POLYHISTOR

STUDIES IN THE HISTORY AND HISTORIOGRAPHY

OF ANCIENT PHILOSOPHY

PHILOSOPHIA ANTIQUA
A SERIES OF STUDIES
ON ANCIENT PHILOSOPHY
FOUNDED BY J.H. WASZINK AND W.J. VERDENIUS
EDITED BY

]. MANSFELD, D.T. RUNIA

J.C.M. VAN WINDEN

VOLUME LXXII

K.A. ALGRA, P.W. vAN DER HORST, D.T. RUNIA (EDs.)

POLYHISTOR
STUDIES IN THE HISTORY AND HISTORIOGRAPHY
OF ANCIENT PHILOSOPHY

POLYHISTOR
STUDIES IN THE HISTORY AND
HISTORIOGRAPHY OF ANCIENT PHILOSOPHY

Presented to Jaap Mansfeld

on his Sixtieth Birthday
EDITED BY

KEIMPE A. ALGRA
PIETER w. VAN DER HORST
DAVID T. RUNIA

EJ. BRILL

LEIDEN NEW YORK KOLN

1996

The paper in this book meets the guidelines for permanence and durability of the
Committee on Production Guidelines for Book Longevity of the Council on Library
Resources.

ISSN 0079-1687
ISBN 90 04 10417 8

Copyright 1996 by EJ. Brill, Leiden, The Netherlands

All rights reserved. No part qf this publication may be reproduced, translated, stored in
a retrieval .rystem, or transmitted in ai!JI form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without prior written
permission .from the publisher.
Authorization to photocopy items for internal or personal
use is granted by EJ. Brill provided that
the appropriate fies are paid direct!J to The Copyright
Clearance Center, 222 Rosewood Drive, Suite 91 0
Danvers MA 01923, USA.
Fees are subject to change.
PRINTED IN THE NETHERLANDS

CONTENTS
Introduction......................................................................................

vii

PART ONE: The Presocratic Tradition

M. ScHoFIELD, Anaxagoras' Other World Revisited................

de Chio .........................................................................................

21

J. BRUNSCHWIG, Le fragment DK 70 B 1 de Metrodore

PART Two: Plato

K. A. ALGRA, Observations on Plato's Thrasymachus: the

Case for Pleonexia........................................................................
M. VEGETTI, Kompsoi Asklepiades: la critica de Platone alla
medicina nel III libro della Repubblica ...................................

M. BALTES, ryovev (Platon, Tim. 28 B 7). 1st die Welt real

entstanden oder nicht? ..............................................................

41
61

76

PART THREE: The Academy and the Peripatos

]. M. DILLON, Speusippus on Pleasure..........................................
99
L. M. DE RIJK, On Aristode's Semantics in De interpretatione 1-4 115
I. G. Kmn, Theophrastus Fr. 184 FHS&G: some Thoughts on
his Arguments............................................................................
135
T. DoRANDI, Ricerche sulla trasmissione delle Divisioni
Aristoteliche... ... ........ ........ ....... ..... ...... ........ ..... ...... .. ......... ....... ... ....
145
PART FouR: Hellenistic and Early Roman Philosophy

J.

BARNES, The Catalogue of Chrysippus' Logical Works........

R. J. HANKINSON, Cicero's Rope...................................................
C. J. CLASSEN, Aristipp und seine Anhanger in Rom..............
P. H. ScHRIJVERS, Lucretius on the Origin and Development
of Political Life (De Rerum Natura 5.11 05-1160) .....................
H. B. GoTTSCHALK, Philosophical Innovation in Lucretius? ..

169
185
206
220
231

vi

CONTENTS
PART FIVE: Imperial Philosophy

B. INwooD, L' oikeiosis sociale chez Epic tete...............................

T. L. TIELEMAN, The Hunt for Galen's Shadow: Alexander of
Aphrodisias, De anima 94.7-100.17 Bruns Reconsidered.....
P.-L. DONINI, Doti naturali, abitudini e carattere nel De Jato di
Alessandro...................................................................................
D. N. SEDLEY, Alcinous' Epistemology.......................................
P. W. VANDER HoRST, 'A Simple Philosophy': Alexander of
Lycopolis on Christianity.........................................................

243
265
284
300
313

PART Six: Dialectic and Doxography

J.

N. M. BALTUSSEN, A 'Dialectical' Argument in De anima

A 4: on Aristotle's Use of Topoi in Systematic Contexts.......
A. A. LONG, Theophrastus' De sensibus on Plato........................
D. T. RuNIA, Additional Fragments of Arius Didymus on
Physics.........................................................................................

333
345
363

PART SEVEN: History of Scholarship

J.

GLUCKER, The Two Platos of Victorian Britain......................

W. M. CALDER III, Drei Briefe Wilhelm Diltheys an Ulrich
von Wilamowitz-Moellendorff (1908-1910)...........................

385

407

BIBLIOGRAPHY
Stepping into the Stream: a Bibliography of the Publications of
Jaap Mansfeld 1964-1995...........................................................

418

INDICES
Index of Ancient Authors..............................................................
Index of Modern Scholars ..... ...... .... .... ... ........... ............. ... .... ..... ...

431
434

Copyright 1996 by E.]. Brill, Leiden, The Netherlands

INTRODUCTION
XPll yap EU flUAa 7t0AAOOV Yotopa~ cptA.oo6cpou~ aviipa~ dvm ...

Naar wijsheid strevende mannen dienen heel vee! zaken

onderzocht en in hun mars te hebben ....
Nach Weisheit strebende Manner miissen sehr viele
Dinge erforschen ... 1

Of Immanuel Kant it has often been observed that he spent his

entire life in the territory of Konigsberg, but that this did not prevent his fame from spreading throughout the whole of Europe and
beyond. In a somewhat similar way Jaap Mansfeld has never lived
anywhere else than Utrecht and its vicinity, and has never been
associated with any other institution than its University, but this
has not stopped him from becoming a dominant force in the study
of ancient philosophy in the final quarter of the 20th century.
Jaap was born in Utrecht and gained his first knowledge of
Greek and Latin at its Stedelijk Gymnasium, where he was taught
by Peter Breemer, later a life-long friend. In 1954 he enrolled at
Utrecht University, began to study Classics and Philosophy, and
completed his Bachelors and Masters degree in rapid time. Mter a
brief period in Bern, where he studied under Willy Theiler, he
obtained his doctorate in 1964 with an innovative study on the
poem of Parmenides. Supervisor of his research was his teacher
Cornelia de Vogel. About this time lecturers began to be appointed
at Dutch Universities, and Jaap soon obtained a position in the
newly established Philosophical Institute. He did not have much to
do, for Professor De Vogel was not about to allow a young whippersnapper to take on the heavy responsibility of giving lectures in
her department. For a time Jaap was enthusiastic about the new
wave of democracy that spread through the Dutch universities in
the early 70's, and served as chairman of the elected University
council. But in 1973 De Vogel retired and he was swiftly appointed
as her successor. Since then he has held the chair of Ancient and
1 Clement of Alexandria, Stromateis 5.141, = Heraclitus fr. B33 Diels-Kranz.
Translations by Jaap Mansfeld in Heraclitus Fragmenten (Amsterdam 1979) 26,
Die Vorsokratiker (Stuttgart 1983-86) 1.249.

viii

INTRODUCTION

Patristic (later Medieval) philosophy. On various occasions he has

been Dean of the Faculty-continuously since 1990-, building it
up into a flourishing institution within the larger University
context.
Such are the bare bones of a successful Academic career. We
briefly mention the details, first because not everyone knows
them, but more importantly because they form the backdrop to
what has above all motivated the contributors to this book, his
enormous contribution to the study of ancient philosophy. All
those who know Jaap personally are well aware of his dislike of
any form of pomp and circumstance. It was with some trepidation,
therefore, that we decided we did not want his 60th birthday to go
entirely unnoticed. When we approached Jaap's many friends
and colleagues in the field, we discovered that they were of the
same mind. In many cases they were already heavily committed,
and the last thing they wanted was yet another request to contribute to a Festschrift. But when they heard that it was for Jaap, they
quickly gave in. The result is this collection of papers, which all of
us-publisher, editors and contributors-offer to him with feelings
of respect, gratitude and friendship.
When we first wrote to the contributors, we asked them to submit
articles with the following double focus: (a) on historical aspects of
the study of ancient philosophy; (b) where possible on subjects that
have been dealt with by Jaap in his many writings or that will be
of particular interest to him. The result of our request can be seen
in the twenty-four papers contained in this volume. They cover
virtually the entire length and breadth of the field of ancient philosophy. Since in most of them reference is made to contributions
made by Jaap himself, they are in themselves an impressive testimony to the staggering range of Jaap's interest and participation in
the research on the history of ancient philosophy. All will agree
that he himself has put into practice, in a manner befitting the end
of the twentieth century of course, the Heraclitean injunction
which we have exploited for the title of this Festschrift.
The papers in this collection have been organized largely in
chronological order. As things turned out, there are few papers on
the Presocratics, despite Jaap's major contributions to this area. This
is perhaps an indication of a decline of activity in the field. In
contrast the many papers on Hellenistic and Imperial philosophy

INTRODUCTION

lX

correspond not only to Jaap's own interests and research, but also to
a major concentration of effort by other scholars at the present
time. A special section is reserved for papers in the area of doxography, in which-as various scholars point out-Jaap's influence
is perhaps more dominant than in any other. At the end of the
volume we include a complete bibliography of Jaap's scholarly
writings up to the end of 1995. Here too we felt a Heraclitean title
was appropriate, for, given the large number of contributions
awaiting publication and other works in progress, we cheerfully
predict that this bibliography will one day deis volentibus appear to
be very incomplete.
For the editors it has been particularly gratifying that there
turned out to be a common thread through most of the papers submitted, i.e. the indissoluble link between the ideas and themes of
ancient philosophers and the writings in which those ideas and
themes have been presented, whether directly or through the
intermediation of the tradition. This has been the chief insight
which Jaap has developed in a myriad of ways during the past
thirty years. Its continued exploration has given the collection of
essays a modicum of unity and an additional sense of purpose.
The chief purpose of the volume, however, is to honour Jaap. We
hope that this gift will inspire and challenge him in the continuation of his scholarly activities. May he long remain faithful to the
saying of Heraclitus that forms the Leitmotiv for this collection.

On a more practical note we would like to offer our warm thanks to

the publishing firm of E. J. Brill in Leiden and especially its editors
Julian Deahl and Albert Hoffstaedt. They responded with enthusiasm to our initial suggestion and so made the publication of this
book possible. We also thank Mrs Gonni Runia-Deenick for efficiently coping with all our demands in the area of type-setting. The
most difficult task we confronted was to impose a measure of
standardization on the diversity of conventions used by the various
contributors (e.g. Diogenes Laertius abbreviated as DL, D.L. or Diog.
Laert. etc.). We are the first to admit that we have not been entirely
successful in this area and have no choice but to beg the reader's
clemency. Because of this problem we tried to limit the number of
abbreviations used throughout the book. For those that remain the

INTRODUCTION

reader is advised to consult L 'an nee philologique in the case of

Journals and the well-known Lexica of Oxford University Press
(Liddell, Scott and Jones, Glare, Lampe) in the case of ancient
authors and standard reference works.

The editors
Utrecht & Leiden, May 1996

PART ONE

THE PRESOCRATIC TRADITION

ANAXAGORAS' OTHER WORLD REVISITED

MALCOLM SCHOFIELD

Very short papers are not what his readers most immediately
associate with the name of Jaap Mansfeld. But his piece entitled
'Anaxagoras' Other World' runs to less than three full pages of text,
and the notes cover only half a page more. 1 Perhaps its brevity is
one of the reasons for its neglect. Schofield in his light revision of
Raven's chapter on Anaxagoras in The Presocratic Philosophers does
not refer to it.2 Nor do more recent articles such as Inwood's or
Furth's. 3 The neglect is unfortunate. Of the difficult text Mansfeld
takes as his topic, 'Anaxagoras' Other World' seems to me much
the most persuasive account available in the scholarly literature. In
what follows I shall advance further considerations in favour of its
interpretation of the mysterious 'other world', and against some of
the alternatives favoured in other quarters.
The text in question is as usual preserved by Simplicius, and
since its analysis by Hermann Frankel known to afficionados as
Fragment 4a. 4 Here is the text, followed by a translation. Anaxagoras' prose as often contains ambiguities. For example, in the first
sentence there is a present infinitive in oratio obliqua, which I have
taken as probably representing a present indicative, although I
once thought it reflected an imperfect. My version borrows from
Furley's and Raven's5 (in KRS):6
'tOU'tCOV & ou'tco<; x6v'tcov XPil 001ce\v EVEtVat noA.A.a 'tE Kat
nav'tota EV nncrt 'tOt<; cruyKptvOJ.lEVOt<; Kat (J1tEpJ.la'ta 7tUV'tCOV
XPTJJ.lU'tCOV Kat iOEa<; 1taV'tOta<; EXOV'ta Kat xpota<; Kat ft&ova<;. Kat
av8pc07tOU<; 'tE OUJ.17tayi\vat Kat 'tU aAAa scpa ocra 'lfUXilV EXEt. Kat
Mansfeld (1980) 1-4.
Kirk, Raven and Schofield (1983) 352-89.
3 Inwood (1986) 17-33; Furth (1991) 95-129.
4 Frankel (1955) 287 and n. 1. For information about Simplicius' citations
see Diels-Kranz (1952) 59 B4: the relevant passages are in his commentaries
on De Caelo 609.3-11 and Physics 34.28-35.9 and 156.2-4, 157.9-16.
5 Furley (1989) 56; Raven in Kirk, Raven and Schofield (1983) 369 and
379.
6 Text as in KRS nos. 483 and 498, except as indicated in n. 7.
I

MALCOLM SCHOFIELD

'tOt<; "(E av6pffinounv etvat Kat nOAE1<; OUVT)JlJlEVa<; Kat Epya

Ka'tEOKEUaOJlEVa, roonep nap' TtJltV, Kat iJeA.t6v 'tE au'tototv dvat
Kat OEATJVT)V Kat 'tCx aAAa, roonep nap' ilJ!lv, Kat 'tTJV yilv aU'tOtOt
<pUE1V noA.A.a 'tE Kat naV'tOta, rhv EKetVOt 'tCx ovfto'ta
OUVEVE"(KaJlEVOt Et<; 'tT)V OtKT)OtV XPCOV'tat. 'taU'ta JlEV OUV Jl01
AEAEK't<Xt nept TJl<; anoKptotoc;,
ouK &v nap' TtJltV JlOVov
anoKptSEi..,, &A.A<x Kat &A.A.n.
I

'

"

,...

,..

";'

on

These things being so, it is right to think that there are, in all
the aggregates, many things, of all kinds, and seeds of all
things-[seeds] having forms and colours and savours of
every kind; and that human beings were formed and the
other living creatures that have soul; and that among the
humans there are cities that have been constructed and
buildings that have, been manufactured, as is the case with
us;7 and that they have sun and moon and the rest, as is the
case with us; and that the earth grows for them many things,
of all kinds, of which they collect together the best into their
dwelling and make use of them. This, then, is my story about
the separation: that separation would have taken place not
only with us, but elsewhere too.
1. Simplicius' interpretation

Simplicius was puzzled by the passage. The first time he presents it

in the Physics commentary, in his discussion of Physics I 2, he
treats it as good evidence for his Platonist interpretation of Anaxagoras. This makes Anaxagoras explain the generation of our
perceptible kosmos as a replication of a purely intelligible kosmosthe 'other world' specified in Fr. 4a. That kosmos is generated
timelessly in its turn from an original intelligible unity, which is
how Simplicius reads Fr. 1. On his account, Fr. 1 presents the
original unity of things not as temporally prior to the present kosmos
we inhabit, but as ontologically more primitive. What Fr. 4a then
does is describe the generation of the intelligible kosmos from that
ur-condition, adding as it goes along comparison with the way
things are with us, in our derivative perceptible kosmos. 8
7 Sense and at one point textual reading in this clause are disputed. I
follow Sider (1981) 69-70, and in particular his decision to read O"UV'I]IlllEVa.~,
not O"UVCflJcrtllEVa.~.
8 Simplicius in Physica 34.18-35.21.

ANAXAGORAS' OTHER WORLD REVISITED

Everybody nowadays dismisses Simplicius' interpretation as

hopelessly ahistorical. It is one thing to see in Anaxagoras
anticipations e.g. of Plato's idea of participation in Forms, as some
recent writers have suggested;9 it is another actually to make him a
Platonist, or rather a Neoplatonist. Of more importance for our
present inquiry is that Simplicius himself evidently had doubts
about his account: not a global doubt about Anaxagoras' Neoplatonism, but a local doubt over whether Fr. 4a really was good evidence
of his conceiving of an intelligible kosmos. Already in the discussion of Physics I 2 he concedes that some might think Anaxagoras
was talking there not of an intelligible kosmos but of other parts of
the earth we inhabit. 10 By the time he reaches Aristotle's sustained
treatment of Anaxagoras in Physics I 4 Fr. 4a is no longer Simplicius' prime evidence for ascribing to him the idea of an intelligible
kosmos, as it was earlier: now he puts greater weight on Fr. 12 and
Fr. 14. He would clearly like to enlist Fr. 4a too, but spends most of
his space on it attacking first the possibility that Anaxagoras is
describing an earlier phase of our present perceptible kosmos, and
secondly once again the suggestion that other parts of the earth are
in question. At the end of the discussion he concludes aporetically
with the remark: 'Yet it would be worth inquiring further as to
whether this or some other stance on these issues is correct.' 11
One curious feature of Simplicius' treatment of Fr. 4a is the
feebleness of the alternatives that he considers to his own preferred
interpretation, and their power nonetheless to unsettle him. The
idea that Anaxagoras might be talking about an earlier phase of
the perceptible kosmos is decisively refuted by Simplicius' own
observation 12 that he uses the present indicative to describe the
behaviour of the humans described in the text (although most of
the verbs are infinitives or participles, 'make use of near the end of
the passage is not). His reasons for discounting other parts of the
earth are admittedly less compelling. He argues that if that is what
Anaxagoras had had in mind, he would have said not 'they have
sun and moon' but they have 'the sun and the moon'. 1 3 Not
necessarily, we might reply, noting that Anaxagoras anyway
9 E.g. Denyer (1983) 315-27; Furley (1989) 62-4.
10 In Physica 35.9-13.
11 In Physica 157.5-24.
12 In Physica 157.18-20.
13 In Physica 157.20-4.

MALCOLM SCHOFIELD

talks of 'the earth' in this context. In the commentary on Physics I 2

Simplicius also takes the reference to 'seeds' as showing that an
'intelligible prototype' of our world (to use Mansfeld's nice formulation) is meant. 14 But he does not repeat the argument when he
comments on I 4, and it will certainly not sway the modern
reader. Some scholars of our era, notably Cornford15 and Guthrie, 16
have actually supposed that the 'other parts of the earth' reading of
Fr. 4a is correct. This must be wrong, for a reason pointed out by
Vlastos.17 Talk of the separation is for Anaxagoras in this kind of
context talk about cosmogony. Reference to its happening 'elsewhere' cannot therefore relate to the possibility or necessity of
civilization in other places on earth, but has to concern the creation
of one or more other worlds, which are envisaged as containing
civilizations like our own. Given that Fr. 4a comes from an early
part of Anaxagoras' book (as Simplicius says it does), 1B in which
the main theses of his system were evidently enunciated and
(some of them) argued, it is hard anyway to see what the point
would be of discussing the existence of civilization elsewhere on
earth. It is quite unclear what light it could throw on the general
themes of the mixture of all things and their separation out which
Anaxagoras was developing here.
If Simplicius felt as he plainly did that even so some uncertainty
hung over Anaxagoras' meaning in Fr. 4a, 19 it seems fairly safe to
infer that that must be at least in part because he could find nothing
in the surrounding context or indeed elsewhere which helped to
clarify the issue. This conclusion is in fact the most interesting
and important thing to emerge from consideration of Simplicius'
evidence. Its significance is muted if Simplicius had access only to
a selection of Anaxagorean texts from what he refers to as 'Book I of

14 In Physica 35.13.

15 Cornford (1934) 7-8.

16 Guthrie (1965) 314-15.

17 Vlastos (1975) 354-5.

18 In Physica 34.28-9, 156.1-2.

19 Raven (in Kirk, Raven and Schofield (1983) 380) is certainly correct in

concluding that this is Simplicius' 'considered view'. Vlastos (1975) 358 tries
to downplay the importance and reduce the scope of Simplicius' doubts, referring inter alia to their absence in the De Caelo commentary (608.31-609.12).
But that commentary probably antedates the Physics commentary: so Hadot
(1990) 289-90. Vlastos is right however that Simplicius betrays no sign of
being tempted by a plurality of worlds interpretation.

ANAXAGORAS' OTHER WORLD REVISITED

On Nature' .20 Most scholars, however, think that the range of

material he can cite from the early parts of Anaxagoras' work, and
his quite numerous indications of the order within it of texts he
cites, suggest that he had access to a complete copy of Anaxagoras,
or at least 'Book I' of Anaxagoras, himself. If they are right, it is
plausible to infer the further conclusion (drawn e.g. by Fdinkel2 1
and Vlastos)22 that Fr. 4a was the only passage in which Anaxagoras spoke of another 'separation'.

2. A plurality of universes?
In his discussions of Fr. 4a Simplicius nowhere entertains the
possibility that Anaxagoras was arguing for the existence of other
physical and in principle perceptible worlds like our own, such as
are very clearly attested for the atomists Leucippus and Democritus,23 and more controversially for e.g. Anaximander.24 This is
surprising, since prima facie it is an obvious way to read the passage.
But Anaxagoras does not appear on standard ancient lists of proponents of a plurality of worlds, nor does Aristotle ever number
him in that company (although he does not explicitly attribute to
him the belief that there is only one world, as Simplicius and
Aetius do). 25 It is true that Simplicius once cites a passage of
Theophrastus which in a comparison between Anaximander's
infinite and Anaxagoras' mixture credits Anaxagoras with the
generation of 'the worlds' as a result of the activity of mind. But he
does not exploit this claim of Theophrastus; and we can probably
not determine whether it was Theophrastus' considered view of
20 As argued by Schofield (1975) 11.
21

Frankel (1955) 288 n. 1.

22 Vlastos (1975) 357.

23 E.g. Diog. Laert. IX.31

A40).

(= DK 67 A1), Hippo!. Ref 1.13.2-4 (= DK 68

2 4 E.g. Stob. Eel. 1.22.3 (= DK 12 A 17), Simp. Phys. 1121.5-9. For a case
against the correctness of this attribution see e.g. Kirk in Kirk, Raven and
Schofield (1983) 122-6. I believe the attribution to be sound: see Schofield
(1997) for a brief statement for the defence, Burnet (1930) 58-61 for a fuller
argument, and Conche (1991) ch. 5 for a thorough refutation of scepticism on
the issue.
25 For the standard lists see the texts cited in n. 24; for Anaxagoras as
proponent of a single world see Stob. Eel. 1.22.2 (= DK 59 A63), Simp. Phys.
178.23-6. The grounds on which Anaxagoras was regarded as a one world
theorist may well have been dubious: very likely the mention of 'the one
kosmos' in Fr. 8, which however is probably primarily a reference to its
internal unity.

MALCOLM SCHOFIELD

Anaxagoras, or a careless intrusion in his account of a position of

Anaximander's which was not replicated by Anaxagoras.2 6
So ancient testimony gives strangely little support to those
modern scholars such as Burnet and Barnes27 who take Fr. 4a as
asserting or implying belief in a plurality of worlds. Or perhaps
not so strangely. In atomism it makes perfect sense that the
random movements of atoms should over infinite time result in
the chance formationof an infinite variety of different universes.2 8
In Anaxagoras' theory creation of a universe requires the activity
of mind. Why should his divine mind want to create several
contemporaneous universes that are all the same? For that separation produces the same results elsewhere as it does with us is what
Fr. 4a insists. Given a divine mind at work, one would expect
either one world only, the best of all possible worlds, or different
worlds, achieving the goal of optimal variety. A mind which has
no wish or no option but to replicate the same product simultaneously over and over again forfeits its claim to be a mind.29
There is also a textual problem. If Fr.4a is to be construed as
making claims about a plurality of worlds, where does Anaxagoras
announce that he is introducing a reference to plural worlds? It
can only be either in a passage lost to us but closely preceding the
fragment, or in its first sentence, presumably in the phrase 'in all
the aggregates'. These two options probably collapse into a single
hypothesis, viz. that 'aggregates' (the word cruyK:plVOf.lEVa. occurs
only here in the fragments) must here be understood as 'worlds'; 30
and that shortly before Fr. 4a he has affirmed the production of a
2 6 See Simp. Phys. 27.15-17 (= Theophr. Phys. Op. Fr. 4, in Dox. 479.7-9).
On the other hand it might be that Theophrastus' mention of 'worlds' in the
plural is explained by his allegiance to the Mansfeld interpretation of Fr. 4a.
27 Burnet (1930) 269-70; Barnes (1979) 2.294 (n. 17).
2B See e.g. Hippol. Ref 1.13.2-4 (= DK 68A40). Barnes (1979) 2.294 (n. 17)
points out that when, as this text reports, Democritus denied that all worlds
have a sun and a moon, he very likely intended to be denying something he
took Anaxagoras to have been asserting. (The text also uses the 'in our case'
locution found in Fr.4a of Anaxagoras.) But this does not (pace Barnes) help
us much with the interpretation of Fr.4a. Democritus' remark would be
pointful on any of the readings of Anaxagoras' position examined in this
pa~er (leaving aside Simplicius' Neoplatonist version).
9 The Stoic divine providence (memorably discussed by Mansfeld (1979)
129-88) is not a counter-example to this thesis. It aims to produce the best
possible world, but the only way it can make this achievement permanent,
given the refractoriness of matter, is to replicate the world it creates in
successive incarnations by a process of conflagration (h:nuprocrtc;) and renewal.
30 For this suggestion see Lanza (1966) 202.

ANAXAGORAS' OTHER WORLD REVISITED

plurality of aggregates as the principal outcome of the process of

separation, in such a way as to make it tolerably clear that he is
talking about worlds. The rest of Fr. 4a will then have to be
interpreted as effectively governed by 'in all the aggregates', i.e. as
spelling out the developments which must be supposed to occur
within them all.
None of this is impossible. But positing a lost part of Anaxagoras'
book to do the key work in the hypothesis, with precious little
encouragement from the reactions of its ancient readers, is
precarious speculation. And as I argued before, it yields a philosophically unappetizing theory.

3. A Gedankenexperiment?
A novel attempt at interpretation, taking a quite different approach,
was offered by Frankel in the 1950s.3I His proposal found some
weighty supporters in the shape of Gregory Vlastos and David
Furley.32 In essence Frankel's Anaxagoras says not that there is a
world besides our own (as in Simplicius) or a plurality of such
worlds (as on the view examined in Section 2), but that given the
initial conditions of an original mixture operated on by the
separating activity of mind, any other world it created would be just
like ours. I.e if mind decided to intervene at some other point of the
original mixture, exactly the same process of separation would
occur there as resulted in our universe. On this reading of
Anaxagoras the point of Fr. 4a could be expressed in terms of
contemporary possible worlds metaphysics as follows: what is true
in our world is true in every possible world-which is a way of
arguing that things in our world are necessarily the way they are.
Frankel's interpretation puts a lot of emphasis on two features of
the Greek. First, it stresses that Fr. 4a begins with a thesis not about
reality, but about the correct theory of reality. Starting from the
assumption that 'these things are so', it states what we must then
suppose (8oKetv) to be the case.33 In effect this is already a claim not
that things happened in a certain way, but that given the conditions specified in the theory elaborated so far, no other outcome is

31

Frankel (1955) 284-93.

32 Vlastos (1975) 354-60; Furley (1989) 56. I accepted the Frankel interpre-

tation myself in Schofield (1980) 102-3.

33 Frankel (1955) 290; cf. Vlastos (1975) 357.

10

MALCOLM SCHOFIELD

possible. Second, Frankel argues that in the crucial assertion at the

end of the passage Anaxagoras uses the construction of optative
with av in a manner familiar in early Greek, not as a muted, polite
form of assertion, but as a true potential, designed once again to
insist on the inevitability of what is being envisaged. To say that
'the separation would have taken place not only with us, but
elsewhere too' is a bit like saying: 'I had to do it-anyone else
would have done the same'.3 4
Frankel's story about Frag. 4a has some undeniable attractions. It
is consistent with the ancient testimony that makes Anaxagoras
one of the physicists who talks of a single kosmos only. It is alert to
some conceivably crucial details of the text. And its account of
Anaxagoras' train of thought gives the fragment a good point in the
context. In Mansfeld's pithy formulation, 'things in our world are
as they have to be'.35
In his critique of Frankel's interpretation Mansfeld concentrated
on the linguistic arguments. I think his objection to what it claims
about the use of the optative with av misfires. He points out
that elsewhere in Anaxagoras (Fr. 6) the construction is twice
employed, not to conduct a Gedankenexperiment, but to express 'a
necessary inference' .36 But on Frankel's account the final sentence of
Fr. 4a where the construction occurs is not part of the Gedankenexperiment. Rather, it tells us what we may conclude from the
Gedankenexperiment. As Mansfeld's own account of the account
already brings out, a necessary inference is precisely what on that
view Anaxagoras should at this point be trying to draw from the
argument he has been developing in the rest of the passage. So
Mansfeld's counter-evidence from Fr. 6 actually turns out to give
welcome support to Frankel.
On OOKEtV, 'suppose', Mansfeld makes the telling point that on
Frankel's view this word really does two quite different jobs in
Anaxagoras' argument.37 On the one hand, taken as governing the
account of sun and moon, agriculture and urban life, it introduces
what Frankel calls the Gedankenexperiment. Here it has to mean
something like 'counterfactually suppose' what he construes as the
imaginary scenario of a duplicate civilization. But in the first
34

Frankel (1955) 288-9; cf. Vlastos (1975) 355-6.

35 Mansfeld (1980) 1.
36 Mansfeld (1980) 1.
37 Mansfeld (1980) 1.

ANAXAGORAS' OTHER WORLD REVISITED

11

clause of the fragment it must function quite differently, for as

Mansfeld says (and Frankel concedes)3B 'the fact that seeds etc. of
all things are contained in all compounds ... is a main feature of
Anaxagoras' theory, not a "Gedankenexperiment "'. This time OOKEtV
must be equivalent to 'hold' in an assertoric mood. Although
Mansfeld does not elaborate on its significance, his diagnosis of
ambiguity is quite devastating for the Frankel interpretation. For
how are readers meant to know as they progress through the passage that when they reach 'and that human beings were formed',
they are to take the implicit OOKEtV governing the clause as switching without warning to a counterfactual mood? The indicative at
the end of the supposedly counterfactual section ('make use of')
gives an opposite signal.39 Nor is av attached to any of the infinitives in the sequence. Frankel argues that archaic Greek does not
use av with the infinitive even when it carries counterfactual force.
This enables him to claim that the absence of the av marker is no
evidence against his hypothesis that Anaxagoras' infinitives are
potential. 40 But equally it can provide no evidence for the hypothesis either, which has to rest solely on an unmarked shift in the
sense of an unexpressed OOKEtV.

4. Furley 's variation

The only way out for adherents of the Frankel interpretation is to
abandon the whole idea of a shift in the sense of OOKEtV, and with it
the notion that Anaxagoras conducts a Gedankenexperiment about

an imagined duplicate world. This is the route taken by David

Furley in his acute discussion of Fr. 4a. 41 Furley takes OOKEtV to
have more of the flavour of '(reasonably) expect'. On his version of
Frankel's reading, Anaxagoras is doing the same thing throughout
the passage: viz. spelling out the thesis that, given the initial conditions, it is 'only reasonable to expect' that the world has turned
out exactly as it has and as we know it-with just the variety in the
aggregates (seeds of all sorts, humans, animals, civilization,
heavenly bodies) that we are familiar with. That is to say, Anaxagoras is not describing another world which he compares with
ours, but asserting that there had to be a world just like ours. Thus
Frankel ( 1955) 290.
was pointed out by Strang (1975) 379.
4
Frankel (1955) 290 and n. 2; cf. Vlastos (1975) 357.
41 Furley (1989) 56 and n. 37.
38

39 As

12

MALCOLM SCHOFIELD

when he reiterates 'as is the case with us', this is the 'as' not of
comparison but of exemplification or illustration.42
This highly ingenious reinterpretation of Fr. 4a, mostly tucked
away in a footnote, is of course rather startling. Everybody-Simplicius and modern readers alike-has supposed that the fragment
is talking about another world or another civilization. The disagreement has been over the identity and status of that world or
civilization. On Furley's view everybody (including Frankel) has
got the passage wrong. It is not about another world at all.
The part of Fr. 4a which is hardest to accommodate within
Furley's exegesis is the concluding sentence:
This, then, is my story about the separation: that separation would
have taken place not only with us, but elsewhere too.
The contrast 'not only with us, but elsewhere too' seems designed
to round off an account of another world that has been compared
with our own: first Anaxagoras tells us about another world just
like ours that is also produced by the separation, then like the
preacher he tells us that he has told us. Furley has to give a more
convoluted explanation of the role of the sentence. 43 He must say
that the point of the main part of the fragment is to argue that the
initial conditions must produce a world such as ours is. We can
then infer that if those conditions were to obtain at some other
place or time, the same would happen or have happened there.
Anaxagoras has not spelled out that inference in the main passage:
it is not 'the story about separation' he has actually told. But if we
assume it was in his sights all along, Furley might suggest, it is
not too hard to understand why he should have claimed that he
has. As I say, this is pretty convoluted.
There is also a whole cluster of difficulties with the suggestion
that by 'it is right to think (8oKEtv)' Anaxagoras means 'it is only
(i.e. uniquely) reasonable to expect', i.e. something amounting to a
claim that what follows is the necessary consequence of the conditions just referred to ('these things being so'). The expression could
in appropriate circumstances mean something as strong as that, but
it cannot be said to be obvious that it does here. It is also rather
puzzling that Anaxagoras should wish to make such a strong claim
4 2 My gloss on Furley: not his own formulation.
43 Again, this account of Furley's position is partly extrapolation from his
very brief discussion of the text.

ANAXAGORAS' OTHER WORLD REVISITED

13

about the outcome of the initial conditions he has evidently been

articulating. For the most part he is content in the surviving
fragments either simply to assert that things are or were so (as e.g.
in Fr. 1 on the original state of things or in Fr. 12 on mind's cosmogonic activity) or to argue that they are so, not that they could not
conceivably be otherwise (as e.g. in the proof that everything has a
portion of everything in Fr. 6). It is hard to see why he would be
motivated to be so insistent on necessity in this instance, yet
paradoxically without ever explicitly and decisively employing
any of the rich vocabulary of necessity now available in Greek
philosophy. Puzzlement mounts when we recall that among the
initial conditions being referred to at the beginning of Fr. 4a is
presumably the separating activity of mind. It would seem odd of
Anaxagoras to be making the implicit assertion that mind would
or could have produced no other world than our own. Why ever
not? It is not as though he went out of his way to portray the kosmos
as the best possible ordering of things: Plato and Aristotle notoriously complained precisely that he did no such thing. 44 Perhaps they
went too far in suggesting that there was no teleology in Anaxagoras' system at all. Frankel was right to draw attention to the
thoroughgoing anthropocentric focus of our text, with 'sun, moon,
stars and earth ... mentioned only as existing for men, i.e. for
human uses', and the reference to earth's bounty particularly striking.45 It seems likely that Anaxagoras conceives of these features of
the kosmos as providentially designed by mind. But mind could
surely have devised a kosmos which was not providentially constructed, or providentially constructed but in some other way.
We may conclude that Furley makes Fr. 4a say something
which comports rather strangely with Anaxagoras' general style
and philosophical outlook; and which, had he wanted to say it, he
would have said differently.

5. Microscopic worlds
'I believe', said Mansfeld, 'that the inquiry should start elsewhere' .46
The Anaxagorean idea he proposed as the key to Fr. 4a is what he
identified as a theory about the infinity, or more specifically the
44 Plato Phaedo 97B-99D; cf. Arist. Metaph. 985a18-21.
45 Frankel ( 1955) 285-6, 292-3.
46 Mansfeld (1980) 2.

14

MALCOLM SCHOFIELD

infinitely small, but which I should prefer to call the thesis that
complexity is no function of size. 47 It is the idea not merely that
there is no lower limit on how small a thing can be ('there is no
least', Fr. 3 and 6), but that the small is just as complex as the large
(Fr. 6). The crucial text is Fr. 6.
And since, too, there are portions equal in number belonging to
both the large and the small, in this way too all things will be in
everything. Since the least cannot be, none of them could be
separated, nor come to be on its own; but as in the beginning, so too
now all things must be together. And in all things there are many
even of the things that are separating off, equal in number in both
the larger and the smaller.

Like most of Anaxagoras' prose, this is not without its ambiguities

and densities, some of which have inevitably been ironed out in
translation. There remain irreconcilable differences over its analysis.4B Nonetheless most interpreters would agree that the main
proposition advanced here is Anaxagoras' central doctrine that in
everything there is something of everything, and that the passage
focuses on a particular way of enunciating that doctrine. For Fr. 6
is chiefly concerned with the way the world is now, when there
has been and continues to be separation from the original mixture.
And it formulates the doctrine as something true both of larger
things and smaller things in this differentiated world. So formulated it states that anything, whether larger or smaller, contains as
many ingredients (even ingredients which are being 'separated
out') as anything else.
What Mansfeld proposes is that it is this thesis about the larger
and the smaller which underlies Fr. 4a. Assume that 'the larger',
taken collectively, is the world of macroscopic objects all around us
and including us, and that 'the smaller', taken generically, is
the way things are or might be at the microscopic level. 49 Then in
See Schofield (1980) 79-94.
For a very different reading of Fr. 6 from mine see e.g. Inwood (1986)
17-33 (especially 31-32), broadly endorsed by Furth (1991) 95-129 (especially
111-19).
49 Mansfeld (1980) 2-3 consistently speaks only of the possibility of worlds
'far below the level of sense-perception'. This is perhaps because of the
empiricism he detects in Anaxagoras' allegiance to the principle that the
phenomena are our window on the unapparent (Fr. 21). But Fr. 4a is not
couched in terms of 'might'; and it seems to me that if its argument depends
on Fr. 6 Anaxagoras must be arguing for the existence of micro-worlds. I shall
assume this version of the interpretation in what follows.
47
48

ANAXAGORAS' OTHER WORLD REVISITED

15

Fr. 4a Anaxagoras is best understood as encouraging us to believe

that at that microscopic level there is or might be a duplicate of the
world-system we are familiar with at the macroscopic level-i.e.
the system consisting of earth and heavenly bodies sustaining
humans and other animals, and in particular human civilization.
Separation-his point would be-occurs not just at the macro-level.
If you could penetrate below the level of the perceptible, you would
find that separation has operated and is operating there too, in just
the same way and to just the same effect as 'with us'. Suggesting
that our entire differentiated world-system is replicated at the
micro-level is a dramatic and unexpected way of making that
point, designed to get the reader to sit up and take notice.
Mansfeld's proposal50 coheres well with various features of the
fragments. In the first place, preoccupation with the infinite, the
small, and the infinitely small, and with the need to think of
larger and smaller in the same way, surfaces one way or another
in many of them (notably in Fr. 1, 2, 3, 6, 12; 5 and 7 pursue related
concerns, albeit no longer altogether intelligibly). Indeed the very
first sentence of the work signals Anaxagoras' decision to place
such concerns at the heart of his enterprise:
All things were together, unlimited both in quantity and in
smallness-for the small was indeed unlimited.
Secondly, it is not unlikely that Anaxagoras had articulated his
specific thesis about the equal complexity of the smaller and the
larger before he reached the passage which survives as Frag. 4a.
We are told by Simplicius that Fr. 4a came 'shortly after the
beginning of Book I of On Nature', and again that it came shortly
after Fr. 2, placed by him in its turn shortly after Fr. 1. 51 But these
are relatively vague indications, and need not preclude the possibility that Anaxagoras made some other substantive points before
Fr. 4a. Presumably there was rather more about cosmogonic
separation than the brief statement about air and aiSllp contained in
Fr. 2. And the first sentence of Fr. 4a seems to presuppose that the
important doctrine of a portion of everything in everything has
already been enunciated:
50 Mansfeld (1980) 3 points out that the main idea of the proposal (viz. that
Anaxagoras is talking in Fr. 4a about micro-worlds) was anticipated in 'an
apparently forgotten paper', viz. Leon (1927) 133-41. It is also advocated by
Strang (1975) 379 n. 28.
51 Simp. Phys. 34.28-9, 156.1-2.

16

MALCOLM SCHOFIELD

These things being so, it is right to think that there are, in all the
aggregates, many things, of all kinds, and seeds of all things.
Given that the aggregates are things in a world or worlds where
separation has taken place and is still occurring, it is hard to see
how Anaxagoras could justify the claim he makes here unless
'these things being so' includes a reference to the contention that
in everything there still is something of everything, even though
the original mixture is no more. But so paradoxical is that
contention that he must have mentioned some arguments in its
favour when introducing it. The reference to seeds at the beginning of Fr. 4a makes it likely that among them was a proof of the
sort reflected in the scholium which gives us Fr. 10: 'How could
hair come to be out of not hair, and flesh out of not flesh?', and
which specifically talks about seeds containing such bodily parts.52
Our evidence, however, suggests that when Anaxagoras argued for
the doctrine of everything in everything he presented a set of
proofs, 53 so the one recorded in Fr. 6 (which announces itself as
one of a series: 'And since, too, ... in this way too') was probably one
of those accompanying the proof from seeds. The reference to
separation in Fr. 6 is also consistent with its having occurred in the
same part of Anaxagoras' book as Fr. 2 and Fr. 4a. It is perhaps
significant that Fr. 4a begins in argumentative fashion. This too is
a feature it shares with the material in the fragments and testimonia on the everything in everything doctrine, in contrast with
the explanatory narrative style used to propound the major claims
about the original mixture and about the cosmogonic activity of
mind. 54
Mansfeld's proposal that Fr. 4a is in effect an elaboration of
Anaxagoras' conception of 'smaller' things equal in complexity to
'larger' things suggests an explanation of some opaque locutions in
Fr. 12. Anaxagoras speaks there of mind's control of 'all the things
that have soul, both the larger and the smaller', and again of how
all mind is the same, 'both the larger and the smaller'. It sounds as
though he has something specific in view when he employs these
5 2 Fr. 10 is contained in a scholium on Gregory of Nazianzus (see PG
XXXVI 911; cf. 521). I discuss it in Schofield (1975) 14-24 and Schofield (1980)
133-43. My use of Fr. 10 here is, I think, in line with the valuable
observations in Mansfeld (1982) 361-3.
53 Cf. e.g. Arist. Phys. 187a26-b7.
54 Anaxagoras' expository and argumentative styles are one of the major
themes of Schofield (1980).

ANAXAGORAS' OTHER WORLD REVISITED

17

expressions. If Mansfeld is right, and in Fr. 4a Anaxagoras has

already given an account of the microscopic world contained in
the 'smaller' things, we can identify what specific reference he is
intending. The larger things that have soul will be humans and
other animals in our macroscopic world, the smaller things that
have soul the corresponding humans and other animals in the
microscopic world of Fr. 4a; and so mutatis mutandis for the larger
and the smaller mind.
One difficulty with the other interpretations of Fr. 4a we have
examined was that none offered an account of it which cohered at
all well with the other evidence we have of the content, structure
and intellectual style of Anaxagoras's book. Mansfeld's interpretation integrates the fragment and its 'other world' with many of the
other fragments, and gives it a thoroughly intelligible point at just
the juncture in Anaxagoras' book that it occupies. In the context of
his story about cosmogonic separation, Mansfeld's Anaxagoras
develops the line of thought of Fr. 4a in order to give explicit and
striking expression to the idea that 'the smaller' is just as complex
as the 'larger': separation is not, as one might say, skin-deep. 55
So the microscopic world's reading of Fr. 4a scores heavily on its
strategic advantages. And it does not contradict the impression
most ancient readers seem to have gained that Anaxagoras did not
believe in the kind of plurality of worlds to which Anaximander
and the atomists were committed: 'micro-worlds' within a single
macrocosm don't make him a recruit to their party. Nor is there
the problem faced by other modern interpretations that mind is so
constrained as to be able or willing to undertake only one kind of
cosmic programme. Mansfeld's Anaxagoras takes no stance on
that issue. All he is insisting on in Fr. 4a is that however the
cosmogonic separation initiated by mind works out at the macrolevel, we have to suppose it works out in just the same way at the
micro-level.

55 Mansfeld's account of Fr. 4a thus presents a strong challenge to the contention of Inwood (1986) 17-33 (especially 22) that the small is by definition
mixed and as such not 'separated out', or of Furth (1991) 95-129 (especially
111-112) that (i) the latent is 'small' (ii) that is latent of which there are
'few' or 'less' (iii) latent things are 'together' and 'commingled', not 'separated out' or 'distinguished'. On the other hand Mansfeld's interpretation is
consistent with the suggestion in Sorabji ( 1988) 61-66 that for Anaxagoras,
'the small' is like an infinitesimal powder.

18

MALCOLM SCHOFIELD

The interpretation needs to be able to show that it makes

satisfactory sense of the detail of the passage as well as its strategic
purpose. Here the only hard question it faces is where or whether
the text itself indicates that it is talking about a micro-world. The
key phrases are once again those of the opening clause:
These things being so, it is right to think that there are, in all the
aggregates, many things, of all kinds.
The most obvious way of taking 'aggregates' is as perceptible
compounds in the separated world with which we are familiar.
Then Anaxagoras is to be construed as claiming that within any
such aggregate or compound is a micro-world consisting inter alia
of all the things mentioned in the rest of the fragment-humans,
animals, heavenly bodies, urban civilization, agriculture. The
space within a perceptible object is 'the smaller'; and by relying on
the principle of equal complexity in the larger and the smaller he
will have believed himself entitled to argue that that internal space
is as heavily populated as the external space in the world about us.
The crucial moves in this reading of Fr. 4a are (a) the decision
to take 'in all the aggregates' as governing the whole sequence of
clauses that follow, and (b) the inference that the inventory of
things contained within an aggregate is designed to illustrate the
complexity of 'the smaller'. The decision is not something the text
forces upon us, but equally it doesn't go against the grain of the
prose, and justifies itself by its results-it enables us to see how the
passage introduces the other world which the concluding sentence
clearly takes it to have been talking about. The inference is not
explicitly legitimated by anything said in Fr. 4a. It depends on the
supposition that Anaxagoras has been talking about the complexity
of 'the smaller' shortly before, and on the reader's making a
connection between that idea and the introduction of a world
within an aggregate. But I hope I have shown reason to think both
the supposition and the connection are attractive.
Neither (a) nor (b) is entirely obvious, however. This is just as
well, since otherwise there would be the difficulty of having to
explain why Simplicius and subsequent commentators failed to hit
upon either (a) or (b). Until Mansfeld, 56 of course.

56 And Leon (1927) 133-41 and Strang (1975) 379 n. 28.

ANAXAGORAS' OTHER WORLD REVISITED

19

Vlastos wrote:57
Had Anaxagoras travelled further along the road of infinity, he
might have glimpsed the enchanting prospect of a world in
every seed, and so worlds within worlds ad infinitum, matching or
bettering Leibniz's doctrine that 'each portion of matter may be
conceived as like a garden full of plants and like a pond full of
fishes; but each branch of every plant, each member of every
animal, each drop of its liquid parts is also some such garden or
pond' (Monadology 67). Anaxagoras' imagination took him far, but
not so far.

If Mansfeld is right-and I have tried to show why and that he isin Fr. 4a we see Anaxagoras' imagination taking him exactly
where Vlastos thinks it did not reach.58

StJohn's College, Cambridge

57 Vlastos (1975) 359.

58 This paper is a development of material presented in a lecture in the

University of Fribourg in October 1995 to celebrate the publication of the

French translation of Kirk, Raven and Schofield ( 1983). I am grateful to
Dominic O'Meara and Helene-Alix de Week for their kindnesses on that
occasion. It is a great pleasure to be able to offer these thoughts to Jaap
Mansfeld, from whom like so many colleagues I have learned an enormous
amount not only about ancient philosophy but about how to work on it.

20

MALCOLM SCHOFIELD

BIBLIOGRAPHY
Barnes, J. (1979) The Presocratic Philosophers, Vol. 2 (London).
Burnet, J. ( 1930) Early Greek Philosophy, 4th edn. (London).
Conche, M. ( 1991) Anaximandre: Fragments et Temoinages (Paris).
Cornford, F.M. (1934) "Innumerable Worlds in Presocratic Cosmogony",
Classical Quarterly 28:1-14.
Denyer, N.C. (1983) "Plato's Theory of Stuffs", Philosophy 58: 315-27.
Diels, H. and Kranz, W. (1951-2) Die Fragmente der Vorsokratiker, 6th edn., 2
vols. (Berlin).
Fninkel, H. (1955) Wege und Formen Jrilhgriechischen Denkens (Munich).
Furley, DJ. (1989) Cosmic Problems (Cambridge).
Furth, M. (1991) "A 'Philosophical Hero'? Anaxagoras and the Eleatics",
Oxford Studies in Ancient Philosophy 9: 95-129.
Guthrie, W.K.C. (1965) A History of Greek Philosophy, Vol. 2 (Cambridge).
Hadot, I. (1990) "The Life and Work of Simplicius in Greek and Arabic
Sources", in Aristotle Transformed, ed. R. Sorabji (London) 275-303.
Inwood, B. (1986) "Anaxagoras and Infinite Divisibility", Illinois Classical
Studies 11: 17-33.
Kirk, G.S., Raven, J.E. and Schofield, M. (1983) The Presocratic Philosophers,
2nd edn. (Cambridge).
Lanza, D. (1966) Anassagora (Florence).
Leon, P. (1927) "The Homoeomeries of Anaxagoras", Classical Quarterly 21:
133-41.
Mansfeld, J. (1979) "Providence and the Destruction of the Universe in early
Stoic Thought", in Studies in Hellenistic Religions, ed. M.J. Vermaseren
(Leiden). Reprinted in his Studies in Later Greek Philosophy and Gnosticism
(London 1989).
- - (1980) "Anaxagoras' Other World", Phronesis 25: 1-4.
- - ( 1982) Review of Schofield ( 1980), Mnemosyne 35: 360-366.
Schofield, M. (1975) "Doxographica Anaxagorea", Hermes 103: 1-24.
- - (1980) An Essay on Anaxagoras (Cambridge).
- - (1997) "The lonians", in The Routledge History of Philosophy, Vol. 1, ed.
C.C.W. Taylor (London).
Sider, D. (1981) The Fragments of Anaxagoras (Meisenheim am Glan).
Sorabji, R. ( 1988) Matter, Space and Motion (London).
Strang, C. (1975) "The Physical Theory of Anaxagoras", in Studies in Presocratic Philosophy, Vol. 2, ed. R.E. Allen and DJ. Furley (London) 361-80.
Vlastos, G. (1975) "One World or Many in Anaxagoras?", in Studies in Presocratic Philosophy, Vol. 2, ed. R.E. Allen and DJ. Furley (London) 354-60.

LE FRAGMENT DK 70 B 1 DE METRODORE DE CHI0 1

jACQUES BRUNSCHWIG

La piece Ia plus celebre du dossier de Metrodore, le representant le

plus important du "scepticisme atomistique" post-democriteen, est
certainement laformule qu'on pouvait lire, selon Ciceron2 et Eusebe
de Cesaree,3 au commencement de son traite De la Nature. Cette formule, selon Eusebe, "donna de f:kheuses impulsions a Pyrrhon,
lequel vint par Ia suite"; le scepticisme paraissait s'y radicaliser en
se prenant dans son propre champ. Son aspect provocant et sa
situation en tete de livre lui ont valu une grande celebrite, 4 a en
1 Jaap Mansfeld est I'un des editeurs de Ia Cambridge History of Hellenistic
Philosophy, actuellement en preparation a Ia Cambridge University Press. Les
collaborateurs de ce volume se sont reunis a Utrecht en aout 1991, ce qui m'a
donne I' occasion de visiter cette ville pour Ia premiere fois. J'etais charge de
rediger un chapitre sur les debuts du scepticisme hellenistique. A cette
occasion, j'avais presente une etude assez longue sur I' ensemble du dossier de
Metrodore, etude dans laquelle j'examinais les questions qu'il pose traditionnellement: Ia theorie de Ia connaissance de Metrodore, apparemment
partagee entre "scepticisme" et "rationalisme", Ia compatibilite entre son
"scepticisme" et son adhesion a l'atomisme de Democrite, le sens de son
interet pour les problemes meteorologiques. De tout cela, il ne restera guere
que deux ou trois pages dans Ia Cambridge History. Pour exprimer a Jaap
Mansfeld mon admiration et mon amitie, j'ai repris, bien plus a fond que je
ne I'avais fait en 1991, l'examen d'un point limite du dossier de Metrodore,
le fragment DK 70 B 1. On raille souvent les auteurs qui, pour contribuer a un
volume de Melanges, envoient ce que l'on appelle un "fond de tiroir".
J'espere que Jaap Mansfeld voudra bien s'apercevoir que l'etude qu'on va lire
est, pour le moins, un fond de tiroir longuement revisite et entierement
retravaille.-J'exprime ici mes remerciements a Tiziano Dorandi et a Carlos
Levy, qui m'ont aide a completer rna documentation et a clarifier mes idees
sur les temoignages de Philodeme et de Ciceron, ainsi qu'a Jean-Pierre
Lefebvre, qui m'a reveJe le sens d'une abreviation dans le commentaire de
Langerbeck (1935) 122, eta Richard Goulet pour son secours informatique.
2 Acad. pr. II 73 = DK 70 B 1.
3 Praep. ev. XN 19 9 = DK 70 B 1. La source de ce passage n'est pas
Aristocles de Messine, a Ia difference des larges extraits du IIE:pl qni\ocro<jl(acde cet auteur qui figurent dans le contexte. Cette source est de bonne qualite,
comme le montre Mansfeld (1992) 33 n. 24 en comparant Ia version donnee
ici du debut du traite Des dieux de Protagoras (XN 19 10) et celle, moins
complete, qu'Eusebe donne ailleurs de ce meme debut (XN 3 7).
4 Le rapprochement qu'Eusebe effectue entre Ia formule de Metrodore et le
debut, non moins celebre et provocant, du traite Des dieux de Protagoras repose
aussi sur le sensualisme qu'il prete aux deux philosophes (XN 9 8), et peut-

22

JACQUES BRUNSCHWIG

juger par les nombreuses references et allusions qui lui sont faites
dans Ia doxographie.
La ran~on de cette celebrite est que Ia parole de Metrodore nous
est transmise sous des formes tres diverses, plus diverses qu'on ne
le dit parfois.s La tache que je me fixe, et qui ne semble pas avoir ete
faite jusqu'ici de fa~on systematique,6 est d'etudier les temoignages
qui Ia concernent, de les comparer, et d'essayer de voir si l'on peut
identifier Ia version qui a les meilleures chances d'etre authentique.Je presente d'abord ces temoignages, dans un ordre qui n'est
pas chronologique (ni a l'endroit ni meme a l'envers), mais qui est
plutot celui de leur longueur et de leur complexite croissantes.7
A= !J.T)OEva !J.T)aEv en(cnacrSm (Epiphane, Adv. haer. III 2,9 = DK 70
A 23, DDG 590,35).
B = !-LD [doe]vm !J.T)l>' m'rro Tou[To (Philodeme, Rhet. fr. inc., Pap.
Here. 224, II 169 Sudhaus = DK 70 A 25).
c = eA.eye !J.T)O' m'rro TOUT' ELOEV<ll, <5n oul>Ev oToe (Diogene Laerce,
IX 58= DK69A2, 72 A 1).8

etre sur Ia tradition selon laquelle Metrodore et Protagoras auraient ete tous
deux les disciples de Democrite (Clem. Alex., Strom. I 64 = DK 70 A 1).
5 Pour presenter Ia parole de Metrodore, beaucoup de commentateurs se
contentent de citer l'un des temoignages disponibles. D'autres traitent assez
cavalierement Ia question de savoir comment les harmoniser. Ainsi ZellerMondolfo (1969) 314 n. 69 cite Eusebe et ajoute que !'on trouve "Ia meme
chose" chez Sextus, Diogene Laerce, Epiphane et Ciceron; cf. aussi, entre
autres, Ernout et Robin (1926) 2.227. Le plus influent de cet ecrasement des
differences a certainement ete le collage opere par DK 70 B 1 entre le texte
d'Eusebe et une retroversion partielle du texte de Ciceron. J'y reviendrai
naturellement. Je signale en passant que Shorey (1919) 393-394 porte, malgre
son titre aujourd'hui trompeur, sur ce qui est maintenant DK 70 B 2.
6 Ce qui s'en approche le plus est Langerbeck (1935) 121-123.
7 Je ne compte pas ici parmi les temoignages sur Ia parole de Metrodore
Ia celebre refutation du scepticisme chez Lucrece (IV 469-477), bien qu'un
nombre considerable de commentateurs aient considere que Ia version du
scepticisme visee par Lucrece (et deja par Epicure) etait celle de Metrodore
(voir entre autres Ernout et Robin (1926) 2.226-227, Bailey (1947) 3.1238,
DeLacy (1971) 605 n. 30, Burnyeat (1978) 204, Isnardi-Parente (1984) 106-121,
Gigante ( 1990) 79). Je me rangerai plutot parmi les rares qui en doutent
(Vander Waerdt ( 1989) 242 n. 48). En effet, Lucrece objecte a son adversaire
sceptique qu'il "ne peut pas savoir qu'on ne peut rien savoir, puisqu'il professe
ne rien savoir"; puis il lui concede tactiquement ( 473) qu'il possede un tel
savoir au second degre. Cet adversaire, contrairement a Metrodore, pretend
done savoir qu'on ne peut rien savoir. Metrodore parait avoir voulu precisement desarmer, par anticipation, ce genre d'objection (cf. Straume-Zimmermann et aL (1990) 414).
8 Curieusement, C n'est pas repris dans DK 70, le chapitre consacre a
Metrodore.

LE FRAGMENT DK

70

B 1 DE METRODORE DE CHIO

23

on

D = oMv '(cr(.L'V, oM' auTo TOUTO '(cr(.L'V

oMv '(cr(.L'V (Sextus
Empiricus, Adv. Math. VII 88 = DK 70 A 25).
E = oMd~; TJ(.LW'V oMv oll>ev, oM' auTo TOUTO, n6Tepov o'(l>a(.LE'V j)
oux o'(l>a(.L'V (Eusebe, Praep.Ev. XIV 19 9 = DK 70 B 1).
F = (i) nego scire nos sciamusne aliquid an nihil sciamus, (ii) ne id
ipsum quidem nescire aut scire <scire> nos, (iii) nee omnino sitne aliquid
an nihil sit (Ciceron, A cad. pr. II 73 = DK 70 B 1) .9
Reprenons maintenant l'examen de ces temoignages. 10
A

La version d'Epiphane, saint homme generalement renomme

pour la faible valeur de ses informations, est la plus courte de celles
qui subsistent de l'enonce de Metrodore. La doxographie contenue
dans son ouvrage Contre les hhitiques contient des notices sur les
opinions physiques, cosmologiques, ethiques, epistemologiques des
philosophes; mais celle concernant Metrodore ne retient que ses
conceptions epistemologiques. En voici l'integralite: "Metrodore de
Chio a dit (i) que personne ne sait rien, mais (ii) que nous ne savons pas
avec exactitude les chases que nous crayons connaitre, et (iii) qu 'il ne faut
pas attacker de valeur aux sensations; (iv) car toutes chases relevent de la
croyance".ll L'originalite de ce temoignage tient ace que la portee
de l'enonce d'apparence brutalement sceptique Ai est immediatement precisee par le complement que lui apporte Aii: si pour
Metrodore il existe des choses (au pluriel: TaUTa) que nous croyons
connaitre (yLvwcrxL'V), et que pourtant nous ne savons pas avec
exactitude ou avec certitude (axpL[3iJ<; oux EnLO"T<l(.L8a), la position
qui lui est attribuee est probablement que ce qui nous est refuse est
un savoir au sens fort (En(crTacr8at), equivalent a la certitude que
notre opinion sur telle ou telle question particuliere est justifiee de
9 La troisieme occurrence de scire est donnee de seconde main par les mss
AB (voir l'apparat critique de Plasberg (1922) 62). La plupart des editeurs (sauf
Plasberg lui-meme) introduisent ce mot dans leur texte, avec des crochets
obliques (ainsi Straume-Zimmermann et al. (1990) 188) ou meme sans (ainsi
DK 70 B 1, Reid (1885) 262, Rackham (1933) 560).
10 J'espere que l'ordre dans lequel je les etudie n'a pas induit d'effets
artificiels dans !'interpretation que j'en donne. Je puis en tout cas assurer le
lecteur que j'ai aborde cette etude sans avoir Ia moindre idee du resultat
auquel elle aboutirait.
11 M. 0 Xio<; l:<jlT] !J.T]l>Eva !J.TJI>EV Erra-raa9at, UAAU Taura,
lloXOU!J.E:V
YlVWCJXE:lV, O.xplj3Gi<; oux EmaTU!J.E:Sa oM Tat<; ataef)ae:m Be:i rrpoaEXElV" 6ox~al
yO.p ECJTl Ta rrO.vTa .

24

JACQUES BRUNSCHWIG

appropnee par les faits. II se peut que cette optmon soit

"objectivement" justifiee; 12 mais nous ne pouvons en etre certains.
Vraisemblablement, Ia raison pour laquelle nous ne pouvons en
etre certains est que, de cela meme, nous ne pouvons avoir qu'une
croyance ou une opinion; mais cette subsomption d'une opinion
sous une opinion portant sur cette opinion meme n'est presente, au
mieux, qu'en filigrane dans l'ensemble de A.
Si cette interpretation est correcte, Aiv, qui explique ce qui precede (yap) par Ia domination universelle de Ia o6XTJO"l~, se relierait
fort bien a Ai et Aii. 13 En revanche, l'attaque contre les sensations
(Aiii) fait figure d'element perturbateur. En effet, Ai et Aii ne se
rapportent pas particulierement a telle ou telle source pretendue de
connaissance, et I' on ne voit pas bien en quoi I' explication fournie
par Aiv serait specifiquement pertinente a l'egard des sensations. 14
Comme l'hostilite de Metrodore envers les sensations est attestee
dans d'autres temoignages doxographiques, mais sans etre rapportee a sa proclamation sceptique, 15 on peut faire !'hypothese d'un
collage maladroit opere par Epiphane, ou par sa source, entre
plusieurs renseignements d'origine diverse.
Dans leur apparat critique, Diels et Kranz invitent le lecteur a
comparer A avec le vers 4 du celebre fragment 21 B 34 de Xenophane. lis y sont evidemment conduits par l'echo assez direct que
l'on peut entendre entre Ia formule du Colophonien, o6xo~ o' rrl
fa~on

1 2 Cette possibilite caractense ce que certains philosophes anglo-saxons

actuels appellent une epistemologie "externaliste": cf. Hankinson (1995) 19 et
118-119, qui se rHere a Bonjour (1980) 53-73; le terme et le concept ont ete
introduits par Armstrong (1973). Une telle epistemologie admet que les
conditions qui font d'une croyance un savoir peuvent etre satisfaites en fait,
independamment de Ia conscience qu'en a le sujet connaissant, et done sans
qu'il sache ou meme qu'il puisse savoir si elles le sont. On peut done, selon
cette epistemologie, savoir des choses sans savoir qu'on les sait.
13 L'echo l>oxoil~t:v-lloxf}crt:L rattache lui aussi Aiv a Aii.
14 Langerbeck (1935) 122, sans doute le commentateur le plus indulgent a
I'egard d'Epiphane, rapproche Aiv de Aii, sans preciser ce qu'il entend faire
de Aiii: "Der Ton liegt ganz auf dem oux O.xpL~W(; unseres Wissens, bzw. dem
l>oxf}crt:L t:ivm der Dinge".
15 Aetius IV 9 1 (cite incompletement par DK 70 A 22, compli:tement par
DDG 396, 12-16): "les sensations sont fausses, selon Pythagore, Empedocle,
Xenophane, Parmenide, Zenon, Melissos, Anaxagore, Democrite, Metrodare, Protagoras, Platon". Cette notice, pour ce qui concerne Metrodore, prend
sans doute appui sur son adhesion a l'atomisme democriteen; elle est en
contradiction au mains apparente avec Eusebe, Praep. Ev. XIV 2 4 (absent de
DK) et 19 8 (DK 70 B 1), d'apres lequel, selon Metrodore et Protagoras, "il ne
faut se fier qu'aux sensations du corps".

LE FRAGMENT DK

70

B 1 DE METRODORE DE CHIO

25

nacrt TETUXTat, et celle qui est ici pretee a Metrodore, ooxf}cret yap
ecrn Ta TTU'VT<l. 16 Mais on peut certainement aller plus loin que ce
rapprochement ponctuel: 1' Ia position resumee ci-dessus peut en
effet constituer une interpretation tout a fait decente du fragment de
Xenophane. 18 Ce ne serait pas le seul exemple d'un rapprochement
ancien entre Metrodore et Xenophane: Mansfeldl9 a brillamment
montre que plusieurs anomalies de Ia doxographie d'Hippolyte20
(Xenophane y vient apres Democrite, et son explication de Ia salinite de Ia mer est mise en contraste, en plein milieu du chapitre qui
lui est consacre, avec celle de Metrodore) viennent de ce que, dans
l'esprit d'Hippolyte, les deux penseurs devaient etre associes pour
une raison plus importante que leur divergence d'avis a propos de
Ia salinite de Ia mer: a savoir, leur commune inclination au scepticisme. Mansfeld ne cite pas A dans ce contexte; mais on peut assez
aisement voir en A le resultat d'une sorte de contamination doxographique, explicitant le sens du scepticisme affiche en Ai par des
considerations (Aii et Aiv) d'inspiration plus ou moins directement
xenophanienne.
B

La version de l'enonce de Metrodore conservee dans un fragment

de la Rhetorique de Philodeme21 est a peine plus longue que A. Dans
le cours d'une enumeration qui parait avoir pour fonction de denoncer quelques celebres paradoxes philosophiques, que l'Epicurien
exprime de fa~on tres condensee, et qu'il tient manifestement pour
insoutenables,2 2 Philodeme glisse celui-ci, rapporte a Metrodore de
16 Si cette hypothese d'un echo est correcte, A temoignerait que i:nl nii<n,
chez Xenophane, a ete interprete par certains, des I'Antiquite, comme un
neutre (alors que Ia plupart des modernes y voient un masculin).
l7 II y en a d'ailleurs peut-etre un autre: cf. O.xptj3w<; dans Aii et cru<f>E<; chez
Xenophane.
18 Voir par exemple Lesher (1978) 1-21, Hussey (1990) 18-22, Lesher (1992)
155-169 et 182-186. Hankinson (1995) 53 considere que l'une des interpretations possibles de l'enonce sceptique de Metrodore (qu'il reproduit cependant,
ii cet endroit, d'apres Ia version de Sextus Empiricus) lui attribue "une position coherente avec celle de Xenophane, peut-etre meme une version de cette
derniere".
19 Mansfeld (1992) 32-36.
2 Ref I 14 = DK 21 A 33, DDG 565, 20- 566, 10.
21 Pap. Here. 224, II 169 Sudhaus = DK 70 A 25. Le contexte est cite par
Cronert (1906) 192 et Gigante (1990) 79.
22 'Tout est dans tout" (Anaxagore), ~e tout est un" (Parmenide, Melissos),
"parce que les sensations sont fausses" (le texte s'interrompt ici). Bizarre-

26

JACQUES BRUNSCHWIG

Chio:2 3 "personne ne donnerait son accord a la formule <<ne pas meme

savoir cela meme",24
Dans cette formulation, interessante bien qu'elliptique, ou
mieux, parce qu'elliptique, s'introduit une expression caracteristique de toutes les versions qu'il nous reste a examiner: f.Llll>' mho
TOUTO. Cette expression a pour fonction, selon toute apparence, de
faire d'un non-savoir ou d'une ignorance de premier degre ("ne
pas savoir 1 ", disons une "ignorance 1 ") l'objet d'un non-savoir ou
d'une ignorance de second degre, d'une meta-ignorance ("ne pas
meme savoir 2 ne pas savoir 1", disons une "ignorance 2"). Mais ce
qui est tout afait remarquable, en comparaison surtout de quelquesunes des versions que nous examinerons par la suite, c'est que
l'ignorance 1 ne fait pas l'objet d'une assertion explicite independante. Tout se passe comme si l'ignorance 2 etait prise tellement au
serieux qu'il n'est meme pas possible de designer son objet comme
etant precisement une ignorance 1 C'est d'ailleurs logique: si nous
parlons d'un objet que nous ne connaissons pas, il n'y a pas de
bonne raison de le designer comme ceci (par exemple comme une
ignorance), plutot que comme cela.
J'introduirai ici une reflexion que je dois a l'etude de Myles
Burnyeat sur la refutation du scepticisme chez Lucrece.25 Supposons qu'un sceptique (peu importe ici qu'on l'identifie ou non a
Metrodore) admette les deux propositions suivantes:
(1) "II n'y a aucune proposition qui soit connue comme vraie".
Done (2) "La proposition (1) n'est pas une proposition qui soit
connue comme vraie".
ment, Gigante (1981) 64 subordonne a Ia proposition de Metrodore toutes les
autres propositions de ce passage, dont i1 donne Ia paraphrase suivante: "in
Filodemo possiamo solo ricordare che nella Retorica viene rispinta Ia nota
proposizione di Metrodoro di Chio che nulla sappiamo ne se tutto e in tutto
(con Anassagora) ne se uno e il tutto (con Parmenide e Melisso) perche le
sensazioni sono false". Cette construction parait exclue par Ia syntaxe du
passage et par l'ordre dans lequel sont presentes les divers enonces paradoxaux (Anaxagore, Metrodore, Parmenide et Metissos).
2 3 Le papyrus contient ici une lacune: >cuTa TO[v) XeTov Mf)T[ .... La place
disponible etant petite, Wilamowitz (1899) 636 a suggere de lire MTJT[pCiv],
sobriquet dont etait affuble Metrodore dans une comedie d'Antiphane (DK 70
A Ia). Cronert (1906) 192 propose de lire Mf)T[pwvu], autre sobriquet ironique,
qui remplirait mieux Ia lacune. De toute fa~on, !'identification avec notre
philosophe est assuree par le toponyme.
24 oM' liv) )(UTa TO[v] Xetov Mf)T[pwvu] o~oi\.oyWLTJ TWL ilD [daE]vm ~TJ8' U\JTO
TOU[TO.
25 Burnyeat (1978) 197-206, en particulier 204-206.

