- III)
Jv' = — a, Jv" = a, Jv2= o.
>=3 (i+nv=**+m) Jv' = a, Jv" = — a, JvE = o.
^ (i + ii + mv=-a) Jv' = o, Jv" = — a, JvE = — a.
>*< (iv=-»+ n + in) Jv' = — a, Jv" = o, JvE = — a.
^ (IT—•+ II + IHV=~") Jv' = — a, Jv" = -a, Jv£ = — 2 a.
General formula for
triple counterpoint: Jv' = — a, a, Jv" •= — a, a, Jv£ = — a, — 2 a.
The Difference Between Triple Counterpoint at the Octave

and Triple Counterpoint at Other Intervals
§ 286. A formula for triple counterpoint at the octave can be compiled, based on the
general formula. Substituting 7 for a is obtained:
Jv'= -7, 7; /»■= -7, 7; Jvlp -7, -14.
In the first two indices Jv- 7 must be discarded, since the derivative is only the
octave equivalent of the original (§ 47); therefore these indices are simple and not
compound: Jv'~ —7 and Jv"- —7. As regards /t>X= —7, —14, the latter value is super
fluous, because of Jv- —7 (§ 50). And so the formula for triple counterpoint at the octave
assumes this form:
JV'= _7, /„•■= _7, /„S= -7.
Therefore this counterpoint does not need a compound index. This answers the first
question brought up in § 284.
06s.—The formula for triple counterpoint at the octave, in the way it is here pre
sented, is inconvenient for practical application, since 7> is the limiting interval be
tween the outer voices (for /vS= —7) and restricts the freedom of their movements.
Another formula, more useful because it contains /t>X= — 14>, is given in § 288.
§ 287. The absence of compound indices is characteristic of triple counterpoint at
the octave. Triple counterpoint at all other intervals absolutely requires all three double
indices, as the execution of such problems is in the highest degree difficult. If, for
example, a= 11 appears in the general formula, then is obtained:
Jv'- -11, 11; Jv'- -11, 11; JvH- -11, -22;
which in conformity with the statements in §§ 47 and 50 reduces to:
Jv'- -11, 4; Jv"- -11, 4; Jvl- -11, -8.
COMBINATIONS GIVING MORE THAN ONE DERIVATIVE 185
Obs. i.-The compound indices in this formula, especially the very difficult index
]v- —8, makes the execution of problems so hard that it would scarcely be possible to
cite a single example in which all five derivatives are obtained by shifting the voices
at the twelfth. To obtain the derivatives lacking it is customary to turn to the easier
shifts at the octave, then, by combining with shifts at the twelfth or some other interval
to get what Marpurg called "mixed triple counterpoint." But for such insoluble problems
it is better to substitute easier ones.
Obs. 2.—Due only to the application of elementary mathematics, and associated with
this, an exact terminology, it is possible to explain the difference that exists between
triple counterpoint at the octave, on the one hand, and all other intervals, on the other.
Theory up to this time, not taking advantage of mathematical resources and lacking an
accurate terminology, was deprived of the means whereby could be analysed the complex
relations that the shifting of voices exhibit in triple counterpoint, and was necessarily
compelled to abandon these problems as admitting of no solution.
Triple Counterpoint at the Octave

§ 288. Therefore triple counterpoint at the octave is a case unique of its kind, for,
requiring no index except Jv= -7 (or its more practical equivalent Jv= -14), it admits
of shifts according to all six diagrams. Since each pair of voices is written in double
counterpoint at the octave each voice can function as bass and can take part in two
combinations, because the other two voices, shifting at the octave, can exchange places.
There are six of these permutations. The formula is given in § 286 but its rigid ap
plication is inconvenient, due to the fact that in it /v£ = -7, whereby 7> is the limiting
interval for the outer voices and prevents freedom of movement. In order for the limiting
interval to be 14> it is sufficient for derivatives ^4and t>^ (§ 285) to
take ]v^L= —14 instead of /t>X= —7 for shifting the outer voices. In this way Jv = —7 is
removed altogether from the compound index /t>£, so that the formula as statedin § 286
is reduced to:
Jv'- -7, Jvn= -7, /vS= -14.
This formula, applied in the following examples, gives entirely natural voice-move
ments without any crossings. It is of far greater usefulness from a practical point of
view than the formula in § 286.
Ex. 188
I+IWII
Original
ft &
'
W TFT ft? nm
^^
_UL -^
m m m &
£=£ m 3E
I+IP^+lII Jv-.-7
Jy"=7
JvIlO
1st Derivative uvt;7
iJ.JJ
TNfrfcrr £
i11 . ' TTT
i
SlU-^ &-*. te
W
m
BI r s i££ B Z
E§ i= £ Tt~
Jv"=-7
JvE;0
2nd Derivative
1
see f^s Si PP I'jj"
M^ JJ iîi it i J J i
7577" HE
3^^ I P=^
mir
Iv-7 + IJv:,7+Ilr-7Jv.:0
Jv"=-14
3rd Derivative Jv£_--i4
Ar IIIJY=-7, .
i EOTTft
«—»—
IT--/
^ <fcri g
£*S J i"J
^m ^
JJV-7
s -O-
fitt^
3E
r-M+n+in jv-h
Jv"r o
.TvZ r-14
4th Derivative
4L
k% -
4 3E
A
^< III
*U IX i HE
f fufcrr
r V=-14 '
+ II + iir JV' =-7

JV" =-7
J»£ = -14
5th Derivative
mmmm*u ,v = -7
it
ffr i
ie
-^
^
HE
§ 289. Since in triple counterpoint at the octave each voice is written in double
counterpoint at the octave with the other two, neither fourths nor fifths may be used
as consonances, but must conform to the conditions of Jv= -7, i.e. 3 and 4^ There
is a single exeception where a fourth is allowed with the bass (cf. § 231).-*
10
IE ±±^^+35.
$
X
§§
III
The shifts of this illustration in triple counterpoint at the octave show the use of
consonant fourths and fifths:
10
x 10
pfrrrrrffT i §
• dmm s?= 1 HE
X
3§E
FPf fP n^ r
*
f^fTr
i
a fI "o
^^
n i
>s<
TT
iXW :Sfc
X
4
P
Aside from such cases as these, neither fourths nor fifths may be used in the capacity
of consonant intervals. If the condition 8 is added, obligatory at Jv= —14, it is also
necessary, in avoiding crossings, to make Jv' and Jv" conform to the limiting interval
7>, and /v£ to 14 >. By these conditions all the rules of triple counterpoint at the octave
are exhausted. Therefore the impossibility of the free use of fifths in places where ties
are absent results in incomplete harmony. This counterpoint attains harmonic com
pleteness by the use of suspensions and complex forms of resolution, as was seen in
the illustrations in § 289.
§ 290. Otherwise, harmonic completness can be attained only by reducing the
number of derivatives, for in practice it is hardly ever necessary to use all six per
mutations. For instance the free fifth may be used in combination I + II, but then voice
I can not appear as lower voice of the derivative. Therefore two derivatives are relin
quished: •^^■^^J and 2^^<Zit and there remain the original and three derivatives:
• * , J-3"""*-^ , •>--^^--», *-^^^* • Also in combination I + II the fourth
may be released from its limitations, using it as a perfect consonance, but not entirely
releasing the fifth from its limitations (4 instead of 4, since the fourth becomes con
sonant). In this case voice II can not appear as lowest part in the derivative, therefore
these two derivatives are eliminated: *^x^* and *^>^^*, with these remaining:
If both situations apply to combination
I + II and both 4 and 3 are used (the latter as a perfect consonance), neither voice I nor II
can function as lower part, and four of the derivatives are relinquished, leaving only
Here, in two-voice combinations where the bass participates,
8 may be admitted instead of 8, but in this case the voices must not shift in such a way
as to contract the ninth to a second. If combination I + II contains parallel fourths, any
shift to these voices is possible, and the original combination remains, i.e. it is within
the province of simple counterpoint.
This is so clear that no special illustrations are needed. The cases designated, not
yielding all six permutations, might be called incomplete triple counterpoint, or else
triple counterpoint with a specified number of derivatives; for instance, triple counter
point with three derivatives, or with one derivative, etc.
Obs.—Those who have fully mastered triple counterpoint at the octave will have no
difficulty with quadruple counterpoint at the same interval, where shifting the voices
gives twenty-four permutations. Here also each voice is written in double counterpoint
at the octave with each of the other voices. With such an abundance of derivatives
several can be discarded, leaving plenty for obtaining the degree of harmonic com
pleteness desired (§ 290).
§ 291. From a practical point of view triple counterpoint at the octave is of greater
significance than any other aspect of three-voice vertical-shifting counterpoint. The
ability to manipulate it is an indispensible requirement of good contrapuntal technique.
Its use was more general in the era of free counterpoint, beginning with the first half
COMBINATIONS GIVING MORE THAN ONE DERIVATIVE ig9
of the seventeenth century, but it has not lost its importance for our time. In some
textbooks are found many examples of this counterpoint, taken from the works of Bach,
Handel, Mozart, Haydn and other composers. Some very valuable suggestions as to the
use of triple and quadruple counterpoint at the octave, free style, are given by Prout in
his Double Counterpoint and Canon, Chapter IX.
CHAPTER XIV
COMBINATIONS GIVING MORE THAN ONE DERIVATIVE

(Continued)
§ 292. The theory of three-voice vertical-shifting counterpoint was explained in the

preceding chapter, together with a detailed study of one phase of it: triple counterpoint
at the octave. Some other phases of it, where voices shift at other intervals, are given
in the present chapter. It was seen how few conditions were required in certain cases in
order for shifts in triple counterpoint at the octave to be practicable. For instance, to
shift the middle voice an octave higher, so that it exchanges position with the upper
voice, requires only one condition: the absence of parallel fourths; also to avoid crossing
these voices the distances between them must be limited to 7>. Three-voice combinations
in which I + II use both of these conditions are constantly found. In this way almost
every combination in which the shifting of the voices in double counterpoint at the octave
was not intended nevertheless admit of it. From this point of view further examples would
refer to what Marpurg called "mixed triple counterpoint" (Obs. 1, § 287J. But this term
is unnecessary, it refers to octave shifts and how they are obtained. To avoid needless
complications it will not be introduced into the conditions of future problems. These
can, when necessary, be treated as individual cases.
§ 293. Triple counterpoint at the octave, as has been seen, has two main features:
(1) it gives all six diagrams when the voices are shifted at the octave; and (2) each voice
is written in double counterpoint with each of the other two voices at the same interval.
It is advisable to dwell for awhile on the second of these attributes, and inquire: How is
a derivative obtained when each voice is written with every other in double counterpoint,
not at the octave but at some other interval, which for the present will be designated by
the letter o? This question is easily answered by the help of the general formula for
triple counterpoint (§ 285); all that is necessary is to remove all indices from it except
Jv = a, namely: /v'=o, /v"=a and /t>2 = -2a. This excludes from the total number of
derivative combinations the three diagrams in which these indices are found:^_
land "* . Hence there are two derivatives : J?^S a^ t^><2that

are obtained at the given conditions, i.e. the three indices Jv' =-a, Jv" =-a, /t>X = —2a.
§ 294. Applying a= 11: , N
Ex. 189
Jv' = -ll, Jv" = -ll, Jv2 = -ll
Original
f
Ir r "I
eg
w
r r ,
(§§ 267-8)
h
m i i
m£
fFff
£
mm s
190
Jvr-7 + I]V-t7+J]]V=-4 Jv..0
Jv"z-ll
JvSr-n
1st Derivative „V,-4
Ê
p i JUJJ
|v:-7
j. j ii
X&- inn
m 3E
>^
iJi iJJ
pi dy
«&
*
3S p=W-
.ifrMfrr- I P
II Vlf7
r=-«+n+ni *=-»
Jv"; 0
JvS;-ll
2nd Derivative
L
»^f=^ Ja j
Pg zz P 3E
>S< III A. i A- J J
a
^
r wm i m # £ *
§ 295. In this example only the two outer voices shifted at —a, forming two deriv-
atives:*~^^^^~»and*^^>^-». Now to these shifts is added an upward shift of the

middle voice at the same interval: IIV_"*" . Hence the number of derivatives is increased
to three, and to the two former diagrams is added one more: ►. Also Jv" becomes
a compound index: Jv"~ —a, a.

In Ex. 190 a- — 9. Uniting the indices of each derivative (§281) and taking Jv= 2
instead of ]v = 9 (§ 47) gives: Jv'=-9, Jv" =-9, Jv2 = -9.
Pi
Ex. 190
I+[|+ |l|--» Jv:0
a ij1 ^
JV :-9
1st Derivative jvX—9 X§§ 267-8)
u HI-
r-*+iMn jp:«
2nd Derivative j^"- <,
Jv : g
JvE:0
3rd Derivative
JriA^jHa
In this example the three-voice combinations, original and derivative, are empty if
regarded separately. Harmonic completeness can be attained by the inclusion of a fourth
voice, when a shift of one of the voices is added to the original. This involves no
difficulty, due to the fact that If is absent as between the upper and each of the inner
voices, also between the inner voices themselves.
§ 296. Hitherto all voices have shifted at one and the same interval. Ex. 191 has
three derivatives, the first resulting from the shift Iv~ 2, the second from IIV_ ±1 , and
the third from Iv= ~2 and IIV= ±11 together. The first derivative takes only the direct
shift. Combining the indices for each derivative (§ 281) and taking Jv- 2 instead of
Jv- 9 and Jv- 4 instead of Jv - 11 (§ 47) gives Jv'- 2, Jv"= -11; JV2= 2, -2.
(S3)
Ex. 191
Jv'rS, 4; Jv"r-ll; 3\Z:2r't-
Original wmm&tmim
i A \-
p0 -f-r
MM; ~£e± im £=l m

r-2 mi mii jv-.2
1st Derivative jf^
s
T~Z rr I
#Ffr ^ «r
^ rrr^rrr
£ I
gfgr^ at H i=£ i£g i£ £
I*|]*"" 'III •>>'= "
Jv .--11
2nd Derivative jv-S-o
± i
«=y J^ ^m
itiUL JJJJ I J iJJl J J J Jil
HI
^3
«r- rf r Pi E?P rrr-ifr r1-
|r-.-t,||T=;u, m jv.:9
Jy"l-U
3rd Derivative Jv S--2
fee ij,] J%_f

*
55 VhLU
i^^
F"f r
The first derivative has /v'=2, an index corresponding to an imperfect consonance.
In three-voice vertical-shifting counterpoint, as in two-voice (§ 196, [1]), when the upper
voice is shifted at an imperfect consonance this voice may be executed together with
its own duplication, forming four-voice counterpoint:
wmî^m
The same applies to the second derivative, where duplicating the upper voice
at Iv * gives this:
&=&=*=
MM etc.
^fe j-fy r'rr-ii

Further details as to duplications will be discussed in the next chapter.
Extracting Derivatives from a Given Combination
§ 297. Up to this point the indices for a given three-voice derivative have been
obtained by applying the formula for derivative combinations. The opposite problem is
now to be considered: how derivatives may be extracted from a given combination.
To do this exact values of Jv', Jv" and /i>£ must be determined. If any of these
indices are compound, i.e. if they conform to the conditions of several indices, then
their sums must be taken. In obtaining their values in this way the combinations must
be united which would correspond to the formula Jv' + Jv" + JvJ, The derivative com
binations obtained from a given original are controlled by these values.
§ 298. The conditions of the indices that satisfy the requirements for Jv', Jv" and
Jvl. in the following example will be determined.
Ex. 192 Original
feSa A^=A
B5
sî
n
S=?Z
£±
m ¥=m-
^
This example was written with the object of obtaining the largest possible number
of derivatives. Similar motion is absent from all three pairs of voices, therefore the
usable consonances are very limited in number. For instance between voices I and II
only three consonances, used without preparation, are found: 0, 2 and 7. In the third
and sixth measures are fifths between voices I + II, used according to a limitation of a
variable consonance (4). Sixths are absent altogether. Examining in order the com
binations I + II, II + III and I + III it will be observed that each of these combinations
satisfies the conditions of the indices of the fifth column: JJv = —9, —2, 5. Besides this
they all admit of a shift at ]v= —11. Finally, Jv' fully meets the requirements of Jv= —7,
while Jv" and Jv 2 meet those requirements conditionally: because of the free use of
fifths, the combinations to which these indices refer (II + III and I + HI) can be shifted
at Jv= —7 only according to diagrams in which the voices appear in the derivative as
upper and middle. These diagrams are, for / ""[^fot/îxi

The JJv common to all three indices have been pointed out. To them must be added
Jv= 2, but only to one Jv'. Therefore each of the three indices combines the conditions
as follows:
JV = -11, -7, -9, ±2, 5;

/w" = -ll, -7*, -9, -2, 5;
/i>S=_11, -7*, -9, -2, 5;
* conditional.
From these values must be put down a formula for each separate derivative. The
indices may amount to a considerable number, in view of the fact that voices of the
derivative may be shifted an octave and therefore take Jv= — 18 instead of /i>= — 11;
Jv= -16 instead of Jv = -9; /v = -14 instead of Jv = -7, etc. Similarly an index with the
direct shift may be substituted for another index with the same shift, to the value of
which is added 7, for example instead of Jv = 2 to use Jv = 9 etc. Also /t>= 0 or Jv = 7
may be freely taken for any combination.
§ 299. Now, with the resources given, formulas for derivative combinations can be
compiled of which each presents the equation Jv' + Jv"- JvtL. Any index is retained of
which the value is a term in this equation for /t>X. For instance, taking Jv £= —11 (or
its equivalent Jvl.= -18), from the lists of combined indices are selected for Jv' and
Jv" those of which the sum would be equal to —11 or —18:
Jv1 - -11 Jv' =-16 Jv' = -9

Jv" - 0 Jv* = 5 Jv"= -9
JV2= -11 Jvj= -11 JvJ= -18 etc.
These correspond to derivatives 12, 13 and 14 at the end of this section.

The original combination has as many derivatives as there are formulas compiled
by these methods. Following are fourteen derivatives, arranged in order of the diagrams,
each with its formula.
1st Derivative v=5
Iv~5+ 11 + III
Jv1 =5
m
4
w& XE
JV" =0 A «
Ji£ = 5
III ?
2nd Derivative
Iv = 5 + IIv=^3+niv=2
Jv' =2
Jv" =5
mm ff &
XT
iiJrh
r
an
XX
xx
JvS = 7 gag £=d JSZ ^

mv = 2
3rd Derivative r*^ ut iir-3.

1v:-4<,r;3+]11V,4jv._,
Jv":7
Jv£-0
4th Derivative
lv,-/, + jjv:.-&+j]|v,?.jv.,.,,
Jv": 7
JvZ:-?
lllv;1
96T 3iaiiH3AN0D iNiOdaaiNfiOD
^ flFF f=p s
6-:..Af
it
SAf t< l|LjT$j§
>~<
i
^ p rr pif r 3E
3T
i-:.jiii ^ it
'?.--z*r
IS,.=.ai i
n--M
lfl£ aATJBAUaQ ^< Pi
^5 3T
3E
;JU+teJI+freJ 0:,AfT,. ^
8T"=.Af
«i-=SAr ^P ^
f=P^ HE
_»:a1I ^ ;fafe" 9-J -y
th8 9AI}eAU3G Pi ^nn

ja*oJi*«j =S p 31 ^ 3E
9i--.v.r
9I--2Af
gfj _jl r"r' ii EP =Ff^
fMri^
£?;aII
SE
M16 3Ai}eAuaa ffi ^jit 33=

-:a1II+X.^;aII+5-za1 2-.AfU
sB i
3T.
^
F? rr^ff
SI-:..Af ^
Al- 91-3
m ^1 f=f 31
3E
zm
^^^ i J-
I MX) aAiiBAuaa Pi
--^F=f, jr
~n~
S!:v.fo-=AlII+iT-.JI^--.J
ft
f=F f^ fL-rrT f
:..Ar 9i
T^^l 31
±zk 31
HE
±± IE
Mill aAi}BAuaQ
^3fa
-Lin »i fr? ?
J J AJr-i
n
'■•:.Arw--AIII+7.T-AlItl
--..Ar 91
M--SAf
i J^l -u Hi
2T=aI
12th Derivative ^S< \ $1$ £ ^J1

rvr-8+rprT3+Up-3Jv,=.u &^--b-l: m ■4JA
Jv"=Q v;-3
JvSz-ll
lite w M ^ ^
F=F m
13th Derivative jj ftJt "^^
-e-
r^+ir^+iir^jv'i-w ^ ~
Jv'i |
JvH-ll
»< „ J2! ±U Ji
E*s ^ f Ê
14th Derivative i><^ tt
fe- IF fet ^=
f ^F r
rv---7tjjvrT2+rjp---Ujv-..H
Jv"=-9 GsiTT
jvr-
18 [ gg j J J PP 3= il ^
jv-ir
CHAPTER XV
DUPLICATIONS IN THREE-VOICE COUNTERPOINT
§ 300. Continuing the order of subjects as given in Division A, the next is the dupli
cation of three-voice vertical-shifting counterpoint at imperfect consonances. Here the
requirements are more limited, so that the original combination yields only one derivative
in which the number of voices is increased as the result of duplication. The following
cases will be dealt with in succession: (a) the duplication of one voice—four-voice coun
terpoint; (b) the duplication of two voices— five-voice counterpoint; (c) the duplication of
all three voices— six-voice counterpoint.
(A) The Duplication of One Voice;

Four-Voice Counterpoint
§ 301. In a former chapter were found cases of where one of the outer voices was
duplicated, namely: I . In accordance with § 222, 8 (and -8) is an obligatory con
dition for every two-voice combination that does not include the bass. For example
this progression:
XT
II
xr
m III
although it gives a correct derivative at the shift Iv -^ + IIV 0 + mv=0
will still not admit of the duplication Id 12. Fourths also must be considered, it is
not necessary that a three-voice derivative, apart from the duplication, shall exhibit
correct three-voice counterpoint as far as fourths are concerned; it is enough that the
total, consisting of the original and the four-voice derivative, shall be correct. Therefore
in the latter the free use of fourths is allowed, which are made possible by means of the
duplication; fot example:
198
DUPLICATIONS IN THREE-VOICE COUNTERPOINT 199
Ex. 193 n,v:-9
The writing of such combinations is facilitated considerably by the use of consonant

fourths, although they render impractical the use of the particular three-voice derivative
I + II + IIIV= ~^.
§ 302. In duplications at imperfect consonances the original combination must conform
to the conditions of the Jv corresponding to the consonance giving the duplication. As
already known, all JJv corresponding to imperfect consonances are those of the second
group ( JJv), i.e. indices not admitting of similar motion (except possible hidden pro
gressions). But in certain cases duplicating at imperfect consonances does not preclude
the use of similar motion, in particular, parallel motion. Progressions of thirds do not
prevent duplication at sixths, nor do progressions of sixths prevent duplication at thirds
(§ 212). The result is a series of first inversions.
Td= -12
The following example shows the bass duplicated in sixths: I + II + III*"
Ex. 194 nvrs-n
ft
Here it is necessary to conform to these indices:
Jv1 = 0
2Jv" = -12
2Jvl = -12
Although Jv" and Jvâte both indices of the second group (JJv), duplicating the bass
in sixths does not prevent parallel motion in thirds between the voices to which the
indices refer, namely: I + II and I + III.
As for I + II, parallel motion is possible between these voices both in thirds and
sixths, because Jv'- 0. Therefore the thirds and tenths in Jv" and Jv5L give derivative
consonant fourths, functioning as fixed consonances.
06s.— Such an example, where the voice duplicated is the bass, is easier to write
than where the voice duplicated is highest in the original combination. In the latter
case consonant fourths are prevented between soprano and alto and between sopra
no and tenor.
§ 303. In Ex. 194 the duplication in sixths allowed parallel motion in thirds at a
Jv, In the next example duplication in thirds allows parallel motion in sixths, also at a
9
Jv (in combination I + II).
Ex. 195
_L_ ^ ^—s , ^ ss
II »**
w ^
*=*.
-vr
fct
I S=* IT
1H-
n wt=HF= r=f =? fe=^
m he-
S
^
^
This example îllustrates the duplication I + IId--t-2 + III. Indices:
Jv' = -2
Jv"= 2
Jvl = 0
Into the conditions of Jv' is inserted a change: the sixth, giving a consonant fourth
(5 —2= 3) becomes a fixed consonance. Parallel sixths are found in the second measure
and in the last two measures.
§ 304. The duplication I 5 also allows parallel sixths between the alto and tenor;
this example sounds rather strange but can not be considered incorrect.
Ex. 196
Formula: Id_ 5 + II + III; indices:
JV =5
Jv" = 0
JvIf5
There is a change in the conditions of Jv'- 5: the sixth, giving a consonant fourth
(5 + 5 - 10) remains a fixed consonance. Parallel sixths in the original combination are
found in the second measure.
§ 305. Often such a combination allows a shift that was not planned for when it was
written. In finding these unintentional combinations it is usual to arrive at them quickly
enough. Whenever similar motion between voices is absent a duplication that is more or
less tentative can be thought of. If parallel motion in thirds is found the only possible
duplication is that in sixths. Unless such thirds come underneath, the given combination
will not admit of any duplication at all. It is the same with parallel sixths, the number
of duplications is limited to a possible duplication at the third (less often at the lower
sixth), as in Ex. 196. This example can give the following shifts:
DUPLICATIONS IN THREE-VOICE COUNTERPOINT
201
(B) Simultaneous Duplication of Two Voices;

Five-Voice Counterpoint
§ 306. It was seen that four-voice counterpoint obtained by duplicating two voices
in imperfect consonances includes three derivative combinations and that an index re
ferring to the original two-voice combination is a compound Jv, the result of uniting
three indices (§ 216). Exactly the same happens when two voices in a three-voice com
bination are duplicated, three derivatives result, and the index is compound, uniting
the conditions of three indices.
Obs.—lt all three indices are different the index is triple; if two of them are alike
and the third does not equal zero it is double; if two are alike and the third equals
zero it is simple.
§ 307. The duplication Id= ~16 + IId= *2 + HI will be taken. Here are five voices:
I + Iv~ ~16 + II + IIV_ & + III. I and II are duplicated. The Jv' of each must show the
result of combining three indices. Therefore combination I + II has three derivatives:
I + IIV = 2, Iv = -16 + II and Iv = -16 + Hv = 2 (here are all combinations of voice I with II
except I + II, which is the original). The indices for each of these derivatives are:
(I + IIV-2) Jv'= 2
,jv--16 + n) /„.= _16
=v-16 + nv=2)
0V /v'=-14
Hence Jv' appears as a compound index: /d'= 2,-16, —14. The two remaining indices are:
(nv- -2 + nI) Jv"= -2

(Iv--16 + III) jv^_i6
Compiling a table for Jv':
3
Jv= -2: —X
A
Jv = -16 (= -9) :? p.
8
(-)
-x (-)
Jv = -14: 3
—X
A 8_
-x (-)
2^= -2<, -16, -14> :
—X r4> (_)
(-) 8
3 can not be released from signs, since at Jv= —9 it is a fixed dissonance. p.

At Jv"= —2 the sixth, giving a consonant fourth, becomes a fixed dissonance. 5 and 6
appear as changed in the conditions of Jv= —2. 8 is excluded from the original com
bination, since II when duplicated becomes an inner voice. The following is written
according to the preceding plan:
202
Ex. 197
Jd^d,^
lVz-lb"
spf trr
§ 308. The next example has the duplication Id=-2 + II + IIId~ ~2. The first two
indices are: Jv' = -2, Jv" = —2. JvZ refers to the two voices duplicated (I + III) and
therefore takes a compound index:
(I + IHV==~2) JV2--2
(Iv = -2 + III) Jv1=-2
v = _2 + inv = -2l) Jvl=-4
(I
Since Jv?. refers to the outer voices Jv= —4< must apply. Hence ^/vX=—2, —4<. Table:
Jv = — 2: 5 6
—X c—>
c— > -X —X —X
jv = — 4.-=: 5 6 7 8
- XI C—5 —X
c—> —x —x —x
'^Jv£ = — 2, — 4-=: 5 6 7 8
—X c—5 —< —
3 does not appear in the table since it is beyond the limiting interval 4<.
Ex. 198
(C) Simultaneous Duplication of Three Voices;
Six-Voice Counterpoint
§ 309. When all three voices are duplicated, in each pair of combinations both voices
are duplicated, consequently they give three derivative combinations. Therefore each
index is compound, consisting of three indices (cf. Obs., § 306).
§ 310. Indices for this problem I d=-9 + nd=^9 + III d = -9
DUPLICATIONS IN THREE-VOICE COUNTERPOINT 203
Beginning with Jv':
(I + IIV=~9) /»'- -9
av=_9 + n) /v'- -9
(Iv=_9 + nv = -9) /v.= _18
Hence: 2/v'--9, -18.

Passing to /t>":
(II + IIIV = -9) /»"--9

(IIV=9+III) /»■- 9
(IIV = 9+IIIV = -9) /»"- 0
Hence: 2Jv"- -9, 9 (or Jv"- -9,2; § 47).

Last, Jv"&
(I + IIIV= -9) /»2= -9

(IV=_9 + III) A>2= -9
(Iv= -9 + mv= -9} /t>2=_18
Hence: 2/v2= -9, -18.

Therefore the three indices are:
Jv' = -9, -18

/v" = -9, 2
/t>£=-9, -18
A table for the double indices of Jv' and Jv~ is given in § 190; (2). It remains to
compile a table for Jv"- —9, 2:
Jv= -2: ■■', —x
4
X C—)
C—3
Jv = — 9: 8 8
C—3
C—3 —X
'-'Jv"=-_ 9^ 2: 3 4 8
— C—3 c—3
The following six-voice example presents the solution of this problem:

Ex. 199 ft** , ^
=îfF
DerivaUveJI] vr-9
m S :nr
U! ^ *> ^ fci
Ji
^3|E
FW ^
lV=-9
J.
ff rr"f^P H
ff fTF t
Ê nzz i& a £5$ §| S i i£ He
J II
Original
^fTT *T
^fncrr cr
PS ^ I S 5S
III
Notice that in neither the original combination (I + II + HI) nor in the derivative is
there any crossing of voices. But the middle voices of the original and deriv
ative (II + IIIV = ) continually cross. Such crossing gives more room for the voices,
without which they would be too cramped in their movements.
§ 311. The duplication of all three voices completes the study of three-voice vertical-
shifting counterpoint, and with it the study of vertical-shifting counterpoint in general.
PART TWO
DIVISION C
TWO-VOICE HORIZONTAL- AND DOUBLE-SHIFTING COUNTERPOINT

PART TWO
DIVISION C
TWO-VOICE HORIZONTAL-SHIFTING AND DOUBLE-SHIFTING COUNTERPOINT
CHAPTER XVI
GENERAL PRINCIPLES
§ 312. The vertical shifting of voices was investigated in Part One. The present
subject is that of combinations giving derivatives by means of the temporal shifting of
voices, i.e. by changing the relationship between the time-intervals of their entries. This
kind of shifting is called horizontal, and counterpoint admitting of it, horizontal-shifting
counterpoint. Double-shifting counterpoint is the use of the vertical and horizontal
shift at the same time. As this counterpoint is written by the same methods as horizontal-
shifting counterpoint alone the two will be dealt with together. In classifying complex
counterpoint in this way all possible cases of voice-shifting are exausted.
§ 313. To shift a melody by a whole measure, or by an intergral number of measures, does
not change it in any essential respect. But to shift it by some fraction of a measure—at a
half-measure, or a measure and a half—transfers the melody from one part of a measure
to another. This is the only change that a melody undergoes in the horizontal shift. All
other changes not resulting from either the vertical or the horizontal shift such as
augmentation, diminution, contrary motion, etc. are outside the province of shifting
counterpoint, though in the ensuing chapters they will be referred to when necessary
for explaining their connections with shifting counterpoint.
§ 314. Comparing the following two combinations-
Ex. 200 (*)
*E*EE uzzn ■ a rl * «» ~n-

(a)
3S m
(b) Ê^ JE
n—r. *=£ 3E HE
JOZ -o-
m m
it will be observed that each of them consists of the same two melodies (I + II) on the
same degrees. But the melodies enter differently: in (a) voice I enters first, in (b) voice
II. This concerns the horizontal shift. Regarding combination (b) as the derivative, it can
be investigated as the result of shifting either I or II, the voice not shifted not being
concerned in the process. In the first case the derivative results from shifting I two
measures to the left.
207
The Basic Version
§ 315. The possibility of a two-voice horizontal shift such as shown in Ex. 200 is
arrived at as follows:
An original is written in two-voice canonic imitation, and to it is added a third con
trapuntal part. From this three-voice combination, called the basic version, is taken the
counterpointing voice, first in union with the voice of initial entry, or voice
to be imitated (proposta) and next with the imitating voice (risposta). This was
how Ex. 200 was obtained. The voices of the basic version will be designated: proposta,
P (plural PP); risposta, R (plural RR); counterpoint Cp. In the basic version, Ex. 200
(b) the distance of entry between P and R is two measures, the entrance interval O.
Basic Version Palestrina

GO
mE= re zrzz f-±J-* WTT
££ HE
a b
" 6> 3E
2;l,f ■ -— a
From this are taken the two combinations Cp + P and Cp + R of Ex. 200 (a).
§ 316. Therefore the combination of voices that serves as the basis for horizontal-
shifting counterpoint, and of which the indispensable attribute is canonic imitation, is
termed the basic version. By means of it are solved problems in horizontal-shifting
counterpoint and, as explained later, in double-shifting counterpoint.
§ 317. For two-voice horizontal-shifting counterpoint with one derivative the basic
version consists of a two-voice imitation at the unison (P + R) and the counterpointing
voice (Cp). Hence two two-voice combinations result: Cp + P and Cp + R, of which the
first is taken as the original, the second as the derivative.
06s. -Two-voice canonic imitation is properly within the province of simple counter
point. With the exception of the infinite canon this imitation does not require the appli
cation of complex counterpoint. Its scope consists of: first, the writing of P as far as the
entry of R; this section of P is then transferred to the other voice, forming the beginning
of R. The voice that first enters continues as counterpoint with the' entry of R; this
continuing melody is also transferred to the later entering voice, etc.*
§ 318. A special case of the basic version is where P is silent at the entry of R, so
that instead of a three-voice combination only two voices are actually present:
Ex. 201 Palestrina
Cp
HE
E £
m. H.
j i aJ U *
±m
m^frw
'A more complete explanation is given in "The Technique of Canon, § 15. —Tr.
GENERAL PRINCIPLES 209
§ 319. In the basic version both the imitating voice (/?) and the counterpoising
voice (Cp) are written in simple counterpoint, as no application of complex counter
point is needed.
§ 320. Counterpoint that admits of horizontal shifting can hardly be regarded as
complete and independent in itself. Ordinarily it is a constituent part of a passage of
greater dimensions and for a larger number of voices.* Therefore it is not considered a
disadvantage when the voices that in the original combination end together with a ca
dence do not do so in the derivative, where in ending at different times one voice is
silent, as in Ex. 200 (b). It is assumed that such vacant places may be filled up by
other participating voices. If they do not, and the necessity arises to fill the vacancies
anyway, the voice about to end continues as long as conditions demand.
§ 321. Another matter is that combinations admitting of the horizontal shift, and
intended to enter into the structure of passages of larger dimensions, may in the basic
version be found where the two voices concerned do not make correct two-voice counter
point without a third supplementary voice. This fact impedes the free use of such combi
nations, and with their application in view it is first advisable to be clear as to what kind
of basic version may be used independently, which requires a supplementary voice, and
what is inadmissable. Hidden progressions and unprepared fourths need a lower sup
plementary voice; a ninth resolving to an octave ( 8 ) is not allowed at all (§ 83).
Basic Version and Imaginary Risposta
§ 322. As seen in the preceding section, a basic version consisting of correct three-
voice counterpoint may be composed to two-voice combinations that do not form correct
two-voice counterpoint. But a basic version, complete in itself, may present a combination
that is not contra puntally correct. A cantus firmus may be taken, transferred to another
voice, entered at any favorable time and at any interval that is not allowed between
the voices. In order to distinguish the risposta of such a fictitious imitation from the
more legitimate kind, it will be termed an imaginary risposta or imaginary voice, and will
be indicated by the initial letter and four dots: R To an imitation with imaginary
risposta is added Cp that forms a correct two-voice combination separately with P and
also with R , for example:
Ex. 202 (a)

Basic Version
Cp
r g d p r\ ,jjPrid [E
=H=
o "-
C.f. s
B....
Hence two combinations are obtained with horizontal-shifting voices: Cp + P and

Cp + /?...., of which the first is the original, the second the derivative:
•The infrequent cases where this counterpoint is independent (termed counterpoint with and
without rests) will be discussed in Division D.
(b) Original (Cp +P)
Obs.—In view of the fact that the voices of a basic version may not form a canonic
imitation that is contrapuntally valid, the term "basic version" is more suitable than
"basic combination." The word "combination" implies a relationship that is contra
puntally correct, while "version" may include a fortuitous relation of voices, such as
the preceding imitation with an imaginary R.
§ 323. The combination Cp + R will be termed the imaginary combination. In
contradistinction to imaginary, Cp + P will be termed the real combination, and each
voice entering into it a real voice. Both imaginary and real combinations are contra
puntally correct. On the contrary, P and R of the basic version do not combine, though
they may form correct counterpoint accidentally.
§ 324. The real combination corresponds to the original, the imaginary to
the derivative.
§ 325. To indicate which voice of the basic version is R (= imaginary voice), the
voice number will be followed by four dots. Therefore the formula of an imaginary com
bination shows two forms: I + II or I + II ; the voice unaccompanied by dots is Cp.
But such an expression as this is an incomplete formula. In writing a basic version
giving a predetermined horizontal shift it is not enough to know which voice has the
imaginary risposta. The number of measures or fractions thereof that the proposta must
be shifted to the right, in order to yield the risposta, must be known. This, relating to
the horizontal shifting of voices, will be discussed in the next chapter.
§ 326. Working out the Cp to a previously written P and R that does not form a
correct combination is a process which though unusual contributes materially to the
development of the polyphonic spirit. A problem in horizontal shifting where the cantus
firmus is given can be solved otherwise only if the cantus gives a correct canonic
imitation accidentally. Therefore, referring to problems on a cantus firmus (as in the
following section), it will be assumed that the corresponding basic version retains the
imaginary risposta and that P and /?.... of this version are written so as to adjust to
both. But there is another and easier method, giving better musical results; the basic
version, retaining the imaginary R, is divided vertically into parts that are written one
after another. This will be explained more fully in § 376.
Conditions of a Difficult Problem on a Cantus Firmus
§ 327. The degree of difficulty shown by a two-voice problem on a cantus firmus

with horizontal shift or with this and a vertical shift at the same time depends on the
nature of the intervals that happen to occur between P and R of the basic version. The
further apart the voices are— unison to octave—the easier the problem. Under these con
ditions, to write counterpoint against the two voices is almost equivalent to writing it
against one. The Cp offers no difficulty when there are consonances between P and R
Difficulties arise if among the consonances are found dissonances. As their number
increases embarrassments increase, especially if two or more dissonances come in
succession. Finally, parallel dissonances can make a problem entirely unworkable. For
instance, with a row of seconds or sevenths it is quite impossible to write a voice that
will make correct counterpoint with each voice of the dissonance. In working out exer
cises favorable cases should be chosen— consonances prevailing over dissonances—and
problems not attempted in which progressions such as these are found:
=eac
33^:
fp
r Ft *
Obs.— It is necessary to remember that these statements of difficult conditions refer
only to those versions where the proposta functions as the given voice, and P and /?....
are written beforehand, together with Cp. These conditions have no meaning for those
basic versions where, in the absence of a given voice, P and /?.... are written in sections,
as well as Cp. More will be said about this later (§ 376). Basic versions written ac
cording to these methods uniformly use intervals that are admissable between P and R
Basic Version for Double-shifting Counterpoint
§ 328. From a basic version with canonic imitation at the unison (with either real
or imaginary risposta) is obtained two-voice counterpoint giving a derivative by means
of only a horizontal shift. If an imitation at some other interval is taken for the basic
version and its associated Cp, then from such basic version is obtained the original
combination Cp + P with a derivative Cp + k (or Cp + R....), representing a simultaneous
horizontal and vertical shift of the voices. Such counterpoint is termed double-
shifting (§ 312).
§ 329. In a basic version for double-shifting counterpoint Cp may appear as either
outer or inner voice. If Cp is an outer voice the derivative combination will show the
direct shift (* ~~* ); if Cp is an inner voice, the inverse shift ( £>-«C^).
§ 330. A basic version for double-shifting counterpoint requires no changes in the

rules of vertical-shifting counterpoint; Cp is written with P and R (or R....)
in simple counterpoint.
§ 331. Exs. 203 and 204, each in double-shifting counterpoint illustrating a basic
version with real R, are taken from the works of Palestrina. In both of them Cp is an
outer voice, therefore the vertical shift is direct (§ 329).
Ex. 203 (a) Basic Version Palestrina
SE ~rr S sUfc ^ JEk rn

73T T3T
Cp
IP^ m m
R imitates P at the third above and enters two measures later. P + R, constituting
part of a correct three-voice texture, do not form correct two-voice counterpoint by them
selves, because of the two unprepared fourths. They require a lower supplementary
voice (§ 321). From this basic version is obtained:
Original (Cp +P)

0>)_ 4-
» 3E
■xr
* t £=Ê
r^^fr-nr^f rr etc.
ii
Derivative (Cp +R)
\>t -
aU± 3E
etc.
ii M 'I ' "i rrrff

The derivative is obtained by shifting I two measures to the right and a third higher,
or what amounts to the same thing, shifting II two measures to the left (§ 413) and I
a third higher.
In the second measure of Ex. 204 is found a hidden fifth and the free entry of a fourth
between P and R, therefore the combination P + R requires a supplementary voice. The
derivative was obtained by shifting I a half-measure to the right and a fourth lower (or
a half-measure to the left and I a fourth lower).
Ex. 204 Basic Version Palestrina
Original (Cp +P) Derivative (Cp +R)
3E
'i- i" ±M& *
*
4 3E
"Hrrrrrr !■'•«*
* «- iif rrrrrf etc.
The combination Cp + R, from the basic version, is found three times in Palestrina,
each time transferred to other degrees and with long note values in the upper voice:
IE
JL
nn
*
In two of these instances both voices shift at the half-measure (§ 335):
i^
§ 332. In the two preceding examples Cp appears as the lower voice, therefore the
shift in the derivative is direct. In the next example Cp is an inner voice, hence the
shift is inverse (§ 329). /?.... enters one measure later than P, at the seventh above:
Ex. 205 Basic Version
ipsE
Original (Cp +P)

■■ û j
m "&C ¥
Derivative (tp +H)
X -».
-
In the derivative the c.f. is shifted one measure to the right and a seventh higher.
§ 333. In the preceding example the c.f. in the derivative appeared as shifted to
another degree. In order to retain it on the same degrees as in the original the Cp would
have to be shifted a seventh lower. Such a shift, not intended when the basic version was
written, might easily result in an augmented fourth or a diminished fifth, f-b,in the third
and fourth measure:
& JE etc.
W ■ f r r r r r
Such situations can be avoided as follows: in the basic version take for the imaginary
voice a key-signature that gives the first note the same relative position in the scale
as the first note of the c.f., using a separate staff for the purpose. For instance the c.f.
of the basic version of Ex. 205 begins on g, the fifth degree of C major, and in the
imaginary voice it begins on f. To make this f the fifth degree of a major scale a signa
ture of two flats must be taken for the imaginary voice, as in Ex. 206. Here, in writing
Cp simultaneously with P and r? its relation to P must be according to C major and
to R according to B-flat:
Ex. 206 (a)
Basic Version
Cp _
The derivative Cp + R must be transposed so that its key-signature is the same
as that of the original; the c.f. then appears on the same degree:
Original (Cp +P) ,
m 3E
s *
Derivative (Transposed Cp + R....)
x: II.
This process is used in some of the problems in Appendix C.
Shifting Voices from One Part of the Measure to Another
§ 334. In Ex. 203 was an instance of where a voice was shifted from one part of the
measure to another; this happens when a shift is made at some fractional part of a
measure, such as a half-measure, one and one-half measures, a third-measure, etc. A
voice of the basic version must be so written that when it is shifted to another part of
the measure no rhythmical errors will result. For instance, when a shift is made at the
half-measure it is bad to tie over a note that occupies a whole measure: .
4l~° 4 °I P ' because these give incorrect ties between two notes of which
the first is shorter than the second: A ^TL \A - |U« 4j'To •
In triple time it is also bad to shift a long tied note one-third of a measure to the
bad
right: V O■ Q WOO 3 O* O ; since ties give:
The same thing happens when a shift is made to the third beat-
bad T» It0* Hf
since these ties give 3 q'Tq 1 and jfl p q* .
The following rule is recommended for avoiding the rhythmical figure PrP
shifting at the half-measure do not use the rhythmical sucession ' » » 1 p » • I S-
Obs.— Exceptions to this rule are sometimes found in the strict style:
Ex. 207 ■» Palestrina
BH Et
O O
in- car-na
m
tus est
Ex. 208
y^ # I
v i . j I , Ibid.
Te - su Chri
Je - ..-..- ste
§ 335. When two voices at Jv= 0 are shifted from one part of a measure to another,
forming a contrapuntal combination which except for rhythmical conditions refers to each
voice separately, some conditions must be considered that refer to their united effect.
These conditions have reference to the place that suspensions and their resolutions
occupy in the measure.
The resolution of a prepared dissonance must not appear at any part of a measure
which would cause the resolution when shifted to come on an accented beat. For instance
in shifting at a half-measure it would be bad to write: -<p_

,h>J 3 rJ , since the
IE
r t
dissonance would appear on the weak half of the measure and the resolution on the accent:
if | ~ n q . On the contrary a resolution of the same dissonance on the

rHT-î
second quarter: 3f£ is entirely practicable, for shifting at the half-mea-
rfrr
sure gives: .). J .1
T?=f
Also in triple time with a shift at one-third of a measure to the right it would be bad
to write: I ^ , which would give:

±± ; shifting it at
> KJ^ g
two-thirds ot
of a measure:
measure 8 — : would give:%— "'
When a shift is made at the half-measure a passing-note cannot be used on the third
quarter: *P a _ because when shifted it would come on the accented beat:
m
r r,i rf
Guided by such considerations it is easy to understand that similar situations must
be avoided in the basic version when shifts are made at one-quarter of a measure, or
three-quarters, etc.
Functions of the Basic Version
§ 336. By means of a basic version are obtained the derivative combinations for
horizontal- and double-shifting counterpoint (§ 328). But the functions of a basic version
are not limited to these. Its application to other categories of complex counterpoint will
herewith be examined.
§ 337. Of the two imitating voices in a basic version P may be considered as a voice
not shifting, and R as a voice shifting horizontally. At each shift of this voice a new
problem arises that requires a new basic version. But at all shifts the fundamental
problem is one and the same: To two voices that imitate must be added Cp that forms
correct two-voice counterpoint with each of them. Shifting R in the direction of P can
bring it in such a position that their entries coincide, for the term imitation, in its
generally accepted meaning, applies to these two voices. A further shift of R in the
same direction makes it the earlier entering voice, i.e. R becomes P and the previous
P is changed to R.
Returning to the case where the entries of P and R coincide, if at such a shift R is
found on the same degrees as P, simple counterpoint results, for because of the complete
coincidence of the two voices Cp is written not to two voices but orily to one. If the
voice that shifts horizontally also shifts vertically at the same time to other degrees,
chen at the coinciding entries is obtained vertical-shifting counterpoint; of the two
combinations Cp + P and Cp + R one is taken as the original, the other as the derivative.
Therefore, to the two functions of the basic version already known may be added a third:
it may serve also as a basis for vertical-shifting counterpoint. But in practice this is
of little utility; it is mentioned only as one of the functions of the basic version, not
with any idea of recommending its use.
Obs.—The coincidence mentioned—the entries of P and R in the basic version—de
pends on the possibility of a consonant and uniform relation between the intervals of
the original and derivative, thereby allowing the application of the general rules gov
erning intervals, the shifting of voices, etc. Hence arises the study of vertical-shifting
counterpoint. By mastering this study a command over tonal combinations and a facility
in writing derivatives will be acquired that can be obtained by no other means. Counter
point textbooks often recommend the direct solving of problems by the help of supple
mentary voices, without previous study of double counterpoint. This method is almost
worthless, as the basic version in which the entries of P and R coincide include also Cp
as middle voice, agreeing with the other two, and forming correct counterpoint with each
of them individually. Superficially this method of applying vertical-shifting counterpoint
would appear to be the simplest, leading at once to the desired result and requiring no
preliminary training except a knowledge of simple counterpoint. This delusion could be
kept up as long as easy problems only were attempted. But passing to problems with
difficult indices, or to those combining the conditions of several indices, or to the
simultaneous shift of three voices, etc., the fallacy of the method would at once become
apparent. Having no other resources than supplementary voices, the situation would be
hopeless, and all attempts to solve such problems would in most cases end in failure.
§ 338. In a two-voice imitation R can reproduce P in a variety of ways; in augmen
tation or diminution, in contrary motion, etc. Such imitations are written in simple coun
terpoint, according to the method described in § 317, Obs. Applying these exceptional
forms of imitation to the basic version, derivatives are obtained where one appears in
augmentation, diminution, contrary motion, etc.
§ 339. In the derivative of the following example I gives notes of twice the value
of those in the original:
Ex. 209 (a)

Original
j g j jj ° *¥+^. m gag i
m 'T*;
m m
m i 3
m 5a= 1 Eg mm gg: -p-'-p-t
TT-
3^25
^^3
This counterpoint is taken from a basic version where R imitates P in augmentation:
(b) Derivative j (in augmentation)
w&II
wsm
^£ «—» =*^
*= ^^
zzz£ nz
K %b, pPm -e-
=£
72= — =t32=I*ZZZ
zee
*3:
>l.^ tar
#--"-!► s . P' »
? -s*
„ (in augmentation)
rPE ^
Basic Version
^ # etc.
IR-c
S
Cp
An example in diminution would be obtained by taking the derivative as the original.

In the next example I in the derivative is- a shift of the same voice in the original,
one measure to the right, and in contrary motion.
Ex. 210
Original
Derivative
41- -^r
h 3*E
F#
m s
x J (in contrary motion)
s m m rEi\ ew
^
Basic Version
m
-e«- ^*- etc.
(in contrary motion)

R....
^
^^
This combination is taken from the following basic version, which, like the preceding,
was written according to the method indicated in § 376.
The exceptional forms of imitation may be united in one voice; for example R may
imitate in augmentation and contrary motion at the same time, giving a corresponding
imitation of this voice in the derivative.
Obs.—li it is necessary that both voices of the derivative shall be taken in contrary
motion the problem can not be solved by means of a basic version. It refers to the
province of metamorphosed counterpoint, having special conditions. The use of excep
tional forms of imitation in the basic version establishes the boundaries that shifting
counterpoint can explore. The matter is referred to here only for the purpose of explaining
the functions of the basic version in complex counterpoint in general.
§ 340. The student who has assimilated the comprehensive materal in this chapter
will have acquired the technique necessary for the successful handling of a wide variety
of derivatives that can be applied in multi-voice work— inventions, fugues, compositions
with text, etc.* For such practical purposes the knowledge acquired is sufficient.
But for the solving of problems with a predetermined shift of voices, for finding
their corresponding basic versions, and to prepare for the study of three-voice horizontal-
shifting counterpoint, it is first necessary to master the contents of the next chapter.
•Examples of the horizontal shift in contrary motion will be found at the end of this Division.
Cf. also Habert, "Die Lehre von der Nachamung," § 13.
CHAPTER XVII
NOTATION OF HORIZONTAL-SHIFTING VOICES; POSITIVE AND

NEGATIVE MEASURES; THE SHIFTS; INDEX FOR HORIZONTAL-SHIFTING
COUNTERPOINT; FORMULAS; METHODS OF WORKING OUT EXERCISES
§ 341. The concepts introduced into the study of vertical-shifting counterpoint—ad

dition, subtraction, positive and negative shifts, contrapuntal indices etc.—are found
also in the study of horizontal-shifting counterpoint.
§ 342. In the vertical shift the unit of distance is the second; in the horizontal
shift (right and left, referring to the notation) the measure is the unit of distance. When
a voice is shifted from one part of a measure to another the necessity was seen for
using fractions. This concerned the metre in which a given combination was written;
shifting at a half-measure, or at one and one-half measures etc. referred to duple time; at
a third or at two-thirds of a measure, to triple time.
§ 343. The voices of an original combination are indicated according to the order of
their entries; I is the earlier entering voice, II the later. In adding any number of meas
ures to the distance between their entries I may be advanced to the left or II to the right.
Subtraction causes each to shift in the opposite direction. In other words, I is the voice
for which positive shift is to the left, II the voice for which the positive shift is to the
right. Expressed in a diagram: ++- I -». — +- II-* +
§ 344. Shifting horizontally, a voice may either recede from the other or approach
it, or even take up such a position that II will precede I. The distance (temporal) between
voices that enter in the order II-I is termed negative. In this way measures will be
positive or negative, analogous to the positive and negative intervals of vertical-
shifting counterpoint.
§ 345. The treatment of measures as positive and negative allows the distinction of
direct and inverse shifts also in horizontal-shifting counterpoint. At the direct shift
the quantities that define the distance between the entries of voices in original and
derivative combinations, in terms of measures or parts of measures, have the same
signs; at the inverse shift the opposite signs (cf. § 25).
06s.—The signs and ^X^ refer only to the vertical shifts of the voices.
§ 346. If the voices of the original combination enter either in the order I-II or II-I, in
both cases the upper voice (I) is considered as the voice of earliest entry, in the sense
that shifting it to the left is a positive operation; and the lower voice (II) as that of
later entry, with its positive shift to the right. The only difference is that in the first
case the distance between the entries is a positive number of measures, in the second,
a negative number.
Obs.—This method removes the difficulties that arise when the lower voice enters
before the upper. As voice of first entry it would have to be indicated by the figure I, as
lower voice by II, besides which it would require a special sign, as these refer to both
the vertical and the horizontal shift.
§ 347. The following diagram shows the positive and negative shifts of upper (I) and
lower (II) voices in both vertical and horizontal directions:
219
I-+ -4- II
+
Each voice retains its sign in the derivative. The voices of the original combination,
I + II, will be regarded as before, in vertical order, so that positive intervals are obtained
between them, i.e. I is in fact the upper voice, II the lower. As for their relative positions
in horizontal order, the distance between their entries in the original combination can
only be positive (when the voices enter in the order I-II), or negative (if the order of
entry is IH), or equal to zero (if both voices enter simultaneously).
Notation of the Horizontal Shift: h, hh;

Formulas for the Derivative Combination
§ 348. The horizontal shift of a single voice will be indicated by the letter h (plural
hh). The quantity indicating the direction and number of measures in the shift is united
to h by the sign of equality and placed after the roman numeral instead of v. For exam
ple the expression in_1 + Hn~'2 means that both voices shift to the left, the upper one
measure, the lower a half-measure; I '2 + II ^ means that the upper voice shifts
one and one-half measures to the left, the lower two measures to the right, etc. The
sign + means that the voices shifted form correct counterpoint. The preceding expressions
are formulas of derivative combinations for horizontal-shifting counterpoint (cf. § 21).
Obs.—In horizontal-shifting counterpoint vertical shifts are excluded only at Jv=0.
§ 349. A voice that does not shift horizontally is represented by its figure with
either h omitted or with h- 0.
§ 350. In the formula for the derivative combination for double-shifting counterpoint
the letter v is placed underneath h. For example the formula of the derivative combination
h= 0 h=l
I +11 means that the upper voice (I) does not shift horizontally, but shifts
vertically a ninth below; the lower voice (II) shifts one measure to the right and at the
h=-l h=0
same time a sixth below. The formula Iv_ u + IIV means a shift of the upper voice
(I) one measure to the right and the lower (II) a fourth below, etc.
Index for Horizontal-Shifting Counterpoint: Jh
§ 351. To define the distance between the entries of the voices of the derivative
the algebraic sum of their hh is added to the number of measures in the original dis
tance (cf. § 26).
§ 352. The algebraic sum hh of two voices is called the index of horizontal-shifting
counterpoint; similar to vertical-shifting; it is indicated Jh; Jhh (cf. § 27).
§ 353. Jh may be added to the original distance between any two notes (cf. § 26). As
a result the derivative distance is obtained between the corresponding notes.
§ 354. To indicate what shift of voices is made at a given Jh the formula of the
derivative combination is enclosed in parentheses and after it is placed Jh, e.g.:
(I +H ) Jh- VA. If theh=formula
0 also
h= 1 includes a vertical shift, after the Vparenth-
eses first comes Jh, then Jv: (Iv_ ~8 + IIV- 5) /*= 1, Jv= -3.
NOTATION OF HORIZONTAL-SHIFTING VOICES 221
§ 355. If to a horizontal shift is added a vertical shift at Jv- 0 the result has to do
with horizontal shifting only, not with double shifting (06s. ,§ 348).
§ 356. If one and the same combination admits of shifts at two or more J Jh the
corresponding figures, separated by commas, follow the sign of equality. Such a Jh,
referring to one original and more than one derivative, is called a compound index. It
may be double, triple etc., according to how many simple indices are present (cf. § 28).
For instance, Jh- 1, %; ]h- 2, —2 are double indices; Jh- 1, J4, —1 is a triple index, etc.
§ 357. A shift of voices I + II at a given Jh may be substituted for another and give
as a result the same distance between entries, if the algebraic sum of the shifts remains
without change (cf. § 29). A voice may appear as shifted to other parts of measures,
depending upon the value of h of the corresponding fraction; this change is the only one
inserted into the horizontal shifting of a melody.
§ 358. If for one of the voices h- 0, then for the other Jh- h. In other words, when
only one voice is shifted Jh~ h of that voice. This follows from the fact that Jh equals
the algebraic sum hh of both voices (§ 352).
§ 359. If a shift of one voice is substituted for a shift of the other with the same
value (positive or negative) the distance between entries is in both cases identical.
Therefore the expression in = l + n and I + II are equivalent. It is the same if h is
equal to a fraction; whichever voice it represents is merely moved from one part of
a measure to another.
The Formula a + Jh= b
§ 360. Adding Jh to the number of measures equal to the original distance yields the
derivative distance (§ 351). Indicating the original distance by a and the corresponding
derivative distance by b, the formula (cf. § 31) is expressed by the equation a + Jh- b.
Hence the value of Jh is determined by the equation Jh- b — a, i.e. Jh is equal to the
derivative distance, to which the original is added, taken with the opposite sign.
Conditions of the Shifts
§ 361. Jh may be either a positive or a negative quantity or may equal zero (cf. § 24).
At Jh- 0, after shifting both voices of the original combination the distance between
their entries in the derivative remains unchanged, even though the voices may shift
from one part of the measure to another.
§ 362. The conditions of the shifts follow from the equation a + Jh = b (§360) and
from the statements in § 343 (cf. § 34):
(1) If a and Jh are both positive or both negative the shift is direct.
(2) If of the two values of a and Jh one is positive and the other negative, then—
(a) If the absolute value of a is greater than that of Jh the shift is direct;
(b) If the absolute value of a is less than that of Jh the shift is inverse;
(c) If the absolute value of a equals that of Jh the result is 0, i.e. the coincidental
entries of the voices.
§ 363. From the equation a + Jh- b it follows: it a- 0, b- Jh.
§ 364. Any combination in simple counterpoint may be regarded as the result of a
horizontal shift of the voices at Jh- 0 (cf. § 44). In general, Jh in horizontal-shifting
counterpoint has the same functions that Jv does in vertical-shifting.
Application of h and v to the Basic Version;
Formula of the Imaginary Combination
§ 365. By means of the sign h can be exactly designated by how many measures in
the basic version R (or R , § 322) stands from P.
§ 366. R is only a shift of P to the right. Therefore h is a positive quantity if P
(therefore also R) is voice II (P II, R II); and vice versa, at P I (and R I) h is a negative
quantity (§ 343). Indicating by a the number of measures that R shifts relative to P, the
combination Cp + R, taken as the derivative, shows two phases:
Cpl+R IIh= a or R Ih= _a + Cp II.
Obs.—The Roman numerals I and II designate the upper and lower voices (§ 347).
Therefore P I indicates P, forming the upper voice of the original combination. Cp II
indicates Cp, consisting of the lower voice of this combination, etc. Adding the Roman
numerals to P, R, and Cp enables the shifts applied to these voices to be indicated.
§ 367. If R is imaginary (R ) the preceding formula takes these forms:
l!\7.-a + II or I + II^T.3;
they are called formulas of imaginary combinations (§ 323).