LE FRAGMENT DK

70

B 1 DE METRODORE DE CHIO

27

De (1) etde (2), on ne peut certainement pas deduire logiquement:

(3) "La proposition (1) est fausse".
En effet, la proposition (1) pourrait etre vraie sans etre connue
comme telle. Cependant, Burnyeat suggere de fa~on tres eclairante
qu'il faut replacer ce genre de discussion dans un contexte dialectique. Dans un tel contexte, la position du sceptique qui admet (1)
et (2) devient fragile, parce que "no sense is left to his furnishing reason,
evidence or proof for ( 1)". Tel est le sens qui donne sa plus grande
plausibilite a !'argumentation anti-sceptique de Lucrece: en un sens
large du terme, la position de ce sceptique tombe sous le coup d'une
m;ptTpom); dialectiquement, sinon logiquement, elle "se refute ellememe".
Mais, s'il en est ainsi, la consequence a tirer de !'analyse de
Burnyeat est que le principe de charite invite a privilegier, s'il s'en
trouve, les temoignages qui n'attribuent pas a Metrodore !'assertion
de (1). B est l'un de ces temoignages, et nous verrons qu'il yen a
d'autres.
Remarquons encore que B ne donne pas de sujet a l'infinitif ~i)
dotvat. Quine sait pas? On n'en sait rien.

c
Diogene Laerce ne consacre aucune notice a Metrodore, et c'est
dommage. 11 est impossible de savoir pourquoi il ne l'a pas fait,
puisque le philosophe de Chio ne lui est pas inconnu. 11 le mentionne au debut de sa notice sur Anaxarque (IX 58), a l'interieur
d'une "succession" democriteenne qui fait de lui, d'une part le
maitre de Diogene de Smyrne, lui-meme maitre d'Anaxarque,
d'autre part l'eleve de Nessas de Chio, ou selon certains, de Democrite lui-meme.26 A cette occasion, Diogene Laerce mentionne
!'opinion sans doute la plus frappante de Metrodore: "il disait ne
meme pas savoir cela meme, qu 'il ne savait rien". 27
26 Sur les variantes de cette succession, cf. Eus. Praep. Ev. XIV 17 10 = DK 69
A 1 (Metrodore eJeve de Nessas, lui-meme eJeve de Democrite), id. ibid. XIV
17 8 = DK 70 B 1 (Metrodore eJeve de Democrite), Clem. Alex., Strom. I 64 =
DK 70 A 1 (Metrodore eleve de Democrite). Ces variantes sont sans doute
Iiees a des incertitudes sur Ia chronologie exacte de Metrodore. Je ne puis
m 'attarder ici sur ce probleme (voir aussi DK 70 A 2 et Alfieri ( 1953) 28); une
datation relativement haute (fixant Ia naissance de Metrodore vers Ia fin des
annees -420) semble etre Ia seule compatible avec Ia mention faite de lui par
Ant~hane (cf. n. 23), dans une comedie que les specialistes datent de -388.
2

MrrcpoMpou

TOU

X{ou, 0(; Eii.Ey f1TJ6' UlJTO

TOUT'

dMvm,

on oMEv oT6.

28

JACQUES BRUNSCHWIG

Par rapport a A, C introduit d'abord une variante qui ne se retrouvera nulle part ailleurs, et qui est en elle-meme assez suspecte: elle
fait parler Metrodore de sa propre ignorance, non de celle des
hommes en general. Faut-il supposer qu'a tel ou tel moment de la
transmission a circule une version de l'enonce de Metrodore qui,
comme B, ne contenait aucune precision a ce propos, et que la
version de Diogene Laerce resulte d'une conjecture destinee a
combler cette lacune ? On peut au moins se poser la question.
On observe aussi et surtout, en C, un embryon d'expression
autonome de l'ignorance 1 ("cela meme, qu'il ne savait rien"). Cette
expression met dans une lumiere plus vive la difficulte deja
evoquee: si je ne sais2 meme pas que je ne saisl rien, pourquoi
designer par 'je ne sais 1 rien" 1' objet sur lequel porte mon ignorance2, plutot que, par exemple, par 'je sais 1 quelque chose" ? Une
position coherente, semble-t-il, serait plutot de dire: 'je ne sais2 pas
si je sais 1 quelque chose ou si je ne sais 1 rien ".
D

Un pas decisif dans l'evolution vers une assertion autonome de

l'ignorance 1 est franchi par la formule que Sextus Empiricus
attribue a Metrodore. Ce dernier figure au nombre de ceux qui,
selon le Sceptique, ont "aboli le critere";28 du moins, precise-t-il au
moment de s'en expliquer plus en detail, un nombre non negligeable de gens ont dit qu'ill'avait fait.29 Leurs raisons tiennent tout
entieres dans l'enonce qu'ils lui pretent, et qui se presente ici sous
la forme suivante: "(i) Nous ne savons rien, (ii) nous ne savons meme
pas cela, que nous ne savons rien ". 3
Cette formulation est probablement la plus claire et la plus
transparente de toutes celles qui nous sont transmises (ce qui ne
signifie pas necessairement qu'elle soit la plus digne de confiance
28 Xenophane, Xeniade, Anacharsis, Protagoras, Dionysodore, Gorgias,
Metrodore, Anaxarque, Monime (Adv. Math. VII 48).
29 Cette precision est donnee en Adv. Math. VII 87, a propos des trois
derniers philosophes de Ia Iiste reproduite dans Ia note precedente. Elle peut
s'expliquer, soit par le fait que Sextus estime ne pas disposer d'une documentation suffisante en ce qui les concerne (les paragraphes VII 49-87 contiennent
en effet, par contraste, des textes et des commentaires assez developpes concernant Xenophane, Xeniade, Anacharsis, Protagoras et surtout Gorgias), soit
parce qu'il n'est pas sur de !'interpretation qu'il convient de donner aux brefs
enonces qu'il est en mesure d'attribuer a Metrodore, Anaxarque et Monime.

30

OUI>ev '(oflE\1, oM'

UUTO TOUTO

'(OfJ.E\1

lSn ouMv '(OfJ.E\1.

LE FRAGMENT DK

70

B I DE METRODORE DE CHIO

29

sur le plan historique-peut-etre meme au contraire, bien entendu). On y voit notamment que I' incertitude qui affecte le sujet de
I' en once en B, et qui etait levee au profit de Ia premiere personne
du singulier en C, se trouve maintenant levee au profit (assurement
plus plausible) de Ia premiere personne du pluriel: c'est nous, les
hommes en general, qui sommes les victimes d'une ignorance
redoublee.
Mais surtout, D contient (pour Ia premiere fois dans l'ordre de
notre examen) une expression explicite et parfaitement assertorique
de notre ignorance 1 universe lie (Di), ainsi qu 'une expression claire
de Ia subsomption de cette ignorance 1 par une ignorance 2 (Dii).
II importerait beaucoup, pour notre enquete, de savoir quelle a ete
Ia genese de cette formulation. En suivant Ia suggestion de
Burnyeat, je tenterai de replacer le probleme dans un contexte
dialectique. On peut au mains imaginer, entre Metrodore et un
interlocuteur que je nommerai Antidore, deux enchainements de
dialogue qui auraient pu aboutir au meme resultat, par des voies
tres differentes.
Le premier enchainement serait le suivant:
Metrodore: "Nous ne savons 1 rien" (= Di).
Antidore: "Si tu l'affirmes, c'est que nous savons2 au mains ceci,
que nous ne savons 1 rien (ton scepticisme est inevitablement ce
qu'on appellera un jour un metadogmatisme negatif) ",31
Metrodore: "Non, nous ne savons2 meme pas cela, que nous ne
savons 1 rien" (= Dii).32
3! Barnes (1992) 4252 n. 54 et 4254 n. 72 a tres utilement distingue le
"dogmatisme negatir, qui consiste a soutenir des doctrines, mais seulement
des doctrines de forme negative, et le "metadogmatisme", qui consiste a
soutenir des doctrines, positives ou negatives, sur le statut cognitif de certaines
propositions. Bien entendu, comme le souligne Barnes, on peut etre metadogmatique negatif, si l'on soutient des doctrines negatives sur le statut cognitif
de certaines propositions. Si l'on ne distingue pas les niveaux de savoir ou
d'ignorance, !'objection d'Antidore tend a montrer que Ia position initiate
de Metrodore se refute elle-meme.
32 D a ete souvent rapproche de !'attitude d'Arcesilas par rapport a !'ignorance socratique (Arcesilas negabat esse quidquam quod sciri posset, ne illud quidem
ipsum, quod Socrates sibi reliquisset, Ciceron, Acad. I 45). Mais cela ne signifie
pas necessairement que Metrodore ait influence Arcesilas: Langerbeck (1935)
121-122 estime ajuste titre que les versions D etC peuvent resulter de !'insertion de Ia position de Metrodore dans un environnement platonico-socratique
qui l'eloigne irremediablement de son environnement democriteen originaire. Seton Straume-Zimmermann et al. (1990) 414, l'agnosticisme de Metrodore s'exprime de fa~on a desarmer !'objection de Ia nepLTpom'J, mais sa
formulation parait aussi "in irgendeiner Beziehung zur Aporetik des plato-

30

JACQUES BRUNSCHWIG

Antidore pourrait alors objecter, comme on l'a indique, que si

nous ne savons 2 pas que nous ne savons 1 rien, il n'y a aucune raison determinante de dire que nous ne savons 1 rien; or, c'est ce que
Metrodore avait commence par dire. En admettant qu'il ait ajoute
Dii a Di pour eviter de tomber sous le coup d'une m:ptTporrl), sa position, telle qu'elle est exprimee par la conjonction deDi et de Dii,
pourrait done tomber sous le coup d'une rrptTporrl) de second degre.
On peut cependant tenter d'eviter cette consequence en imaginant un autre enchainement dialogue:
Metrodore: "Nous ne savons 2 meme pas cela meme" (cf. B).
Antidore: "Que signifie cette enigme ? Qu'est-ce que ce Cela
meme>> que nous ne savons 2 meme pas ? Est-ce <<nous ne savons 1
rien>> ? Dans ce cas, ta formule reviendrait a dire <<nous ne savons 2
rien, me me pas cela, que no us ne savons 1 rien>> (= D). Es-tu
d'accord sur ce point ?".
Metrodore: "Non. Ce <<cela meme>>, dont je dis que nous ne le
savons 2 meme pas, ce n'est pas que nous ne savons 1 rien (rna
position serait alors fragile); c'est si nous savons 1 quelque chose ou
si nous ne savons1 rien. Ne confonc:\ons pas une completive avec
une interrogative indirecte. Je me resume: nous ne savons 2 meme
pas si nous savons 1 quelque chose ou si nous ne savons 1 rien".
Antidore pourrait sans doute montrer que cette formulation, si
elle evite Ia contradiction, risque d'entrainer une regression a
l'infini. Mais le schema que nous venons d'imaginer aurait au
moins !'interet de montrer: (i) que D pourrait etre, non !'expression
directe de la pensee de Metrodore, mais une reconstruction conjecturale de cette pensee; (ii) que cette reconstruction pourrait etre
concurrencee, non sans avantage, par une autre dans laquelle
!'objet de l'ignorance 2 serait exprime par une interrogative indirecte double, et non par une completive.
E

L'interrogative indirecte double que nous avons vu poindre en

filigrane, dans notre discussion de D, apparait en toutes lettres dans
E, la version donnee par Eusebe. Apres avoir donne un rapide
nischen Sokrates zu stehen". Cette expression prudente est cependant quelque
peu dementie par Ia suite de leur commentaire, ou ils affirment avec assurance qu'Arcesilas "hat den Satz Metrodors geradezu im Hinblick auf Sokrates
interpretiert und iibernommen".

LE FRAGMENT DK

70

B 1 DE METRODORE DE CHIO

31

aper~u des principes democriteens adoptes par Metrodore, Eusebe

dit que ce dernier a commence son traite De la nature par une entree
en matiere du genre que voici (dcrj3oA.u Toto.tm;J): 33 "(i) Aucun d'entre
nous ne sait rien, (ii) meme pas cela meme, si nous savons ou ne savons
pas".34
E presente un certain nombre de caracteres communs avec D:
I' assertion explicite de l'ignorance 1 (Ei), la determination du sujet
de cette ignorance 1 (non pas simplement "nous", comme en D,
mais "aucun d'entre nous"), et bien sur la formule caracteristique
ou&' O.UTO TOUTO. La difference a ne pas negliger est l'absence de la
completive OTL oul}f::v '{crp.EV et la presence de !'interrogative indirecte double rron:pov o'{l>O.p.EV Tl OUX. o'{l><J.p.EV.
Avant de voir quel parti l'on peut tirer de cette difference, il
importe de resoudre l'ambiguite syntaxique de Eii. 11 est possible
que !'interrogative indirecte serve a expliciter le contenu de ou&'
auTo TOUTO. La formule signifierait alors: nous ne savons 2 meme
pas cela meme, i.e. si nous savons 1 ou ne savons 1 pas. Dans cette
hypothese, les verbes o'{l>o.p.Ev et oux. o'{l>o.p.v n'auraient pas de
complement, et signifieraient, pris en un sens absolu, "nous avons
un savoir" ou "nous n'avons pas de savoir". Cette construction a ete
presque unanimement adoptee. 35 Pourtant, l'emploi absolu de
dMvo.t ne parait pas atteste. 36 De plus et surtout, la comparaison avec
D invite a penser que ou&' o.\no TOUTO a pour fonction de reprendre
ce qui precede, plutot que d'annoncer ce qui suit (quitte a repeter
explicitement ce qui precede dans ce qui suit, comme c'est le cas
en D). Sur ce modele, on serait conduit a construire oul>' auTo TOUTO
comme se rHerant a ce qui precede, c'est-a-dire a !'expression de
l'ignorance~o et comme etant grammaticalement le complement,
lance en prolepse, des verbes o'{l>o.p.Ev i) oux. o'{l>o.p.Ev. L'interpretation

33 Eusebe signale peut-etre ainsi que sa citation n'est pas textuelle.

34 OMd~; Ti!-LGiv oMev oTliEv, oM' a\rro ToiiTo, rrorEpov o'(ooj..I.Ev il oux o'(ooj..I.Ev.
35 Voir par exemple Ia traduction de DK (1952) 2.234 ("Niemand unter
uns weiB irgend etwas, nicht einmal eben das, ob wir wissen oder nicht
wissen"). Traductions similaires chez Bailey (1947) 3.1238, Cappelletti (1979)
109, Barnes (1982) 559, Des Places (1987) 169, Dumont (1988) 946, Vander
Waerdt ( 1989) 239 n. 39. Quelques traducteurs paraissent avoir senti Ia difficulte de cet emploi absolu des verbes o'(ilaj.LEV et oux o'(ilaj.LEV, auxquels ils fournissent des complements dans leurs traductions: ainsi Brochard (1923 2) 47-48
("Nous ne pouvons rien savoir, pas meme si nous savons quelque chose ou
rien"). De meme Zeller-Mondolfo (1969) 313-314, Sedley (1983) 14.
36 Sauf au participe (o! dlloTE(;: ceux qui sont au courant, les gens bien
informes).

32

JACQUES BRUNSCHWIG

a laquelle on aboutit pour E devient la suivante:

"(i) Nous ne savons

rien, (ii) nous ne savons meme pas si nous savons ou ne savons pas cela
meme (i.e. que nous ne savons rien)".
L'interet de cette construction est qu'elle introduit un troisieme
niveau dans l'etagement des degres d'ignorance. En effet, si l'on
veut preciser les choses, on aboutit au resultat suivant: nous ne
savons 3 meme pas si nous savons 2 ou ne savons 2 pas que nous ne
savons 1 rien. L'enchainement dialectique qui pourrait faire comprendre la genese de cette formulation serait le suivant:
Metrodore: "Nous ne savons 1 rien" (= Ei).
Antidore: "Mais au moins nous savons 2 que nous ne savons 1
rien".
Metrodore: "Non, nous ne savons3 meme pas si nous savons 2 ou
si nous ne savons 2 pas que nous ne savons 1 rien" (= Eii).
Antidore pourrait une fois de plus objecter: "mais alors, quelle
bonne raison as-tu de dire que nous ne savons 1 rien (= Ei) ?". La
vulnerabilite a cette objection est decidement le prix a payer,
semble-t-il, si l'on tient a detacher de fa~on autonome, en tete de
formule, !'expression assertorique d'une ignorance!> comme c'est
le cas avec Di et Ei.

F
11 est temps d'aborder le temoignage de Ciceron, le plus long, le
plus complexe, et aussi le moins eloigne de Metrodore dans le
temps. 11 figure dans une longue enumeration des diverses autorites qui peuvent appuyer, a des titres divers, la position de l'Academie philonienne. Apres avoir parle de Democrite, Ciceron passe a
Metrodore, "celui qui a le plus admire" ce dernier. 37 Confirmant
!'indication d'Eusebe, il dit qu'au debut de son livre Sur la nature,
Metrodore avait ecrit ce que Ciceron traduit en latin de la fa~on
suivante: (i) nego scire nos sciamusne aliquid an nihil sciamus, (ii) ne id
ipsum quidem nescire aut scire <scire> nos, (iii) nee omnino sitne aliquid an
nihil sit. 38

37 Is qui hunc maxime est admiratus. Malgre Mansfeld (1992) 33 n. 25

("Greatest pertains to the importance of Metrodorus"), i1 ne semble pas que
l'on puisse comprendre ici que Metrodore a ete le plus important des
admirateurs de Democrite.
38 Sur Ia troisieme occurrence de scire, voir ci-dessus, n. 9.

LE FRAGMENT DK

70

B 1 DE METRODORE DE CHIO

33

Ce passage pose des problemes d'etablissement du texte, de

construction syntaxique et d'interpretation. Comme ceux-ci sont
concentres dans Ia section Fii, on peut deblayer le terrain en
traduisant les sections Fi et Fiii, qui ne soulevent pas de problemes
philologiques particuliers:
(Fi) 'je dis que nous ne savons 2 pas si nous savons 1 quelque
chose ou si nous ne savons 1 rien". II est capital que !'assertion
initiate d'ignorance posee par Fi soit ici posee d'emblee au niveau
de l'ignorance 2 (par contraste avec Diet Ei), et que I' objet de cette
ignorance 2 soit exprime par une interrogative indirecte double
(comme dans Eii) et non par une completive (comme dans Dii).
(Fiii) "<Je dis aussi que nous ne savons> pas non plus le moins
du monde39 s'il existe quelque chose ou s'il n'existe rien". Cette
ignorance "ontologique" est entierement nouvelle, par rapport a
toutes les versions examinees jusqu'a present; il faudra y revenir.
Mais d'abord, examinons de fa~on plus precise Ia section Fii.
Quel que soit le texte adopte, il faut bien reconnaitre avec Langerbeck que Fii est "stilistisch unschon und gedanklich unklar", 40 Je
dresserai d'abord une petite typologie (non exhaustive 41 ) de quelques-unes des traductions donnees de Fii, afin de faire apercevoir
I' embarras des traducteurs et Ia variete parfois deconcertante de
leurs ten tatives.
A tout seigneur tout honneur: il convient de citer tout de suite
l'influente retraduction en grec d'un morceau de F (a savoir Fii et
Fiii) que DK ne craignent pas d'ajouter a leur citation de E: 42 ouo'
o:{n:o TO p.T) dMvm xal43 TO dMvat o'{6ap.Ev 44 (on Ecrnv). 45 On
39 Ou peut-etre: "nous ne savons pas non plus s'il existe absolument quelque
chose ou rien".
40 Langerbeck (1935) 122.
41 Je ne mentionne que pour memoire Ia traduction aberrante de Cappelletti (1979) 108: "Niego ... hasta que sepamos esto mismo: si en absoluto existe
algo o nada".
42 L'apparat critique (2.234) indique que cette reconstitution est due a
"Diels + Friedlander". Je n'ai pas trouve dans quel livre, article ou compterendu se trouve Ia contribution de Friedlander, et je ne sais done pas que lies
sont les parts exactes de chacun des deux erudits dans le resultat final (le nom
de Metrodore ne figure pas dans I' index de Friedlander ( 1969)). Cependant,
on peut deduire des indications d'Aifieri (1936) 335 que l'idee d'ajouter Ia
retroversion de Fiii a E revient a Diels; par elimination, on peut inferer que
c'est Friedlander qui a eu l'idee d'en faire autant pour Fii.
43 Sic ( le latin porte ici aut) .
44 J'ai deja signale (n. 9) que DK, dans leur citation de Ciceron, acceptent
I' occurrence de scire ajoutee par A2B2.
45 L'addition de Fii et de Fiii a E aboutit ainsi a une cascade assez suspecte

34

JACQUES BRUNSCHWIG

remarquera que DK n'ecrivent pas ouo' O.UTO TO\ho, ce qui leur

permet d'accrocher directement ouo' auTo aux infinitifs substantives qui traduisent nescire et scire. Cette decision a pour consequence que, dans le texte composite obtenu par le collage Ei + Eii +
Fii + Fiii, la liaison par ouo' o.uT6 n'a plus pour fonction de subsumer une ignorancen sous une ignorancen+l Autre consequence:
les infinitifs substantives TO ~n t0f.vut xo.l TO dOf.vut n'ont pas de
complement. 46 Ces deux consequences, on l'a remarque plus haut,
ne vont pas sans difficultes. De plus, !'expression TO ~n dOf.vut xo.l
TO doevut o'{oo.~v n'est guere satisfaisante, ni sur le plan syntaxique ni sur le plan semantique; c'est ce qui a conduit DK a ajouter
la parenthese (on crnv). On ne sait d'ailleurs pas exactement quel
sens ils veulent donner aux parentheses qui entourent on crnv:
s'agit-il d'une glose ou d'une conjecture a introduire dans le texte ?
La deuxieme hypothese est probablement la bonne, si I' on en juge
par la traduction allemande qui est donnee de ce texte a la meme
page ("noch eben vom Nichtwissen und Wissen wissen wir, daB
es ist").
Beaucoup de traduc<eurs et de commentateurs de Fii ont ete plus
ou moins influences, sur ce point, par DK. Les uns acceptent exactement I' interpretation suggeree par la retroversion de DK et l'adjonction de on crnv. 47 D'autres prHerent implicitement remplacer
l'adjonction de DK, on crnv, soit par Et crnv,4B soit par Tt crnv.49
Quelques traducteurs s'ecartent pourtant du schema de DK: au lieu
de construire nescire aut scire en un sens absolu, ils lui donnent pour
complement !'affirmation exprimee par Fi, attribuant ainsi a la
clause ne id ipsum quidem5 la fonction retroactive qui est celle de
ouo' O.UTO TO\ho dans les textes grecs.51
de clauses introduites par "meme pas", celle de Eii (oul>' auTO TOUTO), celle que
I' on tire de Fii ( ne id ipsum quid em) et celle de de Fiii.
46 C'etait deja le cas pour o'{l>ap.Ev et oux o'{l'ia(.l.EV dans Ia traduction DK de
Eii (cf. ci-dessus, n. 35) .
47 Ainsi Mansfeld (1992) 33, qui semble bien ne pas accepter !'insertion
du scire supplementaire ("[I deny] ... even the mere fact that we do not know
or do know"). Traduction similaire chez Rackham ( 1933) 561.
48 Ainsi Dumont (1988) 945 ("Nous ne savons meme pas s'il existe un
ignorer et un connaitre"). Cf. deja Goedeckemeyer (1905) 3.
49 Ainsi Bailey (1947) 3.1238 ("[I deny] ... even that we know what is
ignorance or knowledge"); de meme Appuhn (s.d.) 429-431, Barnes (1982)
559, Conche (1984) 1821.
5
Correctement a mon sens, puisque id doit etre Ia traduction ciceronienne de TOUTO.
51 Cf. Alfieri (1936) 334 et Dal Pra (19752) 1.52, dont Ia traduction suppose

LE FRAGMENT DK

70

B 1 DE METRODORE DE CHIO

35

Devant un texte aussi difficile, on peut s'etonner que les tentatives de correction conjecturale aient ete relativement rares.52 La
plus interessante me parait etre celle de Langerbeck, qui, apres
avoir critique le montage opere par Diels, remarque que !'insertion
de <scirf!> n'arrange rien, et note tres cursivement: "Offenbar ist das
nescire von Cicero der stilistischen blendenden Antithese zuliebe
eingefiigt".53 Dans sa hate, il a oublie de dire que si l'on supprime
nescire, il faut aussi supprimer aut. Mais il me semble que si l'on
suit jusqu'au bout la piste ouverte par Langerbeck, on aboutit a un
texte et a un sens tout a fait satisfaisants. Le texte de Fii devient:
"(nego) ... ne id ipsum quidem scire nos". Le sens qu'on peut donner a
ce texte (Langerbeck lui-meme ne l'explicite pas) est le suivant: 54
"Ue nie) ... que nous sachions3 cela meme (i.e. Fi, i.e. que nous ne
savons 2 pas si nous savons 1 quelque chose ou rien) ".
Les avantages de cette hypothese sont multiples. Tout d'abord, on
peut imaginer sans peine une genese dialectique de l' ensemble Fi
+ Fii, a partir de l'enchainement suivant:
Metrodore: "Nous ne savons 2 pas si nous savons 1 quelque chose
ou rien" (= Fi).
Antidore: "Mais cela, nous le savons3 (i.e. que nous ne savons 2
pas si nous savons 1 quelque chose ou rien) ?"
Metrodore: "Non, nous ne savons3 meme pas cela meme (i.e.
que nous ne savons 2 pas si nous savons 1 quelque chose ou rien) (=
Fii).
Deuxieme avantage: si l'on suppose (plus ou moins litteralement) authentique !'original grec de nego scire nos sciamusne aliquid
an nihil sciamus, ne id ipsum quidem scire nos, on voit assez bien
comment le texte de Ciceron a pu se corrompre en se compliquant:
dans un premier temps, !'alternative nescire aut scire se substitue au
simple scire, soit pour batir une antithese seduisante (comme le
que !'on conserve <scire>("[Io affirmo] ... che non sappiamo neppure se sappiamo o non sappiamo questa cosa stessa"); de meme, mais sans <scire>, Barigazzi
(1969) 291 et Straume-Zimmermann et al. (1990) 189.
52 U ne autre hypothese, suggeree par Carlos Levy (per litteram), serait que
Fi soit Ia traduction du texte de Metrodore par Ciceron, et Fii une traduction
alternative de ce meme texte, effectuee soit par Ciceron lui-meme, soit par un
correcteur posterieur. II n 'y aurait done pas lieu de chercher une difference de
sens entre Fi et Fii. La difficulte est alors que ne id ipsum quidem perd sa valeur
de surenchere par rapport a ce qui precede.
53 Langerbeck (1935) 122.
54 Rappelons qu'en syntaxe latine, quand Ia locution ne ... quidem suit une
negation, elle ne l'annule pas.

36

JACQUES BRUNSCHWIG

pense Langerbeck), soit plutot, peut-etre, par contamination avec

Eii; dans un second temps, l'adjonction d'un scire supplementaire
parait necessaire a Ia construction syntaxique de Ia phrase.
Troisieme avantage: toujours dans Ia meme hypothese, !'apparition des variantes presentees par les differents temoignages s'explique assez bien. La version de F a laquelle nous arrivons contient
a Ia fois une interrogative indirecte double dans Ia section (Fi)
(sciamusne aliquid an nihil sciamus) et une completive simple dans Ia
section (Fii) (scire nos). Ces deux elements se separent et se
presentent isolement, !'interrogation double dans E, Ia completive
dans D. En outre, F ne contient pas d'assertion independante
d'ignorance 1 universelle, mais seulement une assertion d'ignorance2, ce qui est aussi le cas pour B. L'assertion d'ignorance 1
commence a se detacher dans C, qui permet de voir comment l'on
aura eprouve le besoin d'extraire une telle assertion de l'assertion
metrodorienne de l'ignorance 2. Une fois ce mouvement parvenu a
son terme, l'assertion d'ignorance 1 s'enonce expressement (avec
tous les dangers dialectiques que cela comporte pour Ia position de
Metrodore) dans D et dans E. Dans A, enfin, il ne reste plus qu'elle;
pour cacher sa maigreur, on l'accompagne d'un commentaire
inspire de Xenophane.
II reste a dire un mot de Fiii, present dans F, absent de tout le
reste de la documentation. Peut-on attribuer a Metrodore le scepticisme ontologique de Fiii ? Oui, repondent implicitement DielsFriedlander, par le seul fait d'integrer la retroversion de Fiii dans
DK 70 B 1. lis ont ete suivis, ou meme precedes, par un nombre
considerable d'erudits.55 Une voix dissonante56 suggere qu"'il est
fort possible que cette idee [i.e. celle de Fiii] soit une addition
personnelle" de Ciceron. Je suis assez enclin a adopter pareille
55 Cf. Goedeckemeyer (1905) 3, Dumont (1988) 946, Dal Pra (19752) 1.52,
Barnes ( 1982) 559. Mansfeld ( 1992) 33 declare expressement que I' en once
sceptique de Metrodore est "partly preserved in Greek and complete in
Cicero's Latin" (y compris Fiii). Langerbeck lui-meme, qui n'est pourtant pas
indulgent pour Ia retroversion de Diels, et qui admet que Ciceron a pu
rassembler plusieurs enonces a partir de sa source pour obtenir de belles
antitheses, estime que Fiii est "gedeckt durch die Intention der ganzen
Stelle". En effet, juste avant d'aborder le cas de Metrodore, Ciceron (Acad. pr.
73) avait presente Democrite com me le tenant d'un scepticisme radical ( verum
esse plane negat). Puisque Democrite avait denie "l'etre du vrai", ajoute
Langerbeck (1935) 123, "il n'est que trop vraisemblable que Metrodore aussi
avait affirme un tel non-savoir a propos de l'etre".
56 Ernout et Robin (1926) 2.227.

LE FRAGMENT DK

70

B 1 DE METRODORE DE CHIO

37

hypothese: en effet, !'ignorance exprimee par Fiii redevient une

ignorance~> alors que nous avons trouve dans Fi et dans Fii des
raisons de penser que Metrodore n'avait decrit que des ignorances
de degre superieur; de plus, il serait assez etrange qu'une declaration aussi fracassante que Fiii n'ait laisse aucune autre trace dans la
tradition 57.
Concluons en laissant ouverte la question posee par Fiii. 11
semble en tout cas que Fi + Fii, historiquement et philosophiquement, constituent la version la plus plausible de la parole de
Metrodore.
U niversite de Paris I

REFERENCES BIBLIOGRAPHIQUES
Alfieri, V.E. (1936) Gli Atomisti. Frammenti e Testimonianze (Bari).
--(1953) Atomos Idea (Firenze).
Appuhn, Ch. (s.d.) Ciciron: De la Divination - Du Destin - Acadimiques (Paris).
Armstrong, D.M. (1973) Belief, Truth and Knowledge (London).
Bailey, C. (1947) Titi Lucreti Cari De Rerum Natura Lilni Sex (Oxford).
Barigazzi, A. (1969) "Epicure et le scepticisme", in Actes du VIII' Congres de
['Association Guillaume Bude: 286-93.
Barnes, J. (19822) The Presocratic Philosophers (London).
--(1992) "Diogenes Laertius IX 61-116: The Philosophy of Pyrrhonism", in
Aufstieg und Niedergang der romischen Welt, ed. W. Haase (Berlin) 11.36.6,
4241-4301.
Bonjour, L. (1980) "Externalist theories of empirical knowledge", Midwest
Studies in Philosophy 5: 53-73.
Brochard, V. (19232) Les sceptiques grecs (Paris).
Burnyeat, M. F. (1978) "The Upside-Down Back-to-Front Sceptic of Lucretius
IV 472", Philologus 122: 197-206.
Cappelletti, AJ. (1979). Ensayos sobre los Atomistas Griegos (Caracas).
Conche, M. (1984) "Metrodore de Chio", in Dictionnaire des Philosophes, ed. D.
Huisman (Paris) 2.1821.
Cronert, W. ( 1906) Kolotes und Menedemos (Leipzig).
Dal Pra, M. ( 19752) Lo scetticismo greco (Roma-Bari).
5 7 Pour dHendre l'authenticite metrodorienne de Fiii contre les doutes
d'Ernout et Robin, Conche (1984) 1821 ecrit: "Metrodore a pu s'apercevoir de
Ia contradiction qu'il y avait, chez Democrite, a poser des atomes reels pour
expliquer un sensible sans vraie realite, et etre ainsi conduit du scepticisme a
l'egard des sens au nihilisme ontologique". Mais Metrodore, autant que nous
sachions, n'a pas repudie !'ontologie atomistique; de plus, le "nihilisme
ontologique" n'est que l'une des branches de !'alternative dans laquelle Fiii
nous laisse en suspens.

38

JACQUES BRUNSCHWIG

DeLacy, Ph. ( 1971) "Ou !J.MA0\1 and the Antecedents of Ancient Scepticism",
in Essays in Ancient Greek Philosophy, ed. J.P. Anton and G.L. Kustas
(Albany) 593-606.
Des Places, E. (1987) Eusebe de Cisarie: La Preparation ivangilique, Livres XIVXV (Paris).
DDG = Diels, H. ( 1879) Doxographi graeci (Berlin).
DK = Diels, H. und Kranz, W. ( 19568 ) Die Fragmente der Vorsokratiker (Berlin).
Dumont, J.P. (1988) Les Prisocratiques (Paris).
Ern out, A. et Robin, L. ( 19622) Lucrece: De Rerum Natura. Commentaire exigitique
et critique (Paris).
Friedlander, P. (1969) Studien zur antiken Literatur und Kunst (Berlin).
Gigante, M. (1981) Scetticismo eEpicureismo (Naples).
- - (1990) "Quelques precisions sur le scepticisme et l'epicurisme", in Le
scepticisme antique - Perspectives historiques et systimatiques, ed. AJ. Voelke
(Geneve-Lausanne-Neuchatel) 69-83.
Goedeckemeyer, A. (1905) Die Geschichte der Griechischen Skeptizismus (Leipzig).
Hankinson, RJ. (1995) The Sceptics (London).
Hussey, E. (1990) "The beginnings of epistemology: from Homer to Philolaus", in Epistemology, ed. S. Everson, Companions to Ancient Thought: 1
(Cambridge) 11-38.
Isnardi-Parente, M. (1984) "I democritei e l'antiscetticismo di Epicuro (Ratae
Sententiae XXIII-XXIV)", in La storia della filosofia come sapere critico - Studi
in onore di Mario Dal Pra , ed. F. Angeli (Milano): 10&-121.
Langerbeck, H. (1935) t.OEIE EIIIPPYEMIH: Studien zu Demokrits Ethik und
Erkenntnislehre (Berlin).
Lesher, J.H. (1978) "Xenophanes' Scepticism", Phronesis 23: 1-21.
- - (1992) Xenophanes of Colophon: Fragments (Toronto).
Mansfeld, J. ( 1992) Heresiography in Context: Hippolytus' Elenchos as a Source for
Greek Philosophy (Leiden).
Plasberg, 0. ( 1922) Academicorum reliquiae cum Lucullo (Leipzig).
Rackham, H. (1933) Cicero: De Natura Deorum - Academica (Cambridge Mass.London).
Reid,J.S. (1885) M. Tulli Ciceronis Academica (London).
Schiiublin, C. und Graeser, A. (1995) Cicero: Academica (Leipzig) (non vidi ).
Sedley, D. (1983) "The Motivation of Greek Skepticism", in The Skeptical
Tradition , ed. M. Burnyeat (Berkeley-Los Angeles-London) 9-29.
Shorey, P. (1919) "Note on Metrodorus, fragm. 1", Classical Philology 14: 393-4.
Straume-Zimmermann L., Broemser F., Gigon 0. (1990) Hortensius - Lucullus
- Academici libri (Miinchen-Ziirich).
Vander Waerdt, P. A. (1989). "Colotes and the Epicurean Refutation of
Scepticism", Greek, Roman and Byzantine Studies 30: 225-267.
Wilamowitz-Mollendorf, U. v. (1899) "Lesefriichte", Hermes 34: 601-639.
Zeller, E., Mondolfo, R. (1969) La filosofia dei Greci nel suo sviluppo storico. Parte
I, volume V: Empedocle, Atomisti, Anassagora a cura di A. Capizzi (Firenze).

PART TWO

PLATO

OBSERVATIONS ON PLATO'S THRASYMACHUS:

THE CASE FOR PLEONEXIA
KEIMPE ALGRA

1. The second half of the first book of Plato's Republic, which

contains a discussion between Socrates and Thrasymachus on
justice, is undoubtedly among the most widely studied parts of the
work. This popularity is not surprising, for the episode not only
makes for pleasant reading, but also constitutes an intriguing piece
of philosophical dialogue. It is intriguing because it sets an important philosophical challenge which the rest of the Republic tries
to answer, whereas the dramatic presentation is apparently such as
to have given rise to quite divergent interpretations of what exactly
Thrasymachus is made to say. 1 The present paper does not pretend
to offer an entirely novel interpretation of the Thrasymachus
episode. It rather contains some observations which support and
elaborate existing scholarly interpretations, most notably those of
Kerferd on Thrasymachus' definitions, Kahn on the way Rep. I fits
into the rest of the work, and Rutherford on the literary and dramatic elements.2 By bringing these various approaches together, I
hope to be able to contribute to an account of Thrasymachus'
position which is both coherent and securely based on Plato's text.
I have had to be selective in my references to the extensive
scholarly literature on the Thrasymachus episode. Only those
studies are mentioned which I took to be directly relevant to the
points I wish to make. I have also had to refrain from adding
detailed discussions of ( 1) the question of the genesis of the Republic
and of the ancestry of book I, (2) the question of whether Plato's
Thrasymachus represents the sophist Thrasymachus of Chalcedon, and (3) the relative merits and demerits of the various
approaches to Plato (more purely 'analytic', or more literary and
'Wilamowitzean'). A few dogmatic statements to mark off my
1 For a convenient overview of the various ways in which the position
defended by Thrasymachus, and in particular his definitions of justice, have
been interpreted and labelled see Cross/Woozley (1964) 29-41 and Kerferd
(1947) 545-548.
2 Cf. Kerferd (1945) and (1981), Kahn (1993), and Rutherford (1995).

42

KEIMPE ALGRA

position on these matters must here suffice. As to (1), although the

view that Rep. I is to be traced back to an originally separate 'Socratic' dialogue, which some have even christened Thrasymachus, still
does find some support, 3 it is now generally assumed to rest on
poor evidence. For our present purpose, many of the items which
are at stake in the context of this discussion may be ignored, except
for the following points. First, book I is definitely less 'Socratic'
than it may at first blush appear to be. Although it formally ends in
aporia, it has Socrates defend what is, albeit in nuce, a dogmatic
position. 4 Secondly, as was already pointed out by Wilamowitz, we
shall have to assume that, even if Plato originally wrote a Thrasymachus as a separate piece, he must have redesigned it subsequently to fulfill a meaningful role in the Republic as a whole. 5 He
explicitly refers to it as a prooimion to the rest of the work (357a2),
and he refers back to Thrasymachus' position on various occasions
in the later books of the Republic. 6 Moreover, as has now been forcefully argued by Charles Kahn, the first book contains so many
detailed anticipations of themes that are further worked out in the
course of the Republic that even the hypothesis of a secondary
reworking becomes rather implausible, because in that case we
would have to assume that Plato re-wrote more than half of the
whole piece. 7 As to question (2), I shall be assuming that even if the
character of Thrasymachus in Rep. I could be proved to be identical
with the historical Thrasymachus of Chalcedon-which it clearly
3 A sustained, though not very strong, defense of the hypothesis of a
separate Thrasymachus was provided by Von Arnim (1914) 71-109, and (1927).
A qualified and more sensible version of the hypothesis was defended by
Wilamowitz (1920) 2. 182-184, and Friedlander (1964) 2. 45-60. It is still
defended by Vlastos (1991) 250, and in a slightly modified form, by Gigon
(1976) 82, who thinks that Rep. I in its present state represents an abbreviated
version of an earlier work, with a number of additions designed to make it
fit into the whole of the Republic. For a more extensive survey of the history of
the 'Thrasymachus' hypothesis and connected issues, see Zeller (1922) 2. 1.
488 n. 1 and 555-556; Guthrie (1975) 437, and Kahn (1993) 131-133.
4 This point has been well brought out by Gigon (1976) 81-82.
5 Wilamowitz (1920) 2. 182-185. According to Wilamowitz Plato did not
publish the original dialogue, because he had in the meantime written the
Gorgias which more or less covered the same material in a better way, and
later re-worked it for the Republic.
6 The position defended by Thrasymachus in book I is referred to in Rep.
II, 358a-c; 367a5; 367c2; 368b5; VIII 545a8; IX, 590d3
7 Cf. Kahn (1993) 136 and 139. The fact that much of book I actually foreshadows what is to follow later on in the Republic has of course been noted by
others as well, most notably by White (1979). That there were any such
prefigurations in book I was flatly denied by Von Arnim (1914) 72.

OBSERVATIONS ON PLATO'S THRASYMACHUS

43

cannot-historical accuracy was obviously not Plato's primary

aim. He rather appears to have presented Thrasymachus as the
fountainhead of a dialectical position to be refuted. 8 This means
that we should first and foremost study Thrasymachus as a persona
in Plato's dialogue (as the title of this paper indicates), and that we
should realize that the discussion between Socrates and Thrasymachus is basically Plato's own creation. That should alert us to the
fact that the way Socrates and Thrasymachus interact may be
determined by Plato's overall aims rather than by the actual contents of what the two characters are made to say. In other words,
the dialogue should be compared to a (skilfully staged) piece of
work of a dramatist, rather than to the report of a game of chess. 9
Finally, as to question (3), we may note that in 1975 Guthrie
could still write that 'in discussing the views attributed to Thrasymachus the most recent practice had been to consider all possible
alternatives as they appear to a philosopher today, and by exhaustive examination of the dialogue endeavour to decide which of
them is being maintained by Thrasymachus'. He added that 'such
clarification can be most valuable, yet may err by neglecting (as it
is never wise to do with Plato) the dramatic situation and emotional tension between the speakers' . 10 Since Guthrie wrote these line
there has been a growing tendency to complement the predominantly analytical approach of Plato's philosophical arguments-an
approach which in the case of the Thrasymachus episode came
down to focusing primarily on the definitions of justice and their
alleged philosophical import-with an approach which takes
account of the larger context, philosophical as well as dramatic.I 1 It
8 It is therefore methodologically preferable to refrain from using what
little is known of the philosophical ideas of Thrasymachus of Chalcedon to
interpret the text of Rep. I, as has been done by Guthrie (1971) 3. 92-97, and
esp. 297, who tries to square the picture of Thrasymachus which arises from
Rep. I with the evidence of Thrasymachus of Chalcedon fr. 1 Diels-Kranz (a
speech to the Athenian assembly in which Thrasymachus appears to speak in
praise of justice and which would thus seem to be at odds with Rep 1.) by
interpreting the Thrasymachus of the Rep. as a desillusioned and bitter man,
a claim for which there does not appear to be any clear evidence in the text.
9 Cf. the pertinent remarks in Reeve ( 1988) 276, note 5, and Annas ( 1981)
56. The recognition that neither Socrates nor Thrasymachus are writing
their own lines should make us skeptical about the possibility of separating
an allegedly 'original' position of Thrasymachus from additions which are
supposedly due to Plato; an example of such an approach is Maguire ( 1971), on
which see below, n. 39.
to Guthrie (1971), vol. 3, 91 .
11 An extremely useful example of this approach is Rutherford (1995); but

44

KEIMPE ALGRA

will be clear from what follows that I happily endorse the latter
approach.
The present paper is structured as follows. Section 2 contains a
summary which will serve as a basis for the analysis which is to
follow. Section 3 discusses the way the Thrasymachus episode
connects with the rest of the Republic. Section 4 will deal with
Plato's literary and dramatic presentation. Section 5, finally,
contains some observations on how Thrasymachus definitions of
justice interrelate and how they are to be interpreted.
2. After the vivid introduction of Thrasymachus as a persona (336b337a, on which see section ( 4) and after some initial quarrelings
between Thrasymachus and Socrates about method (337a-338c),
the discussion gets started with Thrasymachus' claim (put forward
with characteristic self-consciousness: CJ.')..J..a d oiHc E1t<XtVEtc;) that
justice is 'the advantage of the stronger' ('to wu KpEt't'tovoc; [... ]
OUJ.lq>Epov, 338c). Of course put like this the claim lacks philosophical precision. Hence the Socratic method is applied in order to get
things straight. It turns out that Thrasymachus is not speaking of
corporeal strength, and that by 'the stronger' he means the person
or persons in power ('tO 'ttlc; K<X9EO'tT\1CUtac; apx.ilc; OUJ.lq>Epov, 339al) . 12
In others words, in the context in which he launches this first
definition he is primarily thinking of justice in the political sphere,
of justice in the state.
see also Stokes ( 1986), where, however, the Rep. is not discussed; Muller
(1981) represents a similar approach, but pays little attention to the Thrasymachus episode.
12 Note, first, that the phrase -ro -rl\c; Ka8E1Hl]K'Iliac; &pxl\c; OUJ.upepov clearly
figures as a further specification of the first definition ('the advantage of the
stronger'). as may be inferred from what immediately follows: autl] OE ltOU
KpU'tEl, OOO'tE ou~~aivn -rep 6p8&c; A.oyt~O~Evcp ltavtaxou dvat to auto OiKatOV, tO tOU
KpEi ttovoc; OUJ.HpEpov, 339a2-4). It does not involve any shift of position on
Thrasymachus' part. Secondly Thrasymachus makes it quite clear that he
regards it as an empirical fact that rulers legislate in their own interest
(338e1-2: -ri8Etat OE YE toile; v6~ouc; EKtXOtl];, cXPXft ltpO<; to autfl OU~<pepov). It follows
that, when at 339b7 he assents to Socrates' suggestion that justice is 'obeying
the rulers' (1tEi8eo8at [ ... ] to'ic; iipxouotv) he takes it to be implied all the way
along that these rulers as a matter of course serve their own interest.
Accordingly, pace Maguire (1971) 143, we cannot read 339b7 as containing a
new and independent definition of justice. Thirdly, the notion of to au~<pepov
is prominent right from the beginning of Thrasymachus' account; this
means that the label 'conventionalism' is not very suitable to describe this
position, pace Annas (1981) 40. What may count as justice is never a matter of
mere convention.

OBSERVATIONS ON PLATO'S THRASYMACHUS

45

Socrates now points to the possibility of the rulers being mistaken as to their own interests. Thrasymachus rejects the suggestion of Cleitophon that he can dodge this objection by specifying
that justice consists in whatever the rulers (rightly or wrongly)
think to be in their own interest. Instead he boldly claims that the
rulers qua rulers-i.e. qua possessing their specific skill ('tEXVll)are never mistaken. Interestingly, Socrates does not challenge this
claim as such. He rather takes advantage of the fact that Thrasymachus is willing to speak of the ruler's job as a 'tEXVll, and of a ruler
'in the strict sense', and of the fact that he is prepared to compare
the ruler's art to the art of, say, a doctor. Socrates does so, however,
by shifting the ground. He retorts that anyone exercising a 'tEXVll in
the strict sense is not acting in his own interest. By way of an
elenctic procedure which Thrasymachus is unable to withstand,
he concludes (342e6-ll) that the ruler, qua ruler, will always be
guided by the interests of his subjects. This conclusion is presented
as a reversal of Thrasymachus' own account (o 'tOU Otlcatou /..6yor;
Eir; 'touvav't{ov 1tEptEtO"'tTJKEt, 343a2), which it obviously is to the
extent that the rulers are now said to decree whatever they decree
in the interest not of themselves but of others, viz. their subjects.
However, instead of going along with this conclusion Thrasymachus accuses Socrates of naivete (he doesn't 'know about sheep
and shepherd', 343a8). It is now his turn to shift the focus from the
political level to the level of individual human beings who, he
suggests, are not as Socrates pictures them. And he vigorously
restates his own position in what I shall henceforth refer to as his
'speech' (343bl-344dl). As his primary definition he now proposes
that justice is 'what is good for someone else' ( a'A/..6'tpwv aya06v,
343c2) in the sense that it is contrary to the subjects' own interests.
He contrasts a number of examples of just and unjust behaviour to
show that the unjust, if successful, always gain the advantage, and
repeats that justice is the advantage of the stronger, whereas
injustice is 'one's own profit and advantage' (344c8).
Socrates replies with a series of questions (345b7-347el) designed
to show that the example on which Thrasymachus has based his
case-a shepherd tending his flock in his own interest-was
wrong, because the shepherd (qua shepherd) turns out to act in the
interest of his flock. In general, he argues, this goes for every
'tEXVll, including governing according to the rules (Ka'A&r; 'tTl 'tEXVU,
347al), which is why good rulers are so difficult to get. In fact they

46

KEIMPE ALGRA

are only willing to govern if they get a proper compensation, or

rather in order to evade the punishment of being ruled by others
who are less competent. In a city where this risk would have been
eliminated-i.e. in a city consisting of good people only-everybody would prefer the status of being a subject to that of being a
ruler (347dl-8). For the time being Socrates continues to disagree
with Thrasymachus' contention that justice concerns the advantage of the ruler, but he announces that this subject is to be further
treated later on (347e2: af..)Jx 'tOU'tO JlEV OlllC<Xt eit; ailett; O'lCE'IfOJleSa).
He now turns instead to what he takes to be the more important
issue, viz. Thrasymachus' claim that the life of the unjust is better
and happier than the life of the just. This will be the subject of the
rest of the discussion (347e7-354all). A brief survey may here
suffice. Socrates comes up with three arguments which are rather
different in kind. First of all, he tries to establish that Thrasymachus' claim that injustice is a kind of apE'tft gets him into
trouble, for this would involve that justice is a JCaJCia, a view so
counter-intuitive that not even Thrasymachus is willing to endorse
it (348b8-349a8).13 Furthermore, the injust cannot be superior in
<ppOVT]O'tt; either, since <ppOVT]O'tt;, which is the possession of a
particular skill, entails that one only wants to outdo those who are
inferior insofar as they do not have that skill. This criterion applies
to the just, not to the unjust who want to outdo everybody (the words
used are 7tAEOV EXEtV and 7tAEOVEJC'tEtV, both connected to the noun
7tAEOVE~ia), i.e. both the just and the unjust (349bl-350dl0). Secondly (350d-352b), Socrates challenges Thrasymachus' contention that
injustice is a source of strength. It appears that injustice, on the
contrary, involves faction, hatred and disunity, on the level of
states as well as between individuals, and even within a single
individual. Thirdly, each thing has its proper function (epyov);
exercising this function well is what virtue (apE'tft) amounts to;
living is the function of the soul; since justice is the the particular
virtue of the soul (this is an element which in the present context is
question-begging, and which Thrasymachus need not accept),
living well amounts to living a life of justice. The final section of
the book shows us a Thrasymachus who is more or less silenced,
but who remains unconvinced, and a Socrates who admits that he
has not as yet done his job properly: he has been hopping from
13 See on this Annas (1981) 47-49.

OBSERVATIONS ON PLATO'S THRASYMACHUS

47

subject to subject without clearly realizing that the main question,

the definition of justice, needs to be answered first. This, of course,
is what will be his task in the rest of the work.
3. The passages we have just summarized contain various
elements which foreshadow the more elaborate discussions which
are to follow in the rest of the Republic:
(1) Socrates' claim that true and competent rulers do not rule in their
own interest but in the interest of their subjects anticipates the discussion
in books III and IV of the characteristics of the rulers in the ideal state
(starting with the claim at 412c13-4 that the rulers will have to be
~TJOEJlOVa<; tTl<; 7tOAEoo<;), and in book VII, 540b (apxovta<; [... ] tTl<; n61.Eoo<;
EVElCU ).

(2) Related to (1) is Socrates' insistence (345b7-346e1) on the fact that a

good ruler's ruling will never be entirely voluntary, which foreshadows
his claim later on (519d-521b) that the rulers in the ideal state will take on
their job only unwillingly. Indeed, as we already noted, he even
explicitly states that this is a subject which will be postponed until later.
(3) Socrates's claim (352d-254a) thatjustice is a virtue of the soul in the
sense that it constitutes its health and its proper functioning, is not
justified in the context of Rep. I and indeed only becomes intelligible
against the background of the moral psychology worked out in book IV,
in particular 444a-e.
(4) The idea that unjustice is a source of strife ( crtacrt<;) at the level of
states or smaller social groups (350d-352b) might in principle be justified,
independently of the philosophical commitments of Socrates or Plato, by
an appeal to common sense-for this reason it is within the limited
context of Rep. I certainly the most powerful argument raised against
Thrasymachus-but the idea that this same analysis applies to the injust
individual is neither an element of common sense, nor a claim to which
Thrasymachus needs to be committed. 14 In fact it anticipates the thesis of
the parallelism between (the parts of the) polis and (the parts of the)
individual which is developed later on in book IV (368d-e and 435e-436a);
the conception of stasis within an individual soul is of course further
developed in the context of the theory of individual justice being a
healthy harmony between the three parts of the soul (441c-442d), and in
book VIII-IX, in the extensive account of the decline of the ideal state and
in the subsequent analysis of happiness in relation to psychic
constitution (576b-592b, esp. 588e-589a).
(5) The general theme of the unresolved controversy between
Thrasymachus and Socrates in the final part of Rep. I (viz. whether
justice or injustice is more beneficial to its possessor) is taken up again at
the end of book IX, in 588b ff., where the question is answered on the
basis of the findings of the Rep. as a whole (59la-592b; see also X, 612b).
14 See also Annas (1981) 53, who does not consider the possibility of an
'anticipatory' function of this passage.

48

KEIMPE ALGRA

(6) The idea of injustice as pleonexia (which occurs in Thrasymachus'

speech in 34lb-344d, and in 349a-350c) and the idea that people are
naturally inclined to pleonexia and will accordingly only behave justly if
they have to (an idea which runs through the whole of Thrasymachus'
account; on which see my remarks in the next sections) recur at the
beginning of book II, in 358a-362c, where Glaucon slightly reformulates
Thrasymachus' position by way of a challenge to Socrates; the concept of
pleonexia itself reappears in the discussion of the tyrant in book IX (574a8;
586bl).
(7) Thrasymachus' reference to tyranny as the ultimate form of
i~ustice (tEAEWtcltll aouda, 344a) has its more elaborate counterpart in
book VIII-IX, where Plato presents his account of the decline of the ideal
state as a means to make clear how perfect justice stands to perfect
injustice (tyranny). His aim is to get us in a position to decide whether
Thrasymachus or Socrates was right (545a: 'iva tov aOtKrotatov i06vtE<;
avn9ro~EV tij} oucmotatcp Kat i]~'iv tEAEa i] <JKE\jft<; ft, n&<; 7tOtE i] IXKpa'tO<;
OtKato<JUVTI1tPO<; aOtKiav tl]v aKpatov EXEl EUOat~ovia<; tE 7tEpt tOU lfxovto<;
Kat a9A.t6tlltO<;, 'iva ~ 8paau~axcp 1tEt96~EVOl OtcOKW~EV aOtKiav ~ 'tip vuv
7tpO<patVO~EVq> A.Oycp OtKatO<JUVTIV).
On the basis of this list-which, I believe, covers the mam
issues, but which may not be exhaustive1 5- I would like to make
the following observations.
First, I take it to be obvious that we are here dealing with a
number of instances of what Charles Kahn has labelled 'proleptic
composition'. In introducing this label Kahn took his cue from a
1971 study on Aeschylus and his imagery by Anne Lebeck. In
Aeschylus, according to Lebeck, we witness that 'the significance
of a recurrent image unfolds in successive stages' and 'in its early
occurrences the image is elliptical and enigmatic'. 16
In the present case, however, we need not resort to such remote
parallels in Aeschylean drama, interesting though they may be
in themselves. As we noted, right at the beginning of the second
book Socrates remarks that at this point he thought the discussion
was completed, but that it turned out that the whole thing had
really been nothing but a prooimion (f.yro f..LEV oi>v taum et7tWV cilf.lrtV
A6you nmtA.A.ax8m 'tO 8'~v apa, roc; tOtK, 7tpOOlf.ltoV, 357al). Now we
15 Kahn (1993) 136-140 provides a somewhat differently structured list,
which at any rate does not include my items (5) and (6). On the other hand
he provides two extra examples which I have here left out, because I found
them not quite as obvious and convincing as the other ones: (1) the perfect
knowledge which Thrasymachus' rulers are supposed to possess serves to introduce, according to Kahn, the general theme of the perfect ruler (op. cit. 138),
and (2) the claim that different technai have different capacities serves to prepare the distinction between knowledge and opinion in book V (esp. 477d1).
16 Lebeck (1971) 1.

OBSERVATIONS ON PLATO'S THRASYMACHUS

49

know from ancient rhetorical treatises that a proem was supposed

to attract the attention of the readers or the public and to announce
the theme(s) of the main body of the work. As Aristotle puts it
(with special regard to drama and epic poetry), 'in prologues and
epic poetry a foretaste of the theme is given, intended to inform the
readers of it in advance instead of keeping their minds in
suspense'. 17 But we can also turn to Plato's own words. In the fourth
book of the Laws the Athenian tells his two companions that ideal
laws need to have preambles (prooimia) and that their own conversation from early morning onwards, while walking through the
Cretan mountains, has in fact been such a prooimion. He then adds:
'What do I mean? I want to say this: in all kinds of speeches and
other vocal performances there are proems and, so to speak,
preliminary exercises (1tpOOtJ.LUX 'tE EO''ttV ICCl.t crxeoov oi6v 'ttVE~
avmnvftcrn~) containing a skilled first approach which is useful
with regard to what is going to be accomplished (exoucra{ nva
vtexvov mxeip11crtv xpflcrtJ.LOV 7tpo~ to J.LEAAov 7tEpa{vecr8at) .18
The two main elements of this description-viz. (1) a proem is a
kind of preliminary exercise and (2) it must be useful with regard
to what follows-both neatly apply to the discussion of book I. It is
clearly a preliminary exercise in so far as it broaches the subject of
justice and shows some of the main items with which a full
discussion should deal. It is also useful with regard to the rest of the
work, in so far as Plato's own view as well as the view diametrically opposed to it (Thrasymachus') are both introduced in such a
way as to attract the attention of the reader.
Like the 'proleptic' passages in Aeschylus discussed by Lebeck
and referred to by Kahn, a proem will often contain elements
which are more or less enigmatic in so far as they merely
summarize or announce subjects which will be worked out more
fully later on in the work. In this respect Aristotle's claim, quoted
Rhet. 1415al3-14. See also Rhet. Alex. 29.1.
Leg. 722d2-6. Note that the attached example of a musical prooimion- i.e. a
prelude, which precedes the melody proper (v611o~: of course the choice of this
homonym of v6~to~ in the sense of 'law' is deliberate) - is also used by Plato in
Rep. 532d7 ('let us now move to the melody proper and perform it just as we
have performed the prelude') and in Tim. 29d5 ('we liked your prelude, now
proceed with the melody'). For examples of Plato's use of npooi~ttov (and
npoot~tux~eaeat) in the general sense of 'introductory conversation', see Phdr.
266d7; Lach. 179al; Rep. 432e8; for the use of prooimion in the sense of a 'preamble' to a law, see the whole context of Leg. 722-725. On the prooimion of the
Tim. see now Runia ( 1996).
17
18

50

KEIMPE ALGRA

above, that proems are meant to clarify matters and to mznzmzze

suspense, should be qualified. Jaap Mansfeld, to whom this article
is dedicated, has convincingly argued that in the case of some
Presocratic proems we encounter a form of intentional unclarity,
designed precisely to strengthen the element of suspense.19 In the
case of Rep. I we are arguably dealing with a weaker, but comparable, form of intentional unclarity: the diametrically opposed
positions of Socrates and Thrasymachus are clear in themselves,
but the arguments needed to clinch the issue will only be provided
later on. And here also this procedure arguably creates a kind of
suspense. Like Glaucon, at the beginning of the second book, the
average reader will probably sympathize with Socrates's position
rather than with that of Thrasymachus. And like Glaucon, he will
ask for more: he will not be satisfied with the avowed inadequacy
of Socrates' critique nor with the formal aporia which concludes
the book.20
If, then, the arguments in book I sometimes appear to be
truncated and if the discussion as a whole often strikes us as
incomplete, we should realize that we are only dealing with an
avaKtVTt<H<;, a first sounding of the important Leitmotive of the rest of
the work. There is accordingly no need to to go along with Gigon's
hypothesis that what we have before us is an abridged version of a
more elaborate original and separate dialogue.2 1 Indeed, the recognition that Rep. I has a prefatory character may help us to understand why the piece is not all that aporetic, despite its 'Socratic
style' and despite Socrates' final admission that he still 'knows
nothing' (354b59-cl). Thrasymachus is introduced in order to set
out a position which is the exact counterpart of the position Socrates
is going to work out later on in the Republic, and this means that
Plato could not use a neutral, non-knowing Socrates in book I, but
opted instead for a Socrates who opposes Thrasymachus with
conviction and vigour on all accounts (on whether the rulers'
ruling is in their own interest, on whether injustice is profitable).
In addition the prefatory character of the Thrasymachus episode
arguably offers an explanation of why Thrasymachus is made to
19 Mansfeld (1996).

Cf. Glaucon's words at 357a5: ltOtEpov iJJlCx~ ~OUAEl 001CelV ltEltEllcEVat f\ ro~
UAT]9&~ ltEt<Jill Otllt!lvtl tpomp aJlElVOV eatt oiK:awv dvat f\ aOtKOV. And 358b2-6:
20

E>pacrUJlllXO~ yap

JlOl <patVEtat 1tpcpahepov tOU oeovto~ U1t0 crou WcrltEp O<pl~ KT]AT]Si\vat,
EJlOl o oUltCO Katc'x vouv f] U1t00El~l~ yeyovev ltEpt EK!ltEpoU.
21 Gigon (1976) 82-83.

OBSERVATIONS ON PLATO'S THRASYMACHUS

51

come up with two different-although, as I shall try to show

coherent-definitions of justice ('the advantage of the stronger' in
338c, and 'what is good for someone else' in 343c). It is usually
assumed either that this reflects the fact that Thrasymachus is
simply a confused thinker, or that he merely states his position in
an obscure way in 338c but is forced by Socrates' criticisms to come
up with a better version in 343c.22 As I noted in the introductory
section, however, neither Thrasymachus nor Socrates are writing
their own lines. We should therefore perhaps first ask whether
Plato had a reason for this 'double' presentation. If we realise that
the discussion is supposed to be a kind of proem, we are presented
with such a good reason. As is well known, there are two main
strands in Plato's discussion of justice in the rest of the Republic.
The investigation is carried out at two different levels: justice in the
state (with the concomitant issues of who should rule and why)
and justice in the individual (with the concomitant issue of what
constitutes the happy life). Of course the two strands are connected
in various ways, most noticeably by Plato's anthropology (i.e. his
tripartite psychology). My suggestion is that Plato wanted his prooimion to contain an exposition of the position diametrically opposed
to his own on both accounts, i.e. he wanted Thrasymachus to define
justice both at the political level ('the advantage of the stronger')
and at the level of the individual ('what is good for someone else').
4. On the basis of the evidence discussed in the previous sections it
may reasonably be suggested that Thrasymachus' position is to be
characterized as being essentially a defense of pleonexia. The term
7tAEOVE~{a as well as the connected verbs 7tAEOV EXEtV and 7tAEOVEK'tEtV all occur prominently, and exclusively, in those passages in the
Rep. which refer to Thrasymachus' position, i.e. his own speech,
the rephrasing of his position by Glaucon and Adeimantus in book
II, and the description of (Thrasymachean) tyranny in book VIIIIX.2~ We appear to be dealing with key-words. The typically
Thrasymachean motif of unlimited self-assertion appears to have
22 See below, nn. 39 and 40.
2~

In the Thrasymachus episode in book I we find 7tAEOVE~ia at 359c5, 7tAEov

at 343d5; 349b3-el2 (4 times) and 7tAEOVEKtEtV at 344al; 349 b8-350cl (10
times). In the 'paraphrase' in book II we find 7tAEOVEKtEtV at 362b7 (2 times)
and 365d6. In the context of the discussion of tyranny in book VIII-IX we find
7tAEOVE~ia at 586bl, and 1tAEOV exnv at 574a8. This is the extent of the use of
these terms in the Republic.

EXElV

52

KEIMPE ALGRA

been deliberately opposed to Socrates' (and Plato's) own stress on

order and limitation. And just as Plato's own two accounts of
justice (political and individual) are linked by what we might call
his anthropology of restraint and self-control (characterised by
such words as KOOf.LO~, KOOf.ltO~ and llf.LEpo~),2 4 so also Thrasymachus' two definitions may be seen as linked by what we might
call an implicit anthropology which is governed by the notion of
pleonexia: people are naturally inclined not to respect other people
and their possessions, to intrude into the areas of others and to set
the rules for them, and to disregard any limits that might check
such behaviour. I want to suggest that this general picture of Thrasymachus defending pleonexia is skilfully supported by the way
Plato has fleshed out the persona of Thrasymachus dramatically.
In general, it is worth noting that Plato's characterisations of
Socrates' interlocutors are usually more than simply the icing of
the cake. Their character and behaviour are often made to contrast
strongly with Socrates, thus underlining the fact that they are committed to an entirely different philosophical outlook or way of life.
One of the most striking examples is Alcibiades in the Symposium.
Others are Polus and Callicles in the Gorgias. But also in a more
general way-i.e. apart from any intended dramatic contrast with
the persona of Socrates-Plato's personae are often adapted to the
ideas they are made to express. Here again the Symposium offers a
most vivid example in the persons of the tragic poet Agathon and
his foil Aristophanes. But one may of course add Euthyphro in the
dialogue bearing his name, (exemplifying piety), Laches and
Nicias in the Laches (exemplifying various conceptions of courage),
the youthful quartet of Hippothales, Lysis, Ctesippus and Menexenus in the Lysis (exemplifying erotic philia), and Charm ides in
the Charmides (exemplifying sophrosuni). 25
24 Most notably in 410e1-3, where the philosopher is said to have a nature
which is ~l!EPO<; and which, if properly trained, will turn out to be ~l!EPO<; tE
lCCXt JCOIJI!to<;. For JCOIJI!O<; and JCOIJI!to<; see further 329d4; 331bl; 339ell (lCOIJI!tOu tE
Kal. civopeiou); 403a7 (1Cocr11iou te JCCXtJCcxA.ou); 408a8 ({rytnvouc; te JCcxtJCocrlliouc;); 430e6
(IJCI)(ppocrUVTl as lCOIJI!Oc; and eyJCpcitetcx); 443d4; 486b6 (1COIJI!to<; opposed to aOtlCo<;);
503c4; 539d4 (lCOIJI!lOU<; [... ]lCCXt IJtCXIJll!OUc;); 587b3 (lCOIJI!lCXl btt8ul!tCXl opposed to
tupcxvvtlCCXt ElttSUI!tCXt). For ~l!Epoc; see e.g. 416c2 (rulers educated so as to be
~liE POl towards their subjects); 486b11 (wuxit OtJCcxicx tE JCcxl. ~11epoc; opposed to
OucrJCotvcovTltO<; JCCXt ciypicx); 57lc4 (A.oytcrttlCOV JCCXt ~l!Epov lCCXt iiPXOV); 588c-591 b.
25 Some of the dramatic and literary aspects of the Laches, Protagoras and
Symposium are discussed in Stokes (1986). On the Lysias see Hoerber (1945).
Woodruff (1982) contains a useful analysis of the literary aspects of the Hippias

OBSERVATIONS ON PLATO'S THRASYMACHUS

53

If we now return to Rep. I, we may note that in a number of respects the dramatic picture which Plato draws of Thrasymachus is
reminiscent of other scenes in Platonic dialogues which feature sophists. This goes for Thrasymachus' criticisms of and impatience
with Socrates' dialectical procedure,26 for the way in which Socrates' rational elenchus is opposed to the much more self-assertive
form of reasoning which is typical of the rhetorical practice of the
sophists, i.e. the set speech, 27 for Thrasymachus' scoffing approach,
especially the way he laughs at Socrates,2 8 and for his sullen and
insincere resignation at the end of the conversation of Rep I. 29 But
there is more to it than just the familiar opposition between Socratic
philosophy and sophistical rhetoric. Plato's portrayal of Thrasymachus contains a number of interesting traits which, I believe,
are meant to present this persona as in some respects being himself
an exemplum of the philosophical position he is made to defend.
At 336b-338b Plato starts out with a vivid sketch of Thrasymachus' behaviour during the conversation between Socrates and
Polemarchus, and of the way he interrupts the conversation. He is
pictured as a man who is conspicuously lacking in self-restraint:
he is only prevented from intruding into the conversation between
Socrates and Polemarchus by the others (u1to 'tO>V 1tapaKa8TJJ..lEVrov
OtEKCOA:UE'to) and as soon as he gets the chance he leaps in like a
wild animal, ready to tear his opponents apart ( W<mEp 8TJplOV ... W<;
Otap1taooJ..lEVo<;, 336b2-3). Socrates indirectly compares him to a wolf
in 336 d6-7 ('if I had not seen him before he saw me, I really
believe I should have been struck dumb'), 30 and to a lion in 34lcl
Maior (suggesting, among other things, that Hippias exemplifies a certain
outlook on the theme of the kalon); for a different literary analysis of the
same work, see Ludlam (1991). A most valuable guide in these matters is now
Rutherford (1995).
2 6 Cf. Critias in the Charmides, esp. 165e, and the indignant outbursts of
Polus and Callicles in the Gorgias: 467b10; 473a1, 480e (Polus); and 490d1;
491al; 494 e7; 497b6 (Callicles).
27 Cf. Gorg. 4491>-c; see also Prt. 334c-335a. See also Nicholson (1974) 230.
Thrasymachus himself delivers a kind of set speach of considerable length,
pooring out a flood of words over his company (344d: romtEp paA.aveuc; TtllOOV
KamvtA.ftaac; Kat tiilv ciltrov b.Spoov Kat noA.uv tov Alryov), after which he has to be
prevented by the others from leaving the stage right away.
2B Cf. Polus in the Gorgias. On this 'rhetorical' use of laughter, see
Rutherford (1995) 155; De Vries (1985) 378-381.
29 Cf. the way Callicles is made to behave at the end of his exchange with
Socrates in the Gorgias (esp. 510a1; 513e1; 516b).
30 On this reference to an element of popular superstition, see Adam
(1969) 1. 24.

54

KEIMPE ALGRA

('do you think me foolish enough to beard a lion and to try to outwit Thrasymachus?'). I believe these descriptions and comparisons
are there on purpose. Later on in the Republic the tyrant, after all
Thrasymachus' ideal example of unbridled pleonexia, is compared
to a wolf (565d-566a), and in general Plato is wont to use the
irrational ferocity of animals as a metaphor for the irrational parts
of the soul (the appetitive element drags us romtp SrJpiov (439b4) ,31
the tripartite soul is depicted as a triad of man, lion and manyheaded beast (588b-59la). 3 2 And like Thrasymachus, the raging
bestial tendencies of the lower part of the soul also have to be kept
in check from outside (i.e. by the rational part).33 Thrasymachus,
in other words is depicted as himself being in some ways the
embodiment of the unlimited pleonexia (as against rational
moderation) which he is made to defend.
This general characterization is further supported by what
Thrasymachus is actually made to say. He attacks Socrates for
merely asking questions in his customary way instead of coming
up himself with anything like a detailed account of justice-i.e. for
being not as self-assertive as Thrasymachus himself is-and
makes clear that he regards Socrates and Polemarchus as a couple
of wets who are only too ready to give way to each other (336c: 'tt
E'l>11Hism8 7tpo<; aA.A.fJA.ou<; u7toKa'taKAtVoi1Vot il11'iv au'tO'i<;;). 34 He
proceeds to set the rules for the conversation which is to follow:
Socrates should know that a number of potential definitions of
justice are not allowed (336c6: Kat o1tro<; 110111il Ep'i<; on 'tO 8ov EG'ttV

11118, O'tl 'tO roq>EA1110V 11118, O'tl 'tO AU(J1 'tAOUV 11118, O'tl 'tO Kp8aAov
11118' on 'tO cru11cppov). Socrates' reaction in its turn reinforces the
reader's impression of Thrasymachus as an aggressive character
(336b6: f::yw 't Kat 6 IIoA.11apxo<; 8dcrav't<; 8t7t'tofJ81111V, 336d5: f.yro
aKoucra<; f.~mAaYllv Kat 7tpocr~AE7tOlV au'tov Eq>O~OU1111V, and 336e2,
with what may be a pun on Thrasymachus' name: "Q 8pacru11ax,
3 1 See also 430b8, 535e4, 571c5. At 4llel people whose mind has not been
proferly cultivated by education are also said to be CO<J7tEp !hJptOV.
3
One may also compare the description of the common people, the demos
with their irrational desires, as 'the great beast' (Rep. 493b-c), and the lionlike strong men of which Callicles speaks in Gorg. 483e. See on this Rutherford (1995) 161, who provides an excellent discussion of the animal metaphors.
33 For the use of KroA.unv in this connection, see 439c6-9; and 589b3 (tO. o
&ypux a7toKroA.urov <pumem).
34 For the meaning of Eul](lil;ecree and imoKa.ta.KAtvOJlEVot see Adam (1969) 1.
23 and 24.