Obs.—The formula of an imaginary combination in two-voice counterpoint is at the
same time the formula of the basic version according to its construction. But in three-
voice counterpoint, as will be explained in Division D, three formulas are necessary
for the purpose of the imaginary combination; united, they constitute the formula of
the basic version.
§ 368. To simplify and generalize to the furthest extent, use will be made of the
formula for the imaginary combination for every basic version, irrespective as to whether
its R is imaginary or not. This is entirely admissable, because the actual R may be
regarded, like the imaginary, as fortuitously forming correct counterpoint with P (§ 323).
Stated in this way, the formula for the imaginary combination, representing every
factor in the basic version, does not determine beforehand whether the latter is written
with the real or the imaginary R.
§ 369. For double-shifting counterpoint the formula of the imaginary combination
must indicate at what interval R shifts— above or below—according to its relationship
to P. Indicating this interval by b, the two views of the formula are:
h= a h=~a
I + Iiy.7. ±b and iy.7. ±b + II.
In this way, for the basic version, Ex. 203 (a)—
h--2 Palestrina
-e e-
»
P\
PP etc.
IM " \
Cp II
h=-2
the formula of the imaginary combination is IV.T.. * + II.
For Ex. 204 (a)-
P\
* 331
It
w etc.
^^
Cp
wm
hzzli2
the formula of the imaginary combination is iY... + II.
For Ex. 205 (a)-
etc.
h= 1
it is I + Iiy.7.-,6 etc.
§ 370. The imaginary combination corresponds to the derivative (§ 324). This cor
respondence obtains in the fact that the distance apart at which the voices enter, also
the order of intervals formed by their union, are identical in both combinations. In other
words, both imaginary and derivative combinations show a shift of the original at Jh
and Jv alike. Both imaginary and derivative combinations may, therefore, be regarded
as a shift of the other at Jh- 0 or at Jv- 0.
General Formula of Horizontal-Shifting Counterpoint
§ 371. The formula of the derivative combination states the requirements of a given
problem; the formula of the imaginary combination indicates the kind of basic version
that solves the problem but does not determine in advance how it shall be written (§ 368).
Therefore in solving problems the formula of the derivative combination must be con
verted into the formula of the imaginary. Since Jh is identical in both combinations
(§ 370) and in the combination Cp +/?.... only one voice shifts: R , and h- 0 for the
other voice; thus R has h- Jh (§ 358). Therefore—
cP + R^r.Jh
is the general formula for horizontal-shifting counterpoint, solving all problems in

this category.
§ 372. Corollary to this, whether h (therefore also Jh) is a positive or a negative

value, the formula may be double (§ 367):
I + II.h.T./A= a or l^=.Ih= ~a + II.
Obs.—Since the interval of imitation in the basic version for horizontal-shifting

counterpoint is the unison (0), in the formula for the imaginary combination of this
counterpoint t>= 0 is always understood.
General Formula of Double-Shifting Counterpoint
§ 373. To convert the formula of the derivative into that of the imaginary combination
t>= Jv is added to the imaginary R, in the same way that h = Jh was (§ 371). Hence the
general formula for double-shifting counterpoint is:
h-/A
CP + R?.7.Jv.
§ 374. This also has two forms (§ 372):

h= Jh = a h= Jh = -a
I + 11 v= Jv or I v= Jv + II.
§ 375. Therefore, for the solution of problems in horizontal-shifting or in double-

shifting counterpoint it is necessary that the problem be stated as a formula for the
derivative (§ 348), completed by Jh and Jv; hence a positive or a negative Jh can be
taken for the corresponding formula of the imaginary combination (§ 367), writing in
h= Jh and t>= Jv for the imaginary voice (§§ 371, 373). Examples of formulas are given
at the beginning of each problem in Appendix C.
Obs.— The original combinations (hence also derivatives) studied have consisted of
two different melodies. But it also happens that the original may consists of only one
melody, imitated canonically. Such an original, as a special case of the combination
I + II, will be examined, indicating each voice by its figure, and the shift in the deriv
ative. The following is from Palestrina:
Ex. 211
Original Palestrina
£ 3E ^
fe crn
-O-
J,U rr
II TT W
Derivative
,h:0
Li ^JU
b)
f IE 3E nr: znc
II
h=i ♦ ¥¥ 3L
h= 0 h= 1
(IV_ 7+ Hv_ -*) Jh- 1, Jv- 0; formula of the imaginary combination: I + II^T.1. Cc
responding to this formula is a basic version representing a three-voice canon:
Basic Version
c)
1
-»
J J , O O n J J 9. ■ £ J
i
3E 3E =&:
n ^
" o:
l
¥? n
S h:l (8)
ii
The combination I + II is the original (a), and I + II 1» shifted an octave higher, the
derivative (b). The basic version, though contrapuntally correct, is not found in Pa-
lestrina, only the two-voice combinations (a) and (b). In view of the fact that three-voice
canonic imitation can not be explained here, future examples will be limited to where
voices I + II represent two different melodies.
Methods of Working Out Problems
§ 376. Two methods have been indicated of writing basic versions— with real and
with imaginary R. In both cases Cp correlates to the two voices, previously written, of
the basic version. A third method is possible, not utilizing previously written voices
and therefore applicable only to problems without a given voice. This method consists
in writing all three voices—P, R......and Cp— by sections in definite order (v. diagram
below). The whole problem is broken up into columns, each of which represents the
distance from the beginning of the exercise to the entry of the imaginary voice. Melody
P is represented by the bold-face figures A, B, C, etc. Underneath these Cp is indicated
by the corresponding small letters. The divisions of ft..... .are indicated by the same
capital letters as those for P, each letter followed by dots: A...... B , C , etc. A
problem is worked out as follows:
First column 1 is written (Cp + P). Next the section written for P is transferred to the
second column, where it forms the beginning of the imaginary voice. With the entry of
R each succeeding column is written in the same order, beginning with the imaginary
combination (i.e. Cp is written to R ), which thereafter is real; to Cp is written P,
which is then transferred to the third column, etc. The diagonal arrows from each column
to the next show the trans ferrence of P to the imaginary voice. The curved arrows inside
the columns show the order in which the melodies are count erpointed to one another; the
voice to which an arrow points is written as counterpoint to that from which
the arrow starts:
1) 2) 3 4) 5 6
JR...
A" A- A- A-
P: a/
a
In going from one section to another care must be taken that the melody of a given
section must each time form a natural continuation of the melody in the preceding section.
§ 377. In the following illustration (the beginning of Example No. 2 in Appendix
C, formula of the imaginary combination In = ~2 + II)the divisions into sections are lettered
in accordance with the preceding diagram. The composition of the basic version can
here be traced step by step.
i!'.v* !*ID
etc.
etc.
etc.
Obs.— In this illustration, also in the examples in Appendix C, the indications P,

ft.... and Cp axe omitted. The voice forming Cp is indicated by a star to the left of the
figure (*II). After the figure for the imaginary voice is placed in parentheses the figure
of the voice counterpointing with it: Ir.T. (+ II).
§ 378. Therefore there are three methods of working out problems in horizontal-
shifting and double-shifting counterpoint.
(1) In the first method, to a basic version that imitates canonically is written a
contrapuntal part that is first of all useful in combination with it; this has the advantage
over the other methods in that when applied the basic version not only gives two-voice
combinations of horizontal-shifting melodies but also presents them enriched by new
contrapuntal material. The disadvantage of this method is that a two-voice combination
extracted from three-voice counterpoint may not sound full enough without the assistance
of a supplementary voice.
(2) In the second method a basic version with /?.... is previously written with P, to
which Cp is adjusted. This is a difficult and mechanical method, though useful from a
purely technical point of view, and indispensible in solving a problem with a voice
given beforehand. Otherwise, if such a problem does not show just what kind of horizontal
shift is required, this method can be avoided and a search made for such entry of the
imitating voice as will yield a real and not an imaginary R. In this event the problem
is solved by the first method.
(3) As for the third method, explained in § 376, giving both combinations-a two-voice
original and its derivative-it is hardly ever applied in composition.* Its principal use
is for the working of exercises, for which purpose it excels both other methods in the
extreme ease with which it employs all possible shifts of three voices, giving two-voice
combinations that sound full and do not require supplementary parts. The examples in
Appendix C were written by this method. The same process, but more complicated and
using a larger number of voices, will be applied in three-voice horizontal-
shifting counterpoint.
Obs. -Exs. 209 and 210 were written by this method. In applying it to imitation in
augmentation each successive section is as many times the length of the preceding as
the increased values of the imitating voice.
•Because it is unknown? Method (3) has unrealized possibilities. —Tr.

CHAPTER XVIII
INTERCHANGE OF ORIGINAL AND DERIVATIVE;

PROBLEMS ON A CANTUS FIRMUS
Interchange of Original and Derivative Combinations

in Horizontal-Shifting Counterpoint
§ 379. When the original and derivative combinations are interchanged (cf. § 68) the
formula of the derivative, and corresponding to it, the formula of the imaginary com
bination, is altered as follows:
(1) In the formula of the derivative combination h of each voice, therefore also Jh,
change their signs (+ or — ) to the opposite.
(2) In the formula of the imaginary combination, because of the conversion of Jh to
positive or negative or vice versa, h of the imaginary voice, equal to Jh, also requires
the opposite sign and therefore is transferred to the other voice (§ 372), which becomes
the imaginary voice of the new formula. Since R (therefore also P) may be either of the
two voices (depending on which combination is taken as the original), for horizontal-
shifting counterpoint with one derivative there are two basic versions available for the
solving of problems.
06s.—This does not refer to those combinations where one of the voices is a c.f.; in
this case a problem is solved by the use of only one basic version, namely that in which
P serves as c.f. (§ 326).
Problem No. 1 in Appendix C has the formula (I 1 + II) Jh- 1 for the derivative
combination; hence the formula of the imaginary combination is: I + Ilr.T. .
The beginning of the basic version (cf. 06s., § 377) is as follows: (I + II is the
original, I + II?.T. the derivative):
*I i xs: etc.
II
»
ffiftft-
w etc.
etc.
•Taking the derivative for the original, the signs for the derivative formula are changed:
(I + II) Jh- -1 and a new formula obtained for the imaginary combination: ilVr.-* +
II. This formula corresponds to the following second basic version (original I + II,
derivative Ih.T.-1 + II):
ih.:lMi.
227
Interchange of Original and Derivative

in Double-Shifting Counterpoint
§ 380. Here two cases are to be distinguished, according to whether Jv has the
direct or the inverse shift.
§ 381. Case l.-Jv has the direct shift: ,_ _, (Cp is an outer voice of the basic
version, § 329). The rules concerning h and Jh (§379, [1]) are extended to include also
v and Jv, viz.:
(1) In the formula of the derivative combination the signs for h and v of each voice,
therefore Jh and Jv, are changed to the opposite.
(2) In the formula of the imaginary combination h and v of the imaginary voice change
their signs, these being transferred to the other voice, which now stands as the imaginary
voice of the new formula.
Problem No. 6 in Appendix C has the following formula:
h= 1 h= 0 h= 1
(,v= 0 + nv- 1) Jh= 1, Jv- 1; I + Iiy.T.1.
The formula of the derivative combination at the exchange of derivative and original
is converted to:
h= -1 h= 0
(Iv_ 0 + nv--l)A=_li/t;=_1.
h = -1
hence the new formula of the imaginary combination is: I. .7."" + II.
Corresponding to this formula is the beginning of the second basic version:
As to the key-signature of the imaginary voice cf. § 333.

§ 382. Case 2—Jv has the inverse shift: £x^ (Cp is the inner voice of the basic
version, § 379). Exchanging original and derivative combinations:
(1) In the formula of the derivative h and v of voice I are transferred with their signs
to voice II, and h and v of voice II to I; Jh and Jv therefore remain without change, hence:
(2) The formula of the imaginary combination remains without change.
Exchanging original and derivative in example No. 7, Appendix C, the former I now
appears as II and vice versa:
Original (former derivative)

INTERCHANGE OF ORIGINAL AND DERIVATIVE 229
The previous formula of the derivative was:

h= 0 h= -1
(IV~ "6 + HV~ ~2) Jh- -1, Jv= -8;
converted to:
h= -1 h= 0
av_ _2 + nv- -6} Jh- _lf Jv= _g.
Jh and Jv remaining as before. The formula of the imaginary combination is the same
h = —1
for both: lY.7. + II. It must not be forgotten that I refers now to the former II and II to
the former I. Corresponding to the new formula is the second basic version:
§ 383. Therefore to obtain the second basic version the previous derivative is taken
as the original combination. If the vertical shift is direct ( * * ) the formula of the
imaginary combination changes thus: h and v with change of signs (+ and —) are trans
ferred to the other voice, which now functions as the imaginary voice of the new formula.
At the inverse shift ( ^XC~ ) the formula of the imaginary combination remains
without change.
Problems on a Cantus Firmus
§ 384. In problems on a c.f. the method described in § 376 of composing the basic
version cannot be used, since it does not admit of the writing of a voice in advance.
These problems are written with the aid of a basic version in which the c.f. forms the
proposta, and when shifted, the imaginary risposta. To these voices, previously written,
is then added Cp. Therefore only one basic version is possible for a problem
on a cantus firmus.
Obs.—Strictly speaking, a second basic version is possible for a problem on a c.f.,
but for this canonic imitation must be used—a very difficult matter and one not related
to the subject under discussion.*
§ 385. Of the two voices of the original combination, if the c.f. is the upper (c.f. I)
and in the formula of the derivative Jh is negative; or if the c.f. is the lower voice
(c.f. II) and Jh is positive, the problem is no different from the preceding; the c.f. shifts
to the right at h- Jh, forming /?.... , the usual method of composing the formula of the
imaginary combination according to the basic version.
§ 386. If the c.f. and Jh are found in the reverse relationship, that is, c.f. I at a
positive Jh or c.f. II at a negative, the c.f. must shift to the left and therefore cannot
form R In this case the derivative combination must temporarily be taken for the
original, consequently its c.f. and Jh stand in the same relationship as that described
in § 385, and the problem is solved in the usual way. Taking the derivative for the
original and writing the basic version according to the formula of the imaginary com
bination, the original must be considered not as Cp + P but as Cp + R . If the c.f.
appears on other degrees or on other parts of the measure than those given, all com-
*It is dealt with in the author's sequel to the present work, "The Technique of Canon," -Tr.
binations shift in vertical and horizontal directions (at ]h- 0 and at /t>= 0), in such a
way that the c.f. appears as at first. The derivative combination is arrived at by applying
the requirements of the problem to the original by this method.
(1) Direct Shift.-
h= 1 h= 0
,= -1
(C.f. I v=0 + 11 )/*=l,Jb--l.
In view of the fact that the c.f. is I and Jh is a positive quantity (§ 386), the original
is replaced by the derivative. fh= 1 and Jv= -1 are changed to Jh= -1 and Jv= 1 (§ 381).
Hence the formula of the imaginary combination is:
h= —1
C.f. IV.T. 1 + II.
The basic version corresponding to this formula is:
Ex. 212 (a) Basic Version
Taking the imaginary combination (Cp + /?.....) in the capacity of original, it is

shifted so that the c.f. appears on the appropriate degrees, and the desired original
is obtained:
Original
(b) en
Shifting the voices of the original according to the requirements of the problem:
h= 1 h= 0
C.f. Iv~ ° + IIV- -1, yields-
Derivative
06s. -In the preceding basic version the c.f. (= P) was taken in the register specified
in the problem. But this is not essential; it may be transferred to other octaves, taken
with other key-signatures, etc., whatever is desired. For instances, in the basic
version of the following problem the c.f. is taken an eleventh higher, otherwise the
imaginary voice would appear too low.
(2) Inverse Shift.-
INTERCHANGE OF ORIGINAL AND DERIVATIVE 231
h= -V/2 h= 1 h=-h
Sift + II.
(Iv- "2 + c.f. IIV- ~n) Jh= -%, Jv= -13; iy.T.-AJ
In view of the fact that the c.f. is II and that fh is negative (§ 386), replacing the
original by the derivative does not change the formula of the imaginary com
bination (§ 382).
Ex. 213 (a) Cf.l
The key-signature of the imaginary voice is so calculated that the combination

Cp + R represents a transposition of the original (§333). With reference to shifting
the c.f. at the half-measure, v. §§ 334-5.
From the basic version the imaginary is retained as original, but shifted so that the
c.f. is in the register and at the part of the measure required by the con
ditions of the problem.
(b) Original
w &
p ' o
m
¥ pm
xsz
fs @
m^^ s:
C. f. II
Shifting the voices according to the requirements specified:

h= -IM h= 1
Iv="2 +c.f. nv=-ll
is obtained:
. . Derivative
<c) h- 1
-„v—11
iLlMi'i'ii-iiTrriV^s
CHAPTER XIX
ANALYSIS
Composition of the Basic Version

According to Original and Derivative
§ 387. In analysing horizontal-shifting and double-shifting counterpoint it is not

enough merely to indicate by how many measures and at what intervals the voices of the
derivative shift. Analysis must give in the resulting basic versions the factors of both
combinations—original and derivative. There are two basic versions for a combination
with one derivative, because each of the two combinations may be taken as an original.
Taking one of the two-voice combinations as the original it is necessary to indicate
how the voices in the derivative shift horizontally and vertically, to find Jh and Jv, to
compile a formula for the imaginary combination (§ 375) and to write accordingly R to
the original combination. Using a previous derivative for the original, the formula of the
imaginary combination must be changed to agree if the vertical shift is direct, or retained
if the shift is inverse (§ 383), and the second basic version written in conformity with it.
Obs. —The basic version may be found by other means than using the formula of the
imaginary combination. Taking one of the voices as Cp, one of the combinations can be
mentally adjusted to the other so that this voice coincides in both, the entry of R is then
defined. The coincidence must be complete, and to assure this the Cp in the derivative
must be on the same degrees and on the same parts of the measure as in the original. If
this condition is not satisfied in the derivative the adjustment must be made by applying
the method already given.
§ 388. Either of the basic versions will solve the given problem. But it is of primary
importance that the basic version shall represent correct counterpoint; in analysing they
must be abstracted from the general background and their occurrence in the composition
noted. Consisting of the same two melodies but in highly complex relationships (since one
of the melodies imitates canonically), such a basic version enriches the contrapuntal re
sources of composition to an important extent. Often the result of previously written two-
voice combinations, it not only retains its own melodies but also separates them from and
unites them with other melodies, thus becoming a crucial factor in all thematic invention.
§ 389. The two basic versions referred to in § 387 differ in this regard, that the
voices of one of them consists of Cp + P, in the other of Cp + R (or Cp + /?....) and
vice versa.
§ 390. These basic versions may appear as one of three cases: (1) RR of both ver
sions are imaginary; (2) one basic version contains the imaginary R, the other is con-
trapuntally valid; and (3) both basic versions are contrapuntally valid.
§ 391. Case 1: RR of Both Versions Imaginary.—This situation represents the methods
of composition explained in § 378, (2), (3); though useful for scholastic exercises they
are not applicable in composition. It can hardly be assumed that a composer, having in
view a combination with horizontal-shifting voices, would begin by writing a basic
version with the aid of an imaginary R, or by adjusting Cp to voices not contrapuntally
associated, or composing the whole according to its parts— preferring these mechanical
expedients to the far more natural and simple process of writing a contrapuntally correct
basic version and taking from it the required combination. It would be difficult to cite
an instance of horizontal-shifting counterpoint where analysis would show two basic
versions each having an imaginary R. The following is from Palestrina:
232
ANALYSIS 233
Ex. 214 (a)Original Palestrina
/I I
znz J.I i° .I,|,JJJJ,IJ-7JJ. g
■ '-err
H
i" r ' r r TT
It admits of this derivative (not found in Palestrina):
Derivative
(b) J.li UijUj.l^J.J
p J " ■" ^r f ■ » rr'iT»
The formula of the derivative is: (L + IIh=1) Jv" lj that of the imaginary I + II!1".1.
According to the formula of the imaginary a first basic version may be written with
imaginary R:
(c) 1st Basic Version

C}> oJj|Q J J J JJJJ J—,
nc
f *E
ff
/>'I1
etc.
Then the former derivative is taken as the original and the former original for the deriv
ative, the corresponding changes being made in the formula of the imaginary combination
(§ 383): I. .7. + II. According to the latter a second basic version is possible, also
having an imaginary R:
2nd Basic Version
■ fei n —
f¥W
If horizontal-shifting counterpoint of which both basic versions has an imaginary R
were found in a composition it would be reasonable to consider it as entirely fortuitous,
as in the preceding example.
§ 392. Case 2: One Basic Version Contrapuntally Correct.— In Ex. 215 the voices of
the derivative also shift vertically, but at Jv= 0, therefore the shift is practically only
horizontal. ..
Ex. 215 (a) Original palestrina ( E"£j
" o
T
^ -»—m-
4 W
WIS,
h=V/i h= 0
(Iv-7 +nv--7)/A=1i/2(/t;=o
Formula of the imaginary combination: I + II....1'2. Corresponding basic version:

(c) 1st Basic Version

CP . l ^^V» O i-
-e-
1SE #
o ■
i 8 1
2$*
-©.
Rrf"W
O-
The second basic version, as it contains the imaginary ft, is not given here; it will be
referred to later in connection with double-shifting counterpoint.
Ex. 216 is from the five-voice Gloria in excelsis (Qui tollis) of Palestrina's mass
Ad coenam Agni providi. These three two-voice combinations are found:
Ex. 216 (a) (c)

Original 1st Derivative 2nd Derivative
u Palestrina • ih:0
OCa!) Ud U
m pw
m-7
znaap
w ;rrrr
Here the two derivatives differ only in their vertical shifts (the first derivative is the
separation at the octave of the second) and ]h is identical for both, as is seen by their
formulas. That for the first derivative is:
h= 0 h= -JS h= -Vi
(Iv= 4 + nv- 0} /4_ _l/2> Jv=4. ,y- 4 + „
For the second derivative:

h= -V2 h= 0 h= -Vi
(iv= ~7 + iiv~ 4) fk= -%, fr--3; iY". + H.
Since the horizontal shift is the same for both derivatives, one of them must be selected
for the basic version. Combination (c) is chosen, because its basic version is con
tra puntally correct:
Cp II
ANALYSIS 235
Hence the original is Cp + P; applied to the requirements of the problem as expressed
by the formula for the second basic version, this derivative is obtained, and by separating
the voices an octave, the first derivative. This basic version is found twice inPalestrina,
the first time with a change of the initial note of- R, c instead of d; .the second time
shifted a fifth below, with the initial note of R shortened.
Ex-217 Original Palestrina
Original
(a)
311 jJ iJtJ.IJ
-o^- V BE=3= =
Pff
(b)
Derivative
II rr
_ ftS3*
i m- Ji ±£*±
wm B3 f r FFF "O-
h= —154 h=0 h=-l%

av--4 +iiV=7)M=_iyj>/t,= 3. jy- 3 +n
The following basic version corresponding to the formula of the imaginary com
bination, contains the original. Cp + P is taken an octave lower:
1st Basic Version Palestrina

(c) „ lhr-17
Jt lvis
E 3£ p\
w
W^^T 3E
cP U
zai
m
s 3E
HI-
If the last measure of Cp is discarded the second basic version shows correct coun
ty Hi
terpoint (cf. § 393). Formula of the imaginary combination: I + VlC.
2nd Basic Version

(d)
Cp 1 .
-■ e- JJ- 11 J ° fej
£ 3E
f=f
/>ll
r r
* 3E
Tii^sFP
Such instances, where an unimportant change in one of the voices gives a new basic
version that is contrapuntally correct, should be observed in analysing, as they are ap
plicable in composition.
§ 393. Case 3: Both Basic Versions Contrapuntally Correct.—
Ex. 218 Derivative

Original Palestrina ,h:-ri
(b)
"mm *eS rj" Ud a • =55=

^ -»—a-
.V, - n I
i&2
h= _iy2 h= o h= -il/2
(Iv=-l +nV=4)A=_11/2(/i>==3;iy- * + II.
1st Basic Version

(c) r fei
HE =*K
$
iA! T I1 r' ,1 =§=

3T
3S=
Taking the derivative as original and changing the indices (§ 383) to: Jh- V/2, Jv= —3
gives: I + IIV... . Hence:
2nd Basic Version

(d)
1 j ■ C- 'j ' J -
^
*fe£
fe^ «>—rf
-O CT-
P\\
Note the possibility of a second basic version with only a horizontal shifting of the
voices— not found in Palestrina:
wiO^JI u -o-
II F
Ex. 219
Original Palestrina Derivative
Ji <>•♦
^■O'Tr-'r
Re
In the basic version the original is found an octave lower:
1st Basic Version
a- .&.
—_^*—f _—O—1 " o
(c -■—o-
Cp 1 m HE
3E
**
/»II
ANALYSIS 237
Taking the derivative for the original and retaining the previous formula of the
imaginary combination because of the inverse shift (§ 383) gives:
2nd Basic Version
(d)
^
Cp
w
±±.
r\\ W -p1
Ex. 220 Original Palestrina
(fa;
Derivative
p^±tmum
(b)
I Jul
Jv:-7
h = l
(I»—T + n) jh^^ Jv = — 7; i + n*~
1st Basic Version
In order for this basic version to be contrapuntally correct up to the end it was
necessary to interrupt Cp on the penultimate note, indicated*, after which the voice
progression is free. The unprepared fourth in the third measure requires a supplementary
voice. Taking the derivative as original and retaining the formula of the imaginary com
bination (because of the inverse shift) gives the second basic version, in which all
melodies appear complete:
2nd Basic Version
Ex. 221 to Palestrina

(b)
3a;
~n~
to
=: tei«-n^
Ê
in:
nr
Ul
—e-
II tr "ff JX
h=— '/., h=— >/i

(I + IIT=3 ) Jh = — Va; Jv — 3; IT.=8 +11.
1st Basic Version

r nV;?
# 11 ^
Exchanging as before and altering the formula of the imaginary combination to
h= %
I + IIV... .* gives:
2nd Basic Version
(d)
I
=
\ <* ^
HE
fT
m />U
nsr 3E
4 (4)
The slight alterations at the end of P make this basic version contrapuntally correct.
Obs. -The examples quoted from the works of Palestrina prove that horizontal-
shifting counterpoint played no unimportant role in the era of the strict style. These
examples may serve as material for the student's exercises (cf. § 423).
Analysis of Combinations on a Cantus Firmus
§ 394. It remains to discuss briefly how to find the basic version when one of the
voices is a cantus firmus. In the basic version the c.f. must be P (§ 384). Therefore it
is immaterial which of the two-voice combinations is taken as the original. If the c.f. is
I, Jh must be negative; if it is II, Jh must be positive (§ 385). Of the two two-voice
combinations the original will be whichever fulfils one of these conditions (§ 386).
Ex. 222 Original
II . I 1 I I 1 l I 1 I I I J J -
w 3tE 'J Jl.1'1 J.I |J,JJJJJ|^.;1>) j
Derivative
(b)
m *r,—: ^s o .-= .-==
c/iih--2
V: 2
(lT = -2 + C.f.IIv = 22) Jh = — 2, JV = 0.

ANALYSIS 239
Because the c.f. is II and Jh negative, the derivative is tafcen as the original. Instead
of the former_ Jh, Jh- 2 is obtained (§ 383). Formula of the imaginary com
bination: I + II..T. . Writing R to the derivative gives:
Basic Version
(c)
£ ±^=S=jA £
ps% 3E -J ffJ#J ° =W=
m Cf.P II
Cp + R, taken a third higher, corresponds to the original, and Cp + P to the derivative.

Ex. 223
Original .
(a)
Derivative
b). DfeS
(b)
Kn~
■I. JJJiJJ^JJJJ^
3=^= «iii
r/I-
(Cf. lv=-i + IIv=-7) Jh = iV2, Jv = -8.
Since Jh is positive and the c.f. is I, the derivative is taken as the original. Because
of the inverse shift, instead of c.f. I is obtained c.f. II; the indices remain without
change (§ 382) and the formula of the imaginary combination is:
h= £
Hence: i + cf. iiy=.-8.
Basic Version
C/.R Uk$
Combination Cp + R shifting at the half-measure and an octave lower, corresponds

to the original, and Cp + P to the derivative.
06s.—In analysing it is no disadvantage that a second basic version may not be
contra puntally correct, in which case the correct combination must be extracted. This
basic version involves canonic imitation of a c.f. (Obs., § 384), and refers to the study
of canon. Instances are comparatively infrequent and present considerable difficulty
in their working-out.
CHAPTER XX
COMBINATIONS GIVING SEVERAL DERIVATIVES
§ 395. A three-voice basic version (with one R) consists of the original combination
and one derivative with horizontal-shifting voices. If R imitates P at the unison the
derivative shows only the horizontal shift. If the imitation is at any other interval the
derivative shows a double shift. Such a basic version, consisting of an original with
one derivative, is written in simple counterpoint and does not require application of the
conditions of any vertical index (§ 330). But if more derivatives are needed, without
the addition of a new R in the basic version— i.e. retaining the three-voice texture—they
can be obtained by writing the voices of the basic version according to the rules of any
Jv. In meeting these requirements the voices will give a shift or a duplication at the
given Jv, but this new derivative will consist of only one vertical shift of the voices:
either a shift of the two-voice original, or of its derivative, or of the basic version,
assuming that it is contra puntally correct. The possibility of these shifts is dependent
only on the conditions that the corresponding voices must satisfy the requirements of
the given Jv, not that they shall enter into the structure of the basic version. Therefore
at the vertical shift of the combination Cp + R its Jh always remains unchanged; in other
words, in obtaining new combinations by this method the derivatives are the result of
a simple and not a compound Jh (§ 356).
§ 396. If, on the contrary, to increase the number of RR in the basic version and to
write Cp not to a two-voice imitation but to one of three or four voices, means that each
additional R gives a new two-voice derivative with a new Jh. In contrast to the preceding
case, this gives derivatives at a compound Jh. It is understood that here the separate
combinations or the entire basic version must conform to the rules of the Jv concerned,
and that the new derivatives obtained in this way give only one vertical shift of voices.
§ 397. Both cases will be investigated in this chapter, first the combinations with
several derivatives at a simple Jh, later those at a compound Jh.
06s. -If vertical-shifting counterpoint is regarded as a function of the basic version
(§ 336) and problems shifting in this counterpoint are solved by means of supplementary
voices, that is, with coinciding entries of P and R in the basic version, then every
two-voice derivative requires a separate R, shifting at some Jh or Jv or both at once. In
such cases the number of RR in the basic version is equal to the number of two-voice
combinations. Such a basic version is theoretically possible, but entirely useless for
the practical solution of problems.
Combinations with Several Derivatives at a Simple Jh
§ 398. The application of Jv for obtaining derivative combinations at a simple Jh

may be an elaborate process. These cases will present nothing new to those familiar
with part I of the present work. In analysing combinations at one and the same Jh but
at different JJv it is necessary to find the contrapuntally correct three-voice basic
version, and to indicate what kind of vertical shift, as well as the duplications included
in its two-voice combinations (Cp + P and Cp + R) apply in the given situation, and
what is possible in general; then to see if the same basic version admits of any deriv
ative combinations. If the second basic version is also contrapuntally correct the relation
240
COMBINATIONS GIVING SEVERAL DERIVATIVES 241
to the two-voice combination already analysed should be written out and studied. Often
the basic version sought for will be found in the course of analysing compositions; in
other cases it must be deduced from the existing two-voice combinations, as explained
in the preceding chapter. The following analysis of two examples from Palestrina will
supplement what has just been discussed.
§ 399. In the first motet of Palestrina, 0 admimbile commercium (published in 1569),
in the part for the lowest voice is found the following two phrases, separated by rests:
Ex. 224
2, Palestrina
lar. gi.tus est no. bis lar . g-Ltus est no - bis
In a mass of the same title, published thirty years later (posthumous), in the Sanctus
these two melodies are combined in the following way.*
Ex. 225 mm. 4-5&

4- •Palestrina
—
a) mm. 6-7
KF^Z =^ si
i it it =
$
3^
nm
di mm. n-12
IB 1 p a^
'J I I r °°1
The two-voice combinations (b) and (d) represent vertical shifts of (a); combination
(b) at Jv= —11, (d) at Jv= —4 (with the first note of twice the original value). But the
three-voice combination (c) shows a more complex relationship of these two melodies,
for it is the basic version, in the construction of which (a) and (b) enter. Consequently
in combination (c) is found one of the melodies imitating canonically, at the same time
the other (lower voice) acts as counterpoint to the imitation. The relation of the two-
voice combination (a), (b) and (d) to the basic version is:
Combination (a) = (Cp + P) shifting at Jv= —11;

(b) = (CP + P);
" (d) = (Cp + P) shifting at ]v= —7.
So, these derivatives show vertical shifts of one and the same combination: Cp + P.
From the basic version can be written both combinations: Cp + P and Cp + R, though
the second of these Palestrina did not use:
•In these quotations the small notes in parentheses are variants of the melody as found in
Palestrina. The large notes in brackets are continuations of the melody, absent in the original.
Derivative
/fife
^F^-^H
0.11 r ffrf
Formula of the derivative combination
h=-y2
(Iv- 4 + 11) /*=-%,*>- 4.
Formula of the imaginary:
h=-%
tff. 4 + II.
This formula corresponds to the first basic version (c).
The two-voice combinations (e) and (f) also admit of vertical shifts. The absence
of consecutive sixths enables a shift to be made at Jv= —11. Moreover, combination
(e), where fifths are absent, satisfies the conditions of Jv= —7. The absence of similar
motion in (f) admits of a shift and duplication at Jv= —9. The indices of
these combinations are:
(Cp + P) Jv= - 7, -11; (Cp + R) Jv= -2, -9, -11.
Cp + P also admits of duplications (§ 113):
Id= "9 + II; I + IId= "9 and Id= ~9 + n*= -»
Numerous manipulations of combination (f) are possible; Palestrina did not use them,
nor is the combination itself found, apart from the basic version. Here again may be
emphasized the usefulness of analysing derivative combinations, irrespective of whether
or not the composer employed them.
§ 400. Combination (c) of Ex. 225 is the first basic version. To obtain the second
basic version the former derivative (f) is taken as the original (§ 381), with a corres-
h=-%
ponding change in the formula of the imaginary combination; instead of IV.T. +11 taking
h= Vi
1+ IIV "*, which yields a second basic version that is contrapun tally correct.
2nd Basic Version
1B= £Sfc
M>ET
?m p\
=£ 4=t
The former original appears here as derivative, therefore to express the relations
of the two-voice combinations of § 399 to the second basic version Cp + R must be
substituted for Cp + P:
Combination (a) = (Cp + R) shifting at Jv= -11;
(b) = (Cp + R);
" (c) = (Cp + R) shifting at /v= —7.
In this basic version Cp is found in double counterpoint at the twelfth in relation to
each of the other two voices: (Cp + P) Jv= -11, (Cp + R) Jv~ -11, and can, therefore, be
shifted a twelfth below, which gives a shift of the basic version according
to diagram 5)^<
Shift of the 2nd Basic Version

hi
i^ I
g i ': U
: u JCC
Cp wm
This combination, similar to the two preceding, can also function as a basic version
in relation to the two-voice combinations in § 399. It is found in another part of the
same mass, the Agnus Dei, modified thus:
Ex. 226 Palestrina

mm. 2-5
3E nr a
f=
^m Cp
a rj
w
§ 401. Ex. 227 is from Palestrina's mass Regina coeli, where the following com
binations are found:
Ex. 227 Palestrina
";> JifJy- etc.
II
b
-<U -e-
etc.
1&J '1 Jtrr r 'rrrrfrr m
Basic Version
c)
Cp\
— _ gini
ÊÊE $ & a
etc.
*^
BE J J _o_ 3E
^5=^
p\\ ^
Of these (b) is a shift of (a) /v=— 11. Here it is given with literal exactness; in the
original it appears with unessential changes of note-values for the first note of each
I ' i
- J 4=A\- etc. Combination (c) shows the basic version; Cp + P
gn ZEC
TT
in it corres-
ponds to combination (a), shifted a fourth higher. This basic version contains moreover
the combination Cp + R, it is not found in Palestrina, and it requires a supplementary
voice below because of the unprepared fourths. This combination, (d), is derived from
(a) at /*= 1 and Jv= -4.
fe.
(d)
I ME
nSS,
TOM
j J j j '"' i j J
r^ (3)
etc.
iiW+n--i ) Jh= i, Jv = — 4.
So, in these combinations there is only one index of horizontal-shifting counterpoint:
Jh- 1. The vertical shifts and duplications of the four preceding combinations will now
be listed, also which of them is to be regarded as the original:
Combination (a): A>- -9, -11; Id= ~9 + 11; I + IId= "9; Id= -9 + nd- -9 At these
duplications d= —2 must be substituted throughout for d 9.
Combination (b) has the same indices and takes the same duplications.
Combination (d): Jb- -12, 2; I + IId= 2 or I + IId= 9.
The remaining combination (c) is the basic version, of which the voices are numbered
in the order of altitude: I, II, III.
M «— & g^
p
ff
us ingjjjgii in 1 1
It admits of the following vertical shifts; IIV= ±2; IIv~ ±9; IIV=:+5;IV= -2 + IIV= ±2 +
III; IV_2+II + IIIV^~2; I+IIv=±9+III v~ ~2; and these duplications: IId_±9; IIId— 9-
I + IId_ ±9 + IIId 9. These duplications also must take 2 instead of 9. This gives the
following five-voice derivative:
iir l±LL
(0
£3E f-\rt ff
etc.
j-j.
in
P TT -0-4,°-
To those familiar with Part 1 of this work these multi-voice derivatives present
nothing new. They are shifts at various JJv, and do not depend on whether or not a
given combination admits of a horizontal shifting of voices.
Combinations with Several Derivatives at a Compound Jh
§ 402. If a three-voice imitation instead of a two-voice is taken for the basic version,
i.e. if another R is added (RR according to the order of entries are indicated R., R2,
etc.), the basic version consists of an original (Cp + P) and a derivative (Cp + R) with
the addition of a second derivative: Cp +#,» having its own Jh. Adding R (i.e. making
a four-voice imitation) yields three derivatives: Cp + R , Cp + R , Cp+R3, each with
a special Jh, etc.
§ 403. The number of derivatives each having a separate Jh is equal to the number
of RR in the basic version (§ 396). Therefore Jh will be double (§ 356) at two RR, triple
at three, etc.
§ 404. The number of voices in the basic version equals the number of RR + 2, adding
here the two voices forming the original combination.
§ 405. If R, entering into the composition of a given basic version, imitates P at the
unison, the voices of this combination show a shift that is only horizontal (as the vertical
shift is only at Jv= 0); at other intervals of imitation the voices will be dou
ble-shifting (§328).
§ 406. If Cp in relation to P and the given R is an outer voice its combination with
this R gives the direct shift; if Cp is between these two voices its combination with R
gives the inverse shift (§ 329).
§ 407. Here the conditions will be limited to two derivatives at a double Jh. The
basic version, therefore, must be four-voice (§ 404) and will contain two RR (§ 390).
§ 408. The J Jh that enter into the composition of a double Jh may have the same
signs (both positive or both negative) or have opposite signs. These two cases cor
respond to basic versions of various forms.
§ 409. First, when the signs of both JJh are alike, the basic version consists of a
three-voice imitation with Cp:
Basic Version Palestrina

Ex. 228
CpA .1 <Wiaiii
(a) -3SZ
^^ r^rfT r r p
hri
*,II
JBIEBE £fei r r rfeg
re r
K Ilh
Hence an original (Cp + P) and two derivatives (Cp + R and Cp + R ). Since the
imitation is at the unison both derivatives contain only horizontal-shifting voices.
Original (C,, » f)
- I »
jj-l^ d
(b) ofc= fJfff+% t» ^
1st Derivative (Cp\B^

(c) ^F^-r-U-^* A ^~±l<t gj ^
2nd Derivative (cP+BJ
(d) ir^ I«—*» . o «KU*«i
^prf
Formula of first derivative:
0 + IIh= Vl) Jh= ft; its imaginary: I + Il!\T.1/2.
Formula of second derivative:
(I + IIh= l) Jh- 1; its imaginary: I + II^T.1.
Because of the fact that Jh in both derivatives is positive the imaginary voice is
the same in both formulas: II Shifted at the half-measure (IIv~/2), it forms R and
takes part in a combination corresponding to the first derivative; shifted at the whole-
measure (II.... ) it forms R and enters into a combination corresponding to the second
derivative. In these situations, where JJh are both positive or both negative, the Jh of
the lesser absolute value refers to the first derivative, that of the greater to the second
derivative. In analysing such cases the order of arranging the two derivatives should
be the same, i.e. the derivative of the lesser absolute value should come first.
Obs.—Multi-voice canonic imitation may appear in many different forms. It refers to
the study of canon, a subject exhaustively treated in the sequel to this book. For those
who have not studied canonic form multi-voice imitation at the unison, as found in Ex. 228
(a), is entirely accessible and does not require the application of complex counterpoint.
The process of writing it is similar to that of writing two-voice canonic imitation (Obs.,
§ 317). The only difference is that in writing a part P is transferred not to one voice
only but to two or three, according to the number of RR.
§ 410. Second, when the sign of one Jh is positive and the other negative the basic
version corresponds to other forms. For a negative Jh R must be R I, for a positive, R
II (§ 366), consequently P I and P II are found in one and the same basic version. In
other words both voices of the original combination I + II are PP, and each has one R.
The result is two simultaneous two-voice canons, the so-called double canon. The
voices of the original combination I +11 are indicated by two P's, the first according
to the time of entry P", the second Pb. If both PP enter simultaneously it does not
matter which is used. Each P has its own R, indicated by the same small letter: Pa and
ft" is one of the canons, Pb and Rb the other.
§ 411. The following double canon is from the Kyrie of Palestrina's Missa
Papae Marcelli:
Ex. 229 (a) Palestrina

J=sfe
HP
h-1
*Mlv-Io
-0 1 .—'—I—| .—ar
Taking this double canon as the basic version, in which P" + Pb forms the original
combination, the pairs of combinations R* + Rb must be discarded, as this is only a
shift of P' + Pb at Jh= 0; Pm + R' and Pb + Rb are left, but each of them contain only
one imitating melody, while the derivative combination must consist of both voices of
the original. Excluding these there remain, besides the original P" + Pb, two derivatives-
P" +RbandPb +/?»:
Original (P"+Pb)
v>We«=Uz^J.J Jo m .©#
—nr i ' r ^¥^ _u~
1st Derivative (p«+r*)

(c) 3E
3jE±E=
ih=-l
lv = 7
s £ 3tt:
2nd Derivative (pi+R»)
wBf'JJ'V% ^=
F W
h=-l
(I + II) Jh= —1, Jv= —7; its imaginary::iy.T. ? + II.
a + iih-1)/A= 1; its imaginary: I + II?h=l
Compound index of horizontal-shifting counterpoint: Jh- -1,1. Since one of these values
is positive and the other negative the imaginary voice in one case is I in the other
II...., and the basic version is therefore a double canon (§ 410).
Obs.— A simple method of composing a double canon consists in writing part of the
combination P* + Pb and transferring it to the same or to other degrees at Jh- 0 and
Jv= 0, forming the beginning of Ra + Rb, which the first two voices accompany in simple
counterpoint. This counterpointing section of P" + Pb is also transferred to the other
two voices, forming the continuation of R" +Rb, etc. The composition of such a double
canon is analogous to that of the ordinary two-voice canon, each pair of voices cor
responding to a single voice in the two-voice canon. Other phases of the double canon,
where R* + Rb consists of a shift of P' + Pb not at fh= 0 and Jv=Q but at other in
tervals* can be applied only to a kind of basic version now unfamiliar-dealt with in a
course in canonic form. An exception may be made for the shift at Jv= -7 (cf. Ex. 230).
§ 412. In those cases where both JJh have the same signs the derivatives are ar
ranged in order according to the absolute values of Jh, writing the first derivative at the
Jh of smallest value, the second Jh at the largest (§ 409). The original and the second
derivative are termed outer combinations, the first derivative the inner combination.
§ 413. Taking one of the outer combinations as the original, a basic version in the
form of a three-voice imitation + Cp is obtained by the method indicated in § 409. Getting
now the two basic versions by the same method, it is seen that their only difference is
that P of one forms Cp in the other. Here the reciprocal relation of the outer combinations
•If in this case Jh^ 0 Ra x+ °b

R contains a third derivative with a separate Jh.
is such that if either is regarded as the original the other is the derivative. Hence there
is the same relation between the formulas of the derivative combination (and the imagi
nary) as there is when the original and derivative are exchanged (§§ 380-2).
§ 414. Taking the first derivative (inner combination) as the original, a basic version
is obtained in the form of a double canon. Therefore in a double canon adopted as a basic
version the inner (/"* + Pb) is the original combination and the outer (PB + Rb and
Pb + RB) the two derivatives. In the following is applied all that was said in §413
about outer combinations.
§ 415. The double canon from Palestrina's mass Inviolata illustrates this use of the
basic version. In this canon RB + Rb consists of a shift of PB + Pb at Jv= —7:
Fx 230 1st Basic Version (double canon)

(a) A j?aI1v:_' Palestrina
Pb\\
Formula of first derivative (PB+ Rb):

h= 1 h= 1
(I + Hv~ _7) /*= 1, Jv= -7; 1+ II^.—7.
Formula of second derivative (Pb + /?"):
(Ih= _1 + II) fk= -1; Ih.T._1 + II.
Taking each of the derivative combinations (outer, § 412) in turn as the original, two
basic versions are obtained, each in the form of a three-voice imitation + Cp. First
taking P* + Rb for the original:
Original
■ U-j J 3E
IT
1st Derivative
^rm
41 J J, J ,_U
ZEC
inn
ihTT
lvz-7
2nd Derivative
W^PE 1! J j j , J J --J.
ihr 2
lv;-7

h= 1 h= 1
(Iv= -7 + n) Jk- lf frm _y. j + ny- -7
h= 2 h= 2
(Iv- ~7 + II) /*= 2, Jv = -7; I + Iiy.T.—7.
The second basic version (three-voice imitation + Cp) is constructed according to
the formula of the imaginary combination:
2nd Basic Version
^TT-f
Finally, taking the combination Pb + R" as the original:
2ndgderivatirve1r iSt ?e^*?^5 Hso » 2nd Derivative (former

former 1st derivative) original) Ilhv:?7
CO - II J 1 J, (8) =r 1 1 J I 1 (h) x "v'-r
wm- etc.
II |h,i grrr'rr r
a + Hh~ X) /A= 1; I + IlKT.1.
h= 2 h= 2
(I + iV* ~7) Jh=2,Jv=-7;l+ Iiy.7 ~?
The third basic version is constructed from the same formula of the imaginary as
before. P of this version is the same voice as Cp of the preceding (§ 413).
3rd Basic Version ■fc :*

0) 1 xr:
^PISp
zm- P\\
fi rj Z 0-
- _U
The formula of the imaginary combination for the second derivative in both basic
versions is the same, because of the inverse shift (§ 382).
§ 416. Of the three basic versions given in § 415 Palestrina used only the first, the
double canon. The other two were obtained by the process indicated, and could have
served as material for the composition, on an equal footing with the first. These three
basic versions are unusual in that all of them are contrapuntally correct. For instance if
the double canon in Ex. 229 were taken as a basic version only one of the two remaining
basic versions would give a correct union of voices; this version is obtained if the
second derivative (Pb + R") is taken as the original. According to the formula of the
imaginary combination, which the reader can easily find himself, the following basic
version of the first form is obtained:
250
2nd Basic Version

ih'-i
J=± ~n~
erf Ty
*=^
u
P\
a n;
WP r
o ,o
m^ "o-
etc.
cP\ II
§ 417. It is impossible to indicate the characteristics that would enable one to

determine in advance which of the possible basic versions exhibited correct counterpoint
and which contains R.... , and therefore useless as material for composition. This can
be done only by writing out all the versions. Very often only the basic version as
originally written contains correct counterpoint. For example the four-voice combination
in measure 4 to 8 of the following passage from the Gloria of Palestrina's mass Aeterna
Christe munera satisfies all the requirements of a basic version:
Palestrina
JBBI
$=^
«= Cumsancto Spm
pm-tii in g-lo-ri . a
JM.
He 3SE
(Cj.)
s—o-
WfrfE£
we
Cum sanctoSpui tu in£'lo_n
(PL . a De . i Pa tris
IS * &-f
<e i If8 d £
in glo_ri.a De.i
P Pa
tris A men, in . g-loj-i .a Dei PaJris.A men.

§ 418. In continuing with the study of canonic form using multi-voice imitation no
difficulty will be experienced in applying canonic imitation in four or five voices to the
basic version. But from a practical point of view this is hardly necessary. The number
of derivative combinations available increases according to the degree of advancement
made in the study of shifting counterpoint. It must not be forgotten that in combinations
such as have been studied, for example, the conditions of various JJv may be applied
and new derivatives obtained with only one horizontal shift. Instead of being too con
cerned about methods of obtaining new derivatives a far more important matter is now
to learn how to arrange them to the greatest advantage from the esthetic point of view.
Arrangement of Contrapuntal Combinations

in the Reverse Order of their Composition
§ 419. With the increase number of derivative combinations intended to serve as mate
rial for composition, the important question arises as to the order of their presentation.
The most expedient way of assigning different combinations of the same melodies is
according to the degree of their complexity; at first melodies comparatively simple in
structure, and for elementary thematic work more complicated ones. The progressive
increase of interest throughout the composition and the ease or otherwise with which
parts are heard is what governs this apportionment of contrapuntal material.
Proceeding on this plan, it is not infrequent to change the order in which these com
binations arise, presenting first the ones that are written later and vice versa. For ex
ample it is advantageous to place a contrapuntally correct basic version later than the
two-voice combinations extracted from it; in general, a multi-voice imitation, previously
written, after a two-voice imitation on the same themes. In fugues with more than one
subject—double, triple etc.—the combination of subjects is written first of all and is
generally reserved for the concluding section, from which working-out of the subjects in
the preceding sections is taken. In the same way subjects of a stretto fugue can be
extracted from a previously written canon.
§ 420. The following basic version is from the Benedictus of Palestrina's mass
Jam Christus astro ascenderat:
Ex. 232 Palestrina
(3)
* <3)1U
m
PPP :? »- nE
3E
rrr
-O
SP*
In combination Cp + P is found the free use of fourths, requiring a lower supple
mentary voice. On the contrary, Cp + R presents an independent two-voice combination, at
the beginning of the Benedictus (Ex. 233) with the voices brought an octave closer to
gether. In order of entry P is the first subject, Cp the second. After the two-voice union
of these subjects, on the latter half of the fourth measure, the lowest voice enters, with
an imitation of the second subject. On its final note in the seventh measure the basic
version begins, presenting a more interesting union of the same two subjects. Here the
first subject enters again, forming a canonic imitation in the outer voices at the distance
of one measure. Beginning with the ninth measure the second subject joins the canon, it
is unchanged up to its final note. To these three voices a fourth supplementary voice is
added, an active, pliant melody that forms an excellent counterpoint to the notes of the
basic subjects.
Ex. 233 Palestrina
w 3E
2nd subject -C^
w *
lst subject Be
xj:
ne
TT
di - etus qui
-O-
ve . nn
-xa&L
Hi =t^=
Be .
3»~
jO.
ne . di
ti:
etus qui ve . nil, qui
l-Cp)
fe ri o-
Be ne . di
Basic Version
F=£=
cr
Be lie di etus qui ve
IE
ti :
qui
m £
ve _ nit.
3E
be ne. di etus qui
m ve
-X51
nit, be.ne.di
_^_-
etus qui
f2^=
ve
.&-
—e- IDC c^
PF XT—
_etus qui ve nit, be ne di etus qui
i j^C£L
*
TJZ
18 L,
.nit, be . ne di . etus qui ve
-©-
m:
IF ve nit be ne . di etus qui ve. nit
T
t^h=Ff 3JZ
a
nit, be ne di . etus qui ve nit
&^=n:
ve nit. qui ve nit. etc.
In reading these scores the question will inevitably arise as to how this gradual
increase of interest shown by the various unions of subjects is brought about. The
answer is simple: the combination of greatest complexity— the basic version— is written
first: from it are taken the simpler combinations, placing them first, then the basic ver
sion, appearing as the culmination of the entire thematic material.
§ 421. At the beginning of Ex. 234, from the second part of Palestrina's motet
Manifesto vobis veritatem, the following combinations are found:
mm. 1-4 mm. 3-5

X lJ,v = -7 [ | I
i 3T
11 i rf
r W= ffi^P?
The initial note of II in combination (a) is here shortened in order that the melody shall
conform to future appearances. Formulas:
h= -1 h= -1
(I + 11 v=-7-) M= -1, /v= -7; iy-.-7 + II.
The first basic version follows the formula of the imaginary combination-
1st Basic Version
fe^t^y 3E
m:
Cptt
SX---
B \ vr-7
It is found in measures 8-10 of Ex. 234, shifted an octave higher.

The derivative is taken as original for the second basic version, and in accordance
with § 382 retains the formula of the imaginary combination:
2nd Basic Version
etc.
K Iy.=.-7
Retaining the imaginary R, this combination cannot be applied in the work. If R is

shifted a twelfth higher (cf. combination (e) ) or a tenth (f), a correct contrapuntal union
is obtained in certain places. The first of these (with a shift of R at the twelfth) is
found in measures 4-5, all voices shifted one-half measure:
?^£-~
fei ZKJZ
fTTT
r,l[| T I I •
The second combination (with a shift of R.... at the tenth) is found in measures 6-8:
But in both instances the melodies that enter first must soon be discontinued, since
they cease to form a correct union. Hence it is clear that combination (a) must be re
garded as the principal basicversion, for in it the subjects form a correct contrapuntal
union from beginning to end. This version, as the more complete, follows the above-
mentioned variants of the second basic version (mm. 8-10). After this combination (a)
and (b) again appear in the former order: Cp + P (mm. 10-13) and Cp + R twice (mm.
12-17), but with this difference: that whereas these combinations first appear on the
same degrees as in the basic Version, here they are on other degrees; the result is an
increase of harmonic interest. In measures 17-19 only one theme (P) appears, but in a
more complicated setting: instead of the former two-voice imitation it is here three-voice,
and instead of a distance of one measure between entries, a distance of one-half measure
(stretto announcement):
Ex. 234 Palestrina
Cp\P __ CB+JL , , , , Cp+R
- o nrrpt 3T
ja
m
*=^
»=? iu IAl P le.lu ija AI 1
»— ia.
^~
E ja Al P le.lu P
IE
I--Pj
lMM£ -ja.
m f r=jr>
le
s^ Al le.lu
«:
■Ja
xc
AI te
(-Pi
§ 422. No difficulty will be found in analysing the following examples from the
Benedictus of Palestrina's mass Tu est pastor ovium. Observe the absence of similar
motion in combination P +/?., making possible a duplication at the tenth; this is found
at the beginning of the extract (mm. 1-3) and again immediately after the basic
version (mm. 11-12).
Ex. 235 Palestrina

i-Hp
F=£
*
fee H£&L
IN no . ini.ne Do
C-PJ
&— -^
ê
in no . mi . ne Do
PEÊ
W <-p)
m £
EEf__.-idj-_r
in no . mi . ne Do
rar. he
mi
±*
f—E—=
mi . ni. in no . mi . ne
m Do
$
~^
nn _ in, TsEl.
in
tr\ I
no . mi.ne
m Do
^=^=
W w PP w?m
te^+^ m
3E
' ^, Basic Version

p a e-
« » MZH « • «-» a b
* mi. in in no . mi. ne s
Uo
F=e=
ini
-d-■ g> n DEC

# mi . ni, in no . mi.ne Do mi
-4
*£ 3E 3n «=—#
» in no . mi ne Do nu
mm in no
=t
mi-iie Do
ppp a s
mi . in,
-j p-
ZEC
ni in no mi.ne Do mi
fcfll
-fi (»■
BE r r 'r "
# . m, in no mi ne Do
IF HE
.ni,
i-pi
mm za.
in no .
m
mi . ne
22=
Do
no
(-P>
m .ni. in
rc &*-
no . mi.ne Do
o'
<*—p-
nu
=W=^r-
ni.
$ mi ni, Do
mnn ni.
*^^
w E=E
111 no . mi.ne Do
^^
nu
s ni.
m mi . in,
3T
Do~
30Z
ini ni.
1
§ 423. The composition of such examples, where from a previously written basic
version is taken thematic combinations of different degrees of intricacy, presented in an
order of progressive interest, is of the utmost profit to the student. At first the basic
versions given in Chapter XIX may be used. These exercises should preferably be
set to a text.
Obs.—Imitative forms have more significance when the same melody is set to the same
text. As a subject is interwoven with other voices, each with its own words, both music
and words appear as the expression of a single purpose, a fact that facilitates the
understanding of contrapuntal forms and the meaning of texts. The words assist in
disengaging the principal subject from amid the surrounding melodies and acquaintance
with the melodies inevitably forces the listener to remember their close association
with the words.*
§ 424. The ways described, by which the parts of a composition are allocated in the
reverse order of their origin, lead to widely different results. For instance, a two-voice
canonic imitation is often found, followed by a multi-voice imitation—obviously written
first—on the same themes. In the following example the two-voice imitation in the first
two measures was undoubtedly extracted from the three-voice imitation in measures 3-5:
Ex. 236 Meyer
§ 425. A fine example showing the application of this method is the Finale of
Mozart's C major Symphony (K.V. 551, "Jupiter"). Here a five-voice combination, in
cluding the cantus firmus, was first composed in quintuple counterpoint:
Ex. 237
= etc.
From this the themes of the Finale are taken, and the original combination itself, with
shifts in quintuple counterpoint, is found in the coda as the climax of the movement
(cf. m. 36 from end).
•With reference to the setting of texts, c.f. J. Quadflieg, "Ueber Textunterlage und Text-
behandlung in Kirchlichen Tonwerken," in Haberl, "Kirchen musicalisches Jahrbuch. 1893.
pp. 109-128, and 1906 (XX Jahrgang), pp. 197-233. Also cf. in Bellermann's work already re
ferred to. the chapter ("Vom Unterlagen der Textworts." [Cf., for more accessible references, the
following: "The style of Palestrina and the Dissonance," by Knud Jeppesen, translated by
Margaret W. Hamerik (Oxford University Press), pp. 34-42; "Counterpoint," by Jeppesen, trans
lated by Glen Haydon (New York: Prentice-Hall, Inc.), pp. 158-60 and 255-61; "Contrapuntal
Technique in the Sixteenth Century," by R.O. Morris (Oxford), Chapter VII. —Tr.]
Obs. -It is of interest to observe that in the Chris te eleison from Palestrina's mass
Jesu nostra redemptio not only the same c.f. is found, but also contrapuntal parts that
recall the themes from Mozart's C major Symphony:
Ex. 238 Palestrina
etc.
The entry of the first two voices has a certain affinity with the beginning of the
fugato in the Finale (m. 36); the first notes of motive (a) in the upper voice are re
produced literally in the symphony (voice IV of the five-voice combination), and motive
(b) in the tenor recalls Mozart's subject I in contrary motion.
§ 426. Another example of thematic material presented in the reverse order of com
position can be cited from the last movement of Beethoven's Ninth Symphony. The theme
to which the words "Seid umschlungen Millionen" are set, first appearing in the Andante
maestoso, in the counterpoint, previously written at Jv=—7, to the principal theme of
the movement, "Freude schonei Gotterfunken"—
Ex 239 AlleG'10 ene><rico.
Beethoven
J i.i J
etc.
Mhlun . pen
and then taken as a separate theme which appears earlier.