OBSERVATIONS ON PLATO'S THRASYMACHUS

55

llTt xaA.mo<; llfltV tcr8t). 35 According to Gigon this portrayal of

Thrasymachus and of Socrates' reaction is 'beinahe peinlich': 'Der
Dichter Platon hat hier, so mochten wir wohl meinen, des Guten
ein wenig zuviel getan'. 36 This may be true if we judge the passage
merely from the point of view of naturalistic aesthetics. But such a
judgement should at least be qualified once we recognize that
Plato's dramatic characterization of Thrasymachus serves a
philosophical purpose and that Thrasymachus' behaviour, at least
at the beginning of the discussion, parallels the philosophical
position which he is made to defend.37
5. The picture which I have been drawing thus far will call for
further clarification, precisely in so far as it presupposes that
Plato presents Thrasymachus' position as a clearly recognizable
and basically coherent (though of course ultimately untenable)
counterposition to his own. Mter all it has often been assumed that
Thrasymachus does not present a coherent position at all, or at least
that his two definitions are strictly speaking incompatible. Such
'incompatibilist' interpretations usually focus exclusively on the
definitions and on what may be inferred from them (i.e. they
study the definitions in isolation from their context), and they
claim that these definitions can be written out as two basically
unconnected items, viz. :38
(a) justice is the advantage of the stronger/ruler;
and

(b) justice is another one's good.

35 We know from Aristotle (Rhet. 1400 b 20) that the Greeks were
certainly sensitive to the connotations of this name; on the possibility that
Plato is here making a word-play, see Adam (1969) 24.
36 Gigon (1976) 48.
3 7 We may note that later on his aggressive behaviour loses some of its
edge as his self-confidence apparently diminishes (in 350d he is presented as
sweating and blushing). Gigon (1976) 82 finds this strange ('Die Wandlung
vom furchterregenden Wtiterich zum herablassenden Weltmanne bleibt
etwas befremdlich'), but Plato's point is presumably that Thrasymachus' blunt
self-assertion is unable to stand up to the rationality of the Socratic elenchus.
We may perhaps even go so far as to see the development of the persona of
Thrasymachus from mere aggression, via sullen resignation (at the end of
book I), to a kind of friendly cooperation (later on in the Republic, cf. 545a,
590d) as exemplifying the 'domestication' of the irrational faculties under the
guidance of reason.
38 An extreme example of such a purely 'analytical' procedure is Reeve
(1988) 16-19.

56

KEIMPE ALGRA

Thus presented, and taken at face value, the definitions would

certainly appear to add up to an incoherent position. Take, for
instance, the example of the typically Thrasymachean ruler who
rules exclusively in his own interest. According to definition (a)
he will be just (for he pursues his own interest). According to (b)
he will be be unjust (for he ignores the good of others) .39
An interpretation which tries to square the two definitions has
been proposed by George Kerferd, and has been taken over by various other scholars. According to Kerferd (b) is entirely consistent
with (a), provided one thing is accepted, namely that the statement
that justice is to seek the interest of the ruler was made by looking
at the matter from one viewpoint only, namely the viewpoint of the
ruled. 40 I believe that this interpretation goes a long way into the
right direction, but requires further qualification on some points. 41
In particular, like most other interpretations it still presupposes that
we are indeed dealing with two fully independent definitions, given
in entirely different contexts. This is why Kerferd can claim that
definition (a) is in fact incomplete and occurs in a context which
(unlike the context of (b)) discusses the matter at issue only from a
partial perspective. Kerferd accordingly adds that 'once we ask
what about justice from the viewpoint of the ruler, we can no
longer say that justice consists in seeking the interest of the ruler,
we must say that it consists in seeking the interest of the ruled'. 42
However, if we really take the context into account and turn to
the way Plato actually presents these definitions, we are in for a
surprise. Plato has Thrasymachus actually combine the two defini39 Incompatibilist interpretations along these lines have been provided by
Guthrie (1971), vol. 3, 94; and Gigon (1976) 60. Other interpretations of Thrasymachus' position (as presented in his definitions) as basically incoherent
and confused Oowett, Sidgwick, Cross and Woozley) are summarized by
Nicholson (1974) 213. According to Maguire (1971) the inconsistency is basically due to Plato, who distorts Thrasymachus' original position by adding (b);
on which see above, n. 8.
40 Cf. Kerferd (1947) 560-561; and (1981) 121-122. An interpretation rather
similar to Kerferd's has been defended by Nicholson (1974). Annas (1981) 4546 takes a middle position between 'incompatibilists' like Guthrie (see
previous note) and the view defended by Kerferd by arguing both that the two
definitions are formally in conflict, and that they reflect the same spirit.
According to Annas, 46, Thrasymachus is being 'driven from a partial and
muddled to a clear formulation'
41 Apart from qualifying remarks which I have made in the text, I would
like to note that I see no justification for the thesis of Kerferd (1947) 561 that
Thrasymachus' injustice is a moral obligation.
42 Kerferd (1981) 122.

OBSERVATIONS ON PLATO'S THRASYMACHUS

57

tions in the course of his speech: according to Thrasymachus

Socrates does not know that justice' is 'in truth another one's good,
the advantage of the stronger and the ruler' (aA.M'tptav aya.Sov 'til>
ovn, 'tOU KpEt't'tov6<; 't Kat &pxov'to<; crt>J.uppov, 343c3-4). And when
in the second book Adeimantus refers back to Thrasymachus'
position he simply follows in Thrasymachus' footsteps and connects the two definitions: 'Or should we agree with Thrasymachus
that justice is another one's good and the advantage of the stronger'
(aA.AO'tptOV aya.86v, O"UJ.Uptpov 'tOU KpEt't'tOVO<;, 367c). Two things are
worth noting. First, neither here nor in the first book is there any
suggestion that the two definitions are incompatible and that
Thrasymachus' position is internally incoherent, although it is of
course suggested that this position is wrong. 43 Secondly, these
passages show that the two definitions are not independent, as was
implicitly supposed by Kerferd, but that they apparently belong
together as two sides of the same coin, as two ways of saying the
same thing: (a) is so to speak placed in apposition to (b) as an
explanatory tag (aAAO'tptOV aya.86v, O"t>J.Uptpov 'tOU KpEt't'tOVO<;, 367c).
This does not suggest that (a) is written from a more limited
perspective, or that the cases described by (a) are only a subclass of
the cases described by (b). 44 Both definitions apply to everybody,
and they are meant to describe reality as it is, not to cover every
imaginable situation (including the situation where a ruler might
be just). Now as things are the fact that justice (at the level of the
43 It is true that in 345b7 ff. Socrates asks Thrasymachus to stand by his
own words from now on, or at least, if he does change his position, to do so
openly. This passage has been adduced by Annas (1981) 46-47, in support of
her view that the two definitions are according to Socrates/Plato formally
incompatible and that Thrasymachus is driven from the confused position
expressed by the first definition to the clearer view expressed by the second.
Yet, as becomes clear from Socrates' immediately following words, he is not
referring to Thrasymachus' having presented different definitions but to the
fact that after having introduced the notion of the physician 'in the strict
sense' he has not in a like manner discussed the shepherd in the strict sense
(345b9<3: vuv o op~ ...
1:ov ~ w..11a~ ia.1:pbv 1:0 1tpro1:ov opt~oJ.l.Evo~ 1:ov ~ w..11a~
1tOtJlEVO. OUlCEt\ cpou ot1v Utl'tEpov a!Cpt~~ <pUAU~a.t). In other words, the instability
which Socrates ascribes to Thrasymachus does not concern the latter's main
position, as crystallized in the two definitions, but the arguments by which
he has defended his main position. It will be clear from all this that I do not
see any reason to go along with Maguire (1971) 155, who claims with regard
to the two definitions that 'Plato is at pains to indicate the logical difficulties
in holding both positions simultaneously'.
44 For the contrary view, however, see Nicholson (1974) 224 and 227 and
Annas (1981) 45-46.

on

58

KEIMPE ALGRA

individual) is someone else's good, comes down (at the political

level) to justice being the advantage of the stronger/ruler. In other
words, Thrasymachus consciously and exclusively links justice to
the weak: what justice amounts to for the really strong is a question
which is not even considered. 45 For them, in actual practice, there
simply is no such thing as justice, because there is no need for
them to consider the interest of others.
This picture appears to be confirmed if we turn from the
definitions to the larger context of Thrasymachus' story and the
way it is summarized at the beginning of book II. Immediately
after the definition of justice in 343c, we are told that injustice is
'the opposite' (1:ouvav1:iov) and that it 'rules over those who are truly
simpletons and just' (leal apxu 'tO>V ro<; UA'J18ro<; EU'J18ucrov 'tE Kat
OtKairov, 343c6-7). 46 The just are simply identified with the losers
who have to obey the rules set by others. 4 7 In the paraphrase in
book II they are referred to as those who are neither able to act
injustly themselves nor to escape the injustice done to them by
others (359al). 48 Those, on the other hand, who end up on top in
4 5 This claim is not invalidated by the fact that Thrasymachus at 343e1
refers to the example of just people taking some office (apxiJv nva). The way
they get themselves into trouble shows that they do not qualify for the position
of 'real', i.e. strong, rulers (cf. 343b5: o'l roc; aA.Tt9iilc; llpxouow).
46 This phrase, and indeed the whole context of Thrasymachus' speech
may be taken to invalidate the rather relativistic interpretation of Nettleship
(1964) 32, which claims that 'if the doer of unjust things is strong enough,
then what he does is called justice by weaker men'. A similar view has now
been defended by Reeve (1988) 19: 'By breaking all the laws, the tyrant reveals
himself as completely unjust. By simultaneously overthrowing them, and the
power behind them, he destroys the standard by reference to which he is
unjust. Justice is now his advantage'. On the basis of the observations which I
have here made, we may object that although the ruler determines what may
count as just, his own unjust deeds and commands are and remain unjust
from each and every perspective. It is only the obeying of the ruled to his
unjust commands - not those commands themselves - that is to be called just.
47 See also 348c-d, where Thrasymachus refers to justice as yevvaia euiJ9na
and where the successfully injust (o'i YE tEMCO~ [... ] otoi tE aoucEiv' 1tOAEt~ tE Kat 9vTt
OUVUJ.lEVot av9pclmmv ucp' ECXUtOU~ 7t0tEtcr9at) are called cpp6vtJ.Lot and aya9oi.
48 It is to be noted, however, that in Glaucon's paraphrase the perspective
has changed slightly. Pleonexia is still described as the motive force behind
human behaviour, and the best life is still described as a life of unbridled
self-assertion, but it is now argued that justice is a kind of contract by which
the weak secure, by way of a second-best solution, at least some protection
against the pleonexia of the strong. Nothing of this sort was maintained by
Thrasymachus, where it was argued that the contents of justice are completely
determined by the self-interest of the rulers. There is no justification in the
text of book I for the claim of Gigon (1976) 61 that 'der Starke wird mit ihm
(scil. dem Begriff der Gerechtigkeit) seine Starke aufs Spiel setzen, der

OBSERVATIONS ON PLATO'S THRASYMACHUS

59

the struggle for power resulting from what we might call Thrasymachus' implicit anthropology of pleonexia, 49 can do what they
want. This is why Thrasymachus ends his speech by claiming
that injustice is not only stronger and more powerful than justice,
but also more indicative of freedom (H.eu8eptro'tepov). Justice,
apparently, is a matter of compulsory self-restriction, a matter of
limits and bonds. The really strong and powerful, with the tyrant
as the most shining example, are free of such bonds. For them
justice does not exist. 5o
6. We may summarize our findings as follows. The general
picture which emerges from our analysis is that of Thrasymachus
putting forward, apparently on empirical grounds, the case for
pleonexia. This position foreshadows Plato's vivid sketch of the
quasi-freedom of the tyrant, whose passions rule in complete
lawlessness (EV
avapxig. Kat aVOJ.Ltg., 575al). It thus serves as a
counterpoint to Plato's own insistence in his moral psychology on
limit and moderation, elements which in the last analysis, and in
ways which need not concern us here, appear to be connected to
his metaphysical conception of the Good. 51 The overall dramatic
characterization of Thrasymachus as a persona confirms this
picture by presenting him as animal-like, aggressive and incontinent, in short as in some respects an embodiment of the position he
is made to defend. It appears that Thrasymachus' two definitions of
justice belong together and are both connected with his general
anthropology: justice sets limits to one's tendency towards pleonexia, and one only lets this happen if one has to. Consequently for
the really strong justice does not exist. 52

1tacrn

Utrecht
Schwache dagegen hat die Aussicht, den Starken bei ihm behaften zu
konnen'.
49 This anthropology is made explicit, however, in Glaucon's paraphrase
in book II, 359c4: ou'x Ti]v ltAeOVE~iav' 0 Jtacra lj)UO"tc; OunKEtV ltEij)UKEV roc; tiya86v.
50 We may recall that according to Thrasymachus the ruler 'in the strict
sense' makes no mistakes, i.e. he is always successful in ruthlessly pursuing
his own interests. Note the way this feature is underlined by Plato's use of the
vocabulary of power and success (tov ueyal..a Ouvauevov JtAeOVEKtE'iv, 344al; tiOtKia
tKav&c; ywvouevn, 344c5-6; Ouvacrew o aOtKElV 1\ tcp A.avSavEtv 1\ tcp OtauaxecrSat,
345a5; Ot 'YE 'tEAEwc; [... ) oto{ 'tE aOtKElV, 348d5).
51 On which see especially Rep. 500c.
52 I would like to thank David Runia and Teun Tieleman for their
comments on an earlier draft of this article.

60

KEIMPE ALGRA

References:
Adam, J. (1969) The Republic of Plato: edited with Critical Notes, Commentary
and Appendices by James Adam; second edition with an Introduction by
D. A. Rees, 2 vols. (Cambridge).
Adkins, A. W. H. (1972) Moral Values and Political Behaviour in Ancient Greece
(London).
Annas, J. (1981) An Introduction to Plato's Republic (Oxford).
Arnim, J. von (1914) Platons ]ugenddialoge und die Entstehungszeit des 'Phaidros'
(Leipzig).
--(1927) Platons Dialog 'Thrasymachos', Mededeelingen der Koninklijke Akademie van Wetenschappen, afdeeling letterkunde deel63 (Amsterdam), 201-232.
Classen, C. J. (ed.) (1976) Sophistik (Darmstadt).
Cross, R. C. I Woozley, A. D. (1964) Plato's Republic; A Philosophical
Commentary (London).
Friedlander (1964) Platon, 2 vols., 3. Auflage (Berlin).
Gigon, 0. (1976) Gegenwiirtigkeit und Utopie; eine Interpretation von Platons
'Staat'; Erster Band: Buch I-IV (Ztirich/Mtinchen).
Guthrie, W. K. C. (1975) A Histury of Greek Philosophy, 6 vols. (Cambridge).
Hoerber, R. G. (1945) 'Character Portrayal in Plato's Lysis', The Classical
]oumal41, 271-273.
Kahn, C. H. (1983) 'Drama and Dialectic in Plato's Gorgias', Oxford Studies in
Ancient Philosophy, 1: 75-121.
- - (1993) 'Proleptic Composition in the Republic, or Why Book 1 was never a
separate dialogu~. The Classical Quarterly 43: 131-142.
Kerferd, G. B. (1981) The Sophistic Movement (Cambridge).
- - (1947) 'The Doctrine of Thrasymachus in Plato's Republic', Durham
University ]oumal40; repr. in Classen (1976) 545-563.
Lebeck, A. (1971) The Oresteia: a Study in Language and Structure (Washington
DC).

Ludlam, I. (1991) Hippias Major: an Interpretation (Stuttgart).

Maguire,]. P. (1971) 'Thrasymachus ... or Plato', Phronesis 16: 142-163.
Mansfeld, J. (1996) 'Insight by Hindsight: Intentional Unclarity in Presocratic Proems', BICS, 42: 225-232 .
Muller, G. (1981) Platons Dialog vom Staat: Kunstform und Lehrgehalt, Wiesbaden.
Nettleship, R. L. (1964) Lectures on the Republic of Plato, (London); reprint of the
second edition of 1901.
Nicholson, P. P. (1974) 'Unravelling Thrasymachus' Arguments in 'The
Republic", Phronesis 19: 210-232.
Reeve, C. D. C. (1988) Philosopher-Kings; The Argument of Plato's Republic
(Princeton).
Runia, D. T., (1996) 'The literary and Philosophical Status of Timaeus'
Prooemium', forthcoming in the Proceedings of the International Plato
Society conference, Granada 1995.
Rutherford, R. B. (1995) The Art of Plato: Ten Essays in Platonic Interpretation
(London)
Stokes, M. C. (1986) Plato's Socratic Conversations (London).
Vries, G. J. de (1985) 'Laughter in Plato's Writings', Mnemosyne 38: 378-381.
White, N. P. (1979) A Companion to Plato's Republic (Indianapolis).
Wilamowitz-Moellendorff, U. von (1920) Platon, 2 vols. (Berlin).
Woodruff, P (1982) Plato: Hippias Major (Indianapolis).
Zeller, E. (1922) Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung, 5.
Auflage, Band II, 1 (Leipzig).

KOMPSOI ASKLEPIADES:
LA CRITICA DI PLATONE ALLA MEDICINA
NEL III LIBRO DELLA REPUBBUCA
MARIO VEGETTI

Jaap Mansfeld ha sostenuto a piu riprese (cfr. nota 17) che il metodo
di Ippocrate cui Platone si riferisce nel celebre passo di Phaedr.
270c-d trova in Arie acque luoghi il suo testo principale di riferimento. Un'analisi della violenta critica che Platone rivolge nel
libro III della Repubblica alia medicina dietetica contemporanea
non puo portare conferme o smentite dirette a questa tesi, e quindi
contribuire alia soluzione della vexatissima 'questione ippocratica'.
La polemica platonica costituisce tuttavia un documento significativo degli sviluppi della medicina nel IV secolo, e della loro
valutazione nell'ambito della discussione etica e politica. Da questo
punto di vista, qualche indicazione sia pure indiretta su quella
questione puo venir ricavata dalla Repubblica: si tratta, da un lato, del
rilevante impiego platonico del linguaggio diagnostico e psicoetnologico di Arie acque luoghi, dall'altro del principale bersaglio
della polemica del libro III, che, attraverso Erodico, sembra colpire
soprattutto il Regime. Di questi dati andni forse tenuto conto nell'interpretazione della 'testimonianza' del Fedro, e sembra plausibile
pensare che essi possano almeno marginalmente confortare Ia
lettura propostane da Mansfeld.
1. II passaggio dalla prima citta (definita <<Sana, uytfl~. 372e7)'
alia seconda, che <<vive nel lusso>>, le ragioni e le forme della
patologia che affiigge quest'ultima, infine le vie possibili per il suo
risanamento, vengono discussi facendo largamente ricorso aile
categorie diagnostiche e terapeutiche proprie della medicina dietetica del V e IV secolo.
La polis lussuosa soffre di gonfiore da infiammazione (q>AqJl<ltvoucra, 372e8). Non si tratta soltanto dell'impiego metaforico di
termini medici in ambito morale, che e consueto in Platone.
L' eziologia di questa malattia endemica e indicata da Platone nella
cattiva dieta e nella corrispondente mancanza di esercizi (apyia Kat
Otat'ta, 405d); quanto al primo aspetto, si tratta soprattutto della dieta

62

MARIO VEGETTI

carnea, introdotta nella 1tOAt~ tpu<p&cra 1 (373b-c). Ora, questa nesso

diagnostico ed eziologico e diffuso nella medicina dietetica, a
partire dai piu antichi Luoghi nell'uomo (L.H.) e Arie acque luoghi
(A.A.L.) fino al piu recente Regime e allo stesso Erodico, principale
bersaglio della critica platonica. II termine <pAqJlaivetv ricorre con
grande frequenza in L.H.; per il gonfiore da infiammazione si
vedano soprattutto 13.1 e 29.1. II Regime nelle malattie acute (37) menziona l'alimentazione carnea (KpETJ<payiTJ)2 tra le cause di flatulenze
( <pucrat: come si vedra piu avanti, questa termine gioca un ruolo
centrale nella pole mica platonica).
L'eccessivo riempimento di cibo (7tATJOJlOVft), non compensato da
adeguati esercizi (7t6vot) produce, secondo il Regime, tanto JCatappot
(3.70.1: anche questa termine e impiegato polemicamente da Platone), quanta <pucrat (3.74.1). Era questa, a quanta attesta l'Anonimo
Londinense, anche Ia dottrina di Erodico: le malattie nascono
dall'eccesso di alimentazione non compensato da 7tOVOt sufficienti
(9.20 sgg.).
La medicina dietetica offre dunque a Platone gli strumenti per Ia
diagnosi di una nialattia sociale che e al tempo stesso morale, per
Ia dedizione intemperante ai piaceri del lusso, ed organica, per i
danni che questi piaceri producono alia salute pubblica.
Ma essa offre anche le indicazioni necessarie alla sua terapia.
Platone insiste sulla necessita di <<purificare>> Ia polis (Ka9aipnv,
OtaJCa9aipetv, 399e): egli gioca indubbiamente anche sulla valenza
religiosa del termine (c'e un miasma collettivo da scongiurare), rna
e chiaro che in questa contesto dietetico il riferimento principale e
quello aile purghe dimagranti dei medici. In questa sensa il
termine e frequentissimo in L.H. (cfr. soprattutto 13.4, 28.1), e
ampiamente usato nel Regime (2.40-85). La <<purga>> morale e intellettuale che Platone somministra alia polis tryphosa mira a mettere
sotto controllo l'abuso dei piaceri del lusso. Ad essa fa seguito Ia
costruzione di un ambiente in cui sia possibile il risanamento e Ia
1 Questo termine non appartiene a! linguaggio della medicina. II participio con valore aggettivale si trova, in questo stesso senso, in Aristoph. Nub. v.
48. In Xen. Mem. 1.6.10 Socrate accusava Antifonte di identificare Ia feliciti
con Ia tpuq>'ft. Lo stesso Antifonte sembra aver elaborato una teoria fisiopatologica della <pAEYilOVft (DK B29a).
2 Ne qui ne altrove nei testi medici c'e tuttavia una condanna dietetica
dell'alimentazione carnea (anche Ia focaccia d'orzo, per es., puo causare
physai). Le pro prieta alimentari dei diversi tipi di carni sono ampiamente
discusse nel Regime, 2.46-49.

LA CRITICA DI PLATONE ALLA MEDICINA

63

formazione armoniosa del ceto politico-militare cui verra affidata Ia

guida della nuova citta. Questi giovani dovranno crescere, dice
Platone, <<in un luogo salubre>> (ev uytEtVql 't01tql), in senso morale,
intellettuale e anche estetico, da cui siano positivamente influenzati
come da <<un'aura che reca salute, proveniendo da luoghi benefici>> (a:upa cppoucra U1t0 XPllO''tmV 't01t(l}V uyietav' 401 c).
Anche a questo proposito, non e difficile individuare i riferimenti medici della terapia platonica. II termine aupa, raro in prosa,
compare in A.A.L. 6. Ed e ben noto come questa stessa opera dedichi
una grande attenzione all'influenza dei venti sulla salute psicofisica delle popolazioni (cfr. soprattutto cap. 5). Piu in generale, e
altrettanto noto che l'autore 'ippocratico' ritiene essenziale il ruolo
che i nomoi, cioe un huon assetto costituzionale, possono giocare, in
aggiunta ai fattori ambientali, riguardo ai costumi (11911) e alia
dnvuxia di individui e popoli (capp. 16, 23).
Gia da quest'ultimo punto di vista, il rapporto fra la terapia
platonica della polis e il pensiero medico-politico di Arie acque luoghi
esce da un contesto solo metaforico, com'era quello fra l'aura
educativa e i venti salubri del trattato medico. Ci sono tuttavia altri
riscontri, piu marginali rna proprio per questo forse piu significativi. I suoi <<atleti della guerra>>, scrive Platone, saranno esposti
durante le loro campagne a frequenti <<mutamenti>> (JlE'tapoA.ai)
nella dieta, nel clima e nelle stagioni, e al tempo stesso dovranno
risultare <<insonni>> ( &ypu7tvot), acuti nella vista e nell 'udito e di
salute costante ( 404a-b). Platone aggiunge altrove che il carattere
psicologico dominante di questi guerrieri sara l'impulsivita
animosa, lo spirito 9uJlOEt0E<;, che a seconda dell'educazione puo
volgere verso Ia selvatichezza (1:0 &yptov) collerica o verso una
benevola mitezza UlJlEPO'tll<;) (410d-e).
Questo intero complesso concettuale e linguistico relativo ai
rapporti fra condizioni ambientali e caratteri psicologici ha il suo
unico e puntuale precedente proprio nelle analisi di Arie acque
luoghi. Qui e sistematicamente discussa !'influenza delle metabolai
stagionali, e delle EK7tAE~Et<; psico-somatiche che ne conseguono,
sulla condizione morale e psicologica dei diversi popoli. Gli Asiatid, il cui clima e esente da bruschi mutamenti, mancano percio
di coraggio, di capacita di sopportare gli sforzi, di thymoeides (il
termine none altrove attestato prima di Platone) (12: 'tO OE avOpEtOV

lCO.t 'tO 'tO.Aat1tropov lCO.t 'tO EJl1tOVOV lCO.t 'tO 9uJlOEt0E<; OUlC av ouvat'tO ev
'tota.U't'[l cpucrn eyyivecr9at).

64

MARIO VEGETTI

Essi sono inoltre pili miti (TtJlEpron:pot) degli Europei, perche non
esposti alle metabolai e alle conseguenti ekplexeis che inselvatichiscono il carattere, eccitano lo thymoeides, tengono desta la mente come accade appunto peri popoli d'Europa (16: 't11v oprf1v ayptoucr8ai 'tE
Kat 'tOU ayvcOJlOVoc; Kat 8UJlOEtOeoc; JlE'tEXEtV (... ] ai. yap JlE'ta~oA.a{ Eicrt
't&v 7tUV't<OV ai. E7tEydpoucrat 'tltV yYcOJlllV). La frequenza e la violenza
dei mutamenti rende al contrario questi uomini selvatici, impulsivi, coraggiosi, disposti agli sforzi (23: 'tO 'tE &yptov Kat 'tO UJlEtK'tOV
Kat 'tO SuJlOEtOec; v 'tOtaU't!l q>UO"Et yyivE'tat [... ] a1to 'tllc; 'taA.amcopillc;
Kat 'tiDV 1tOV<OV ai. avOpEtat), duri, tonici, insonni, indipendenti di
giudizio, acuti nell' esercizio delle tecniche, migliori combattenti
(24: crKATlpouc; 'tE Kat icrxvouc; [... ] Kat ev'tovouc; [... ]. 'tO 'tE pyanKov
EVEOV v 'tft q>UO"Et 'tft 'totaU't'{l Kat 'tO &ypu1tVov [... ] Kat iowyvroJlovac;,
'tOU 'tE aypiou JlUAAov JlE'tEXOV'tac; f\ 'tOU TtJlEpou, ec; 'tE 'tac; 'texvac;
o~u'tepouc; 'tE Kat O"UVE'tco'tepouc; Kat 'tCx 7tOAEJlta aJlEtvouc; EupftcrEtc;).
Esattamente il quadro dei caratteri, dunque, che Platone richiede
ai suoi <<atleti della guerra>>, preoccupandosi per altro di temperarne l'asprezza <<europea>> con il ricorso all'addolcimento che puo
venire da una buona educazione musicale.
La diagnosi della patologia sociale tracciata nei libri II e III della
Repubblica, la sua eziologia, Ia sua terapia, sembrano cosi largamente ispirate ai testi della medicina dietetica; in particolare, il
rapporto fra il linguaggio psicologico di Platone con quello di Arie
acque luoghi e cosi stretto da rendere improbabile l'ipotesi di una
coincidenza casuale.3 Abbiamo qui, come altrove, un Platone
attento lettore dei testi di medicina, e pronto a fruire della loro concettualizzazione metodica tanto come riferimento metaforico
quanto come strumento analitico.
2. Tanto pili sorprendente risulta il violento attacco alla medicina
contemporanea a base dietetica che Platone Iancia a partire da 405a.
11 punto di partenza e il nesso che viene stabilito fra patologia
sociale e sviluppo della medicina4 (parallelo a quello, in ambito
3 Si puo inoltre rilevare un'eco della critica ai sacrifici di A.A.L. 22 nel
discorso di Adimanto (II 365e sgg.).
4 Puo essere interessante notare che Ia denuncia di Platone non e soltanto
basata su di un'indignazione moralistica. C'e stato in effetti nel IV secolo un
peggioramento della salute pubblica in Grecia dovuto all'abbandono dell'agricoltura, con il conseguente mancato drenaggio delle paludi, e alia sovrappopolazione urbana. Ne consegui un aumento delle malattie malariche, delle
gastro-enteriti, del tifo e della tubercolosi. L'eta media al momento della

LA CRITICA DI PLATONE ALLA MEDICINA

65

morale, tra decadenza dei costumi ed espansione della pratica

giuridica). L'abuso intemperante dei piaceri alimentari determina
una rottura del quadro nosologico tradizionale, che includeva
secondo Platone soltanto lesioni traumatiche (come le ferite sui
campi di battaglia) e malattie a decorrenza stagionale (e7ten:ux
voa~J.UX'ta: qui e ancora chiara }'influenza di Arie acque luoghi 2 e di
Epidemie I-III) (405c).
La diffusione di nuove malattie interne determina la proliferazione dei medici e dei loro ambulatori, che costituiscono
dunque un effetto e un sintomo della malattia morale della societa,
rna che, come vedremo, finiscono per diventarne una delle cause.
In queste circostanze, la medicina, come l'avvocatura, viene ad
assumere un'indebita 'aria di distinzione' (aeJ.Lvuvov'tat), motivata
anche dal fatto che molti uomini liberi>> vi si dedichino con
impegno professionale, cosa che Platone sembra trovare sorprendente (405a).
Questa sorpresa va interpretata come un segnale del crescente
umore polemico di Platone; mai altrove, nei numerosissimi riferimenti alla medicina presenti neUe sue opere (e del resto anche nel
I libro della Repubblica) egli ha messo in dubbio che la techne
medica sia degna di uomini liberi. E c'e soprattutto il celebre passo
delle Leggi in cui la medicina praticata dai liberi viene contrapposta
a quella degli schiavi come modello per l'azione del huon legislatore nella sua opposizione rispetto al tiranno (IV 720a-e).
11 nuovo quadro patologico, causato dalla dieta malsana e dalla
inattivita fisica proprie della polis lussuosa, comprende peUJ.La'ta e
7tVEUJ.la'ta, flussi>> e <<arie>> che ribollono nei corpi come in paludi
limacciose ( 405d). Si tratta di termini medici diffusissimi, che
indicano processi morbosi relativi (in linguaggio moderno) all' apparato respiratorio e soprattutto a quello digestivo. Di fronte a queste
novita nella patologia, <<i raffinati Asclepiadi>> (KOJ.l'lfOt 'AaKAT)7ttaOe<;, 405d) sono chiamati a un lavoro di precisazione linguistica e
concettuale delloro lessico fisio-patologico, vista l'inadeguatezza di
quello tradizionale. Due osservazioni sono qui necessarie. KoJ.L'IfOt
ha certamente una sfumatura polemica, indicando un eccesso di
morte per gli uomini passo dai 45 anni in epoca classica ai 42,4 nel IV e III
secolo. Si vedano i dati raccolti in M. Grmek, Les maladies a l'aube de la
civilisation occidentale (Paris 1983) 32, 141-42, 157-58. E dunque vero che lo
sviluppo della medicina scientifica coincide in Grecia con il declino delle
condizioni generali di salute.

66

MARIO VEGETTI

raffinatezza che diventa 'sofisticazione'5 (e infatti il principale

bersaglio dell'attacco, Erodico, e definito Sofista a nessuno inferiore>> insieme con un altro dietetico, Icco di Taranto, in Prot.
316e).
In secondo luogo: quando Platone parla di Asclepiadi, non
intende sottolineare la discendenza della professione medica dal
clio guaritore (alla cui tradizione essa viene anzi poco dopo esplicitamente contrapposta), bensi designare la corporazione ( Kotv6v) dei
medici, ben nota nella vita pubblica dell'Atene dei suoi tempi. 6 Si
trattava di un gruppo numeroso di professionisti colti, socialmente
prestigiosi e ben visibili, l'appartenenza al quale segnalava l'affermazione e la competenza del singolo medico. Quando Platone
cita lo stesso Ippocrate egli non manca di qualificarlo come Asclepiade (Prot. 31lb, Phaedr. 270c), per rafforzarne l'autorevolezza. C'e
dunque un'intenzionale ironia polemica quando Glaucone, approvato da Socrate, definisce a loro volta 1tavu KOJ. l\jfOt, davvero
raffinati>> ( 408b), i figli di Asclepio, la cui medicina tradizionale e
stata contrapposta a quella moderna dei 'sofisticati' Asclepiadi.
Per tornare appunto alle loro innovazioni terminologiche, esse
vengono esemplificate da Platone con le parole cpucrat, che specializza 7tVEUJ..La-ra, e Ka-rappot, che tecnicizza pEUJ..La'ta. 7 Flatulenze>> e
<<Catarri>>: termini che Glaucone qualifica prontamente come
<<nuovi e assurdi>> (Katva Kat a-ro1ta, 405d) .8
Non tanto <<nuovi>>, si puo osservare, al tempo in cui Platone
scriveva Ia Repubblica ( ed e improbabile che egli pensasse qui alla

5 KOJ.l'lfO'tEpot sono per es. definiti in Theaet. 156a i seguaci di Protagora

sostenitori di una gnoseologia empiristica. Cfr.. anche Phil. 53c.
6 Sui koinon dei medici ateniesi, cfr. S. Aleshire, The Athenian Asklepieion
(Amsterdam 1989); V. Nutton, "The medical Meeting place", in Ancient
Medicine in its socio-cultural context, eds . PhJ. Van der Eijk et al.. (Amsterdam
1995) 1.3-25 (4). A Delfi era riconosciuto e privilegiato un koinon degli Asclepiadi di Cos e di Cnido: cfr. J.. Bousquet, "Delphes et les Asclepiades", Bulletin
de Correspondance hellinique, 80 (1956) 579-93 .
7 0. Gigon, Gegenwiirtigkeit und Utopie (Ziirich-Miinchen 1976) 309 sgg.,
nota che questi termini, che definiscono le Zivilisations-krankheiten,
compaiono per Ia prima volta in Platone a! di fuori della letteratura medica
(rivestendo cosi un netto valore citazionale).
8 Antica medicina polemizza a sua volta contro gli inventori di una K:<XtVTJ
u7t69Ecrt~: il termine vale di nuova invenzione, e anche straniero (cfr.
Festugiere ad loc.). Puo quindi darsi che Platone rivolga deliberatamente
contro i medici il !oro stesso linguaggio polemico. Ma lo stesso Socrate
(Eutyphr. 3b) definisce cl't07t<X le accuse che gli vengono rivolte di introdurre
dei 'nuovi' e 'stranieri' (K:<Xtvou~).

LA CRITICA DI PLATONE ALLA MEDICINA

67

sua 'data drammatica'). Essi compaiono infatti in testi che a

Platone, come si e visto, dovevano essere ben noti e ormai vecchi di
qualche decennio. Ka'tappoo<; e usato in un gruppo di scritti piuttosto
vicini tra loro, come Luoghi nell'uomo (29.2), che offre inoltre
un'elaborata teoria dei p6ot, <<flussi>> (9-10), Male sacro (6, 7, 8-dove
e identificato appunto a pEUJ.l<X, 13-dove e discusso il rapporto fra
7tVEl>J.La'ta, <<Venti>>, e x:a'tappot che scendono dal cervello), e infine
in Arie acque luoghi (3-dove si parla di <pAE'YJ.l<X EmKa'tappeov, 10).
Quanto a <pucra, il termine compare in Luoghi nell'uomo ( 45.3),
Antica medicina (10, 22-dove viene tracciata una sorta di anatomofisiologia delle physai), Regime nelle malattie acute (23, 37).
Del resto, secondo la pur discutibile testimonianza dell'aristotelico Menone riassunta nell'Anonimo Londinense, il concetto di
<pucra occupava un ruolo centrale nell'eziologia dello stesso lppocrate (5.35 sgg.).9
Nulla di particolarmente 'nuovo', dunque, nell'uso di questi
termini in quanto tali. Ma si possono forse indicare alcuni testi,
questi si nuovi al tempo in cui Platone scriveva Ia Repubblica, e forse
anche dal suo punto di vista <<assurdi, che documentano il lavoro
linguistico degli Asclepiadi e le sue conseguenze in termini
medici: essi costituiscono, con ragionevole margine di probabilita,
il termine immediato di riferimento dell'aggressione polemica di
questi passi della Repubblica. Dal primo punto di vista, sono senza
dubbio significative le precisazioni del capitolo 3 dei Venti, un'opera
appunto dedicata al ruolo eziologico delle physai, che tracciano
nell'ambito dei pneumata una distinzione fra l'aria esterna, aer, e
quelle interne, appunto le physai: 7tVE'I>J.L<X'ta 'tfx J.LEV EV 'tOtcrt O"cOJ.l<XO"t
<pUO"<Xl K<XAEOV't<Xl, 'tfx o E~(J) 'tO>V crroJ.La'troV aftp.
Lo 7tVEUJ.l<X esterno, secondo questo autore, entra nel corpo
insieme con i cibi, e Ia relativa 7tATJO"J.LOVTt determina le flatulenze
interne (7); le <pucrat so no a Ioro volta causa dei pEUJ.l<X't<X organici.
C' e dunque in questa recente opera della medicina dietetica una
9 A prima vista, questa testimonianza sembra riferirsi allo scritto sui Venti.
Qui, tuttavia, le physai sono legate all'immissione di aria esterna, mentre
Menone, aristotelicamente, le mette in rapporto con i residui (neptcrcrrolla'ta)
dei cibi ingeriti: cfr. G.E.R. Lloyd, La questione ippocratica, in Id., Metodi e problemi della scienza g;reca (trad. it. Roma-Bari 1993) 348-50. Per quanto Ia testimonianza sia inattendibile, e l'Anonimo sembri aver riassunto in modo impreciso il testo di Menone, essa documenta comunque l'esistenza di un'autorevole dottrina eziologica delle physai tra Ia fine del V e l'inizio del IV
secolo.

68

MARIO VEGETTI

precisa traccia del lavoro di sistemazione linguistica alia quale si

riferisce polemicamente Platone.
II nodo concettuale e terminologico che egli ha di mira nel
nostro passo della Repubblica non e tuttavia reperibile tanto nei Venti,
quanto nel Regime, dove credo esso possa venir rintracciato nella
sua completezza.
Anche quest' opera distingue tra 1tVEUJ.l<X, respiro dell'aria esterna,
e <pucra, flatulenza interna, che si connettono nel processo digestivo
(2.40.2). Ma e soprattutto importante Ia sequenza che viene istituita
fra prevalenza dei cibi sui 1tOVot, e conseguente 1tAl10'J.lOVll, e formazione di Ka'tappot (3.70.1), da un lato, dall'altro di <pucrat, sempre
dovute alia 1tAl10'J.lOVll (3.4.1). Qui sembra pienamente articolato quel
rapporto fra eccesso dietetico, mancanza di esercizio fisico, e
formazione di <pucrat e Ka'tappot, nell'ambito dei 1tVtllJ.l<X'ta e dei
ptllJ.l<X'ta, che Platone individuava nei recenti e discutibili sviluppi
della medicina dietetica.
E che il Regime possa venir considerato il pili diretto bersaglio
della polemica platonica 10 sembra confermato da altre considerazioni, relative alia collocazione sociale della medicina. Platone,
come vedremo, vorrebbe che Ia terapia dei ricchi venisse assimilata a quella dei poveri e degli artigiani, 11 con il comune obbiettivo
di ripristinarne il pili rapidamente possibile Ia capacita di erogare
prestazioni socialmente utili, lavorative nel caso dei secondi, politiche in quello dei primi (406d-407a).
L'autore del Regime, al contrario, sostiene che Ia sua oiat'ta (invero J.l<XKpa Otat'ta, per dirla con Platone in 406d3-4), none fruibile
dalla massa (1tA.T18o~) dei poveri, che devono per forza guadagnarsi
Ia vita ( E~ aVcX"fKTl~ Ota'tEAEtV 'tOV ~iov, e cfr. in Platone 406c: epyov 0

10 Diventa da questo punto di vista rilevante il problema della collocazione

cronologica di Regime. R. Joly, Recherches sur le traite pseudo-hippocratique du
Regime (Paris 1960) 203 sgg., lo data intorno al 400 (cfr. anche Ia sua introduzione all'ediz. Bude, 1967).]. Mansfeld, The Pseudo-Hippocratic Tract peri
hebdomadon ch. 1-11 and Greek Philosophy (Assen 1971) 25 n. 116, 105 n. 195,
considera il trattato precedente il Timeo, rna lo colloca dopo il 370 in base al
calendario implicato in 3.68. Neppure questa datazione piu bassa escluderebbe
Ia possibilita che Platone abbia conosciuto i1 Regime durante Ia stesura della
Repubblica. Si tratta in ogni caso di argomenti altamente congetturali, date le
modalita 'fluide' di pubblicazione e di circolazione di questi testi di scuola.
11 G. Cambiano, Platone e le tecniche (Roma-Bari 19912) 149 sgg., mette in
luce come Ia polemica di Platone colpisca anche il carattere 'classista' della
medicina dietetica, fruibile, per ragioni di tempo e di costo, solo da parte dei
rice hi.

LA CRITICA Dl PLATONE ALLA MEDICINA

69

&vayJCal.ov pya~Ea8at). 11 nuovo regime e invece messo a disposizione dei ricchi, i quali so no piu attenti alla salute (3.69): esattamente, dunque, quello stravolgimento di interesse, dal bene pubblico alia salute privata, che Platone condanna nella medicina
dietetica. E proprio in questo modo che essa si trasforma, da
sintomo ed effetto, in concausa della malattia sociale.
3. L'eroe negativo di questa metamorfosi della nuova medicina,
appropriata alia polis tryphosa, e individuato da Platone nella figura
di Erodico di Selimbria. 12 Erodico era stato accusato nel Protagora,
come si e visto, di essere un sofista travestito>> e potente; nel Fedro
(227d), Platone ironizzava sulle lunghe e faticose passeggiate cui
Erodico sottoponeva se stesso e presumibilmente i suoi pazienti (un
eccesso di esercizi per cui evidentemente era famoso e anche
aspramente criticato da altri medici) . 13 Tutto cio dipendeva probabilmente dalla sua teoria secondo Ia quale era necessaria riportare
il rapporto fra alimentazione ed esercizi al suo equilibria naturale
(Anon. Lond. 9.20 sgg.).
Nella Repubblica, Erodico viene accusato di aver fatto deviare Ia
medicina dalla sua efficace tradizione terapeutica trasformandola
in un controllo minuzioso e ossessivo del regime di vita dei malati.
Fattosi medico dopo esser stato maestro di ginnastica (1tatOo-rpi~rt~),
egli ha mescolato>> e confuso la ginnastica con la medicina
( 406a-b): si allude evidentemente alia cura degli esercizi fisici
raccomandata da Erodico. Ha praticato con rigore ossessivo il
rispetto della dieta consueta e prescritta ( 406b), di lunghissima
durata ( 406d). In questo modo, ha trasformato Ia medicina da
terapia in <<pedagogia delle malattie>> ( 406a), 14 in una voao-rpO<pia
12 Le testimonianze su Erodico sono raccolte in H. Grensemann, Knidische
Medizin. Teil I (Berlin-New York 1975) 15 sgg. Cfr. ora A. Jori, "Platone e Ia
svolta dietetica della medicina greca. Erodico di Selimbria e le insidie della
techne", Studi italiani di filologia classica 11 (1993) 157-95, che propone una
cronologia di Erodico fra il 460-50 e il 380.
13 Cfr. Epidemie 6.3.18. Probabilmente a causa della polemica di Platone,
Erodico era gia diventato proverbiale a! tempo della Retorica di Aristotele (1.5
136lb 4-6: molti sono sani nel senso in cui si dice sano Erodico; nessuno
invidierebbe Ia !oro salute, perche devono astenersi da tutte le soddisfazioni
umane o dalla maggior parte di esse).
14 Nel Timeo, a! contrario, Platone avrebbe raccomandato di Jto:too:yroynv
omito:u; le malattie croniche, invece di irritarle con farmaci, nei limiti di
tutta Ia crxoA.{J disponibile (89c-d): un ritorno 'erodiceo', dunque, da annoverarsi fra le molte tensioni nel pensiero platonico determinate dalla polemica di Repubblica.

70

MARIO VEGETTI

( 407b), che prolungano un' esistenza malaticcia distogliendo I' attenzione e le energie dei cittadini dalla vita sociale per concentrarle in
modo morboso-noi diremmo ipocondriaco-sulle condizioni di
salute e le loro minime variazioni quotidiane.
Va detto che molte delle accuse platoniche colpiscono l'intera
medicina dietetica, compresa quella di ispirazione 'ippocratica', e
non soltanto la figura di Erodico. La continuita fra ginnastica e
medicina era per esempio stata fortemente sottolineata in Antica
medicina (3, 4, 7), e del resto condivisa da Platone (Crito 47b; Gorg.
464b; una contrapposizione tra le due arti, forse proprio in polemica
contra Erodico, era stata pen) stabilita in L.H. 35). I pericoli derivanti
da un mutamento della dieta consueta erano stati enfatizzati nella
stessa opera (10-11), e confermati in Regime delle malattie acute (26).
Perche dunque Platone ha deciso di fare di Erodico il suo principale bersaglio polemico, il capofila emblematico della generazione degli Asclepiadi che li rende complici dei vizi della citta del
lusso?
Una risposta puo forse venire proprio dagli altri due dialoghi in
cui il medico di Selimbria e citato esplicitamente.
Nel Protagora, all'attacco contra Erodico sofista corrisponde la
menzione di Ippocrate di Cos come maestro per antonomasia di
medicina, al pari di Policleto e Fidia per la scultura (3llb-c). II
Fedro, che si apre con !'ironia su Erodico, culmina nel celebre
elogio del metoda di Ippocrate, cui si deve ispirare Ia riforma
dialettica della retorica (270b-c).
Erodico sembra dunque configurarsi come l'anti-lppocrate: concentrare sui suo nome l'aggressione polemica poteva significare
per Platone mettere al riparo da essa Ia medicina ippocratica, che,
nonostante le sue ovvie tendenze dietetiche, manteneva ai suoi
occhi un rilevante val ore metodologico ( e anche, stan do al primo
libro della Repubblica, deontologico).
Del resto, piu specificamente, se e vero-come qui si e cercato di
dimostrare-che Platone individuava gli eccessi degenerativi e
socialmente pericolosi della medicina dietetica in opere come il
Regime, puo darsi che egli fosse al corrente di quello stretto rapporto
fra quest'opera e il pensiero di Erodico, che e stato a piu riprese
ipotizzato dai moderni.I5
15 Un rapporto fra Erodico e il Regime era stato sostenuto gia da C.
Fredrich, "Hippokratische Untersuchungen", Philol. Untersuchungen 15 (Berlin
1899) e in seguito riproposto da K. Deichgraber, "Die Epidemien und das

LA CRITICA DI PLATONE ALLA MEDICINA

71

Questo puo comportare una conseguenza di rilievo per la stessa

'questione ippocratica'. Se esiste in Platone una contrapposizione fra
Ippocrate ed Erodico, e se quest'ultimo puo essere awicinato alla
dietetica del Regime (al quale e se non altro accomunato dalla stessa
polemica platonica)' e allora da escludere che l' opera attribuibile a
lppocrate secondo il <<metodo>> descritto nel Fedro sia proprio il
Regime, secondo la tesi sostenuta da W.D. Smith.l6 Per contro,
l'ampio uso che Platone fa di Arie acque luoghi nella diagnostica
sociale e nella psicologia collettiva del libro III puo confermare che
questo testo sia fra quelli riferibili a lppocrate aHa luce della
testimonianza del Fedro.11
4. Torniamo ora alla discussione dellibro Ill. Come si e visto, la
medicina dietetica era servita a Platone come strumento diagnostico per la malattia della polis tryphosa; sul piano sociale, il suo
sviluppo era stato considerato come un effetto della degenerazione
morale di questa citta, con i suoi abusi alimentari e la sua 'pigrizia',
e infine anche, negli eccessi erodicei, a sua volta come una causa
capace di confermare quella degenerazione.
Una riforma della medicina e dunque indispensabile, parallelamente aHa <<purificazione>> della citta. Essa sara mirata ad un
corretto uso politico delle procedure terapeutiche, destinate soltanto
ad un sollecito ripristino delle prestazioni lavorative e pubbliche dei
cittadini, ricchi o poveri che siano.
Da questo punto di vista, Platone ritiene necessaria un ritorno
alia medicina tradizionale, che si ispira a quella praticata sui campi
di battaglia dai <<figli di Asclepio>>, contrapposti cosi: alla corporaCorpus Hippocraticum ", Abhandl. d. Preuss. Akad. d. Wissensch., Phil.-Hist.
KJasse, 3 (Berlin 1933) 58-64. R. Joly, introd. a Du regime (Paris 1967) nega
che Erodico possa venir considerato l'autore dell'opera, in base alia critica ai
ginnasti di 1.24, rna ammette che le sue dottrine sono riprese e sviluppate nel
Re.ftme (p. 13).
6 Cfr. W.D. Smith, The Hippocratic Tradition (Ithaca-London 1979) 44 sgg.
Smith con divide Ia datazione del Regime proposta da Joly (p. 60).
17 Per questa tesi, dopo Deichgraber, op. cit. (n. 15), cfr. soprattutto J.
Mansfeld, "Plato and the Method of Hippocrates", Greek Roman and Byzantine
Studies, 21, 1980, 341-62, e Id., "The historical Hippocrates and the origins of
scientific medicine", in Nature Animated, ed. M. Ruse (Dordrecht 1983) 49-76.
Per una posizione simile, rna con argomenti diversi e meno convincenti, cfr.
R. Joly, "La question hippocratique et le temoignage du Phedre", Revue des
Etudes grecques, 74, 1961, 69-92, e Id., "Platon, Phedre et Hippocrate: vingt ans
apres", in Formes de pensee dans la Collection hippocratique, eds. F. Lasserre-P.
Mudry (Geneve 1983) 407-22. La testimonianza del Fedro e discussa in M.
Vegetti, La medicina in Platone (1969, Venezia 1995) cap. IV.

72

MARIO VEGETTI

zione professionale dei moderni Asclepiadi. Se essi non si dedicavano aile raffinatezze dietetiche proprie di questi ultimi, era perch:
avevano ben chiari i compiti sociali e collettivi della loro arte, non
certo per <<l'ignoranza e l'inesperienza>> che i sofisticati medici
moderni attribuiscono ai Ioro predecessori tradizionali ( 406c: e
forse possibile scorgere in queste parole una confutazione polemica
di Male sacro 1-2, che appunto attribuiva a1tupia e ignoranza ai
guaritori ciarlatani legati alia superstizione religiosa).
Ma il richiamo ad Asclepio e ai suoi figli e evidentemente solo
polemico ed ironico. Lo sottolineano due pungenti battute di Glaucone: <<davvero raffinati (KOJ.HJfOt) li descrivi, i figli di Asclepio>>
( 408b); e <<Un politico (7tOAtttJCOV) ne fai di Asclepio>> (407e).
AI contrario, Platone appare molto serio, e perfettamente informato, quando prescrive i requisiti per Ia medicina adatta alia citta
purificata.
(a) Essa deve trattare solo malattie ben circoscritte (v6or11.1.a
a1tOKEKptJ.1.EVov, 407d). Qui Platone si oppone alia tendenza della
medicina moderna, inclusa quella ippocratica, a farsi carico
dell'intero organismo malato, e, al limite, dell'intera esistenza del
paziente, a scopo sia terapeutico sia profilattico. Una tendenza, del
resto, che egli aveva condiviso e approvato. Secondo il Carmide
(156b-c), i <<buoni medici>> (aya8o1. ia1:poi, qui sicuramente gli
'ippocratici'), se si rivolge Ioro uno che ha una malattia agli occhi,
<<gli dicono che non si puo cominciare a curare solo gli occhi, rna
che bisognerebbe curare anche Ia testa se si vuole guarire gli occhi;
e dicono ancora che e un'assurdita pensare di curare Ia testa per se
stessa senza tenere conto dell'intero corpo. Cosi [ ... ] cercano di curare e guarire Ia parte applicando un regime all'intero corpo>>
( otai'tat E7tt 1tiiv 1:0 o&J.I.a). Ma i buoni (e ippocratici) medici del
Carmide, bench: esemplari metodicamente, non sono socialmente
utili nella citta sana>> della Repubblica.
(b) II trattamento di queste malattie locali deve essere rapido e
immediatamente efficace: esso non impieghera dunque diete rna
solo farmaci catartici, cauterizzazioni e incisioni chirurgiche
( 406d). Questa triade e naturalmente canonica, rna anche qui Platone corregge gli ippocratici e se stesso. Arie acque luoghi, per esempio,
aveva sottolineato Ia necessita di astenersi dal somministrare
cpapJ.I.aKov, 'taJ.I.VEtv e JCaiuv durante i mutamenti di stagione (11). E
anche Platone nel Protagora (354a) aveva indicato come rimedi
standard della medicina cpapJ.I.aKov, JCauot<; e 'toJ.I.ft, aggiungendovi

LA CRITICA DI PLATONE ALLA MEDICINA

73

tuttavia anche le restrizioni dietetiche (AtJ..LOK'tovia), che qui vengono omesse in polemica contro Ia medicina dietetica e probabilmente come inutili in una polis dove il regime alimentare e gia
politicamente controllato.
(c) La medicina inoltre dovra rifiutarsi di trattare quanti siano di
costituzione malferma e incapaci percio di rendersi socialmente
utili, lasciandoli morire>>, proprio come i giudici dovranno mandare a morte coloro che siano incurabili nell'anima (410a).
II punto di vista di Platone e qui perfettamente chiaro. L'esistenza
individuale non ha alcun valore se non in funzione della sua
utilita sociale; sopravvivere senza essere in grado di contribuire al
benessere comune non ha alcun senso ne personale ne pubblico.
Questo punto di vista non ha naturalmente paralleli nella deontologia medica, anche se vi compare a volte (cfr. peres. p. technes 3) il
rifiuto di affrontare Ia terapia degli incurabili, dovuto soprattutto alia
preoccupazione che la loro morte possa venir addebitata al medico
invece che alia malattia. E qui il caso di accennare che il parallelo
platonico fra medicina e giustizia, 18 con I' obbligo per Ia seconda di
mandare a morte chi sia moralmente irricuperabile, e per Ia prima
di lasciar morire gli invalidi permanenti, era destinato ad inaugurare una lunga tradizione di pensiero tanto etico quanto medico.
Essa sarebbe culminata, con Galeno (Quod animi 11, K 4.815-6}, in
un singolare rovesciamento. 19 Poiche le devianze morali sono
secondo Galeno determinate da malformazioni organiche, spetta al
medico-che si appropria cosi anche della funzione del giudice
platonico-di diagnosticare chi sia incurabile nel corpo e (dunque)
anche nell'anima e di decretare Ia condanna a morte di coloro che
risultano tanto <<irrimediabilmente malvagi>> da non poter essere
rieducati dalle Muse stesse ne migliorati da Socrate o da Pitagora
(e neppure, s'intende, curati dallo stesso Galeno). In Platone, c'e
ancora parallelismo, e non convergenza, fra sentenza di giustizia e
diagnosi medica, anche se, come si e visto, la patologia morale influenza quella organica e viceversa. La nuova medicina, frutto di
una purificazione della citta, contribuira comunque a sua volta a

18 II parallelo fra medicina e giustizia, rna senza riferimento ai rimedi

estremi del nostro passo, era stato elaborato in Gorg. 478a sgg.
19 Per questa vicenda, anche in rapporto al riferimento aristotelico alia
punizione medica o politica (Eth. eud. 1.3), cfr. M. Vegetti, "Cura dei
tribunali, punizioni della medicina", in Immaginario e follia, ed. F. Rosa
(Trento 1991) 29-36.

74

MARIO VEGETTI

depurarla dagli individui inutili, al contrario di quella dietetica che

l'aveva invece intossicata prolungando, alia maniera di Erodico, la
loro malsana esistenza.
5. L'atteggiamento platonico verso la medicina nei libri II e III
della Repubblica risulta dunque piuttosto articolato e complesso. Da
un lato, Platone non dimentica Ia fruibilita metodica del modello
medico: essa sara certo enfatizzata soprattutto nel Fedro, e nella
Repubblica viene messa in secondo piano per il privilegio che i suoi
libri centrali riconoscono aile matematiche, rna anche qui, come
si e visto, Ia medicina offre utili strumenti diagnostici e modelli
terapeutici per Ia patologia sociale della polis tryphosa. Anche sui
piano etico, il rapporto medico-paziente, che dal Gorgia aile Leggi
costituisce un possibile paradigma del potere giusto, non manca di
venir richiamato, nello stesso senso, nel I libro del nostro dialogo.
Dall'altra parte, pen), Platone attacca violentemente il ruolo
sociale della medicina dietetica in quanto solidale alia patologia
morale della polis tryphosa. In un 'regime di salute pubblica', quale
e quello necessario alia purificazione della citta, questa medicina
non puo a sua volta che essere destinata all'evacuazione purgativa,
diventando cosi vittima dei suoi stessi canoni terapeutici. II necessaria ridimensionamento della funzione pubblica della medicina e
al tempo stesso una premessa e una conseguenza del risanamento
della citta: nella polis sana, moderata e temperante, serviranno
certamente me no medici, me no iatreia, me no diete (come anche
meno tribunali e avvocati). E in tanto, le energie sociali necessarie
a questo risanamento andranno ricuperate distogliendo l'attenzione
dalla vicenda privata e individuale delle terapie dietetiche per
reinvestirla nella sua prioritaria destinazione pubblica.
Ma in questa drastica svolta, Platone cerca comunque di salvare
il salvabile, e ne e un segno il fatto che Ia sua polemica risparmi
qui il nome di lppocrate. Serviranno ancora buoni medici ( 408c-d),
presumibilmente di ispirazione 'ippocratica' nel metodo e nell' ethos
(gli agathoi iatroi menzionati nel Carmide). Essi dovranno pero essere
'politicizzati', al pari dell'Asclepio rivisitato da Platone, e di tutti gli
altri cittadini-pazienti. Che Ia buona medicina>> descritta nel
Carmide e neUe Leggi sia suscettibile di questa politicizzazione,
secondo il modello della sbrigativa terapia dell'artigiano>> a base
di farmaci, incisioni e cauterizzazioni approvata nel libro III, e un
serio problema concettuale, che Platone non affronta in questa

LA CRITICA DI PLATONE ALLA MEDICINA

75

sede, sotto l'urgenza del progetto di riforma etico-politica e anche

sanitaria della polis. In seguito, il ridimensionamento di questa
progetto consentin1 alia medicina di riassumere, in forme variahili, Ia sua funzione di paradigma metodico e di modello etico,
senza l'asprezza delle tensioni prodottesi nella Repubblica.2
Pavia

20 Una versione leggermente modificata di questo testo compare anche, in

forma di pre-print, nel II fascicolo di Platone. Repubblica, traduzione e commento a cura diM. Vegetti (Dipartimento di filosofia, Pavia 1996).

fEyOVEV (PLATON, TIM. 28 B 7)

1ST DIE WELT REAL ENTSTANDEN ODER NICHT?
MATTHIAS BALTES

Von allen Fragen, die der Platonische ''Timaios" aufWirft, ist iiber
keine so viel diskutiert und gestritten worden wie iiber die hier
gestellte. Das gilt fiir die lateinische und griechische Antike, aus
der uns zu diesem Problem mehr AuBerungen erhalten sind als
zu allen anderen 1 Das gilt in ahnlicher Weise fiir das lateinische,
griechische und arabische Mittelalter2, die beginnende Neuzeit
und die neueste Zeit. Wie schon in der Antike, so zeichnet sich
auch unter den heutigen Interpreten kein Konsens ab: Die einen
sind nach wie vor der Meinung, Platon lehre im "Timaios", die
Welt sei zeitlich oder, wenn schon nicht zeitlich, so doch wenigstens in einem einmaligen Schopfungsakt entstanden, wahrend
die Gegenseite ebenso beharrlich die Auffassung vertritt, der
Schopfungsakt, von dem der "Timaios" spricht, sei nur ein bildlicher Ausdruck fiir ein immerwahrendes Schopfungs-Geschehen, fiir eine creatio continua oder perpetua: die Darstellung entfalte
das, was sich im Kosmos immerfort gleichzeitig abspiele, in
einem zeitlichen Nacheinander, urn auf diese Weise die innere
Struktur des Kosmos und der Vorgange in ihm einsichtig zu
machen3. Beide Seiten konnen sich dabei nicht nur auf Passagen
1 Vgl. M. Baltes, Die Weltentstehung des Platonischen Timaios nach den antiken
Interpreten 1-11 (Lei den 1976-78); H. Dorrie - M. Baltes, Der Platonismus in der
Antike III (Stuttgart-Bad Cannstadt 1993) 296 ff.
2 C. Baeumker, "Die Ewigkeit der Welt bei Plato", Philosophische Monatshefte 23 (1887) 514 ff; M.-D. Chenu, "Die Platonismen des XII. Jahrhunderts",
in Platonismus in der Philosophie des Mittelalters, hgg. von W. Beierwaltes
(Darmstadt 1969) 274 ff. 287; M. Fakhry, "A Tenth-century Arabic Interpretation of Plato's Cosmology", journal of the History of Philosophy 6 (1968) 15 ff =
Philosophy, Dogma and the Impact of Greek Thought in Islam (Brookfield, Aldershot
1994) VI 15 ff; F. van Steenberghen, "La controverse sur l'eternite du monde
au Xllle siecle", L 'Academie Royale de Belgique. Bulletin de la Classe des Lettres et
des Sciences Morales et Politiques, 5e serie, 58 (1972) 267 ff; H. Simon, "Weltschopfung und Weltewigkeit in der jiidischen Tradition", Kairos 14 (1972) 31
ff; R. Sorabji, Time, Creation and the Continuum (London 1983) passim; ders.,
Matter, Space and Motion (London 1988) 259 ff; Dorrie-Baltes III 298 mit
Anm.4.
3 Die Hauptvertreter der heiden Richtungen werden vorgestellt bei W.
Scheffel, Aspekte der platonischen Kosmologie (Leiden 1976) XI ff und R. Hackforth, "Plato's Cosmogony (Timaeus 27 D ft)", Class. Quart. N.S. 9 (1959) 17; die

1ST DIE WELT REAL ENTSTANDEN ODER NICHT?

77

des 'Timaios" selbst berufen, sondern auch auf Stellen aus anderen Platonischen Dialogen; vor allem der "Staat", der "Politikos",
der "Philebos", der "Kritias" und die "Nomoi" spielen in dieser
Diskussion eine groBe Rolle 4 Beide Seiten ziehen fiir ihre unterschiedlichen Ansichten zudem die unmittelbaren Platonschiiler
heran, Aristoteles einerseits und Speusipp, Xenokrates und Herakleides Pontikos andererseits 5 Schon in der Schute Platons, der
Akademie, war man sich also offenbar nicht einig dariiber, wie
der Schopfungsbericht im 'Timaios" zu verstehen sei. Wenn nun
aber schon Platons direkte Schuler in dieser Frage uneins waren,
wie sollen wir da in der Lage sein, die Sache zu entscheiden?! Die
Situation scheint vollig verfahren, das Ratsel unlosbar zu sein.
Bei unserem Versuch, das Ratsel zu losen, wollen wir uns ausschlieBlich auf den "Timaios" konzentrieren und zunachst seinen
Aufbau, genauer den Aufbau der Rede des Timaios im "Timaios"
betrachten. Unter dem Gesichtspunkt einer zeitlichen oder einmaligen Schopfung ist dieser Aufbau geradezu chaotisch. Denn
alteren Vertreter werden genannt bei E. Zeller, Die Philosophic der Griechen in
ihrer geschichtlichen Entwicklung II 1 (Leipzig 19225, Nachdr. Darmstadt 1963)
792, Anm.1; G. Vlastos, "The Disorderly Motion in the Timaeus" (1939), in
Studies in Plato's Metaphysics, ed. R.E. Allen (London 1965) 379 und H.
Cherniss, f..ristotle's Criticism of Plato and the Academy (Baltimore 1944) 424,
Anm.357. Uber die dort Genannten hinaus treten heute ein
- fiir die nichtzeitliche Auffassung der Weltentstehung im "Timaios": M.
Landmann, Ursprungsbild und Schopfertat (Miinchen 1966) 162 ff; E. Ostenfeld,
"Disorderly Motion in the Timaeus", Classica et Mediaevalia 29 (1968) 22 ff; L.
Taran, "The Creation Myth in Plato's Timaeus", in Essays in Ancient Philosophy,
ed. J.P. Anton and G.L. Kustas (State University of New York Press, Albany
1972) 372 ff; L. Brisson, Le meme et l'autr:e dans la structure ontologique du Timee de
Platon (Paris 1974) 82 f; K. Alt, "Die Uberredung der Ananke zur Erklarung
der sichtbaren Welt in Platons Timaios", Hermes 106 (1978) 460 f;
- fiir die zeitliche Auffassung: F. Solmsen, Aristotle's System of the Physical
World (Ithaca, New York 1962) 51; G. Vlastos, "Creation in the Timaeus: Is it a
Fiction?" (1964), in Studies in Plato's Metaphysics, ed. R.E. Allen (London 1965)
401-419; J. Pepin, Theologie cosmique et theologie chretienne (Ambroise, Exam. I 1,1-4)
(Paris 1964) 85 f; HJ. Easterling, "Causation in the Timaeus and Laws X",
Eranos 65 (1967) 25; T.M. Robinson, Plato's Psychology (Toronto 19952) 64-104,
bes. 68 f; ders., "The Argument of Tim. 27 d ff", Phronesis 24 (1979) 105 ff;
W.K.C. Guthrie, A History of Greek Philosophy V (Cambridge 1978) 301 ff;
Sorabji, Matter 248-259. 273-282; ders., Time 272 ff; J. Mansfeld, "Rez. Scheffel, Kosmologie", Mnemosyne 35 (1982) 168 ff; R.D. Mohr, The Platonic Cosmology
(Leiden 1985) 64 ff; G. Naddaf, "The Atlantis Myth: An Introduction to Plato's
Later Philosophy of History", Phoenix 48 (1994) 202; vgl. dens., L'origine et
!'evolution du concept grec de 'phusis' (Lewiston 1992) 391 ff, bes. 413;
- fiir eine unzeitliche, aber reale Weltentstehung: Scheffel, Kosmologie.
4 Vgl. Baltes, Weltentstehung I 210. 216 mit Verweisen (s. auch die Indices
I 237 f) und II 166 f.
5 Zu diesen vgl. Taran, "Creation Myth" 389 f (zu Herakleides speziell
404, Anm.113); Baltes, Weltentstehung I 5 ff. 94 ff.

78

MATTHIAS BALTES

die Schritte, in denen sich der Schopfungsakt im "Timaios"

vollzieht, sind derart iiber die Rede des Timaios verstreut, daB man
groBte Miihe hat, sie zusammenzustellen. Es sind die folgenden:

1. Die Situation, bevor der Gott eingreift: in dieser existieren drei Dinge,
das Seiende, der Raum und das Werden (ov, xropa, YEvEcrtc;, 51 D 3).
Die yevEcrtc; bezeichnet den Zustand, in dem die Ideen ( ov) auf den
Raum (xropa) wirken und in ihm Spuren {tXVTl ana, 53 B 2) von
Elementen hervorrufen 6 , die sich zwar noch in vollkommener
Unordnung befinden (52 D 3 - 53 A 8), aber doch auch schon
Andeutungen ihrer kunftigen Eigenschaften besitzen (68 E 1 ff),
beispielsweise Schwere, Dichte (53 A 1 f) und Sichtbarkeit (52 E 1). Da
die avroJlaAia in diesem vorkosmischen Chaos zu einer Schuttelbewegung der Chora fUhrt (52 E 1 ff; vgl. 57 D 7 ff), findet in diesem
Stadium sogar schon eine Art Gruppierung der Spurenelemente statt
(52 E 5 fl) 7. Dies ist der Zustand 1tpo tile; tou oupavou YEVEO"Eroc; (48 B 3 ff;
52 D 4; 53 A 7), in dem die avayKTl herrscht (4 7 E 4 f; 68 E 1. 6 f) und
den der Gott vorfindet, bevor er einschreitet (30 A 2-4; 69 B 2 f. 5 ff).
2. Auf der zweiten Stufe tritt der Gott auf, der in dieses Chaos eingreift
und es ordnet, indem er die Spuren der Elemente durch Gestalt und
Zahl formt ( OtEO"XTlJl!lttO"!ltO etOEO"t tE JC(ll apt6jlotc;, 53 B 4 f). Nun erst
gelangen Ordnung und MaBin das Chaos (53 B 1- 7; 69 B 4 f); denn
nun werden die einzelnen Elemente und auch der Kosmos als ganzer
nach den fUnfPlatonischen Korpern gestaltet (53 B 7-56 C 7). Erst auf
dieser Stufe herrschen die mathematischen Gesetze unter den
Elementen ( 69 B 4 f), erst jetzt gibt es einen Kosmos 8 , in dem die
einzelnen Glieder durch Proportion gebunden sind und in
Freundschaft miteinander Ieben (31 B 4-32 C 4). Dies ist der Zustand,
in dem der vouc; die Herrschaft ubernimmt und die UVU'Y1CT1 uberredet,
so daB diese zum Besseren des Ganzen nachgibt (47 E 5-49 A 5; 56 C
5 f). Von nun an herrschen im Kosmos vouc; und q>pOVTlcrtc; (34 A 2; 39
C 2; 47 B 7; 90 C 7 fl).
3. Wo bleibt nun aber in dieser Abfolge die Weltseele? Nach dem, was
der "Timaios" uber die ersten heiden Stufen lehrt, ist die Weltseele
6 Zu diesem Wirken der Ideen auf d_ie xropa vgl. Prokl. In Plat. Tim. I
387,8 ff; 388,16 ff. 28 ff Diehl; L. Robin, "Etudes sur Ia signification et Ia place
de Ia phyique dans Ia philosophie de Platon", in La pensee hellenique des
origines a Epicure, publ. par P.M. Schuh! (Paris 1942, Nachdr.1967) 258; Baltes,
Weltentstehung I 17, Anm.46 und 48; Scheffel, Kosmologie 103 ff. 142, dem ich
jedoch nicht in allen Einzelheiten zustimmen kann.
7 DaB nach dieser Gruppierung der Spurenelemente die ungeordnete
Bewegung aufhoren miiBte, kann ich Tarin, "Creation Myth" 386 ff nicht
zugeben; denn die Ideen wirken ja immerfort weiter auf die xropa; und rufen
in ihr standig neue Spurenelemente hervor, deren aVffiJlClAlCl die xropa;
weiterhin in Bewegung setzt, so daB die Schiittelbewegung andauert. Ebenso
kann ich ihm nicht zugeben, daB die Seele im "Timaios" QueUe aller Bewegungen ist (a.O. 384 u.o.); sie ist nur Quelle aller geordneten Bewegungen.
8" Auf den Gedanken, daB es "the world" schon vor dem Eingreifen des
Demiurgen gibt, kann man nur kommen, wenn man die Leistung des
Demiurgen unerlaubterweise so sehr minimalisiert, wie Mohr, Platonic
Cosmology 66 f es tut.

1ST DIE WELT REAL ENTSTANDEN ODER NICHT?