§ 427. The same thing occurs in Borodine's In Central Asia, where there is a com
bination of two themes: a Russian song and with it a contrapuntal part of distinctly
oriental character. It appears after the separate statement of the two themes:
Ex. 240 Borodine
^m Q'MFSrS^tcT t/f-
etc.
§ 428. Horizontal-shifting counterpoint can be used to advantage in all situations

where two themes are developed together. It is applicable in double fugues, in the
development section of sonata movements etc. Horizontal shifting brings the greatest
variety into the combination of themes, both in their contrapuntal and harmonic re
lationships, especially if the vertical shift is used in connection with the horizontal,
and with the application of some Jv applied at the same time in the separate com
binations. It may be observed that except in those cases where the derivatives succeed
the original the experienced eye of the contrapuntist will often detect the purpose of a
series of combinations. The sharpening of this faculty of contrapuntal thinking should
be one of the chief objects in studying the present work.
§ 429. As an example of horizontal shifting applied to the subjects of a double fugue
the following may be cited, from a fugue by Battiferro, a seventeenth-century composer;
an excerpt is quoted by Marpurg in his Abhandlvng von der Fuge. The two subjects,
combined first in this way:
Ex. 241 Battiferro

(.) »--,■_, 'j J.JJJ J iiX
etc.
later appear with these horizontal shifts:
,ihrl
1IV--2
(b) 5 . a i
¥= w&m etc.
|h-0
lv;-9
b- 1
According to the formula of the imaginary combination I -i- 11*7. the following
basic version is obtained:
(c)
The two-voice combination contained in this basic version admits of the following
shifts: (Cp + P) £--11, -14; (Cp + R) /v=-9, -11. The latter combination admits
also of various duplications in imperfect consonances. Another double fugue by the
same composer, also using horizontal-shifting subjects, is quoted in Dehn's Lehre
vom Kontrapunkt.
§ 430. In Vol. II of Ambros' Geschichte der Musik, chapter on "Die Kunst der Nieder-
lander" he says that the compositions of J. S. Bach recall the contrapuntal methods of
the old Netherlanders. One would therefore expect to find horizontal shifting among the
resources used by Bach. Riemann observed this in his Cathechismus der Fugenlehre
(Teil U, pp. 149, 205) in connection with fugues 17 and 23, Vol. II of The Well-Tempered
Clavichord. Reference may here be made to the following combination from Fugue 18,
Vol. 1 of the same work:
Another reference is to the triple fugue in W for organ*, where there is a combination
in horizontal-shifting counterpoint of the first subject:
Ex. 243
with the third (cf. beginning at m. 30):
An analysis of this fugue will be found in Prout's Fugal Analysis, pp. 171-89.
Horizontal shifting can also be found in Handel-in the double fugues for piano in
g minor and fj+ minor, in the concluding chorus of the oratorio Alexanderfest,
and elsewhere.
§ 431. As an unusual example of the application of two-voice horizontal-shifting

counterpoint by a later composer the following passage can be cited from the Seventh
Symphony, Op. 77, of Glazounow:
Glazounow
Ex. 244
I'HM1 'W
"Organ compositions, Peters Ed., Vol. No. 1. [Cf. also: Bach, Two-Voice Invention No. 11,
mm. 1-2 and 3-5 (horizontal shift and contrary motion); "The Well-Tempered Clavichord," Vol. 1,
Fugue 1, mm. 2-3 and 4-5; Vol. 88, Fugue 4, mm. 55-6 and 67-8; "The Art of Fugue." Contra-
punctus VIII, mm. 39 and 125; Contrapunctus XIV (the last, unfinished, fugue), mm. 147-53 and
167-74 (see note at end of chapter). —Tr.]
fijag
U «r v r
§ 432. The last three examples quoted belong to the free style, in which the prin
ciples of horizontal-shifting counterpoint may be similarly applied. Here the study of
horizontal-shifting counterpoint is to be distinguished from that of vertical-shifting. The
rules of the latter are different, according to whether the style of the composition is
strict or free, but for horizontal-shifting and double-shifting counterpoint this is of no
importance. These rules are of a more general nature and do not lend themselves to a
rigid classification according to strict and free styles. Such matters as the proper use
of ties, the entrance-intervals of voices in the basic version, the extraction from it of
the possible two-voice combination etc. do not affect the question as to whether the
contrapuntal parts of the basic version are strict or free. It is only when the conditions
of JJv are applied that a positive demarcation emerges between the strict and free styles.
Aside from this the study of horizontal-shifting counterpoint can be extensively applied
to problems in contemporary musical art. The introduction of horizontal-shifting into the
technique of present-day composition will open up numerous untried resources and will
enrich it by effects characteristic of this field of complex counterpoint, effects that
are inaccessible by other methods.
[Translator's Note]
An interesting case—perhaps a unique one—of where the basic version was not discovered
until long after the composer's death, is illustrated by the last fugue in Bach's "The Art of
Fugue." Bach did not live to complete it; it breaks off at measure 239, after he had assigned a
separate exposition to three subjects and had combined the first and second, and all three.
Of all the fugues in the collection this unfinished work is the only one that does not contain
the "motto-theme" on which the entire work is based, a fact that has caused misunderstandings.
The first was due to a theory advanced by Wilhelm Rust, who claimed that because this subject
was absent the fugue did not belong to the collection at all, and was an independent work. Also,
several theorists have misinterpreted this fugue for a similar reason; at least three—Prout,
Goetschius and Gedalge—classify it as a triple fugue, which is unjustifiable, considering its
unfinished state. This is still less excusable when it is remembered that long before these
theorists wrote, Martin Nottebohm discovered a way in which the motto-theme in quadruple counter
point with the three subjects developed by Bach up to measure 239, the obvious conclusion being
that Bach was writing a quadruple fugue. Several attempts have been made to complete the fugue,
on the basis of Nottebohm's discovery, but none were successful until Donald Francis Tovey
wrote a thoroughly convincing completion in which the four subjects are combined in both direct
and contrary motion. The basic version will be found in measures 271-6, Contrapunctus XIV, of
Tovey's edition of "The Art of Fugue," Reference should also be made to his "Companion to the
Art of Fugue"; both are published by the Oxford University Press.
DIVISION D
THREE-VOICE HORIZONTAL-SHIFTING AND DOUBLE -SHIFTING COUNTERPOINT

DIVISION D
THREE-VOICE HORIZONTAL-SHIFTING AND DOUBLE-SHIFTING COUNTERPOINT
CHAPTER XXI
THEORY OF THE TECHNIQUE OF THREE-VOICE

HORIZONTAL-SHIFTING COUNTERPOINT
Preliminaries
§ 433. In three-voice horizontal-shifting and double-shifting counterpoint a derivative
combination is obtained as the result of a horizontal shift (or also with vertical) of one
or two melodies of the original combination. Instances of where all three melodies shift
can be accounted for as the result of shifting only two, or if desired one melody may be
regarded as not shifting.
§ 434. It is necessary that all three melodies of the original combination shall be
present in the derivative. Therefore if one of them is replaced in the derivative by a new
melody, even though the remaining two show a horizontal shift, the result is not three-
voice horizontal-shifting counterpoint but only two-voice.
The Basic Version

§ 435. Three-voice horizontal-shifting and double-shifting counterpoint is written
with the aid of a basic version that includes not less than four voices.
§ 436. If to a two-voice canonic imitation (P + R) is written two contrapuntal parts:
Cp and Cp , a basic version is obtained from which two three-voice combinations can
be 'taken: Cp +Cp2+ P and Cpl + Cp2 + R. These two combinations, of which the first
is used as the original and the second the derivative, differ in that the voices concerned
in the imitation appear in the derivative as a horizontal shift.
§ 437. If the imitation is at the unison the voices show only a horizontal shift; if at
other intervals, a double shift (cf. § 328).
(1) Imitation at the unison; horizontal-shifting counterpoint:
Ex. 245 (a) Basic Version
W
*
HE
P
o o m -e~ £
etc.
w 3E £M p
J&t IE
3E
Cp,
ppÔ P
JE
etc.
7 t, 4s m P
265
Original (Cpl + Cp., + P)

I .».
(b)
3|E TE £M :'i.|o
¥ JL
ii HE F^
III etc.
HE
a-o-
F^
S s
Derivative (Cpx+Cp.,+Jl)
Eh i i p-»
(c)
Is£ Se o ^
JL^ 3E ^ P HE
3E r¥^
&
III etc.
-»- m
P=^ £3E
^^
The derivative is obtained by shifting the upper voice one measure to the right.
(2) Imitation at the seventh below; double-shifting counterpoint:

^*-
(a) 3E
CEl
R
B= HE a j m—P.
_^
aa
Original (Cp^Cp.^P)
(b)
s3^ §!PPf£i
I II
^
etc.
JIL
3^ ^
^
£
TECHNIQUE OF THREE-VOICE HORIZONTAL-SHIFTING COUNTERPOINT 267
Derivative r*^ (Cp^Cp.^+B)

(c)
41 =
m
fe s
^
JIL
^^ P^ =£
The derivative is obtained by shifting the upper voice one measure to the right and a
seventh below, according to diagram (2) (cf. Obs., § 442).
§ 438. The composition of such a four-voice basic version presents no difficulties.
If not written in this way three-voice combinations will be obtained that are apt to sound
empty, and impractical without a supplementary voice. The advantage of following the
method is that a basic version will be obtained that represents correct four-voice counter
point, and can be employed in the capacity of thematic material on an equal footing
with the three-voice combinations taken from it.
§ 439. In a previously written basic version an imitation can appear as imaginary R. If
this is done Cp and Cp must be so written that they form correct three-voice counterpoint
with both P and /?.... separately. This corresponds to the second method of solving prob
lems discussed in § 378 and is the least useful for three-voice counterpoint. The difficul
ties involved in writing only one voice in such a basic version have already been noted,
but with two voices, indispensible for three-voice horizontal-shifting counterpoint, these
difficulties increase to an excessive degree. This method can be applied only in those
situations where the simultaneous movement of P and R does not form dissonant har
mony (§ 327), and where P continues only for a short time after the entry
of R for example:
Ex. 247 Basic Version with imaginary Risposta
Hence the original is (Cpl + Cp2 + P) and the derivative (Cp t + Cp + R), where the
middle voice shifts two measures to the right.
§ 440. Writing horizontal-shifting counterpoint with the aid of a four-voice basic
version gives a derivative combination in which the entries of two of the voices keep
the same relationship that they had in the original. These two voices, therefore, may
be regarded as non-shifting, and the derivative combination as having been obtained
from the horizontal shifts of only one voice. To work out solutions of all possible shifts
of voices in both horizontal- and double-shifting counterpoint another method will be
given, consisting of an adaptation to three-voice counterpoint of the third method of
solving problems explained in § 376. A basic version can be written by this method
only if among the JJh no Jh is equal to zero. Six voices are concerned, three real and
three imaginary, and the writing is done by sections, one after another. A detailed
description is given in the next chapter.
Notation of Voices and Shifts

§ 441. From a three-voice original combination I + II + III will first be examined the
shifts of the outer voices I and III. These voices are indicated in the same way that
I and II were in two-voice counterpoint. In the original combination I is always the
highest voice, III always the lowest. In relation to the horizontal shift I is the voice
of earliest entry, though in fact it may enter later than III; III is always considered as
the voice of latest entry. The vertical and horizontal shifts of these voices can be
represented by the following diagram (cf. § 347):
f
+ +-III
i
* — — !-♦. +
That is, for I the positive shifts are upward and to the left; for III downward and to the
h= 1 h= -2
right. Thus the expression Iv~ +IIIV~ * means that the upper voice (I) shifts one
measure to the left (h= 1) and at the same time a fourth higher (v= 3); the lowest voice
(III) shifts two measures to the left (h= —2) and a second below (v= 1) etc.
§ 442. How to indicate the middle voice remains to be explained. The horizontal
shift of this voice is expressed in the same way as was the vertical (§ 241). II in re
lation to I is regarded as the voice of later entry, for which the positive shift is to the
right. In relation to III it is the voice of earlier entry, in which case a shift to the right
is negative. Based on this, a horizontal shift of the middle voice (II), similar to the
vertical shift, is indicated by the use of two signs together, placed one under the other;
the upper referring to the combination of II to I, the lower to its combination with III.
Therefore ± means a shift of II to the right, + its shift to the left. Indicating the number
of measures that II shifts by a, its shifts are shown in this diagram:
IIh=*a and IIh=*a
For example, the expression II +1 means that II shifts one measure to the left, and
II -Z two measures to the right. In indicating a shift of II where it is combined with
one of the outer voices (I + II or II + III) one sign is enough, referring to its combination
with the voice concerned. E.g., the expressions I + II ^ and II * + III both indicate
a shift of II two measures to the right.
06s.—The six diagrams that show the permutations of voices in the derivative (§ 248)
refer only to their vertical shifts, therefore they apply here only to double-shifting
counterpoint.
TECHNIQUE OF THREE-VOICE HORIZONTAL-SHIFTING COUNTERPOINT 269
§ 443. The use of these signs in the formulas of derivative combinations indicates
the shifts that meet the conditions of a given problem. For example, in the three-voice
combination of Ex. 245 the formula of the derivative is:
lh=_i + IIh=o + IIIh=o
Voices II and III are here regarded as the two that remain stationary, but I could have
been considered as the non-shifting voice, in which case the formula for the same
derivative would have been:
ih=o + IIh=:j:i+n.h=-i.
In this way different formulas may refer to one and the same derivative. The essential
similarity of such formulas is proved by the fact that their JJh are identical (cf. § 448).
The Indices Jh\ Jh", Jh2

and their Reciprocal Relationships
§ 444. In relation to the horizontal shift, three- voice counterpoint will be examined,
as before, as the union of three two-voice combinations: I + II, II + III and I + III.
Given a formula for the derivative combination, indicating a horizontal shift, the
corresponding Jh must be found for each of the three pairs of combinations. This done,
three formulas for the imaginary combinations are then compiled, which when united
constitute the formula of the basic version.
Obs In two-voice horizontal-shifting counterpoint the basic version is composed
according to the formula of the imaginary combination, which at the same time shows
the formula of the basic version. But in three-voice counterpoint these two formulas do
not coincide; the formula of the basic version is compiled from three formulas of the
imaginary combination, as there are three pairs of voices involved. If one of the JJh is
equal to zero the formula of the basic version will consist of only two formulas of the
imaginary combination (cf. § 461).
§ 445. Analogous with the indices Jv , Jv" and Jv% (§ 243), those of horizontal-
shifting counterpoint are indicated: Jh',. Jh" and Jh?,. Jh' refers to the combination
I +11, Jh" to II + III and Jh?. to I + III. The latter combination is equal to the sum of the
first two, which can be demonstrated in the same way by which /vS1 Jv' +Jv" was (§ 234).
§ 446. Expressing the relations of JJh by the equation Jh' + Jh"~ MX, the value of
each term can be defined (cf. § 244):
Jh'=Jhl- Jh";
Jh"-]h1-Jh'
Jh2=Jh' +Jh"
Hence it follows that: (1) only two JJh may arbitarily be taken, the third necessarily
follows; (2) one of the JJh may equal zero; (3) all three JJh may equal zero; (4) two
JJh can not equal zero unless the third also does. When all three JJh equal zero the
result is counterpoint in which the relations between the entries of all three voices
remain unchanged, in other words, simple counterpoint, which may thus be regarded as
a special case of the horizontal shift.
§ 447. Without altering the values of the three JJh substitutions of other hh values
may be made for the individual voices; the relation between the entries in the derivative
remains without change (cf. § 245). For example, I * + II ." + III ° may be sub
stituted for I * + II " + III , because the values of JJh are identical, namely:
Jh'-3, Jh"= —2, /42=1. All this was explained in Division B with reference to the
indices of vertical-shifting counterpoint. In general there is a uniform analogy be
tween JJh and JJv.
§ 448. The formulas given in § 443 represented different expressions of one and the
same arrangement of voices in the derivative. The identity of the formulas is seen when
the JJh are found, these being the same in both cases. For the first formula (I 1 +
nh= 0 + mh= 0) the indices are.
( Ih= -1 + IIh=_° ) Jh' = -1
(iih~ ° + mh- °) /&"= 0
For the second formula (Ih~ ° + IIh= +1 + IIIh~ -1) the indices are:
(Ih=0 + IIh=-l )/4i=_!

(IIh_ 1 + IIIh_ _1) Jh«- 0
Jh =-1
Formula of the Basic Version

§ 449. Three-voice horizontal-shifting counterpoint, like two-voice, is written with
the aid of a basic version. According to the three indices Jh', Jh" and /A£ (if in none
Jh- 0) three formulas for the imaginary combination are compiled, each of them in the
same order as was done in two-voice counterpoint: at a positive Jh the imaginary voice
is indicated by the highest numeral, at a negative Jh by the lowest. In the imaginary
voice is written h= Jh. Hence each of the three formulas of the imaginary combination
can appear in either of two forms:
(1) I + IIh = Jh'=a or Ih = Jh'=-a + II;

(2) II + mh = Jh" = a of „h =Jh» = -a+ni.
(3) I + IIIh = Jh2 = a or Ih = JhI = -a+III.
The three formulas of the imaginary combination together constitute the formula of
the basic version (§ 444).
§ 450. In the preblems in horizontal-shifting counterpoint presently to be taken up, as
also in the later problems in double-shifting counterpoint, two forms of the basic version
will be distinguished. The first will refer to those cases in which no Jh- 0, the second
to those in which one Jh- 0.
CHAPTER XXII
THREE-VOICE COUNTERPOINT WITH HORIZONTAL SHIFTS
(A) First Form of the Basic Version for

Horizontal-Shifting Counterpoint: Jh= 0 Absent
§ 451. Problems in which three JJh are found contain a formula for the imaginary
combination for each pair of voices (§444); with Jh=0 absent, the number of these
formulas is therefore three.
The three formulas for the imaginary combination, which when united constitute the
formula of the basic version, define the entries of its RR.... (imaginary voices).
§ 452. Assuming that the problem is: In= ° + iih= ±2 + mh= 1, the JJh for the three
pairs of combinations are written out:
( jh= 0 + nh= 2 j jh, _ 2

(IIh--2 + IIIh~ !) /A"= — 1
m= 1
From these three indices are compiled the formulas for the imaginary combinations,
constituting the formula of the basic version:
i + nh.T.2, iih.T.-1 + in, i + m!1::.1.
Of the three imaginary voices contained in the basic version two are indicated by
the same numeral (II ). These two RR of one and the same P, i.e. of a real voice,
are indicated by the same figure. The third imaginary voice is designated by the other
figure (III....), and therefore functions as R relative to the second P. In this way the
basic version includes a double imitation with imaginary RR Of the three real voices
one of them (in this case I) enters into two of the imaginary combinations, being united
with the other two numerals II.... and III.... (combinations I +11™..- and I + III v7". - ).
These three voices— one real and two imaginary— must, as will be seen, always form
correct counterpoint: this is the derivative. The vertical shift absent, as in the present
instance, they reproduce the progression on the same degrees, and since h is not equal
to a fraction (§ 359), also on the same parts of the measures. Thereupon it is seen that
the other real voice (here III) enters into one imaginary combination (II..T. — + III).
Finally, the last of the real voices (here II) does not participate in any of the imag
inary combinations.
§ 453. According to the preceding it should be observed that with three imaginary
voices one of the real voices enters into combination with two of them; another real
voice combines with one imaginary, but the third real voice remains free and does not
combine with any of the imaginary voices. This is a general characteristic of all basic
versions that do not have Jh- 0 and where, therefore, there are three imag
inary combinations.
To demonstrate this let it be assumed that of the three real voices in the basic
version each one combines with one of the three imaginaries . In this case the formula
of the basic version would be represented by one of these two combinations:
(a) I +11.... , II +III.... , I.... + III or

(b)I.... +11, II.... + III, I + III
271
But neither of these is possible. From the way in which the imaginary voices are dis
posed in formula (a) it is clear that the first and second imaginary combinations have
h as positive but the third is negative. And since for an imaginary voice h- Jh of the
corresponding combination, Jh' and Jh" would be positive but JhZ negative, an impos
sibility, for MS, as equal to the sum of the other two JJh (§ 445), would also have to
be positive. The same reasoning proves the impossibility of formula (b), where JhS.
would have to be negative, but the sum of two positive quantities. Therefore all of the
real voices cannot participate in the imaginary combinations. Neither can any one of
the real voices participate in all three imaginary combinations, since one of the voices
is absent from each of the three pairs of combinations: I +11, II + III and I + III. There
fore only two of the real voices can enter into the structure of a two-voice imaginary
combination. But since there are three of the latter, one of the real voices must enter
into two of them, and the third into one.
§ 454. The following is a solution of the formula in § 452:
h:1 (IÔ+Tjh--,) Jh,:2
lh=o + IIh=.-2 + m I + IIh.:2, II1?::1 + III, I +111"h-l
dih=-2 +inh=i) Jh---i
JhZ:l
Ex. 248 (a)
Basic Version
EftfrHun)
o>)
•£.za
n &S 23p
m
^
*' o
m m s
m m *c: ~yr
THREE-VOICE COUNTERPOINT WITH HORIZONTAL-SHIFTS 273
Derivative
jhrO
(c) _o_
wm mm
41h:12
£ m m rj
e
hrl
3E s
111
Se no
=^
The real voice that counterpoints two imaginary voices is indicated by two stars
(**I); the one counterpoising one imaginary by a single star (*III). The two imaginary
voices that counterpoint one and the same real voice, II (+1) and III (+1), are
written on two adjoining staves, connected by bar-lines. The remaining imaginary voice,
which counterpoints one real, may occupy a staff either above or below the two former.
Observe that the derivative, given separately, is an exact reproduction of this com
bination formed by the real voice entering into the two imaginary combinations (**I), and
the two imaginary voices which it counterpoints.
§ 455. As before, in approaching the solution of problems, it is first of all best to
look over the general features and to place on the supplementary staves a clef for each
imaginary voice, putting it to the right of the clef for the real voice, at a distance equal
to the number of measures indicated by the figure for h of the imaginary voice. Next
write the real combination up to the measure where an imaginary voice first enters
(in Ex. 248, therefore, the first measure). This constitutes the first section of the
problem. From this section the real voices are then transferred to the supplementary
staves, as imaginary voices. This done, the next section is taken up (m. 2, Ex. 248), in
which is the first entry of one of the imaginary voices. At its first entry this rule should
invariably be followed: In each section the imaginary combinations (i.e. the union of the
real voice with the imaginary) are written in advance of the real combination. At this
point (beginning where all the imaginary voices have entered) the order in which the real
voices are composed for each section is as follows:
(1) First write the real voice that counterpoints the two imaginaries (in Ex. 248, **1).
It is composed in this way in order to form correct three-voice counterpoint with them.
(2) Write the real voice that counterpoints one of the imaginary voices (in Ex. 248,
*III). The reason for this is that it must form two-voice counterpoint separately with
its own imaginary and with the real voice of which part is already written.
(3) Finally write the real voice that does not participate in any of the imaginary
combinations (in Ex. 248, II). It is written as counterpoint to the two real voices already
existing, forming with them a correct three-voice combination. When a section is com
pleted in this way the real voices, now established, are transferred to the supplementary
staves as imaginary voices for the ensuing sections. These are composed in the same
order, the real voices again being transferred to the supplementary staves, and so on.
§ 456. It is seen from the preceding that each teal voice forms a contrapuntal com
bination with either of the other two voices. When these two voices are both real or both
imaginary it is written so as to form correct three-voice counterpoint with them. If one
of these two voices is real and the other imaginary, three-voice counterpoint is not
obtained, but the real voice first mentioned forms correct counterpoint with each of them.
§ 457. In regard to the real voice that unites with one imaginary (in Ex. 248, *III), the
following fact should be observed: the union of this voice with the imaginary is exactly
the same as the combination of the two remaining imaginary voice (in Ex. 249 cf. com
bination II.... + III beginning with the second measure, with the combination of the
remaining two imaginary voices, beginning with the third measure of the original). The
importance of this real voice, uniting with one imaginary, is that this combination
guarantees the correctness of the union of the imaginary voices at their recurrence. There
fore it must not be regraded as an independent two-voice combination, but always as
entering into the structure of a three-voice, being reproduced exactly in the derivative. It
follows that when this combination does not include the lowest voice of the derivative,
unprepared fourths may be used in it, but not, on the contrary, 8. In Ex. 248 this com
bination (II + III) can function as independent two-voice counterpoint only because
one of its voices is the lowest voice of the derivative.
§ 458. Since the combination of real and imaginary voices just discussed is only
of secondary importance, being essential only because of the other two imaginary voices
which counterpoint one and the same real (I + II and I + III , Ex. 248), this counter-
pointing of one real voice by one imaginary need not continue to the very end of the
of the problem. To continue the imaginary voice as far as the next to the last section
is enough, and if the derivative has a free ending this voice may be interrupted.
§ 459. In the preceding basic version the correct three-voice union of all three
imaginary voices in the first measures was entirely fortuitous. Generally even two
imaginary voices, counterpointing one and the same real, do not combine with the third
imaginary as in the sixth measure of Ex. 248.
(B) Second Form of the Basic Version for

Horizontal-Shifting Counterpoint: Jh= 0 Present
§ 460. Having examined three-voice horizontal-shifting counterpoint in which none

of the ] Jh were equal to zero, the next to be studied are those cases in which one of
the JJh equals zero. Of three JJh, two of them can not be equal to zero— this was dis
cussed in § 446— as the result of the equation Jh' + Jh" =Jh%. Therefore if only one
voice is shifted its position is changed relative to the other two, and still neither of the
two JJh that refer to the combination of the shifting voices with each other can equal zero.
§ 461. The presence of Jh- 0 means that the corresponding combination represents
only a case of simple counterpoint; that it can not form an imaginary combination and
must be excluded. The result of this is that the formulas to be examined contain not
three but only two formulas for the imaginary combination. These two formulas may
represent one of two cases: the voice common to both formulas can be either (a) real
(e.g. I.... + II and II + III.... ; I + II.... and I + III....), or (b) imaginary (e.g. I + II.... and
II.... + III; I.... +11 and I.... + III).
§ 462. Hence there are two aspects of basic version with two imaginary voices, here
after referred to as b.v. form 2 a and b.v. form 2 b, respectively. B.v. form 2 a contains
two imaginary voices, these counterpoint one real voice, and constituting with it the
required derivative. B.v. form 2 b contains one imaginary voice, which counterpoints
3DI0A-33HHI INIOdHSlNflOD HUM S13IHS-1V1N0ZIH0H SZ2
om; jeaj 'saaiOA \\b aajhi 9ui;n)i)SU03 ath .aAijeAuap ox ajEDipui sith AieuiSeuji 3010A
ui ah; ciseq uoisaaA aq} s|eJaiunu of thoq jeaj sceioA qiiM iphiM }] si pajeposse I|tm
aq pceejd in :sasaq}uajed ""H f) I 4 !(m ""l +) |] + .(m qacg of aq} 3uio3aifo -qns
suijfo si i|jiM3jaq •pa}Ej}snj]i
§ 3.gj, aso^) '/ .A.g mjoj 2 a. aqj.— ciseq uoisiaA smejuoo o»j A\ieui3eun :s30ioa
fe-:qll+o:.,!( »-=*'
0=ql + Jti{II + qlll 0-
iill( ?.= + )0=qlll ■S-W
z-Î + H II + z:qlII
.xa 6^3 (a) ciseg 'onisjaA jsjij j uijo

)H+(7.-i.'|I
)n+)s^ra
^^sr^g
(q) jBui3ioO
i
a3 fe£ ^
sn +-d±d+
%
£gm m »#
III -»«;<; - e£ rm W= *v znz
mM^~ffffjm*^l$P^&i:p^3m
nfc m
3113
Ma ^g
2=5E* iFT" J_ i -r; f^ b: _/? ; UT
£
S *£
(c) aAHBAuaa
' O. rO-
3C
a (t ^ JJJrl^M-«jo
^^
-11 = 2
stetefe
H*»*
si _^nr
^£S 3E
TT
^^FF^F1^
* B £
9LZ 318UH3AN0D XNIOdâiNnOD
^fÊ m ±*
3T s
^^ m=
*===a 31
s 3t
J -frl r
§ -^gp 3SDJ '2 8 'A' Zm±od *<! ahX- DISBq U0ISJ3A SUIBJU03 3U0 AJBUISBUII 30I0A
f-=,i + oîi + 0=,m

„.
rj + )II qr ^
o=MTI) + (ortflll 0="Hr
*--H\
-l
n+-i'ii
f-=,i + in
m
cisbq onisjaA poncas j uuo

-xa qSZ: |-"Î )ni+n+)
)B) IbuiSuO
)q) 3APBAU3G
t-=.d 5-—o— TT ^
dr~«r o—1 oJJ. JW "i—*^j — yjj-^j-j.
|JJ 1
*M* ".=,11
o— r^ J J e- " <j * =^g -°i J 1
ft-% o-nlll J j |J
ST •« .d plr^Ji \j
fi di J ffTfrFtWff 3*
u j irrrrJjj, »r ■*i n
§ -594, uoTjeuiujxag jo b jetcads uoijecijddB jo 3ui}jihs-iB}onzuoi( 'luiodjajunoc

pajjec juiodjajunoc mi*,, puB jnohjiM ,,S}S8J \\\iti aq ahi pafqns jo JajdeiQ 'AIXX
CHAPTER XXIII
THREE-VOICE COUNTERPOINT WITH SIMULTANEOUS

HORIZONTAL AND VERTICAL SHIFTS (DOUBLE-SHIFTING)
§ 466. At certain horizontal shifts in three-voice counterpoint each voice in the

derivative retains the same position relative to the voices, as regards altitude, as it
had in the original. In other words, all three voices are disposed according to diagram
(1): • • If the voices shift not only horizontally but at the same time vertically
their arrangement in the derivative may correspond to any of the six possible diagrams
(Obs., § 442). In solving problems the particular diagram according to which the voices
of the derivative are arranged must be taken into consideration; and in selecting intervals
for the vertical shift care must be exercised to assure enough room for the shifting voices
to move around in without confusion. Here one must be guided by the same considerations
as in the problems in vertical-shifting counterpoint.
§ 467. As before, two forms of the basic version are distinguished. First will be
examined those cases in which among the JJh none is equal to zero, then those in which
one Jh equals zero.
(A) First Form of the Basic Version for

Double-Shifting Counterpoint; Jh= 0 Absent
§ 468. In the formula for the derivative combination the vertical shift of the voices
must be indicated as well as that of the horizontal, besides which v for II must take a
double designation (± or + , § 241):
•3*3 h= 0 h=±3 h= 1
Iv= "3 + IIV= ±4 + IIIV= "10
Preceding the formula is the diagram governing the disposition of voices in the deriv
ative. According to this formula the values of Jh', Jh", Jh'L and those of Jv', Jv" , Jvl,
are then determined:
h= 0 h= 3
( Iv" ~3 + IIV~ 4 ) Jh' = 3, Jv' = 1
h= -3 h= 1
(nv- -4 + niv= -10) Jh„ _ _2t jv» - _14
/tt= l,/t;2=-13
Then, using the JJh and JJv found, formulas for the three imaginary combinations are
compiled, which together constitute the formula of the basic version (§ 444). In it each
imaginary voice takes h- Jh of the given combination and v- Jv of the same combination
(§ 373). In this way the formula of the basic version for the same preceding problem
assumes this form: ._ , ._ « ._ .
i + ny^.iiYf..-14 + in, i + mr.T.-13
277
According to this the clef for each imaginary voice is placed to the right of the
clef for the real voice at the distance indicated by h of the imaginary voice. The voice
that shifts vertically is transferred to a supplementary staff according to the indication
h=3
for its v. For example, in the combination I + IlY.T. the imaginary voice must be trans
ferred to a supplementary staff three measures later and a second below the real voice
h=-2
indicated by the same numeral; in the combination IlY.T. + III the imaginary voice
enters two measures later and two octaves below the real voice etc. See Appendix D,
No. 4, for a basic version written according to this formula.
§ 469. The two imaginary voices that counterpoint one and the same real must be
arranged on the supplementary staves in accordance with the diagram of the given
problem and in the positions which they take in the derivative, as indicated by the
same numerals. Therefore in the basic version referred to (App. D, No. 4), conforming
to diagram (4) ( ^^^^3 ), voice III.... (+ I) is placed on the staff above voice II....
(+ I). The third imaginary voice, not being in contrapuntal relationship with the other
two, occupies a staff either above or below the latter, depending on which it is nearest
to, according to pitch. Therefore in this basic version the separate imaginary voice
II (+ III) is placed underneath the other two.
§ 470. The introduction of the vertical shift does not involve any essential change
in the methods of working out problems, which as before are written in sections, one
after another. The parts for the real voices, written as far as the first entry of an imag
inary voice, are transferred to the supplementary staves, forming the first section, which
determines the size of all the later sections. A real voice, whether it counterpoints
two imaginary or two real voices must, according to the preceding, form simple three-
voice counterpoint with each pair. Counterpointing two different voices, one imaginary
and one real, this voice forms simple two-voice counterpoint with each separately
(§ 323). What was stated in § 457 about the combination of this voice with an imaginary
applies also here.
§ 471. If the combined use of the horizontal and vertical shifts involves any new
difficulties in the solutions of problems these difficulties are only of a kind that are
characteristic in general of vertical-shifting voices. The necessity of avoiding augmented
fourths and diminished fifths in the derivative, both as melodic progressions and as
intervals between any two voices (except as passing notes) adds considerably to the
limitations of these problems. Whereas in two-voice combinations such obstacles can
often be overcome by chromatic alteration, in three-voice combinations this is seldom
available because it can result in a cross-relation. Since one such mistake vitiates
the whole problem, in writing the basic version it is necessary to take into consideration
the possible appearance in the derivative of the forbidden intervals—the augmented
fourth and diminished fifth. This is most easily done when the combination of two real
voices that is reproduced as an imaginary combination, contrapuntally associated (in
the examples the supplemantary staves with connecting bar-lines), has /t>= 0, ±7, or ±14.
In this case a key-signature for the imaginary voices (always the same for both) must
be chosen which causes the voices to appear on the same degrees of the scale as in the
original, and consequently also in the derivative (because of the JJv indicated). By
doing this neither augmented fourths nor diminished fifths will be encountered in the
derivative, for they are absent in the original (cf. Nos. 3, 4, 6 in Appendix D). In order
not to multiply difficulties a remaining precaution is that the real voice which counter
points the two imaginaries has a melody that is valid with both key-signatures (§ 333). At
THREE-VOICE COUNTERPOINT WITH HORIZONTAL AND VERTICAL SHIFTS 279
other J Jv, where both imaginary voices can not be on the same degrees as in the original,
care must be taken that a voice functioning as imaginary does not start on the same
degree of the scale as when it functions as real, and that in each section it does not
form an augmented fourth in either case. The signature for the two contrapuntally asso
ciated voices having been decided, a signature for the derivative is next chosen that
will give the corresponding voices on the same degrees of the scale. The third imaginary
voice must then take a signature which will put it in the same relative scale-position
as this voice has in the derivative.
§ 472. Sometimes during the writing of problems it appears more convenient to
change the signature from that adopted at the beginning of the two imaginary voices.
Such instances can not be brought under any rule. The overcoming of this or some other
difficulty in each individual case helps to build up the experience in contrapuntal writing
which these exercises will develop in the course of time.
§ 473. Before proceeding with exercises it will be of advantage to study minutely

the six examples in Appendix D that have often been referred to.
(B) Second Form of the Basic Version for

Double-Shifting Counterpoint Jh= 0 Present
§ 474. In three-voice horizontal-shifting counterpoint a combination having Jh- 0 is

excluded from the formula of the basic version, as was the case with simple counterpoint
(§ 461). If with the horizontal shift is incorporated also the vertical, this rule remains
in force only for those situations where, except in ]h- 0, the given combination also
has Jv = 0. If, at ]h- 0, Jv± 0, this combination is an instance of vertical-shifting coun
terpoint and enters into the formula of the basic version, not in the capacity of imaginary
but as a real combination, having its own Jv. The formula of the basic version in this
case will consist of three formulas; two of these are formulas of the imaginary com
bination, one the formula of the real combination with determinate Jv . Here for the first
time is found the requirement that in counterpoint with one derivative two real voices
shall satisfy the conditions of vertical-shifting counterpoint. For instance, in the formula
for the basic version in Ex. 251 there is, besides the two formulas for the imaginary
combination, also a formula for the real combination, with the requirement that the real
voices II + III shall be written in double counterpoint at the tenth, namely: (II + III)
Jv= — 16. Exactly the same formula for the basic version is found in Ex. 252; included
in its formula is: (I + II) Jv = 2.
§ 475. The second form of the basic version now to be examined, appears in two
aspeacts (§§ 461-2), which as before are referred to as b.v. form 2 a and b.v. form 2 b.
The characteristics follow by which these two sub-forms are distinguished. A voice
found in both formulas for the imaginary combinations may be either real (b.v. form 2 a),
or imaginary (b.v. form 2 b).
§ 476. Case 1, B.V. Form 2 a.—If the voice that recurs in both formulas of the
imaginary combination is real (e.g. I in the formula I + II I + III or II in I.... + II,
II + III....), there enters into the construction of the basic version two imaginary voices,
designated by different numerals, both of which counterpoint one real. These imaginary
voices are shifted at the given Jv of the real combination, wherein Jh- 0. In counter-
pointing the real voice they form a three-voice combination, which is the derivative.
In Ex. 251 the real voices, II and III, are written in double counterpoint at the tenth
(/t>= —16). The two imaginary voices represent a shift at this Jv. The problem is written
section by section. The first two measures (as far as the entry of the imaginary voices)
are written not in simple counterpoint as before, but so that II and III form double coun
terpoint at the tenth. Then the voice parts are transferred to the supplementary staves
in accordance with vv of the imaginary voices, in addition to which they show a shift
at Jv= -16. Next the second section is written (third and fourth measures). Since the
imaginary combination must be written in advance of the real (§ 455) I is done first of
all in a given section, so that it will form three-voice counterpoint with the two imaginary
voices. Then II and III are counterpointed to I, but so that the two former voices will
be in double counterpoint at the tenth. The voice parts in measures 3 and 4 are now
transferred to the supplementary staves and the same process applied to the third section
(mm. 5-6) and so on to the end of the problem.
,-o (Ivlo+IlvE§) Jhz2 jv..3 .hz* TlInh:2

>^ i$i8+n«a+nw "13 (Tlvi3+Il]v:-13) Jh" = 0 Jv"=-16
1+ HVF.3, MIIyAl3,(U+III) Jv:-IG
Jh£ = 2 Jvlz-13
Ex. 251 (a) Basic Version First Form
my?rw(+l)
Original
Derivative
(b) . ^ mhr O
v;-13 o
IE
$EEE iHLeÊ
=£^
.hi
lv;(p
-£i
fay
I^^^P pBffSRf »
EE
p^33
pa*
U6lg
rt IE <V-
3g 3E
p Siw =f g? ZZZTO
S
§ 477. Case 2, B.V. Form 2 b— The characteristic of this sub-form is that in both
formulas of the imaginary combination the imaginary voice is represented by one and the
same numeral (e.g. II + III.... and I +IIL... ; I +11.... and II.... + III etc.). Since this
voice enters at the same time in both combinations, then in the absence of the vertical
shift it also enters at the same altitude in both. The result is that it could be written
only once, as in Ex. 250, where the basic version has only one supplementary staff.
When the vertical shift is incorporated with the horizontal this imaginary voice must be
taken at a different altitude for each combination, and therefore must have a double
presentation, although the time entries for both combinations are identical, as in the
basic version of Ex. 252. This sub-form of the basic version exhibits still another
peculiarity: the absence of that combination of real and imaginary voices which appears
as the derivative; also the imaginary voices do not represent a shift of the real voices
at the given Jv, as in the basic version of the first sub-form. The derivative, since it
is broken up into its units, can not be written until the basic version is completed; the
shift is then made according to the requirements of the problem.
== teg+nfea+ntfii ,&ffi} *=0 £=■ (MDfcOMntg, I+I0&

(]lv;-2 + l|lvr4) Jh rl Jv'i2
JhSil jv£:4
Ex. 252 (a) Basic Version Second Form
invil(+i)
JV:2
î^^fipi^^P
^j&p-^Hf^z^£$££m ^m
w*m m m f-e- 331 3=t
m&me V O ^^
XI
m *HE=_
In view of the fact that the lower voice does not participate in the real
combination, having Jv= 2, the free use of fourths is possible, which at this
Jv are variable dissonances. Since an original fourth at ]v=2 gives a deriv
ative imperfect consonance (3 + 2=5), it may be regarded as having the char
acteristics of the latter, hence the fourths on the unaccented beats in the third
measure, the two parallel fourths in the seventh measure, etc. Releasing fourths
from their limitations facilitates considerably the solution of problems. Furthermore, the
voices of the derivative are arranged according to diagram (1), consequently I and II, with
a shift at Jv= 2, remain as upper and middle voices in the derivative. Hence it is neces
sary to exclude from the original the interval that gives a derivative 8, namely, 10.
§ 478. If in double-shifting counterpoint all JJv= 0, horizontal-shifting counterpoint
is obtained; if all JJh, vertical-shifting; if all JJv and JJh are equal to zero, simple
counterpoint. Therefore simple, vertical-shifting and horizontal-shifting counterpoint may
all be regarded as special cases of double-shifting counterpoint.
§ 479. Combinations giving several derivatives will not be dwelt upon. To obtain
them with the aid of a basic version, increasing the number of its imaginary RR, would
be too difficult. Each additional R , imposing new restrictions on the real voices that
counterpoint the imaginaries, only multiplies difficulties, and the problems are hard
enough without them. The solution of such problems is possible with the aid of a basic
version that is contrapuntally correct. But increasing their RR also increases the defects
that are characteristic in general of such versions: the unavoidable emptiness of the
three-voice combinations extracted from them. These versions may be obtained rather
by writing two contrapuntal voices (Cp. and Cp2) to a three-voice imitation, which gives
an original and two derivatives (Cp. + Cp2 + P'< Cp, + Cp2 + /?i, Cpi + Cp2 + /?2)> or by
writing a fifth voice (Cp) to a double canon, which gives an original and three deriv
atives: (Cp + Pa + Pb, Cp + R* + Rb, Cp+P* + Rb, Cp + Pb + «a), etc. Theoretically
such cases present nothing new, and examples will not be quoted because of
lack of space.
§ 480. With the study of three-voice double-shifting counterpoint the problems are
ended. The shifting of voices in two directions, vertical and horizontal, has been in
vestigated, and the methods whereby can be solved any problem of shifts in either
direction, as well as all possible combinations of shifts in both directions. The two
chapters following may be regarded as supplementary.
CHAPTER XXIV
COUNTERPOINT WITH AND WITHOUT RESTS
§ 481. In explaining the technique of horizontal-shifting counterpoint numerous

examples from different eras were quoted where both the original and derivative com
binations—sometimes the basic version—entered into the structure of a composition,
serving as thematic material. As supplementary to this study some very infrequent appli
cations of this counterpoint will be examined, in which the entire composition presents
an original combination that can be executed in another way, as derivative. This process
was known in the Flemish Schools as counterpoint "with and without rests." In music
histories mention is often made of a pupil of Josquin de Pres, Pierre Moulu, who wrote
a mass that can be performed in two ways: as written, and also by omitting all rests.*From
this mass, containing movements in from two to five voices, two examples—two-voice and
three-voice— will be quoted, in order to explain the essential features of this process and
to show that such a problem can be solved by means of what has already been learned.
& 482. Rests found in the original are omitted in the derivative. However, not all of
them are omitted, only those larger than a half-rest. This, and all rests of smaller values,
are retained in the derivative. The half-rest, as will be seen in the examples, often
precedes the first note that enters after a voice has been interrupted, and together with
this note is transferred to the derivative. The note with which a new phrase begins may
be of any desired length, but the note on which it ends occupies ordinarily not less than
half a measure, because of the necessity of ending a melody quietly and not cutting it off
abruptly on a short note. If some intermediate notes of a phrase occupy three-fourths of
the measure, followed by a half-rest, and in the next mesaure a rest of the same value
precedes the note of entry:
2=^
A^-<h_L^ ZEE
both rests, equal in value to a whole-rest, must be omitted in the derivative:
t * ES ICE
Such cases are not found in the derivatives of the following examples.
Obs.—The silencing of voices on a long note, whereby the melody ends naturally with a
cadence in one part while the other voices continue their independent movement—the
entry of a voice after a rest where the harmony needs completion, so that both harmonic
and contrapuntal requirements are satisfied (as when the entering voice imitates)—all
•The manuscript of this mass, entitled "A deux visaiges et plus," or "Missa sans pause," is
in the Public Library of Cambria. While the first edition of this book was in course of preparation
the author was enabled to obtain a photographic copy through the efforts of V. A. Bulichev, who
went from Paris to Cambria for this purpose, and to whom the author herewith expresses his
thanks. [Cf. Grove, "Dictionary of Music and Musicians," article Moulu. —Tr.]
284
COUNTERPOINT WITH AND WITHOUT RESTS 285
these are essentials of good contrapuntal style. In analysing it is necessary to pay
special attention to the technical details that enter into the structure of a beautiful
melodic line. It is impossible not to observe the beauty of the entrances in Mozart. In
this connection his later string quartets, especially the one in C major (K. V. 465), should
be the subject of constant and careful study. A finished work of voice-leading can only
be the result of a course of study that starts with simple counterpoint, followed by
imitation, especially imitation as applied in choral-figuration. Reference may again be
made to Bellermann's work on counterpoint, the chapter Von der Nachahmung (pp. 293-5),
also to Bussler's Der Strenge Satz, §§ 29-30.*
§ 483. The omission of rests in the derivative has the effect of shifting a melody
to the left. Such a shift for the upper voice will be indicated by the sign +, for the lower
by —, for the middle by + (§§ 441-2). The omission of rests will be called rest-removal.
This term refers, exclusively, to all rests larger than a half-rest.
§ 484. Since the omission of rests shifts the voices only horizontally, their pitch
remaining as before, counterpoint with and without rests is a special case of horizontal-
shifting counterpoint, not of vertical-shifting; otherwise there is nothing to prevent its
use with vertically-shifting voices also (cf. Ex. 258).
§ 485. If rest-removal were found in one of the voices only at the beginning, this
would be an ordinary case of horizontal-shifting counterpoint. But in the examples to be
studied, rest-removal is found sometimes in one voice, sometimes in another, giving
melodies that are divided into phrases. Therefore an instance of this kind must be re
garded as the association of several examples in horizontal-shifting counterpoint.
§ 486. Each time that a voice shows that rests have been removed there will be
placed above the staff, immediately after the last rest, a figure equal to their total sum,
taking as usual the measure as the unit of value. If the highest voice, no sign will be
placed before the figure, + being understood; if the lowest voice the figure will be
preceded by the sign — ; if the middle voice, by + After this figure is a comma, followed
by h, indicating the shift of this voice in the derivative. The quantity h is equal to the
sum of the first rests: . .
a, h-u
-©-
■S
But following this, rests that again shift the voice in the same direction are indicated
by a new h, equal to the sum of the new rests, added to the preceding h:
Ex. 253 (a) J hrli
l. h-M j.h;3
IS 5 6
r°!■oi it
2, h=5
-OZ
F&^Fi^
■ j ■
The same voice as derivative:

00 ,n:2A lh:3 . 1h=6
Z3E
W*=^ rrr"KH O _
2
•Cf. also: Goetschius: "Applied Counterpoint," Chapters X and XI; Dunhill: "Chamber
Music." —Tr.
The concluding note where each phrase is silent is indicated by a star: *. By numbering
the measures each shift to the left is easily verified. For example, the second phrase
in the original enters on the second half of measure 5, and has h- 2l/i; therefore in the
derivative its entry occurs at the beginning of measure 3. The third phrase in the original
(h = 3) begins on the second half of measure 8, shifted to the second half of measures 5 in
the derivative. Next comes h= 5, therefore measures 14 in the original and 9 in the deriv
ative correspond, etc.
§ 487. Such problems are solved by the methods previously described: with the aid
of a basic version written according to the formulas of imaginary combinations. Here the
only difference is that instead of one formula for the imaginary combination (= basic
version) each problem must have several of these formulas, because of the changing
values of Jh.
(A) Two-Voice Counterpoint With and Without Rests
§ 488. The first formula of the derivative with corresponding Jh, and according to it
the formula of the imaginary combination, are compiled in their entirety, as in all other
cases of horizontal-shifting counterpoint. As long as rest-removal is not in effect in
either voice the imaginary voice is written without interruption. If rest-removal is in
effect in a real voice that is identical with the imaginary, and only in this one voice,
the imaginary voice is written as far as the end of the given phrase, stopping on the note
indicated by the star in the real voice (cf. imaginary voice in Ex. 254, mm. 11 and 22,
and similar places in Ex. 255). After such interruption of the imaginary voice a new
formula for the imaginary combination is substituted for the previous formula; the writing
of the imaginary voice then continues according to the new formula. If in this the ima
ginary voice remains the same voice as it was in the previous formula, it enters some
what later than the first entry of the real voice after the latter is silent, as in Ex. 254,
where the imaginary voice continues the same (I ), and I enters after the rest in
measure 11. The new formula (I..T. ) is now applied, beginning with measure 12; the
same I enters after the rest in measure 21, and a new formula goes into effect in mea
sure 22, etc.
If rest-removal is effective in a voice that is not identical with the voice it counter
points, the imaginary voice is interrupted at the place where the voice counterpointing
it is silent, i.e. at the sign * (Ex. 254, m. 17). At this interruption the previous formula
of the imaginary combination is cancelled, and a new formula takes effect at the next
entry after the silencing of the real voice; if this entry precedes the rest-removal (§ 482)
then a quarter of a measure earlier.
§ 489. Whatever rest-removal is in effect— consequently a new h— a new formula for
the derivative is compiled, the voice concerned takes its new h (immediately after a
rest), and the other voice the h that it had before. Jh is equal to their algebraic sum and
defines the value of h of the imaginary voice in the new formula (cf. the formula for
the next examples).
If the new imaginary formula is identical with the preceding formula its repetition is
not necessary, and the imaginary voice is continued without further interruption. If in
the new formula h- 0 the imaginary voice is discontinued until the next for
mula takes effect.
§ 490. The following example from Moulu's mass, with the accompanying formulas
of derivative and imaginary combinations, illustrate fully these explanations.
Pierre MoulU. Plenisuntcoeli.
Ex. 254 (a) . I*:*(+ll)
Original (with rests)

^TW
I i£M& E f'_Ê^Bg 2f ^p:
IJ.3
£
-t. h:-2
«= ~tv UXI
iiffiHi '
l,hzl
T7-^
an W=
P £= £ wm?
P g> |3 -«-
IP- rrr
j^^^î^â
J J .1 iTf' r J p J fi o g
m
fe g f r r f- u;
• -i,
J, W
u
i^P^pp -**-
ft:? Jh...
h-l
^-"^—f ^ ^l^^^^^ÊEÊ^ ip
l,h:2
- p Tl s> rT
^
h:-3
pp r P¥?
fC CC
P £
•In the examples taken from this mass the liberties in the use of dissonances, characteristic
in general of the old Flemish composers, are not indicated. The reader familiar with the rules
of the srtlct style can easily locate these places himself.
Formulas for the preceding example:
Meas. 3: (I + IIh = _2) /*- -2; A""2 + II

" 11: ah = 1 + Hh=~2) M--1; Lh.7.-1 + II (m. 12)
" 18: (Ih = 1 + IIh = -3) Jh=-2; fiZ'2 + II
" 21: (Ih = 2 + IIh = _3) M--1; ii?.?.""1 + II (m. 22)
(b) Derivative (without rests)
I SE "• y
rrc PS j3 =£
Thr-2
34= 3E -O- 3E
II ^
££P
Ihrl
Egfg Jrf 3E **
P =P
a « »• r» » »—» 77—^
I TT
£=3? * j ,i T~^=r * e e ^ at
^ P<f
**:
^_JEh:-3
* g *
h:2
j I
£= « =qz - rj n—n o 22
£
f 0 P- 0 •- p 3E
K^^ PiP? £
LTi-IL
I *pp? 4 4
*3=
M^f-f-ff 3X:
§ 491. In the preceding example from Moulu (and also in two other two-voice sections
from the same mass, the Crucifixus and El ascendii), Jh in all derivative formulas has
the same sign (negative), therefore the imaginary voice also remains constant: I,... . In
the next example Jh is alternately negative and positive, giving for imaginary voices
I.... and II.... respectively.
Ex. 255
Original (with rests)

^l - ° f
i -A\
S ~rt~
l!'.:-JnD)
^TTTf^^^
3E
US h;-25
3E
jgpg HZ s -t—f as
Meas. 2: (Ih~1 + I1) /A= 1; I -h II*1.- X
n 6: (Ih = l + "h = _1 *> /A = -& *"•~lVl + °
•■ 9: (Ih * 2 + Hh = -1 Vl) /* " V* I + ««" V2
<> 12: 0h = 2 + "h= ~2Vl) Jh= Hfc fcl"* + "
Derivative (without rests)

(b) ,h-.i
3E
zaz HZ
w g ■
£
^-F
H pp^p ffw^fm
♦=F
If g TT
m 3E
lhz-25
IF 3E
m mi
(B) Three-Voice Counterpoint With and Without Rests
§ 492. When rest-removal is in effect in any voice of a three-voice texture two

formulas for the derivative must be compiled, taking the voice concerned alternately in
combination with each of the other two. The procedure according to each formula is the
same as indicated in §§ 488-91. All that was said regarding imaginary voices applies
also here.
Pierre Moulu
I>!..3(+m>
Ex. 256 (a)
«i »v :nr
Hi
h--3- a rO
:3, h:;3
a s ti pE=
^M I W g^
&m
F^»= " o - p w
n W
3,h,3
m
tt^-£ff
■ n
m
g r-i
ifl r;
g £=M
m ^m
u ^m-fi-
i,h--*
w
m TFi
#-P- <»—p~ m
(+U+I1I
j —. ) . , -)-—_ __ _ , , 1 | . —1 -j ,. _ r*i
2,
»1 . n
Jo
1, h-;4
• .
m
O
3E
=?E&
f^^f
-&*- 3E
M-m-U--
^ W^
"*
^g £ it:
25
iai
27
Meas. 4: (II + IHh = ~3) Jh" = -3; II?."-3 + III
G + HIh=-3) ]hl =-3; I^= —3 + III
" 7: a + Hh = _3) Jh' =-3; Ih.= _3 + II
(IIh = 3+ mh = ~3) Jh" = 0
" 12: ah = 3+nh = ~3) Jh' = 0
(Ih = 3 + mh = -3) Jhl= 0
" 15: aih = 3 + IIIh = -4) Jh" = -1; IlKr.-1 + HI
ah=3 + mh —4) jki = -u ih.=-1 + iii
" 18: 0h = 3 + Hh = -4) Jh' =-1; I?.= _1 + II
(IIh = 4 + HIh = -4) Jh" = 0
" 25: ah = 5 + Hh = _4) Jh' = 1; I + H^.T1
(Ih = 5 + mh 4} Jkl= |. I = IIIh = l
**' Derivative (without rests)

TT=-£-ri ■P--P ^ 71 ZZ
£3£ *¥
P .IIh;;3
n ■ IE a a a
g^ 33: 5a
^^
h.-3
111
^^
^F <*=}
£ F»=
£
^-
. 1 h:3 fS^rt
*yg~>^ r^-^ ■-* ■ a a
P P
gf>; # J P fzc »rv a uc 2=â a Or Ft» ^
B P*f^
• HI h = -4
P
ti a _ p SP 9 pg
S^
sppp m. II h::4
-2ii
W. 0
#
zn P 1 It
P
P^ §JPp^ m
gg^g HE =£ m
B g g *-f-0-
lh:b
1s= ^ 0- Jl-g- -p^T5>rrr--

zbg =«=
^^
■EZS
1^ =££h£ fffE
Khz
»i: . I h—«-
2^ £§ fe
§ 493. In conclusion is quoted a three-voice example from Berardi's book,* illus

trating counterpoint with and without rests, the only specimen of the kind that the
present author has found in theoretical literature.
Angelo Berardi
Ex. 257 (a) fr:2 (+ID
^ i*i hr;2
ii tm 3E
m$
h:-l
in m 3E
f= ~o~
« iV:-.^(+ll)_» I1?::2 -+II>
•Angelo Berardi: "Documenti armonici." Libro Primo, doc. 30. "Cantilena con le pause
a senza."
►6Z 31HI1H3AN0D INIOdaaXNOOD
"*m-t-m-4- 4 *
,,IH - u... L_J * -■ o f
,Jtt ■
a h [J ' JJ1 J
ffl o ".
e- nr-" r«- ~M
-*J"J«fr-
i i Njjjj
— *
", _£.. ....... n«
sJ- J —5 " r»
I e i i
fe* ^rg^â
H+),-:,,I + 'in
==«s* E^^^^pE^^TÎ^:^^
^
?i--i
Sat
rt
s
lg^ p £=#I»_J if^TTJ