79

spater als die dort entstehenden Korper, einschlieBlich des Kosmoskorpers. Jedenfalls wird sie auf diesen Stufen nirgends erwahnt.
Dagegen ist der Kosmoskorper nach 30 A 6 - C 1 von Anfang an beseelt,
und nach 34 B 10 ff ist die Weltseele sogar friiher als der Kosmoskorper; dieser wirdja in die schon erstellte Seele hineingebaut (36 D 9 f),
und die Sterne werden erst nach ihrer Erstellung in die Umlaufe der
Seele hineingesetzt (38 C 7 f; 40 A 4 f). Also ist die Weltseele nach
Platon wirklich friiher als die Weltkorper. Aber wann ist sie dann
entstanden? Auf der ersten Stufe oder erst auf einer Zwischenstufe
nach der ersten und vor der zweiten Stufe? Tim. 34 B 10 ff scheint fiir
die zweite Alternative zu sprechen 9.
Auch bei der Erschaffung der Zeit tun sich Schwierigkeiten auf; denn
nach der Erzahlung des Timaios schafft der Gott die Zeit nach der
Weltseele und dem Weltkorper, aber vor den Gestirnen (37 C 6 ff);
doch korrigiert Timaios sich sogleich, indem er sagt, die Zeit sei
zusammen mit dem Himmel entstanden (38 B 6 ff). Obschon also die
Darstellung des Timaios dies zunachst suggeriert, bildet die Erschaffung der Zeit dann doch keine eigene Stufe.
4. Auf der vierten Stufe erfolgt die Erschaffung der Gestirngotter und der
~(lt).lOV~ (39 c 3-41 A 6).
5. Erst danach entsteht auf der fiinften Stufe der unsterbliche Teil der
Menschenseele, der aus den zweit- und drittrangigen Resten der
Ingredienzien der Weltseele geschaffen wird (41 D 3 ff). Dieser Teil
der Menschensede entsteht nach den GOtterseelen (vgl. 38 E 5 f; 40 B
5). Auch die Gotterseelen miissen also aus Resten der Ingredienzien
der Weltseele entstanden sein, doch wird man annehmen konnen,
daB diese Reste nicht zweit- und drittrangig waren. Also ist vor der
Entstehung des gottlichen Teils der Menschenseele und der
Gestirnkorper zwischen der dritten und fiinften Stufe die Entstehung
der Gotterseelen anzusetzen, die der "Timaios" nur voraussetzt, aber
nicht behandelt.
6. Auf der sechsten Stufe zieht der Gott sich zuriick, denn er hat seine

Aufgabe erfiillt, und die Untergotter losen ihn ab. Sie schaffen jetzt
den menschlichen Korper, binden die gottliche Seele in diesen und
fiigen dann den sterblichen Teil der Menschenseele hinzu ( 42 E 5 ff;
69 C 5 ff). Dabei entsteht nur der Mann ( 41 E 3 ff; 90 E 6 ff). Wie die
Stellung des Berichts im Ganzen nahelegt, werden auf derselben
Stufe die Pflanzen erschaffen, die den entstandenen Menschen als
Nahrung dienen sollen (76 E 7-77 C 5).
7. Auf der siebten Stufe entstehen die Frauen und mit ihnen auch die
Fortpflanzungsorgane des Mannes (42 B 5 ff; 76 D 8; 90 E 6 ff).
8. Auf der achten Stufe entstehen die Tiere (42 A 1 - D 2; 76 D 8 ff; 91 D 6
ff).
9 Trotzdem bleibt auch diese Losung schwierig; denn wenn Tim. 48 A
der vouc; die avayJCTl zu iiberreden vermag, muB diese schon von sich aus Seele
besitzen, wenn denn der Satz gelten soli, daB vouc; ohne Seele niemandem
zukommen kann (Tim. 30 B 3). Hat also das vorkosmische Chaos schon vor
dem Eingreifen des Gottes Seele? Oder wird die Seele der avayJCTl in dem
Augenblick mitgegeben, in dem der Gott sie anspricht? Eine unlosbare
Schwierigkeit.

80

MATTHIAS BALTES

Diesen Aufbau der Handlung kann man unter dem Gesichtspunkt

einer zeitlichen Schopfung nur als katastrophal bezeichnen. Es
kann nicht Platons Absicht gewesen sein, den wirklichen Verlauf
der Schopfung darzustellen, und es war auch nicht seine Absicht.
Was Platon mit seiner Darstellung wirklich intendierte, war keine
Entstehungs- und Entwicklungsgeschichte des Kosmos, sondern
eine Analyse der Leistungen der einzelnen Faktoren, die bei der
Schopfung zusammenwirken. DaB dies so ist, hebt Timaios selbst
im Verlauf seines Vortrags mehrfach hervor. So sagt er beispielsweise, daB der Abschnitt 27 D 5 - 29 D 3 als 7tpOOtJ..ltOV zu gelten hat,
in dem die fiir die Darstellung notwendigen Vorfragen geklart
werden (29 D 4-6).
An das 7tpOOtJ..l10V schlieBt sich 29 D 7 - 47 E 2 zunachst ein
Abschnitt an, der die Entstehung des Kosmos unter dem Gesichtspunkt betrachtet, was in ihr die Vernunft leistet; es folgt der
Abschnitt 47 E 5- 68 D 7, der danach fragt, was die Notwendigkeit
( av ay-Kll) zur Kosmogenese beitragt ( 4 7 E 3-5). Am En de dieses
Abschnitts stellt Timaios ausdriicklich fest, daB nunmehr die
Analyse beendet sei und die Synthese der restlichen Darstellung
beginnen konne: "Die Gattungen der Ursachen liegen siiuberlich
getrennt bereit, aus welchen es nun die restliche Darstellung
z.usammenz.uweben gilt" (7tapaKEnat 'ta 'trov ah{rov yv11 OtUAtcruva, ~
<i>v 'tOV 1tiA.omov A6yov 81 cruvucpavSnvm, 69 A 6 f).
DaB Timaios dieses Verfahren wahlt, ist urn so auffalliger, als
er selbst zugeben muB, daB beide Abschnitte, der iiber die Leistung
der Vernunft und der iiber das Wirken der Notwendigkeit, nicht
ganz ohne das Wirken des jeweils anderen Prinzips auskommen
(vgl. 1tATtV ~paxrov, 47 E 3). DaB Timaios dieses Vorgehen trotzdem
vorzieht, hat seinen Grund in der 46 D 7 ff geauBerten Uberzeugung, daB jeder verniinftige Mensch zuerst die vernunftbestimmten Ursachen erforschen miisse und dann erst die durch
die Notwendigkeit vorgegebenen. Denn nur die Suche nach den
ersteren fiihre zu einem gliicklichen Leben (68 E 7 f) 10 .
Die Darstellung ist also nicht durch die zeitliche oder sachliche
Abfolge, sondern durch systematische und axiologische Uberlegungen bestimmt. Timaios weist ja ofter selbst darauf hin, daB
die Reihenfolge der Darstellung nicht der sachlichen Reihenfolge entspricht (34 B 10 ff; 48 A 7 ff; E 2 ff).Ja, gelegentlich hebt er
10

Dazu Scheffel, Kosmologie 126 ff.

1ST DIE WELT REAL ENTSTANDEN ODER NICHT?

81

sogar hervor, daB er urn der Darstellung willen Zusammengehoriges trennen muB (61 C 3-D 5).
Timaios gibt auch klar und deutlich zu verstehen, warum er
dieses Vorgehen wahlt: Er ist iiberzeugt, daB seine Zuhorer auf
diese Weise besser verstehen, was er sagen will (27 C 7 - D 4). Dies
gilt vor allem fiir die Zweiteilung in die Abschnitte "Werke der
Vernunft" und "Werke der Notwendigkeit"; denn erst nachdem
Timaios seine Zuhorer mit den leichter verstandlichen Werken
der Vernunft vertraut gemacht hat, fiigt er die "unerhorte und
ungewohnliche Darlegung" (a:t01t0~ Kat ail~~ ~htl"fll<Jt~, 48 D 5 f)
iiber "die schwierige und dunkle Art" (xaA.e7tov Kat UJ.!UOpov doo~),
die "Aufnehmerin allen Entstehens" (7t<l<Jll~ yevcrero~ t>7toOoxfl, 49
A 3-6) an, eine his dahin vollkommen unbekannte Realitat, iiber
die noch nie jemand gesprochen habe ( 48 B 5 fl).
Die genannte Erklarung gilt in gleicher Weise fiir die Binnenstruktur der einzelnen Teile. So wird im Abschnitt iiber die Werke
der Vernunft die einfachere Entstehung des Kosmoskorpers (29 D
7- 34 A 7) vor der komplizierteren Entstehung der Weltseele (34 A
8 - 36 D 7) behandelt, obschon die Weltseele "friiher" ist als der
Weltkorper 11 Ebenso wird die Entstehung der Zeit nach der Entstehung des Kosmos besprochen (37C6fl),obschon sie doch zusammen
mit dem Kosmos entstanden ist (38 B 6 fl). Aus denselben didaktischen Griinden wird die Entstehung der Gestirnkorper 38 C 3 ff
vor ihrer Belebung behandelt, die erst in 38 E 3 ff erwahnt wird 12.
Mit anderen Worten, die Darstellung der Weltentstehung im
''Timaios" ist durch didaktische Absichten bestimmt; so, wie sie ist,
ist sie OtOacrKaA.ia~ xaptv bzw. craqnweia~ EVEKa. Das haben schon
Speusipp und Xenokrates gesehen. Doch sind diese noch einen
Schritt weitergegangen. Sie haben behauptet, nicht nur die Darstellung der Weltentstehung, sondern auch der dargestellte Akt sei
nichts weiter als ein Zugestandnis an die OtOacrKaA.ia. Wie in der
Geometrie zum Zweck der Unterweisung und der Verdeutlichung
ein Dreieck aus vorgegebenen GroBen konstruiert werde, das doch
niemals entstanden sei, so lasse Platon im ''Timaios" den Kosmos
nur zum Zweck der Unterweisung und urn der groBeren Deutlichkeit willen durch den Demiurgen konstruiert werden. In WirklichVgl. Robinson, Plato's Psychology 65. 68.
Ebenso wird im zweiten Teil des "Timaios" zuniichst die weniger
Vorwissen erfordernde xropa behandelt ( 48 E 2 - 53 B 7), dann erst die
"ungewohnliche Darstellung" (aft611~ A.6yo~) der Elementenlehre, die
erhebliche mathematische Vorkenntnisse voraussetzt (53 B 7 ff).
II

12

82

MATTHIAS BALTES

keit sei der Kosmos ebensowenig entstanden wie die geometrische

Figuri3.
Was ist von dieser Interpretation zu hal.ten? In der Tat fiihrt die
Erzahlung der Weltentstehung, wenn man sie als Schilderung
eines sukzessiven Geschehens auffaBt, zu uniiberwindlichen
Schwierigkeiten, von denen ich nur einige aufzahle:
1. Woher kommt der Gott, der auf der zweiten Stufe so plotzlich
auftaucht? Auf der ersten Stufe wurde er unter den Prinzipien ja nicht
erwahnt. Wenn er aber, woran nicht zu zweifeln ist, Prinzip ist, dann
muB er auch schon auf der ersten Stufe existiert haben; dann aber
kann er nur mit dem auf der ersten Stufe genannten ov (52 D 3)
identisch sein. Nun ist aber mit diesem ov nach 48 E 2 ff das ideale
Vorbild gemeint. Also bleibt nur die SchluBfolgerung, daB der Gott
und das ideale Vorbild identisch sind.
2. Bestiitigt wird dies durch die Aussage, daB das VOTitOV s<Pov alle VOT'Ita
scpa. in sich enthalt (30 C 5 ff; 31 A 4 f). Wenn der Demiurg, was nicht
bezweifelt werden kann, ein VOTitOV scpov ist, dann gibt es nur die
Alternative, daB er entweder ein Teil oder das Ganze des VOTitOV s<Pov
ist. Ware er ein Teil, so ware er unvollkommen (30 C 4 f) 14 Da das
undenkbar ist, kann er nur das Ganze des VOTitOV s<Pov sein; denn er
ist ja~ "~a~ be~te der i~t~lligiblen und immerseienden Dinge" (t&v
VOTitO>V a.Et tE ovtO>V ... to a.ptcrtov, 37 A 1).
DaB der vou~. der die ava:yK'Il iiberredet, der vou~ des Gottes ist, den
Timaios 36 D 8 und 39 E 7 im Voriibergehen erwahnt, kann man vermuten, es wird aber nicht gesagt. Wie er zu deuten ist, werden wir
spater zu klaren versuchen. So vie! scheintjedoch klar, daB die vorher
genannte zweite Stufe der Kosmogenese nicht einfach damit erklart
werden kann, daB plotzlich ein Demiurg auftritt, der vorher nicht da
war.
3. Wohin zieht sich der Gott zuriick, nachdem er sein Werk auf der
fiinften Stufe beendet hat? Ja, zieht er sich iiberhaupt zuriick? Zwar
scheint die Aufgabenverteilung zwischen Gott und Gottern vollkommen klar zu sein; der Gott will mit der Erschaffung des sterblichen
Teils der Menschenseele und des menschlichen Korpers, aber auch
mit der Erschaffung der Tiere nichts zu tun haben ( 41 B 7 - D 3; 42 D 2
- E 4). Gleichwohl wird die Trennung im Folgenden nicht aufrechterhalten; denn an die Stelle der Gotter tritt haufig der Gott (46 C 7;
E 8; 47 B 6; 71 A 7; E 2 f; 73 B 8; 74 C 6; D 6; 75 D 1; 76 C 6; 78 B 2; 80 E 1;
92 A 3). 1st der Gott also doch nicht abgetreten? Sind der Gott und die
Untergotter etwa gar identisch? Sind der Demiurg und die Untergotter vielleicht nur Hypostasen fiir einzelne Stufen des Handelns des
Gottes?
13 Vgl. Speusipp, fg.54 Lang = fg.61 Tarin = fg.94 f Isnardi Parente;
Xenokrates, fg.54. 68 Heinze = fg.153 ff. 188 Isnardi Parente; Baltes,
Weltentstehung I 18 ff.
14 Dieses Problem hat schon der Platoniker Attikos gesehen; vgl. fg.34
des Places.

1ST DIE WELT REAL ENTSTANDEN ODER NICHT?

83

4. Kann man wirklich annehmen, daB der Gott die Weltseele vor dem
Weltleib erschafft, gleichsam eine Seele ohne Aufgabe? Denn diese
Aufgabe erhalt sie ja erst nachtraglich dadurch, daB der Gott den
Weltkorper in sie hineinbaut (36 D 9 f; 38 C 7 f; 40 A 4 f) 15.
5. Kann man wirklich annehmen, daB die Zeit erst mit dem Kosmoskorper entstanden ist? 16 1st dem so, dann waren die vorkosmischen
Bewegungen nicht "in der Zeit". Waren sie dann iiberhaupt noch
Bewegungen? Welchen Sinn hat es dann noch, daB Timaios davon
spricht, daB sie "vor der Entstehung des Himmels" waren (48 B 3 f; 52
D 3; vgl. 37 E 2) und daB der Urzustand und das Eingreifen des Gottes
als zeitliches Nacheinander dargestellt werden (53 A 7 ff; 69 B 2 ff; 30
A 2 fl) 17?
Die Interpreten helfen sich hier mit der Annahme einer ungeordneten vorkosmischen Zeit (so schon der Platoniker Attikos) 18 bzw. mit
einer unendlichen Dauer ("an a1tEtpov of duration"), die nicht periodische Zeit sei, weil sie sich nicht wie diese gemaB der Zahl bewegt
(so Skemp und Hackforth) 19 . Aber kann es in dieser "Dauer"
iiberhaupt ein Friiher und Spater geben, wenn es kein Majl gibt,
woran beides gemessen wird?
Vlastos nimmt analog zu den Spurenelementen "Spuren einer
zeitlichen Ordnung" ('"traces' of temporal order") im vorkosmischen
Dagegen schon Prokl. In Plat. Tim. I 287,11 ff; 104,4 ff; 115,2 f Diehl.
Es ist falsch zu behaupten, daB, "when Plato says that the Demiur~
makes time, he means that the Demiurge makes a clock, nothing more
(Mohr, Platonic Cosmology 54); denn die Zeit ist zwar ihrem Wesen nach ein
bewegtes und der Zahl nach fortschreitendes Abbild der Ewigkeit (37 D 6 f;
38 A 7 f), aber sie ist keine Uhr; denn eine Uhr ist nicht Zeit, sie mi}Jt die
Zeit. Die Interpretation Mohrs ist nicht unvoreingenommen; denn sie geht
davon aus, daB es auch vor dem Kosmos eine 'Zeit' gibt: Platon "means that
time is a clock, a clock by which we measure time, where 'time' here is used
in a colloquial sense, as that about motion and rest which is measurable"
(ebd. 59). Platon kennt diese zweite An von 'time' nicht, und es stimmt
nicht, daB "the technical sense of time and this ordinary sense exist side by
side through the text we are discussing (e.g., 38c2: o o' ... xpovov [?])" (a.O.).
Grotesk wird es, wenn neben dieser ersten Art von 'time' eine zweite
eingefiihn wird, namlich 'time' im Sinne von dem, "which is measurable
as actually having been measured" [Hervorhebung von mir], und wenn gesagt
wird, diese Bedeutung von 'time' "appears in the description of the planets as
'the instruments of time(s)' ( 42d5; 41e5)" (a.O. und 64).
Ganz unmoglich ist die Behauptung Mohrs ebd. 67 f, die vom Demiurgen
geschaffene Zeit unterlie~e keinem Wandel und sei ewig. Platon sagt doch
deutlich, sie sei eine Eilcrov 1C1V1lt6r; ttr; airovor; (37 D 5) und ihre do11 Vergangenheit und Zukunft seien Bewegungen (38 A 2). Weder als bewegt noch
als Abbild der Ewigkeit kann sie selbst wirklich ewig und unveranderlich
sein. Die Zeit ist zwar nicht "in der Zeit", aber das heiBt noch lange nicht,
daB sie wirklich ewig (im Sinne des Vorbildes) ist. Siehe unten Anm.33.
17 Vgl. zudem Tarin, "Creation Myth" 378 ff.
18 Plutarch und Attikos bei Prokl. In Plat. Tim. I 276,32 ff; 286,26 ff; III
37,11 ff Diehl; dazu Sorabji, Time 273 f (dagegen Prokl. In Plat. Tim. I 286,29
ff Diehl; s. Baltes, Weltentstehung II 40 ff). Vgl. schon Epikureer bei Cicero, De
nat. deor. 1,21 und dazu Baltes, Weltentstehung I 26.
19 J. Skemp, The Theory of Motion in Plato's Later Dialogues (Amsterdam
19672) 111; Hackforth, "Plato's Cosmology" 22; ahnlich Guthrie, History V
301 fund Mansfeld, "Rez. Scheffel, Kosmologie" 172.
15
16

84

MATTHIAS BALTES

Chaos an2. Doch von den letzteren sagt der Platontext nichts. An
anderer Stelle2 1 unterscheidet Vlastos die geschaffene Zeit als einen
"einformigen und meBbaren ZeitfluB" ("uniform and measurable
time-flow") von einer vorkosmischen "unumkehrbaren zeitlichen
Abfolge" ("irreversible temporal succession"), einer "quasi-time"22. In
der letzteren gebe es zwar ein Friiher und Spater, aber die Intervalle
zwischen heiden seien nicht meBbar: "If A, B, C, are successive
instants we would have no means of telling if the interval, AB, is as
long as the interval, BC, or longer, or shorter. "23 Ich mochte bestreiten,
daB es iiberhaupt sinnvoll ist, in einem vorkosmischen Chaos von
einem Friiher und Spater zu sprechen. Ein Friiher-und-Spater impliziert in jedem Fall eine zahlbare Abfolge, also ein Voranschreiten der
Zahl nach (Kat' apt8J.I.OV ivat, 37 D 6)2 4 ; d.h. es impliziert eine
Ordnung, die es im vorkosmischen Chaos nicht gibt. Wer immer
also ein Friiher oder Spater im vorkosmischen Chaos konstatiert, kann
selbst nicht zu diesem gehoren, vielmehr betrachtet er das Chaos
gleichsam von auBen, von der Ordnung seines Erkenntnisorgans aus;
er ilbertriigt also etwas auf das Chaos, was nicht in ihm ist.
Auch darf man m.E. nicht hingehen und, wie Vlastos es tut, Vergangenheit, Gegenwart und Zukunft fiir die genannte zeitliche Abfolge ("temporal succession") in Anspruch nehmen2 5 . Denn Platon
bezeichnet diese doch ausdriicklich als xpovou yEyovota EtOTJ, d.h. als
Unterarten der Zeit (37 E 4)2 6 , die zusammen mit dieser entstanden sind:
xp6vou tllUtll ai&va JllJ.I.OUJ.I.EVOU Kilt Kilt' apt8J.I.OV KUKAoUJ.I.EVOU YEYOVEV
dO, (38 A 7 f).
Jedes Postulat einer wie auch immer gearteten Zeit ist fiir das vorkosmische Chaos abzulehnen; denn die einzige Zeit, von der Platon
spricht, ist die Zeit, die "zusammen mit dem Himmel entstanden ist"

und die zusammen mit dem Himmel vergehen kann (38 B 6 f); d.h.
ebensowenig wie es nach einem Vergehen des Himmels Zeit geben
wiirde, hat es vor dem Entstehen des Himmels Zeit gegeben. Die
geordnete Bewegung des Himmels ist ja die Bedingung der
20 Vlastos, "Disorderly Motion" 390. Mansfeld, "Rez. Scheffel, Kosmologie"
171/172 spricht in iihnlichem Sinne von "a faint sort ... of 'matter' of time"
bzw. von "a sort of precosmic time-'matter'".
21 Vlastos, "Creation" 409 ff
22 Vlastos, "Creation" 414. Andere meinen, Platon nehme zwar keine
vorkosmische Zeit, aber eine vorkosmische "duree" an; vgl. Naddaf, L'origine
437 Anm.l. Nichts im '"fimaios" weist auf eine solche hin.
23 Vlastos, "Creation" 411. Ahnlich urteilt Mohr, Platonic Cosmology 58 ff.
63. 64 ff.
24 Das haben selbst Plutarch und Attikos zugegeben; vgl. Prokl. In Plat.
Tim. I 277,6 f Diehl.
25 Vlastos, "Creation" 411 ff
26 Strenggenommen bezeichnet Platon nur "war" und "wird sein" als
Arten der Zeit, aber das schlieBt nicht aus, daB auch das "ist" eine Art der
Zeit ist; denn Platon sagt nicht, daB nur die beiden Arten der Zeit seien,
vielmehr daB sie nur von der Zeit, nicht von der Ewigkeit ausgesagt werden
durfen, fUr die nur das "ist" zuliissig ist (Tim. 37 E 4 fi). DaB dies richtig ist,
zei~t Tim. 38 c 2 f: b ~ (l~ ~l(X tEAoU~ 'tOV 1i7tllV'tll x.p6vov 'YE'YOVOO~ 'tE Kilt rov Kilt
cro~tEVO~. Vgl. C. Eggers Lan, Las nociones de tiempo y eternidad de Homero a Platon
(Mexiko 1984) 177 f.

1ST DIE WELT REAL ENTSTANDEN ODER NICHT?

85

Moglichkeit fiir die Existenz von Zeit (38 C 3 ff; 38 E 4 f; 41 E 5; 42 D 5),

und die kosmische Zeit ist "die ganze Zeit", wie Platon ausdriicklich
betont (36 E 5; 38 C 2). Dementsprechend verwendet Platon den
Begriff xp6vo~ nie, wo er vom vorkosmischen Zustand redet. 1st aber
Zeit nur im kosmischen Zustand moglich, dann kann Platon keine
Schopfung im zeitlichen Sinne gemeint haben.
Ferner ist folgendes zu bedenken: Wenn die Zeit erst. mit dem
Kosmoskorper entstanden ist, dann ist die Weltseele, die ja vor ihm
geschaffen wurde, weder in der Zeit noch mit der Zeit entstanden,
noch erfolgen ihre Bewegungen "in der Zeit", obschon doch eben
diese Bewegungen, wenn die Gestirne in sie hineingesetzt werden,
die Zeit erzeugen und ihre Bewegungen dann 7tpo~ tov auJ.17tavta
xp6vov erfolgen (36 E 4 f). Sind aber nicht einmal die Bewegungen der
Weltseele unmittelbar nach ihrer Erschaffung in der Zeit, dann ist erst
recht ihre Erschaffung nicht in der Zeit erfolgt.
6. Kann man wirklich glauben, daB Platon einen Zustand angenommen hat, in dem es auf Erden nur Manner gab, keine Frauen und
keine Tiere? Diese Frage stellt sich urn so dringlicher, als die
menschliche Natur ( i] av8pco7ttVTJ q>uat~) nach Platon zwiefaltig ist
(Bt7tA.il), d. h. aus Mann und Frau besteht (42 A 1 f). Die menschliche
Natur ware in ihrem ersten Stadium-ohne die Frau-also unvollstandig. Erst durch die Verfehlung des Mannes wiirde sie vollstandig
(42 B 5 f). Zudem muB man fragen, wie denn diese ersten Manner
entstanden sein sollen, wenn nicht aus einer Frau oder Frauen?27
Aber Frauen gab es ja noch nicht2B.
7. Kann man wirklich annehmen, daB der Gott bei der Erschaffung der
Weltseele "Reste" iibriglaBt, aus welchen er anschlieBend die Seelen
der Gotter und Menschen schafft (41 D 4 ff)? Oder daB die Gotter bei
der Erschaffung des menschlichen Korpers erst einmal einen Vorrat
an Knochenmark, Sehnen und Fleisch erschaffen, urn diese erst
danach zu verwenden (73 B 1-75 D 5)?
8. Darf man glauben, daB der Demiurg etwas Unvollkommenes schafft?
Das tut er, indem er nur den Kosmos und die unsterblichen Wesen in
ihm hervorbringt, die Erschaffung der sterblichen Wesen aber den
Untergottern iiberliiBt. Sagt er doch selbst, daB der Kosmos ohne die
sterblichen Wesen "unvollkommen" (atEATt~) ist (41 B8).

Diese und andere Schwierigkeiten machen die Annahme unmoglich, daB der Kosmos nach Platon sukzessive entstanden ist.
Darf man also daraus schlieBen, daB er iiberhaupt nicht entstanden ist? Das zu behaupten ware voreilig; denn es bleibt ja noch
die Erklarung, er sei in einem realen Schopfungsakt sozusagen auf
27 Platons erste Menschen, die Manner-Generation, werden zwar von den
Untergottern erschaffen, also nicht geboren (42 E ff; 69 C ff), aber wie
entsteht die zweite Generation? Oder praziser: Wie entsteht die erste Frau in
der zweiten Generation?
28 Vgl. Prokl. In Plat. Tim. III 282,27 ff; ~93,32 ff Diehl; dazu Baltes,
Weltentstehung II 199 ff. Zur Ambiguitat der AuBerungen Platons tiber die
Entstehung des Menschen vgl. Naddaf, L'origine 393 ff.

86

MATTHIAS BALTES

einen Schlag entstanden29. Aber auch diese Annahme bereitet

uniiberwindliche Schwierigkeiten, die sich zum Teil mit den
schon genannten decken:
1. Platon deutet an keiner Stelle an, daB er eine solche Vorstellung im
Sinn gehabt hat, weder im "Timaios" noch in seinen anderen
Dialogen 30.
2. Timaios sagt 52 A 3: Das, was sich immer gleich verhalt-gemeint
sind die Ideen-das geht nirgendwohin (OUtE de; aA.A.o 7t0l i6v). Wie
ist es damit vereinbar, daB die aJlEptcrtoc; oucricx, das aJlEpu:stov tcxut6v
und das nJlEptcrtov Oatepov, d.h. Ideen, in die Weltseelenmischung
eingehen (35 A)?31
3. Wenn es zum Wesen des VOlltOV scpov gehort, 7tCXpa~EtYJ.1CX zu sein,
dann hat das VOlltOV si!>ov vor dem Kosmos sein Wesen verfehlt; denn
dann wares Vorbild von nichts. 1st das aber undenkbar, dann muB es
das Abbild dieses Vorbildes, den Kosmos, immer gegeben haben 32.
DaB Platon dies wirklich meint, zeigt ein unscheinbares Wortchen in
37 D 7, wo er die Zeit eine cxirovwc; eiKrov, "ein ewiges Abbild" der
Ewigkeit nennt 33 . Nun ist aber die Zeit ebenso existentiell mit dem

29 Vgl. Scheffel, Kosmologie 59. 73 f. 122 f. 130 f. 140 ff. Derselben Ansicht
waren auch manche christlichen Schriftsteller hinsichtlich des Schopfungsberichts der Genesis, beispielsweise Basileios, In Hexaem. 1,6 S.110-112 Giet;
Ioh. Philoponos, De aet. mundi 1,3 S.6,2 ff; 11,13 S.367,19 ff u.o. Rabe.
30 Bei einer schlagartigen Schopfung wiirde der Demiurg des "Timaios"
zudem in ein und demselben Augenblick auftreten und abtreten.
31 Selbst Hackforth, "Plato's Cosmology" 20, der doch der Ansicht ist, daB
die Weltentstehung im "Timaios" im Sinne eines zeitlichen Aktes zu
verstehen sei, halt es fiir "ridiculous" anzunehmen, auch die Mischung der
Weltseele sei als ein einmaliges Geschehen in der Vergangenheit aufzufassen; diese Mischung sei vielmehr nichts weiter als "an analysis of the
cosmic soul's faculties of cognition and motion".
32 Vgl. Prokl. bei Ioh. Philoponos, De aet. mundi II S.24,1 ff Rabe; Baltes,
Weltentstehung I 214 mit Verweisen; II 136. Dieselbe Argumentation IieBe
sich iibrigens auch auf den Demiurgen (vgl. Prokl. a.O. III S.42,1 ff Rabe;
Baltes a.O. II 136 ff) oder die Zeit anwenden (Prokl. a.O. V S.103,20 ff Rabe;
Baltes a.O. II 140).
33 Mit Recht schreibt Tarin, "Creation Myth" 397, Anm.60: "Since the
word airovto~ is applied to the model and to the copy, it cannot mean in the
case of the latter only imperishability: there must be something common to
the model and the copy in order that both may be said to be airovto~. And
Timaeus 37 C 6 - 38 B 5 implies the answer: both model and copy are airovto~,
but whereas the former is et!!rnal and atemporal, the latter possesses the
eternity of infinite duration." Ahnlich urteilt Eggers Lan, Eternidad 175 ff.
Platon selbst ist noch expliziter: Es war nicht moglich, die airovto~ q>'ucrt~
(d.h. den airov) vollkommen auf den Kosmos zu iibertragen (7t<XVttl..ro~ 7tpocra7ttttv}; d.h. der Kosmos konnte nicht schlechthin airovto~ werden; aber ein
KompromiB war moglich: Der Kosmos konnte die til<rov airovo~ anfnehmen,
die Zeit, die selbst wieder airovto~ ist: wei! sie ja nicht "zeitlich" (eyxpovo~)
sein kann. Airovto~ mull auch im letzteren Fall so etwas wie "ewig" heiBen,
d.h. anfangs- und endlos, wei! ja nur so das Ziel erreicht wird, daB die Zeit
der Ewigkeit "moglichst ... ahnlich ist" ('{v' ro~ OJlOtO't<X'tO~ autcp K<Xta OUV<XJltV TI
38 B 8 f; vgl. 37 D 2).
DaB Platon wirklich annimmt, die Zeit sei anfangs- und endlos, zeigt die
Parallelitiit des Satzbaus in Tim. 36 C 1-3:

1ST DIE WELT REAL ENTSTANDEN ODER NICHT?

87

Kosmos verbunden wie die Ewigkeit mit dem Vorbild. Wenn also die
Zeit ein 'ewiges' Abbild der Ewigkeit ist, dann muB der Kosmos ein
'ewiges' Abbild des Vorbildes sein. Folglich ist er nicht entstanden.
4. Spricht schon die Bezeichnung der Zeit als a.irovw<; dKrov gegen _eine
Entstehung derselben, so wird diese Ansicht durch folgende Uberlegung untermauert: Vergangenheit und Zukunft sind Arten der Zeit
(xpovou ... Et&r], 37 E 3- 38 A 8); d. h. sie sind EtOTJ des yvo<; Zeit. Wenn
die Zeit nun gleichsam mit einem Schlag ins Sein gesetzt worden
ware, dann ware sie unvollstandig ins Sein gesetzt worden; denn in
diesem ersten Augenblick hatte es noch keine Vergangenheit gegeben, sondern nur Gegenwart und Zukunft. Das aber ist unmoglich,
denn
a) der Gott hatte in diesem Fall etwas Unvollkommenes geschaffen34,
was undenkbar ist, und
b) es ist unvorstellbar, daB das yf.vo<; Zeit einmal das doo<; Vergangenheit nicht umfaBt.
Also kann die Zeit nicht entstanden sein 35
5. Wenn die Welt das schonste aller entstandenen Dinge und der Gott
die beste aller Ursachen ist (29 A 5 f) und wenn der gute Gott keinen
Neid kennt und alles moglichst gut mach en (29 E 1 ff), ja, wenn er
iiberhaupt in seinem Handeln immer das Schonste will ( 30 A 7 f),
warum hat er dann die Welt nicht schon vorher erschaffen? 36 Dies
gilt urn so mehr, wenn es vor dem Anfang der Welt das vorkosmische Chaos der ersten Stufe gab. Wie konnte der Gott dieses
iiberhaupt mit ansehen, ohne einzugreifen?37 Man wende nicht ein,
daB es nicht sinnvoll sei, vor der Entstehung der Zeit von einem Vor
zu sprechen. GewiB, es gab kein zeitliches Vor, aber doch ein
sachliches: Unordnung vor Ordnung bzw. das Nichts vor dem Sein.
Wie ist das bei einem guten Gott moglich?
6. Gott ist "immer seiend" (34 B 8), ja, er ist das Beste der intelligiblen
Dinge (37 A 1), d.h. er ist keinem Wandel unterworfen (27 D 6- 28 A
2). Wieso kann er sich dann iiberhaupt plotzlich zur Erschaffung der
Welt entschlieBen? Ein solcher EntschluB wiirde doch notwendiger-

to !lEV ... nc:xpa0EtY!1C:X n-avm aimva ecmv OV'

0 o' c:xi) ... TOY an-avm xp6vov yeyovroc. t l((lt rov KC:Xt EO"OUEVOC..
Ewigkeit und Zeit existieren parallel.
Anzunehmen, die Zeit sei nur in einer Richtung ewig, wiirde 1. die
Ewigkeit zeitweise ihrer Ursiichlichkeit berauben und 2. sozusagen ein
amputiertes Abbild erzeugen. Vgl. Proklos bei Ioh. Philoponos, De aet.
mundi XVIII S.607 ,24 ff Rabe (Baltes, Weltentstehung II 162) und auch
Cherniss, Aristotle's Criticism 419, Anm.350.
34 Spiiter tadelte man dergleichen als T1!11tEA~<; 01]!1\0Upyic:x; Ioh. Philoponos,
De aet. mundi 9,10 S.344,15 f.
35 Vgl. in diesem Zusammenhang auch Cherniss, Aristotle's Criticism
419f.
36 Vgl. Porphyrios bei Ioh. Philoponos, De aet. mundi 6,27 S.224,18 ff
Rabe = fg.456a Smith; Porphyrios bei Prokl. In Plat. Tim. I 394,12 ff Diehl;
Proklos selbst: In Plat. Tim. I 367,2 ff. 13 ff Diehl; dazu Baltes, Weltentstehung I
90. 92. 159. 160 f; II 74 ff; Proklos bei Baltes, a.O. II 134 ff.
37 Vgl. Tarin, "Creation Myth" 381; Baltes, Weltentstehung I 215 mit
Verweisen.

88

MATTHIAS BALTES

weise einen Wandel implizieren 38 . Wie kann er sich nach der

Erschaffung der unsterblichen We sen und Dinge zuruckziehen ( 42 E
5 f)? Eigentlich kann er sich doch our immer gleich verhalten; denn
er gehortja zu den ad KCXta tCXUta Kat ffioa{moc; exovta (29 A 1; 35 A 2;
37 B 3; 38 A 3; vgl. 28 A 6 f; 52 A 1). Also kann er sich auch zur
Schopfung our immer gleich verhalten.
7. Wenn der Gott "immer seiend" ist, dann gibt es fUr ihn keine
Vergangenheit und keine Zukunft. Dann kann er auch nicht uber
"den Gott, der einmal sein wird", d.h. den Kosmos, nachdenken (34
A 8 f), zumal die Zeit noch gar nicht existiert. 1st aber dieser Akt
unmoglich, dann ist auch die einmalige Schopfung unmoglich 39 .
8. Wo befindet sich der Gott vor der Weltentstehung? Wo nach seinem
'Abtritt' (42 E 5 f)? Wie verhiilt er sich in die sen heiden Stadien zum
VOTJtOV siiJov? Blickt er auf es? Aber wozu? Blickt er nicht darauf?
Wohin dann?
Wenn man beachtet, daB Platon sowohl im vorkosmischen (52 D 1 ff)
als auch im kosmischen Zustand (50 C 7 ff; vgl. 51 E 6 ff) our drei
Wirklichkeiten anerkennt, das Seiende (to ov, ta ovta, d.h. die
Ideen), den Raum ( nSJlvTJ, xropa) und das Werden bzw. die werdenden Dinge (yevmtc;, to ytyvof.LEVov), wenn man ferner beachtet, daB im
kosmischen Zustand das Seiende als Wirkursache urul paradigmatische Ursache bezeichnet wird (to o9ev a<pOflOlOUflEVOV <pUEtat to
ytyvof.Levov, 50 D 1 f) 40 , dann drangt sich mit Notwendigkeit der
SchluB auf, daB der Demiurg sowohl im kosmischen als auch im
vorkosmischen Zustand mit dem idealen Vorbild identisch ist; d.h. er
tritt our im Zustand der Kosmogenese gleichsam als eigenes Wesen
aus dem idealen Vorbild heraus und kehrt anschlieBend wieder in
dieses zuriick. Da aber ein Wandel, wie gesagt, im "immer Seienden" undenkbar ist, und da, wie Platon ausdrucklich sagt, das Seiende
nichts in sich aufnimmt (52 A 2 f), ist es auch unmoglich, daB das
Auftreten und Abtreten des Gottes real zu verstehen ist. Dann aber
kann der Demiurg-pace T.M. Robinson 41 -nichts anderes sein als
der schopferisch-ordnende Aspekt des Seienden, so wie das Vorbild
der paradigmatische ist. Dieser schopferisch-ordnende Aspekt wirkt
38 Vgl. Prokl. In Plat. Tim. I 288,14 ff Diehl; dens. bei Ioh. Philoponos,
De act. mundi XVIII S.604,14 ff Rabe; s. Baltes, Weltentstehung I 213 mit
Verweisen; II 43 ff. 159 ff.
39 Sucht man cine Erkliirung, dann ist die des Proklos, In Plat. Tim. II
101,3 ff Diehl noch immer die verniinftigste: "Als 'Gott, der einmal sein
wird', hat Platon den in seiner Darstellung (ev tip A.6yc:p) Teil fiir Teil
entstehenden Gott bezeichnet; denn der Gott liiBt die Teile und das Ganze
mit einem Mal (&ep6roc;) entstehen, die Darstellung (o A.6yoc;) hingegen
zerlegt das, was zusammen entstand, liiBt das entstehen, was unentstanden
ist, und zerteilt das, was immerwiihrende Dauer besitzt, der Zeit nach. 'Der
Gott, der einmal sein wird', ist also offensichtlich ein solcher mit Riicksicht
auf die Darstellung (tip A.6yc:p), die mit Analyse und Synthese arbeitet." Vgl. dazu
Baltes, Weltentstehun(5 II 100. 101 f.
40 Vgl. Kat o~ Kat 7tpocremicrat 7tpE7tEt ... to o' oeEv 7ta.tp{, 50 D 1 f; das Seiende
ist das Priigende: 50 C 5; D 4. 6.
41 Robinson, Plato's Psychology 102. Wenn es den Gott zum Gott macht, bei
den Ide en zu sein (Phaidr. 249 C 5 f), dann versteht man, daB der Demiurg
der hochste Gott ist, weil er in gewisser Weise die Ideen ist.

IST DIE WELT REAL ENTSTANDEN ODER NICHT?

89

aber auch im Kosmos. Denn Pia ton sagt ja ausdriicklich, daB dort, wo
der Gott abwesend ist, Unordnung herrscht (53 B 3 f) 42. 1st somit die
Gegenwart des Gottes unabdingbare Voraussetzung fiir jede Ordnung,
dann auch fiir den Kosmos. Der Gott kann also nicht abtreten, soli der
Kosmos bestehen bleiben 43.
Bestiitigt wird das durch eine wunderliche Aussage des Timaios, die
den modernen Interpreten einiges Kopfzerbrechen bereitet hat.
Timaios sagt 42 E 5 f: Nachdem der Gott den Untergottern seine
Auftriige fiir das weitere Schaffen erteilt hatte, "verharrte er in der
seiner Art entsprechenden gewohnten Haltung/Seinsweise" ( EJlEVEV
f.v tip f.a.utou Ka.ta tpo7tov i18Et). Hier driingt sich doch geradezu die
Frage auf, ob sich der Gott vorher-wiihrend seiner Schopfertiitigkeit
-nicht "in der seiner eigenen Art entsprechenden gewohnten
Haltung/Seinsweise" befunden hat. Da man diese Frage nicht
bejahen kann, hat der Gott immer "in der seiner Art entsprechenden
gewohnten Haltung/Seinsweise verharrt". Das lmperfekt EJlEVEV, das
den Interpreten solche Schwierigkeiten bereitet hat, vermag diese
Interpretation zu stiitzen.
Auch folgendes ist zu bedenken: Wie soli das "intelligible Lebewesen" wirklich Lebewesen sein, ohne zu wirken? Also darf es wohl
als gesichert gelten, daB der Demiurg nichts anderes ist als der
schopferisch-ordnende Aspekt des Seienden, so wie das Vorbild sein
paradigmatischer Aspekt ist 44 . 1st dies richtig, dann versteht man
auch, warum Pia ton eigentlich nur zwei Ursachen gelten liiBt ( oo'
a.hia.c; Ei:on), die notwendige und die gottliche (to JlEV ava.yKa.'iov, to of:
8EtOV' 68 E 6 ff)' den vouc; und die UVcX')'K'Il (4 7 E 3 fi) 45.
Allerdings ist eine Einschriinkung zu machen: Der Gott ist nicht das
aktive Prinzip des Seienden schlechthin, sondern nur das aktiv
ordnende Prinzip. Aktiv sind ja auch schon die Ideen, die auf die xropa.
einwirken (50 C 4 f; 51 A 7 f; 52 D 4 ff) 46
Ein vorletztes Argument kann das bisher Gesagte stiitzen: Das aktiv
ordnende Prinzip des VOTjtOv t;<pov nennt Platon auch einfach vouc;. So
sagt er 39 E 7 ff, daB es der vouc; ist, der die vier Hauptideen der
Lebewesen im VOT]tOV t;ipov sieht und im Kosmos zu verwirklichen
strebt, und 47 E 5 ff hebt er hervor, daB es der vouc; ist, der die ava')'K'Il
iiberredet. An heiden Stellen wird nicht erkliirt, wem der vouc; gehort.
42

Vgl. Plat. Polit. 272 D 6 - 273 E 1.

43 Vgl. Proklos bei Ioh. Philoponos, De aet. mundi XVIII S.606,9 ff; 608,11
ff Rabe; Baltes, Weltentstehung II 160 ff.

44 Daher kann man denn auch verstehen, daB Aristoteles dem Platonischen Demiurgen kaum Aufmerksamkeit schenkt (Cherniss, Aristotle's
Criticism 609 f; Tarin, "Creation Myth" 382. 391), ja, daB er sogar sagen
kann, Platon verwende nur zwei Ursachen, tft tE tou ti ecrttv Kat tft Kata tl]v
UAfiV, wobei er unter tft tou ti ecrttv die Ideen versteht; vgl. Arist. Met. A 6 988
a 7 ff.
4 5 Ebenso spricht er 48 E 3 ff bei der Rekapitulation von 27 D 5- 47 E 2 nur
von ouo EtOfl, dem napaony~ta und dem lltllfllliX napaoeiy~tatoc;, wahrend er den
Demiurgen vollig iibergeht. Offenbar sind auch hier Demiurg und Paradi~ma identisch.
6 Es stimmt also nicht, daB "the Platonic Ideas are to be viewed as
standards and as nothing else" (Mohr, Platonic Cosmology 55; Hervorhebung von
mir).

90

MATTHIAS BALTES

Natiirlich hat man, wie schon gesagt, an den vou<; des Demiurgen
gedacht, von dem einmal die Rede ist (36 D 8), aber das ist nicht
mehr als eine Vermutung. Nach unserer Interpretation ist dieser vou<;
der demiurgisch schaffende und ordnende Aspekt des intelligiblen
scpov, der mit dem Demiurgen identisch ist.
Als letztes Argument sei schlieBlich die Beobachtung von Theodor
Ebert angefiihrt, daB im sog. 7tpOOtJltOV (27 D 5-29 D 3), ja, im ganzen
'Timaios" nie bewiesen wird, daB es einen Baumeister (~Jltoupy&;) der
Welt gibt, sondern nur, daB es eine Ursache (atnov) fiir sie gibt 47.
Diese Ursache wird unmittelbar nach ihrer ersten Nennung 28 A 4-6
in die Aspekte Demiurg und Vorbild zerlegt (28 A 6 - B 2), ohne daB
dafiir eine Begriindung gegeben wird. Dadurch entsteht sogar ein
Ieichter Bruch im Gedankengang, der den Leser offenbar aufhorchen
lassen und darauf hinweisen soli, daB die Ursache, von der Timaios
spricht, mehrere Facetten hat: Sie ist absichtsvoll planende ( ~Jlt
oupy6<;) und nach einem rationalen Plan (7tapaoetyJla) verfahrende
U rsache (atnov).
Nach all diesen Uberlegungen versteht man es besser, daB Platon
gleich zu Anfang sagt, schon den Schopfer und Vater des Ails zu
finden sei schwierig, und es sei vollig unmoglich, ihn allen Menschen zu verkiinden (tov JlEV ouv 7tOUJtTJV Kat 7tatepa touoe tou 7tavto<;
eupe'iv tE epyov KCll eup6vta ei<; 1tclVtCl<; aouvatov A.eyetv' 28 c 3 ff). Er ist
in der Tat schwer zu finden, wenn er sich im Zustand der Kosmosordnung gleichsam im gottlichen Vorbild versteckt und erst durch
das Hilfsmittel der Kosmogenese wirklich Konturen annimmt48
Dies also sind die Schwierigkeiten, die sich einstellen, wenn man
annimmt, die Welt sei in einem einmaligen Schopfungsakt entstanden. Sie alle sind nur dann zu losen, wenn man zugibt, daB
Platon auch den Schopfungsakt selbst nicht als zeitlich oder einmalig aufgefaBt hat. Wie vor allem das Beispiel des Demiurgen
zeigt, hat Platon in der Kosmogenese nur ein Hilfsmittel gesehen,
urn in dem immer gleichen und daher nur schwer durchschaubaren Geschehen im Kosmos sauber zu unterscheiden, was in
dieser Ordnung auf die gottliche und was auf die notwendige
Ursache zuriickzufiihren ist: Der gottlichen Ursache verdankt sie
alles Gute (29 E 1 ff), der notwendigen Ursache die Einschdinkungen, die der Verwirklichung dieses Guten auferlegt sind ( 48 A
3; 56 C 5 f; 75 A 7 ff; 77 A 1 f). In der gottlichen Ursache ist der
Aspekt des 7tapaonyJla von dem des OTJJ.ltoupy6~ zu unterscheiden.
4 7 T. Ebert, "Von der Weltursache zum Weltbaumeister", Antike und
Abendland 37 (1991) 51 f
48 Die Untergotter des "Timaios", die oft_mit dem Demiurgen zusammenzufallen scheinen, konnen nach diesen Uberlegungen nichts anderes
sein als 'Ausliiufer' des demiurgischen Aspekts des Seienden, die auf
unterster Ebene wirken.

1ST DIE WELT REAL ENTSTANDEN ODER NICHT?

91

Der erstere begriindet die Existenz der Dinge (52 D 2 fl), der letztere
ihre Ordnung (~tcx.KOaiJ.TJOt<;, 'ta~t<;, 53 A 7 ff; 69 B 2 fl).
Die Argumente, die ich bisher fiir die Auffassung vorgetragen
habe, die Welt sei nach Platons ''Timaios" nicht in einem einmaligen Akt entstanden, sind nur die starksten, die der ''Timaios"
zu bieten hat, und sie sind m.E. unwiderlegbar. 1st somit eine
einmalige Weltentstehung in grauer Vergangenheit undenkbar,
so stellt sich nun mit noch groBerer Dringlichkeit die Frage, wie
der entscheidende Passus 28 B 6 - C 2 zu verstehen ist.
Nimmt man den Beweisgang fiir das Entstehen der Welt 27 D 5 28 C 2 ernst, dann wird in ihm nur nachgewiesen, daB die Welt
ein ytyVOIJ.EVOV ist49 ; denn alle entscheidenden Verbalausdriicke
stehen in diesem Abschnitt im Prasens. Die einzige Ausnahme
bildet yf.yov EV (28 B 7) .-Auf den Ausdruck a1t' apxil<; 'ttVO<;
ap~aJ!EVO<; komme ich spater zuriick.-Man beachte nun, wie die
Begriindung fiir dieses yf.yovev wieder auf das Prasens rekurriert:
Die Welt ist entstanden, weil sie sichtbar, tastbar und mit einem
Korper ausgestattet ist und sich alles dergleichen als ytyvOJ!EVC:X Kat
yEVVTJ'ta/yEVTJ'ta gezeigt hatte.
Man wende nicht ein, wie man es zu tun pflegt, daB der
Ausdruck yEVVTJ'ta/yEVTJ'ta die Vergangenheit impliziere. Tate er es,
dann ware der Beweisgang nicht einwandfrei. Aber er tut das
ebensowenig wie bei den im gleichen Satz verwendeten Verbaladjektiven opa'tO<; 'KC:Xt a1t't0<; (28 B 7). feVVTJ'tn/yEVTJ'tn meint dem
Zusammenhang nach nichts anderes, als daB die unter ihm
subsumierten Gegenstande in der einen oder anderen Form die
Fahigkeit zum yEvvaaSat oder yiyvea8at haben5.
49 Dies ist die Auffassung vieler antiker Interpreten; vgl. Baltes, Weltentstehung I 86 f. 108 mit Anm. 116; 121 f. 128 f.
50 Vgl. R. Kuhner- F. Blass, Ausfii.hrliche Grammatik der griechischen Sprache.
Erster Teil: Elementar- und Formenlehre, II (Leipzig 1 1892, Hannover 1978) 288 f;
E. Bickel, "Die q>EUK'ta der Stoa bei Ausonius", Rhein. Mus. 86 (1937) 287 f; E.
Schwyzer, Griechische Grammatik I (Mtinchen 1939) 501.
Zu diesem Gebrauch des Verbaladjektivs vgl. 1tpt 8e&v opa-rrov mi yevv1)-rrov
( 40 D 4); YVV1J'rOV, TrEfPOP1JJlEVOV aei, ylyvOJlEVOV 'r EV 'ttVt 't01tCp ... (52 A 5 f). In
heiden Fallen ist nicht die Bedeutung "entstanden/geworden" intendiert.
Auffallig ist der Gebrauch in 28 B 1 f: on li' iiv d.; yqov6.; (sc. j3A.1tEt b
liTJ!ltoupyo.;), "fEVVTJ'tcp 1tapalieiy11an 1tpocrxpro11Evo.; ... Wenn hier yqov6.; und
"(EVVTJ'tcp dasselbe bedeuten, dann ist der Ausdruck tautologisch. Der Vergleich
mit 28 A 6 f zeigt jedoch, daB "(EVVTJ'tcp im Gegensatz zu 'tO K!l'tll 't!lU'tlX exov zu
verstehen ist; dann aber liegt die Bedeutung "dem Werden und Entstehen
unterworfen" naher als die Bedeutung "entstanden/geworden". Ambivalent
ist schlieBlich die Ausdrucksweise in 37 D 4, wo mit 'tcp "(EVVTJ'tcp entweder
"das entstandene" oder "das dem Entstehen unterworfene (All)" gemeint

92

MATTHIAS BALTES

Wenn das richtig ist, dann ist auch ye:yovev (28 B 7) nicht im Sinne
eines einmaligen Schopfungsaktes zu verstehen-fiir diesen hatte
man ja auch eher die Form E"(EVE'tO erwartet. reyovev besagt vielmehr, daB das Werden und Entstehen im Zustand der Vollendung ist: Die Welt ist als eine entstandene, yeyovro<; ecrnv, wie
Platon vom Kosmos selbst sagt (31 B 3)51.
DaB dies wirklich so zu verstehen ist, legt eine weitere Uberlegung nahe: Wenn alles, was sichtbar, tastbar und korperlich ist,
in einem einmaligen Akt entstanden ist, dann ist auch das
vorkosmische Chaos in dieser Weise entstanden52, das ebenfalls
sichtbar (30 A 3; 52 E 1), tastbar und korperlich ist (vgl. 30 A 3 ff; 52
E 1 fl) 53. Da das vorkosmische Chaos nach Platon aber offenbar nur
yevecrt<; im Sinne des standigen yiyvecr8a.t ist (52 D 3), folgt daraus,
daB auch die yevecrt<; des Kosmos nur im Sinne des standigen
yiyvecr8a.t zu verstehen ist. 1st dem so, dann versteht man auch
besser, warum 29 D 7 yevecrt<; und 'tO nav gleichgesetzt werden (Ot'
t]vnva. a.i'tia.v YEVeO"tV Ka.t 'tOOe 'tO nav 6 O"UVtO"'ta<; O"UVEO"'tTIO"eV; vgl.
E 4): Der Kosmos ist yevecrt<;.
Man wende nicht ein, das vorkosmische Chaos werde zwar als
yevecrt<; bezeichnet, nicht aber wie der Kosmos als yeyov6<;. Das ist
nicht wahr; denn 28 C 5 ff wird die' Moglichkeit erwogen, daB der
Demiurg die Welt nach einem entstandenen Vorbild (yeyovo<;
na.paOetyJla.) geschaffen hat. Dieses Vorbild kann nur das vorkosmische Chaos sein, in dem schon eine gewisse Ordnung der
Spurenelemente herrschte (52 E 5 ff). Nur diese kann mit dem
entstandenen Vorbild gemeint sein, denn sonst gab es ja nichts
(52 D 2 fl)5 4 .
Nun konnte man darauf hinweisen, daB Platon ja nicht nur
danach fragt, ob der Kosmos entstanden sei oder nicht, sondern
danach, ob er immer war ohne jegliche &.pxn oder ob er geworden
sei, beginnend von einer gewissen &.pxn. Hier scheint der Gegensatz klar und deutlich zeitlich zu sein. Darauf konnte vor allem das
lmperfekt ~v im Gegensatz zu dem Perfekt yeyovev hindeuten. Aber
sein kann.
5! Oder, wie Prokl. In Plat. Tim. I 290,23 ff Diehl es wortreich nachweist,
der Kosmos ist ad ytyYOflEVO~ UflU Kat YEYEV1)flEVO~. Vgl. dazu Baltes, Weltentstehung II 52 ff.
52 Vgl. Prokl. In Plat. Tim. I 283,31 ff Diehl und dazu Baltes, Weltentstehung II 30 ff; auch Zeller II 1, 793.
53 Beides ist Voraussetzung fiir das 30 A 3 ff und 52 E 1 ff geschilderte
vorkosmische Geschehen; vgl. Prokl. In Plat. Tim. I 387,10 ff Diehl.
54 Vgl. Mansfeld, "Rez. Scheffel, Kosmologie" 171.

IST DIE WELT REAL ENTSTANDEN ODER NICHT?

93

das lmperfekt ist iiberhaupt seltsam; denn nach dem Argumentationszusammenhang erwartet man ein crnv: Ist der Kosmos
immer oder ist er geworden? Also ist das lmperfekt nur kolloquial
zu verstehen wie auch andere lmperfekte im "Timaios" (29 E 1; 37
D 3). 1st dies richtig, dann lautet der Gegensatz: 1st die Welt immer
seiend oder das Ergebnis eines Entstehungsprozesses? Uber den
Beginn des Entstehungsprozesses ist damit noch nichts gesagt,
und auch die zweifache Betonung der apxfJ weist nicht unbedingt
in diese Richtung. Denn schon in der Antike hat man auf den
auffalligen Gegensatz apxilv xrov ou~h:uia.v-a1t' apxflc; nvoc;
ap~aJ.LEVoc; aufmerksam gemacht55 . Es scheint, daB Platon mit
diesen Formulierungen darauf hinweisen wollte, daB der Begriff
apxfJ zu den vieldeutigen Begriffen (7tOAAa.x&c; AeyOJ.LEVa.) gehort
und er apxfJ nur in einem ganz bestimmten Sinne verwendet. Da
die Methode der begrifflichen Differenzierung in Platons Akademie gepflegt wurde56, wuBte jeder Platonschiiler, daB auch der
Begriff apxfJ viele Bedeutungen hat und daB Platon hier nur auf
eine bestimmte apxfJ hinauswill (apxfJ nc;)' ohne allerdings klar
anzugeben, welche apxfJ er wirklich im Sinne hat. Aber immerhin war der Schuler davor gewarnt, apxfJ hier einfach in der
Alltagsbedeutung zu nehmen.
Aber weist nun nicht gerade das ap~aJ.LEVOc; ganz entschieden in
die Richtung der Alltagsbedeutung? Hier haben wir doch den
erwarteten Aorist. Aber was fiir einen Aorist? Keinen Indikativ,
der mit dem Augment als Vergangenheitszeichen allein mit
Sicherheit das einmalige vergangene Ereignis hatte ausdriicken
konnen. Hatte Platon geschrieben: ilp~a.to, so ware klar gewesen,
daB er den einmaligen Akt in der Vergangenheit gemeint hatte;
ebenso wenn er formuliert hatte: EYEVE'tO a7t' apxflc; nvoc; ap~aJ.LEVOc;;
dann namlich hatte das voraufgehende Tempus auch auf das
Partizip abgefarbt. Zusammen mit dem Perfekt yf.yovev besagt das
Partizip ap~aJ.Levoc; nicht mehr, als daB eine apxfJ dem jetzigen
Zustand voraufgeht, d.h. zugrunde liegt.
Auch der Hinweis auf 28 c 2 f: 'tcp o' a.?> YEVOJ.LEV:p <pa.J.LEV U7t' a.hiou
nvoc; avayKT\V dva.t yevf.cr8a.t hilft nicht wirklich weiter; denn auch
hier liegen keine finiten Verbformen vor, und strenggenommen
besagt dieser Satz nicht mehr als: "wenn er nun aber entsteht,
55 Prokl. In Plat. Tim. I 279,30 ff; 285,21 ff und Baltes, Weltentstehung I 143 ff
56 Vgl. Arist. Met. !J. 1/2; vgl. z.B. auch E. Heitsch, "Argumentation und

Psychagogie", Philologus 138 (1994) 219 ff.

94

MATTHIAS BALTES

dann entsteht er, wie wir gesagt haben, notwendigerweise durch

eine Ursache". Da der Ausdruck tip ... yevoJlEVq> das to ytyvOJlEVOV
von 28 A 4 aufnimmt, weist die Aussage nicht auf einen
einmaligen Akt, sondern auf den speziellen Fall hin 57 .
Damit ist klar, daB der Ausdruck yyovev (28 B 7) nicht im Sinne
eines zeitlichen oder einmaligen Aktes verstanden werden darf.
Was also ist mit ihm gemeint? Er will besagen, daB die Welt
Ergebnis eines Entstehungsprozesses ist, der von einer Ursache
abhangt, die sich zerlegen laBt in einen aktiv schaffenden, einen
ordnenden und einen paradigmatischen Aspekt. Diese Ursache ist
die apxi) seines Entstehens. 'Apxf) bezeichnet also nicht den zeitlichen Beginn des Entstehungsprozesses-eine solche Annahme
wiirde in ausweglose Schwierigkeiten fiihren, wie wir gesehen
haben-vielmehr bezeichnet es das Woher des Entstehens im
ursachlichen oder ontologischen Sinne.
Warum aber ist Platon im "Timaios" so anigmatisch? Warum sagt
er nicht frei heraus, was er denkt? Mochte er seine Leser hinters
Licht fiihren? GewiB nicht. Bevor ich eine Antwort zu geben
versuche, mochte ich auf folgende auffallige Erscheinungen
hinweisen: Platon laBt seinen Hauptunterredner Timaios immer
wieder darauf hinweisen,
1. daB die Darstellung des Gemeinten schwierig sei (27 D 3 f; 28 C 3 ff;

48 c 2ff);
2. daB es von seiner Seite eine groBe Anstrengung erfordere, seine
Gedanken adaquat zum Ausdruck zu bringen (27 D 3 f);
3. daB es unmoglich sei, die Darstellung vollkommen widerspruchsfrei
zu halten (29 C 4 ff)5B,ja, daB sie nicht einmal frei von Zufallen sei (34
c 2f);
4. daB die Mitunterredner des Timaios zwar hochgebildete Leute sind59
57 Auffallig ist, daB im Folgenden (28 C 5 - 29 B 1) keine Aoriste vorkommen, sondern nur Imperfekte und Perfekte, die die Dauer und den
Endzustand der Handlung festhalten und nicht auf die Einmaligkeit deuten.
Aoriste begegnen erst wieder ab 29 D 7 ff, d.h. vom Beginn des Schopfungsmythos an. Wie jede Erzahlung muB auch dieser Mythos selbstverstandlich
auf weite Strecken im Aorist erzahlt werden.
58 Dazu gehort etwa, daB Timaios sich gelegentlich unscharfer kolloquialer Ausdrucksweise bedient, so beispielsweise, wenn er 28 B 6 statt n:o'tEpov
ecrttv ad von n:on:pov !Jy_ aei spricht; desgleichen ist das Imperfekt in 29 E 1 und
37 D 3 unscharf gebraucht (vgl. Prokl. In Plat. Tim. I 278,29 ff Diehl und
Baltes, Weltentstehun~ I 133 f).
59 Der "Timaios ist der einzige Dialog Platons, in dem die Unterredner
einigermaBen gleichrangig sind (vgl. Tim. 19 E 8 ff; T. Szlezak, "Gesprache
unter Ungleichen. Zur Struktur und Zielsetzung der platonischen Dialoge",
Antike und Abendland 34 (1988) 99-116), und dennoch betont Timaios immer

IST DIE WELT REAL ENTSTANDEN ODER NICHT?

95

und ein gehoriges Vorwissen besitzen (53 B 7 ff), Timaios a her

trotzdem auf sie Riicksicht nehmen muB (27 D 2), ja, daB auch ihnen
'die letzten Dinge' in der gegenwartigen Form nicht mitgeteilt
werden konnen (48 C 2 ff); denn diese konnten nur Gotterlieblinge
erfahren, die die Zuhorer trotz ihrer Bildung offensichtlich nicht oder
zumindest nicht aile sind (53 D 6 f).
Ja, nicht nur 'die letzten Dinge' konnen ihnen nicht mitgeteilt
werden; wird doch an zahlreichen Stellen auf eine weitere Priizisierung auch auf niedrigerer Ebene ausdriicklich verzichtet (38 B 3
ff; D 6 ff; 40 C 2 ff; 59 C 5 ff; 86 B 8 f; 89 D 7 ff; vgl. 68 B 6 fl). Ferner
wird immer wieder auf die Schwache der menschlichen Erkenntnis hingewiesen, die eine exakte Darlegung unmoglich macht (29
C 8 f; 68 D 2 ff; 72 D 5 ff), und gelegentlich wird die Vorlaufigkeit
der getroffenen Entscheidung herausgestellt (54 A 7 ff; 55 D 4 ff).
So nachdriicklich wie in keinem anderen Dialog streicht also
Platon immer wieder heraus, daB die Sachverhalte, die darzustellen er sich anschickt, nicht so einfach sind, wie sie erscheinen.
Besonders auffallig ist sein Hinweis, daB der Demiurg schwer zu
finden sei und daB es, wenn man ihn gefunden habe, unmoglich
sei, ihn allen Menschen zu verkiinden (28 C 3 f). Das ist vordergriindig doch eine wirklich seltsame Aussage; denn verglichen
mit anderen Gegenstanden-dem idealen Vorbild, der Seele, der
xropa, der Zeit-gehort der Demiurg zu den GroBen des "Timaios",
die man am ehesten zu verstehen glaubt. Und doch sagt Platon, er
sei schwer zu finden und schwer mitzuteilen. 1st das nicht ein
deutlicher Hinweis darauf, daB Platon seine Darstellung nicht
prima facie verstanden wissen will? Wie sich das Wesen des Demiurgen erst dem Iangen Nachdenken iiber die Aussagen des
"Timaios" erschlieBt, ebenso steht es mit dem idealen Vorbild,
ebenso mit der Zeit, ebenso mit dem vorkosmischen Chaos, ja, mit
allem. Platon fordert offenbar den mitdenkenden Leser, der nicht
beim vordergriindigen Wortlaut hangenbleibt, sondern iiber diesen hinaus zu dem vordringt, was Timaios eigentlich meint (U
Ot<X.VOOUJ.l<X.t, 27 D 3).
DaB Platon so ver!ahrt, hat einen besonderen Grund: Die Frage
danach, ob die Welt entstanden sei oder nicht, war in seiner
Schule, der Akademie, offenbar heftig umstritten. In der Topik
nennt Aristoteles zweimal als Beispiel fiir ein q>Ucrtx:ov 7tp6PATJJ.l<X.
wieder die Schwierigkeiten, die sich einem angemessenen Verstiindnis des
Gesagten entgegenstellen.

96

MATTHIAS BALTES

die Frage, 7t6n:pov 0 ICO<JJ.LO~ tOtO~ 11 ou (A 11 104 b 8; A 14 105 b 24

f). Diese Frage hatte in der Akademie offenbar dieselbe Aktualitat
wie die, 1t6u:pov ft ftoovi'J aipe'tov 11 ou, die Aristoteles im gleichen
Zusammenhang anfuhrt (A 11 104 b 7). In De caelo widmet
Aristoteles dem Problem der Weltentstehung drei ganze Kapitel (I
10-12), in welch en er sich auch mit der Ansicht seiner Mitschuler
auseinandersetzt, die Welt sei nach Platons Lehre nicht entstanden. Platon selbst kennt die Diskussion, wie einzelne AuBerungen
in seinen Dialogen zeigen 60, und er nimmt selbst Stellung dazu,
nicht nur im "Timaios", sondern auch im "Politikos", doch tut er
es in heiden Dialogen anigmatisch. Warum?
Offenbar will er durch beide Dialoge zu dieser Frage Stellung
beziehen, aber nicht in der Weise eines au'to~ cpa, das aufgrund
des groBen Ansehens, das Platon im Kreise seiner Schuler
genoB6 1, die Diskussion, wenn nicht beendet, so doch gefahrdet
hatte. Was Platon mit seinem "Timaios", seinem "Politikos", ja
mit all seinen Dialogen wollte, war: Probleme zu benennen oder
aufzugreifen 62, den Leser zum Mitdenken und Nachdenken anzuregen und Hinweise zu geben, wo nach seiner Meinung die
Losung zu suchen sei. Eben dies tut er im "Timaios".
DaB Aristoteles ihn anders verstanden hat als Speusipp, Xenokrates und Herakleides Pontikos, hat Platon nicht beunruhigt; auf
diese Weise blieb die Frage als ein Problem im Raum, kam das
Nachdenken daruber nicht vorzeitig zur Ruhe. Ja, nicht nur in der
Alten Akademie kam diese Frage nicht mehr zur Rube, sondern
die ganze Antike, das Mittelalter und die Neuzeit hindurch bis auf
unsere Tage, wie wir anfangs gesagt haben. lch glaube, daB eben
dies Platons Anliegen war: das Problem zu formulieren, zum
Nachdenken daruber anzuleiten und Wegweisungen fur eine
mogliche Losung zu bieten. Jeder sollte seine eigene Losung
finden-ganz dem Wort Goethes entsprechend: "Was du ererbt
von deinen Vatern hast, erwirb es, urn es zu besitzen!"
Westfalische Wilhelms-U niversitat, Munster
60 Plat. Staat 546 A 2: yevoJ.LEVCJ> navti. cp9opa ecrnv (vgl. Tim. 41 A 8: to J.LEv oilv
Oft oe9ev 7tUV A.utv); Phaidr. 245 D 3 f: E7tEl0ft OE ayEVl)tOV EO"tlV, Kilt aotacp9opov
auto avayKl) dvat; vgl. Nomoi 781 E f; Epin. 981 E f; Proklos bei loh.
Philoponos, De aet. mundi XVII S.589,6 ff Rabe.
61 Zu Platons Ansehen in der Akademie vgl. M. Baltes, "Plato's School,
the Academy", Hermathena 155 (1993) 8 f. 18.
6 2 Vgl. Philodemi Academicorum historia Y 2 ff S.126 Dorandi = S.152
Gaiser.

PART THREE

THE ACADEMY AND THE PERIPATOS

SPEUSIPPUS ON PLEASURE
jOHN DILLON 1

The task of rescuing from oblivion, and properly evaluating, the

philosophical contributions of Plato's nephew and successor as
head of the Academy, Speusippus, is one that is by no means over
yet. Aristotle, whether through pique or genuine conviction, gives
the distinct impression of not taking him seriously,2 and modern
scholars have in general been content to follow his lead.
I have already ventured to come to the defence of his metaphysics,3 and I am glad to observe that other scholars are coming to
share my view that he is a thinker to be taken seriously. 4 What I
wish to come to the defence of on the present occasion is a basic
principle of his ethics, intimated to us in a text of Clement of
Alexandria, 5 namely the doctrine that happiness, or well-being
(EuBaqwvia), for man consists in 'freedom from disturbance'
(aoxft.:r,cria), 6 and that pain and pleasure are evils in an equal
1 This essay I offer, with some apprehension, to a great master in the art
of interpreting doxographic materials. With Speusippus all, sadly, is dependent on doxographic reports (in the broader sense), many of them hostile.
2 Good examples of this attitude would be Metaph. Z 2. 1028b21ff. (on his
first principles); PA I 2, 642b5ff. (on his logic); and EN VII 14, 1153blff. (on
his ethical theory).
3 'Speusippus in Iamblichus', Phronesis 29 (1984) 325-32. I was in fact only
developing insights first contributed by Philip Merlan in his stimulating
work From Platonism to Neoplatonism (The Hague 1954, 19602, 19683), and
criticizing what seemed to me the excessive scepticism of Leonardo Taran in
his otherwise very valuable edition of the fragments of Speusippus, Speusippus
of Athens (Leiden 1983). An important recent contribution both to our understanding of Speusippus and of the ancient interpretation of Plato's Parmenides
is contained in an article of Jens Halfwassen, 'Speusipp und die metaphysische Deutung von Platons Parmenides', in Hen kai Ptethos, Festschrift Karl
Bormann, edd. L. Hagemann & R. Glei (Wii.rzburg 1993).
4 Notably Margherita Isnardi Parente and (despite what seems to me to
be his perversity on various points) Leonardo Taran in their respective
collections of the fragments; and R.M. Dancy, particularly in 'Ancient NonBeings: Speusippus and Others', Ancient Philosophy 9 (1989) 207-43, repr. in Two
Studies in the Early Academy (Albany 1991).
5 Str. 2.133.4 = Fr 77 Taran.
6 This term, if Clement can be trusted, would seem to have been invented
by him. It is attested later for Epicurus and his followers, which is slightly
bothersome, but it would be odd of Clement to employ this precise term in

100

JOHN DILLON

degree, extending away from this 'mean' in an indefinite series of

gradations in either direction, so to speak.
It will be no harm, perhaps, to quote Clement's testimony in full,
since we have little enough apart from this to go on in reconstructing Speusippus' position.
l:7tEUO't7t7tOc; tE 0 TIA.citrovoc; cioel..cptliouc; 'tlJV euoaq.J.oviav <!Ill crt v E~tv
eivm teA.eiav EV toic; Kata cpucrtv exoucrtv ii E~tv ciya8rov, ~c; Iii!
Katacrtcicreroc; a7tavtac; J.lEV civ8pcimouc; ope~tv EXEtv, crtoxcil;ecr8m OE
toile; ciya8ouc; ti\c; cioxA.11criac;. eiev o ai. apetal. tijc; euOmJ.loviac;
ci1tepyacrn Kai.
Speusippus, the nephew of Plato, declares that happiness is a perfect
state in the area of what is natural, 7 or the state of (possession o)8
goods, which is a state for which all men have a (natural) impulse,
while the good aim at freedom from disturbance. It would be the
virtues that are creative of happiness.
We see here a very summary (but, we hope, basically accurate)
sketch of Speusippus' doctrine. He recognises that most men strive
for happiness, but seems to introduce as a qualification of this that
the good aim at aoxA.11cria. There must be some ellipse, though,
here, I think, doubtless as the result of Clement's compression of
his source. The state (Katacrtacru;) for which all men have ope~t~
may be euoaqwvia, but Speusippus may well have gone on to say
that most men believe, foolishly, that this lies in the acquisition
and enjoyment of pleasure, while only the wise understand that it
really resides in attaining freedom from disturbance. In that case,
Speusippus would grant this much to the proponents of hedonism
(such as Aristippus, founder of the Cyrenaic School, and his own
colleague Eudoxus of Cnidos), that pleasure is indeed a natural
object of striving for all living creatures, but he would deny that
that proves what it is claimed to prove: that pleasure is the good for
man. Instead, Speusippus wishes to maintain that for man, as a
characterizing Speusippus' doctrine, if it were known to be an Epicurean
term, unless it was in fact Speusippus' coinage.
7 In view of the problem arising below as to what sort of cllti cru; the
liucrxepet~, or 'cranky ones', of Phlb. 44a were 'experts' (lietvoi) in, we may
note the ambiguity of the word here. It could refer either to human nature or
to the 'nature of things'.
8 This represents a rather waffling attempt to render what seems to be a
variation between the meanings of the two uses of est~ here. It seems almost
incredible that Clement would use est~ in two different senses, 'state' and
'possession', in the space of one line, but the ambiguity may be more apparent
in English than in Greek.