T- ?2--q
® n ZEC
-W-
j seyniujo joj ath Suipacajd :ajduiBxa
.Z 0 + (i=hiii S¥/ = :i- x-Uql +) (III

iO + (x-=Mm ..V/ = u- .=qll x- +) (III
.£ I) + (2-=MII .¥/ *--
Z--H1 + II
MII) = 2 + HHI = (x- u¥/ II + TqHI T
= 'l
.9
(z_=HII+yT=hI) = ;t/ii- e/ix_î + ii a*«o
.1/
(x-=mih+^-mi) 5¥/ = zA- "I+Ki—H1
.L hii) = c + C%x_=hiii .»/ = M H + =ÎH zA •<«) (8
HIII+yi-HI> = Vxx- x-=.î + III
3V/ = x-
•8 MII+%-Hl) = (yiZ- .f/ = z- n+t--¥
MIl) = t/lZ + HIII = .%I_ II + x*=*în
1 ..V/ = I
y.x»HI) + (yiz-=MII x-.=.Î + II •*) (.11
.u .v/ = x-
v,i=mi) + miii (yi— XV/ = 0
.z\ HC-1Ifl + lm 5tt—( ..V/ = 0
yii-HI) + (y.3-=MII
3Y/ - T- • ql = x- + Ill
Obs.—Stated in this way, the formulas show that the real and imaginary voices must
agree. But the same problem can be solved by a similar method, with the aid of all three
formulas, because of the fact that the voice entries fall definitely into three divisions:
(1) from the beginning to measure 6; (2) from the entries of the voices after the rests
in measures 6-8 up to measure 10; and (3) from the new entry of the outer voices in
measures 11-12. These divisions may be written as three different problems. But in this
case the measures where the divisions coincide would have to be written without the
corresponding formula, then compared with the derivative in order to avoid inaccuracies.
(b) Derivative (without rests)

-4 L
" 5
w 1 hr;2
zee
¥=
lg ~n~
rrrr HE
m It1 nr
i j ° £
■ it:' -
w
- r>
\h-.-M
w f p> fljJ
zai <**e
m mh;.ii
t'-f-w
^
£
B?k " jOZ
mi
. I",**
acn 3T
s
^m. mh:-^ papo -A

m ^=3
* ITU
-w
S 494. Compositions featuring this kind of counterpoint— "with and without rests"—are
little more than historical curiosities. On the contrary, more freedom in using counter
point admitting of horizontal shifts, as one of the resources of thematic development, is
of much greater scope, as is proved by the numerous shifts in the examples quoted. This
counterpoint was known to the composers of the Flemish Schools. Passed on to the old
Italian and German composers, it was extensively used at the time and is still a resource
of contemporary music (cf. Ex. 244, from Glazounow). These facts should have aroused
the interest of theorists and led them to an investigation of this counterpoint—to study
its characteristics and to establish its place in the other categories of complex counter
point, and if only in a general way to explain how such contrapuntal combinations are
brought about. What light does present music theory throw on this matter?
§ 495. The earliest known examples of this counterpoint in theoretical literature are
found in Vicentino's L'Antica musica, published in 1555, three years before the pub
lication of Zarlino's treatise. These examples illustrate double-shifting counterpoint.
Following the terminology of his time, the author calls his counterpoint "with and
without rests;" he bases his treatment on a common characteristic of double counter
point (the inverse shift) and disregards the other characteristics by which the horizontal
shift is defined, thereby referring his examples to the province of double counterpoint, to
which he devotes a chapter. Vicentino uses the term "double counterpoint" in the same
sense in which "complex counterpoint" is here employed; he divides the entire domain
of counterpoint into two categories: simple and double, a classification still found in
the latest works.* Vicentino's examples are chiefly of historical interest—they can not
be regarded as models to imitate. Observe in both derivative combinations the proximity
of fifths and the hidden fifth at the beginning of Ex. 259, evidences that the author had
not fully mastered the limitations inherent in the problem.
Ex. 258
Vicentino
L
lf^
Original
£=
-H-— HE
ft ZEE
Derivative
lv;-7
Ex. 259
Original
44- i=i
1i1 f
Derivative l]hf J„
Ûi
m i^-f-i j j J
lv;-4
Irreproachable as to technique is the question from Berardi, Ex. 257, which as

already stated is the only three-voice specimen of this counterpoint found thus far in
theoretical literature. It dates from the late seventeenth century (Documenti armonici,
published in 1687). This example proves that the traditions of the Flemish composers
regarding the resources of complex counterpoint were not yet lost. The length of this
example— fifteen measures— precludes any idea that the horizontal shift of three voices
may have been a mere coincidence. The writer could not have been unacquainted with
•It is hardly necessary to dwell on the fallacy of such a classification. Since the character
istics of double counterpoint are represented in the inverse shift, cases where the shift is direct
would be excluded from the domain of both simple and double counterpoint.
the methods that yield such combinations. Had his skill been of a purely empirical nature
it would have been concerned with problems of a single definite form, or would even
have generalized to a greater extent, but no light is thrown on these questions in his
text. The author limits himself to the statement that he could have quoted other com
positions of no less interest than this (he does not say that the example is by himself);
in his own words such examples prove that "musica est res profunda", but as to how
they are written he is silent. No doubt this is an illustration of a fact often mentioned
in music histories—that the old contrapuntists kept their knowledge secret, passing it
on to the elect only verbally. Certain expressions used by Berardi in reference to the
example quoted seem to confirm this supposition, for instance: "per dare un poco di
lune, come si devono comporre simile artificii ..." If this was the attitude of the old
contrapuntists their efforts were certainly crowned with sucess as far as horizontal-
shifting counterpoint was concerned. That this was the state of contrapuntal study in
the eighteenth century is clearly shown by the fact that one of the best-known theorists,
Kirnberger,* in combining two melodies in double counterpoint, included, among other
shifts, the following combinations in double-shifting counterpoint:
Ex. 260
Kirnberger
II" i_° i »—
rFfKfft
Q_
Se 3E
SBE ^ £ m
He places these examples in the division entitled "Von dem doppelten Kontrapunkt",
but did not accompany them with any remarks nor did he say anything about horizontal
shifting. If Vicentino, the sixteenth-century theorist, incorrectly referred horizontal
shifting to the category of double counterpoint he at least classified such examples,
calling them "counterpoint with and without rests." Even the eighteenth-century theo
rists took no notice of the horizontal shift when it appeared among other shifts and
apparently gave no examples from which to quote.
Of the theoretical literature of the nineteenth century, Habert's work on imitation is
worthy of notice. In analysing the imitations of Palestrina Habert does not pay enough
attention to the numerous cases of horizontal shifting where two subjects are concerned.
Referring to these cases he says that at first glance they seem to be in double counter
point, whereas in fact they are not.** In this quite correct observation is seen the first
•J. Ph. Kirnberger: "Die Kunst des reinen Satzes," II Theil, II Abtheilung (1777), s. 177-81.
5-ter Abschnill, "Von dem doppelten Kontrapunkt."
**"It would appear that double counterpoint were present, but such is not the case," Habert,
"Die Lehre von der Nachamung, 1899, S. 208.
step in advance since the time of Vicentino in the study of horizontal-shifting counter
point. Habert treats this feature of horizontal shifting as an accident peculiar to the
combination given; he advises to continue horizontal shifting only as far as it forms
counterpoint with the other voices (p. 208), but he gives no indications as to how
shifting the subjects horizontally gives a contrapuntal combination.*
If it is added that Riemann called attention to the horizontal shift in contrary motion
in two of the fugues from The Well-Tempered Clavichord, as also did Prout to the
shifting of the two subjects in the E£7 organ fugue of Bach (Ex. 243), it is clearly seen
that more than three centuries have elapsed before any material advance has been made
in the study of horizontal-shifting counterpoint. It may be asserted without exaggeration
that the questions brought up in Part Two of this work have hitherto not only remained
unsolved but have not even been stated, and that the study of them as here presented
is the first of its kind.
•And this same policy of evading an important issue is still prevalent in the twentieth century.
The statements of at least two eminent writers confirm this. Dr. Percy Goetschius ("Applied
Counterpoint," 1902, pp. 277-8) quotes two examples in horizontal-shifting counterpoint and refers
to two more in a footnote. In the text he says (p. 277): "These are purely experimental, and
cannot be obtained by any method of calculation." Also Dr. Donald Tovey, referring to a case
of the horizontal shift in Contrapunctus vm of Bach's "The Art of Fugue," says: "The abnor
mality in this relation . . . consists in the fact that it is not only at a different pitch, which in
itself would mean inversion in a double counterpoint other than in the 8ve, but that it combines
at a different point, a device for which the books have no name" ("A Companion to The Art of
Fugue," pp. 19-20). The place he refers to is m. 125, cf with m. 39. —Tr.
CHAPTER XXV
CONCLUSION
§ 496. The study of shifting counterpoint is a most important and extensive phase
of complex counterpoint. Closely associated with it in the present work is counterpoint
that admits of duplication in imperfect consonances. In discussing the functions of the
basic version the method for obtaining derivative combinations with altered note-values
and other changes of form was indicated (§§ 338-9), cases that also refer to the province
of complex counterpoint.
A summary of its divisions was given in the Introduction, and of these only one is
absent from this work: metamorphosed counterpoint. This phase of the subject, found in
the older textbooks (Vicentino, Zarlino), also in later ones, is applied—among others— by
Bach, in two of the fugues from The Art of Fugue. It can be explained very concisely
and a sufficient mastery of it should take only a comparatively short time. Best of all, it
can be studied in Marpurg's Abhandlung von der Fuge.* With the exception of metamor
phosed counterpoint all other kinds of complex counterpoint are represented in this book.
§ 497. Since multi-voice counterpoint in the strict style, both simple and with certain
changes that have already been dealt with in the study of three-voice counterpoint, there
would be no purpose in explaining the theory of shifting counterpoint for a larger number
of voices. Such a theory would present nothing new, and the difficulty of working out
problems would increase to such an extent that although they might be theoretically
possible they would often prove incapable of solution. At the beginning of the study of
shifting counterpoint attention should be concentrated on two-voice counterpoint (Di
visions A and C, especially the former). These divisions may be studied independently,
in succession, or concurrently. As to specific instructions, see Chapter VIII, also
§ 275, and the observations to §§ 290, 340, 423 and 482.
§ 498. In passing from strict to free counterpoint it is necessary to have a good
command of harmony. The harmonic attitude toward music, which coincides with the
beginning of the seventeenth century (rise of opera, the homophonic style, figured bass),
appeared as the exact antihesis of strict counterpoint. Multi-voice music was not the
result of combined melodies, but that of harmonies (chords) in succession. Instead of
all voices sharing equally in the texture of a composition, one principal voice pre
dominated, accompanied by chords—the system of ecclesiastical modes was superseded
by the present tonal system. The synthesis of these opposing forces resulted in free
counterpoint. The ideal relationship of these elements— counterpoint and harmony— is
that by the union of voices that move freely and independently harmonies arise that
satisfy all the requirements of tonality and modulation.
In the free style the use of intervals is entirely different from that in the strict, and
varies according to the harmonic context. At times their use is in agreement with the
strict style, at others there is more or less departure, finally all connection with the
strict style is broken, as harmonic function takes control.
•Part II, Chapters 3 and 4. [See also the articles Augmentation, Cancrizans, Diminution,
Recte et recto, Rovescio, in Grove's "Dictionary." —Tr.]
299
But the possible effects of the strict style on the free is not abrogated; dissonant
intervals, for example, that enter into the structure of chord-figuration, can be made to
conform to its conditions, so that these effects do not prevent the chords remaining in
the tonal system (cf. §§ 232-5).
The general features of the strict style, penetrating in some form or another the free
style, exert on the latter the most salutary influence. The result of this influence is
logic and consistency, and a good harmonic basis is secured for both instrumental and
vocal music, especially vocal.
The variable conditions of using intervals that are not subject to exact control
prevents in the free style the statement of rules for vertical-shifting counterpoint in the
simple and definite form that the strict style renders possible. It may be said that an
independent theory for shifting counterpoint does not exist in the free style. Certain
features common to both the strict and free styles may serve as guides, for instance in
double counterpoint, where the practical application of the strict rules are of value, as
in the establishing of limiting intervals, the forbidding of similar motion at indices of
the second group, etc., often the rules of harmony apply. The strict rules offer something
tangible and permanent, those of the free style vary according to the opinions
of different theorists.
The dependence of contrapuntal textures on a harmonic basis which explains and
justifies it and which requires constant control from the harmonic standpoint for both
original and derivative combinations, gives the exercises in complex counterpoint of
the free style an original character that distinguishes them sharply from the analogous
exercises in strict counterpoint. The assimilation of the new possibilities available
in shifting counterpoint of the free style should not be so much a theoretical consid
eration as a matter of practice and skill, acquired by empirical methods. As for hori
zontal-shifting counterpoint, the rules are the same for both the strict and free styles.
§ 499. Skill in the handling of counterpoint is of such advantage in composition that
the beginner should devote part of his time to its study. The object will be attained
with the least expenditure of time and labor by following the methods given in this
book; the result will be greater assurance and freedom in part-writing, skill in improving
the musical content of the individual voices, the development—of special value in
thematic work-of the ability to extract derivative combinations, the profit to be gained
by acquaintance with an infinite variety of contrapuntal forms as means of artistic
expression, and the subordination of the whole realm of tonal material to the creative
imagination. Quoting Goethe:
Wer sie nicht kennte,

Die Elemente,
Ihre Krafte
Und Eigenschaft,
Ware kein Meister
Ueber die Geister.
["Those who do not know the elements and technique of their craft can never be
master over its spirit." —Faust, Part I]
The old masters did not scorn exercises in counterpoint. If their custom of giving
each other puzzle-canons to solve evokes a smile, the time and work spent on the
solution of such problems is not entirely to be regretted. Though the secrets of this
skill were jealously guarded it must be admitted that at the basis of such feats was a
CONCLUSION 301
real understanding of the importance that they had for the development of contrapuntal
technique. Therefore the composers of former times were able to write the most com-
licated scores rapidly, easily solving technical problems that were beyond the powers
of later musicians. Amid this atmosphere of contrapuntal research, with its experimenting
with new and intricate interweaving of voices, arose the majestic figure of Johann
Sebastian Bach, who reached the furthest limits in his use of complex counterpoint as
the means of realizing a high artistic purpose.
§ 500. The object of the present work is not to analyse and classify examples from
musical literature but to develop the deductive faculty. The points of departure are: The
concept of voice-shifting and the principle that in complex counterpoint both the original
and the derivative combinations must satisfy the requirements of simple counterpoint.
The analysis of the concept of shifting governs the general plan of the book and it
includes in its scope shifting counterpoint in all of its many phases, where each form
has a definite place, irrespective of whether or not it is used. The consistent application
of these principles gives the system of rules for shifting counterpoint.
The deductive method can be applied without difficulty in a given case, owing to the
fact that these same fundamental qualities of voices—their capacity to shift, and the
resulting relationship— are completely subject to mathematical treatment. Basing the
study of shifting counterpoint on elementary algebra renders possible exhaustive state
ment together with a conciseness otherwise unobtainable. A cursory view of the program
of shifting counterpoint leads to the conclusion that musical literature uses only a few
of the resources available. Evidence of this is that many of the examples quoted (most
of them in Division B and nearly all in Division D), both in the text and in the appen
dices, refer to forms of shifting counterpoint never before actualized in music.
It is believed that the study of this book will lead to the conviction that only on the
basis of mathematics can be built a clear and rational theory of shifting counterpoint;
that it is no longer possible to revert to the redundant, vague, inaccurate and confused
explanations of former times, and that only the method of mathematics can rend the veil
of semi-mystical secrecy that for such a long time has obscured the study
of convertible counterpoint.
L
APPENDIX A
Thirty-Two Two-Voice Examples at all Indices from
/t> = l to Jv~6 and from /t> = —1 to Jv = —13 Inclusive

eoe
X=Af
xT,_
Z v8 x-x-xT—
9 9 £, 8 hjiA 33U3J3J3J O} SIl]J 3|qe}
T=Afai+T=Ai)
d
pue oj ssoi{} jo aip Suipacecns
0 I S
X— X— X— —
8 L 8 f 'Af *p "sqO § frSI
—X — X
]BUl3lJO xi
I JrÛP^ <fejl
01
fF *
5
t^ZT
^T
II
i P*
L
—X X
z
—X
■I d 3ATJBAU3Q
M ,=.I
g^ -»-
^^si( It**
"cja = JBUlSlJO
p=l=3
—<
ss
0, X
-«v w ^^
JC }X3U 3]duiBX3
T-=Af2
d
X X — —X X —X
6 * 9 £ 8 6
I) + i--Ar(U s»iuipB jo
*r "t 8 " * Z " xe WIhs jb 9=Af §) )9S
a3uBii3ja}ui jo ibui8iiO P«e veijBAuaa §) '89 :)[I] AD = j_ + )h Af = I—
IbuiSuO x~ x- ' jOf
ii m :.—lS>~»-^r:T
i £
—X
—X
—> aAI}BAIJ3Q
,»; II
,-=»!
P=^ i ? £= nr
E^SM "DJ3 - jbut3uo jo
II Sinpceaid ajdtueex
m =i -r»
'6 *
¥ -«» <Jb«
W£
S=^f2
„I1 9 1
(lN«=*l) Z=Af 0 2 9 L
8=pI + )[I]'%T§)H
I X8 8
jeurSuO 1 x*
âiM o °sL,,° 4b° " sfc==z^
3tr
i
II .7? * - ■ J = =■=3= zaz Ti ■
3E #-r-^
i
X
8 . ,1
» ■9 ' ~rr
asm *—« 3E
»: rsi:
-re-
£ "-:ra~ B J I f o|
ZEE
x±
8 0 8
9 ' ig Tr
*
J - J feN
•<* 4
"TT-
J ■rifRjJjJ
T iS
3T
xe
3AI}eAU3Q
0S 8
»
3E
** 3E
a ZEC
"cja = jeuiSyjo
II jo }exu ajduieex
m .-—-e =t=t
s - - 1J 3E
SOS
S-=Af,
a
Y9S fi 01
Z * L 6 +»-=*! H §) '961 )[I]
£ *y (Jj ^J|+[) 2=Af J0S}IUIPB
ytip }b Af = S §) :)9S
aSueipjajU! jo ibui3u0 pus sAijBAuaa §) '89 <W (H*t-J) A1 S" aDU8H I . 11+ §) '961 )[I]
Fut3uO "T
8 9
r> >" . . 3 e- * HE
1 S o -e ^ " o
J
II *£ - " J J 3r r «■ rj Jdfd
ZOZ
i
d
9
e< p*«*
mm m m ^^
mr
..n£ g s 3E
01 8 0
~n~ i' 4 .y
«-? si
1g=^ i=^
nr: Eg d -g * 1
^ ff**3
3£ feH "U
"9 ¥ 01
SATJBAUaQ
01 8 j-=a1
53=^ ^^3
t=fc=*
^^qt 31=3:
''cja = jbuiSuo jo
11 Suipacajd ajduieex
:^T 1 1 =w
-»*. - j 1 M =n
90£
SÂfS
—
d
X — X X—
e s 9 L 8
(II+«=aI) E=Af 3d * d
1*0 X9 8*£
]but3uO
.9 e
*W •EC
£f^ or
3t±at
ii oe
SI *=£
aPF *=¥
M
-r»
W£ •i
X *?
—x
2
3|P
6» •"
*±* -TT-*
^
-O-
*3t
S£ ^ ^
"o
vetjBAuaQ
e-Ai
jjûi 3T
tO"
^
:?HF^
-r&
m "0}3 = JBUlgUO
* 7«
= 11 jo iexii ajdujeex
Si 3E
L0£
8-=Af.
— -XX— — X
9 S fi 01 tl
? 9 A
T*« *« iToi 1) (11+
sjiuips jo jjhis }b *=Af §) (99
aSuBhcjajui jo ibui3uO puB aAijBAuaa §) '89 ^W (II+s-=aI) 8"=Af
]bui3uO —
»« »
I pi a * r; ?-E*
II zc
1 xi
"XT"
si ^ JJW 1,1 *=| J îjji JJ<I1> jW
a .gl ■
^g u i
8 ,
—X
«
—X
"8 11
feFF^
K
01 JS
•<* 4 "XT
m} ±M
m ftd mm
3APBAIJ3Q
8-=AI
•IL
ffi: j-jji; a -o
SP ax
tm 31
'*cja = jeuiSuo jo
3upapaid
-7&-
fi i HE
II
80c
fr=Af\
d
— X X— — X — —
9*82 8
(n*wi) t-Af 0 d S L
I
—X
8 */
—X —
8
—X
^ [' .■ -ri^JU Jjj

3
II ^B^ P **=-**
^
I
—X
J»
31
'l"-'^^jl P==E
5
t^
-O-P-
S Tt~ -O- 3CE
3E
tr 31
XT V 8
^1
?t-»— ti
8 9 * 4
s*S rJ O t±*
i «£» fe3
îS 3r •--■# e£ HE
3E 22Z
"«
„8
+=3l
pp zn~f~ ig 1 *
"• 1 **
E^jsfig; S* ps f*=*F ^ 1^
rr
8ApeAUdQ J,
iEE*£3 ~nr
^
^ '*0}3 = JBUl3U0
II jo jexu ajdujeex
^^3
/*<L ? 3 g
S *3==3 =*F»
fc5fc±
^-SAP
= -<*,-
sjiuipe jo jjish jB (9vS)ti-=Af
0 2 _ »/
3AI}6AU3Q
x. II
B-—
■ir^W* ff^.^%4^ «
■*• 7 a
,vl
u n\
^
4,11 r
60S
■>— ".x*"- — —
9 £ 8 01 ZI V V11
*9 T 8 Tl sjirape jo jjnjs JB «=Af S) (9?
a2ueipj3}UI JO JBUlSlJO piIB 3AP8AIMQ §) '89 •([!] +*-=a1) (II *"=Af
IbuiSuO Of-"
"3J3 = JBUl3uo JO
Suipacaid ajdiuxaa
aSueipjajui jo jbuûq P"B veijBAuaa §) 1B9 K[Z] (¥--aIItI) *f*•—
— _ ♦--aIIx*
310
\Jv=5
3 5 6 8
(r5+n) Jv=5 0 2 4 7 Id=5+II (§196, [1])
1 4,
Original
i-
mHE ¥=^
zaz
p^m
ii ?m ~o- 2= g ■ a~r
r r « p• • r;
j| j .1 g 1 ^P
-e r _—y e=Epq
I, J J J J J p pS
3 lx
Derivative
IT=5
*ES g II
etc., = original
of next example
gg-o o- o q ^--c 3-^-
II
\Jv=-5:
ld-6>+ILI + IId-5>(§l%,[2]) —X— X — X
2 3 4
0 P (l + Il)jv=-5>
12 3 admits of shift of Jv=-i2(§r>6)
Original — X —X— X
—X —X -X —X Derivative
114' 3 , 3 3 4
i» MMvJlUll ife
II tv,-S
lx ix Ix Ix ix
311
Mvr-5-
2. & (1 • II) Jr-,>-

admits of shift at Jv--z (%S6)
Interchange of Original and Derivative (§ 68, [1]):(IT=5 + II) Jv=-6<
Original
13 11 8 10
I £ $E!E -e- :g d J J J zfc=
S " 6» *—TT- n:
II S ~rr- @
—x
5" li 10
3E
S3 m
ff3 ^?
~rr-
i J J J J J f^
8 »,
Derivative
P o
etc., - original of
preceding example
gp J~«> 'V
II
Id=-5>*II, Mld="6 "(§196. [2])
Interchange of Original and Derivative (§ 68, [2]); (I + IP"5) Jv=-5>
Original Derivative
I, »
—X
rrrrrf^VW
—X X X —X
312
2Jv=6
—x—X — x x-£-x
2 3 4 6 7 8
(P'Îl) J„6, P 5 P
ix IxlxA S.
Original
—x
- 4 So* V
I fe»# i F=#= 3E # i V^L
IE zaz
II ^^
3
—x
ax idx
r
?Ct2Z ESI he -o- fi»
n;
S 2 '
—x
2»
—>
-eH
«,
9' J IF
Derivative
r r E2-
-r WJ' r *t £*# £e
IF etc., = original
of next example
3cc;
p
=^*: *^ ^P^
2
—x II
Jv=-6> — 2-x
(r-6+n)jv=-«> 3 5 6 (I + 11) Jv=-«>
2 4
Original 0,1 1 admits of shift of Jv=-i3>(§5f)
1
^3 i- J7?..-IJ-'-^-lor.-i-~f^
3E
II ^¥¥
i==ri 1 1 rfl_3ort
fflc
■■ '■J, -Qu
-etc.
I I o I r"
313
Mv=-6-
sVlO* U 13* lV
P _I
7 8 9 10
—X — X — X —
14
(Ml) Jt=-6<
—X
admits of shift of Jy- i ( §56)

Interchange of Original and Derivative (§ 68, [1]):(IV ~B + il) Jv=-e<
Original
—P * X
p K
10 JO 10
i l-ri*e £= £ r~# I n
£ fe
w
ll s^ 9
:rc
—x
p
X —X
10 1C
jo: -e a> pa
W 3f ?ps -p-1—m
&-
Ik P
10
8
—X
'8
—X
14.
m +W
Derivative
|V;-S
9 10
j 1——.
H-ll-^-:^ fc^l £
etc., = original of
ii preceding example
^- 12= -^ m
S ;^<e ■"
«,
Interchange of Original and Derivative (§ 68, [2]):(I+ Hv="6) Jvr-6>

Original
Bf r*r4T iL j■■ J-fjp

.1 J . n m liJ
f-îfr- U' 1
>liv.ii j=
N i O 3 „ . . T"r¥
1 I
Derivative
!|v-«
S 3E
y a „ j^ 7=£Z
etc.
^^ W
H£
£-=AP.
d
M'Mi*i) 0 Z d S £
I 8 V sjiuipe jo m^s jo *i-=Af g) (98
jeuiSiJO
,1 »
a »»
a""< 5 i J J' ' =»=* 0 ■ e- 33C
3E
II ^ 3E
d d
g'i «l
i^=£ r j J J
tzrr-J
sa •m
...
o °
*
^3 §Ê3 j 5 a ■3
fq «> £3*=
ec=S 3E
I If I *
3AIJBAU8Q
t-^Jl
V'«l P{ :**
* ''cia = ieui3uo
X
jo jexu ajduieex
¥ -«: yg » ^-»
315
Interchange of Original and Derivative (§ 68, [2]):(iv=7+II) Jv=-7

Original
t ?m
i aaas ~"~
-*»- zr n
ii (B^^tJ: 33=
3x _1_
-e^- » ax:
• W » r^
IP £
■ p- ■
ip^-h-^ ^=?
6 4*
P—o~
^£fz=m o d
I if £^g££ IE » P p F==»
1 4
Derivative
*
^nr^ etc., = original of
X preceding example
w^ «r ftlli xe:
II
-P-TT
316
2Jv=-8
—x — —x—x—x
(It=-8+II)jt=-8 3 6 6 7 8
0 12 3
—x —x —x —
6
—x
Original 7X —x —X
3 6
P-:
^m i ■ o
ii m r> •
i
—x
P=ppe r rfr . r 5
—X
Derivative
etc., = original
of next example
317
Interchange of Original and Derivative (§ 68, [2]): (1 + 11'""*) Jt=-8

Original
7'
Z_J^ fc b 3
m £
jQj
P^P P T rfr r
in:
nlSS nc
P
i.
»* 6'
1^ 1 ci o i
p^f ~n~ «:
g^PPP* fl 31= ^- ri I g * _P z=r£3

f
».
Derivative
* = -H
II
roc
i^ P etc., = original of
preceding example
Jl:
z£:
&E
P^P P
8T£
6"=Afz
° Z 60iQ9* $WM 9Wssod §>H+6-=pI )861

(e-JI*!) 6-=Af
i 8 „ % „ (terooz§)«-=Piw
IBUlSliO
zaz tf3=n a » S ' i * --" KZ

I SEB ^
TV-
II ^S *i
I K
8 9
=3 31=
J - *J g
si » 3E
^
i
It N - — ©_ | ~S-A —T.
__—.
ij| J-f-j—
# * p • rJ
- J 1
i*dj~^— J J i * >).j ■L-î-ji*-i-
7~ !—i—1—1 rrttT-nrri£8- TI — ■ — -"■
F¥=„, ■.■*'-■
rr-ir- -i
h~ 9
zee
I SE
'"Dja = IBUi3uo
jo jexu ajduixaa
dV • -J-j- W ^^1 ^^
6T€
a3uBhaia}Ui jo ibui3uq pue veiiBAuaa §) '89 (lI+6-,AI(:)[2] *"*t

jbui3uO
V K JT 9
3E
S=* »| ii
I ^ 5 *=^=*
3C
ii seh ¥
* r g JJ
$ »^»
3C
^cr
« gH
i^ #*-■— ^m
aser
mm^m.
* *=3
fi
3AI}eATJ3Q
II
$ n jj ¥* 31 m ¥
'*cja = jbdiSijo jo
3utpacajd ajdiueex
^m -A F=*
oze
0I-=APi
d d d
OI-=A'(lI+oi-=J)
d 9
0 I Z 8 V
—X — X — X —X —
L a
— X —X
IBUISUQ
.IT I ^* TT
„s
"IV
II mm IT
I
—X
4
/i( J
^ o nx 3=*
m P=£=P
— it X8
m =ri -g ' •
d
T T
.* 9 9
wP a J a i P "XT"
^f
F«=tHf
a 35= 33=
T ^
3AIJBAU3Q
rr— ±
?£ ~n-
m E -»
'*cja = |eui3uo
X jo jexu ajduieex
•xs
^g 01
arc: 33Z
-A
321
Interchange of Original and Derivative (§ 68, [2]): (I+HT=~10) Jtz.i0

Original _p_ _
6 9"
I m SEEE
~n~
«=* r j s
II 3S 3E
i,
p.
— X
6 2 6
«*
£
[J i ■! rir r J P a*
g « '•
^M ■ pa " r ir fJ
7
X
9' 2X
•£ a I o —7*-
£ TI~
&» p ti
*^ g j g
Derivative
IP*-1"
IB £ m P
IDC
etc., ■ original of
preceding example
Z3£
n-=AP«
R 9 fi OT
+„-=*!) )II "-** 0 Z » £ 6 II
1 C 9 8
{butSuO
3C
^F^Tt 31
i
fea==j]F3t
II " J d * 3r £=P==P 3E
«
1 ^
8 Is 8
-rO- ~rO- -e-
^
P zc
rJJî4
x¥
«S 01 N
CT
g>—g SP 4 awm
a
•3T
:**
a «■ J JJJJ :«:
T T
3AT}eAII3Q
II
iS
*=ac -»-
^S zxn
'*cja = jeuiSJuo
X jo }exu ajdmeex
sa^ ;zr
II-=aI i rFFrr
323
Interchange of Original and Derivative (§ 68, [2]);(I+IIV_~ ) Jr=-n

Original
8 6*3
i ps r r r^xg •"• pj -r* T-
gg=g I "
g^g S -o-
[Bia'-t- zfcr*=
^t—p £? ^^
3X- zs
■ 5 IT
IS
&
£
10 _ 10
g ■ o «r
PPi S -ii.
fe*3
^^mm i1 3
Hirzrnz IZZ
—<
Derivative
IV--H
an
P^ X
r ' r • '?
etc., = original of
preceding example
=sr
«e
ST-=Af2
d
=A'(lI+«-,J( ^T■ un 8 6 01 u [l i (Vozg)„_=pII+I

— X —X — X — —X — X X ) —(
jbutSuO xH
>* '—or
^TT
i ■ ' 8 53^ *=¥ s
ft
ii ^#s «33t
„°J
ar
fi i ^g
a 3E
*
a o
i P
■«
T&
£
fi
an
mX a P n
£ 1
*8
-r&
d
t
3AI}eAU3Q
II
a »• I3 la X
&*
''cja = ieui3uo
jo jexu ajduieex
EXE un
'11"z J " Si 21
X— =4
S3£
a3ueipjajui jo jeuiSuO pue aAijBAuaa §) '89 I):)[Z] (ti-=AlI+ ZI"=Af

|bui3uO
^
^^
zm -j- * i * «*
S^ ^gg
—*
oi n 01
■7TfntmJJ 3E
^
5
•■» '» -»- or
i -o- g, »-g
xz "01
II .1 IT
^1 1 r .«* 'J— 1— 1 L \ r<5*- «-■ P :
7-o- ■ —e p J J |
=£-eJ— rfJ ^ UJ 1—4
, rJ~ f
1 'J (9 •f/~T"rJ r f r 1
V *« *6
ZI--AI1
01
a 3E =»
It x
-9-
-o—*
m ''cja = jbui3ijo jo
3uipceajd ojdujBxo
^ 3E
*£ =tt *m m
9ce
T d
£ 9 ,9 „« flj ,CIW
Af(lKE|-,Al) ?. * 0 II joj sonpBcijdnp *P § SOZ
BI-
10 8 <i L V 01
—X — — —X —X — —
jbui3ijO
nr
I ad J I ■ i f
#^3
9 ^
±=za: Tl~
II SI o*
—
res:
TJ
ncr »• -o
h «
~rr 6 o
wm j m:
8 *9
cm
01
=P=¥= S £
°-1 x«
B-+- mtft±
3 S3 ^
a it i ^•J s 3E
01 ,1
" o -o-
'L^***^ *'[f
^
^4fc«L=3 i^tp aê ii rcr

01
i
aAiiBAijaQ
4' II
eai ^P & J ■ n
SE ^=
'*0}3 = JBUl3lJO
x jo }xau ajdujBex
a 5S ss ei-=Ai
IDC
ZZ£
a2uBip.ia}ui jo ibuiSuO pue veWeATJaa §) '89 (gi-=aII+I):^^ 8I"rAf

jbutSuO
—ef r,i
o-—zr—-*
i Si 3X
*
31
3E
3C
II 3S 3E
fe
SI QI *S
KJ cH •
31
3r
»
jlS " .<» us:

31
=^
'I
a r I rrJ 3E
^
^^^
^7 f f » J 30Z ■! ■»
i "i
fr
S
—X
8 4 9
"JT^g § r » ■i m
£3£ 3E zz: ~n~
3AT}BAU3Q
II -=A 81
m ^H
%
stt 31
'"cja = |bui3uo jo
auipaoajd ajduieex
T^rM^ m 3T
3X
APPENDIX B
TO PART ONE, DIVISION B
Twelve Examples in Three-Voice Vertical-Shifting Counterpoint,
Two to Each Diagram
Jv= 0, No. 5 (Jv<), Nos. 3, 6 (JvZ) Jv= -6>, No. 3 (/t>')

Jv= 1, No. 10(/v") /t;= _7>, No. 10(/w'—14), No. 8
/f= -1, No. 8 (/«') (/t;» = _14)
/t> = 2, No. 1 QvZ) /„ = -7<, No. 4 (/rl)
/v= -2, No. 2 (/v1 and /t>2), No. 9 (/vn) /v= -8, Nos. 7, 8 (/t>2 = -15)
/t> = 3, No. 4 (/t;-) /i>= -9, No. 9 (/t;'=-16), No. 11 (/v,
/t;= -3, No. 1 (/„') /t>")„No. 12 C/v2 = -23)
/v= 4, No. 6 (/«;'= 11), No. 2 (/vM) /« —10, No. 4 (/«')
/«. _4<, No. 7 (/v1) /v—11, No. 12 (/v1), Nos. 6, 7 (Jvn),
Jv= 5, No. 1 C/f") Nos. 9, 11 (/fV = -18)
]v= -5>, No. 5 (/t>") /u = -12. No. 12 (/«")
]v= -5<, No. 5 (/vl) ]v = -U, No. 10 (/«£)
/v= 6, No. 3 (/«") /e-^.14, No. 10 (Jv<), No 8 (/v")
331
Ml.
*.> (*)
Jv'=-3
rMr'Mn .Iv '- 5 jv'3-3: 5 % .2, 1 V i,
*> Rule
fi
c (Siss) .HTrf"
"J .. B (S 253) (§§ 267-8)
Original
i mg
,''l>A i1 -- T^
II,
i J-il i iAJ
? f5
ja:_
£
1U
^?
JJ J_l JJ ii
^WTfE^P^; 3f^p^^ fSJf=FS
III
g§£ iip^ ft* 9 ^
U: JJji ii-i
-Ui MkMk
sEîTrfp-ff.7nn^d(^^^ m m
Derivative ,V = 2
J^MJ: Jd^Ly-pJ.
^rfhf T=T5
ID
f ff ^ r=7 r
ggi >---fr# sp sSP g
LA A J=i
ffi
* PF^ ^^ffrfffrf ^ ff o- ^
mm
pîii M w- -er
m ±* 33C
£££
'SsNT
(« (.» (««»
l( I (■)
:2"=>f fw? (S-t9ï§S)
(.. n (îa5S)V
jeuiSuO (,.. " a S) «as
r fff ï'r Cf ff
^
±^ -i
iti-^ ^^yé
m «uâ ^^
^3^ 1 è ^
m ^ ^ *qJj
3AIJEAU3Q
rTTT f^f m ^p^

W:=Jij:rJ^J:^ Jliii^^J
Si 2=A
g é É* ^ IXE
=jii
sin:
<9-=Af * , ) )— Il— — *— =5^ (->
iEui8uO
i
*
ss
n ig ^É "p-* m
r„
=^3Ê
i=3
m
rrp r^r p*r
êi é àÉmà m 3£
III
3At}eAuaQ
II 9±=A
s mm m mm
X ii MÉàdm•àl m
^mê m èm é%
333
JNI4 i.)
Jv=-ir> CI)
J/J3 Jv=-10: ttAtfijp
7(h!
— x — x — xl—I — (— )
.Iv2=-7< *) Rule r(8w«) ,
•*' ,. D(S259)
Original (SSJ67-8
L
B^gP i F 4* ^ î ^
10 7 10
lli^MHJ jjjj>-i jjjj.

HI 1
ipe H¥f rm EC XC
Derivative
JNS5.
(f)
v = -S Jv = 0 (e>
><MMII Jv'=-5> (S4j2«5 7-s)
Jv£=-5<
Original
<^^pf
Derivative
~ or ^ iUf-|rH'jrj£(pg=^
<Ki
xr—
^mz fi !=^ ij=*g^|fe
P
II
334
^6.
0)
— i+ir=±"+iii Jv'= 11 (Jv=4)
Jv=- 11 (§§2S7-S)
JvZ = 0
Original
fa J J^J J J m
y^Ff
ii r?
U m
rr rf r r
^s in ^
f ^m m
-fr-^
hd J=L X^L i=^ ^ ^ XE
r RBF r^ P^ ^T f=ff
g r r p ? l=i H^ D=
Derivative
i SifeiÊ £
±j
X J.
4^ A±k
ii v=±ll
p r—r
# ■ # ■ i o-
* p
I yp^rU # P^P
ff =§=
335
JG7.
ni
>*£ r-**ii*iir-u Jv"-u
JvI=-15(jr.-8):o;|HI'xI" 5,«, 7 ) —x
(— H
JvE=-i5(.Iv=->.j
Rule D(§25S) (§§J67-s)
Original
12
LL 3L=t
&
*
n"T r rTTT FFF r
^ s
in
o
*^—i.
i rr T
32=
r- -ff rr rrr
ÊEE^ §fe=l
=8=
S ^ £ =fc
Derivative
|V -4
IIP
*
S^ -Tr
oon U
S=S II
!brr- £ fe£ ^
336
J«8.
. , (») (W (*) (•)
-^ Iv=-8 + ir±7 + nlV=-7
Jv =- 14 JvZ=-15(cf. N"7) Jj^£ !=§
JvS=-i.-,
§2"-<>)
Original
-_ii ibk
$
S
n ^
^
m f^ ^
3S £^ p ■
P
m
fc^j ±d= ££ -J-rJ-^t
F¥ r^ f rf ■=¥
Wff
s PP i PP
Derivative
ft/'7ijiJ' s TT
iV=-8
•»*
-B • ■ ffl
gjj f ■ j | £ ^
II V=±7
^P
^^
njr^rr r rr \
a .to
337
J^9. (k) <t)

Jr— ia
Jy=-2 Jt=-2(cm N2 8 Jy = -2)
£y, 'Ml ^
Jy£=-is(Jv;-9)
(SSss'-s)
Original
Jjjyjjq
h? fi f^ FS55
Derivative
II
OZ
ffc^r
&=< in w fr- rfri
r r
* n* H-
^s
spp m f£ S
.NslO Ml
tbl
-J '
Jy=-14
2<r-"-iHir" Jy'= i Jv-i: (o 1 2 3 ■hi
Jv2 = -13 M=
•) Rule d(§159)
Original (§S *«'-•)
£ hti
7^ Ff
IW* I riE £ Hi
m
Derivative
ii
iSE i I mi
s 10
ii
.111™"
PS y=j j dig ±=4 Ij 1 Jii
fSPfT f*
338
JVsll
Xiv=-9+iwif=-* Jr'=-9
Jv^-9 3^L (§§SS7-«)
Jv£=-i8 (Jv--n)
Original
I,
!4
I $^44 U Êk
&F^
s-=—m-
r rrf
^
n rrr rr ^
is m m m f
Derivative tnT=-9
r
3± ± &skk ;te
f rr fFTIf
•X'H rrr rr
?«!» ■ g
V=-9 PS i m %
J* 12. (») (b) (c)
xr-.4+ir±8+iir=-9 Jv=-u
Jy-- 18 §§»81-8)
Original
1,
'.ft* -ll -T.UrJJ.il" k& gg
;-rrF WWm =w fi
n%r— is £ HZ
#F £
HI
Derivative nlT=-9
!§&
X II"43
.&--& iE32SI mm ^
U JJ
u
^m^^grFfff fe
frF f^
n:
APPENDIX C
TO PART TWO, DIVISION C
Eight Two-Voice Examples:
Nos. 1-5, Horizontal-Shifting Counterpoint;
Nos. 6-8, Double-Shifting

341
/Tl
Ka
11 "H / T1
u
P 1
T "
(11
IT f'Tl
(]J
si
ill
(J ;
C 1
')
1,
J (
1 (I
3-
'"-
e [ ,
"5 1 1
m -a 01
O o Q D
342
he
/.]
(I ll
11 ~Î
1 "1
(1 a 1 1
<
1
ll
II
1
X *Tn
II
CO X
Z 7
1
«
ll
/NT
1 "
ll
ll
«
B
■iA
_<■■ -N
:\—|l
i.. j
i
ii i
3+ -r f
> L 1 - ! " 'nil
i c + -c ■
o o o
343
(1
«
1, * TI
'Ti
/
/■-
• I 1
If l( I
• A 1
I * ii
i it
/cn (i
«
I
« 1 IP 1
l /1 T\
i
I)
II
^
1
1\ T
ll
CO.
; n L
fi fi
T T
•1
«
«
r
( J *.. 1 HI
i 1 1
IT c*"
IT
n
***
1 'i i
1 1
—
;
i • [._ 1 1
LL. i
■
■
a '> *H8
; ■ >
> .«iL
+
i
CO
c
•a '
•c
o
<;
it
(i
ilS
t1
11;
1
ii Ii
(1
s
Q lL
II
"3c
a .1
°B
o 1
• n Ee
4)
Q
E
!
-
1"
344
L J?:
/ii
1
f
:
e ll
.
:
■ft

m
e xI •a
■a>
a
>
X
•2
•c
t B X o o
V
APPENDIX D
TO PART TWO, DIVISION D
Six Examples in Three-Voice Double-Shifting Counterpoint
One to Each Diagram

347
: : [
B
+
ii _
» i ._ i
C- ii
! - -
: «i -
I
/!- / :
I. 
/'.
1
c
•
>
ft „ ■ i
fe» ""»> "->w>
-il
i -i
I"
,
£ 8 •-3
t 1..
II i v
i
J=>
i
= cq
c ■
/*
t (
S i
ii
Hi
• / . . i_
< !s
( i
t- 1j
«>■ 1 c ":: i
5 i . ■
+
•—
~ i.
M it ,1
AS
—— i
! i
111 ■ '
"7c
.cv <
— 1
>
•a S> *i
!■ 'l -J
!
t i j
;
5i J1 ■« -
a5 ^'i.: 5 ; ••
348
"it" >
1
e+
i
I .
^J
—V '*
1
*
2>
l\ g
XI if
i£.x0 I
o s
6fr£
CO ► SP
+ ♦ IJ
c
I I
XI fi
XI
<B. a.n
s.
o a
350
:;■
3
+
V
(
+ +
i
i
s
> E3
•c
o ii2
ISE
w ?•) ?)
'?V>
'•
"4»
i; M
cn O" OS
m-
<3"
©a
X
I
».:
o
352
t I
■ ■
(4 04 • J 1 ■ I
-: • | H
i
I. :
i i.
ii n «
3 i
• (1
/"-
i
(; ii
i |I, l,
:. «j
IL
1
« L._
... i
1. (,
! II,
1
1 II IL ..
.... » 1 Ii
/-- -4- "
1 « 1 (,
in 1,. (1 ii Li I
X r* T
4 V
►* ii r- 1 1
*•
! .11 ?
i!
CO 4 i
% !
i 1 Ii. (i
+ +
,„ (1 .ii.
o1? 7©
ah J«> « . (
1
i
II
«
< i.- yll 5
IL
1
: ».. '<
ii i
i
ii 
(
: I
ii i
: i
| >
1 !1 ! 1" ! a ■
.2 << 1
•e •J
ill i1 ii - B
IIB 1 1 V
o Q .. 1 j 1 1
INDEX
(Numbers refer to paragraphs)
Addition, 341 double at the 7th, 236

Augmentation, 338 double at the 8ve, 64, 206
imitation in, 339 double at the 8ve (Jv = -7), 39
double at the 10th, 276, 474, 476
Basic version, 315 double at the 12th, 207, 236, 273
composition of, 387—393 275, 400
four-voice, 438 double at the 12th (Jv = -ll), 39, 64
functions of, 336-340 double-shifting 312-314
special cases of, 318 double-shifting, general formula for,
sub-form of, 477 373-375
three-voice, 395, 449, 451-459 florid, 234
free style, 274
Canon, double, 410, 411, 414, 416 horizontal-shifting, general formula
Cantus firmus, 326 for, 371-372
problems on, 327, 384—386 multi-voice, 262
Choral-figuration, 482, 498 multi-voice, verification of, 262—269
Combinations quadruple, 291
arrangement of contrapuntal, 419 ff. quintuple, 425
derivative, 29 simple, 361, 464
derivative, formula for, 350 three-voice horizontal-shifting,
five-voice, 425 433, 444, 474
four-voice, 417 three-voice vertical-shifting, 237—238
giving several derivatives, 395 ff. triple, 28, 283-291
imaginary, 379, 386, 415, 461, 468 two-voice, 367
imaginary, formula for, 365—370 two-voice, basis of strict style, 263
interchange of original and derivative, two-voice vertical-shifting, exercises in,
379-383 171-184
on a cantus firmus, 394 vertical-shifting, analysis of, 242
original, 343 vertical-shifting, special cases of,
original and derivative, 24 270-280
Compound Jh, 402 ff.
Consonances Derivative
duplication of, 195-202 at a compound Jh, 402
fixed, table of, 108 combinations giving more than one,
imperfect, duplication of, 216 292-2%
negative and positive, 15 extracting from a given combination, 297
perfect and imperfect, 14 faulty, 146
positive and negative, 16 in triple counterpoint, 285
variable, 133—134 Derivative combination, 23
variable, supplementary sign for, Diminution, 338
135-137 Dissonances
Contrary motion, 338 fixed, 123-124
imitation in, 339 fixed, preparation of, 131—132
Counterpoint fixed, substitution of signs, 125
basic version for double-shifting, positive and negative, 17, 18
328-333 variable, 138-140, 162
double, 28, 31, 49, 50 variable, distribution of, 155
353
variable, limitation of, 141 three-voice, 409, 413, 479
variable, similar motion to, 143 Index
Distance compound, 28
derivative, 360 contrapuntal, 341
negative and positive, 344 double, 187-190
Distance, unit of, 342 equal to zero, 43
Double counterpoint, (see also vertical horizontal-shifting counterpoint, 351—359
shift) negative, 48
zero column, 110 polymorphous, 193
Duplications, 398, 401 positive, 47
at imperfect consonances, 302 sign for, 27
at the 10th, 422 value of, 31
in five-voice counterpoint, 306—308 Indices
in four-voice counterpoint, 301 columns of, 51—60
in six-voice counterpoint, 309—311 Jv' Jv« Jv2, 243-249
in three-voice counterpoint, 300 triple, 191-192
uniting conditions of, 203-208 with direct shift, 200
with inverse shift, 200
Entries, order of, 343 Intervals
addition of, 2
Form algebraic sum of, 19
canonic, 411, 418 central, 149, 150
imitation, 423 compound, 7, 8, 9
Fourths limiting, signs for, 38, 39
augmented, 333 negative, 4, 169, 170
consonant, 210 negative in original combination, 168
consonant, chord containing a, 218 original and derivative, 24, 63
derivative, 212, 218, 225, 259, 260 positive and negative, 11, 13
free use of, 420 subtraction of, 3
in duplication, 301 tied, duplication of, 209
in two upper voices, 251 variable, preparation of, 146
negative, 258
parallel, 220, 224, 477 Key signature, 380, 386, 471
rules for, 255-257 change of, 69
Free style, 432
Fugues, 419 Measure, positive and negative, 345
double, 428, 429 Melodies, allocation of, 65
stretto, 67 Motion
tonal answer, 66 parallel, 196, 303, 304
triple, 430 similar, forbidden, 207
Harmonic series, 11 Ninths

Harmony, dissonant, 215, 217 derivative, 199
rules for, 252-254
Imitation table for duplications, 201
canonic, 212, 316, 375, 384, 394, 420 Nota cambiata, 164
exceptional form of, 339 variable consonance used as, 164
multi-voice, 424 Notation of voices and shifts, 441-434
multi-voice canonic, 409
INDEX 355
Problems, methods of working out, Suspensions, 269

376-378 complex forms of resolution, 85, 86
Proposta as given voice, 327 duplication of, 232-233
Puzzle-canon, 499 duration of, 87
insertion of consonant note, 163
Relation, reciprocal, 413 ornamental resolution of, 160
Relationship, reverse, 386 preparation of, 86
Rests, 399 symbols indicating usage of,' 88-90
counterpoint with and withoutc 481 ff . Third, imitation at, 331
omission of, 485, 486 Tie-sign, 93
Risposta, imaginary, 322—326 "Tritone", 69
Shifts
at a fraction of a measure, 334 Unison, 1, 409
at the 9th, 277-279 derivative, 105
conditions of 361—365 horizontal shift at, 437
direct, 2, 68, 386 two-voice imitation at, 317
direct and inverse, 25, 46
direct, tables of indices, 153, 154 Vertical shifts
horizontal, 378 fixed consonances, 97, 98
in four-voice counterpoint, 274 four possibilities, 96
inverse, 23, 68, 382, 386 indices of first groups, 103
inverse, tables of indices, 149 indices of second groups, 102
mixed, 23 without suspensions, 112—116
negative, 10 without suspensions, three voice, 119
positive and negative, 341 without suspensions, two rules for, 117
positive and negative, diagram for, 347 Voice
signs for, 35, 36, 41 crossing of, 248
simultaneous of two voices, 26 duplication of two using suspensions,
unintentional, 305 221, 222
vertical, 10, 21, 399 (see also double horizontal shift of a single, 348
counterpoint) imaginary, 333, 376, 452, 455-459, 491
Seventh chord, 229 imaginary, clef for, 468
Signs imitating, 319
in derivative combinations, 443 outer, shifting of, 241
supplementary x, 167 real, 455-459
Simple counterpoint receding, 39, 196
dissonances, 77—85 silencing of, 482
limitations, 71—76 simultaneous duplication of two, 213—228
Six-four chord, 229, 230, 231 supplementary, 146, 321, 378, 401, 420
Subtraction, 341 , 343 three imaginary, 459
Supplementary parts, 64
Zero, 347, 440, 467, 478
(Continued from front flap)
to contrapuntal combinations, and by restating
certain essential rules in terms of the conven
tional symbols of mathematics. This enabled
me to take into consideration a far greater mass
of relevant facts, and to bring them under the
control of a comparatively small number of
general principles."
Assuming a knowledge ot simple counter
point on the part of the student, Taneiev pro
ceeds to develop on an algebraic basis a theory
of complex counterpoint. The mathematical
system clears up a number of difficult points that
in other counterpoint texts are either slurred
over or not mentioned at all, but which occur in
composition. Far from resulting in "mathemati
cal" music, the Taneiev approach to counter
point actually releases and stimulates the imagi
nation, pointing the way to endless possibilities
that otherwise would never have been con
ceived. That this was true of Taneiev's own orig
inal work is attested by Rimsky-Korsakov, who
wrote : "It would seem that his method ought to
result in a dry and academic composition, devoid
of the shadow of an inspiration; in reality, how
ever, his Oresteia proved quite the reverse — for
all its strict premeditation, the opera was strik
ing in its wealth of beauty and expressiveness."
Taneiev deals exhaustively with twenty-three
ways of writing double counterpoint, not by
endless lists of rules and exceptions, but by alge
braic equations that eliminate all trial- and-error
methods and give positive results. He also treats
of triple counterpoint at intervals other than the
octave — something that virtually no other writ
er on counterpoint has done. Though applying,
according to the title, to the strict style of the
Polyphonic Period, the principles of Convert
ible Counterpoint — as Taneiev himself points
out — can be extended to the free style of later
times and to the modernism of today and of the
future. The study and mastery of the Taneiev
system will result in a command over the re
sources of composition that can be obtained in
no other way, for, by the application of Tane
iev's method, an inventory of what is still un
tried in counterpoint would run into astronomi
cal figures. Although an advanced work, Con
vertible Counterpoint may be successfully at
tempted as soon as the principles of two-voiced
mixed counterpoint have been grasped.
Critical Comment
"This project for making the Taneiev Counter "Taneiev is the rising star of Russian music, still
point available in English is a very good one and de unrecognized, but destined to be recognized as are
serves encouragement. My best wishes for its suc almost all of the truly great, not in his lifetime but
cess." after death."
Serge Rachmaninoff Leonid Sabaneiev
"Taneiev 's glorious treatise on counterpoint "I have long and eagerly awaited a translation of
means to musical science as much as Newton's the Convertible Counterpoint, for I know enough
Principia to cosmology. . . . Professor Brower is a of Taneiev's work—including a number of his com
finely equipped musician and his translation is a positions which are not well known in this part of
noble and astonishing piece of work." the world—and am sufficiently familiar with him
Lazare Saminsky historically and with reports of the merits of this
treatise to be sure that it should be available in an
"To make available to the English-speaking English translation. It is an event of the utmost im
world such a monument as the famous counter portance that a translation is to be published and
point treatise of Serge Taneiev is to render a service that the responsibility of justly translating this valu
of inestimable value to the art of music." able text has been entrusted to G. Ackley Brower."
Walter Piston Philip Greeley Clapp
"Taneiev was a good teacher, and his

treatise on counterpoint— one of the best books
of its kind—was highly valued by me in my
youth."
Igor Stravinsky
The Author
Serge Ivanovitch Taneiev (1856-1915), pupil, ber of works for chamber-music combinations.
friend, and critic of Tchaikovsky, succeeded the His greatest contribution to music, however,
latter as professor of instrumentation at the was in the field of the theory of composition;
Moscow Conservatory, followed Nicholas Ru and among his students at the Moscow Con
binstein as chief professor of pianoforte, and on servatory, where he expounded the system
Tchaikovsky's recommendation was made Di treated in his Convertible Counterpoint, are to
rector of the Conservatory in 1885. His musical be numbered such illustrious names as Rach
works include four symphonies ; an opera, Ores- maninoff, Gliere, Scriabin, Siloti, Spendiaroff,
teia; a cantata, John of Damascus; and a num Vassilenko, and Sabaneiev.
The Translator
G. Ackley Brower graduated from the Insti in English, but several years passed before he
tute of Musical Art, New York, in 1912, received was able to devote the necessary time to the
the post-graduate certificate in composition in work of translation, which he completed in
1914, and subsequently from the University of 1932. He has also translated another book of
Toronto thr> A* „C \t.
Bac. In 1 02? Lazare Taneiev's, The Technique of Canon. . ,
11 ii 11 mi 1 to Taneiev's great His original works include a piano sonata, sev
ecognized at once eral smaller piano pieces, a concert overture "In
ribution to music the South," a string quartet, and a symphony.
ling like it existed
9"780828"3T4152I
ISBN 0-8283-1415-2

Convertible Counterpoint

Uploaded by

Copyright:

Available Formats

Convertible Counterpoint

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Convertible Counterpoint

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What are some of the challenges Taneyev encountered in studying counterpoint?

What are some of the challenges Taneyev encountered in studying counterpoint?

THE ENGLISH TRANSLATION OF THE MOST DISTINGUISHED

TREATISE EVER WRITTEN ON MUSICAL COMPOSITION

IN THE STRICT STYLE

Translated by G. ACKLEY BROWER

With an Introduction by SERGE KOUSSEVITZKY

BOSTON • BRUCE HUMPHRIES . PUBLISHERS

J Music scholars who are acquainted with the

IN THE STRICT STYLE

BOSTON • BRUCE HUMPHRIES • PUBLISHERS

Printed in the United States of America

In studying the difficult and involved treatment of counterpoint—especially double

TWO-VOICE VERTICAL-SHIFTING COUNTERPOINT

Dissonances: (a) Passing and Auxiliary; (b) Suspensions 57

Tonal Combinations Resulting from the Duplication of Suspensions;

XII SPECIAL CASES OF THREE-VOICE VERTICAL-SHIFTING COUNTERPOINT 174

XVI GENERAL PRINCIPLES 207

Application of h and v to the Basic Version; Formulas of the Imaginary

THREE-VOICE HORIZONTAL-SHIFTING AND DOUBLE-SHIFTING COUNTERPOINT

(A) Shifting Counterpoint

(3) Vertical and horizontal shifting together, hence double-shifting counterpoint:

" 'i fiinam ^r+^

In this work double-shifting counterpoint is included in the divisions devoted to

The subdivisions of shifting counterpoint are therefore:

(1) Vertical-shifting (upward or downward);

(B) Duplicated Counterpoint

An original three-voice combination yields derivatives of four, five, or six voices,

TWO-VOICE VERTICAL-SHIFTING COUNTERPOINT

TWO-VOICE VERTICAL-SHIFTING COUNTERPOINT

The Notation of Intervals

Addition and Subtraction; Negative Intervals

M*j t-Tr 2= 7-2 = 5

§ 4. A negative quantity therefore refers to an interval of which the lowest tone be

-o" -& -©- -©-

Successive Series of Intervals;

vals, for example:

by a fifth; adjacent intervals from the series: ~ra n_::za

Order of Intervals in the Successive Series

§ 14. It is desirable to dwell at some length on certain peculiarities in the order of

First the distance is to be measured between positive dissonances only or negative

4(='int.) I lit Int) I

THE VERTICAL SHIFTING OF CONTRAPUNTALLY COMBINED VOICES;

THE SHIFTS; INDEX OF VERTICAL-SHIFTING COUNTERPOINT

Notation of the Vertical Shift: v, vv. Formulas

Obs.— The sign +, used as indicated, refers to addition as meaning a combination of

Original 1st Derivative 2nd Derivative

Symbol: Zx^ This shift is what is commonly known as "double counterpoint,",

II J.J- J.I J.J

§ 24. The combination of melodies I + II forms a series of intervals:

HiII '^JJJ i^H 3r

Intervals of the Derivative combination: 7 10 8

Intervals of the derivative combination: -2 -3 4 -7 -5 -4 -2

Index of Vertical-Shifting Counterpoint (Jv)

The Formula m + Jv = n; Inferences Therefrom

Limiting Intervals; Their Signs (<, >)

Signs for the Shifts. (<, >, <>, <)

THE GROUPING OF INDICES