SPEUSIPPUS ON PLEASURE

101

rational being, pleasure as an end must be transcended, as a

consequence of dialectic, and supplanted by a rational striving for
aoxA.TJc:ria.
Speusippus appears to have used a number of arguments to support his position (which seems to have been developed in opposition to that of his colleague Eudoxus, who maintained that pleasure
was the only good, as being the natural object of appetition) ,9 one of
which in particular comes in for trenchant criticism from
Aristotle, 10 but this does not go to the core of his doctrine, which is
not addressed by Aristotle-not unnaturally, since his purpose is
primarily polemical. It is only, I would maintain, if one considers
Speusippus' ethics against the background of his metaphysics, that
its inner logic becomes apparent. II
9 It seems fairly clear to me that Philebus, in Plato's Philebus, is intended
substantially to represent Eudoxus, and that Eudoxus is being gently teased
throughout the dialogue (the reference at the end, 67B, to drawing deductions from the behaviour of animals, is particularly pointed). We do not, I
think, have a title for any work he composed on this subject, and it is possible
that he advanced his views merely orally, though we do not need to assume
that. (By saying that Philebus is intended 'substantially' to represent Eudoxus,
by the way, I do not wish to imply that Philebus is intended as a portrait of
Eudoxus, nor that his position is a full or fair representation of that of
Eudoxus. Philebus is rather crass person, as Eudoxus was not, and his hedonism is really a sort of 'lowest common denominator' of the overall hedonist
position, which does not involve the use of certain arguments which we know
to be distinctive of Eudoxus. All I wish to claim is that it is Eudoxus' revival of
the hedonist position within the Academy-and Speusippus' attempts to
counter him-that has provoked Plato to compose this dialogue.)
As for Speusippus, his doctrine was no doubt set out primarily in his
treatise On Pleasure, perhaps in response to Eudoxus' exposition of his doctrine;
but his dialogue Aristippus very probably also dealt with the same theme, this
time featuring Aristippus, follower of Socrates and founder of the Cyrenaic
school, and a notable propounder of hedonism, but no doubt also aimed
primarily at Eudoxus. I see the Philebus as Plato's ironic commentary and
judgement on this ongoing dispute between two of his most distinguished
followers. There is no reason to assume, after all, that the views of either of
these men were published only after Plato's death, or even later than the
composition of the Philebus-a point which is of some importance to my
argument. After all, by the generally agreed date of the Philebus (c. 355 Be),
Speusippus was already in his fifties (he is generally agreed to have been
born c. 410-8 BC), and Eudoxus (born c. 390) probably already dead, in his
mid-thirties.
10 At ENVII 14, 1153b1-7 (=Fr. 80 Tariin), and X 2, 1173a5-28 (=Fr. 81
Tariin).
II It is interesting, in this connection, that one of the few positive things
reported by Diogenes Laertius (on the authority of an unknown Diodorus,
in Book I of his Memoirs ) in his very wretched 'life' of Speusippus (IV 2) is
that he 'was the first to discern the common element in all subjects of study
(E:v 1:o'ic; JJ.a8t1JJ.aot E:8Eciomo 1:0 JCotvov), and tried to coordinate (ouvq>JCEirooE)

102

JOHN DILLON

I hope that it may now be taken as reasonably certain that ch. 4

of Iamblichus' De Communi Mathematica Scientia preserves (by one
means or another, at either first or second hand) a conspectus of
Speusippus' metaphysical position. 12 There we find a metaphysical
scheme involving a pair of principles, One, or Unity, and Multiplicity, very much on the lines of those attributed to Plato in his
oral teaching 13 (and which are in effect present in the Philebus in
the guise of Limit and the Unlimited). It is by the action of Unity
on Multiplicity that every level of being in the universe is
generated.
It is not, fortunately, necessary to our purpose on this occasion to
go through all the details of this multi-level process (which is
certainly complex, though not, I think, deserving of the satirical
treatment which Aristotle gives it in Metaph. A 10). All I want to
focus on is what seems to me to be Speusippus' basic intuition,
reflected in the Philebus, that, on the human level, arete consists in
the imposition of peras, 'limitedness', on the indefiniteness of
human emotions.
What I conceive to be his view is actually well expressed, much
later and on the basis of a more elaborate theory of the virtues, 14 by
Plotinus in Ennead I 2.2.15-18, as follows: 15
The virtues do genuinely set us in order and make us better by
giving limit and measure to our desires, and putting measure into
all our experience; and they abolish false opinions, by what is
altogether better and by the fact of limitation, and by the exclusion
of the unlimited and indefinite and the existence of the measured;
and they are themselves limited and clearly defined' (trans.
Armstrong).
them as far as possible with one another.' Diodorus is, as I say, otherwise
unknown, but the fact that he is writing 'memoirs' (a7tOJ.lV11J.lOVEUJ.la'ta), in
which Speusippus features in Book I, might indicate that he was a hanger-on
of the Old Academy, and thus not a bad source.
12 On this see my article 'Speusippus in Iamblichus' (above, n. 1) 325-32,
in which I endeavour to defend Philip Merlan's hypothesis (in From Platonism
to Neoplatonism, ch V) against the scepticism of Leonardo Taran, op. cit. pp. 86107.
13 Cf. Ar. Metaph. A 6.
14 Plotinus is speaking here of the 'civic' (7tOAl'ttKai), as opposed to the
'cathartic' or 'purificatory' virtues, which is a distinction unknown to Speusippus. For our purposes, for 'civic' virtue one may understand virtue in general.
15 Ai. ... ape'tai... Ka'taKOCJJ.lOU<Jl J.lEV OV'tO><; Ka\. OJ.lELVOU<; 1tOlOUCJlV opit;oucrat
Ka\. J.lE'tpoilcrat 'ta<; E7tt9uJ.tia<; Ka\. oA.ro<; 'ta Tta911 J.lE'tpoilcrat Ka\. veullet<; Msa<;
acpatpoilcrat 'tc9 oA.ro<; OJ.leLVOVl Ka\. 'tc9 ropicr9at Ka\. 'tcOV OJ.lE'tpO>V Ka\. aopicr'tO>V esro
eivat Ka\. 'tO J.lEJ.lE'tP11J.lEVOV, Ka\. UU'ta\. optcr9etcrat.

SPEUSIPPUS ON PLEASURE

103

The repetition of various verbal and nominal forms of the

concepts horos and metron in this passage is quite remarkable. One
does not need to postulate any direct influence from Speusippus on
Plotinus here, of course. He is merely reflecting a deeplyingrained Platonist doctrine. I pick it out merely because I find it a
particularly clear expression of what I conceive Speusippus'
position to have been.
Speusippus, to reiterate, is criticized by Aristotle 16 for presenting
pleasure as just as much of an evil as pain, in his argument that,
because pleasure is the opposite of pain, it does not follow that if one
of them-pain-is an evil, the other is a good, since both may be
evils, opposed to some third thing, which is a good-even as the
greater may be opposed both to the less and to equal. Aristotle does
not object to this argument as such-he can hardly do that, since it
is integral to his own argument in favour of virtue as a mean
between extremes-but tries to pick holes in it on the specific
ground that, if something is an evil, it is to be avoided (<!>EUKtov),
and men do not in fact seek to avoid pleasure, while they do seek to
avoid pain.
This is not, however, an argument which Speusippus needs to
accept. His position could well be that pleasure is indeed a thing
that is to be avoided, even though men do not in fact generally
seek to avoid it. What men are seeking is in fact happiness
( euSaqwvi a) 17, and they wrongly imagine this to be attainable
through pleasure, or even to consist in pleasure. What all too few of
them realise, but which all should realise, is that the state which
constitutes true happiness is that in which the soul has reached the
perfect balance between pleasure and pain, which is attained by
the scientific application of peras to the unlimited spectrum of more
and less that runs from extreme pain to extreme pleasure, either
condition constituting a grave disturbance of the organism, and of
the soul which presides over it. It is this median state which
Speusippus denominates 'freedom from disturbance' (aoxA.Tlcria),
and which Epicurus later termed 'katastematic pleasure', and the
Stoics xapa, 'joy' .18
l6

ENVII 13 1153b1-7 (=Fr. 80 Tarin); ENX 2 1173a5-28 (=Fr. 81 Tarin).

17 That Speusippus held this (fairly obvious) fact to be the case is attested by

the doxographic report of Clement (Fr. 77 Tarin), where he declares that

eudaimonia is a state for which all men strive (il~ liT, Katacnaae~ ihtavta~ IJ.EV
avepomou~ ope!;w txew ).
18 For the Stoic conception of xapa (as euA.oyo~ e1tapm~) see D.L. 7.116, =

104

JOHN DILLON

To a certain extent, this whole argument may seem to turn on a

semantic quibble as to what constitutes ~ooviJ, but this may not after
all be the case. Speusippus would not wish to deny, I think, that his
state of aoxA-TJcria is in some sense pleasant, but his objection to
'pleasure' in its usual connotation in Greek is that it is essentially a
process, and an open-ended, disorderly one at that, while what he
is aiming at, and recommending, is a steady state. The Stoics later
developed a term for this to rival ~ooviJ, to wit xapci, but Speusippus
did not have this at his disposal, or rather did not get round to
inventing it. 19 Instead, he fastened on this rather negative-sounding term for his ideal state-without, however, I am sure, intending it to be taken as something merely negative.
And this brings me to the main topic of this discourse, the
identity of 'the enemies of Philebus' in Phlb. 44A-D. A great many
scholars before this, of course, have taken up positions on this
question, many in favour of the identification with Speusippus,
notably Doring,20 Wilamowitz,21 Philippson,22 Burnet,23 Taylor,24
FriedHinder,2 5 Gauthier & Jolif,26 Diiring,27 Kramer,2B and most
recently Malcolm Schofield.29 That is a pretty impressive line-up, I
think it would be generally agreed, but Leonardo Tanin, in his
collection of the fragments of Speusippus (pp. 78-85), nothing
daunted, proposes to dismiss the identification as based on misunderstandings and inadequate evidence. Taran is not alone in his
scepticism, 30 of course, but he is the most recent negative voice, and

sw 3.431.

19 But see below, n. 35, on the suspicious concentration of uses of xaipetv in

Phlb. 43-4.
20 'Eudoxos von Knidos, Speusippos, und der Dialog Philebos', Vierteljahrschriftjur wiss. Philos. und Sozio[. 27 (1903) 113-29 (esp. 125-7).
2( Platon (Berlin, 1919) 2.272-3.
22 'Akademische Verhandlungen iiber die Lustlehre', Hermes 60 (1925)
444-81, esp. 452-3 & 470-4.
23 Greek Philosophy, I: Thales to Plato (London 1914) 324f.
24 Plato, The Man and his Work (London 19374) 409-10, 423 with n. 1, 434-5;
A Commentary on Plato's Timaeus (Oxford 1928) 455-6.
25 Plato vol. 3: The Dialogues, Second and Third Periods (Princeton 1969) 339 &
540, n. 61.
26 L'Ethique a Nicomaque (Louvain-Paris 19702) 2.2.777, 788, 801.
2 7 Aristoteles. Darstellung und Interpretation seines Denkens (Heidelberg 1966)
457, n. 157.
28 Platonismus und hellenistische Philosophic (Berlin 1971) 205-9, with n. 88.
29 'Who were oi. liucrxepetc; in Plato, Philebus 44Aff.?', Museum Helveticum 28
(1971) 2-20 and 181.
30 August Dies, in his Bude edition (pp. LVII-LXII) also rejects it, for one,
but for reasons which Taran rightly dismisses ('not consistent with what we

SPEUSIPPUS ON PLEASURE

105

a powerful one, and it is he with whom I propose to deal here. He

begins with what seems to me to be a misconception of his own:
Because of his extreme anti-hedonism, several scholars have
ascribed to Speusippus the doctrine espoused by Philebus' enemies
in Plato's homonymous dialogue (44B-D), according to which pleasure is merely the cessation of pain. This conception of pleasure as
something purely negative (my italics) can hardly have been part of
Speusippus' doctrine, however, since for him the neutral state
between pleasure and pain coincides with the good. This neutral
state he must have identified with the freedom from disturbance
(aoxA.Ttcria) which for him is the necessary condition of virtue and
happiness. Hence, according to Speusippus the virtuous man must
free himself from both pain and pleasure; and, if pleasure were
nothing but the cessation of pain, he would not have considered it
an evil, nor would he have thought that the virtuous man must
avoid it as such.

It seems to me that Taran's mistake here is to take Plato too

literally, even as he frequently seems to me to take Aristotle too
literally in his attacks on Speusippus31 . There is certainly nothing
negative about Speusippus' doctrine on pleasure, but that does not
mean that his uncle's satirical presentation of it might not seek to
present it as negative-even as he is doubtless distorting in various
ways the hedonistic arguments of Eudoxus.
The dispute between Speusippus and Plato, after all, can be
viewed as at least partly a semantic one. The mere cessation of
know of Speusippus' character from the sources'). Hackforth too (Plato's Examination of Pleasure, Cambridge 1945, pp. 87-8) finds it impossible to reconcile
the position of the liuaxepe'ic; with what he feels he knows of Speusippus' doctrine, and opts cautiously for Grote's suggestion that they were 'Pythagorizing
friends' of Plato. But, as I shall argue, he is wrong to make an absolute
contrast between the assertion of the liuoxepe'ic; that pleasures are nothing but
escapes from pain and the belief of Speusippus that pleasure is real, but an evil.
3
As a notable instance of this one might take his interpretation of
Aristotle's criticism of Speusippus at Metaph. N 4, 1092all-17 (=Fr. 43 Taran),
where Aristotle is drawing as a polemical conclusion of Speusippus' position
that his first principle 'would not even be any sort of being (roo'tE J.lTJiiE ov 'tt
eivat 'tO v au't6)'. Even if, as Taran asserts (and I do not agree), the cOo'tE with
infinitive construction implies that Speusippus himself did not draw this conclusion, that would not in itself guarantee that that was indeed the case. See
my discussion in 'Speusippus in Iamblichus' (above, n. 3). Another example is
his treatment of Theophrastus' polemical criticism of Speusippus (Metaph.
llalB-26 = Fr. 83 Taran) for limiting the good in the universe to a little patch
in the middle, with vast stretches of evil on either side of it, which is a
malicious conflation of his doctrine that 'good' only arises at the level of soul,
the One not properly being describable as 'good' (see below), with his ethical
doctrine of the good lying between the twin evils of pain and pleasure. Taran
takes this far too seriously ( op. cit. (n. 3) 444-9).

106

JOHN DILLON

pain could, after all, be presented as a type of pleasure; but it need

not be, and Speusippus chose not to present it as such. He wishes to
reserve the term 'pleasure' to describe one of the equally reprehensible extremes which flank his ideal state, whereas Plato, it
would seem, wishes to extend the notion of 'pleasure' to take in also
the state which Speusippus sees as transcending both pleasure and
pain, and even to give it the title of 'true' or 'pure' pleasure. This
would not be the only area, we may note, in which Speusippus
ventures to oppose what he feels to be a looseness in terminology
on his uncle's part. On the important question of whether goodness
can properly be predicated of the first principle, he is at odds with
him (at least if we may take into evidence the survey of metaphysical doctrine preserved in Iamblichus' De comm. math. sc. ch. 4,
as I contend we may). There (p. 16, 10-14 Festa) he declares: 'The
One should not be called either fine or good, by virtue of the fact
that it is superior to both fineness and goodness; for it is only as
nature proceeds further from the first principles that, first, the fine
makes its appearance, and secondly, and at a level of elements
which are at a yet further remove, the good.'
Now Plato could argue that he is only calling the first principle
the Good in the sense of being the ultimate object of all striving, but
such terminology could nonetheless be criticized as involving
misleading connotations-and in particular as tending to an improper characterization of the indefinite dyadic principle opposed
to the One as 'evil'. Whether this somewhat impertinent correction
of the Master was aired before or only after his death we have no
means of knowing, but there is no reason to assume that Speusippus did not advance it while Plato was still alive. At any rate,
this seems to be the line he is taking on the question of whether the
perfectly balanced intermediate state which he postulates between
pleasure and pain may itself properly be described as pleasure. He
plainly feels that such a characterization would be seriously misleading, and prefers to reserve the term iJooviJ for those unlimited
and mobile areas on the spectrum stretching out on the opposite
side of his ideal state to painful ones.
On the other hand, he was left with certain terminological
problems. It is unfortunate that Speusippus felt constrained to use a
negative term to describe his ideal state, instead of, say, something
like the later Stoic xapa, but that does not mean that he intends
anything negative by it. Indeed, he undoubtedly wished to claim

SPEUSIPPUS ON PLEASURE

107

that this state was productive of profound satisfaction, not to say joy.
He may, however, for want of a better adjective, have been constrained to describe it as T]ou, 'pleasant', though denying that it
involved any form of T]oovt1, and this could have given his uncle
the chance to stick the knife in.32
Let us look now closely at the crucial passage of the Philebus, and
see if we can divine just what sort of position is being criticized.
We may begin at 43c. Socrates has just secured Protarchus' agreement that not all changes in our constitution produce pleasure or
pain, but only fairly considerable ones. He continues ( 43c8-dl0) :33
Socr.:
become
Prot.:
Socr.:
(aA.um)~

In that case, the form of life I mentioned just now would

a possibility again.
What one is that?
The one we said would be without distress or enjoyment
'tE lWl

aveu XUPflOVWV).

Prot.: Certainly it would.

Socr.: To sum up, then, let us posit three forms of life, one of
pleasure, one of distress, and one of neither. What would you say
on the subject?
Prot.: Just what you have said, that there are these three lives.
Socr.: Now not being in distress would hardly be the same as
enjoying oneself (xaipnv), would it?
Prot.: Of course not.
Socr.: So when you hear people say that the pleasantest thing
(ilotcrtov mivtffiv) is to live all one's life free from distress, what <h
you think they are saying?
Prot.: They seem to be saying that not being in distress is
pleasant ( il<\u).

Let us pause here for a moment. Somebody, or some class of

person, is being accused of claiming that the life without pain is
the most pleasant (i]ou) of all.34 If this is Speusippus, then one would
32 An interesting reflection of how difficult it is to avoid using parts of
when one is in need of a positive value word may be seen at Phlb. 66a
(where, indeed, Plato may be making some linguistic play with this fact, as
is suggested by Dies, op. cit. Intro. p. LXXXIX).
33 I borrow here the translation of J.C.B. Gosling, Plato, Philebus (Oxford,
1975).
34 This, of course, is not the first time that this argument has come up. In
Book IX of the Republic (583c-585a), Plato criticizes those who postulate that
pleasure consists essentially in cessation of pain (1taiil..a AU1tTJ~, AU1tTJ~ a1tai..A.ayti), whereas that is like thinking you have reached the top of a mountain,
say, when you have in fact only struggled up from the bottom as far as the
mid-point. But the opponents here sound more like hedonists of some sort
than anyone maintaining Speusippus' position, unless he is being seriously

itllU~

108

JOHN DILLON

like to think that he is being deliberately misrepresented; otherwise, he is being notably careless in his language. Speusippus
does, admittedly, have a problem, as I have suggested above. He
needs some positive value word to characterize his ideal psychic
state, but he must be wary of employing words of the same root as
i]Bovt1 for that purpose. The fact, however, that Plato also uses
XUtpEtV repeatedly 35 in this passage to describe this state might
indicate that this was in fact the preferred term of the supporters of
this position, and that the Stoics were actually anticipated in their
technical use of this term. The difficulty here is, however, that
there is no obvious adjective from xapa available to characterize the
preferred state,36 and this might have led Speusippus to fall back
incautiously on i]Bu, which would allow Plato to win a terminological argument against him.
Let us continue, however (43el-44b3):
Socr.: Take any three things now, say one gold, one silver, one
neither, just to have fine names.
Prot.: All right.
Socr.: Could the one that is neither possibly become either gold
or silver?
Prot.: How on earth could it?
Socr.: Similarly, it would be a mistake for anyone to believe and
therefore to say that the midway life was either pleasant (i]ouc;) or
distressing; at least if we are to be strict.
Prot.: How could it be?
Socr.: Yet we find people who say and believe these things.
Prot.: Certainly.
Socr.: Do they then think they are enjoying themselves ( xaipetv) on the occasions when they are not in a distressed condition?
Prot.: That's what they say, at any rate.
Socr.: So they believe they are then enjoying themselves
(Xaipnv), or they wouldn't say it.
Prot.: Probably.
Socr.: Yet they are making a false judgement about enjoyment
if, that is, enjoyment and lack of distress are two quite different
things.
Prot.: But they turned out to be quite different things.

misrepresented.
35 Five times between 43d3 and 44a5, together with XCXPJ.10vai once, at 43c7.
36 There is, of course, as the editors pointed out to me, xapto~. which,
though a rather poetical word, is actually presented as a synonym for itou~
(along with tEp1tvo~) at Prot. 358a, but Speusippus would still have the problem
of defining a special meaning for this term; otherwise he might as well use
itou~.

SPEUSIPPUS ON PLEASURE

109

Socr.: We have a choice, then. We could hold as we did just

now, that there are three alternatives, or that there are only two, first
distress, which we would say was a human evil, and secondly release
from distress, which, being itself good, 37 we should call pleasurable.

I think we can see various instances of unfair argumentation in

this passage. First of all, to produce 'gold' and 'silver' as examples of
two extremes between which there is to be a third thing which is
neither of them is profoundly tendentious. First of all, they are not
opposites; and secondly, they are selected in such a way that any
other item in the same class (presumably of metals) is going to be
worse than they are-unless, perhaps, it be the fabled orichalc!
Further, Socrates is attempting to convict the protagonists of this
position of wishing both to condemn pleasure, and to commend
their chosen intermediate state as pleasant, which it would certainly be inconsistent of them to do, without allowing that they may
have wished to claim that their intermediate state was characterized by a condition of mind far superior to pleasure, precisely as
being the result of the imposition of limit and order on the disorderly and unlimited spectrum of sensations of which pleasure is
one 'wing', so to speak. They may have had difficulty in finding
an adjective to describe this (and so may have incautiously fallen
back on 'liM), but it rather looks as if their preferred verb/noun for it
was xaipetv/xapci.
This, after all, would by no means be the only time that Socrates
was less than fair to a position of which he disapproved. But let us
press on. Socrates now gets round (44b4-c2) to identifying a group
which he describes as 'the enemies of Philebus', who are
characterized as being 'experts in natural science' (OEtvol. 1:a 7tEpt
<jrucrtv) and as being afflicted with a certain 'crankiness arising
from a not ignoble nature' ('tt~ OU<JXEpEta cpucrEox; OUJC ayevvou~) :38
3 7 This may be a rather weak rendering of auto to\ito ciya9ov ov. Perhaps,
'being just what it is to be good'?
3 8 I borrow the translation of Gosling here, with some amendments. The
particular point of the repeated mention of liucrxepei~/ouaxtpeta has been well
explored by Malcolm Schofield in his article, cited above (n. 29). His thesis is
not strictly provable, I think, but I am attracted by the notion that liucrxep~~
was known to be Speusippus' favourite adjective for describing a difficulty, and
that a gently teasing allusion is being made to that which we are meant to
pick up on ('we', of course, being Academic auditors or readers of the dialogue). Certainly, Aristotle's four uses of liucrxep- when discussing doctrines of
Speusippus in the Metaphysics (1086a2-5; 1090a7-10; 109la33-bl; 109lb22-5)
would seem to back up this conjecture.

110

JOHN DILLON

Prot.: Why are we raising this question at this stage, Socrates? I

don't see what you are getting at.
Socr.: Don't you know the real enemies of Philebus here?
Prot.: Who are you referring to?
Socr.: People with a considerable reputation as experts in the
science of nature,3 9 who deny any real existence to pleasures. 40
Prot.: How do they do that?
Socr.: According to them, what Philebus and his friends at
present call pleasures are nothing but cases of release from distress
(l.:urcrov arcoq>vyai) .

The question to be faced here is whether the position stated here

is in any way compatible with what appears from our other
evidence to have been Speusippus' doctrine. On the face of it, one
would be compelled to doubt it. Mter all, Speusippus is not attested
to have denied the existence of pleasures; rather, he declared that
they were just as much of an evil as pains, which would seem to
recognise their existence at least to the same extent.
But then one asks oneself, could anyone in his right mind have
denied the existence of pleasures? Surely what we have here from
Plato is the conclusion, rather tendentiously put, of a polemical
argument against pleasure. The argument, I submit, would go
something like this (and a complementary argument could in
theory be employed against the substantial existence of pain, except
that there was no need to wean people from an irrational attachment to pain!). Any pleasure you care to name, if you analyse it,
can be discerned as the result of freeing the organism from some
form of distress or other, resulting from a disequilibrium in one
39 The real meaning of the appellation lietvo\. I..EyOJ.!EVOt 'tCx Jca'ta <j>ucnv is a
troublesome question. It does not necessarily refer, as it is often taken to do, to
expertise in, or enthusiasm for, what we would call' natural science', though
it may include that. It can just as well refer to what we would term 'the
human sciences', including ethics. I take it, though, to be a good-naturedly
teasing reference to Speusippus' well-known concern to give a comprehensive
account of the relations of all branches of knowledge to one another, attested
by Diogenes Laertius (IV 2; cf. above, n. ll)-with which may be linked his
assertion, criticised by Aristotle at A.Po. B 13, 97a6ff. (= Fr. 63 Tarin), that
knowledge of each thing requires knowledge of its differentiae in respect of
everything else, which would lead to the sort of frenzied classificatory
activities satirized by Epicrates in his famous fragment (Fr. ll Kock), and
exemplified in the many volumes of Speusippus's Homoia.
40 There is some problem here as to what the Greek ('to 1tapa1tav i]liova<; ou
<j>aow Elvat) really means. Gosling wishes to translate 'who completely deny
that they (sc. the pleasures arising from cessation of distress) are pleasures'.
This would admittedly make better sense, but it is not the most natural
meaning of the Greek, I think. I give a translation compatible, I hope, both
with the Greek and what I take to be its real meaning.

SPEUSIPPUS ON PLEASURE

111

direction. What is called 'pleasure' is simply the tilting of the

disequilibrium in the other direction. So pleasures do not have a
substantial existence, in that their manifestation is always a byproduct of the removal of pain. You do not experience a sensation of
pleasure except in the context of the relief of some painful organic
imbalance or other.4I
Now this is a pretty tendentious argument, certainly, but at least
it is not manifestly insane, as an outright denial of the existence of
pleasure would be; and I think that it is a position that Speusippus
could have taken up, consistently with his known views. 42 The
great problem that it leaves him with-and this is where his
'crankiness' (oucrxepeta) comes in-is that he is forced to deny that
what Plato would wish to term 'pure pleasures' (51aff.), such as
those of smell or hearing, where there has been no previous
unpleasantness, or indeed most pleasures of the mind, are to be
described as pleasures at all.
But of course this is just what Speusippus wishes to deny, and it
here that his semantic dispute with his uncle becomes acute. Plato,
at 52c-dl, actually makes very much the distinction that Speusippus must have made, but he makes it between 'impure' and
'pure' pleasures, whereas Speusippus would have made it between
'pleasures', which he would regard as intrinsically unmeasured
41 We seem to get a reflection of this argument is the admittedly rather
casual remark of Socrates at the beginning of the Phaedo (60b), as he sits up on
his bed and rubs his legs after they have been released from the irons, that
pain and pleasure seem in a way to be Siamese twins, springing from a
common root.
4 2 The position of certain 'sophisticated persons' (Ko~voi nvec;) later, at
53d, that pleasure is always a state of coming to be (yeveatc;), never of being
( ouaia), is an interesting problem. The original argument, that pleasure is 'a
kind of smooth motion, issuing in sensation' is pretty certainly attributable to
Aristippus (cf. D.L. 2.85), and is thus a hedonist argument (see Dies' excellent
discussion op. cit. LXII-X), but it is less clear that Aristippus would have
wished to characterize pleasure as a genesis, as it lays him open to the counterargument presented here. The identification of kinesis as a genesis sounds
more like a hostile deduction. We must recall, in this connection, that
Speusippus composed an Aristippus (though whether before or after the Phlb. is
unfortunately uncertain). In such a work he must inevitably have sought to
counter the main feature of Aristippus' position. My suggestion would be that
Plato is here borrowing an argument from Speusippus (though he makes
Socrates ironically thank the original proposers of the argument at 54d4-5,
for providing material for their own refutation). Even if Speusippus' Aristippus
is judged to be the later work, it seems to me that he must have used this
argument, and that this is reflected later in Aristotle's reference to it at EN
VII 12 1152bl2ff.

112

JOHN DILLON

and disorderly, and sensations arising out of his ideal state of

equilibrium:
So now we have an orderly sorting out of the purified pleasures
from what should be called unpurified cases. As a further point we
ought to attribute to violent pleasures disorderliness ( cXJlEtpia), and to
those that are not, orderliness ( EJ.1J.1Etpia). Those that admit of great
degrees and intensity, whether becoming such commonly or only
rarely, we should put in the category mentioned earlier of that
which is indeterminate (to ihtetpov yevo~) and more and less and
in varying degrees permeates both body and soul. The others we
should put in the category of ordered things.
Socrates goes on (52d5ff) to argue that the disorderly and indeterminate class of pleasures can properly be declared 'false', and only
the ordered ones 'true', for very much the reasons that oi oucrxep1~
back in 44b-c were criticised for denying the substantial reality of
pleasure. Even there, we may note, the position of the oucrxep1~ is
not absolutely rejected; they are rather treated, ironically, as
'inspired prophets' (J.HlVtet~, 44c5), who have grasped an intimation
of the truth through a certain natural talent, but have not worked it
out dialectically. This, it seems to me, is a very suitable form of putdown for Plato to use when dealing with a bumptious nephew. In
fact, their two positions are not that far apart; it is just that Plato does
not see the sense of denying the title of 'pleasure' to those states of
the human organism which he has identified as 'pure' pleasures,
whereas for Speusippus the essential distinction between these
states and what he wants to designate 'pleasures' is precisely that
they are states, and the others are processes, or motions, admitting
of indefiniteness and 'more and less'.
Plato's contention is that the position of the oucrxepei<;, while
doing them credit, has not been properly thought out. It seems to
me, on the other hand, that there is a certain amount to be said for
it, even if it presents difficulties. After all, if we bear in mind
Speusippus' Pythagoreanizing world-view (which is not, indeed,
very different from that presented in the Philebus), it is important
that, in the sphere of ethics, one postulates the existence of perfect
states, representing the imposition of peras on the disorderly substratum of apeiria, and that it is such states that generate eudaimonia.
The problem of how to describe the by-products of such states,
which may include the enjoyment of the smell of roses, or the
sounds of birdsong or Beethoven's string quartets, as well as of
philosophizing, is something that he may not have quite mastered,

SPEUSIPPUS ON PLEASURE

113

though it seems possible, as we have seen, that he made use of the

verb xaipetv in this connection.
To complete our exegesis of the passage 44b-d, however: we may
derive from it (or rather from its continuation, 44e-45d) a further
polemical argument against pleasure which may well have been
advanced by Speusippus. Socrates, at any rate, with his usual irony,
thanks the oucrxepet~ warmly for supplying him with it.
The argument goes as follows. If we want to get a clear view of
the nature of a given thing, we should look at it in its most extreme
or unadulterated form. In the case of hardness, for example, we
should examine the hardest things that we can find, if we wish to
acquire an accurate idea of what hardness is. In the case of
pleasure, then, we look for the most extreme forms of pleasure, if
we wish properly to comprehend the nature of pleasure. But then
we are led to admit that people suffering from illness, either
physical or mental-the argument glosses over the question
whether all illnesses qualify in this regard, or only certain onesexperience more pleasure in more extreme forms (as a relief from
their various pains or distresses) than healthy people. Thus it is
argued that it is the most unbalanced natures that experience the
most extreme pleasures. And so it is indicated (I hesitate to say
proved ) that pleasure is essentially connected with imbalance in
the organism, and is thus the antithesis of a desirable state.
I see no reason not to accept that Plato is in fact borrowing this
argument from Speusippus, and using it for his own purposes. He
makes Socrates thank the oucrxepet~ for it, as I say. If anything, it
brings home to us how close their positions really were. This does
not, however, mean that Plato was any the less displeased with
Speusippus for taking up such a position. Plato would not appreciate
his nephew taking up a position on his 'right', so to speak, on such
a sensitive question as the status of pleasure. Mainly, though, he
seems to have regarded this super-austerity of Speusippus as
deficient in dialectical rigour. The sensations of the virtuous and
self-controlled have enough in common with those of the dissolute
and intemperate to merit being given the same generic name. A
more correct way of evaluating them is to make the distinction that
he works out in the latter part of the dialogue between 'adulterated'
and 'pure', 'false' and 'true' pleasures, rather than being left with a
class of sensations for which no proper name at all had been developed. This is, after all, just the role that the 'heavenly tradition'

114

JOHN DILLON

set out in 16-19 in meant to fulfil, making the correct distinctions

within the previously vague and amorphous concept denominated
'pleasure'.
It may after all turn out, though, that Speusippus has the best of
this argument. The Stoics were deficient to none in logical rigour,
and they, it seems to me, sided, in retrospect, with Speusippus. The
Stoic theory of Eumi9Etat, or 'equable states', after all, is precisely
developed to provide a set of sensations, including pleasurable ones,
which are appropriate to a wise man who is quite free from 7ta9oc;,
and that is the sort of Pythagorean sage that Speusippus seems to be
envisaging. For them, the experiences of the sage are not to be
confused with those of the vulgar, despite superficial resemblances
between them, and that is surely very much the position of
Speusippus before them.43
Trinity College, Dublin

43 By way of an irreverent concluding footnote, one may record the (doubtless malicious, but possibly accurate) report of Diogenes Laertius in his Life of
Speusippus (4.1) that the philosopher was 'prone to anger and vulnerable to
pleasures' (opyiA.oc; Kat itllovrov i;tto>V ~v). If so, he would not be the first or last
philosopher whose personal life did not quite come up to his principlesthough it must be said that the instances given by Diogenes to back up this
assertion are puerile and incoherent in the extreme.

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETATIONE 1-4

L.M. DE RIJK
1. Introduction*

By and large, in De interpretatione Aristotle is concerned with our

capability to speak about all that presents itself to our mind. From
chapter 4 onwards, he deals with the statement-making expressions (affirmation and negation), which are the main tools for
conveying our thoughts about things. This discussion is prepared
(chapters 1-3) by some important observations concerning the basic
elements of such expressions, viz. OVOJ.1<X and pflJla. The present
contribution contains some comments on Aristotle's view of the
proper nature of statement-making as put forward in De interpretatione.I First, I would like to highlight Aristotle's, what Sir David
Ross has called2 'frankly 'representative' view of knowledge' by
discussing the terms OJ.1otroJ.1a and npuyJ.1a. Next, I will discuss what
is meant by a term's 'time-connotation', and finally I will examine
the semantics of oVOJl<X, pflJla and A.Oyo~.

2.1. Meaning and Representation: lVhat does OJ.lOlOOJ.la refer to?

In the opening chapter, Aristotle's attention is focussed on the
representative character of expressions and thoughts, respectively:
'Spoken utterances are symbols of 'affections in the soul' (i.e.
thoughts), and written marks symbols of spoken utterances' (16a34), [ ... ] what these (sc. expressions) are in the first place signs ofaffections of the soul-are the same for all people; and what these
affections are 'likenesses' of-things-are surely the same' (16a6* Twenty years ago Jaap Mansfeld and I worked together on the Festschrift for our common teacher and his predecessor in the chair of Ancient
and Patristic philosophy at Utrecht. It now gives me great pleasure to be able
to contribute this little offering to his Festschrift, on the occasion of his 60th
birthday.
1 The present author has greatly profited from the rich literature on the
subject, especially the pioneering work by John Ackrill (1963), E. Montanari's (1988) extensive commentary on De int., chs 1-4, and the exemplary work
by Hermann Weidemann (1994).
2 Ross (1949) 25.

116

L.M. DE RIJK

8). The author is not interested in the psychological aspects of

making utterances, which aspects he regards (16a8-9) as 'not
belonging to the present subject', but to 'the work on the soul' (i.e.
De anima III 3-8).
The pivotal notion of OJ.Loico~ou:x ('likeness') has already drawn the
attention of the Ancient Greek commentators of De interpretatione. 3
The English word 'likeness', however, does not cover the entire
semantic area of Greek OJ!Ot<OJ.La. In addition to the basic sense of
'likeness', 'image', 'replica', the Greek word as used by e.g. Plato
and Aristotle connotes the idea of 'being substitutable for' or
'representative of the object the thing called OJ!Ot<OJ.La is said to be
the likeness of, to the effect that an object's nature may be
designated and clarified by its OJ!Ot<OJ.La. In order to illustrate this
wider meaning I shall give some significant passages from
Aristotle.
In Nicomachean Ethics VIII 10, 1160b21ff. Aristotle claims that
there are resemblances between different types of political constitutions and the relationships between the members of a household;
e.g. the relation between a father and his sons corresponds with a
monarchy, whereas that between man and wife may stand for the
aristocratic constitution, and so on for timocracy and democracy.
The general idea is that the different types of household may
elucidate the different natures of their respective counterparts in
politics. This discussion is introduced as follows: 'One may find
resemblances (OJ.LOtcOJ.La'ta) to the constitutions, and, as it were,
patterns (7tapaoiyJ.La'ta) of them even in households (1160b21-2).'
In Politics VIII 5, in which the nature of musical expression is
discussed, it is argued (1340a16-8) that 'there is clearly nothing
which we are so much concerned to acquire and to cultivate as the
power of forming right judgements, and of taking delight in good
dispositions and noble actions.' 4 With regard to the role of musical
expression, Aristotle points out that qua representations (OJ.LOtcOJ.La'ta)
of anger, gentleness, courage and all other kinds of qualities of
character, rhythm and melody are practically equivalent in their
effects on our souls to the affections they represent. We know from
our own experience, he argues (a23-5), that the propensity of
3 See Weidemann (1994) 134-51, and also Ackrill's (1963) 113 criticism of
'grave weaknesses in Aristotle's theory of meaning'.
4 Oxford Translation (B. Jowett, revised by W.D. Ross). Most of the
translations have been taken from the Oxford Translation.

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETATIONE

117

feeling pleasure or pain at mere representations is not far removed

from our experience of the same feelings when brought about by
reality ('tllV aA.ftSnav). So OJ.lOlCOJ.l<X, which is juxtaposed to JllJ.lllOt<;
(al2) and JllJ.lllJ.l<X (a39) 5 as their equivalent, stands for some entity
that is regarded as representative of something else, with which it
has its effectiveness in common.
In Metaphysics A 5 Aristotle discusses the importance attached to
mathematics by the Pythagoreans, 'who were the first to take up
mathematics (and) not only advanced this study, but also having
been brought up in it thought its principles were the principles of
all things (985b24-6).' In numbers the Pythagoreans seemed to see,
Aristotle argues (985b27-32), as many 'OJ.lOHflJ.l<X't<X to the things
that exist and come into being, more than in the traditional basic
ontological elements, fire, earth and water, such and such a
modification of numbers being justice, another being soul and
reason, another being opportunity, and similarly almost all other
things being numerically expressible.' 'Since, again', he continues
(985b32-986a2), 'they saw that the modifications and the ratios of
the musical scales were expressible in numbers; since, then, all
other things seemed in their whole nature to be modelled
(aq>roJ.lot&crSat) on numbers, and numbers seemed to be the first
things in the whole of nature, they supposed the elements of
numbers to be the elements of all things.' In the Pythagoreans'
view, the various entities could adequately be represented by the
corresponding numerical ratios. In this context, the notion of
OJ.lOlCOJ.l<X goes beyond just likeness; rather, it expressly connotes the
idea that different numerical ratios are representative of, and may
stand for, different entities and that, accordingly, the nature of
these entities can be represented by the corresponding number.
To my mind, it is this pregnant notion of OJ.lOlCOJ.l<X that is found
in the opening lines of De interpretatione, in which expressions are
said to primarily refer to thoughts ('affections of the soul'), which,
in their turn, refer to 7tpnyJ.la'ta. In order to determine more precisely Aristotle's view of 'representative thought' the next subject
that presents itself for examination is the proper meaning of
7tp&yJ.la.

5 The genuine meaning of J.llJ.lTJO't~ and its cognates ('representation') is

thoroughly discussed in Kardaun ( 1993), who has convincingly shown that
the common rendering 'imitation' is incorrect.

118

L.M. DE RIJK

2.2. Meaning and Representation: What does nparJ.La properly stand for?
Serious doubts may be raised against the rendering 'actual thing'
for 7tpftnux.6 According to Boethius, Alexander of Aphrodisias
already put forward the question of why-given the claim in De
interpretatione, ch. 1 that expressions act as names referring to
things-, Aristotle says that utterances are primarily (16a6 7tpclnrov)
signs of thoughts rather than of actual things. 7 Alexander thinks
that Aristotle possibly means to say that, although expressions are
names of things, we do not use them in order to signify things, but
rather to signify the thoughts ('affections of the soul') we have of
things in our mind. Aristotle justly claimed, he says, that utterances are primarily signs of thoughts, because they are properly
used to signify the thoughts we have of the things involved.s Some
lines further on, our spokesman Boethius rightly comments that
in calling the OJlOU:OJ.UX'ta 'affections of the soul' Aristotle means to
say that to think is precisely 'to receive an object's proper image in
the soul's reflections'.9
Alexander-speaking for Aristotle-appears to reject the opposition between thinking and reality that is implied in the question
'do the utterances refer to either thoughts or (actual) things?' Not the
objects as such are the proper objects of signifying but the thoughts
that are formed of the (actual) things; in Boethius words 'quae ex
rebus nobis innatae sunt animae passiones'. In other words,
6 E.g. Ackrill's rendering of 1tPcl'YiliXtiX at 16a7 and 17a38 (as in Cat. 5,
4a36ff; 12, 14b19ff.); cf. Boethius II 22.2-6. It is righty rejected by others, e.g.
Weidemann (1994) 138-9.
7 Boethius II, 40.30-41.7: [ ... ] quaerit Alexander: si rerum nomina sunt,
quid causae est ut primorum intellectuum notas esse voces diceret Aristoteles?
Rei enim ponitur nomen, ut cum dicimus 'homo', significamus quidem
intellectum, rei tamen nomen est, idest animalis rationalis mortalis. Cur
ergo non primarum magis rerum notae sunt voces quibus ponuntur potius
quam intellectuum?
8 Ibid., 41.7-13: 'Sed fortasse quidem ob hoc dictum est, inquit, quod, licet
voces rerum nomina sint, tamen non idcirco utimur vocibus ut res significemus, sed ut eas quae ex rebus nobis innatae sunt animae passiones.
Quocirca propter quorum significantiam voces ipsae proferuntur, recte eorum
primorum esse dixit esse notas.' The flavour of innatism ('innatae sunt
passiones animae') in Alexander's exposition would surely not have amused
Aristotle.
9 Ibid., 43.12-16: 'Similitudinem vero passionem animae vocavit, quod
secundum Aristotelem nihil aliud est [omitted by Meiser; see however the
older editions, e.g. Migne, PL 64, 414 B13] intellegere nisi cuiuslibet subiectae rei proprietatem atque imaginationem in animae ipsius reputatione
suscipere.'

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETATIONE

119

neither actual things (or states of affairs) by themselves nor our

intellections as such are under consideration, but 'things' (whether
or not actually existent) in as far as they are conceived of and referred to
by expressions. 10 Admittedly, the Ancient thinkers, realists as they
were, believed that there is a real world existing independent of
human thinking, and that man is capable of grasping it by
thinking, but this does not alter the fact that, when actually thinking, a realist's thinking, too, does not go beyond things as conceived
of; to conceive of a thing as it is independent of conceiving is flatly
contradictory.
The notion of 7tpa:yJ.La ('thing or state of affairs as conceived of)
deserves our special attention. W. Wieland already claimed that,
generally speaking, to Aristotle 7tpayJ.La has its locus naturalis only
within what he calls 'the linguistic horizon' and merely is that
which one has in mind in each case when using a word or a
phrase.ll I could not agree more. This 'horizon' of 'things (states of
affairs) taken as being conceived of, and as such obtaining or notobtaining' is frequently found in Aristotle. 12 Wieland (loc. cit.)
rightly adduces De interpretatione 7, 17a38ff. to corroborate his view.
In this passage, Aristotle distinguishes between universal and
particular 7tpayJ.La'ta, and claims that the universal one is 'that
which is by its nature said of a plurality of things and the
particular that which is not'. Now, it is not actual things that are
said of something else, but their concepts; moreover, unlike Plato,
Aristotle rejects the actual existence of universal things, because to
him, whatever exists is particular. 13 Therefore, 7tpiX.yJ.La must stand,
not for 'actual thing', but for the content of an expression, i.e. a
thing (rather a state of affairs) as conceived of, irrespective of its
actuality.

lO For that matter, also Weidemann's (1994, 148-9) diagram of two dovetailed semiotic triangles is based, it seems, upon a mistaken opposition of
'thing' and thought, and thus seems to obscure the crucial point.
11 Wieland (1970) 170-1 (dealing with Aristotle's Physics): '[ ... ] das pragma seinen Ort nur innerhalb des sprachlichen Horizontes hat'; '[ ... ] es ist nur
das, was man mit dem Wort oder mit der Rede jeweils meint'. See also
Weidemann (1994) 139.
l2 For a more detailed survey of its occurrence in Aristotle see De Rijk
(1987) 36-39.
13 See our note 21 below. In Cat. 10, 12b5-16 and in Metaph. t:. 29, 1024b19ff.
even nonexistent (false) states of affairs are called np&wa. See De Rijk (1987)
38-9; Weid..:mann (1994) 138.

120

L.M. DE RIJK

3. On time-connotation and timelessness

In the remaining part of the opening chapter (16a3-9), Aristotle
goes on to clarify the distinction between statement-making
expressions on the one hand and their constituents, OVOf.!<X and pflf.!a
on the other. The decisive difference between them is that statements are susceptible of truth-values, whereas oVOf.!CX.'ta and pllf.l<X't<X
are not. In order to have a truth-value, a thought, and its linguistic
expression, require the addition of 'is' or 'is not', because 6v6f.!a'ta
and pfJf.!a'ta by themselves, like thoughts when taken by themselves, are neither true nor false (16a9-15).
A serious difficulty arises at 16a18, where this addition of 'is' or
'is not' (dvat and f.!ll dvat, respectively) is qualified by the phrase 11
anA.&c; llK<X'tU xp6vov ('either simply or with reference to time').
Ackrill prefers the common explanation of this phrase in terms of
the opposition 'present time versus past and future times' to taking
anA.&c; to allude to the timeless or to the omnitemporal present
tense. Yet he feels that the distinction between present time and
past and future times 'would not be very happily expressed by the
disjunction 'either simply or with reference to time'.' 14 The same
phrase is also used at A.Pr. I 15, 34b7-8 and, somewhat modified, at
Topica I 5, 102a22-30 and De anima III 10, 433b5-10. To my mind,
these parallel passages provide clues that point to the alternative
interpretation of this phrase, viz. as referring to the opposition
between time-connotation and timelessness. 15
Some preliminary remarks should be made on Aristotle's use of
the word anA.&c; (simpliciter). Aristotle defines it on two occasions. At
Topica II 11, 115b29-35 it is described by the claim (b29-30) that
'what is honourable simpliciter is that which without any addition
you are prepared to say is honourable.' The honourable simpliciter,
e.g. to honour the gods, is opposed to what is honourable to some
people, such as to sacrifice one's own father, which is only
honourable secundum quid. A similar opposition between simpliciter
and secundum quid is also found elsewhere. 16 Another definition of
the word occurs at GC I 3, 317b5-7: 'The word 'simpliciter' either
14 Ackrill (1963) 115.
15 Weidemann (1994) 157 adduces only A.Pr. I 15, 34b7-8, but does not

re9ard it as a parallel, and joins Ackrell's interpretation.

6 E.g. SE 5, 166b22-167a20; 6, 168b11-16; 25, 180a23-31; Phys. I 1, 184a18; I 7,
190b 2 ('substances, i.e. anything that can be said to be without qualification');
GCI 3, 317b2-5; b11-13; Metaph. ll 11, 1018b11; Z 3, 1029b6.

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETATIONE

121

refers to what is primary, regarding any appellation of what is, 17 or

to the universal in the sense of 'all-comprehensive'.'
In the second definition there is a shift from 'in an unqualified
sense' to 'essential' or 'by nature'. The juxtaposition of Cx7tA&<; to
<pUOEt is frequently found. So when discussing (Anal. Post. I 2,
71 b34-72a5) the requirements for the premisses of a demonstrative
proof, Aristotle remarks that they should depend (ultimately) upon
primitive truths which are 'prior and more familiar', meaning, not
that they are so in relation to us and, thus, nearer to sense
perception, but that they are so by nature, simpliciter and far away
from sense perception. 18 In Topica VI 4, there is a similar association of true definition (through genus and differentiae) with a thing's
essential nature (141b24), which is 'prior simpliciter and requires
exceptional understanding to be grasped' (141b13-14).
The word Cx7tAOO<; also occurs in De caelo I 9 where Aristotle discusses the contradistinction between something's shape (jlop<pft,
which term is used as an equivalent of dooc;) in itself and the one
in combination with matter. This distinction is telling, Aristotle
argues, since e.g. the form (as such) of the sphere is one thing and
the golden or bronze sphere is another, for when we state the
essential nature ('to 'tt nv dvat) of the sphere we do not include its
material in the definition, because it does not belong to its essence.
If the universe is a particular, there will still be a distinction
between the being of 'the particular universe' and 'universe' taken
by itself (a1tA.&c; i.e. qua dooc;). So there is a difference between 'this
universe' (i.e. ours) and 'universe taken qua dooc;; the latter stands
for form and shape (dooc; Kat jlop<pft), the former for a form that is
combined with matter. 19 In Phys. VIII 7, 261b4-5, with respect to
l7 Ka8' i:KO:crtTJV KUtTJyopiav toli ovtoc;. The common rendering 'within each
Category' (Oxford translation) is senseless. Aristotle means to say that if you
name an entity, say stone or health, and so on, a1tA.&c; it is stone, health etc.
in an unqualified sense (e.g. 'health simpliciter at Top. VI 4, 142all). In the
sense of 'all-comprehensive' 'stone a1tA.&c;' means the class of all stones. For
the key notion KUtTJyopia (= 'name' or 'appellation/nomination' rather than
'predicate') see De Rijk (1980) and (1988) passim.
18 A.Post. I 2, 71b33-72a5: 'Things are prior and more familiar in two
ways: for it is not the same to be prior by nature and prior in relation to us,
nor to be more familiar <simpliciter> and more familiar to us. I call prior and
more familiar in relation to us what is nearer to sense perception, prior and
more familiar simpliciter what is further away. What is most universal is
furthest away, and the particulars are nearest; and these <two kinds> are
mutually opposite'.
19 Cael. I 9, 277b31-278a15.

122

L.M. DE RIJK

becoming and passing away, a similar opposition between their

being as such (a1tA.&c;) and their being instantiated in particulars
(Ka8' EKCX.cr'tov) is made.
The general idea in all these passages is that the phrase '[x]
taken a1tA.&c;' is used by Aristotle to stand for an entity irrespective
of its being instantiated in matter, and that its counterparts2 stand
for [x] as instantiated in matter. Now, any material instantiationone must keep in mind that for Aristotle, anything sensible is
instantiated in matter2I-is temporal. Therefore, the notion of
a1tA.&c; naturally comes close to that of timelessness. 22
In point of fact, one sometimes finds the notions of a1tA.&c; and
'universality' opposed to the temporality of particular instantiations.
Thus, when discussing the nature of the universal syllogistic
premiss (e.g. 'A belongs to all B'), Aristotle explains (Anal. Pr. I 15,
34b7ff.) that we must understand the phrase 'belonging to all of a
subject' not with reference to any time (llflKa'ta xp6vov opicrav'tac;),
but a7tA.&c;, hence 'timeless'' for that is the way to get hold of the
material genuine syllogisms are made of. The author expressly
insists that in such cases 'universality must be taken simply', that
is not limited in terms of time. 23 The phrase (ou xp6vcp Otopl~OV'tac;
at b18) seems to rule out time and temporal instantiation and to
imply timelessness, rather than merely temporal indefiniteness or
omnitemporality. To Aristotle of course, unlike Plato, no transcendent domain of being is implied when he uses the notion of
timelessness. In his view, this notion is merely instrumental in
our conceiving the things of the outside world, viz. our taking (at
times, e.g. when framing apodeictic syllogisms about them) their
immanent natures abstracted from these natures' inherence in
these things, in order to draw, and to justify, non-contingent
conclusions about them.24
Aristotle's remark at 16a18 ('either simply or with reference to
viz. '[x] Ka'ta f.J.Epo~. '[x] fJ.Efll'YfJ.EVov', and so on.
Cael. I 9, 278all: 'whatever is sensible subsists altogether in matter.'
22 Ackrill (1963) 115 also considered (and rejected) the suggestion that
anA.&~ 'alludes to the timeless or to the omnitemporal present tense'; see
above, p.120.
23 A.Pr. I 15, 34b17-8: 'tO Ka86A.ou A.Tjn'tEOV anA.~ Kat ou xp6vcp Otopt~OV'ta~.
24 For Aristotle, all epistemonic inquiry starts from particulars of the
outside world, but they are always taken, then, as instances of the universal
nature(s) immanent in them. Therefore, the 'anA.&~-consideration', by which
they are taken simpliciter or detached from their being materialised, is of
utmost importance for the epistemonic inquiry. See De Rijk (1995) 78-102.
20

21

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETATIONE

123

time') should be interpreted along the same lines, i.e. in terms of

timelessness vs. temporality. What he means to say is that the
'being' added to verbal expressions or thoughts, which provides
them with truth-value, may either be the timeless being belonging
to the connotatum of the term involved, or it may refer to the
temporal instantiations and instances.25 Thus, statements of various
content may be made.26
This interpretation will find some support in Aristotle's discussion of i>ilJla., which in chapter 3 of his treatise, is said to 'additionally signify time' .27

4. On the semantics of OVOf.la ('name)

In the opening line of De interpretatione, Aristotle had announced
(16al-2) that it was necessary to 'posit' what an OVOJ.l<X, a PTlJl<X, a
A.Oyo~ etc. is. In ch. 2, 16al9 he starts the discussion concerning the
first component of the statement-making expression, which is
called OVOJ.l<X. In the wake of Plato, Aristotle's OVOJ.l<X is used to stand
for a one-word expression, or 'significative word', always with the
strong connotation of 'referring to' or 'representing' some thing
(whether connotatively or denotatively) .2 8 It may be rendered
'name' in English, in the general sense of 'designation'. Its
semantic function is to indicate something and to bring it up for
discussion.
Name is defined (2, 16al9-21) as [1] 'an articulate sound, [2]
significant [3] by convention, [ 4] without time, [5] none of whose
2 5 'Manhood' and 'triangularity' are 'instantiations' inhering in the
'instances' 'man' and 'triangle', respectively.
26 Cf. the opening lines of the next quotation from Boethius.
2 7 The first of the three explanations of our phrase mentioned by Boethius
(II, 51.2-10) seems to be very close to the one argued for above: '( ... ] vel
simpliciter vel secundum tempus. Hoc vero idcirco addidit quod in quibusdam
ita enuntiationes fiunt ut quod de ipsis dicitur, secundum substantiam proponatur, in quibusdam vero hoc ipsum 'esse' quod additur, non substantiam sed
praesentiam quandam significet. Cum enim dicimus 'Deus est', non eum
dicimus nunc esse, sed tantum in substantia esse, ut hoc ad inmutabilitatem
potius substantiae quam ad tempus aliquod referatur.' Cf. Trendelenburg
(1892) 53: 'anA.i.i:H; simpliciter earn affirmationis vel negationis necessitatem
indicat, quae quasi maior omni tempore dominetur; cujus modi genus praesenti efferre solemus; 1ca;-ta xp6vov contra enunciatum certo quodam tempore
cohibet.'
28 For Plato's semantics of ovojHl and Pllllll see De Rijk (1986) 217-82;
(1987) 27-33; 53. -It should be remarked that in these passages, my unfortunate rendering 'imitation' and 'to imitate' for lltlll]O"t~ and j1tj1E'icr8at must be
corrected into 'representation' and 'to represent'; see Kardaun (1983).

124

L.M. DE RIJK

parts is significant separately.' The ingredients [1] and [2] are

quite in line with what is said in the opening chapter, and [3]
alludes to the expositions in Plato's Cratylus.29
From the semantic point of view the differentiae indicated by
phrases [ 4] and [5] are the most interesting, because they are
meant to mark off the name against PllJ.l<X and A.6yo~. (Following
Aristotle, I will postpone the explanation of phrase [ 4] 'without
time' to the next section). With feature [5] Aristotle, to my mind,
only intends to contrast the name,-as at 3, 16b6-7 the PllJ.l<X -qua
one-word expression with A.6yo~, because, when taken by themselves, their parts ('letters')-unlike those of A.6yo~-are semantically incomplete and, accordingly, do not fulfil the requirement of
being properly significative.30
Expressions such as 'not-man' give rise to some difficulties.
They are not genuine names, nor are they a A.6yo~ or a negative
statement. On the other hand, they are significant by themselves,
the only difference to ordinary names being that they indicate
only a vague, indefinite connotatum3I and may denote, consequently, 'a wildly various range of objects' .3 2. So they are not
entirely denied the label 'name' and are called by Aristotle
'indefinite names' .33
Finally, Aristotle rules out the inflexions of names (i.e. oblique
cases of nouns, such as 'Philo's' or 'to-Philo') as candidates for
counting as names. They do not bring up the entity signified by
the name properly, as e.g. 'Philo's' may primarily refer to his wife,
or a son or daughter of his, or one of his possessions or properties,
anything other, that is, than the man himself. That is why only
the name in the nominative form is called KAll<Tt~, i.e. 'calling
case', as in A. Pr. I 36, 48b39-49al, where Aristotle claims that the
terms of the syllogism are always to be stated in 'the calling-upforms of the names' (Kata ta~ KAft<TEt~ to>V 6v6Jlatrov), e.g. 'man',
'good', 'contraries', not 'of-man', 'of-good', 'of-contraries'.

29

For the semantic impact of this dialogue see De Rijk (1986) 217-53.

30 Weidemann (1994) 166.

31 See De int. 10, 19b9: 'for what it signifies is in a way one thing, but
indefinite.' Obviously, when actually used, its denotatum may be quite precise,
e.g. in 'this not-man is a tree'.
32 Ackrill (1963) 118. See also Boethius II, 62.3-19; 69.29-70.3.
33 Boethius II, 62.19-63.14. For Plato too, the indefinite name has a
descriptive force. See De Rijk, (1986) 165ff.; 170ff.

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETATIONE

125

5. On the semantics of pfiJ.La ('attribute)

As a technical tool piiJ.UX (literally 'what is said' or 'assertible') is
identified as 'that <sc. PiiJ.ux> which additionally signifies time, no
part of which is significant separately, and which is3 4 always a
sign of things qua being said of something else'.
The wording of this formula requires some comment. It must
strike the reader that the first part of the definition of Myoc;, which
will be presented by Aristotle in chapter 4, runs entirely parallel to
that given of ovoJ.La (in chapter 2) in that each of them commences
with their common generic part, viz. 'articulate significative sound
which .. .'. What Aristotle says here about {:rilJ.La, however, does not
start from that generic part, he rather posits (notice the use of 8ecr8at
at 16al, and its being alluded to at 19bl4) 35 PiiJ.La as 'the one which
... etc.'36 This wording strongly suggests that Aristotle is not defining piiJ.La in general, but instead identifies the PiiJ.La as intended in
the present treatise as 'that <sc. piiJ.La> which additionally signifies
time, no part of which ... etc.'. Quite understandably so, because
e.g. Plato's PiiJ.La is not necessarily a one-word expression37, so that it
would not meet the second condition ('no part of which is significant in its own right') for being a PiiJ.La in the sense Aristotle has
determined just now.
The features [1] and [3] are interesting from the semantic point of
view. What is meant by 'to additionally signify time'? Aristotle
explains (16b8-9) this by opposing uytaivn to (the name) uyina: 'By
'additionally signifying time' I mean this: health is a name, but ishealthy a PiiJ.La, because it additionally signifies something as
obtaining now.' He apparently intends to say that as a name uyieta
refers to the entity 'health', but in bringing it up only qua form,
whereas uytaivet refers to health as a form actually inhering in
something else, hence as a form that is materialised or actualised
and manifests itself in temporal conditions. According to Aristotle,
34 The Greek language does not allow one to connect more than one
relative clause with an antecedent, and instead has the paratactical construction esti de ('and it is'). See Kuhner-Gerth (1904) 2.432f.
35 At 10, 19b14 Aristotle qualifies the notion pftJ.UX by saying 'i>TtiHl'tU
according to what we laid down'.
36 It is worthwile to note that Boethius finds it useful to begin his
comments on this lemma by presenting the 'integra definitio' (II, 66.3-6:
'Verbum est vox significativa secundum placitum, que consignificat tempus ...
etc.').
37 See De Rijk (1986) 222-234.

126

L.M. DE RIJK

no form exists unless it is materialised in something else as its

subject-substrate. 38 This need not prevent us, however, from referring to a form as such, irrespective of its being immanent in a
substrate. For that matter, Aristotle's vivid discussion with Plato
about his Master's Separate, Transcendent Forms, as well as his
own reflections in his Metaphysics about true substance, could not
possibly take place without bringing up 'forms as such' in abstracto.
Now, in such discussions, names designate entities regardless of
their being immanent in temporal substrates ('without time'),
whereas by using pfu.ux'ta this ontological state is focussed on.
The third feature ('is always a sign of things qua being said of
something else') is closely linked up with the first one, and is in a
way its counterpart. Whereas the first raises the question of an
entity's ontological state as something that is immanent in a
subject-substrate and hence temporally determined, the third
refers to the matching grammatico-semantical situation, namely
that, in keeping with this ontological state, the entity is also
examined inasmuch as it is assigned to this subject-substrate. 39 Thus
Aristotle implicitly presents his basic view-opposed to that of
Plato-that what is signified by a pftJla is a non-subsistent entity. So
a pftJla always implies a reference to the subsistent substrate which
the significate it designates inheres in. 40 Quite naturally, then,
these two features of pft Jl<X are explained by Aristotle as being
closely related. 41
Next, Aristotle deals with two kinds of pftJla-like expressions
which, however, are not entitled to bear the label properly, viz.
negative expressions such as oux- uytatVEt, and inflexions of PllJl<X,
i.e. its preterite and future tenses. He proposes to call the former
a6pta'ta PTtJl<X't<X. Though they have time-reference and their
significates include a connotation of temporality and always their
Cael. I 9, 278a1l.
39 Though the addition (at 16bl0) ~
imoKElllEVq> (found in Porphyry and
some manuscripts, and defended by Montanari ( 1988, 193-9) is perhaps not
authentic, it is perfectly in keeping with what Aristotle has in mind in this
chacter.
4
The two conditions are explained by Boethius (II, 66.9-22; 67.7-21; 68.131) along the same lines.
41 '[ ... ] because it additionally signifies something as obtaining now, and
is always a sign of what is said of something else, that is, of a subjectsubstrate'(16b9-10), reading, with a number of manuscripts (and Porphyry),
Kct9' tpou AEYOilEvrov instead of \mapx6vtrov (Minio-Paluello). See Weidemann
(1994) 173.
38

ev

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETA TIONE

127

holding of something else, they are different42 and can only be

classed as pftJ.W.'ta. in a qualified sense.
There is some discussion about the precise meaning of
'indefinite pflJ..La.'. Ackrill considers it a 'misnomer' .43 It should be
noted at the outset that Aristotle's account of it (at 16bl5) is rather
obscure: 'Because it [sc. the indefinite pflJ..La.] holds indiscriminately of anything, whether being or not being.' Ackrill and
Weidemann are of the opinion that the indefinite pflJ..La. is in fact a
sentence-fragment, containing an ordinary pflJ..La. together with the
negative particle, which serves to make the statement of which it is
part a negative one.44
I prefer, however, to take a6ptO''tOV PllJ..l<X. in a similar way as
'indefinite name'. On this assumption, it must refer to an expression that equals an ordinary pflJ..La., save for the latter's definite way
of referring to the significate involved. The indefinite pflJ..La. thus
signifies some kind of being in an indefinite manner, e.g. 'being
not-healthy' (16bll) or 'being not-just' (20a31). One should be
aware that, just as in the case of indefinite names, there is no
empty designation involved, but really an antic value, despite the
minimal information given about its nature. The 'characteristic'
designated by the indefinite pflJ..La. is so vague even that 'it can be
said of things that are and things that are not, indiscriminately',
meaning that e.g. 'not-healthy' and 'not-just' may stand for any
existent thing that lacks health as well as for what does not exist at
all.
That precisely reflects the double meaning of a6ptcr'tOV elsewhere in Aristotle. In Physics II 5, 196b23ff., when discussing, for
instance, the different ways in which a thing's cause may be
indicated, he distinguishes the (determinate) per se designation
from the plurality of indeterminate, incidental designations. E.g.
the cause of a house (architect) may be per se designated as
'housebuilder' or by one of the man's incidental attributes as 'the
pale entity' or 'the musical'; in the latter case the cause is an
indefinite one ( UOptO'tOV, b 28), not qua cause, but as to how it is

42 I agree with the common interpretation of this passage, but, to my

mind, it is necessary to remedy the harsh construction by supplying (at b13,
after {mapxu) the words (litacpEpEt lib, which may have been omitted due to
ha~lography.
3 Ackrill (1963)

120; also Weidemann (1994) 177.

44 Ackrill (1963) 120-1; Weidemann (1994) 177. See alsop. 133 below.

128

L.M. DE RIJK

being referred to. In Physics V 1, 225a20-25, the author uses both

senses of the indefinite term 'not-being', which may stand for
unqualified not-being ('what is not'), as well as 'what-is-not' taken
in a certain sense, viz. 'what is not-white' or 'what is not-good'.
In my view, it is wrong to assume that the phrase 'holding
indiscriminately of what is and what is not' refers to the existence
or non-existence of the subject of the sentence in which the
indefinite PllJlU is actually used. Rather, only the applicability of
the indefinite PilJla's own significate itself is concerned, irrespective of its actually occurring in a sentence. Of this significate it is
said that it either applies, or does not. 45 In this connection it should
be kept in mind that in chapters 2 and 3 Aristotle is discussing
OVOJlU and PllJlU as potential ingredients of a statement-making
expression and thus he focusses on the things they refer to, each in
its own specific way. Of course, to determine what, in accordance
with their specific nature, each of them contributes to the A.6yo~,
and the statement-making expression when "is' or 'is not' is added'
(16a17-8), remains the ultimate aim of the observations made in
these chapters.
The other candidates for the label 'pilJla in a qualified sense' are
its inflexions. In fact, they fulfil the condition of being said of
something else, but do not meet the requirement of holding now,
that is, at the time of the utterance, but refer to the inherence of the
characteristic in the past or the future. Ackrill is right in remarking that the opposition of the inflexions of the PllJlU to the ordinary
Pll JlU is not at all parallel to that between names and their
inflexions. Unlike name inflexions, which do not refer to what the
name stands for, and only indirectly connote it, the inflexions of
PllJlU really refer to the same entity as the PllJlU, albeit with a
different time-reference.
The two first lines of the next passage (16b19-20), which deals
with the PllJlU when uttered just by itself, give rise to less difficulty
than is commonly assumed, if only they are interpreted along the
forementioned lines. The phrase 'when uttered just by itself is
wrongly taken to mean 'when taken out of a certain statemental
context' .46 Rather it only is about the absence of the two features that
4 5 Therefore, the reference made by Ackrill (1963, 120) and Weidemann
(1994, 177) to Cat. 10, 13b27-35, where it is said that if Socrates does not exist
'Socrates is not sick' is true, is beside the point.
46 Cf. Weidemann (1994) 178.

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETATIONE

129

make up the specific difference between OVOIJ.CX and Pll!J.CX, viz. the
latter's additional time-reference and its being said of something
else, including its attributive use, as e.g. in 'white man'. On this
assumption, the Pll!J.CX is merely a name, because, just like a name,
it only brings up its significate as an on tic value, 47 detached from
its being immanent in something else and without implying the
subject-substrate of which it is said. The fact that the Pll!J.CX in that
position does signify something does not yet eo ipso imply that the
entity involved is, or is not, the case. This is quite in line with what
has been said earlier (16al3-8) about OVOIJ.CX'ta and p~IJ.a'ta, 'when
nothing further (viz. 'is' or 'is not') is added'.
What remains to be discussed are the final sentences of this
chapter, in which the author tries to support his claim about the
Pll!J.CX when taken on its own, by an 'anticipative refutation'48
concerning the verb 'to be'. Not even the verb 'to be' (or 'not to be'),
which on the face of it seems to really imply 'being-the-case',
affords a counter-example to invalidate my claim, Aristotle
argues,49-not even if you use, without a further addition, the term
('be-ing'), which qua substantivated neuter word does imply the
idea of something actually being there.5 Why not? Because,
according to Aristotle (16b23-5), even the verb 'to be', including its
nominal participle, 'is nothing' (an empty notion, that is), and
merely helps to signify51 a certain cruv8ecrt<;, which cannot be
conceived of without the things combined.
A final remark should be made on the rendering of Pll!J.CX. The
rendering 'verb' does not reflect adequately the semantic function

ov

4 7 In Aristotle's words: 'it signifies something: the speaker arrests his

thinking and the hearer acquiesces (i.e. 'comes to understand')'. Aspasius
rightly defends this interpretation on the level of name-giving: when
hearing a single word, the hearer's mind is not in suspense and, like the
speaker, the hearer has a determined entity in his mind. This operation has
nothing to do as such with framing statements; see Boethius II, 74.9-33. Cf.
Aristotle, Phys. VII 3, 247bll-2: 'We are said to know and to understand
owing to our thought coming to acquiesce and arriving at a standstill.' For
the historical context of the expression see De Rijk ( 1986) 268, n. 26.
48 I refer by this term to Aristotle's habit of supporting a claim by invalidating a seemingly strong counter-example. Cf. Ackrill (1963) 114; 123;
Weidemann (1994) 156; 163; 181.
4 9 I have to leave Weidemann's (1994, 180-5) interesting attempt to
emend the locus curruptus 16b21-3 out of consideration now.
50 The sentence
f.av
ov et1tU~ 'lftAOV should be read in parenthesi, pace
Weidemann (1994) 186.
5! For this sense of 7tpOO"O"TUUltVE1V cf. ch. 10, 20a13. For cruv6ecrt~ see below,
p.131-133.

ouo'

to

130

L.M. DE RIJK

Aristotle has in mind when using pflJ.La, because it indiscriminately stands for any attributive determinant of an entity referred
to by a name, whether it is nominal52 or verbal. Therefore, the
neutral rendering 'attribute' seems to be the most appropriate in
English.53

6. On the semantics of A.6ro~ (both 'statable complex' and 'statement?

Chapter 4 opens with the definition of A.6yo~, the definiens of
which has been arranged in a similar way as that of OVOJ.L<X : 'a
significant articulate sound some part of which is significant in
separation-as an expression, not as an affirmation' ( 16b26-8). This
addition, which aims at marking off the main difference between
Myo~ and its components, oVOJ.L<X and pflJ.La, is explained by Aristotle
in the next lines where he points out, quite in line with 16b21-2,
that e.g. the name 'man' taken as part of a A.6yo~ signifies something, but leaves the question of its being the case undecided,
whereas when 'something' (viz. 'is' or 'is not') is added, it makes
up an genuine affirmation or negation. Next, the non-statementmaking significant expressions, such as prayers, threats and
commands54 are ruled out, because they all lack truth-value.
The Greek word A.6yo~, which is the verbal noun deriving from
a root that, roughly speaking, always connotes an orderly procedure, discernment, or the speaker's intention to 'clear up' some
situation, has several meanings. 55 One of its primary meanings is
'account', which includes 'speech' (as a means of 'giving an
account' or 'accounting for'), as well as the more sophisticated
'argument'.
It should be kept in mind, however, that A.6yo~ does not stand as
such for 'statement', and may also refer to phrases, such as 'beautiful horse' (16a22) or a definiens, such as 'two-footed land animal'
(17a13). As is clear from chapter 5, 17a9ff., the sentential functor
'is' or 'is not' should be added to the name and the attribute-which

52 Nominal i>~wrca. are found at 16a15; 20a32; 20b2.

53 The rendering 'Aussagewort', which some German interpreters rightly prefer to 'Verb' or 'Zeitwort', hits the mark, provided it is not associated
(as in Weidemann, passim) with the notion 'sentential predicate'. That this
association is wrong is clear from 1,16a15, as well as from its function within
the A.6yo~. which has a broader sense than 'sentence'; see our next section.
54 Cf. Poet. 19, 1456bl0-l.
55 For a survey of its various senses see De Rijk (1986) 225-31.

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETATIONE

131

together make up a A.6yoc; or 'assertible (or statable) complex'-to

turn the A.6yoc; from simply an assertible complex into a statementmaking expression. It is really important to distinguish between
these two senses of A.6yoc;. 56
In order to make sense of Aristotle's semantics of A.6yoc; it is very
important to understand correctly the anatomy of the Aristotelian
statement-making expression. The Aristotelian proposition is
commonly taken to be an expression with the (modern) 'S=P '
structure, to the effect that, like the copula in 'S=P ', the functor 'is'
('is not') is held to be dyadic, and to act as that which combines
( copulat in Latin) the subject-term with the predicate-term. However, there is an alternative view57 to the effect that the verb 'is' is a
monadic functor which, when added to an 'assertible or statable
complex' (i.e. a A.6yoc; expressing a state of affairs), makes up a
statement by solely asserting that this state of affairs is really the
case. In my view of things, any dyadic function (as 'copula') is out
of the question, since the A.6yoc; by itself is already a composite
expression combining ovojla and Pllll<l. Much evidence may be put
forward in support of this view.
[1] To begin with, Aristotle never speaks of 'is' as copula, nor is this
verb ever brought up under a special name. Besides, 'is' and 'is not'
are always said to be added to the combination already formed of
an ovojla and a Pllll<l, whereas there is never any suggestion that
Aristotle takes these functors as themselves performing the combination of an ovojla and a Pillla. 58
[2] Secondly, it is relevant to recognise what Aristotle means
when he says (at 16b24) that 'it (the verb 'is', and 'is not' as well (!);
see 16b22) helps to signify some CJUV8E<Hc;, which cannot be
thought of without the things combined.' Now, the word cruv8Ecrtc;
is basically used either to refer to a compound structure ('composite') of physical things or to the mental act of uniting two or
more concepts. This logico-grammatical use is twofold: either
cruv8ecrtc; only refers to an (alleged) composite, or it asserts this composite as being the case or 'obtaining'. 59 The term Ota ipEcrtc;,-which in
56 They are opposed by Aristotle as Myo~ and Myo~ anocpaV'tll(O~ (17a8ff.; for
Plato's Myo~ dp~JlEVO~ at Tht. 189a3-4, see De Rijk ( 1986) 309-16).
57 Matthen (1983) 125-31; De Rijk (1987) 43f.; 50-2; 60, n. 60.
58 Notice that in as far as 'is' is concerned, Aristotle always uses (16a15;
16a18; 16b30; 17a12; 19b25; 19b30;19b38;20a9; 20a14; 20a36; 21a21;21b6; 21b21;
cp. 21b27-30) the passive verb npoa'ti8Ea8at, never the active verb auv'tieEvat.
59 See De Rijk (1987) 45-50, where more evidence from Aristotle's other

132

L.M. DE RIJK

De int. occurs only once, viz. in the opening chapter (16a12-4)-,

must be explained as the denial of the cruv8ecrt<; referred to by the
assertible complex, rather than as a separation of its components.
Thus, the procedure of speaking truly (aA.TJ8euetV) and speaking
falsely ('lfEUOecr8at) is basically a matter of asserting or denying
rightly-or wrongly, respectively-an assertible complex, to the
effect that to assert or to deny amounts to giving one's assent to, or
dissent from one and the same assertible complex, respectively.
For example,
'obtains (obtained, will obtain) [man plus paleness]'
or
'does (did, will) not obtain [man plus paleness]',
meaning indiscriminately *'a man being pale is (was, will be) the
case' (='a man is [was, will be] pale') or *'a pale man existing is
(was, will be) the case' (='there is [was, will be] a pale man'). 60
A happy consequence of the alternative view is that the term
t>1tOKetf.1EVOV can be taken to stand for 'substrate' (referred to by the
name), rather than for sentential subject, as the 'S=P' device would
suggest. Thus, it is not necessary to oppose, as Weidemann does, 61
the use of this term at 19a38 to that at 21b28-2, where his rendering
'sentential subject' will not do at all.62
Aristotle's semantics is often censured, but a lot of criticism may
turn out to be beside the point, if the alternative interpretation is
followed. Here are three examples:
-Aristotle is commonly reproached for failing to distinguish
between the 'copulative' and 'existential' uses of 'to be' 63 . This
distinction, however, turns out to be anachronistic, because the
functor meaning 'obtains [ (x )f)' is monadic and combines both
uses indiscriminately.
works is adduced.
60 For the whole statement procedure see De int. 1, 16a12-8; A. Pr. I 46,
52a24-38; Metaph. l:l 7, 1017a31-5; E 4, 1027b18-1028a1; De an. III 6, 430a27-b6
(this passage includes time-differentiation), and the discussion in De Rijk
(1987) 48-50. At Phys. v 1, 225a21, non-being taken KCX'ttl cruv9EO'tV ~ litcxtpEcrtv
appears to concern 'that-which-is-not' in the sense that it is the assertion of a
false assertible complex or the denial of a true one. Cf. Metaph. E 2, 1026a35;
E> 10, 1051b1; N 2, 1089a28.
6! Weidemann (1994) 158.
62 'Being' is said (at b29) to act as llltOKElJlEVOV to the modifiers Ouvcxcr9cxt
and EvOEXEcr8cxt in a similar way as in 'being a pale man', 'pale' and 'man'
do; see De Rijk (1987) 35 and 55, n. 25.
63 Ackrill (1963) 119; 123; also Weidemann (1994) 174; 190.

ON ARISTOTLE'S SEMANTICS IN DE INTERPRETA TIONE

133

-Ackrill's 64 claim that Aristotle fails to distinguish between the

predicative and the assertive functions of the pf\J.w. (as used in a
statement) is also pointless, for a similar reason; see our foregoing
analysis of the statement-making expression.
-The problems raised by Ackrill and Weidemann65 concerning
Aristotle's (putative) use of the indefinite pf\J.ux in statements disappear as soon as the 'S=P' structure is abandoned.

7. Conclusion
Aristotle's semantics of the statement-making expression as occurring in De interpretatione is quite modest indeed, but serves its
purpose well enough for the narrow scope of issues he has to deal
with, given his main intention, viz. to set against Plato's ontology
his own view of how to speak about 'things' that really are the
case. 66 The semantics as set out in De interpretatione is suitably supplemented by the semantics of name-giving found in the Categories
and the doctrine of categorisation 67, which enables Aristotle to
develop a method of scientific inquiry, without having to take
recourse to Platonic transcendent Forms.68

BIBLIOGRAPHY

( 1) Primary works
Aristoteles graece ex recensione I. Bekkeri. Edidit Academia Regia Borussica. Aristotelis Opera, vol. I-II (Berlin 1831; reprint Darmstadt 1960)
Ross, W.D. (ed.), The Works of Aristotle translated into English. (Oxford, 1928ff.)
Aristotelis Categoriae et Liber de interpretatione . Recognovit brevique adnotatione
critica instruxit L. Minio Paluello (Oxford, 1949 19562)
A.M.S. Boethii Commentarii in lilnum Aristotelis IIepi. EPflTJVetm;. Recensuit Carolus
Meiser. Pars posterior secundam editionem et indices continens (Lipsiae
1880)

64 Ackrill ( 1963) 119; 150-1.

65 Ackrill (1963) 120-1; Weidemann (1994) 177. See also above, p. 126.
66 For Plato's view see Soph. 260e4-264b9, and De Rijk (1986) 194-354. It is
noteworthy that in the technical treatise A.Pr. the syllogistic premisses are
not framed along the lines of De int., but by using the verb imapxetv instead of

dvat.

67 See De Rijk (1988) passim.

68 See De Rijk (1995) 78-102.

134

L.M. DE RIJK

(2) Secondary literature

Ackrill, J.L. (1963) Aristotle's Categories and De Interpretatione. Translated with
Notes (Oxford).
Bonitz, H. (1870) Index aristotelicus (Berlin 1870; reprint Graz 1955).
Kardaun, Maria S.J.M. (1993) Der Mimesisbegriff in der griechischen Antike. Neubetrachtung eines umstrittenen Begriffes als Ansatz zu einer neuen
Interpretation der platonischen Kunstauffassung. Koninklijke Nederlandse Akademie van Wetenschappen. Verhandelingen Afdeling Letterkunde,
Nieuwe Reeks, dee! 153 (Amsterdam/Oxford/New York).
Kuhner, R. - Gerth, B. (I, 18983 ; II, 19043 ) Ausfuhrliche Grammatik der griechischen Sprache. Zweiter Teil: Satzlehre. 2 Bde (Hannover/Leipzig; reprint
Hannover 1983).
Matthen, M. (1983) 'Greek Ontology and the 'Is' of Truth', Phronesis 28: 113-35
Montanari, E. (1988) La sezione linguistica del Peri hermeneias di Aristotele. Vol. II:
Il commento (Firenze).
Rijk, L.M. de (1980) 'On Ancient and Medieval Semantics III: The Categories as Classes of Names', Vivarium 18: 1-62.
- - ( 1986) Plato's Sophist. A Philosophical Commentary. Koninklijke Nederlandse Akademie van Wetenschappen. Verhandelingen Afdeling Letterkunde,
Nieuwe Reeks, dee! 133 (Amsterdam/Oxford/New York).
- - ( 1987) 'The Anatomy of the Proposition. Logos and Pragma in Plato and
Aristotle', in L.M. de Rijk and H.A.G. Braakhuis (eds.), Logos and
Pragma. Essays on the Philosophy of Language in Honour of Professor
Gabriel Nuchelmans, (Nijmegen) 27-61.
--(1988) 'Categorization as a Key Notion in Ancient and Medieval Semantics', Vivarium 26: 1-18.
- - (1995) 'Ockham as the Commentator of His Aristotle: His Treatment of
Posterior Analytics' in Fernando Dominguez, Ruedi Imbach, Theodor
Pindl and Peter Walter (eds.), Aristotelica et Lulliana magistro doctissimo
Charles H. Lohr septuagesimum annum feliciter agenti dedicata (The
Hague) 77-127.
Trendelburg, F. ( 18929 ) Elementa logices aristotelicae (Berlin).
Weidemann, H. (1994) Aristoteles, Peri Hermeneias. Uebersetzt und erlautert.
Aristoteles. Werke in deutscher Uebersetzung hrsg. von Hellmut Flashar.
Band 1, Teil II (Berlin).
Wieland, W. ( 1970) Die aristotelische Physik. Untersuchungen iiber die Grundlegung der Naturwissenschaft und die sprachlichen Bedingungen der
Prinzipienforschung bei Aristoteles. 2. durchgesehene Auflage mit
einem Nachwort (Gottingen).

THEOPHRASTUS FR. 184 FHS&G:

SOME THOUGHTS ON HIS ARGUMENTS
IAN

G. Kino

This is an old hobby horse, much ridden and rightly so, for
arguments concerning the eternity or destructibility of the cosmos
became a stock topic over a long period of the history of Greek
philosophy, the response and arguments coloured by the philosophical stance of each proponent. In modern scholarship, this fragment, located at Philo, De aeternitate mundi 117-149, has evoked
widely, not to say wildly conflicting views with regard to context,
provenance, influence and reliability from scholars such as Zeller
and Von Arnim on, 1 and has been much revived in the last
decade, with an important article by David Sedley on 'Theophrastus and Epicurean Physics' about to appear. Since Jaap Mansfeld
has himself richly engaged in the struggle, 2 it will seem imprudent or even impudent for an intruder to this particular paddock to
offer him some thoughts on the matter. The imprudence will be
only too obvious to him, but I hope he may excuse it remembering
our common fascination with the problems of fragmentary
evidence and a shared appreciation of their extreme difficulty and
danger; and the impudence will have to be forgiven in the light of
a long and much-valued friendship. And my contribution is very
limited, and skates over or ignores many detailed issues.
As with all fragments, it is best to keep distinct reporter, context
and reference. Since the only reporter who names Theophrastus
for these arguments is Philo, not only should we allow for Philonic intrusion, but also in general distinguish later influence of the
1 E. Zeller, 'Der Streit Theophrasts gegen Zeno iiber die Ewigkeit der
Welt', Hermes 11 (1976) 422-9; H. von Arnim, SW 1.106; but on the other
hand, H. von Arnim, 'Der angebliche Streit des Zenon und Theophrastos',
Jahrbii.cher fur classische Philologie, 39 (1893) 449-467.
2 E.g. 'Providence and Destruction of the Universe', Studies in Hellenistic
Religions, ed. MJ. Vermaseren (Leiden 1979) 129-188; 'Physikai doxai and
Problemata physika' in Theophrastus, His Psychological, Doxographical and Scientific
Writings, Rutgers University Studies in Classical Humanities 5 (New Brunswick 1992) 63-111; 'A Theophrastean Excursus on God and Nature and its
Mtermath in Hellenistic Thought', Phronesis 37 (1992) 314-335.

136

IAN G. KIDD

arguments themselves from the original Theophrastean Sitz im

Leben. To begin with, Philonic authorship has been denied for Aet.,
but I am persuaded that David Runia's arguments for genuineness
are valid;3 so we do have a general context of authorship, and this
offers some control over language and presentation.
The presentation of the arguments should first briefly be put in
local context. Aet. begins (like Aristotle's De Caelo 1.10, but quite
differently) with a tripartition of views: that the cosmos is created
and destructible, for which Democritus, Epicurus and most Stoics
are named (7-9); that it is uncreated and indestructible, citing
Aristotle as innovator, or perhaps Pythagoreans (10-12); that it is
created but indestructible, the view of Plato (13-16). Mter interpretations of the Platonic position, 4 Philo launches into a main
exposition of the Peripatetic view and argument (20ff). But it
becomes clear that the main attack is against the Stoa (4 7ff). At 55
he refers to the theories of the 2nd cent. BC Head of the Peripatos,
Critolaus. At 75 he returns to the main attack on the Stoics, citing
Boethus of Sidon, Panaetius and Diogenes of Babylon, seeming to
add incidentally at 113ff an attack on atomistic modes of destruction. Then it is at that point (117ff) that Theophrastus is brought on
with his classification of the four main arguments used by proponents of the coming-into-being and destructibility of the cosmos,
and the counter arguments against these.
What is immediately obvious is the 2nd cent. BC weight of
evidence in Philo. Plato is there, of course, but Timaeus was common coin. But the case of Aristotle is curious. He is named only
once as the possible innovator of the uncreated and indestructible
theory (10-12). When we come to the main arguments for this
(20ff), Philo appears to know nothing of the important arguments
in De Caelo 1.10-12, but presents others which have now been
classified as fragments of Aristotle's De Philosophia (F19abc Rose
and Ross; F916 Gigon). This may well be so, but no particular
source is named. Earlier (10-12), where Aristotle was named,
Philo seems to make a point of saying that what he had read was
Ocellus, presumably meaning the Ilepl. 'tfi<; 'tOU nav'to<; cpucrero<;,s a
work again of the mid-2nd c. BC apparently, and something of a
3 David T. Runia, 'Philo's De Aeternitate Mundi: The Problem of its Interpretation', Vigiliae Christianae 35 (1981) 105-151.
4 For the strategic importance of this for Philo, see Runia, art. cit.
5 Ed. R. Harder, Neue Philologische Untersuchungen 1 (Berlin 1926).

THEOPHRASTUS FR.

184

FHS&G

137

Peripatetic/Pythagorean mixture (though Philo regarded it as older

than Aristotle). So clearly it has to be kept in mind that the
Theophrastean evidence may be filtered and embroidered.
The fragment itself is structured in three parts: (1) (lines 1-4,
11 7) 6 the explicit statement that Theophrastus said that those
asserting the coming-to-be and destruction of the cosmos were led
astray by four arguments cited as the greatest or most important
(J.!Eytcr'ta), then succinctly listed: (i) unevenness of the land; (ii)
withdrawal (nvaxffipllcrt~) of the sea; (iii) dissolution (ouiA.ucrt~) of
each of the parts of the whole; (iv) destruction (c:p8op6.) of whole
classes (lca'ta yev11) of land animals. (2) (4-89, 118-131) The
second part fleshes out each of the four arguments in turn. This
part is led into by indirect speech still dependent on Theophrastus
(4, 118), but immediately lapses into straightforward narrative
indicatives, interspersed with things like 'as I said' (12, 119), and
'they say' (18, 120; 80, 130). (3) The third section (90-204, 132149) presents arguments to counter the previous arguments, again
in order. The language at least of this section could quite naturally
be taken to be from Philo, but it is Philo's practice to present the
arguments of others in his own way. 7 Clearly we have to accept
Part (1) as Theophrastus, but in what follows, how much is Theophrastus, and how much Philo? This obvious question has interested scholars since Von Arnim accepted only the first sentence as
Theophrastus. 8 Given the complexity of transmission, the answer
is unlikely to be simple.
There is little doubt Parts (2) and (3) have at least been contaminated linguistically, and in detail, and by embroidery. Certainly
Philo expresses himself in his own language. Much has been
made of the Stoic vocabulary in Argument (iii) (45-52, 125) as
being also Philonic. But I can give a more clear-cut example. In
Argument (i) we read that if land did not come to be but was
always there from eternity, it would have been flat, worn smooth
6 In references to the fragment, I give the line numbers of F184 FHS&G
(on this edition see below n. 25) followed by section numbers of Philo, Aet.
7 Runia, art. cit. (n.3) 130.
8 Von Arnim, art. cit. (n.1). Compare also, J.B. McDiarmid, 'Theophrastus on the Eternity of the World', TAPA 71 (1940) 239-247. W. Wiersma
(Mnemosyne III 8 (1940) 234-243) argued again that only Part (1) can be
trusted, since Parts (2) and (3) can be shown linguistically and in prose
rhythm to be strongly Philo, and they do not always match consistently with
(1).

138

IAN G. KIDD

by the rain (12-13, 119). But the Greek for 'flat' is not, as one
would have expected, oj.LaAO<; or 1tdHov, but AEcocp6po<; (13, 119):
everywhere would have been broad highway, or motorway. This
extraordinarily flamboyant phrase9 brought me up short until I
discovered that it was a favourite metaphor in Philo for the smooth
main highway leading to virtue and to God. 10 Of course, later
language does not disprove earlier argument. But some of the
language seems tied up with the argument, like the Stoic (or
Philonic) vocabulary in Argument (iii) (see below). And indeed
some arguments themselves, such as the latter part of the extended
argument in Argument (ii), appear out of tune with the earlier part
(see below). There are suspicious details of examples or illustrations which reappear in later popular clusters (e.g. in the counter
arguments for Argument (ii), see below); and most notoriously,
from animal marvel lore, the extended with graphic and technicolour detail but hardly apposite illustration of nineteen lines to
Argument (iii), of snakes in India twining themselves round
elephants, and by draining their life blood, killing them, only to
perish themselves when the dying elephant fell on top of them
(59-77, 128-129). It is true that Theophrastus himself was into
marvel literature (Frs. 218, 236 FHSG); but much of the strange
illustrative detail in Philo has its cluster of popular echo in
Diodorus Siculus, Strabo, Seneca (NQ), Pliny, Lucan, and Aelian.
But, of course, later embroidery of detail or illustration in the
expanded arguments of Part (2) does not disqualify the spines of
argument from being original. Indeed the transmission of philosophical arguments, like the transmission of anecdotes, invites
such embroidery. And it can be proved with certainty that the
expansion of arguments in Part (2) does not come from Philo,
because the extended argument (iii) is reproduced in Lucretius 5.
235-317, and Philo's extended argument (iv) recurs in Lucretius
5.324-350. Furthermore, these two arguments follow in Philo and
9 In earlier philosophical Greek I can only bring forward the Pythagorean yvro1111= 1:a~ 't A.Ero<p6pou~ lllt Paoi1;;uv (58C6, 1.465.32 D-K; D.L. 8.17). Philo
himself quotes it at Prob. 2.
1
For the metaphorical use relating to the highway leading to virtue and
to God: Opif. 69, 144; Leg. III 253; Post. 31, 102, 154; Deus 61, 143, 165, 182; Agr.
101; Conf 19; Her. 70; Congr. 28; Fug. 203; Abr. 7;jos. 142; Mos. II 138; Decal. 50;
Spec. I 355; Spec. II 202; Spec. III 185; Spec. IV 62, 155, 167; Virt. 51; Legat. 216.
Philo also uses Aro<p6po~ literally, e.g. Det. 2; and twice referring to Moses'
parting of the waters: Mos. I 177; Contempt. 86.

THEOPHRASTUS FR.

184

FHS&G

139

Lucretius in the same immediate sequence. There must therefore

be an ultimate common source. The common source cannot be
Epicurus, because Philo's source is certainly not Epicurean, but
anti-Epicurean, and rather Peripatetic/Pythagorean with noticeable
2nd c. BC influences. Also, apart from embroidery, the framework
of argument in Argument (iii) on destruction of elements, for
example, is different in Philo and Lucretius. The evidence suggests
the possibility of different intermediary sources or influence, but a
common ultimate source; and the simplest solution for that is
Theophrastus.
As for the counter arguments of Part (3), they were the whole
point, not only of Philo introducing Theophrastus' four arguments,
but of Theophrastus in the first place introducing the four most
important arguments so that he could counter them, because the
arguments supported a dogma to which he was opposed. So, I am
prepared to believe that at least the spines of the arguments do
derive from Theophrastus.
If the argument is granted so far, it would be interesting to
survey very briefly the peculiar characteristics of the four arguments themselves with their counter arguments, attempting carefully to distinguish original spine from later accretions. This is
particularly important because of an initial major misdirection in
the study of this evidence by Zeller, 11 who assigned the arguments
selected and attacked by Theophrastus to Zeno of Citium; and the
whole fragment was then sanctified by Von Arnim as a Stoic
fragment (SW 1.106).12 We must, however, distinguish between
the Theophrastean Sitz im Leben and the subsequent history of the
argument. Both are interesting, but should not be confused. So, to a
brief review of the arguments and their counters:
Argument (i) (4-17, 118-119) is a cracker. If land did not
come to be but was there from eternity, it would have been worn
smooth with no mountains or hills; for rain from eternity would
have worn it level (4-17, 119). Although, of course, mention of the
destructive power of rain on earth occurs elsewhere,I 3 its use as a
cosmic argument is not surprisingly lacking elsewhere. It smacks
II Arl. cit. (n .1 ) .

I2 A position still entertained until at least quite recently: e.g. A. Graeser,

Zenon von Kition (Berlin 1975) 187-206.
I3 E.g. Lucr. 5.255f; but there it makes earth pass into another element
(water), and so relates more to Argument (ii).

140

IAN G. KIDD

of rhetorical debate. It is however a geological argument derived

from observation, and is countered by meteorological observation
and theory: the continual but slow (and hardly perceptible)
cyclical evolution of geographical features, whereby mountains
are not only worn down but recreated by fire and earthquake (93ff,
132ff).l4
Argument (ii) (18-40, 120-123) depends on the observation of
the constant diminishing or withdrawal of the sea. This is again
established by geographical evidence: rising of islands, shell
deposits inland, etc. All this occupies nineteen lines of text. The
disturbingly brief concluding argument takes up only four lines: if
the sea diminishes, so too the earth will diminish. This seems
crazy on observational grounds; if sea diminishes, the land area
increases, and vice versa. But we now seem to be talking of sea,
earth, air diminishing in the sense of being used up (O.vaA.ffi8ftcre'tat, 38, 123) and ending up in the one oucr{a, fire. Now, ignoring
the fearful confusion between geographical/ meteorological phenomena, elements and principles, this seems a brief Stoic tail later
pinned onto, and wagging, the donkey of the first part. For after all,
the first part is already in Aristotle. Indeed, that the sea had a
beginning, and the wetness finally all dried up by evaporation,
was attributed by Theophrastus himself to the Presocratics Anaximander and Diogenes of Apollonia (according to Alexander,
F221 FHS&G), a theory countered by Aristotle in Meteor. Il.lff. But
the actual tying of the argument of the sea drying up to the
creation and destructibility of the cosmos is countered already by
Aristotle in Meteor. 11.3, 356b6ff. And he specifically links the argument of sea perishing by drying up to Democritus (356b10ff =
68A99a Diels). As for the observational evidence for the sea drying
up and changes of sea and earth areas, Aristotle cites it (Meteor. 1.14,
351a19ff) as evidence for an orderly cycle where one geological
element reciprocally gives way to another. Interestingly, he has
no mention of Theophrastus' inland shell deposits, but they do turn
up in Olympiodorus, Comm. on Meteor. on 351b22 (p.ll6 Stuve). The
Theophrastean counter arguments (129-158, 138-142) are again
based purely on geographical observation of sea encroachment. It
is true that the examples chosen by Philo were later popular: for
Sicily being previously attached to Italy: D.S. 4.85.3; Strabo 6.1.6;
14

Earthquakes also certainly interested Theophrastus.

THEOPHRASTUS FR.

184

FHS&G

141

Sen. NQ 6.30.3; Pliny NH 3.86; but these make clear that the
tradition is an old one. For Helice and Bura exampled as cities
swallowed by sea, compare Polyb. 2.41.7; Strabo 8.7.2; Ovid, Metam.
15.293f; Pliny NH 2.206; Pausanias 7.24.5f and 12f; 7.25.8; Amm.
Marc. 17.7.13. Interestingly for Theophrastus, the disappearance of
Helice and Bura was also associated with earthquakes as well as
sea. But although popular around Philo's time, the cited instances
of Bura and Helice are referred by Seneca (NQ 6.23.4; 7.5.3-4)
specifically to Aristotle and Callisthenes, Aristotle's nephew. The
story of Atlantis (150ff, 141) of course comes from Plato, Tim. 24E,
25CD. 15
Argument (iii) (41-79, 124-129) seems a more general argument, that of parts/whole correspondance in genesis and destruction. If parts are perishable, the whole will perish. But all parts of
the cosmos are perishable, therefore the cosmos is perishable. This
is illustrated by running through the four elements, but again by
observation as meteorological phenomena: the strongest stones
crumble and dissolve; water if stagnant become noisome; air can
become tainted, fire extinguished. Historical reference for this
argument surfaces strongly in hellenistic philosophy. It is certainly Epicurean, for it recurs in considerable detail in Lucretius 5.235317; but Stoic too, for Philo presents it in Stoic garb. It is true that
some of the Stoic vocabulary is also Philonic, but it can hardly be
denied that the framework is also Stoic in physical destruction
through the loosening of the 7tVEUJla'tuco<; 1:6vo<; as E~t<; ( 4 7), and the
death of water as if the removal of soul. In any case the argument
is independently given as Stoic by D.L. 7.141 ( SVF 2.589) . 16
However, the main Theophrastean counter argument to this (160ff,
143), is that Argument (iii) would hold if and only if all parts
perished together at the same time. But that is not so in the cosmos
when an element individually changes to another. But that is
normal Aristotelian doctrine frequently expressed,l7
15 D.T. Runia, Philo of Alexandria and the Timaeus of Plato, Philosophia
Antiqua 44 (Leiden 1986) 84f thinks the reference may come from Philo
rather than Theophrastus. But this is debateable, and even if it does come
from Philo, does it matter?
16 Which incidentally displays exactly the same reaction to the Theophrastean counter argument that Sedley claims for Lucretius and Epicurus, i.e.
change of elements into each other is destruction.
17 Ar. GC 11.4, 33lal2ff; De Caelo I.IO, 280al6ff; III.6, 304b23ff; Meteor. II.3,
357b27ff.

142

IAN G. KIDD

Argument (iv) (80-89, 130-131) is based on the coevality of the

human race and the cosmos: if the cosmos were eternal, the
human race would be eternal. But man's origin is recent because
the arts and crafts must be coeval, and they are clearly recent.
Therefore the cosmos was generated. 18
Again, this argument shows up in Epicurean garb at length in
Lucretius 5.324-350. But it is hardly confined to that. The argumentative link between the eternity of the human race to the
eternity of the universe is in Ocellus, TI~:pt 'tll<; wu na.v'to<; qn)cr~:ro<;,
38ff Harder, whom Philo had been reading, and so is likely to
have gone further back in Peripatetic discussion. Anyway, the
counter arguments to this, dissociating the history of the two by
maintaining that the human race is continuous, although liable to
partial destruction by fire and water, while the arts are periodically
destroyed and have to be recreated, certainly have earlier reference. The cycles of destruction from fire and water are linked to
Philolaus (Stob. 1.20.1, p.172.6ff W.); 19 while the continuity of
human beings but destruction and rediscovery of 'tEXVa.t are
already in Plato (e.g. Tim. 22c-23b; Laws 3.677a-d), and Aristotle
(Meteor. 1.14, 351b8ff; Metaph. A.8, 1074b10ff; PoL 11.8, 1269a4ff).
Even in such a brief and superficial skating over the four
arguments selected by Theophrastus, and even with the savagely
curtailed evidence available to us today, what strikes one is the
range of philosophical provenance involved. The allocation of the
arguments to precise origins seems impossible, and the attempt to
do so perhaps misguided. I strongly believe that we oversimplify
the historical transmission of ideas, forgetting not only written
evidence we no longer possess, but also the strong influence of oral
communication and discussion. The form and range of the arguments suggest that Theophrastus probably selected types of argument that were cullable from the whole range available to him
(after all they were the four JlEytcr'ta.); i.e. from predecessors (socalled Presocratics, Plato), his own master (Aristotle), and surely
contemporary models from debate, which might include incipient
Epicurean and Stoic discussion; I do not mean settled dogma, but
exploratory oral discussion.2 It need not even refer to full-blown
It is interesting to note that the logical form of the agument is: P, Q;
.: -P, i.e. modus tollens. So also Argument (i).
9 See now Carl A. Huffman, Philolaus of Croton (Cambridge 1993) 26lff.
20 We know that much of Zeno's teaching was oral; I.G. Kidd, Posidonius,
18

-~

THEOPHRASTUS FR.

184

FHS&G

143

theories, nor even to influence in one direction or another, but

rather centred on contemporary reaction to a common topic. This
was certainly one such, namely 7tO'tEpov KOOJ.!O<; <xtOto<; T1 ou, as we
perfectly well know from Aristotle, Top. 105b24; 104b9, b17. And
that context ( Top.1.14) makes clear that such propositions were the
subject of common formal debate, both rhetorical and philosophical. There were clearly a number of such fashionable common
topics producing lively reaction in the contemporary Schools, and
not only in natural philosophy. For example, in moral philosophy
one could hazard: the relativity of goods; whether virtue can be lost
or not; the moral status of anger. Moreover, the structure of the
Theophrastus fragment in Philo, comprising several arguments
with their counter arguments supplied for debate, is typical of
Theophrastus, as Jaap Mansfeld has shown.2I
That is old ground. What is much more interesting is why one
should choose these four arguments to illustrate the topic of the
eternity or destructibility of the cosmos; and they are underlined
as the four most important (J.!Eytcr'ta). For they are indeed a strange
collection. Furthermore, the stark fact is that the character of the
arguments as given by Theophrastus and their counters, is in the
sharpest contrast to those not only of Plato in Timaeus, but, and
more relevantly, to those of his master Aristotle on the subject.
Aristotle's examination of the different views on the topic, and his
own arguments in De Caelo 1.10-12 are constructed completely
through logical and conceptual analysis, and opponents are
attacked for logical inconsistency. And this is also true, although
the details are different, of the so-called 'Aristotelian arguments',
usually taken to be from De Phil., from earlier in Philo, Aet. 20ff.
But in marked contrast, the arguments that were selected by and
appealed to Theophrastus for selection, and so also the arguments
(which is perhaps even more important) by which they were
countered, depend on observation of geological, geographical and
meteorological phenomena with empirical appeal of instance, not
on metaphysical argument. I have argued at length elsewhere,2 2
that this is precisely the characteristic approach of his Meteorology.
So this attempt to penetrate to the Sitz im Leben of the fragment

vol.II Commentary on Fl65.4EK; cf. T91, T99, Fl66EK.

21 Mansfeld, 'Physikai doxai and Problemata physika', (see above n. 2) 67ff.
22 'Theophrastus' Meteorology, Aristotle and Posidonius' in Theophrastus, His
Psychological, Doxographical and Scientific Writings, 294-306.

144

IAN G. KIDD

suggests that it tells us more about Theophrastus himself than about

the arguments or their sources.
As for the Nachleben of the arguments, the result is even more
striking, for far from being transmogrified into a Stoic fragment, it
seems to have been the Epicureans who took most account of it.
This is no monopoly, of course, for the evidence indicates derived
grounds of argument gong on in all the major Schools, on what
after all continued to be a central topic. But the extensive use of
Theophrastus' arguments by Lucretius chimes in exactly and
significantly with what happened to Theophrastus' Meteorology,
about which we now know much more from the invaluable find
of the Arabic translation extract, edited by Hans Daiber.23 I have
argued elsewhere,24 that it was Theophrastus' Meteorology (which is
marked by similar characteristics of argument, very different
from Aristotle, as those of our De Aeternitate fragment) which was
again extensively used by Lucretius, and that this is in sharp
contrast, for example, to the Stoic Posidonius, whose Meteorology
was a recognisable development from the Meteorology of Aristotle.
This position is now reinforced by the history of the Theophrastus
fragment from Philo. One wonders whether Lucretius had direct
access to Theophrastus. Whether he did or not, the splendid new
edition of Theophrastus' fragments25 has opened new light on the
transmission of ideas in the 1st c. BC.
St. Andrews

23 H. Daiber, 'The Meteorology of Theophrastus in Syriac and Arabic Translation;' in Theophrastus: His Psychological, Doxographical and Scientific Writings,
16&-293.
24 Kidd, art. cit. (n. 22). So also, J. Mansfeld, 'A Theophrastean Excursus
on God and Nature and its Aftermath in Hellenistic Thought', Phronesis 3 7
(1992) 314-335.
25 Theophrastus of Eresus, Sources for his Life, Writings Thought & Influence, edd.
W.W. Fortenbaugh, P.M. Huby, R.W. Sharples, D. Gutas, 2 vols., Philosophia
Anti qua 54 (Lei den 1992).

RICERCHE SULIA TRASMISSIONE DELLE

DIVISION/ ARISTOTEL/CHE
TIZIANO DORANDI

MuUa Dircaeum levat aura cycnum,

tendit . . . quotiens in altos
nubium tractus ...

Chi abbia una qualche familiarita con I' opera di Jaap Mansfeld sa
bene come anche una sola sua nota possa celare quantomai
inaspettati e suggestivi spunti per ulteriori succose ricerche. Questa
e il caso del mio contributo il cui germe e gia tutto nella frase di
una nota dell'articolo: 'Physikai doxai e problemata physica: Da
Aristotele ad Aezio (ed oltre) ': 1
Cio che vale per questa trattatello [scil. lo pseudo-aristotelico nept
apet&v Kilt K!lKtrov 0 nept apetilc;] vale anche per i Placita e per le
Divisioni aristoteliche. Bisogna tentare di rendere conto delle pluralita
delle tradizioni in gioco; ha poco senso cercare di ricostruire una
singola Urquelle.

E prendendo lo spunto da questa frase che vorrei, dunque, indagare Ia tradizione delle cosi dette Divisioni aristoteliche (d'ora in poi:
DA) in vista di una nuova edizione-mia o di altri-che sostituisca
quella vecchia e ormai superata di Hermann Mutschmann.2
La trasmissione delle DA e duplice. Da un lato esse sopravvivono
in tre codici medievali, dall'altro sono conservate, in maniera
affatto indipendente, da Diogene Laerzio. Si tratta di redazioni
diverse di un unico corpus o, forse, di singoli e distinti corpora, che
Caro ]aap ti chiederai perchi anch'io ti abbia "tradito" accogliendo ['invito a
partecipare a questo volume miscellanea: "ll cui diritto all'esistenza" tu-wilamowitzianamente-riterrai "non ancora del tutto accertato". Ma come dimostrare, seppure in
maniera affatto inadeguata, un minimo della affettuosa riconoscenza e della stima
infinita che mi legano a ter Il tema del mio intervento trova origine in una delle tue
illuminanti ricerche e ha come recondita speranza quella di aprire un dialogo con te: che
non sia solo ideate! Perchi-e mi auguro di non apparire troppo presuntuoso e irriverente
nella mia scoperta allusione-almeno una volta l'anno bisogna io appaghi il desiderio,
che mi tormenta tutto l'anno, di discutere con te di storia della filosofia antica.
I Mansfeld (1993) 370 n. 80.
2 Mutschmann (1906).

146

TIZIANO DORANDI

vanno considerati e studiati, il pill delle volte, ciascuno per se. Ma

prima e opportuno descrivere i testimoni.
Comincio con Diogene Laerzio (DL). A conclusione della Vita di
Platone (3.80-109), DL trascrive un corpus di 32 divisioni che
presenta come la testimonianza di Aristotele sulla maniera in cui
Platone divideva le cose (3.80: or{lp1 OE, <j>TJOtV 'AptcrtOtEATJ<;, Kat ta
7tpayj.lata tOUtOV tOV tp07tOV. Cf. 109: &o Kat ta 7tprota or(lpEtKata tOV
'AptcrtotEATJV). In questo caso, la tradizione delle DA e inscindibilmente legata a quella dell'intero complesso delle Vite laerziane e
deve essere, pertanto, studiata in tale contesto.
Le ricerche recenti di Knoepfler 3 hanno gettato nuova luce sul
dibattutissimo problema della storia del testo laerziano e i suoi
risultati, sebbene siano fondati sull'indagine di una sola biografia e
quindi richiedano ulteriori verifiche sistematiche sulle restanti
parti dell'opera, considererei, in linea di massima, convincenti. Li
ripropongo in forma estremamente sintetica.
Tra i manoscritti vetustiores si distingue B (Neap. Burb. gr. III B 29:
s. XII), la cui superiorita deriva non solo dal fatto che e il pill antico
dei manoscritti integri, rna anche da quello che discende da un
codex perduto scritto in onciale (dunque anteriore all'anno 800 ca.).
Nonostante i suoi numerosi errori minori, la tradizione che questo
manoscritto offre-non ancora contaminata (se si escludono gli
interventi seriori di B 2) dalle correzioni cui ando soggetto l'antenato comune diP (Paris. gr. 1759 s. XIII ex.) e di F (Laur. 69.13: s. XIII
in. )-e superiore a quella dei rappresentanti della cosi detta
"vulgata" e a quella degli altri due vetustiores, cioe p (con il quale e,
tuttavia, strettamente imparentato) e F, un manoscritto bizzarro e
variamente apprezzato. I tre manoscritti integri pill antichi ebbero un
archetipo in comune (Q), dis tin to da quello dei codici degli
excerpta, che rimontano tutti a C) (Vat. gr. 96: s. XII in.), un testimone,
comunque, non molto lontano da n. Le differenze che caratterizzano i due rami della tradizione portano a concludere che
l' autographon excerptoris, modello di C), e che risale al s. XI o XII, non
era quello di BPF, cioe n, che viene quindi a occupare la posizione
di subarchetipo in rapporto al vero e proprio archetipo X, capostipite
comune di C) e n. C) discende da X attraverso l'intermediario di un
manoscritto traslitterato e forse, in pill punti, corretto (X) verso il
1100. Simili conclusioni e i risultati delle ricerche pill recenti su
3

Knoepfler (1991).

DIVISIONI ARISTOTELICHE

147

alcune famiglie della "vulgata" (il cui capostipite Knoepfler

designa con la sigla a, da non confondere con la prima classe di
manoscritti di Martini), hanno consentito allo studioso di
delineare i rapporti tra i codici laerziani in uno stemma. 4 Per
quanto concerne la localizzazione spazio-temporale dei testimoni,
Knoepfler distingue un ramo italo-greco, rappresentato da BPF e i
loro discendenti, e un ramo orientale rappresentato da ~ copiato a
Bisanzio. Si puo supporre che la tradizione orientale non rimase del
tutto sconosciuta in Occidente e che esercito una influenza indiretta sulla formazione della "vulgata", che si sarebbe ugualmente
diffusa in Oriente per l'intermediario di P, dopo essere stato
corretto, sulla base di a (recensione P 4 ); dall'Oriente, P sembra sia
stato riportato in Italia da Guarino Veronese. A Firenze, da P 4 fu
copiato il Laur. 69.35 (H), uno degli esemplari utilizzati da
Ambrogio Traversari per la sua traduzione latina.
A partire da quanto sopra esposto Knoepfler trae alcuni principi
sui quali deve fondarsi la costituzione del testo di DL e, di
conseguenza, quello delle DA nella recensione che da DL ci e stata
trasmessa.s La ricostruzione dell'archetipo (X) sara possibile solo
quando ~ e conservato; in tutti gli altri casi bisogna contentarsi di
ricostituire il subarchetipo (0), il cui testo puo, talvolta, coincidere
con quello dell'archetipo. A fondamento del testo e da porre la
lezione di B correggendone gli errori veniali con l'aiuto di PF
(quando vi sia concordanza). L'accordo di B con Po con F riporta
alla lezione del subarchetipo anche qualora ~ appoggi Ia lezione
dell'altro manoscritto. Nei casi in cui i tre codici primari divergano
fra loro e manchi la testimonianza di ~ ne sia possibile dare
meccanicamente la preferenza a B, si optera per la lezione che e
pili vicina a quella di B. Con il ricorso alla testimonianza della
"vulgata", l'editore potra altresi rendersi conto di quale manoscritto,
fra Fe P, trasmette il testo autentico (immune cioe dalla revisione
di co' nel caso di P o di co nel caso di F). Tale situazione rende,
infine, ben evidente che, se talora si riesce a correggere un errore
di n grazie al ricorso a ~, una lezione corrotta di X si potra
emendare solo con l'eventuale ricorso alia tradizione indiretta o al
iudicium. Ma con un testo come quello di Diogene Laerzio l'invito
Knoepfler (1991) 154.
Knoepfler (1991) 163-164. Condicio sine qua non per Ia validita di queste
conclusioni e l'accettazione dello stemma tracciato da Knoepfler, una cui
verifica e in corso da parte di chi scrive.
4

148

TIZIANO DORANDI

alla cautela e a non abusare dell'arte congetturale none mai superfluo.

Questo e sufficiente a proposito della recensio laerziana. Passiamo
ai tre manoscritti.
M (Marcianus gr. 257 s. XIII ex. o s. XIV in.). 6 Questo codice, unico
testimone conosciuto dell'altra redazione delle DA fino a Mutschmann, presenta sotto il titolo duxipecn~ 'Ap1cr'tO'tEAOU~ (non dunp0"1~ 'Ap10"'tO'tEAou<; Rose/Mutschmann) una raccolta di 69 divisioni
(ff. 251 r-254r). II testo delle divisioni in M e sfigurato da numerosi
errori, imprecisioni e lacune; anche se in misura minore di
quanto fin ora supposto.
Stretta parentela con M mostra A (Parisinus gr. 39 s. XIII), che
tramanda un gruppo di sole 39 divisioni sotto il titolo dta.tp0"1~
'Aptcr'tO'tEAOU~ (ff. 168v-172v).7 Rispetto aM, A conserva un numero
ridottodidivisioni (omette [22], [27], [35]-[37], [42]-(44], [46], [48][57], [59]-[69]) ;8 e infestato da diverse lacune; abbrevia il testo di
[41], dopo di cui presenta, inoltre, due frammenti allotri (f. 172' 7172 v 4) ;9 qualche volta restituisce solo i primi due dei quattro
termini di cui e costituita una divisione 'normale' ([1]-[3], [7], [9],
[38]-[39]); in altri casi manca Ia sola ricapitolazione ([4]-[6], [8],
[10]-[12]); in altri ancora il terzo termine e mescolato, seppure solo
in parte, con il primo ([11], [58]).
Infine L (Leid. Voss. gr. Q 11 s. XIV, prima meta). Esso trasmette
(ff. 92r-96r) un corpus di 61 divisioni rispetto alle 69 di M sotto il
titolo anonimo Tiepl. 01<X1pcreco<; scritte da due mani distinte, rna
databili entrambe alia prima meta del XIV sec. 10 L none ne fratello
ne discendente ne imparentato direttamente con M come dimostra
la diversita di errori e lacune. La differenza pili evidente e nel
numero minore di divisioni conservate: mancano in L le otto
divisioni [35], [63] e [64]-[69] Ia cui assenza-se si eccettua [35],
accidentalmente omessa-puo essere voluta: [63] costituisce, infatti,
Cf. Rossitto (1984) 13 s. n. 3.
7 Boudreaux (1909), che per primo ne scopri l'esistenza. Indico il manoscritto con Ia sigla A invece che con P (come finora fatto) per evitare confusione con il Paris. gr. 1759 di Diogene Laerzio, ugualmente siglato P.
8 La numerazione, fra parentesi quadre, e quella di M, che contiene Ia
raccolta piu completa.
9 E non tre come scrive Boudreaux (1909) 222 n. 5 seguito da Moraux (1977)
126. La seconda e Ia terza aggiunta costituiscono, infatti, un unico testo. Cf.
infra, 150-152.
10 Moraux (1977), che per primo ne ha scoperto l'esistenza.
6

DIVISIONI ARISTOTELICHE

149

in un certo modo, un doppione di [51], mentre [64]-[69] formano

un gruppo compatto i cui membri sono saldamente legati l'uno
all'altro con caratteristiche che non sono le stesse che ispirano le
restanti divisioni. AI posto di quel nucleo, L riporta una appendice,
nella quale Moraux scorge tracce di influenza neoplatonica, e che e
costituita da una Oux.{peat<; 'lfUXll<; JlE'ta opt<rJlOU e da un opo<; vo6<;. 11
Alcune lezioni di L possono servire talora a correggere quelle
corrotte di M o A.
II testimone pili sfruttato delle DA, anche perche conosciuto da
pili tempo, e M. Esso e stato posto a fondamento delle edizioni di
quell'opera, come unico testimone della tradizione parallela a DL,
fino a quella recente curata da Gigon,12 che continua a ignorare sia
Asia L.
E ormai accertato che M A L tramandano una seconda
'redazione' delle DA. I tre manoscritti devono, comunque, essere
indagati singolarmente al fine di metterne bene in risalto le
caratteristiche portanti e meglio definirne gli eventuali legami e
rapporti.
Comincio con M A. Per Boudreaux, A, per certi aspetti, inferiore
a M, e a questo manoscritto imparentato, come prova Ia presenza di
errori congiuntivi e il fatto che le divisioni sono riportate nella
stessa sequenza; esso risale, tuttavia, a un modello comune in
maniera probabilmente indipendente. Lo dimostrerebbero i casi in
cui A fornisce Ia lezione originale, sfigurata in M; quelli in cui
aiuta a eliminare dal testo di M alcune intrusioni o a ripararne
omissioni, tra cui anche il testo della ricapitolazione di cinque
divisioni ([26], [28], [30]-[32]). II caso, infine, di [10] dove A presenta un testo superiore a quello raffazzonato di M. Per Moraux, 13 Ia
parentela fra i due codici e evidente anche se assai lontana. A
tramanda, infatti, una scelta di divisioni spesso in forma abbreviata
soprattutto nella terza parte di quella che e Ia struttura canonica
completa di una divisione: 1. Indicazione dei membri della
divisione (OtapEt'tat ... Ei<; -rp{a, -raaapa); 2. Enumerazione di questi
membri; 3. Spiegazioni relative a ogni membro, spesso con una
serie di esempi introdotti da otov; 4. Ricapitolazione, che riproduce,
in linea di massima, I' enumerazione di 2. In A Ia ricapitolazione
11 So no riprodotte da Moraux ( 1977) 126. Esse dovranno figurare nella
nuova edizione delle DA.
12 Gigon (1987) F 83.
13 Moraux (1977) 101, 103-104.

150

TIZIANO DORANDI

compare solo per le divisioni [26], [28], [30]-[32], casi in cui,

curiosamente, manca negli altri testimoni. Poiche le ricapitolazioni svolgono, pili o meno, Ia stessa funzione dell'enumerazione,
c'e da chiedersi se queste cinque aggiunte risalgano al modello di
A oppure siano state introdotte da un anonimo redattore. 14
La dipendenza di A e M da un capostipite comune appare sicura.
Resta, semmai, da vedere fino a che punto sia lecito correggere il
testo di M a partire dalle lezioni di A e viceversa; se cioe le
differenze che distinguono i due manoscritti siano da imputare al
fatto che A ha trasmesso meglio il suo modello oppure se derivino
da una recensione (?) diversa con cui A o il suo antenato (o antenati?) hanno contaminato il testo comune anche aM, o meglio se
non siano interventi personali del suo copista (che in questo caso
avrebbe svolto anche il ruolo probabile di redattore). Si tenga conto
che Ia datazione dei due manoscritti impedisce di postulare una
derivazione di A da M e che una derivazione di M da A e altresi
impossibile.
Prendo in considerazione le caratteristiche di A rilevate da
Boudreaux. La presenza di un numero inferiore di divisioni,
I' omissione di uno dei membri che costituiscono Ia divisione tipo,
il rimaneggiamento della struttura o Ia tendenza a abbreviare puo
bene essere imputata sia al modello di A sia al copista (redattore?)
stesso. Se alcune omissioni e errori comuni a A e a M si spiegano
come verificatisi in maniera accidentale, diversamente da Boudreaux giudicherei, invece, i luoghi in cui A conserva Ia lezione
corretta rispetto a M. Non e affatto escluso che le lezioni di A siano
frutto di congettura (lo confermerebbe il fatto che, nella quasi
totaliti delle occorrenze gli studiosi successivi avevano proposto le
medesime correzioni). Lo stesso discorso puo val ere, talvolta, per
quei casi in cui A consente di eliminare intrusioni o riparare
omissioni in M. Per l'aggiunta delle cinque ricapitolazioni, ha
ragione Moraux a lasciare Ia questione in sospeso. Quanto a [10],
non parlerei, con Boudreaux, della versione di A come della
genuina rispetto a quella rimaneggiata di M, rna piuttosto di un
caso di tradizione ritoccata (rispetto a M e a L), non esente essa
stessa da eventuali corruttele (NdA.ou).l5
Qualche chiarimento relativamente ai rapporti fra M e A ritengo
possa venire dallo studio delle due inserzioni allotrie dopo [ 41] . La
14

Moraux (1977) 104 conclude con un non liquet.

15 Per Ia questione, vd. Rossitto (1984) 157-158.

DIVISIONI ARISTOTELICHE

151

seconda (1tapax:aA.ro 7tprotov 7t<xvnov ... 8Ecp ta tEtOAJ.lTJJ.ltVa), come

sono riuscito a appurare, e una citazione letterale dall'omelia di
Severiano di Gabala, In incarnationem Domini. 16 Piu difficile l'identificazione della prima-che trascrivo:
eyro eiJl.l eyro eiJl.l mi em~ av 1cmampaaa-re eyro eiJl.l (LXX Is. 46,
4) Set yap av-rov f3aC1lAeVElV axpl~ o{j av lJfj nav-ra~ 'fOV~ exfJpov~
av-rov vno n6t5a~ av-rov (NT 1 Cor. 15.25). to em~ ou xpovou
~1lAO:mKov, &A.A.' i~icoJ.UX tf\c; EMac; ypa<pf\c;. G'll(J.LEicoom) ta1ha
~ux to mi OVIC ey{vmcrKa avn)v em~ o{j E'fEICE 'fOV viov av-rij~ 'fOV
npro-ro-roKov (NTMatth. 1. 25) npmo-roKo~ EK veKpiiJv (NT Col. 1.
18) . d7tf. K<Xl tac; OlCtcO avacrtacrEtc;.l7

L'd1tt mi porterebbe a pensare che siamo di fronte a un appunto

anonimo per un sermone. Si tratta, dunque, di due frammenti
derivati da un contesto cristiano la cui estraneita alle DA e indubitabile, rna la cui sopravvivenza ben si spiega in questa sezione di A,
il solo manoscritto in cui le DA sono precedute e seguite da
materiale di contenuto prevalentemente teologico. 18 Resta da
chiedersi a chi debba essere imputata l'inserzione di questi brani, a
quel punto del manoscritto, a interrompere la sequenza delle DA: se
al copista (redattore?) di A o gia al suo modello. Se si accetta la
seconda ipotesi e si mantiene ferma una derivazione di A e M da
un comune antenato, sebbene lontano, la mancanza di questi
frammenti in M si potrebbe spiegare sia ammettendo che M abbia
volontariamente omesso quelle parti estranee oppure presupponendo un testimone intermedio fra A e il suo modello. Ma se,
come ho suggerito, il primo frammento e quanto resta di un
anonimo appunto per una predica, ben piu probabile mi appare
l'eventualita che i due passi siano stati aggiunti direttamente dal
copista di A. Quello che mi sembra derivi, comunque, in maniera
evidente dalla identificazione di queste due citazioni e dalla stessa
struttura di A e la prova concreta di una sicura circolazione delle
DA all'interno di milieux cristiani, d'altra parte, gia presupposta. 19
l6 PG 59. 697, 69-698, 11 (riferita parzialmente e con varianti anche da
Fozio, Bibl. cod. 277. 520a 27-42 [8. 142, 27-143, 42 Henry], che Ia attribuisce a
Giovanni Crisostomo). Cf. Clavis Patrum Graecorum 2 n 4240.
17 Posso richiamare a confronto: Didym., Jr. Ps. F 1054, 5 ss. Muhlenberg;
Thdt., interpr. in Pauli ep., PG 82.356, 25-28; Didym., Trin. 3, PG 39.832, 39-833,
5. Perle OKtro avacrtacru~: Georg. Mon., Chron., 1. 315, 1-316, 2 De Boor.
18 Sulla struttura composita di A, cf. I' Inventaire sommaire di Omont, 1. 7. Ma
Ia mescolanza eterogenea dei testi in esso contenuti meriterebbe di essere
ap~rofondita anche in rapporto con Ia presenza delle DA.
9 Cf. Ia bibliografia citata alia n. 44.

152

TIZIANO DORANDI

Vengo ora ai rapporti, ben pili complessi, di L con A eM (o

AM). La scoperta di L ha gettato luce sul problema della tradizione
delle DA e non soltanto per il contributo che da questa terzo codice
puo venire per la costituzione del testa, rna anche per meglio spiegarne la diffusione e i canali della trasmissione. L e un manoscritto indipendente da A e M: la differenza pili notevole che distingue L daM consiste nel fatto che il prima contiene un numero
inferiore di divisioni. Per Moraux2 l'assenza, in particolare, delle
ultime sei ( [64]-[69]) dipenderebbe non dal fatto che il modello di
L era mutilo, quanta piuttosto dal fatto che L risale a un ramo della
tradizione, che non conosceva l'esistenza di questa appendice; essa
sarebbe stata nota solo al tronco da cui derivo M. Come A, L non e
esente da errori, rna consente anche, in molti casi, di correggere
luoghi corrotti di M e di colmarne lacune.2 1 Pur con la dovuta
cautela dettata dal fatto che alcune divergenze non possono essere
imputate alla negligenza del copista, rna tradiscono l'intervento di
un redattore, si deve ammettere l'importanza del contributo di L per
sanare talvolta il testa difficile e malamente alterato delle DA.
Le differenze fra i tre manoscritti risultano dalla collazione
pubblicata da Moraux. 22 I criteri che lo studioso enuncia a proposito
dell'uso che si puo fare dei tre testimoni per la ricomposizione
dell' Urtext delle DA non possono, tuttavia, assumere il valore di
principi validi in assoluto:23
Comme L represente une autre branche de la tradition, il a souvent
conserve le texte original la oli M presente une lacune. Et quand
c'est L qui a, par distraction, laisse tomber un morceau de texte entre
deux mots identiques, on trouve d'ordinaire le texte complet dans M
(eventuellement dans P [=A]). La comparison des deux branches
de la tradition permet done, dans bien de cas, de reconstruire le
texte de !'archetype, que ni MP [=A] ni L n'ont conserve dans sa
purete premiere.

Sul fondamento di queste premesse, Moraux propane un nuovo

testa parziale per [36], [53], [56] o completo per [22], [38], [40], [45],
[51], [54] combinando insieme lezioni di M e di L. Non posso
Moraux (1977) 102.
Una delle caratteristiche di L e Ia presenza di titoli per le singole
divisioni escluse [26]-[27]. Nel caso di [13], [16], [19] e [41], i titoli sono
riportati a margine. A ha il titolo solo di [ 1], M di [2]. Per quanto riguarda
DL ci sono sporadiche tracce di titoli in B e in P (sistematicamente cancellati
dal quarto correttore p4).
22 Moraux (1977) 105-114.
23 Moraux (1977) 114-115. Le esplicazioni in corsivo sono mie.
20
2I

DIVISIONI ARISTOTELICHE

153

discutere nei dettagli tutti i casi i cui Moraux ha operato sulle DA.
Noto solo che talvolta i suoi tentativi hanno dato origine a una
ricostruzione dubbia sia dal punto di vista testuale sia soprattutto dal
punto di vista filosofico, come ha chiarito la Rossitto nelle pagine
del suo commento.
Piu cautamente e con maggiore verisimiglianza, argomenta la
Rossitto:2 4
L compie interventi molto significativi sui testo, dando l'impressione di volerlo rendere piu coerente, o addirittura di correggerlo:
ad esempio esso ripristina spesso nella spiegazione l'ordine in cui i
termini appaiono succedersi negli elenchi iniziale o finale, o risulta "confermare" congetture di studiosi moderni, il che puo essere
considerato piuttosto una prova di congettura da parte del copista,
che non tradizione genuina o non inquinata del manoscritto.
Riserve avrei, semmai, sulle sue ulteriori conclusioni-ma di
questo in seguito:
II grado di "perfezione" che (L) sembra presentare parla a sfavore di
una sua maggiore fedelta a! testo primitivo rispetto aM, il quale, in
virtu delle sue non infrequenti, rna nemmeno decisive, incongruenze, e della sua rozzezza, rimane in fin dei conti Ia versione
delle Divisioni, tuttora maggiormente affidabile.

Fin qui lo status quaestionis sui manoscritti.

L'edizione di Mutschmann, le cui debolezze non dipendono solo
dalle cattive conoscenze della tradizione laerziana,25 e stata messa
in crisi dalla scoperta delle raccolte di A e di L. Ne tantomeno
soddisfa l'attesa riedizione delle DA approntata da Gigon, insieme
con i restanti frammenti di Aristotele.26 Lo studioso riproduce,
infatti, le due versioni delle DA di M e DL, la prima a partire dalle
edizioni di Rose e di Mutschmann e da una rinnovata autopsia di
M, la seconda tenendo presente il testo stabilito da Hicks, dagli
Editori di Basilea e da Long e suggerendo, in entrambi i casi
sporadiche correzioni.27 Purtroppo, Gigon non ha utilizzato ne A ne
L ne ha apportato sostanziali miglioramenti.
L'edizione delle DA, qualunque sia illoro valore intrinseco, resta,
pertanto, ancora un desideratum nella storia della filologia classica e
24

Rossitto (1984) 25 n. 38.

25 Cf. Biedl (1955) 28. Severa, rna giusta, Ia recensione di Praechter

(1910).
26 Cf. Rossitto (1989) 213-215.
27 Gigon (1987) 318.

154

TIZIANO DORANDI

della filosofia antica. Ma come affrontare il problema? Su quali

principi fondare il nuovo testo critico? Non e possibile rispondere a
queste domande se non si imposta prima correttamente (non dico,
si risolve) Ia questione an cora pili spinosa e ostica della origine di
questa raccolta, del genere letterario e della sua eventuale paternita.
Sono questioni interrelate fra loro e gia ampiamente dibattutte, rna
che solo ora si avviano a trovare una risposta soddisfacente.
Le liste antiche dell'opere di Aristotele attestano l'esistenza di
scritti il cui titolo riporta nell'ambito delle Divisioni: LltatpEcrEt<;
croqncr'tucat 8 (DL 29; H 31 *;A-) Lltatpecrtt<; t~ (DL 42; H 41; A 58*
?) LltatptnKrov a (DL 43; H 42; A-) LltatpE'ttKOV a (DL 63; H -;A
-). E questi dati parrebbero confermati daMe A e da DL: Me A
nel titolo stesso (LltatpEcrtc; 'Aptcr'tO'tEAouc;), DL nelle parole che
premette alla raccolta (3.80: BtflpEt Be, q>Ttcrtv 'Aptcr'to'tEATt<;, Kat 'ta
npayJ.La'ta 'tOU'tov 'tOV 'tponov) e nella conclusione (109: 6>8 Kat 'ta
npro'ta ~hflpn Ka'ta 'tOV 'Aptcr'to'tEATtv). Ma Ia paternita aristotelica
delle Divisioni e Ia loro identificazione con una delle opere citate,
seppure in una versione in larga misura riscritta e rimaneggiata, e
stata spesso accolta con scetticismo.2B
Per primo, Rose pubblico nell' Aristoteles Pseudepigraphus29 il testo
delle DA nella redazione di M insieme con i frammenti degli
scritti di Aristotele che non erano rientrati nel corpus di Bekker,
giudicandoli tutti spuri. Nelle due edizioni successive della sua
opera, Rose, pur mantenendo ferma Ia propria convinzione dell'inautenticita dei frammenti, non riprodusse pili il testo delle DA nella
rubrica intitolata LltatpEcrnc; e si limito a un semplice rinvio a DL (F
1112 = 1143) e aM (F 1122 = 1153).30 Le DA rivendico a Aristotele
Heitz. 31 Egli distinse, peraltro, due raccolte di Divisioni: Ia primauna esposizione di divisioni platoniche-rappresentata dal testo di
M e DL; Ia seconda dalle divisioni citate nei cataloghi delle opere
di Aristotele e nella testimonianza di Alessandro di Afrodisia. M e
DL conserverebbero una raccolta di divisioni di origine peripatetica, interpolata da autori cristiani, dove, accanto a molte cose
indegne di Platone e Aristotele si intravedono ancora dementi pili

2B Dettagliata e documentata analisi in Rossitto (1984) 13-26.

29 Rose (1863) 677-695.
30 Rose2 F 110-112 e Rose3 F 113-115. F 112 =113 e un passo di Alex. Aphr.,
in Top., 242, 1-9 Wallies (CAG 2.2), che Gigon (1987) F 622 assegna piuttosto
alia T&v aya8&v otaipecrt~.
31 Heitz (1868) 91-112, 118-119.

DIVISION/ ARISTOTELICHE

155

antichi, risalenti forse allo stesso Aristotele. Se Suckow32 considero

le DA come opera genuina di Aristotele, in generale, prevalse
l'opinione che si trattasse piuttosto di una raccolta di dottrine
plato niche e aristoteliche (Susemihl), della rielaborazione di
un'opera composta nell'ambito della Academia antica (Zeller) o un
manuale scolastico in essa usato (Christ). Un giudizio pili
favorevole sullo scritto dette invece Hambruch, 33 che individuo in
alcune DA (M [37], [64]-[69]) l'esposizione di regole logiche comuni a Platone e alla sua scuola, che rimontavano alla Academia
antica, rna che erano state utilizzate anche da Aristotele nei Topici.
Lo studioso mise, inoltre, in evidenza alcune analogie fra le DA e le
esposizioni delle lezioni orali di Platone, quale quella Sul bene, e il
Tiept 'taya.Oou di Aristotele, nonche con i frammenti dell'altro suo
scritto perdu to Tiept vav'ttrov. Sulla falsariga dei risultati di
Hambruch si pose Mutschmann nella Praefatio alia sua edizione:
per molte DA e possibile postulare una derivazione diretta da
Platone; altre si presentano come Ia rielaborazione di dottrine dei
dialoghi platonici risalente al suo insegnamento orale e trascritte
dai suoi discepoli, tra i quali lo stesso Aristotele, che vi aggiunse
qualcosa di suo e ne adatto i contenuti al proprio pensiero. E
difficile stabilire a quando risalgono gli interventi di Aristotele: Ia
presenza giustapposta di elementi della fase platonizzante e di altri
di una fase pili matura farebbero pensare a una rielaborazione
diacronica. Alcune DA sono frutto di manipolazioni seriori dovute
all'intervento di editori peripatetici e talvolta anche cristiani. Nel
complesso, ci troveremmo di fronte a un compendia di filosofia a
uso delle scuole Ia cui relativa facilita e il cui livello elementare
sarebbero dovuti proprio a questa destinazione. Se le DA analizzate
da Hambruch erano finalizzate all'insegnamento della dialettica,
le altre erano destinate, invece, aile esercitazioni retoriche. Le
analogie fra alcune DA e i resti del Tiept 'taya.Oou aristotelico furono
approfonditi da Wilpert, 34 che intravide nelle DA una esposizione
academica delle dottrine non scritte di Platone. Lo studioso segnalo
anche interessanti coincidenze e differenze con il libro !1 della
Metafisica e il Tiepl. va.V'ttrov. Le DA, in ultima analisi, so no-a suo
avviso-una rielaborazione delle dourine orali di Platone nata nella
Academia per le cure di un discepolo sui genere del Tiept 'taya.Oou
32 Suckow (1855) 96ss.
33 Hambruch (1904).
34 Wilpert (1941) 236-241, (1949) 108, 149, 183.

156

TIZIANO DORANDI

di Aristotele. Favorevole alia paternita aristotelica delle DA sembrerebbe fosse Ross, a meno a quanto si puo ricavare dalla sua
edizione degli Aristotelis Fragmenta selecta, dove le due redazioni
dello scritto (M e DL) vengono menzionate nella sezione sulle
opere di logica sotto il titolo ~tatptaEt~.35 I rapporti con l'insegnamento orale di Platone ha ulteriormente sviluppati Kramer36 per M
[67]-[68] e DL [32], divisioni nelle quali lo studioso individua una
probabile relazione delle dottrine non scritte di Platone elaborata da
Aristotele.37 E accanto a queste, le DA mostrano corrispondenze
anche con il pensiero di altri Academici antichi, talvolta rivisitato
attraverso il ripensamento di Aristotele. Una svalutazione globale
delle DA ha riproposto Moraux:3 8 si tratterebbe di un manuale
mediocre, senza originalita, ricolmo di piattezze, banalita e inezie,
anche se fondato su un sostrato che alia lontana risale a Platone e a
Aristotele. Per Gigon,39 le DA sono un excerptum, con fini protrettici,
di materiale derivato da distinti dialoghi di Aristotele; il loro
carattere prevalentemente etico mostra stretti parallelismi non solo
con i trattati aristotelici, rna anche con i dialoghi platonici. Esse
sono, dunque, da restituire a Aristotele e da ricondurre nell'ambito
della Academia Antica dove egli soggiorno a lungo, restandone
influenzato e a sua volta influenzandola. Infine, Ia Rossitto 40 ha
cercato di dimostrare che le DA hanno un grado di "aristotelicita"
non inferiore a quello delle altre opere scritte da Aristotele durante
il periodo academico e ne ha spiegato Ia peculiare natura con Ia
loro destinazione scolastica. Innanzitutto le DA mostrano numerose
coincidenze di contenuto con quanto Aristotele afferma nelle sue
opere, soprattutto quelle che possono essere fatte risalire al periodo
giovanile sia quelle frammentarie (Protreptico) sia quelle conservate (i Topici, il libro ~ della Metafisica e le Categorie); rna non
mancano neppure significativi paralleli con le due Etiche e Ia
Retorica. In secondo luogo, le DA presentano analoghe coincidenze
sia con le dourine degli Academici antichi (Speusippo, Senocrate,
Ermodoro) sia soprattutto con alcuni Dialoghi di Platone e con le sue
dourine orali; e, nondimeno, influssi dalla Rhetorica ad Alexandrum
35 Ross (1955) 101-102 = F 12 (DL) e 3 (M).
36 Kramer (1983) 141-142.
37 Gaiser-seguito da Kramer, in studi successivi-allarga il numero a M

[23], [67]-[68] e DL [27], [32].

38 Moraux (1977) 1-127.
39 Gigon (1987) 318.
40 Rossitto (1984) 27-33.

DIVISIONI ARISTOTELICHE

157

di Anassimene di Lampsaco, da Isocrate, dai Magna Moralia e dallo

pseudo-aristotelico De virtutibus et vitiis. Tutto cio induce a pensare
che nelle DA siano confluite: "Classificazioni di carattere dialettico
e retorico pili diffuse all'interno, rna non meno all'esterno,
dell'Accademia platonica del IV secolo", un ambiente con il quale
doveva avere buona familiaritl colui che ha costituito il corpus delle
DA, che si presenta come una raccolta di gruppi di classificazioni
non totalmente eterogenei fra loro, perche legati da divisioni "di
passaggio" e soprattutto perche riconducibili a un: "Denominatore
comune, rappresentato dal loro carattere retorico e dialettico". 41
Non deve meravigliare che I' organizzatore di tutto questo materiale
in un unico scritto sia stato lo stesso Aristotele: egli vi avrebbe
aggiunto proprie dottrine e l'avrebbe uniformato dal punto di vista
terminologico negli anni del suo soggiorno nell'Academia, in
preparazione dei suoi corsi di retorica e di dialettica. II metodo di
lavoro adottato e i fini sono simili a quelli che Aristotele applica nel
libro 11 della Metafisica. La tesi che le DA costituiscano, invece, un
anonimo "manuale di filosofia elementare" e stata suffragata da
Mansfeld42 portando a confronto-accanto ai manuali appartenenti
alia famiglia dei Placita-il caso del IlEpt apE'tWV Kat KO.K1WV 0 IlEpt
apE'tll<; dello ps.-Aristotele sia nella versione abbreviata sia in quella
seriore riveduta con integrazioni stoiche, nota con il titolo Ilept
1ta.Srov e attribuita a Andronico di Rodi.43 L'ipotesi none nuova, rna
i confronti che Mansfeld adduce sono quanto mai significativi e
un'interpretazione delle DA come "manuale a uso scolastico"
spiega, meglio di ogni altra, aspetti della trasmissione di quel testo
e e da questa confortata.
Mail dibattito sulla paternitl aristotelica o meno delle DA svolge,
in un certo qual modo, un ruolo secondario. II corpus (o i corpora?),
messo insieme da anonimi redattori, contiene, effettivamente,
materiale originale derivato da tradizione scritta o orale risalente a
Platone, Aristotele e a altri membri interni o esterni all'Academia.
Quello che mi sembra evidente e, tuttavia, il fatto che, nelle redazioni in cui sono oggi leggibili, le DA difficilmente possono essere
considerate al pari di altri frammenti di opere perdute di Aristotele.
Nemmeno se si suppone una pessima trasmissione, riesco a intra-

41 Rossitto (1984) 30-31.

42 Mansfeld (1993) 317, 326-327.
43 Mansfeld (1993) 370 n. 80.

158

TIZIANO DORANDI

vedere come loro unica fonte uno scritto originario di Aristotele in

cui egli avrebbe organizzato materiale di diversa provenienza (rna
questa vale anche per un qualunque altro au tore specifico). Pure
ammettendo, infatti, che all'origine ci fosse una o piu opere di
Aristotele, gli infiniti passaggi e rimaneggiamenti cui i testi sono
stati sottoposti le hanna talmente sfigurate, che eben faticoso intravederne Ia struttura stessa-non parlo della forma ne tantomeno
della lingua.
Delle DA circolarono gia nella tarda antichita redazioni distinte,
inquinate e interpolate da "redattori" e da "maestri di scuola", 44
arricchite di esempi o semplificate nella loro struttura e nel loro
dettato: Ia raccolta giunta a DL e un accenno in uno scolio allo
Hexaemeron di Basilio45-scoperto da Pasquali, e correttamente
utilizzato da Mansfeld46_ne provano Ia diffusione nel mondo
tardo-antico.
All'origine di queste diverse redazioni bisogna ammettere una
pluralita di modelli non attribuibili a un determinato autore (p. es.:
Aristotele). Si puo pensare a piu scritti (p. es.: i trattati aristotelici
ricordati nei cataloghi delle sue opere, Platone, Anassimene e
Isocrate) dai quali anonimi maestri o redattori hanna ricavato-piu
o meno indipendentemente-il loro materiale; rna anche-e mi
sembra ipotesi piu consona con le realta di cui disponiamo-a una
o piu raccolte anonime, patrimonio di scuola che, come il citato
Tiepl. apE'tOOV lC<XllC<X.Ktrov pseudo-aristotelico, avevano Ia funzione di
manuali: "Per studenti di filosofia, rna anche per studenti di
retorica". 47
A partire da queste premesse, devono essere impostati i criteri
ecdotici per Ia nuova edizione. L'indipendenza di DL nei confronti
di M e stata riconosciuta fino da Rose e ammessa dagli editori e
studiosi successivi, che pubblicarono DL separatamente (Rose,
Heitz, Gigon) o su una colonna parallela al testa di M (Mutsch44 Rose (1863) 678, Mutschmann (1906) XXXIII e Moraux (1977) 121 n. 13
vi hanno scorto interpolazioni cristiane. Ne ridimensiona Ia presenza Gigon
(1987). Cf. Rossitto (1984) 133 e supra 150-152.
45 Schol. in Bas. Hexaem. XIX Pasquali (1910/1986) 200, 4-5 (= 545, 4-5): ev
tate; 'AptatotEA.ouc; Eic; IIA.Ihrova ava<pEpoi!EVIl\c; lhatpEO'EO'\V KEltll\ llUtT\ t&v tEXVOOV fJ
~ta<popci. II riferimento e a [8].
46 Mansfeld (1993) 370-371 nn. 81 e 83.
47 Mansfeld (1993) 370 n. 81. Una situazione simile e stata postulata da
Souilhe (1930) 155-158 a proposito degli opm pseudo-platonici.

159

DIVISION/ ARISTOTELICHE

mann). 4B DL rappresenta una tradizione a se, la cui origine puo

essere ravvisata in una raccolta precedente la composizione delle
Vite laerziane (prima meta del s. III d.C.) e posteriore a Aristotele.
Restano gli altri tre testimoni. Mutschmann, che conosceva solo
M, vi intravide una seconda recensione, che pubblico a fianco di
quella di DL, da lui considerata primaria. La situazione e divenuta
pili complessa con la scoperta di A, prima, e di L, dopo. Gia
Boudreaux riconobbe in A una tradizione vicina a quella di M e
suggeri di utilizzarne il contributo per correggere quelli che riteneva errori o omissioni meccaniche di M, rna che sono piuttosto, in
molti casi, omissioni volontarie e sintomatiche di influenze di
diverse scelte "redazionali" o "scolastiche". Questo significa che,
sebbene i due manoscritti siano tra loro imparentati e, in un certo
qual modo, costituiscano un'unica famiglia, non sempre le loro
lezioni possono essere mescolate a formare un Urtext. Similmente
per Ia testimonianza di L. Le lezioni di L risultano talvolta utili per
riparare guasti di M (e di A) qualora siano corruttele di origine
meccanica; anche in questo caso, si deve, comunque, evitare di
incorrere nella tentazione di ricostruire, a ogni costo, un Urtext. Per
le DA, ci troviamo, dunque, di fronte almeno a due redazioni
diverse e distinte di un testo originale Ia cui forma e struttura spesso
non puo essere stabilita con certezza. Non si deve dare Ia
preferenza, in assoluto, all'una o all'altra delle tradizioni, rna il
contributo di ciascuna deve essere considerato di per se. Questo
comporta, come prima conseguenza pratica, l'invito al futurus editor
di pubblicare il testo delle due redazioni delle DA su pagine
parallele: su quella di sinistra Ia redazione di DL, su quella di destra
il testo MAL ( oppure M A, M L, M o L sol tanto). II tutto puo essere
pili o meno cosi rappresentato:
Pagina sinistra
DL

48

tario.

Pagina destra
M

Le differenze analizza Ia Rossitto (1984) nelle pagine del suo commen-

160

TIZIANO DORANDI

L' editore sara talvolta costretto a riprodurre su colonne parallele

le distinte 'versioni' dei singoli manoscritti perche troppo divergenti; in altri sara sufficiente stampare il testo di M sulla prima
colonna e, a fianco, solo le 'varianti' di A o L; in altri ancora potnl
presentare il ricostruito Urtext limitandosi a segnalare eventuali
minime diffenze in apparato. Il confronto fra i codici consentira
altresi di correggere vicendevolmente corruttele meccaniche, di
migliorare incertezze testuali o risolvere dubbi.
Per rendere piu chiaro il discorso, cerco di definire graficamente la mia suggestione. Una cosa vorrei risultasse evidente e
cioe che questo schema non e ne uno stemma codicum ne ha l'ambizione di rappresentare una ricostruzione, dettagliata e sicura in tutti
i particolari, della nascita e delle tappe progressive di formazione
delle due redazioni delle DA giunte fino a noi. Esso trovera conferme o smentite nel lavoro di cesello che costituisce il momenta
della recensio di qualsiasi testo antico:

(s. IV a.C. ?)

(?)

a
(s. III d.C.)

DL

(?)
(s. XIII)
(s. XIII/XIV)

(s. XIV)

M
L

DIVISION/ ARISTOTELICHE

161

Con XJ, x2 .... , x 0 indico Ia serie molteplice di fonti (anonime o

non) che costituirono il sostrato materiale dal quale derivo Ia
successiva tradizione "scolastica" ( m), da intendere come un
ipotetico bacino collettore, Ia cui ricostruzione risultera spesso
difficile o disperata a causa della fluidita di quell'opera, rna Ia cui
esistenza non puo essere negata. Da m attinsero, in tempi diversi e
con modalita diverse, i singoli redattori (a. ~ y), in maniera diretta
o attraverso modelli intermedi, non sempre ne sicuramente
identificabili o postulabili.
A mo' di conclusione propongo, infine, quattro esempi di come
intenderei dovessero essere pubblicate le DA. Si tratta di una
divisione ([10]) tramandata da tutti e quattro i testimoni, di due da
tre (DL M L [36] eM A L [30]) e di una da due (M L [22]). None
necessario proporre divisioni conservate da DL e M; ne esistono
casi di DA note daM A.49

49 Alia fine di ogni esempio segnalo Ia bibliografia e mi soffermo su una

scelta di problemi testuali.

162

TIZIANO DORANDI

[10] DLMAL
DL
Titulus deest

ouup1-tm of. i] tuyivtta E~ tt&, 'tCttapa. v J.l.Ev, W.v cktv oi. ltpOyOVOt K:aAot Kaya&i Kat
OtK:Cl101, touc; EK: toutrov 'YEYEVVl]J.l.EVOUc; EilyEvEic; <j)CliHV dvat &A.A.o o, eav c1otv oi.
ltp6yovm Otouvau'tEUK:Ottc; Kai iipxovttc; 'YE'YEVTJJ.l.EvOt, touc; eK 'tOUtrov tuytvtic; <pautv
Eivat. &AM> o, eav cktv oi. ltpoyovot OVOJ.l.ClO''tOt, olov altO O"tpCltl]ytac;, altO O'tE<pClVtt&v
ayrovrov. K:Cll yap touc; EK: tOUtrov 'YYEVV1]J.l.EVOUc; EUyEvEic; ltpouayopEUOJ.l.EV. &AM> tiooc;,
eav autoc; nc; Uytvvaoac; tf\v 'lfUVlV K:Clt J.l.YClAO'IfUXOc;, K:Cll toutov Euytvf\ <pClO'tV ivat
K:Cll tf\c; 'YE ruytvdac; ClUtl] K:pattO'tl]. tf\c; iipa EUyEvdac; to ~v altO ltpoyovrov Elt1E1K:IDV. to
OE OUVClO"tiDV. to OE EVOO~rov. to OE altO tf\c; ClUtOU K:ClAoK:aya9\ac;
Bibl Mutschmann (1906) 12-13; Boudreaux (1909) 223-224; Moraux (1977) 106;
Rossitto (1984) 151-158

[36] DLML
DL
Titulus deest

to aya9ov de; tEttapa YEV1] 01CllpEitCl1' J>v Ev J.l.EV AEyOJ.l.EV EtVCll tOV tftV apEtftV E~Vt(l
io{~ aya96v . &AM> OE autilv tilv apttilv K:Cll tftV 01K:Cl10uUV1]V AEyOJ.l.EV aya9ov ElVCll.
tpttOV OE, OlOV O'lttCl K:Cll 'YlJJ.l.VllO'lCl tft ltpOO<popCl K:Cll <plipJ.l.ClK:Cl' tCtClptOV OE <j)ClJ.l.EV lvat
&ya96v, otov auA.l]nK:i]v K:al. ultoKpmK:i]v mi ta totauta. &ya9ou &pa tE-ttapa tt&, eu't{
tO J.l.EV to tftV apttilv EXElV, ettpov OE autil i] apttil. tpttOV OE O'lttCl K:Cll 'YlJJ.l.VllO'lCl ta
ID<pEA.tJ.l.Cl' tEtClptOV OE ClUAl]ttK:i]v K:Cll UnOK:ptnK:T]v K:Cll lt011]ttK:i]v aya9ov AEyoJ.l.EV dvm
Bibl Mutschmann (1906) 30-31; Moraux (1977) 110, 121; Rossitto (1984) 244-247

163

DIVISION/ ARISTOTELICHE

Titulus deest

Titulus deest

Olaipemc; e\ryeveiac;

ounptital T, euyev Eta de;

tpia ecrn yap autilc; Ev ~LEV
a1to evOO~rov Kat ovofacrt&v
yoverov yqoveval, owv a1to
~(lO"lAE(l)V Kilt ap:xovtroV
[yeyoveval] ft aAATJV nva
oo~av e~ovtrov : to OE ,a1to
0"1t0ll0!Xl(l)V Kilt OlK!Xl(l)V,
f'
'
\ ....
...
owv
a1to
.:.evo<provtoc;
Kill\
NEiA.ou Kat t&v towutrov
Kilt 11YIXA07t p7tf, c; Kilt
oiKawc; Kat ta &A.A.a ta tmauta. tile; euyeveiac; (&pa)
!lEv ecrn tO U7t0 evOO~rov Kilt
ovo~tacrt&v 7tpoy6vrov yeyoveval EV OE tO U1t0
0"7tOUOairov Kat OlKairov, v
OE to a7t'aut~c; t~c; tou
fxOVtoc; apEt~c;

OlalpEital T, euyevEta de;

tpia crn yap aut~c; ~tEv
U1t0 evM~rov Kilt OVO!l!Xcrt&v yoverov yeyovevat, v
OE U1t0 0"1t0ll0!Xl(l)V Kilt
OlKairov, oE a7t'autilc; tile;
tOU exovtoc; apEt~c;. ecrn OE
to ~LEV a1tO EvOO~rov, olov
U1t0 ~(lO"lAErov Kilt ap:xovtrov, to OE a7tO cr7touOairov
Kat OlKairov, olov a1to Sevo<provtoc; Kat NEiA.ou Kat t&v
towutrov, to OE a1t 'aut~c;
...
"
'
...
f'
tTJc;
tOll... E"JCOVtOc;
!XpEtTJc;,
OlOV
,,
1'
...
'
!XV 1l YEVV!Xtoc; Kill !1EYilAo7tpnTtc; Kat oiKmoc; Kat ta
&A.A.a ta totauta

ev

ev

ev

OtalpEital T, euyevEla de;

etOTJ tpia ecrn OE IXUt~c; Ev
!lEv U1t0 evM~rov Kilt
ovo~t~crt~v yo~erov yeyo~e

val, EV 0 IX7t0 0"1tOll01XlWV

Kilt OlK!Xtrov, Ev OE a7t'autilc;
tile; tou f:xovtoc; apetilc;. ron
OE tO !lEV U1t0 evM~rov Kilt
f'
OVO!l!XO"trov YYOVV!Xl, OlOV
U1t0 ~(lO"lAEWV Kilt ap:x6vtrov n~titv ttva Kat M~av
exovtrov. a7t0 OE 0"1tOllO!Xirov
Kilt OlK!Xl(l)V' OlOV U1t0
:::evO<provtoc; Kat NEiA.ou Kat
trov tOlOUtrov Kilt !1EYIXAo7tpE1tttc; Kat oiKatoc; Kat ta
totauta
'

,...

L conferma l'espunzione di [yeyoveval] proposta da Jensen. Aperta rimane Ia

questione se NEiA.ou (S. Nilo) abbia sostituito un originale fpl>A.ou, il figlio di
Senofonte morto a Mantinea (Rose, Wendland, Rossitto) oppure se sia
corruzione per NElA.eou o NElKtou (Sudhaus) o NeA.roc; (? Gigon). Cf. Rossitto
[1984], 157-158.

ML

Titulus deest in M ( OmipEcrtc; uya6ou L)

Ol!XlpEtt!Xl tO aya8ov de; tEcrcrapa ecrn yap IXUtOU Ev !1Ev a pEtit Kilt OlK!XlooUVTJ' EV OE tO
exov apEtlJV' Ev OE tO O"ll!l<pEpov' Ev OE tO tEp7tVOV Kilt tO ~OEcr81Xl1tOlOUV. olov apEtlt !lEV
Kilt OlK!XlOOUVTJ on toutrov EKUtEpov aya8ov AEyEt(lt Etvat I' Ev OE to f:xov apEtf,v' OlOV
t1t7tOc; Kill &v8p(l)1toc; Kill ta tOl!XUta Kilt yap tootrov EKIXO"tOV AZyEt!Xl aya8ov dvat, eav
xn tlJV apEtf!v. tO OE O"ll!l<JlEpov' otov YlJ!lVIXO"ta Kat <pap!laKEia Kat t&Ua ta 7tpoc;
irydav Kat EUE~iav. Kat yap tOOtrov EKacrtov AZyEtal aya80v dval ~2&v O"llll<JlEPU. to OE
tEp1tVov Kal. to ~OEcr8at 1towuv, otov imoKpttitc; Kat auATJtilc; Kat t&v ~projl.litrov evta Kat
ta tOtaUta 0 yap {moKpttitc; Kill llUATJtftc;, ou tip ro<pEAElV aya8ov AEyEtal 3 &Ua tip
tEp7tElv [tou &ya8ou dval]
1 M: AZyEtal L - 2M: roc; L - 3M: aya8ov dvat A.tyetal L

L conferma l'espunzione di [tou aya8ou dval] in M (glossa?) proposta da

Wendland. La corretta deduzione di Moraux relativa alia lacuna per saut du
meme au meme di M e fondata su false indicazioni di Mutschmann. La frase
incriminata e, infatti, tramadata da M; l'omissione risale a Mutschmann o a
Wilke (che collaziono M). Cf. Rossitto (1984) 244-245. Non necessarie le
correzioni di Gigon aM: OlKillOO"UVTJ [on] aya8ov AEyEtlll dval tOUt(l)V EKUtEpov e
eue~iav (7tpocr<popa).

164

TIZIANO DORANDI

[30] MAL
M
Titulus deest

A
Titulus deest

ounpei:mt to &A.,eE~ d~
t pia. ecrtt yap autou EV !lEV
to 7tpiXY!la &A.,SE~ dvat, ev
OE tov A.Oyov, v oE Kat to
cruvall<pOtEpOV. ecrtt OE tg
7tp&Yila &A.,eE~ omv
outro~ exov Kat OUK liA.A.ro~.
otov to tov Seov aSavatov
dvat outro~ yap exu to
7tpilylla Kat ouK liUro~. to
oe A.f.yuv 'outo~ f.crttv
aA.11ffit~ A.Oyo~' tou aA.118ou~
E<JttV. ecrtt yap 0 A.Oyo~
1t p&Yila u1to tov A.Oyov
toutov ov. to OE A.f.yuv on
'oi.liv8pro7tot Ell'VUXOt dcrt'
Kat oA.Oyo~ aA.118Tt~ icrn Kat
to 1tp&ylla

ta crUVall<pOtEpa

n to llEv 1tpUYila touto aA.118E~

ll~~

L
otaipecrt~ ayaSou

to !!Ev 7tpilyll<X cXA118~ etvat

ll~~
out~ yap

omittit

exEt .... touwv ov

Kat to 7tpaylla tou aA.,Sou~

lipa to !lEV icrtt to 1tpaylla
aA.,SE~. to OE tOV A.Oyov
dvat aA.11 Sil. to oE ta
cruvall<potepa Kat to
7tp&Jp.a Kat tov A.Oyov
c:XA.11tti\ etvat

Bibl Mutschmann (1906) 49-50; Boudreaux (1909) 223; Moraux (1977) 109;
Rossitto ( 1984) 223-230
L'aggiunta dell'elenco finale in A non prova che questo debba essere reintegrato in MeL. Cf. Rossitto (1984), 225-226. Non necessarie le correzioni di
Gigon in M: to OE AEyEtv, (... ) 'oUto~ ecrttv aA.11ffit~ A.Oyo~' [tou cXA118ou~ ecrnv].

[22] M L
Titulus deest in M

(OtatpEcrt~ til~ trov

ovtrov OVOilacr{a~ L)

OtatpEttat i] trov ovtrov OVOilacr{a Ei~ 1tEVtE. ecrn yap autrov ta !lEV i.Xmep EV 1tp0~ 1tOAAa,
ta OE ro~ 1tOAAa 7tp0~ 1tOAAU Kat 7tp0~ EV' ta OE ro~ clVOilOta 1tp0~ OllOta, ta OE ro~ OllOta
7tp0~ OllOta, ta OE ro~ EV 7tp0~ EV. ecrtt OE to !lEV ro~ EV 7tpo~ 1tOAAU otov taxtcrto~ Kat
llEYtcrtO~ Kat KaAAtcrto~ Kat taUa ta to taut a. ta OE ro~ 1tOAAU 7tp0~ 1tOAAU Kat 7tp0~ EV'
otov 1tA.Eiro taoe trovoe Kat llet/;ro taoe trovoe Kat KaUiou~ o'loe trovoe Kat tliUa tO.
tOtauta. tO. OE ~ cXVOilota 1tp0~ Ollota, otov 1ta~p 1tp0~ ui.ou~ Kat 0Ecr1t0tll~ 7tp0~ oouA.ou~
Kat tliUa tO. totauta. tO. OE ~ OllOta 7tpo~ OllOta, otov clOEAcpOt 7tpo~ aoeA.cpou~ Kat cplA.ot
7tpO~ cplA.ou~ Kat tO. totauta. tO. OE ~ ev 7tp~ ev, otov KaUirov ooe touoe Kat Sacrcrrov ooe
tOUOE Kat ta totauta
BibiMutschmann (1906) 46-7;Moraux (1977) 107-108, 115;Rossitto (1984) 186-192
Ho riprodotto il testo stabilito da Moraux. L conferma Ia lacuna per saut du
mime au mime in M gia individuata e parzialmente restaurata da Rose. II testo
di L puo essere accolto anche se exempli gratia a colmare Ia lacuna; non e
esclusa, infatti, una diversa disposizione degli esempi o un numero variato.
Piu sicura Ia proposta di integrare Ia frase finale nella redazione di L
sfigurato, di nuovo, da una piccola lacuna per saut du mime au mime (oOE ... ooe).

DIVISIONI ARISTOTELICHE

165

... Ego apis Matinae

more modoque,
grata carpentis thyma per laborem
plurimum, circa nemus uvidique
Tiburis ripas operosa parvos
carmina jingo.*
BIBLIOGRAFIA

Biedl, A. (1955) Das Grosse Excerpt IP. Zur Textgeschichte des Laertios Diogenes
(Citta del Vaticano)
Boudreaux, P. (1909) "Un nouveau manuscrit des Divisiones Aristoteleae", Revue
de Philologie 33: 221-224
Gaiser, K. (1963) Platons ungeschriebene Lehre (Stuttgart) Testimonia Platonica
Gigon, 0. (1987) Aristotelis opera. III: Librorum deperditorum fragmenta (Berolini
et Novi Eboraci)
Hambruch, E. (1904) Logische Regeln der Platonischen Schute in der Aristotelischen
Topik (Berlin)
Heitz, E. ( 1868) in Aristotelis Opera omnia graece et latine (Parisiis)
Knoepfler, D. (1991) La Vie de Menedeme d'i:retrie de Diogene Laifrce. Contribution a
l'histoire et a la critique du texte des Vies des Philosophes (Basel)
Kramer, H.J. ( 1983) Altere Akademie in Flashar, H., ed., Die Philosphie der
Antike, 3. Altere Akademie, Aristoteles, Peripatos (Basel-Stuttgart)
Mansfeld, J. (1993) "Physikai doxai e problemata physica da Aristotele ad Aezio
(ed oltre)" in Battegazzore, A.M., ed., Dimostrazione, argomentazione dialettica e argomentazione retorica nel pensiero antico (Genova) 311-382
Moraux, P. ( 1977) "Temoins meconnus des Divisiones Aristoteleae", L 'Antiquite
Classique 46: 100-127
Mutschmann, H. (1906) Divisiones quae vulgo dicuntur Aristoteleae (Lipsiae)
Pasquali, G. (1910/1986) "Doxographica aus Basiliusscholien", Nachrichten der
Akademie der Wissenschaften in Giittingen, phil.-hist. Kl.: 194-228 = Scritti
filologici (Firenze) 1, 539-574
Praechter, K. (1910) Rec. di Mutschmann (1906), Berliner Philologische Wochenschrift 30: 387-395
Rose, V. (1863) Aristoteles Pseudepigraphus (Lipsiae)
Ross, W.D. (1955) Aristotelis Jragmenta selecta (Oxford)
Rossitto, C. ( 1984) Aristotele e altri. Divisioni. In trod., trad. e comm. (Padova)
--(1989) Rec. di Gigon (1987), Elenchos 10: 213-215
Suckow, F.G.G. (1855) Die wissenschaftliche und kiinstlerische Form der Platonischen
Schriften in ihrer bisher verborgenen Eigenthiimlichkeit dargestellt (Berlin)
Souilhe, J. (1930) Platon, Oeuvres completes. Dialogues apocryphes: 13.3 (Paris)
Wilpert, P. (1941) "Neue Fragmente aus llEpt taya8ov", Hermes 76: 236-241
- - (1949) Zwei aristotelische Friihschriften iiber die Ideenlehre (Regensburg)

CNRS, Paris
Una stesura dell'articolo ha letto con Ia consueta disponibilita e sostanziali suggerimenti, soprattutto per Ia costituzione dello 'stemma', il Prof. R.
Kassel (Koln). Di alcuni punti ho discusso con profitto con i miei colleghi e
amici L. Brisson (Paris), R. Chiaradonna e E. Spinelli (Roma).

PART FOUR

HELLENISTIC AND EARLY ROMAN PHILOSOPHY

THE CATALOGUE OF CHRYSIPPUS' LOGICAL WORKS

JONATHAN BARNES

At the end of his brief Life of Chrysippus, Diogenes Laertius

remarks that
since his books have a very high reputation, I have decided to
record here the list of them arranged by subject. They are these.
(VII 189)

There follows an articulated catalogue of Chrysippus' works.

Diogenes has a standing interest in the writings of his subjects,
an interest which he implicitly avows in his preface (I 16). Almost
all the Lives refer to what their subjects wrote; 1 and in the vast
majority of them Diogenes presents a book-list. The list of Chrysippus' writings stands out on three diverse counts. First, it is
articulated into sections and subsections,2 and the articulation is
based on philosophical principles. 3 Secondly, it is incomplete: the
end of Book VII is missing from all surviving manuscripts of
Diogenes4-and with it half the Chrysippean bibliography.5
Thirdly, it is exciting; for it appears to offer us information about
Chrysippus' philosophical activities, and in particular about his
logical activities, which we cannot find elsewhere-it parades his
terminology; it shows where his interests lay; it indicates the
structure which he gave to his philosophical work. 6
1 Exceptions: Myson, Anaximander, Anaximenes, Archelaus, Socrates,
Eubulides, Diodorus, Menedemus of Eretria, Lacydes, Lycon, Menedemus the
Cynic, Melissus, Leucippus, Anaxarchus. Occasionally-and particularly if the
subject wrote nothing (e.g. Arcesilaus, IV 32; Carneades, IV 65)-the reference is brief (e.g. Polemo, IV 20; Bion, IV 47; Clitomachus, IV 67; Eudoxus,
VIII 89).
2 Other articulated lists: Plato (IV 57-66), Heraclides Ponticus (V 86-88),
Antisthenes (VI 15-18), Diogenes of Sin ope (VI 80), and Democritus (IX 46-49).
3 Compare e.g. the list of Democritus' works (drawn up by Thrasyllus): IX
46-49. Contrast e.g. the Platonic tetralogies: III 57-66.
4 On the loss see Mansfeld (1986) 308-312; Dorandi (1992).
5 What survives lists some 350 titles: in all, Chrysippus wrote 'more than
705' works (VII 180; llltEp 1tEVtE l((ll e1ttU1COO"Ul EO"ttV: a bizarre remark-change
1tEvtE to 1tEVtT(KOVt<X? or delete 1tEVtE K<Xt ?)
6 Odd then that it has accumulated so little literature Numerous scattered
comments, of course; and annotated translations in Italian (Gigante (1983);
Baldassarri (1985)), German (Hiilser (1987/8)) and French (Hadot (1994)).

170

JONATHAN BARNES

The catalogue appears to offer such information. But what is its

pedigree? What authority does it possess?
Diogenes does not name his source; but we can guess it. On the
one hand, we know that Apollonius of Tyre 'published the table of
the philosophers of Zeno's school and of their books' (Strabo, XVI 2.
24 (757 C)) ,7 a work which must have included a catalogue of
Chrysippus' writings. On the other hand, Apollonius is an author
whom Diogenes cites more than once.s Diogenes, like many
another scholar, sometimes cites sources which he knows only at
second hand; but nothing hints that he did not himself read
Apollonius; and in any event it is eminently reasonable to believe
that his list came from Apollonius, either directly or indirectly.
What do we know about Apollonius? Next to nothing. 9 His date
is roughly fixed; for Strabo describes him as living 'a little before
us', which presumably puts him in the first part of the first century
BCI 0-and Philodemus cites him in the Index Stoicorum (col xxxvii),
which cannot be dated much later than 50.11
Scholars generally assume that he was a Stoic philosopher; and
a popular view makes him a follower of Apollodorus of Seleucia. 12
He wrote about the Stoics, and part of the work betrayed at least a
quasi-philosophical interest in the subject. No doubt you might
But apart from Dyroff (1896) and Brunschwig (1991) I know of no sustained
study.-The present paper will, I hope, eventually form part of a larger work.
I am indebted to several friends and colleagues; and in particular to Keimpe
Algra, whose comments enabled me to correct a number of errors in the
penultimate version of the paper.
7 eK Tupau ~ 'Avtfnlltpo; Kill J.LtKp{>v np(> ftJ.L&v 'AJtOA.Acl>vt~ b tov nfvllKil K6ei~ t&v
alto ZfJvwvoc; qnA.ocro<pwv Kill t&v PtPA.fwv. There is no reason to think that Strabo's
phrase is, or is inten!ied as, a book-title; and the ltlVIl~ was surely presented
in Apollonius' work On Zeno, which was in more than one book (see Diogenes Laertius, VII 2, 6): so Mejer (1978) 75, n.33; Hahm (1992) 4166 n.207.
8 See VII 1, 2, 6, 24, 28. (For the spurious correspondence between
Antigonus and Zeno at VII 6 see Cronert (1906) 28-29; Dorandi (1994) 9 and

n. 41.)

9 On Apollonius see e.g. Goulet (1989a); Dorandi (1994) 6, 33. -Hahm

(1992) argues that Apollonius was Diogenes' primary source not only for the
catalogue but also for the biographies of Zeno and his six followers and for
the Stoic doxography: like most essays in Laertian Quellenfurschung, this will
persuade no-one but its author.
10 Strabo uses the same phrase of Antioch us (XVI 2.29 (759C)), whose dates
are c. 130-68/7; cf. e.g. VIII 7.5 (387C), XII 3.41 (562C)-events datable to the
early 60s.
H Dorandi (1990) 2334-2335.
12 So Dyroff (1896) 39-41; cf. Cronert (1906) 80, 180; Baldassarri (1985) 1216 (see also Hiilser (1987) LXXXVII).

THE CATALOGUE OF CHRYSIPPUS' LOGICAL WORKS

171

have a philosophical interest in the Stoics without being a Stoic--or

even a philosopher. Nonetheless, it is not a bad guess that Apollonius was a Stoic.
Why a follower of Apollodorus? The argument is this. The
catalogue regularly uses the word t67to<; to mark areas or divisions
of its subject matter; and 'Apollodorus calls these parts (i.e. the three
parts of philosophy) t67tot, while Chrysippus and Eudromus call
them JlEPTJ and others call them yEVTJ' (Diogenes Laertius, VII 39).
So when Apollonius talks of t67tot he is using a terminology
peculiar to Apollodorus; and hence he was probably a member of
his school.
Evidently, the argument is at best frail. In fact it is worthless; for
in the catalogue the word t67to<; is not used in its Apollodoran
sense. Apollonius speaks, say, of 'the logical t67to<; concerning
7tpayJlata' (VII 190), thereby referring to that area or part of logic
which deals with 'objects signified'. Thus logic itself is not designated a t67to<;: rather, logic has t67tot. Apollonius uses the word
t67to<; in its standard sense. He does not use it in its special Apollodoran sense.
Allow that Apollonius was a Stoic, and that he was Diogenes'
source for the Chrysippean catalogue. Did he compile the
catalogue himself?
Diogenes provides book-lists for the other Stoics whose lives he
records-Zeno (VII 4), Aristo (VII 163), Herillus (VII 166), Dionysius (VII 167), Cleanthes (VII 174-175), and Sphaerus (VII 178). The
lists are short, and unlike the Chrysippean list they are not
articulated. But it is an economical conjecture that they all derive
from Apollonius. Now the catalogue of Aristo's works ends with
'Letters -4 books'; after which Diogenes makes the following
comment: 'But Panaetius and Sosicrates say that only the Letters are
his, the other items belonging to Aristo the Peripatetic' (VII 163). 13
The comment presupposes that Panaetius and Sosicrates had a list
of Aristo's putative writings. On chronological grounds it is unlikely that this list came from Apollonius. 14 It follows that there
was a catalogue of Aristo's writings in existence before Apollonius
13 Panaetius is mentioned in similar contexts at II 64 (Aeschines of
Sphettus), 85 (Aristippus), and III 37 (Plato). For Sosicrates see II 85 (Aristip~us) and VI 80 (Diogenes of Sinope).
4 Panaetius died in 110/09 (Dorandi (1991) 41-42); Sosicrates' dates are
uncertain-but he probably flourished in the second century (Giannattasio
Andria (1989) 73-75).

172

JONATHAN BARNES

published his On Zeno. And if for Aristo, surely also for Zeno and
Cleanthes ... and Chrysippus?
Hence Apollonius did not himself compile the Chrysippean
list-or at any rate, he was not the first to compile such a list.
Perhaps there were catalogues in the library of the Stoa, which
Panaetius and Sosicrates had consulted? Perhaps Apollonius
merely prepared these for public consumption, adding a few
embellishments of his own? 15 All this is lamentably vague; but I
doubt if we should speak with confidence of Apollonius as the
compiler of the Chrysippean catalogue.
However that may be, there are more important questions to
broach: in particular, how close is the catalogue to Chrysippus
himself? Did the compiler arrange the works to reflect the scope
and structure of Chrysippus' logical interests, or does the articulation of the catalogue tell us more about the compiler's own
interests? Did he copy Chrysippus' titles, or did he (in some or all
cases) invent titles of his own? 16 It is generally assumed without
argument that the catalogue gives us reliable information about
Chrysippus. Here is an argument. It starts from a remote origin.
The catalogue presents all the minor problems which beset any
text of Diogenes. There are also two major problems. First, the
logical section of the list ends by observing that the total number of
books in Chrysippus' logical oeuvre was 311 (Of.LOU 'ta 7tUV'ta 'tOU
A.oytKou na': VII 198). But if you count the works which Diogenes
lists you will not readily arrive at that figure. Secondly, the ethical
part of the catalogue-the surviving fraction of the ethical partcontains several titles which apparently point to a logical rather
than an ethical content: On Divisions and On Contraries, for example;
or On Dialectic and On the Objections to the Dialecticians. Can these
really be titles of works whose main effort was devoted to ethics?
Why not amalgamate the two problems? Suppose that a few
pages in a manuscript copy of Diogenes worked loose and were
15 Note that Strabo says that he made them public (h9Ei~). not that he
collected the materials for them.
16 We should not assume without ado that Chrysippus entitled his writings. Note e.g. that Plotinus gave no titles to any of his essays-those of his
pupils who managed to get hold of copies gave them different titles (Porphyry, Vit. Plot. 4: see Brisson et al ( 1982), 283, n.1). Similarly, Galen did not give
titles to writings which he sent to his pupils and friends, and he complained
that these works were later in circulation under different titles (Lib. prop. XIX
10-11 K). Many of Chrysippus' works are addressed to pupils and friends.

THE CATALOGUE OF CHRYSIPPUS' LOGICAL WORKS

173

later replaced in the wrong order; and that this disorderly copy is
the source of our modern text. Then if we cunningly rearrange the
pages we may, at one blow, save the arithmetic at VII 198 and
restore to their proper place the logical works which now appear in
the ethical part of the list.17
First, let us look more closely at the arithmetic. The first four
logical 't07tot contain, by my count, 118 works divided into 263
books. (But in at least three places the manuscripts offer variant
readings for the book numbers, 18 so that the transmitted total is
between 261 and 265.) The total includes four works in eight books
which are marked as spurious; and it is not clear whether they
should be included in the arithmetic. In addition to the contents of
the first four 't07tOt, there are the 39 books mentioned at VII 198.
Hence in all there are 303 books, plus or minus two (or 295, plus or
minus two, if we do not count the spurious items). 19 Thus we are
between 18 and 6 short of the desired total of 311.
It is easy to emend the text: for example, read 'ta' ('301 ') instead
of na', and then take the lower of the variant readings. But the
possibilities are so many, and corruption of numerals is so common,20 that it is a waste of time to play the numbers game. Hence it
is tempting to observe that the transmitted numeral is not wildly
out, to conclude that Chrysippus' logical bibliography contained
about 300 books, and to abandon as delusory the hope of establishing a precise figure.
But this leaves aside the second of the two major problems.
Roughly speaking, the apparently logical items in the ethical part
of the catalogue are the contents of the second, third, fourth and
fifth cruv'ta~n<; of the first ethical 't07tO<; and the contents of the first
and second cruv'ta~n<; of the second 't07tO<;. In all, they amount to 26
works in 70 books. Suppose that the text were rearranged so that all
these items appeared in the logical section: then the total number
of logical books would be about 370. (And should we emend na'to 'toa'?2 1)
17 For an example of this sort of thing see Egli ( 1967) 3.
18 I rely on the notoriously unreliable OCT: doubtless there are variant

readings for many more than three titles.

19 I have counted as carefully as I can, and three times. But scholars differ,
even in their addition. E.g. Dorrie (1970) col. 151, finds 298 books; Baldassarri (1985) 5, counts 305; Harlot (1994) 336, gives 300.
20 As Galen observes: Antid. XIV 31-32 K.
21 Egli's reordering of the catalogue purports to yield a grand total of 321

174

JONATHAN BARNES

It would be absurd to deny that some of these 70 books should in

some sense be counted as part of Chrysippus' logical oeuvre. On
Dialectic, to Aristocreon, whatever else it contained, surely included
some Chrysippean reflexions on dialectic;22 the works on definitions, on species and genera, on division must have published
some logical reflections;23 again, there are titles dealing with
etymology and poetry and rhetoric, all of which fell under the
generous Stoic conception of A.oytJCT,.2 4 But it does not follow that
these titles should be transplanted to the logical section; and in
point of fact no simple hypothesis of mechanical corruption is
satisfactory and no simple textual transposition resolves the problem.25 Complex hypotheses are delightful to construct-and impossible to justify. If the text of Diogenes is seriously disorganised,
then we can no longer (or not yet) hope to put it right.
That being so, we might as well ask whether the second major
problem cannot be solved without textual surgery.26 Well, consider
the rubric to the second ethical t67to<;: it is the area concerned with
'reason in general and the arts and virtues which are constituted
by it' (VII 201); that is to say, the rubric announces that the section
will contain logical material. Hence it would be quite unwarranted
to transfer the first <JUVtU~t<; of this t01tO<; to the logical part of the
catalogue-or the second cruvta~t<;, which contains titles bearing
on the arts or virtues of dialectic and rhetoric. Similarly with the
first ethical t67to<;; for this t67to<; deals with 'the articulation of
ethical concepts' (VII 199), and hence with matters on the borders
of logic.
In brief, the rubrics to the first two ethical t67tot suggest that they
collected items dealing with the logical aspect of ethics. No need
then to blench if some of the titles promise a logical content.
A question remains: on what grounds did the compiler of the
items; so he emends to tKo:'. But I am not sure how he does his arithmetic.
Moreover, his selection of 'logical' works from the ethical part is capricious.
22 But see below, p. 177.
23 The topics are within the scope of logic: e.g. Diogenes Laertius, VII 60,
61, 61.
24 E.g. Diogenes Laertius, VII 60, 41.
25 It is enough to note that the rubrics attached to the apparently logical
t67tot are all ethical: if we transfer the titles, we must leave the rubrics
behind; and this scotches the straightforward hypothesis that a folio dropped
out of an early manuscript and was stuck back in the wrong place.
26 'The fact that the arrangement in the bibliography is disturbing need
not entail that it has been disturbed' (Mansfeld (1986) 358 n. 147).

THE CATALOGUE OF CHRYSIPPUS' LOGICAL WORKS

175

catalogue assign, say, On Dialectic to ethics? Assume that the work

said something about logical matters-otherwise its title would be
wholly misleading. Assume, too, that it contained at least some
ethical content-otherwise its position in the ethical section is
inexplicable. Why, then, assign it to ethics rather than to logic?
(No doubt it is absurd even to guess at such questions. But what the
hell, archie, what the hell.)
Chrysippus did not secure the three parts of philosophy each in
a watertight compartment, and his thought often seeped from one
to another-he could touch on physical and ethical matters in his
logical writings, on logical and physical matters in his ethical
writings, on logical and ethical matters in his physical writings.27
It is clear, then, that the compiler will have faced problems once he
determined to order his catalogue according to the standard tripartition of philosophy. Now the logical section contains virtually
no maverick titles: virtually all the titles clearly suggest a logical
content. Perhaps, then, the compiler, convinced of the superiority
of ethics to logic and of physics (or theology) to ethics, decided to
adopt the following principle: any work which dealt substantially
with physical matters would be placed in the physical section of
the catalogue, whatever its title; of the remaining works, any
which dealt substantially with ethical matters would be placed in
the ethical section of the catalogue, whatever its title; and the rest
would be ranked under logic. This is at least a rational principle
and it explains (better than any other hypothesis I have come
across) the puzzling entries in the ethical section. 28
If some such suggestion is right-and, more generally, if the
catalogue in its present form is not seriously disturbed-, then we
may plausibly draw two merry consequences. First, the compiler
must have taken his task seriously; for he could have had no
reason to place a work entitled On Dialectic in the ethical part of the
catalogue unless the content of the work convinced him that there,
despite its title, it belonged. In other words, the catalogue was not
27 Thus in his account of Stoic logic Diodes mentions, inter alia, texts from

On the Soul and Physics (Diogenes Laertius, VII 50, 54, 55).
2 See e.g. Long (1978) 117-118, who appeals to the idea that 'the study of
dialectic, for Chrysippus at least, is an integral part of moral conduct'. Such
an idea might lead you to range all logical titles under the heading of ethics;
but it does not yet explain how you could place some of them under ethics
while leaving most in an independent section on logic.
Ch~sippus'

176

JONATHAN BARNES

constructed on the basis of titles alone: the compiler must have had
some knowledge of the contents of the works he was cataloguinghe must have unrolled at least a few columns of text.
This claim remains plausible even if the catalogue has been
substantially deranged. For some of the titles in the logical section
itself are puzzlingly placed.29 Had the compiler looked no further
than the title of, say, llr8awx IVVTJJ.lJ.lEVa, he could have had no
reason to place the work in the first-and no doubt introductory't07t0~.30 Its position suggests that he must have troubled to form
some idea of its contents.
Secondly, the book-titles must come from Chrysippus and not
from any compiler; for a compiler would surely have given each
work a title appropriate to the position it was to hold in his list. If
you have an untitled work which you intend to place in the first
ethical 't07tO~, you will not call it On Dialectic: you will give it a name
which fits its place-for example, Dialectical Matters bearing on Ethics.
This second claim also remains plausible even if the catalogue
is seriously disturbed. Suppose that the compiler found an untitled
work which, on the basis of its contents, he decided to place in the
first part of the logical section: would he have given it the bald title
Plausible Conditionals?
With these two claims I have returned to my sheep: the
catalogue has an honest pedigree. No doubt the compiler made
mistakes: he certainly missed a few works,3 1 and he probably misplaced others. No doubt the list has been damaged in transit.
Certainly there are titles whose import remains opaque to us. Nonetheless, we may properly boast that we possess a fairly competent
and a fairly well preserved and a fairly intelligible list of Chrysippus' logical writings. In particular, we may confidently use the
titles as evidence of Chrysippus' own logical lexicon; and we may
plausibly discern in the catalogue a general indication of Chrysippus' logical interests-of the topics he did and did not discuss, of
29 You might, of course, hypothesize serious local disruptions (compare the
case of Aristotle's catalogues); but I have not managed to contrive an interestin~ hypothesis in this vein.
0 On the work-the title of which is usually misunderstood (e.g. Hadot
(1994) 338)-see Barnes (1985).-The Mss. give no subtitle to the first 'tonoc;:
von Arnim (1903) II 4, rightly suggested something like 'Aoyu(Qu 'tonou <'tou
1tEpi 't~V ouxp9pcootv 'tiDV A.oyuc&v evvot&v>' (misreported by Hiilser ( 1987) 172).
3! See Hiilser (1987) 190-191.-1 assume that he was trying to give a
bibliography of Chrysippus' works rather than, say, a list of the Chrysippean
holdings in some library.

THE CATALOGUE OF CHRYSIPPUS' LOGICAL WORKS

177

the breadth and depth of his researches, of the relative weight

which he gave to different parts of the subject.
The individual titles must wait on a detailed commentary. Here, a
word on Chrysippus' logical interests. And first let it be underscored that there are limits to the inferences we may draw from
the list. 32 For ancient book-titles-even when they were not deliberately allusive or fanciful-do not invariably give exact information
about the contents of the books to which they are affixed. First, titles
were sometimes incipits. Protagoras' On the Gods was so-called
because it began with the words Ilpt J.!EV 8.rov ... : it was not a work
about the gods at all.3 3 (I have wondered whether On Dialectic, to
Aristocreon did not begin: 'On dialectic, 0 Aristocreon, I here say
nothing; for my subject is ethical .. .' .) Secondly, titles are often
vague, so that one and the same name could be attached to books of
very different stamp: consider, say, IlEpt "Oprov or Ilpt 'OpJ.!fic;. 34
Thirdly, ancient philosophers were less subject-bound than
scholars (modern and ancient) care to imagine. Poor Plato was
incapable of sticking to a single topic: ancient scholars felt obliged
to list his IloA.uuc6c; as a 'logical' work;3 5 modern scholars have
wearied themselves and their readers with fatuous efforts to
discover what the Phaedrus is 'about'. In the case of Chrysippus we
know that there was logical material in works which we should
classify-and which the ancient catalogue classified-as works on
physics; and we might recall the chagrin which Dionysius of
Halicarnassus felt when he contemplated the contents of the books
entitled Ilpt cruv1:a~roc; ( Comp. verb. 31-32).
Nonetheless, the catalogue does allow a few banal inferences.
First, Chrysippus wrote a large number of logical works: at least
130 essays or treatises, amounting to over 300 rolls. Aristotle's
Organon consists of six works in fifteen books. Chrysippus' logical
oeuvre was twenty times the length of the Organon.3 6
Secondly, three aspects of logic received especial attention.
Chrysippus liked puzzles: twelve works in 23 books on the Liar;
32

Brunschwig (1991) 80, is perhaps too sanguine on this point.

33 See Nachmanson (1941) 50-52; compare Dyroff (1896) 14-15. Note also

Galen, Lib. prop. XIX 11 K.

34 See Dyroff (1896) 14-21.
35 Diogenes Laertius III 58; see Dyroff (1896) 14-15.
3 6 But not twenty times the length of Aristotle's logical reuvre, much of
which is lost.

178

JONATHAN BARNES

another nine works in another 26 books on other puzzles; seven

works in 17 books on amphiboly. In all, more than a sixth of his
logical production. Again, there are some eighteen works in 35
books on arguments-on the various forms of inference and
syllogism which the Stoics recognised and investigated. (More
than it seems, since technical logic uses little papyrus.) Finally,
there are some twenty works in 46 books on the elements of
arguments: on the various forms of complex proposition (five works
in 12 books on conditionals); on simple propositions; on predicates
and names. Chrysippus devoted as much paper to the elements of
arguments as to the arguments themselves-and here his work
has no parallel in the Peripatetic tradition, where this aspect of
what we call 'philosophical logic' was scarcely studied.
Thirdly, it is plain-and unsurprising-that most of Chrysippus'
works were technical essays or treatises on particular points within
logical theory. Almost all of these works are wholly lost. What we
are accustomed to refer to as the 'fragments' of Stoic logic are
mainly drawn from summaries and allusions and polemical
caricatures. Much of this material surely derives from handbooks,
and the handbooks themselves derived in the end from the
original works of Chrysippus and his followers. But none of the
subtleties of the originals was conserved: the account of Stoic logic
in Diogenes Laertius is, for us now, richly informative; but it is also
the thinnest of gruels-and although it is one of our main pieces of
evidence for Stoic logic, it is not a piece of Stoic logic.
Finally, what of the organisation of the catalogue? In particular,
does it perhaps reflect the way in which Chrysippus himself
organised his logical studies? 37 Does it perhaps follow the arrangement of topics in one of Chrysippus' own introductory books on
logic? We can no longer read these books; but Diogenes Laertius
offers us two general accounts of Stoic logic, the summary at VII
42-48 and the longer version at VII 49-82, 38 and we can at least
compare the structure of these accounts with the arrangement of
the catalogue.
Here is a pretty crude and tentative description of the three items:
37 So Tieleman (1992) 246; and a hint in Mansfeld (1986) 363-4.
3 8 On which the fundamental article is Mansfeld ( 1986); see also Cell up rica (1989) (but I remain stubbornly convinced that the whole account comes
from Diodes of Magnesia ... ).

THE CATALOGUE OF CHRYSIPPUS' LOGICAL WORKS

179

The Catalogue :39

I LOGIC
A: Introductory
B: ~YJI't
1. Simple a~tcOJ.UX't(X
2. Complex a~tcOJ.La.ta.
3. Non-&~tcOJ.La.'ta.
4. Incomplete AEK'ta
5. 1t'tc00t<;

C: AS..;
1. Questions of usage
2. Elements and syntax
3. Ambiguity

0: A6yot
1. Forms of argument
2. Arguments
3. J.LE't<X1tl1t'tOV't<;
4. Hypotheses
5. The Liar
6. The Liar
7. Other paradoxes
8. The Nobody
9. Sophisms
10. croV'Tl9da.
E: Miscellaneous

II

ETIHCS
F: Introductory
1. Definitions
2. Definitions etc
3. Definitions etc
4. Etymology
5. Poetics
G: Kotv~~
1. General

39 A-H are the toltot, numbered sections are auvta~Et~. My names for tOltOt
derive from the text; I have invented names for the auvta~Et~.

180

JONATHAN BARNES

2.
3.
H: The
1.

Dialectic
The Virtues
Good and the Bad
Pleasure

The Long Account (Diogenes Laertius, VII 49-82):

A*: Epistemology
1. cpav'tacr{a

2. a\cr6l,crt<;
3. evvota
4. The Kpt't'fJptov
B*: ll@t cprovijc;
1. cprovf)
2. AE~t<;
3. O"'tOtXEta

4.
5.
6.
7.
8.

The parts of speech

The virtues of AE~t<;
Poetry
Definition
Division etc

9.

aj!cpt~oJ.{a

C*: ll@t :!t@"f!-'Cl-trov

1. AEK'ta
2. The parts of AEK'ta
3. Complete AEK'ta
4. a~troj!a'ta
5. A.&yot

6. Puzzles
The Summary (Diogenes Laertius, VII 41-48) :

A+: The Parts of Logic

B+: Rhetoric
C+: Dialectic
1. The parts of dialectic
2. cprovf)
3. cruA.A.oytcrj!ot

4. Epistemology
5. Dialectical virtues

THE CATALOGUE OF CHRYSIPPUS' LOGICAL WORKS

181

The summary is too thin to allow detailed comparisons. But it

seems that: C+ 1 answers roughly to A; C+ 2 matches C (and also
B*); C+ 3 corresponds to D and to C* 5; C+ 4 goes with A*; and C+ 5
fits with G. One thing is plain: if the summary derives from a
single source (as it surely does), then the source organised the
material in a manner very different both from the catalogue and
from the long account.
As for the catalogue and the long account, it is worth observing,
first, that they agree in one significant peculiarity: both follow an
unorthodox ordering of the parts of philosophy: <logic, ethics,
physics>. The order <logic, physics, ethics> is ascribed by Diogenes to Zeno and to Chrysippus (VII 41); and according to Sextus,
PHil 13, it was the order favoured by most Stoics. 40 (But Sextus also
says that the order <logic, ethics, physics> was in general the most
common (M VII 22); and at Stoic. rep. 1035A-F Plutarch takes
Chrysippus to task for inconsistently adopting now one and now
another order.41)
But this apart, the long account and the catalogue differ
dramatically. 42 At the first level of specificity, nothing in the catalogue corresponds to section A*, on epistemology, in the account.
Epistemological topics were surely discussed in some of the items
which the catalogue ranges under ethics; and I guess that others
were covered in the lost section on physics, perhaps in a subsection
on psychology. But it is striking that the catalogue, so far as we
know, did not imagine a special place for what many ancient and
modern critics have taken to be the glory, jest and riddle of the
Stoic world. Section C in the catalogue answers to B* in the
account, while B and D together cover C*. In the long account we
are told that 'most are agreed in thinking that the study of dialectic
starts from the area concerned with utterances' (VII 55). The word
'most' suggests that there were at least a few deviant thinkers; and
the compiler of the catalogue may have been among them.
Take next the subdivisions of B* and C*: B* 1 has no clear
counterpart in the catalogue-although some of the items in C 1
are pertinent; 43 B* 2 answers to part of C 1; B* 3, and perhaps B* 4,
40

It is the order which Sextus himself follows in PH and in M VII-XI.

42

See Mansfeld (1986) 357-8, 371-3.

41 On all this see Ierodiakonou (1993).

43 It is generally supposed that Diogenes of Babylon was the first Stoic to

take the study of <pmviJ seriously.

182

JONATHAN BARNES

to C 2; and B* 5 to another part of C 1. B* 6-8 are matched in the

ethical section of the catalogue, namely in F 5, F 2, and F 3; and B*
9 corresponds roughly to C 3. C* 1 perhaps answers to A. C* 2 goes
with B 4 and B 5; C* 3 with B 3; and C* 4-5 with B 1-2. C* 5
corresponds to D 1-2, and C* 6 to D 5-9. There are thus several
subsections in the catalogue which do not appear to be matched in
the account-scarcely odd, given the jejune character of the
account. More noteworthy are the differences in arrangement.
Finally, a sample from the lowest level of specificity-compare
C* 4 with B 1-2. In the table, the left column gives the order of
events within C* 4, the right column the corresponding titles in
the catalogue.44
(i) Simple) (non-simple a~u:oJla'ta B 1: 7tpt a~u.oJ.l(i'trov [7], 7tpt

'tOOV oux U7tAOOV a~tffiJllhrov [8]

(ii) Simple a~te.oJla'ta
a. Negations
b. Denials
c. Privatives
d. Categoricals
e. Determinates
f. Indefinites

(iii) Non-simple a~uoJla'ta

a. Conditionals

b. Quasiconditionals
c. Conjunctions
d. Disjunctions

B 1: 7tpt a7toq>anKrov [10]

B 1: 7tpt 'trov Ka'ta cr'tEpTtcrtv
AeyOJlEvffiV [12]
B 1: 7tpt 'tOOV Ka'tayopeunKrov [ ll]
B 1: 7tpt 'troV aoptO''tffiV a~tffiJlU'tffiV
[ 13], 1tepl. 'tftc; Otaq>op&c; 'trov
aoptO''tffiV [14]
B 2: 1tepl. ai..Tt8ouc; O"UVTtJlJlEvou
[ 18], 7tpoc; 'to 1tepl. aKo/..ou8rov
[20]
B 1: 7tpt 'tou O"UJl7tE1tAE'YJ.livou [9]
B 2: 7tepl. ai..Tt8ouc; ota~euyJlvou
[17], (?) 7tephou ota 'tptrov
[21]

e. Causals
f. Comparatives
g. Conditionals
44 Numerals in bold give the position of the item in the catalogue: the
first logical 't01toc; contains six items, so that [7] is the first item in the second
'to1toc;.

THE CATALOGUE OF CHRYSIPPUS' LOGICAL WORKS

183

h. Quasiconditionals
i. Causals
(iv) Modal a~te.OJ.La'ta
a. Plausible
b. Possible/impossible
B 2: 7tEpt ouva'to>V [22]
c. Necessary/non-necessary
No doubt the differences between the two columns were not as
great in reality as they appear in the table-thus, for example, the
fact that no title in the catalogue corresponds to 'categoricals' in
Diodes' account hardly implies that Chrysippus had nothing to
say on the topic. Nonetheless, real differences remain, and some at
least of them are to be explained by the fact that the account refers
to Stoic logicians who worked after Chrysippus.
From this spattering of factlets, what to infer? Suppose (as seems
pretty plausible) that the arrangement of items in the account
reflects a Stoic handbook on logic. Suppose (as seems mildly plausible) that the arrangement in the catalogue also reflects a Stoic
handbook. 45 In that case, the account certainly reflects a postChrysippean handbook; and the catalogue certainly reflects a
different handbook-and probably an earlier handbook. Hence-a
leap but not an Olympic leap-the catalogue may well reflect a
Chrysippean handbook.
In truth, this final argument is made of gossamer. But the
question of the structure of the catalogue is in any event of less
concern that the questions of its scope and its terminology.
Small beer, alas. And I had wished to offer a large vodka.
Geneva
BIBLIOGRAPHY

Arnim, H. von (1903) Stoicarum Veterum Fragmenta (Leipzig, 1903-05)

Baldassarri, M. (1985) La logica stoica: testimonianze e Jrammenti - II: Crisippo
(Como)
Barnes,]. (1985) 'ni9ANA l:YNHMMENA',Elenchos6: 455-467
Blumenthal, H., and Robinson, H.M. (edd.) (1991) Aristotle and the Later Tradition, Oxford Studies in Ancient Philosophy supplementary volume (Oxford)
45 Other suppositions, of course, are readily imagined-especially if the
compiler was himself an amateur of the subject.

184

JONATHAN BARNES

Brisson, L., Goulet-Caze, M.-0., Goulet, R., and O'Brien, D. (edd.) (1982)
Porphyre: la vie de Plotin, I (Paris)
Brunschwig, J. (1991) 'On a Book-Title by Chrysippus: "On the Fact that the
Ancients Admitted Dialectic along with Demonstrations"', in Blumenthal and Robinson (1991)
Celluprica, V. ( 1989) 'Diode di Magnesia fonte della dossografia stoica di
Diogene Larzio', Orpheus: 10, 58-79
Cronert, W. (1906) Kolotes und Menedemos (Munich)
Dorrie, H. ( 1970) 'Chrysippos', Paulys Realencyclopiidie der classischen Altertumswissenschaft suppt XII (Stuttgart)
Dorandi, T. (1990), 'Filodemo storico del pensiero antico', in Haase (1990)
Dorandi, T. (1991) Ricerche sulla cronologia dei filosofi ellenistici, Beitriige zur
Altertumskunde 19 (Stuttgart)
Dorandi, T. (1992) 'Considerazioni sull'index locupletior di Diogene Laerzio',
Prometheus 18: 121-126
Dorandi, T. (1994) Filodemo: storia dei filosofi-la Stoa da Zenone a Panez.io (PHerc.
1018), Philosophia Antiqua 60 (Leiden)
Dyroff, A. (1896) Uber die Anlage der stoischen Biicherkataloge, Programm des ...
Gymnasiums zu Wurzburg (Wurzburg)
Egli, U. (1967) Zur stoischen Dialektik (Basel)
Giannattasio Andria, R. (1989) I frammenti delle 'Successioni dei filosofi', Universita degli studi di Salerno, quaderni del dipartimento di scienze
dell'antichitii 5 (Naples)
Gigante, M. ( 1983) Diogene Laerz.io: Vite dei filosofi (Rome/Bari, 2nd edn.)
Goulet, R (1989a) 'Apollonios de Tyr', in Goulet (1989b) 294
Goulet, R. (1989b) (ed.) Dictionnaire des philosophes antiques I (Paris)
Goulet, R. (1994) (ed.) Dictionnaire des philosophes antiques II (Paris)
Haase, W. (1990) (ed.) Aufsteig und Niedergang der romischen Welt II.36.4
(Berlin)
Haase, W. (1992) (ed.) Aufsteig und Niedergang der romischen Welt 11.36.6
(Berlin)
Hadot, P. (1994) 'Liste commentee des oeuvres de Chrysippe (D.L. VII 189202)', in Goulet (1994) 336-356
Hahm, D. (1992) 'Diogenes Laertius VII: On the Stoics', in Haase (1992)
Hulser, K.-H. (1987/8) Die Fragmente zur Dialektik der Stoiker (Stuttgart/Bad
Cannstatt)
Ierodiakonou, K. (1993) 'The Stoic Division of Philosophy', Phronesis 38: 57-74
Long, A.A. (1978) 'Dialectic and the Stoic Sage', in Rist (1978)
Mansfeld, J. ( 1986) 'Diogenes Laertius on Stoic Philosophy', Elenchos 7: 295382 (reprinted in Mansfeld ( 1990))
Mansfeld, J. (1990) Studies in the Historiography of Greek Philosophy (Assen)
Mejer, J. (1978) Diogenes Laertius and his Hellenistic Background, Hermes
Einzelschriften 40 (Wiesbaden)
Nachmanson, E. (1941) Der griechische Buchtitel, Goteborgs Hogskolas Arsskrift
47 (GOteborg)
Rist, J.M. (ed.) (1978) The Stoics (Berkeley/Los Angeles/London)
Tieleman, T.L. (1992) Galen and Chrysippus-Argument and Refutation in the De
Placitis Books II-III, Quaestiones Infinitae 3 (Utrecht)

CICERO'S ROPE
R. j.

HANKINSON

Substitute for the time honoured 'chain of causation', so often introduced

into discussions upon this subject, the phrase 'a rope of causation', and
see what a very different aspect the question will wear. (J. Venn, The
Logic of Chance, London 1866, 320)

Metaphor plays a central role in the conceptual scaffolding of our

metaphysics. This is particularly the case the further removed
from the realm of ordinary experience are the phenomena being
discussed and analysed (consider the cases of 'wave' and 'particle'
in current high-energy physics); but it is also true at a less
recherche level. The choice of metaphors is not only conditioned
by the particular presuppositions of those who coin them, but also
affects the way in which that conceptual structure develops and
then ramifies through the wider linguistic community.
In this paper I wish to consider the origins of Venn's hallowed
causal metaphor of the chain. Effects of previous causes become in
their turn causes of subsequent effects, and so on in a continuous
series; every member in the series depends causally upon its predecessor. The notion is ubiquitous; and even where the metaphor is
not actually invoked, the idea that causation involves a sequence of
distinct entities standing in a certain relation to one another is
never far from the surface. 1
Both the metaphor and its associated conceptualization of causation are usually ascribed to the Stoics, and this has been the case
since antiquity. I shall argue that such an ascription is mistaken,
and if accepted encourages a wholly unwarranted and misleading
reading of the Stoics' actual views on the matter, which are (the
polemic of later Peripatetics such as Alexander notwithstanding)
very much closer to those of Aristotle than has been generally
1 Compare Hume's celebrated account: although Hume never employs the
chain metaphor, his definition of a cause as 'an object, followed by another,
and where all the objects similar to the first are followed by objects similar
to the second' (Enquiry Concerning Human Understanding VII 2; cf. Treatise I iii 15)
clearly conceptualizes causes and effects as distinct, discrete entities, something which the chain-metaphor at the very least encourages.

186

R.

J.

HANKINSON

allowed by scholars both ancient and modern. In fact the Stoics'

use of metaphor encourages a quite different picture of the metaphysics of causation, one which stresses the continuity of processes, rather than relations between discrete events. Finally I will
indicate some points of contact between the Stoic position, as I
understand it, and one particular modern anti-Humean account of
causation, that of Wesley Salmon.2
I

Aristotle's treatment of causation and explanation makes little or no

use of metaphor: but I wish to begin by sketching some particular
aspects of it for the purposes of subsequent comparison with the
Stoic account with which this paper is primarily concerned.
Aristotle's famous doctrine of the four causes is at once familiar
and frustratingly elusive. Much modern attention has been
focussed on the legitimacy of translating Aristotle's aitia as 'cause'; 3
I shall not here contribute (except perhaps indirectly) to that rather
sterile discussion, beyond noting that, for Aristotle, ai'ttat properly
so-called are intensional items (i.e. it matters how you pick them
out), while causal contexts in ordinary language at least (pace
Anscombe) 4 appear to be extensional (i.e. they tolerate the substitution within them of co-referring terms); on the other hand
Aristotle allows that the results of such substitutions within the
contexts of aitiai will yield what he calls 'incidental causes', a hint
Kn-ra <rUf.L~E~TJKO<; (Phys. II 3, 195a29-b 12). Thus, if a builder is the
efficient cause of a house's being built, and the builder happens to
be Callias, then Callias is the incidental efficient cause of the
building; to be an incidental cause is simply to be a cause picked

2 See in particular W. Salmon, 'Causality: production and propagation', in

P.D. Asquith and R.N. Giere (eds.) Proceedings of the 1980 Biennial Meeting of the
Philosophy of Science Association, 1980, 2.49-69 (repr. in E. Sosa and M. Tooley
(eds.) Causation (Oxford 1993) 154-77: my references are keyed to their pagination); 'Probabilistic causality', Pacific Philosophical Quarterly 61 (1980) 50-74
(repr. in E. Sosa and M. Tooley (eds.) Causation (Oxford 1993); and 'An "at-at"
theory of causal influence', Philosophy of Science 44 (1977) 215-24.
3 Cf. e.g. M. Hocutt, 'Aristotle's four becauses', Philosophy 49 (1974) 385-99;
G.R.G. Mure 'Cause and because in Aristotle', Philosophy 50 (1975) 356-7; R.
Sorabji, Necessity, Cause and Blame (London 1980) chs. 1-3.
4 G.E.M. Anscombe, 'Causality and extensionality', journal of Philosophy 46
(1969) 152-9; see on this issue J.L. Mackie, The Cement of the Universe (Oxford
1974) ch. 10.

CICERO'S ROPE

187

out under a description which does not perspicuously link the

cause to the effect.
This strongly suggests that Aristotle's principal concern is to
relate causes to their effects in such a way as to make the link
between them intelligible: it is because Callias is a builder (has
internalized the 'tEXVT) of building) that we may anticipate, under
the appropriate circumstances, his producing a house (cf. Metaph. H
7-9). Similarly, it is because Callias's father contained within
himself the human form (and, in virtue of being male, was able to
transmit it), that he produced Callias. Causing, at least of the
intelligible kind, is the transmission of form; aitiai are explanations: but they are causal explanations.
Occasionally things just occur or arise spontaneously, and in
these cases there is no communication of form from one thing to
another. Rather, the form simply emerges from the appropriate
matter in virtue of its inherent tendencies (although some things
never arise without the mediation of form-stones never spontaneously arrange themselves into a house: Metaph. H 9, 1034a9-21).
And while we might say (although perhaps strictly speaking we
should not do so) in the case of such accidentally emergent
structures that they exhibit form, they have not come to do so as a
result of any form-transmitting process.
This is, I think, the key to understanding an obscure passage of
Metaph. Z, where Aristotle is discussing various types of accidental
or coincidental occurrence:
That there are principles and causes which are generable and
destructible without their being for them any process of generation
and destruction is obvious; for otherwise everything will be of
necessity, since what is in the process of being generated or
destroyed must have a cause which is not accidentally its cause.
Will this occur or not? Yes, if such and such happens, but not if it
does not. And this will occur if something else does. And so if time
is constantly subtracted from a limited temporal extent, we will
clearly arrive at the present. This man, then will die by violence if
he goes out; and he will do this if he is thirsty; and he will be
thirsty if something else happens ... For instance he will go out if
thirsty, and he will be thirsty if he eats something spicy; and this is
either so or not; so he will either of necessity die or not die ....
Everything, then, that is to be will be of necessity; e.g. it is
necessary that someone who is alive must some time die, since
something has already occurred <sc. to necessitate it>, namely the
existence of contraries in the body. But whether he dies by disease
or by violence is not yet determined, but depends on the happening
of something else. Evidently then the process goes back to a
particular source but this does not itself indicate anything further

188

R.

J.

HANKINSON

<in the past>. This, then, will be the source of what happens by
chance, and will have nothing else as the cause of its coming to be.
(1: Metaph. Z 3, 1027a29-1027bl4)

This text has been much discussed, and there is little agreement as
to its import. My translation is tendentious (although not I hope
unwarrantedly so). What Aristotle is saying here (if I am right) is
not that some occurrences are uncaused or unnecessitated, 5 but
rather that some cannot be properly explained, where proper
explanation has to do, as we have seen, with the propagation (in
one way or another) of form. It is not that the hapless individual's
death is uncaused-it is rather that the causal chain (as we might
say) that leads from his predilection for hot foods via his
consequent thirst and his desire to slake it to his unfortunate
meeting with the ruffians at the well is not such as to instantiate, in
its totality, any generally-repeating process. 6 That is why the
causes in such chains come to be without any process of
generation: there is nothing, prior to the coincidence, which can
be said in any generally illuminating way to be leading up to it. In
this sense, nothing leads up to it.
The crucial point is that coincidences are intersections of processes: JCtvftcrw;. 7 Aristotle devotes most of the second half of the
Physics to an analysis of JCtvftcretc;: but for our purposes the important
thing about them is that they have beginnings, middles and (most
importantly) ends (Phys. V 1, 224a34-b16, 225a12-34). Not all
5 These do not (necessarily) come to the same thing: see Sorabji (op. cit.,
n. 3), chs. 1-3; I do not, however, agree with Sorabji's interpretation of this
passage. See D. Frede, 'Aristotle on the limits of determinism: accidental
causes in Metaphysics E 3', in A. Gotthelf (ed.) Aristotle on Nature and Living
Things (Bristol/New Brunswick) 1985, for an account with which I am more
in sympathy.
6 Compare Aristotle's other canonical example of coincidence: happening
to run into someone who owes you money and recovering your debt: Phys. II
5, 196b24-197a5; see W. Charlton, Aristotle's Physics I, II (Oxford 1970) 105-11,
for a useful discussion of this and related passages.
7 It is not clear that all such intersections are coincidences for Aristotle,
since some might be held to be predictable from within the working-out of a
particular process-indeed I am tempted to see the generation of substantial
form as having this structure. Here two processes (that in the male and that
in the female) intersect; but while the particular manner, time and place of
their intersection is not predictable from within either process, that there
will be such intersections presumably is; moreover, in the appropriate way,
the creation of a new substantial form is instantaneous, and not itself the subject of generation in time. This is speculative, and needs more working out
than I can give it here. However I note, proleptically, that if this is right,
then the Aristotelian and the Stoic accounts will be even more closely related.

CICERO'S ROPE

189

changes are Ktv-ftcretc;, however (Phys. V 2); in particular, instantaneous changes from one state to another do not qualifY (although
they may, as in the case of the generation of substances, occur as a
result of processes). Processes have an internal cohesion and a
natural direction. They may be as simple as the falling of earth to
its natural place 8 or as complex as those involved in animal
metabolism; but they share a crucial characteristic. Once begun,
their progress towards their ends is natural, intrinsic to the order of
things, and will (other things being equal) result in the process's
successful completion.
When the process fails to fulfil itself, that failure will be neither
uncaused nor utterly inexplicable; but it will be attributable to
factors external to the development of the form itself. Thus not all
acorns become saplings and not all saplings oaks; in some cases
the oak form will have been insufficiently firmly imprinted in the
original woody matter (for various reasons: Gen. Anim. IV 4-6), in
others external conditions will have been unpropitious (drought,
insufficient sunlight) or directly inimical (lightning, squirrels).
All of this suggests that for Aristotle it is the processes which are
basic. 9 Processes are not simply aggregates of discrete events,
describable in some suitable four-dimensional spatio-temporal
language 10 (indeed, I am inclined to think that the concept of an
event, at least as it is deployed in contemporary analytical philosophy, is quite alien to the ancients). And while processes evidently
have temporal stages which stand in relations of succession to one
another, on this metaphysical view there is no temptation to suppose with Hume that what causation amounts to is the mere
sequence of discrete events (and hence to make the causal relation
utterly mysterious), nor either to imagine that somehow each
such event is metaphysically dependent in some way upon its
predecessor. Equally, the difficulty that Russell felt to be fatal to the
entire concept of causation, namely that causally-significant events
8 The fact that Aristotle refers to such natural motions as involving the
perfection of form becomes perfectly intelligible on the interpretation being
offered: Cael. IV 3, 310a32-b1; cf. I 8, 276a23ff.
9 This may sound bizarre: of course on Aristotle's own account it is the
category of substance which is fundamental, and in terms of which
everything else must be explained (cf. e.g. Cat. 1-5; Metaph. H); but the upshot
of my suggestion is that Aristotelian substances (or most of them) may be
re'iarded as processes.
For a sophisticated modern defence of this philosophical position, see
D.Wiggins, Sameness and Substance (Oxford 1980), esp. 25-6.

190

R.

J.

HANKINSON

must be instantaneous and that hence between any two distinct

events conceived of as cause and effect a third potentially disruptive event may be intruded, 11 ceases (insofar as it ever did) to
trouble.
One final point bears noting. We are so inclined to suppose that
causes in general precede their effects that we tend to overlook the
fact that Aristotle hardly treats of the temporal succession of causes
and effects at all. Occasionally Aristotle will offer examples of the
aitia-relation where what is explanatory precedes what is to be
explained, most notably the example of efficient causation at A.Po.
II 11, 94a36-b8: the Persians made war on the Athenians because
the latter had attacked Sardis. But this is a rarity (moreover, the
example itself is unhappy, and badly fits Aristotle's programme in
the passage). 12
In A.Po. II 12, where he apparently deals with temporal considerations, Aristotle is concerned only with the appropriate tenses
to employ in different explanatory accounts-and indeed explicitly says (95a36) that the middle term (which does the explaining
by binding together subject and predicate in the explanandum) must
be coeval with what it explains.
Aristotle, then, for theoretical reasons rarely even suggests that
causes may precede their effects. Perhaps the nearest he gets to it is
text 1: but the whole point of that, if I am right, is to deny that, in
any strong sense, causes do precede their effects. To be sure, as we
would see it at least, earlier events are causally relevant to later
ones-but they are so for Aristotle because they are part of intelligible processes. When those processes intersect in unforeseen
ways there is created a new (instantaneous) cause, but that cause is
itself the cause of something else instantaneous and cotemporal
with it, namely the initiation of a new process. Finally consider
Wicksteed's translation of apxl, Ktvftcrero~. Aristotle's description of
the efficient cause, as 'initiating the process': 13 initiation sounds
like triggering, setting a train of events in sequence-but that is
quite alien to Aristotle's sense. Rather the apxl, is the source of the
11 B. Russell, 'On the notion of cause', Proceedings of the Aristotelian Society
13, 1912/13, 1-26 (repr. in Russell, Mysticism and Logic, New York 1918); see
the discussion in Mackie ( op. cit., n. 4), 143-54.
12 See W,D. Ross, Aristotle's Prior and Posterior Analytics (Oxford 1949) andJ.
Barnes, Aristotle's Posterior Analytics (Oxford, 19942) ad loc.
l3 In Aristotle: The Physics, Loeb Classical Library (Harvard 1920), in
particular at Phys. II 3, 194b29ff.

CICERO'S ROPE

191

process or change, what drives it; and such sources are coeval with
their effects.I 4

n
Metaphor is strikingly absent from Aristotle's treatment of these
issues; 15 not so from the Stoics'. For the Stoics, everything that
occurred in the world was the result of the working out of fate,
iJ,.UXpj.LEvT]:
Chrysippus described (in the second book of his On the Cosmos) the
substan~e (o~cria) of fate as t~e pnel!matic power I~ which goyerns
everythmg m orderly fashiOn (ta~Et tou navtoc; OtOlKTitlKTIV) ....
Elsewhere he describes it in several ways, as the logos of the
cosmos, the logos of the things which are governed by providence
in the cosmos, or the logos according to which things that have
occurred have occurred, things which are occurring are occurring
and things which will occur will occur. ( 2: Stobaeus I 79.1-12, = S VF
2.913, = 55M L&S)
That image of government, OtotlCT]crtc;, is found elsewhere in texts
reporting the Stoic position:
Fate is a continuous cause ( aitia dpOIJ.EVTl) of things, or the logos in
virtue of which the world is administered. (3: DL 7.149; cf. 8 below)
But more important for our purposes is the metaphor of the continuous thread, the a.itia dpoj.LEVT] (cf. the /..6yoc; dp6j.LVoc; of Arius
Didymus Fr. 29, = SW 2.528). Hicks in the Loeb translates: 'Fate is
defined as an endless chain of causation, whereby things are, or
as the reason or formula by which the world goes on', which
seems to me to be multiply misleading. Nothing in the Greek
corresponds to Hicks' 'endless'; 17 'whereby things are' is a needless
14 This is partly what causes Aristotle's notorious difficulties with projectile motion: if the spear continues to move after it leaves contact with the
thrower's hand, something must still be pushing it (Phys. VIII 10, 266b29267all).
15 Unless you count as metaphoric the apxi] in his standard expression for
the efficient cause, UPXTJ lClVlt0'(1)~.
16 I.e. the power manifested in the pneuma, the all-pervasive dynamic
constituent of the universe which is responsible for its cohesion: cf. e.g. D.L.
7.138-9 (= SW 2.634, 470 LS); Cicero, ND 2 22 (= 47C LS); Galen, Plen. VII 524
Kuhn (= SW 2.439, 47F L&S); Plutarch, Comm. Not. 1085c-d (= SW 2.444, 47G
L&S); see also 47 1-M L&S; and see further below.
17 Although the Stoics do think that the sequence of causes is endless (see
n. 18, and text 10, below): Alexander fastens upon this aspect of their theory
in order to argue that such an infinite sequence of causes would render the
cause (or explanation) unknowable, since knowledge just is knowledge of the
primary cause: Fat. 25, 196.1ff Bruns,= SW 2.949. Alexander here confuses the

192

R.

J.

HANKINSON

over-translation; but most serious is the intrusion of the metaphor of

the chain.
The word dpoJlEVTJ is the passive participle of dpnv, to string
together, as of necklaces (cf. Od. 15.460, 18.296); in grammar and
rhetoric, the AE~t~ eipoJlEvTJ is a continuous or running style, as
opposed to the classical periodic construction of balanced, antithetical clauses ( cf. Aristotle Rhetoric III 9, 1409a24ff.). The metaphorical
emphasis here is on the continuity of the process, not the discrete
identity of its parts. IS Cognate with Etpro is the noun EipJlO~, defined
in LSJ as 'train, series, sequence'; and the Stoics regularly
described Fate as an EipJlO~ of causes, seeking to discern an
etymological link between EipJlO~ and EiJl<XPJlEVTJ (cf. Chrysippus,
reported by Diogenianus: SW2.914).
The EipJ.16~ metaphor is undoubtedly Stoic: it figures in all the
major doxographical accounts of the Stoic position, and is probably
original to the school (at all events, I can find no earlier traces of
it). Aetius remarks that
the Stoics say that fate is an EipJ.lO<; of causes, i.e. an ineluctable
ordering ('tal;t<;) and connection ( E1ttcrUVOEcrt<;). ( 4: 1 28.4 Diels, =
SW2.917)
Nemesius objects that:
if fate is a sort of ineluctable eipJ.lO<; of causes, as the Stoics define it
(i.e. an unalterable 'ta~t<; and E1ttcrUV0Ecrt<;), ends will come about not
for the sake of any advantage, but simply in accordance with their
own motion and necessity. (5: Nat. Hom. 37, 299, = SW2.918)
And Alexander notes:
still, they [sc. the Stoics] agree that all things that occur in
accordance with fate do so in accordance with order ('ta~t<;) and
consequence ( aKOAou8ia), and have within themselves some
sequentiality. (6: An. Mant. 25, 185.1-3 Bruns,= SVF 2.920; cf. Alex.
Q]laest. 14)
But, it may be objected, an EipJlO~ is a string of beads (cf, the
passages from the Odyssey cited above): surely the weight of the
metaphor might equally well fall upon the individual items which
are strung together? And if that is the case, then the metaphor of
infinity of temporal sequence with an infinity of metaphysical ordering:
only if the Stoics were committed to the latter (which they are not) would
knowledge, even on this Peripatetic account, be threatened.
!8 Cf. Critolaus, ap. Philo. Aet. 70 (= SVF 2.913): 'since according to the best
students of nature fate is without beginning or end, stringing together the
causes of everything without cessation or intermission'.

CICERO'S ROPE

193

the chain is perfectly appropriate. The Stoics will, after all, be seeking to emphasize the fact that causal sequences are sequences of
distinct events, like the beads on a necklace, somehow joined
together.
I think that this conclusion can and should be resisted. It is
worth reiterating the connection of eipJ.lO<; with etpro and etpOJ.iat: it
is a verbal noun, capturing the process of the stringing together.
More significantly, however, consider the other expressions that
are used alongside eipJ.lO<; ainrov, presumably by way of exegesis.
In the texts cited above (SW 2.917-20), we have ta~t<;, (mcruv8ecrt<;,
and UKOAou8ia. Gellius reports that:
in the fourth book of his On Providence, Chrysippus says that 'fate is a
certain natural, eternal ordering of the whole, in which things are
consequent upon and follow one another, and in which the interweaving is ineluctable'. (7: Noct. Att. 7.2.3, = SW2.1000, = 55KL&S).
Plotinus writes that those who believe in:
the interweaving ( E7tt7tAOKtl) of causes with one another, the etpiJ.~
from above, and the invariable consequence of later occurrences
upon earlier ones ... evidently introduce fate in another manner [sc.
from that of the astrologers]. 19 (8: Enn. 3.1.2, = SW 2.946; cf. ibid.
3.1.4, = SW2.934)
In the course of elaborating an objection (unjustified as it will
turn out) to the Stoic account, namely that it makes every prior
event a cause of every subsequent one, absurdly (since thereby day
would be the cause of night, and summer of winter: Fat. 25, 194.
25ff., 195.13ff. = SW2.948; 22 below), Alexander speaks of them as:
positing a certain binding together ( em<rUvBmt<;) and cohesion
(cruVEXEta) of causes, and holding this to be the reason why nothing
comes to be causelessly. (9: Fat. 25, 195.4-5 Bruns,= SVF2.948).

A page later, he asks:

how could it not be absurd for them to say that causes are infinite,
and that the Ei.p~to<; and E7ttcruv<5Ecrt<; of them has no first or last
19 Plotinus's account is in many ways puzzling. He apparently contrasts
with astrological determinists two groups of people who hold similar views
about the interconnectedness and ineluctability of fate; and then (at the end
of Enn. 3 1 2) he remarks that the second of these two groups should be further
subdivided between 'those who make everything dependent upon a single
principle' and those who do not (cf. ibid. 4-7). It is far from clear where, if
anywhere, orthodox Stoicism is supposed to fit into this picture. But by
Plotinus's syncretist age, original orthodoxies had become watered down,
and distinctions blurred; there is, at any event, little doubt that the language
that Plotinus adopts here originates with the Stoics.

194

R.

J.

HANKINSON

member? (10: ibid. 196.1 Bruns,= SVF 2.949; see 14 below; and nn.
17, 18 above)
'Binding together', 'cohesion' and 'interweaving', along with
'ordering' and 'consequence', clearly do not emphasize the supposedly discrete nature of each cause in the sequence.
Latin texts reporting or reflecting Stoic doctrine support this.
Cicero writes:
I call 'fate' what the Greeks call EiJla.pJlEVT]: and ordering and
sequence of causes ( ordinem seriemque causarum), since cause being
connected to cause brings things to be out of itself; it is an everlasting truth which flows from all eternity. (11: Cicero, Div. 1 125, =
SW2.921, = 55L L&S; cf. Servius ad Verg. Aen. 3.376, = SW2.920)
And Augustine:
'Those (i.e. the Stoics) who call by the name of 'fate' not the
arrangement of the stars but rather the connection and sequence of
causes ( conexionem seriemque causarum) .... attribute the ordering of
causes itself and that particular connection to the will and power of
God the supreme'. (12: m 5.8, = SW2.932)
Moreover, this emphasis is buttressed by considerations deriving from the Stoics' analysis of the causal relation.2 Sextus
Empiricus reports that
the Stoics say that every cause is a body which becomes a cause to a
body of something incorporeal. For instance the scalpel, a body,
becomes the cause to the flesh, a body, of the incorporeal predicate
'being cut'. ( 13: Adv. Math. 9.211, = SW2.341, = 55B L&S; cf. SW 1.89,
2.336)
Causes are bodies, effects are predicates (or rather attributes). Thus
it cannot be the case in any straightforward manner that the effect
of a previous cause may itself in turn be the cause of a subsequent
effect.2 1 Causation is not, then, to be understood on the model of a
sequence of events, each of which brings about its successor, as in a
row of falling dominoes.22 That may be how opponents such as
20 On the Stoic analysis of causation, see M. Frede, 'The original notion of
cause' in J. Barnes et aL (eds.) Doubt and Dogmatism (Oxford 1980); see also
J. Barnes, 'Skepticism and causation', in M.F. Burnyeat (ed.) The Skeptical Tradition (California 1983); and RJ. Hankinson, The Ancient Concept of Explanation,
Oxford, forthcoming, ch. 7; and 'Causation and explanation' and 'Determinism and indeterminism, both in J. Barnes, J. Mansfeld, M. Schofield
(eds.) Cambridge History of Hellenistic Philosophy (Cambridge, forthcoming).
2l As noted by F.H. Sandbach, The Stoics (Bristol 19892) 81-2; I disagree
with Sandbach, however, in his characterization of Stoic effects as events.
22 Although there are ancient analogues to such a notion: cf. Galen, On the
Differences of Symptoms VII 47-9 Kiihn. In such sequences, Galen stipulates that

CICERO'S ROPE

195

Alexander seek to paint their doctrine in order to refute it (day

causing night, summer winter, and so on: SW 2.948); but it is not
how the Stoics themselves saw matters.
At this point we may examine Alexander's treatment of the
issue in a little more detail:
they [sc. the Stoics] say that this cosmos, which is a unity encompassing everything within itself and which is governed by a
living, rational and intelligent nature, has an eternal government
of things in accordance with a certain prior Ei.pJloc; and tal;tc;, with
the prior things being the causes of those which occur after them.
In this manner all of them are bound together with one another,
and nor does anything occur in it without something else following from and being linked to it causally. Nor can any one of the
subsequent events be detached from those which precede it so as not
to be consequent upon and bound to one of them; but rather
everything which occurs is succeeded by something else which is
necessarily causally dependent upon it, and everything which
occurs has something prior to it upon which it is causally dependent. For nothing in the cosmos can either exist or come to be
without a cause since nothing within it is detached or separated
from everything that has come before. For if any uncaused process
were introduced, the cosmos would be torn apart and divided and
would no longer remain a unity, forever governed according to
one ordering and administration; and such an uncaused process
would be introduced if everything which exists an~ comes to be
were not to have preceding causes which they followed of necessity. They say that this would be similar to something's coming to
be out of nothing without a cause, and similarly impossible; and
being of such a kind, the government of the totality goes on
actively23 and without cessation from infinity to infinity. (14: Fat.
22, 191.30-192.17 Bruns,= SW2.945, = 55N L&S)
That is Alexander's report of Stoic doctrine: and it is by no means
entirely unfair. For the Stoics, a cosmos is a totality of totally
interrelated items, no one of which can be affected without some
effect upon all of the others: everything is linked by cr'UJlrta9na,
joint affection, which accounts for the unification of the whole
(Cleomedes 1 1, = SW 2.534, 546), for which the Stoics adduced as
evidence the dependence of the tides and of the behaviour of
animals on the moon, and which formed the cornerstone of their
theological cosmology (Sextus Ad. Math. 9.78-85, = SW2.1013).

each item causes its immediate successor 'primarily and essentially', its more
remote consequents 'secondarily and incidentally', 'thirdly and incidentally', and so on.
23 Reading evEpy&c; with Usener and von Arnim, against the MSS.
reading vapy&c; preferred by Long and Sedley.

196

R.

J.

HANKINSON

The universe as a whole is a rational animal, all of whose parts

are responsive to its overall will, which is the source of its OtohcliCH~
and is identified with God (Cicero, ND 1.39, = SVF 2.1077, = 54B
L&S; DL 9.148, = SVF 2.1022, = 43A L&S). Indeed, God is identical
with intelligence and fate (ibid. 135. = SVF 1.102, = 46B LS), and is
even identified with the cosmos itself (ibid. 137, SVF 2.526, cf. 5278). God's rule of the universe is accomplished by his pervading it
through and through in the form of pneuma, the subtle, dynamic
mixture of fire and air responsible for intelligence and sensation
(Alexander, Mixt. 216.14ff., = SW2.473, = 48C L&S; 224.14ff., = SVF
2.442, = 471 L&S; ibid. 226.34ff., = SW2.475; DL 7 151, = SW2.479, =
48A L&S; cf. n. 16 above). This pneuma is also responsible for the
very cohesion of physical objects (ibid. 223.25ff., = SVF 2.441, = 47L
L&S).24
We now need to see what implications this physics has for the
Stoics' overall view of causation. That the Stoics introduced new
categories into causal analysis is undisputed, although quite which
categories (and what they were intended to do) is a matter of much
less agreement. But it is reasonably clear at all events that the Stoics
distinguished between containing (or sustaining) causes (at na
cruveKnKa), causes which were coeval with their effects and necessary and sufficient for them, and antecedent causes {atna
npoKa'tapK'ttlca). Sextus almost certainly has the Stoics in mind
when he attributes the following account to an unnamed group of
Dogmatists:
causes are containing ( cruvEKnKa) if, when they are present the
effect is present, when they are removed the effect is removed, and
when they are decreased the effect is decreased (thus they say that
the application of the noose is the cause of the strangling); ... and
some of them have said that things present can be causes of things
future as well as their antecedents (7tpoKa.'ta.pKnKa): for instance
protracted exposure to the sun offever. ( 15: PH 3 15-16)
24 Something Galen at any rate thinks requires no explanation, solidity
simply being the property of being cohesive (On Containing Causes 6.3); pace
Galen, there is nothing absurd in treating solidity as a derived property,
dependent upon the presence of continuously active pneuma in sufficient concentrations in the appropriate passive material. Nor does Galen succeed in
showing that the doctrine is viciously regressive on the grounds that pneuma
is, for the materialist Stoics, itself a physical stuff (ibid. 6.4; cf. SVF 2.439-42).
These issues are more fully explored in my opp. citt., n. 20 above. For the
Stoics' division of the material of the world into active and passive components, associated with fire and air and earth and water respectively, see

SVF 2.524, 299, 300, 311, = 44A-C L&S.

CICERO'S ROPE

197

This distinction is elaborated by Clement:

when antecedent causes are removed the effect remains, whereas a
containing cause is one during whose presence the effect remains
and on whose removal the effect is removed. ( 16: Str. 8 9 33, = SVF
2.351, = 551 L&S)
Atna aUVEK'tuca are no longer simply the immanent causes of persistence and coherence:25 they are the sustainers of processes as
well.
But evidently if there is to be a succession of causes in any way
at all, not all causes can be atna aUVEK'ttJCcl on this model. That the
Stoics did allow that, in some sense, causation embodied consecutive relations, is not in doubt. The question is: what form do
those relations take?
Alexander, in 14 and elsewhere, seeks to foist upon them the
unpalatable conclusion that every precedent event is the cause of
every subsequent one; but there is no reason to suppose that the
Stoics accepted this consequence, nor that they were in any way
committed to it. Similar (although less sweeping) concerns are
voiced by Cicero:
'cause' should not be so understood as to make whatever precedes a
thing its cause, but what precedes it actively: the cause of my
playing ball was not my going down to the campus, nor did
Hecuba's bearing Alexander make her the cause of the death of
Trojans, nor was Tyndareus the cause of Agamemnon's death
because he was Clytemnestra's father. ( 17: Fat. 34).
Mere prerequisites should not be confused with causes as such. But
he goes on to note that the Stoics:
say that there is a difference between whether a thing is of such a
kind that something cannot be brought about without it, or such that
something must necessarily be brought about by it. None of the
causes mentioned therefore is really a cause, since none by its own
force brings about that of which it is said to be the cause; nor is that
which is a condition of a thing's being brought about a cause, but
that which is such that when it is present that of which it is the
cause necessarily is brought about. (18: ibid. 36, = SW 2.987)
Finally, we may quote Cicero's report of Chrysippus's own causal
distinctions, deeply puzzling though it is:26
'some causes ', he says, 'are perfect and principal, others are
auxiliary and proximate. Hence when we say that everything takes
25 Which Galen says was the original Stoic sense: Adv. Jul. XVIIIA 278-9
Kiihn, : SW 2.355; cf. idem Syn. Puls. IX 458 Kiihn, : SW 2.356.
26 I offer an interpretation of it in my op. cit. (n. 20), ch. 7, ld.

198

R.

J.

HANKINSON

place by fate from antecedent causes, we should not be taken to

mean by perfect and principal causes, but by auxiliary and proximate causes'. Accordingly he counters the argument which I have
just set out as follows: 'if everything comes about by fate it does
follow that everything comes about from prior causes: not, however,
from principal and perfect but from auxiliary and proximate
causes'. (19: ib. 41, = SW2.974, = 62C L&S)
The 'auxiliary and proximate causes' presumably at any rate
include antecedent causes.27
Elsewhere Chrysippus is explicitly said to have identified fate
with a sequence of antecedent causes. Plutarch objects that if Chrysippus allows that antecedent (rather than perfect or containing)
causes are the source of fate, and that is identical with God, God
will not be all-powerful, since antecedent causes are weaker than
their perfect cousins (Stoic. Rep. 47, 1056b-c, = SlP2.997; cf. ps.-Plut.
Fat. 11, 574d, = SW 2.912; Cicero Top. 59). But Plutarch's objection
can be blunted in the following manner. Consider Chrysippus's
celebrated example of the distinction he has in mind, reported by
both Cicero (Fat. 42-3, = SVF 2.974, = 62C L&S) and by Gellius
(Noct.Att. 7 2 6-13, = SlP2.1000, = 620 L&S): once it has been set in
motion on a smooth surface, a cylinder will continue rolling for a
while. Its beginning to roll is caused by something external to it,
an antecedent cause which predates the rolling (and ceases to exercise its influence before the cessation of the rolling: cf. Clement,
Str. 8.9.33, quoted above). However it continues to roll, according to
Cicero, suapte viet natura, as a result of its own force and nature, or,
in Gellius's version, quoniam ita sese modus eius et formae volubilitas
habet, because its form and rollable shape are thus. No cylinder
will start rolling simply in virtue of its own rollability-yet, once it
is rolling it is its volubilitas (in a state of actuality) which accounts
for its continuing to roll, a persistent, immanent cause which lasts
as long as the rolling: i.e., an atnov crUVEK'ttlCOV.
The Greeks had nothing comparable to the modern concept of
inertia. Things don't just go on and on until something stops them:
if something is happening, something must be making it happen.
There is thus no radical distinction in type between what we
would think of as temporally-extended events (such as the cylinder's rolling) and genuine processes, such as the maturation of an
oak tree. If this is correct, first impressions notwithstanding, the
27 The situation is complicated, however, and controversial: compare the
different accounts offered in the works cited in n. 20 above.

CICERO'S ROPE

199

Stoic account has much in common with that of Aristotle sketched

earlier.
That impression is strengthened by further consideration of the
role and nature of the antecedent cause. By contrast with containing causes (which are internal to their objects and identified at
least partially with their nature or form),28 antecedent causes are
external to the objects affected, and are responsible for the actualization of some particular propensity they have. Thus the irruption of
an antecedent cause into the tranquil life of the object, disturbing it
and rousing it into action, is not something predictable or
explicable from within the object itself on the basis purely of an
understanding of its intrinsic properties. That a cylinder has a
tendency to roll is not accidental to it-that it starts to do so at 2.30
on Thursday afternoon in response to a kick given it by a passing
Stoic is. For all that, sub specie aetemitatis and the total structure of the
world, that kick is not undetermined or accidental-indeed, it is
only to be expected.
That, ultimately, is why Chrysippus wants to identify fate with
antecedent causes: from the point of view of understanding why
certain processes are initiated at certain times, we need to understand their antecedent causes-or, more precisely, we need to
know why and when one particular process embodied in an object
impinges upon another object, setting a second process in train.
The internal propensities of things to act, considered purely in and
of themselves, are inert,29 and cannot account for the dynamic
structure of things. Not that antecedent causes are against nature:
indeed, as part of the overall harmonious totality of things which is
the universe, they could not be. But the Stoics distinguish between
two senses of 'nature':
sometimes by 'nature' they mean that which holds together
(cruv[xoucrav) the cosmos, sometimes that which is reseonsible for
the growth of things on earth. Nature is a constitution ( E~t<;} which
moves of itself, and brings to completion and sustains (cruvexoucra)
what results from it in accordance with the seminal principles in a
determinate time, and produces things of the same type as those
from which they came. (20: DL 7.148, = SW2.1022, = 43A L&S)
28 Frede (art. cit., n. 2(} above, 243) is thus clearly right to describe the
o:tttov uuvK'ttK6v as 'the Stoic analogue to Aristotle's formal cause'.
29 Strictly speaking, this is false: the internal propensities are E~Et~ of the
objects, which are themselves to be identified with the structure of the pneuma
within them, and that pneuma is intrinsically dynamic. But this sort of
dynamism is of a different type to that exhibited by macroscopic objects in
action, even though those actions too are ultimately traceable to pneuma.

200

R.

J.

HANKINSON

This passage resonates with unmistakable Aristotelian echoes:

compare his definition of nature as 'an internal principle of
motion'30 (Phys. II 1, 192b8-193a10); and note the conception of
generation as the transfer of form embodied in the final clauses.
On the other hand, it also illustrates the Stoics' un-Aristotelian view
of the perfect harmony of the whole, which is, in fact, a single
organism. Each natural individual in the world has its own nature
(its own structure and set of functions, which are in fact its
containing causes), just as do the parts of a living body; but the
overall goal of nature as a whole is to work towards their complete,
harmonious interaction. Entirely congruently, for the Stoics,
moral progress consists in the ever-closer assimilation of the
individual's nature and will to the nature and will of the whole, in
order to secure the eupota ~iou, the smooth flow of life which is
human happiness (SVF 2.184, 554; 3.4, 11), and which is what, in
the long run, is accomplished by ineluctable fate:
Chrysippus the Stoic said that there was no difference between
what was necessitated and fate, saying that fate was an eternal,
continuous, and ordered movement. (21: Theodoretus GAG 6.14, =

SW2.916)

III

What is the upshot of all of this? First of all, I have been trying to
bring out the Stoics' own emphasis on the essential continuity of
the processes involved. Where the working out of one process abuts
spatia-temporally against a suitably-disposed body, another process
is set in motion: and here the one (or rather the object in which the
process is embodied) may be said to be the cause to the other of the
new process (entirely congruently with the canonical form of
Stoic causal sentences: see above, p. 194). Thus the Stoics are by no
means committed to supposing, as Alexander would have them
absurdly do, that every temporally precedent stage of every process
is the cause of every subsequent stage,3I although for the Stoics any
particular stage of a process will in a sense be causally relevant to
others, since they will, other things being equal, be inferable from
30 Or 'change', or 'process':

KtVTJC1t<;.

I say 'absurdly': Alexander evidently considers it absurd, and it seems

to me at least to do violence to our ordinary notions of causing; yet some
(notably Mackie, op. cit., n. 4 above, ch. 8) at least entertain the thesis. At all
events, it is irrelevant to the Stoics.
31

CICERO'S ROPE

201

it. Moreover, they reject the Aristotelian orthodoxy that properly so

called no cause may precede its effect: for them, in contrast with
the Peripatetics (if my account above is correct), antecedent causes
are causes not just of the initiations of processes, but of the processes
themselves and of their outcomes.
Nor are the Stoics to be saddled with a picture in which causes
and effects are essentially discrete events, on the Humean model,
whose relations are consequently obscure. Containing causes,
responsible for the continuous persistence and evolution of things
and processes, are themselves continuous. Antecedent causes do
indeed precede their effects, but are themselves parts of other
processes. Alexander claims that the Stoic doctrine forces them to
suppose that winter causes summer (Fat. 25, 194.30ff., = SVF 2.948;
22 below) ;32 but they are under no compulsion to suppose any such
thing. The orderly progression of the seasons (as well as the
alternation of day and night) is simply a continuous, repeating
process. In fact, the Stoics will explain it in much the same way as
Alexander does, as 'being caused by the motion and rotation of the
divine body, and its inclination in virtue of the ecliptic, in accordance with which the motion of the sun is equally responsible for
all of the aforementioned phenomena' (ibid. 195.11-13): DL.7 151-2,
= SVF2.693 (cf. SW2.694-6).33
Alexander's attempt to foist an uncongenial picture of the causal
structure of things upon the Stoics is, however, underlined by his
own use of a particular metaphor-which takes us back to where
we started. Twice he characterizes the Stoic view, in which (so he
alleges) earlier events invariably bring about later ones, as involving events being related to one another ot~v aJ.:uoero<;, in the manner of a chain. Here unequivocally (by contrast with the case of
eipf..16<;, investigated above), we have our time honoured metaphor.
32 My colleague Steve White suggests that this example, as well as that of
night causing day, might have seemed particularly ben trovato for the Peripatetics, with their insistence that opposites could not be the cause of opposites
(a widely-held ancient principle: cf. Plato, Phaedo 101 b); however Alexander
also cites cases which do not involve opposites (walking being caused by
standing up, the Isthmian games by the Olympian: Fat. 25, 194.29ff.).
33 There is, however, a difference of emphasis: for the Peripatetics, the
unity of the cosmos is preserved by the regular heavenly rotation, rather
than, as in the Stoic account, by the interpenetration of intelligent pneuma:
Alexander, Mixt. 10, 223.9ff., 11, 226.24ff., = SW 2.441, 1048 (cf. SW 2.470-5);
moreover, the Stoics reject the Peripatetic postulation of a divine fifth
element, the ether, of which the heavenly bodies are made.

202

R.

J.

HANKINSON

aA:uat<; is not a very common word: but it clearly means a chain in

the literal sense of a set of interlocking links.3 4 Alexander first
refers to

their first thesis [cf. 14 above], namely that of all the things that are,
some become causes of things after them, 35 and things depend
upon one another in this way, namely with the second depending
upon the first in the manner of a chain ( OtKi)v aJ.:uaE~), which is
what they take to be the essence of fate ... ( 22: Fat. 23, 193.4-8, = SVF
2.945)
The second passage occurs immediately after his explanation of
the true causes of summer and winter:
nor is it the case, because day is not the cause of night and nor is
winter of summer, and these things are not interwoven ( EJl7tE7tAEKtat) with one another in the manner of a chain ( aA:uaEroc; OtKi)v ),
that they would come to be without a cause,3 6 nor that if they did
not come to be in this way the unity of the cosmos and of
everything which exists and comes to be within it, would be torn
apart. (23: ibid. 25, 195.13-16, = SW2.948; cf. 14 above)
What are we to make of this? Even if the chain-metaphor as it is
34 Cf. Hdt. 9 74 (the Athenian hero Sophanes secures himself to the
ground in battle with an anchor tied to his corslet by means of a chain);
Polybius 3 82 8, 32 3 6, of chains used for securing captives: note the idiom eis
halusin agein, to throw into chains: ibid. 4.76.5, 20.10.7-8, 21.5.3.
35 Both text and translation here are problematic. Sharples (Alexander of
Aphrodisias on Fate, London, 1983, 71) accepts Bruns' text (ro~ 1t<Xvtmv toJV ovtmv
ait{mv ttvmv ytvollevmv t&v !lEta tauta) and translates 'that all the things that
are become causes of some of the things after them' (i.e. everything that exists
comes to be a cause for at least something subsequent to it); but this seems to
me to be an impossible construal of the Greek. On the other hand, my
translation, while a good deal less tendentious as an interpretation of Stoic
doctrine, does not seem capable of doing the work Alexander requires of it
(his subsequent counterexamples consist of cas.es where something fails to be
the cause of anything subsequent, cases which would not be ruled out by my
translation of the thesis here). Sharples is no doubt influenced by considerations of what Alexander needs to say for his argument (and text 8, to which
Alexander refers here, certainly suggests, although it does not explicitly
state, that the thesis in question is that everything is a cause for some subsequent event). There is an alternative textual tradition (ro~ t&v t&v ovtmv aitimv,
etc.) which would encourage the translation 'of the causes of things, some
occur of things subsequent to them', i.e. some causes precede their effects,
would be equally unexceptionable as an account of Stoic doctrine, but equally
incapable of bearing the weight of Alexander's argument here. Alexander
may perhaps have intended first to argue, on general Peripatetic principles,
against the thesis that any cause can precede its effect, before turning to attack
the stronger claim that everything has some subsequent effect.
36 The text is once again problematic, but the difficulties are not pertinent
to our purposes. See Sharples (op. cit. n. 32) for useful comments ad Zoe., as
elsewhere.

CICERO'S ROPE

203

deployed here does not carry any particular Humean connotations

of the discreteness of the individual events in the chain, it clearly
suggests (and is intended to suggest) a transitivity in the transmission of causal power from one item to the next, in the manner
of falling dominoes (above, p. 194), which is equally inimical to
the picture I have been trying to establish on the Stoics' behalf.
It should be repeated that aA.ucrt~ is not a common word; and it is
very rare indeed in philosophical contexts.37 I know of no text outside Fat. which deploys it as a causal metaphor at all, much less of
any which ascribes it to the Stoics, and for this reason I am tempted
to suppose that it is Alexander's own contribution (or possibly
borrowed from earlier polemical exchanges, now lost, involving
anti-Stoic philosophers, perhaps Academics). If this is right, the
image is not originally Stoic at all, and rather derives from the unsympathetic and unwarranted interpretations of their intellectual
opponents of the certainly authentically-Stoic EipJ.LO~ metaphor.
There is to my knowledge only one text that tells against this
supposition:
fate, which the Greeks call EtJ.Ulp~TJ, was defined by Chrysippus,
the chief Stoic philosopher, in roughly the following terms: 'fate',
he said, 'is a certain everlasting and unalterable sequence (series)
and chain (catena) of things turning around and interweaving itself
through its eternal orderings of consequence, out of which it is
fixed and interconnected'. These are Chrysippus's actual words, if
my memory serves me aright ... (24: Aulus Gellius, Noct. Att. 7.2.1-2,
= SVF 2.1000)
Note that Gellius is quoting from memory here, which he admits
is fallible. Moreover, he is quoting in Latin: he has already made a
mental translation of Chrysippus's original Greek, in contrast with
his practice elsewhere of quoting the Greek directly (as he does a
few lines later: 7 above). I suggest that catena here is his infelicitous
rendering ofdpJ.LO~ (compare SVF2.916-21, quoted above); and that
this constitutes another route by which the image of the chain,
with its unwelcome suggestions of the transitive metaphysical
dependence of later links upon earlier, as well as of their discreteness, found its way into the tradition.
37 I have come across only three instances in all of the Aristotelian
commentators, and although they have metaphorical significance, none has
anything to do with causation. They are Simplicius, in Phys. 10 877.10, 878.23
(discussing various different ways in which things can be joined together);
and Alexander, in Meteor. 3.2 217.7, discussing types of chemical combination.

204

R.

J.

HANKINSON

IV
I hope, then, to have demonstrated that the Stoics did not adopt the
chain as a metaphor for continuing causal sequence and dependence: and for good reasons. John Venn sought an alternative
metaphor, that of the rope. If I am right, this is, in effect, very
much the force of the Stoic eipJloc;, properly interpreted. There is,
however, one text which explicitly mentions ropes:
it is not the case that the things which are going to be simply
spring into existence; rather the passage of time is like the unwinding of a rope ( rudentis explicatio), 38 producing nothing new, but
merely unrolling each thing for the first time. ( 25: Cicero, Div. 1
127, = sw 2.944, = 550 L&S)
The unfolding of time simply brings to light what is already
implicit in the structure of things, a picture reinforced by a passage
a page or so earlier:
since this is the case, nothing occurs which was not going to be,
and in the same way nothing is going to be for which nature does
not already contain causes productive of it. (26: ibid. 1.125-6, = SVF
2.921 = 55L L&S: follows 11 above)
Wesley Salmon, who quotes Venn as the envoi to his article ('Causality', art.cit., n. 2 above), might as felicitously have cited Cicero.
In this article he seeks to undo what he takes to be the damage
done to our philosophical accounts of causation by an uncritical
acceptance of the Humean model. Causation, in his view ('Causality' 155-6), primarily involves two types of process, production and
propagation. Production involves the creation of something new,
propagation the transmission of something already existing. Moreover, production consists in the interaction of processes, 'the means
by which structure and order are propagated or transmitted from
one space-time region of the universe to other times and places'
(169). For Salmon, processes, not events, are ontologically fundamental (170-1).
There is, of course, much in Salmon's position which has nothing whatever to do with the Stoics. Salmon is expressly concerned
with developing an account that will apply to probabilistic causation, 39 a concept without parallel in the ancient world, not least
Intriguingly, von Arnim's index to SVF reads: rudens: = O:A.\}(nc;, repeatGellius's mistake.
9 Cf. the articles cited in n. 2 above.

38
in~

CICERO'S ROPE

205

because the elaboration of any such notion with any degree of

rigour requires an understanding of statistics, something notoriously beyond the ancients' competence. 40 But if I am right there are
structural similarities between the two views that are more than
merely adventitious, and which are instructive in a number of
ways. Salmon's interactive 'productions' are more than merely
superficially reminiscent of the Stoics' antecedent causes;4 1 while
his 'propagations' bear more than passing comparison with the
sorts of process which for the Stoics exhibit containing causation.42
The contributors to this volume were asked to write on some
subject intersecting with the philosophical interests of Jaap Mansfeld. Given the extraordinary catholicity of those interests, such an
injunction was, as the editors themselves pointed out, not a hard
one with which to comply. They did, however, suggest that we
might focus our attention 'on historical aspects of the study of
ancient philosophy', aspects which Jaap has done so much to
illuminate. I have sought to fulfil that brief; and I hope that my
attempts to draw illuminating comparisons between ancient and
modern views will not appear to have adulterated the pure doxographical spirit so clearly distilled by Jaap and his colleagues in
Utrecht.
Thus I do not wish to suggest, absurdly, that all causal wisdom is
owed ultimately to the Stoics. But a crucial aspect of their fundamental metaphysics of the matter seems to me, as presumably it
would also to Salmon and others of his ilk, fundamentally on the
right lines, in contrast with what has been for the past couple of
centuries the dominant account in the Anglo-Saxon philosophical
world (which is in any case the only philosophical world which
takes such things seriously). In philosophy, no less than in the wider world, one who is ignorant of history is doomed to repeat it. 43
University of Texas at Austin

40 See I. Hacking, The Emergence of Probability (Cambridge 1975) for an

excellent historical survey of the concept.
41 As they are also, if I am right, of Aristotle's coincidences, and perhaps
also of his account of other types of generation: see n. 7 above.
4 2 And they too bear comparison with Aristotle's account of the transmission of form.
43 I should like to thank Jennifer Hankinson for helpful comments of
both a strategic and a stylistic nature.

ARISTIPP UND SEINE ANHANGER IN ROM

CARL JoACHIM CLASSEN

Die Letzten, die sich zur Lebensweise des Aristipp und den Lehren
seiner Anhfmger bekannteni, waren-etwa in der Zeit des ersten
Ptolemaerkonigs-Hegesias, Annikeris und Theodoros und eine
Generation spater deren Schuler, die uns fast ausnahmslos nicht
mehr namentlich bekannt sind2. Seit der Mitte des dritten
Jahrhunderts v. Chr. gibt es in den uns erhaltenen Quellen keine
Hinweise mehr auf eine weitere, jungere Generation von
Kyrenaikern, und bezeichnenderweise findet schon kurz vorher
der Spott uber sie mit Alexis sein Ende3. Unter den ubrigen Dichtern erinnert nur der gelehrte Elegiker Hermesianax an Aristipp
und zwar an dessen Liebe zu Lais (frg. IV A 94 Giannantoni = frg.
2, 95-98 Diehl)' wahrend Timon Aristipp und dessen 'tpU<j>Epi] <jlucrt~
ebensowenig verschont wie andere Philosophen (frg. IV A 51
Giannantoni = frg. 27 di Marco).
Von den Philosophen selbst werden in den ersten heiden
Generationen nach Aristoteles-sieht man von Epikur ab-nur
biographische Einzelheiten uberliefert etwa von Phainias in der
Schrift uber die Sokratiker (frg. IV A 1 Giannantoni = frg. 31
Wehrli), vielleicht beilaufig von Chamaileon (frg. IV H 11 Giannantoni4) und von Philipp dem Megariker (frg. IV E 5 Giannantoni = frg. 164 A Doring5); dasselbe gilt auch fur Kallimachos (frg.
IV H 27 Giannantoni = frg. 438 Pfeiffer). Nur Bion von Borysthenes, der zeitweilig Schuler des Theodoros war (frg. IV H 28 =
1 Die Zeugnisse und Fragmente sind zusammengestellt und teilweise
auch erliiutert von E. Mannebach (1961) und G. Giannantoni (1990).
2 Annikeris hatte noch einen Bruder Nikoteles und einen Schuler
Poseidonios (frg. IV G 2 Giannantoni).
3 Frg. IV A 9 Giannantoni = frg. 37 Kassel Austin; zur Bezeichnung
croqncrn\~ s. Giannantoni (1958) 21; 31-32; Classen (1986) 193; 212. Alexis frg.
IV A 71 Giannantoni = frg. 241 Kassel Austin liil3t sich nicht mit Sicherheit
auf Aristipp beziehen.
4 Wehrli schreibt Chamaileon die Bemerkung iiber Theodoros' Tod
nicht mehr zu (frg. 35), s. auch Manne bach ( 1961) frg. 266. Zu den iilteren
Berichten iiber Aristipp s. Classen (1986) 267-271 mit Anm. 13-39 (zuerst
1958) und Doring (1988) 62-69.
5 Eine genaue Datierung ist nicht moglich.

ARISTIPP UND SEINE ANHANGER IN ROM

207

test. 19 und 3 Kindstrand), gibt in seinen Diatribai eine Anekdote

uber Aristipp wieder (frg. IV A 79 Giannantoni = frg. 40 Kindstrand6), wahrend Teles in seinen Diatriben Aristipp ein Dictum
zuschreibt, das sonst anderen zugewiesen wird (frg. IV A 103
Giannantoni = p. 29, 13-30, 1 Hense).
Sogar die Autoren, die in den folgenden Generationen uber die
Schulen der Philosophen und deren Doxai berichten, oder auch
die Philosophen selbst zeigen ein erstaunlich geringes Interesse
an den Lehren der Kyrenaiker: Es wird auf Lehrer-SchulerBeziehungen verwiesen, z.B. von dem nicht sicher datierbaren
Antisthenes von Rhodos in seinen cltA.ocrocjlrov ~taooxai (frg. IV H
13 Giannantoni = frg. 5 Giannattasio Andria), und es werden
Schriftenverzeichnisse uberliefert (frg. IV A 144 Giannantoni =
Sotion frg. 6 Wehrli, Panaitios frg. 123 v. Straaten und Sosikrates
frg. 10 Giannattasio Andria), au8erdem biographische Einzelheiten etwa von dem Historiker Idomeneus, einem Schuler
Epikurs (frg. IV A 15 Giannantoni = frg. 26 Angeli, s. auch frg. 27
AngelF), und vor all em Anekdoten, die oft Tadel an Aristipp und
seinen Anschauungen spurbar werden lassen und Polemik, aber
auch Rechtfertigungen, von Sotion (frg. IV A 17 und 96 Giannantoni = frg. 4 und 5 Wehrli) oder Hegesander in seinen Hypomnemata (frg. 17,26 und 36 Giannantoni = frg. 17,1 und 18 Muller),
wie auch sonst die Berichte und Zitate oft durch Hinweise auf
Gegensatze oder Ahnlichkeiten oder durch Polemik bestimmt
sind 8
6 Zur Formulierung oi. 7tp\. tov Birova vgl. Kindstrand (1976) 150-151 (ad
loc). Eine Anekdote erzahlt auch der Historiker Istros in seiner literarisch
nicht Ieicht einzuordnenden Schrift nep\. ilitotT]toc; a9A.rov (frg. IV E 2
Giannantoni = FGrHist 334 frg. 55 Jacoby); da sie nicht zu einem Anhanger
der Lebensweise Aristipps paBt und auch von einem Kyrenaer Eubotes
erzahlt wird (bei Aelian VH 10, 2, der kurz darauf den Kyrenaer Aristoteles
erwahnt: VH 10, 8 = frg. IV E 3 Giannantoni), ist sie wohl nicht auf den
Kyrenaiker Aristoteles zu beziehen, von dem der Megariker Philipp spricht.
7 Zu Idomeneus s. jetzt M. Erler (1994) 244-246, auch zu Epikur (29-202)
und dessen anderen Schiilern, z.B. Metrodoros (216-221); auch der mit
diesem Diog. Laert. 10, 136 genannte, wohl ins 2. Jahrhundert zu datierende
Diogenes von Tarsos scheint die Kyrenaiker erwahnt zu haben (= frg. IV A
200 Giannantoni); daB Diogenes Laertios dessen Schriften selbst noch benutzt
hat, ist allerdings wenig wahrscheinlich, wie M. Giusta (1963) 121-122 mit
Recht betont.-Von den friiheren Historikern hat wohl nur Theopomp
Aristipp in seiner Polemik gegen Platon genannt (FGrHist 115 frg. 259 =
frg. IV A 146 Giannantoni); sein Vorwurf deutet iibrigens auf die Existenz
von Schriften Aristipps, vgl. auch Doring (1988) 68.
8 Zur Polemik vgl. Antigonos von Karystos frg. IV A 161 Giannantoni = p.

208

CARL JOACHIM CLASSEN

Dagegen bleibt sehr allgemein, was etwa Hippobotos9 oder

Panaitios iiber die Auffassung der Kyrenaiker von der iJoovft sagen
(frg. IV A 172 Giannantoni = Panaitios frg. 49 v. Straaten und
Hippobotos frg. 4 Gigante) oder Kleitomachos zu den von den
Kyrenaikern nicht akzeptierten Teilen der Philosophie (frg. IV A
172 Giannantoni = frg. 10 Mette), anders iibrigens als Sotion (frg.
IV A 168 Giannantoni = frg. 7 Wehrli) 10 So zeigt sich, daB Aristipp
und seine Anhanger bei den Griechen nicht vollig in Vergessenheit gerieten, man ihnen aber andererseits nur ein begrenztes
Interesse entgegenbrachte. Denn wenn uns auch nur wenige
Triimmer der hellenistischen Literatur vorliegen, so diirften diese
doch ein einigermaBen richtiges Bild von dem geben, was man
von Aristipp und den Kyrenaikern wuBte, was man iiber sie
dachte und welche Bedeutung man ihnen beimaB 11
Ahnlich wie Panaitios ist Kleitomachos hier besonders wichtig,
weil er die Briicke zu den Romern schlagt: Er kann nicht nur von
den Erlebnissen der Teilnehmer an der Philosophengesandtschaft
in Rom berichten, er widmet dem Konsul L. Marcius Censorinus
und dem Satiriker Lucilius nach Ciceros Zeugnis philosophische
118 v. Wilamowitz und den nicht sicher datierbaren Theodoros (zu ihm
Mejer (1972) 76; 78-79) frg. IV A 15 Giannantoni; s. ferner PapHerc 418 frg. 4,
10-14 (= frg. IV A 72 Giannantoni; dazu W. Cronert (1906) 19). Parallelen
werden z.B. zu dem Stoiker Ariston notiert (frg. IV A 107 und 166 Giannantoni = SVF I frg. 349 und 353 v. Arnim).
9 Zu seiner Einteilung der Philosophenschulen s. frg. I H 6 Giannantoni
= frg. 1 Gigante; vgl. auch die Aufziihlung der Schulen durch Herakleides,
den 'Pythagoreer' (bei Joseppos, lib. mem. 143 = frg. I H 6 Giannantoni),
dessen Gleichsetzung mit Herakleides Lembos durch H. Usener von Wehrli
(1978) 17-18 mit guten Grunden bestritten wird.
IO Der vor Kleitomachos genannte Meleagros llepl. lio~rov (frg. IV A 172
Giannantoni) begegnet nur hier und ist nicht sicher zu datieren; dagegen
ergibt sich die Datierung von Nikolaos und Sotion (frg. IV A 202 Giannantoni) aus ihrer Polemik gegen Diokles, der seinerseits dadurch datiert wird,
daB Meleagros von Gadara ihm seinen Kranz widmet, s. Martini ( 1905) 798801; Diokles (cf. frg. IV A 30 Giannantoni) wird von Diogenes Laertios
zweimal Magnes genannt 7, 162 und 7, 48, wo zugleich seine Schrift mit
dem Titel 'ElttlipOJ.ltl trov <l>tA.oao<~>rov begegnet, der 10, 11 verkiirzt ('EmlipoJ.l{J)
erscheint, wiihrend Diogenes sonst von Biot trov <i>tA.oao<~>rov spricht (2, 54; 82)
oder Diokles ohne Buchtitel erwiihnt (mehrfach) und stets ohne Herkunftsangabe (mit Ausnahme der oben angegebenen zwei Stellen); Euseb kennt
Diatribai eines Diokles von Knidos (frg. IV H 29 Giannantoni), dessen
Identitii.t jedoch nicht gekliirt ist, vgl. Kindstrand (1976) 169.
11 Bekanntlich erwiihnt auch Ps.-Demetr. de eloc. Aristipp zweimal (288
und 296 = frg. IV A 14 und 148 Giannantoni); doch ist die Datierung dieser
Schrift weiterhin umstritten, so daB hier auf eine Einordnung verzichtet
werden muB.

ARISTIPP UND SEINE ANHANGER IN ROM

209

Schriften 12. Lucilius ist dann der erste Romer, der-soweit die
triimmerhafte Uberlieferung uns zu urteilen erlaubt-Aristipp
erwahnt (frg. IV A 145 Giannantoni = frg. 800 Krenke}l3) und zwar
ein Buch, das Aristipp dem Tyrannen Dionys geschickt habe,
wieder ein Stiick Biographie oder wohl eher eine Anekdote. So
stellt sich die Frage, welche Kenntnisse die Romer spater von
Aristipp und seinen Anhangern hatten und wie sie zu ihnen
gelangt sein mogen.
In seinen friihen Schriften spielt Cicero nur einmal auf Aristipp
an, und zwar auf einen der Ausspriiche iiber den Umgang mit
iiberfliissigem Reich tum ( inv. 2, 176 = frg. IV A 83 Giannantoni);
und eine Anekdote, die z.B. Vitruv (6,1, 1) und Galen (Protr. 5) auf
Aristipp beziehen, fiihrt er sogar nur mit der unbestimmten
Formulierung ut mihi Platonis illud, seu quis dixit alius, perelegans esse
videtur (rep. 1,29 = frg. IV A 50 Giannantoni) ein. Auch spater
begegnen Anekdoten fast nie, nur in den Tuskulanen eine kiihne
Antwort des Theodoros auf eine Drohung des Konigs Lysimachos
(1,102 = frg. IV H 8 Giannantoni) und in einem Brief an Paetus
Aristipps beriihmter Ausspruch iiber sein Verhaltnis zu Lais (Jam.
9,26, 2 = frg. IV A 95 Giannantoni).
Von der Philosophic Aristipps und der Kyrenaiker spricht
Cicero zum ersten Mal in seiner Ubersicht iiber die sokratischen
Schulen (de orat. 3, 61-62 = frg. I H 4 Giannantoni): Neben Platon
und Antisthenes nennt er Aristipp, begniigt sich hier allerdings
nicht (wie bei Antisthenes) mit einem Hinweis auf die besondere
Vorliebe fiir bestimmte Inhalte der sokratischen Dialoge und
einem weiteren auf die von ihm ausgehende Cyrenaica philosophia;
er charakterisiert auBerdem Aristipps und seiner Schiiler Umgang
mit den eigenen Anschauungen knapp, ehe er sie den zeitgenossischen Vertretern ahnlicher Lehren gegeniiberstellt 14 : Das offene
Bekenntnis zur Lust, das jene friiher einmal ablegten (simpliciter
defenderunt), scheint ihm den Vorzug zu verdienen gegeniiber

pr. 2, 137 und 102 (=test. 18 Marx).

Krenke! (1970) 445 verweist auf Diog. Laert. 2, 82 und vor allem 2, 84 (=
frg. IV A 144 Giannantoni); frg. 716-718 Krenke! spricht Lucilius ubrigens
auch von Socratici carti.
14 Nachdem Cicero zuniichst eindeutig vom iilteren Aristipp gesprochen
hat, wiihlt er fUr die spiiteren Kyrenaiker bewuBt das seltene posteri, das er
fin. 5, 13 fUr die spiiteren Vertreter des Peripatos verwendet, von denen er
sagt ... ita degenerant, ut ipsi ex se nati esse videantur.
12 Cicero ac.
13

210

CARL JOACHIM CLASSEN

denen, qui nunc voluptate omnia metiuntur, also den Epikureern, die
sich doch weder voll zur voluptas noch zur dignitas (der Menschen)
bekennen.
Erst in den philosophischen Schriften der letzten Jahre spricht
Cicero ein wenig haufiger von Aristipp, dessen Lebensweise,
dessen Anschauungen und den Lehren der Schuler. Allerdings
beriihrt er nur wenige Aspekte und seine Formulierungen bleiben
meistens sehr allgemein. Eine Priifung der einzelnen Stellen zeigt
zunachst, daB bald Aristipp allein, bald Aristipp und die Kyrenaiker, bald diese allein genannt werden, und zwar oft so, daB sie
mit anderen Philosophen verglichen oder kontrastiert werden.
Genaueres Hinsehen lehrt, daB Cicero nur einmal die Schrift
eines Kyrenaikers anfiihrt, den 'A1t01capn:prov des Hegesias ( Tusc.
1, 84 = frg. IV F 4 Giannantoni), also nie ein Werk des alteren oder
des jiingeren Aristipp, und sehr vorsichtige Formulierungen
wahlt, wenn er die Meinungen oder Lehren Aristipps oder der
Kyrenaiker erwahnt. Wo er sich auf den alteren oder jiingeren
Aristipp bezieht, gebraucht er gern Wendungen wie Aristippi est
oder voluptas illa Aristippi15 oder Ellipsen 16, ahnlich bisweilen auch
bei den Kyrenaikern 17. Fiir den alteren oder jiingeren Aristipp
wahlt er dan eben je einmal dicit und dicere (jin. 2, 19 und Tusc. 2, 15
= frg. IV A 183 und 204 Giannantoni), einmal ducit, einmal ponit in
voluptate ... non dolere: (jin. 2, 41 und 2, 19 = frg. IV A 185 und 183
Giannantoni) und einmal sententia (jin. 2, 19, s. auch Tusc. 2,15 =
frg. IV A 183 und 204 Giannantoni), das er daneben auch einmal
fiir Aristipp und die Kyrenaiker verwendet (sententiae ... :fin. 2, 39 =
frg. IV A 185 Giannantoni) 18, einmal fiir diese allein ( Tusc. 3, 52 =
frg. IV A 208b Giannantoni); fiir sie finden sich auBerdem posuerunt, censuerunt und censent19 , non recusant und negant (jin. 2, 114 und
15 Fin. 1, 23 (= frg. IV A 180 Giannantoni) und 2, 35 (nicht bei Giannantoni); s. ferner fin. 2, 18 und 20 illud Aristippeum (genus); fin. 2, 34-35 ... fines ...
Aristippo simplex voluptas, ... unus Aristippi vel Epicuri ... Aristippo voluptas; fin. 5,
20 ... principes: voluptatis Aristippus (= frg. IV A 183, 184 und 187 Giannantoni),
auch ac. pr. 2, 139 Aristippus ... solum corpus tuetur (s. auch fin. 2, 18) = frg. IV A
179 (und 183) Giannantoni.
16 Fin. 1, 26 und 2, 18 (quam Aristippus) = frg. IV A 181 und 183 Giannantoni.
17 Fin. 1, 39: conclusum est ... contra Cyrenaicos; Tusc. 3,31: accipio ... a Cyrenaicis haec arma = Frg. IV A 182 und 208a Giannantoni.
18 Fur sie zusammen auch simpliciter defenderunt (de orat. 3, 62, s.o.); ... esse
voluerunt, quorum princeps Aristippus, ... unde Cyrenaici (ac. pr. 2, 39) = frg. I H 4;
IV A 178 und 185 Giannantoni.
19 Off 3, 116: Cyrenaici atque Annicerii ... omne bonum in voluptate posuerunt

ARISTIPP UND SEINE ANHANGER IN ROM

211

ac. pr. 2, 76 = frg. IV A 186 und 209b Giannantoni) sowie putant und
putent (ac. 2, 142; 2, 20 und Tusc. 3, 76 = frg. IV A 209a, 209c und
208c Giannantoni). Fiir Theodoros20 (immer zusammen mit
Diagoras) gebraucht er putaverunt, deorum naturam sustulerunt und
deos esse negabant (nat. deor. 1, 2; 1, 63 und 1, 117 = frg. IV H 19a; 19b
und 19c Giannantoni). Dagegen heiBt es in den Tuskulanen a
Cyrenaica Hegesia sic copiose disputatur (1, 83 = frg. IV F 3 Giannantoni), unmittelbar bevor dessen Buch erwahnt wird.
Auffallig ist weiterhin, daB Cicero gelegentlich, wenn auch
selten genug, den alteren Aristipp eindeutig bezeichnet, etwa als
Schiiler des Sokrates (ac. pr. 2, 131 = frg. IV A 178 Giannantoni2I),
nie jedoch den von den Griechen als jlT]'tpooioax:'toc; charakterisierten Enkel (cf. e.g. Strabo 17, 3, 22 und Diog. Laert. 2, 83 und 86
= frg. IV B 1 und IV A 160) und daB er nur einmal einen Aristipp
(den alteren oder den jiingeren) mit den Kyrenaikern unmittelbar
zusammennimmt (fin. 2, 39 = frg. IV A 185 Giannantoni), ihn
sonst aber deutlich von ihnen absetzt22 oder allein von Aristipp23
oder allein von den Kyrenaikern24 spricht; eine Systematik bei
der Verwendung der einzelnen Bezeichnungen ist nicht erkennbar25.
SchlieBlich tallt auf, daB sich das, was Cicero iiber Aristipp und
die Kyrenaiker sagt, auf sehr wenige Themen beschdinkt, und auf
sehr vage und allgemeine Aussagen iiber diese Themen, und weiter, daB er deren Anschauungen in der Regel neben die anderer
Philosophen riickt, sei es daB diese vergleichbare oder entgegengesetzte Positionen vertreten. Diese Beobachtungen erlauben die
Vermutung, daB Cicero weder je ein Werk des alteren noch des
virtutemque censuerunt; Tusc. 3, 28 und 52 = frg. IV A 189; 208 a und 208 b
Giannan toni.
20 Zweimal erzahlt Cicero Anekdoten von Theodorus, Tusc. 1, 102 und 5,
117 (= frg. IV H 8 und 7 Giannantoni) und einmal von Anti pater ( Tusc. 5, 112
= frf IV C 1 Giannantoni).
2 Vgl. ferner de orat. 3, 62 (s.o.), Tusc. 2, 115 Socraticus Aristippus und auch
off 1, 148 si quid Socrates aut Aristippus contra morem ... fecerint (= frg. I H 4; IV A
204 und 83b Giannantoni).
22 Ac. pr. 2, 131; fin. 1, 23; off 3, 116 = frg. IV A 178; 180; 189 Giannantoni.
23 Cf. ac. pr. 2, 139; fin. 1, 26; 2, 18-20; 34-35; 41; 5, 20; nat. dear. 3, 77 = frg.
IV A 179; 181; 183; 184; 185 (amEnde); 187; 161a Giannantoni.
24 Cf. ac. pr. 2, 20; 76 (Cyrenaei}; 142;fin. 1, 39; 2, 114; Tusc. 3, 28-31; 52; 76
(Konjektur) = frg. IV A 209c; 209b; 209a; 182; 186; 208a; 208b; 208c
Giannan toni.
25 Dort wo die Cyrenaici allgemein genannt werden, wird fast stets gegen
Epikur polemisiert.

212

CARL JOACHIM CLASSEN

JUngeren Aristipp gelesen hat, sich aber auch nicht auf Handbiicher mit Darstellungen der einzelnen Philosophenschulen
stiitzt, sondern auf Passagen bei anderen Philosophen, die sich auf
Aristipp und dessen Anhanger berufen oder gegen sie polemiSleren.
Damit wird zugleich die Frage aufgeworfen, welche Anschauungen und Lehren der Kyrenaiker Cicero iiberhaupt kennt und
warum er sie erwahnt.
Die erkenntnistheoretischen Probleme, die Cicero in den Academica erortert, geben ihm AnlaB, mehrfach von den Kyrenaikern
zu sprechen. Zunachst verweist Lucullus in seiner Rechtfertigung
der Moglichkeit von Erkenntnis, die er mit Ausfiihrungen iiber
die Zuverlassigkeit der Sinne beginnt, auf den 'inneren Tastsinn',
der Schmerz oder Lust empfindet und mit dem, wie er hinzufiigt,
fiir die Kyrenaiker das einzige Wahrheitskriterium gegeben ist
( ac. pr. 2, 20 = frg. IV A 209c Giannantoni). Ganz ahnlich klingt
die Beschreibung, die Cicero selbst in seiner Entgegnung von der
Lehre der Kyrenaiker gibt, die er hier als minime contempti philosophi
bezeichnet ( ac. pr. 2, 76 = frg. IV A 209b Giannantoni2 6 ); sie
konnten, wie er sagt, nur erfassen (erkennen), was sie mit dem
inneren Tastsinn spiirten, und er fiigt hinzu, sie wiiBten nicht,
welche Farbe oder welchen Klang etwas habe, sie empf'anden nur,
daB sie irgendwie betroffen seien. Gegen Ende des Dialogs
schlieBlich kommt Cicero noch einmal auf die Position der
26 Die Lesart Cyrenaei iiberrascht (und wurde schon von Lambin 1573
durch Konjektur in Cyrenaici geiindert, das einige Editoren in den Text
aufgenommen haben); denn die Griechen unterscheiden KupT]va\:oc;
(Einwohner von Kyrene) und KupT]Yai:Koc; (Anhiinger der Philosophie der
Kyrenaiker). Doring ( 1988) 35-37 bestreitet dies unter Berufung auf die
Formulierung, mit der Diogenes Laertios 2, 83 (= frg. IV A 144 Giannantoni) die erste Liste der Schriften Aristipps einleitet, und glaubt, daB
Diogenes 2, 86b-93 (= frg. IV A 172 Giannantoni) "referiert, ... was er als
orthodoxe, auf den Auffassungen Aristipps basierende Lehre in seiner Quelle
bzw. seinen Quellen verzeichnet fand" (35). Er erkliirt nicht, warum
Aristipp in den Quellen so oft von den Kyrenaikern getrennt wird (s. auch
o. Anm. 14 und 22) und z.B. iiberall dort, wo es urn erkenntnistheoretische
Probleme geht, Cicero, Plutarch oder Sextus Empiricus nie von Aristipp,
sondern stets nur von den Kyrenaikern sprechen bzw. Aristokles von oi EK
tile; KuptiYTJc; oder oi Kat' 'ApianmtOY (toY KupT]YatoY) (Doring (23): "die nach
allgemeiner Auffassung in der Nachfolge des Aristipp aus Kyrene philosophieren ") und dane ben von oi 1tpl 'tOY 'E1tl1COUpOY ElC tf\c; 'Apt<J'tl1t1t0\J otayroyf\c;
op~tc6~tYOt (frg. IV A 209-219 Giannantoni). Ich bin eher geneigt, an jeweils
einmalige Irrtiimer bei Cicero und Diogenes zu glauben (ac. pr. 2, 76 und
Diog. Laert. 2, 84 = frg. 209 und 144 Giannantoni).

ARISTIPP UND SEINE ANHANGER IN ROM

213

Kyrenaiker zunick (ac. pr. 2, 142 = frg. IV A 209a Giannantoni). Er

beginnt die Schilderung der Uneinigkeit der Philosophen fiir den
Bereich der Dialektik, ehe er eine Reihe weiterer Meinungen
referiert, zunachst, indem er Protagoras und die Kyrenaiker
einander gegeniiberstellt-Protagoras, der die Auffassung vertritt,
daB fiir jeden wahr ist, was ihm als wahr erscheint27, und die
Kyrenaiker, die praeter permotiones intumas nihil putant esse iudicii.
Zum dritten Mal begniigt sich Cicero mit einer sehr allgemeinen
Formulierung, urn die erkenntnistheoretische Position der Kyrenaiker zu charakterisieren, und verzichtet erneut darauf, sie im
einzelnen vorzustellen, zu erortern oder zu widerlegen.
Haufiger erinnert Cicero in den Diskussionen ethischer Fragen
an Aristipp, den alteren oder den jiingeren, und an die Kyrenaiker. Kurz vor der eben zitierten Stelle in den Academica wendet
er sich anlaBlich der Vielfalt der Meinungen, die die Philosophen
iiber Gut und Bose vortragen, nach vielen anderen, die er teilweise
fiir nicht mehr bedeutsam halt (ac. pr. 129: ilia quae relicta ... videntur;
130: abiecti), den en zu, die seit Ianger Zeit und oft vertreten worden
sind: alii voluptatem finem esse voluerunt, quorum princeps Aristippus, qui
Socratem audierat, unde Cyrenaici, post Epicurus, cuius est disciplina nunc
notior nee tamen cum Cyrenaicis de ipsa voluptate consentiens. (ac. pr. 2,
131 = frg. IV A 178 Giannantoni). Wichtig ist hier, wie mir
scheint, der Hinweis, daB den Zeitgenossen eher die Auffassung
Epikurs gelaufig ist-die Kyrenaiker sind offenbar weitgehend
vergessen. Wenn Cicero trotzdem betont, daB Epikur ihrer Auffassung von der voluptas nicht voll zustimmt, so will der Skeptiker
hier die Meinungsvielfalt mit einem weiteren Beispiel illustrieren. Denn die Formulierung voluptas finis est, ist so allgemein, daB
sie fiir Aristipp ebenso wie fiir Epikur gelten kann, und so stellt
Cicero auch beide wenig spater als Alternativen nebeneinander:
labor eo ut adsentiar Epicuro aut Aristippo (ac. pr. 2, 139 = frg. IV A 179
Giannantoni). AnschlieBend erlautert er dann allerdings Aristipps Position, urn ihn mit dem Stoiker Zeno kontrastieren zu
konnen: Aristippus quasi animum nullum habeamus corpus solum tuetur,
Zeno, quasi corporis simus expertes, animum solum complectitur.
Ahnlich kann Cicero Epikur gleich zu Beginn seiner Darlegungen iiber die Hilflosigkeit vieler Philosophen aus verschiedenen
Schulen in den Tuskulanen unmittelbar an Aristipp heranriicken
27 Zu Protagoras s. Classen (1989).

214

CARL JOACHIM CLASSEN

(2, 15 = frg. IV A 204 Giannantoni28), weil er sich auf die sehr allgemeine Aussage summum malum dolor beschrankt, der er ahnlich
knappe Formulierungen zu anderen Philosophen entgegensetzt.
Auch in seiner einleitenden Kritik an Epikur in de finibus
bedient sich Cicero, sobald er sich der Ethik zuwendet, nur sehr
allgemeiner Formulierungen fiir die Kyrenaiker: Der Grundsatz,
daB voluptas und dolor alles bestimmen, was wir erstreben oder
meiden, wird vor Epikur auch schon von Aristipp vertreten und
von den Kyrenaikern besser und freimiitiger verteidigt (1, 23 = frg.
IV A 180 Giannantoni), eine Bemerkung, die er durch einen sehr
scharf kritisierenden Zusatz erganzt. Ahnlich heiBt kurz darauf ...
voluptatem ... summum bonum ... ante Aristippus, et ille melius (se. quam
Epicurus) (1, 26 = frg. IV A 181 Giannantoni). Damit greift Cicero
auf, was er schon bei der ersten Vorstellung Aristipps und seiner
Anhanger in de oratore iiber sie und Epikur sagte: daB sie zusammengehoren und daB die Kyrenaiker ihre Position freimiitiger,
rigoroser (d.h. wohl ohne Riicksicht auf die Konsequenzen,
mogen sie auch noch so abstoBend sein) vertreten, anders als
Epikur und dessen Anhanger, von denen es dort heiBt dum verecundius id agunt, nee dignitati satis faciunt, quam non aspernantur, nee
voluptatem tuentur, quam amplexari volunt (de orat. 3, 6229 )' wahrend
Cicero hier von dem voluptas-Begriff der Kyrenaiker sagt eius modi
esse iudieo, ut nihil homine videatur indignius (fin. 1, 23).
Dagegen laBt Cicero Torquatus ein wenig spater nicht nur den
Unterschied zu den Kyrenaikern durch den Hinweis betonen, der
mit Hilfe der unbewegten Hand der Chrysippstatue gewonnene
SchluB richte sich gegen sie, nicht gegen Epikur, er legt dem
Epikureer noch eine Uberlegung in den Mund, die die voluptasvorstellung der Kyrenaiker deutlicher werden lassen soli (fin. 1, 39
= frg. IV A 182 Giannantoni): ''Wenn allein das Lust sei, was die
Sinne gleichsam kitzele-um es einmal so zu formulieren-und
mit SiiBigkeit auf sie zustrome und in sie hineingleite, dann ...
konne kein Teil des Korpers zufrieden sein mit der Freiheit vom
Schmerz ohne eine angenehme Lustbewegung." Offensichtlich
ist Cicero mit der Definition der voluptas (iJooviJ) als A.eia KiVT\<n<;
vertraut (cf. Diog. Laert. II 86 = frg. IV A 172 Giannantoni), und
dies wird durch mehrere Passagen im zweiten Buch de finibus

28 Zu enervata muliebrisque sententia s. Gorier (1974) 79, s. auch 50.

29

Vgl. auch.fin. 2, 14 (= frg. 1V A 186 Giannantoni): verecundius/constantius.

ARISTIPP UND SEINE ANHANGER IN ROM

215

bestatigt, z.B. eam voluptatem . . . quam Aristippus, id est, qua sensus

dulciter ac iucunde movetur oder ... in ea voluptate, quae maxima dulcedine
sensum moveret oder ... optabile esset maxima in voluptate ... versari, cum
omnes sensus dulcedine omni quasi peifusi moverentur (fin. 2, 18; 39 und
114 = frg. IV A 183; 185 und 186 Giannantoni) oder durch eine
Formulierung in den Academica: revocat virtus .. , pecudum illos motus
esse dicit (2, 139 = frg. A IV 179 Giannantoni).
Die angefiihrten Stellen lassen zugleich zwei andere Aspekte
erkennbar werden, die immer wieder begegnen, wo Cicero von
der voluptas Aristipps oder der Kyrenaiker spricht: Einerseits auBert
er oft scharfe Kritik 30 an Aristipps Deutung der zugleich als finis
oder summum bonum bezeichneten voluptas, da sie menschenunwiirdig sei. Dafiir verweise ich neben de orat. 3, 62 und ac. pr. 2,
139 auf fin. 1,23 (ut nihil homine videaturindignius),fin. 2, 18 (voluptatem
... quam etiam pecudes, si loqui possent, appellarent voluptatem), fin. 2, 40
( ut tardam aliquam et languidam pecudem ad pastum et ad procreandi
voluptatem hoc divinum animal ortum esse voluerunt, quo nihil mihi videtur
absurdius) und fin. 2, 114 qui est autem dignus nomine hominis, qui unum
diem tatum velit esse in genere isto voluptatis? (= frg. I H 4 und IV A 179;
180; 183; 185 und 186 Giannantoni)3I.
Zum anderen riickt Cicero neben die voluptas-vorstellung Aristipps und der Kyrenaiker gern eine oder mehrere andere vom
finis oder vom summum bonum, etwa das Freise in vom Schmerz (non
dolere, vacuitas doloris, sine dolore esse32) oder virtus oder honestas33,
oder mehrere wie doloris vacuitas und Jrui principiis naturalibus (fin. 2,
35, vgl.fin. 5, 17 = frg. IV A 187 Giannantoni). Offensichtlich
erscheint Aristipps voluptas Cicero besonders geeignet, urn eine bestimmte Position in der ethischen Diskussion gegeniiber anderen
zu markieren, und so begegnet er mehrfach neben Hieronymos
und Karneades (fin. 2, 35 sowohl mit Epikur-wie auch ac. pr. 2, 131
neben Hieronymos und Kalliphon-wie ohne Epikur), aber auch
Epikur ohne Aristipp (fin. 4, 49; 5, 87; Tusc. 5, 84). Daraus wird deutlich, daB es Cicero bei solchen Reihungen nur ganz allgemein
30 Diese Kritik spiegelt sich nicht nur in Anekdoten und Ausspriichenauf die Cicero fast vollig verzichtet-sondern auch in der Charakterisierung
der Aristippanhanger als aaco'tot, an die er nat. dear. 3, 77 erinnert (= frg. IV
A 161 Giannantoni).
31 Vgl. dazu auch Gorier (1974) 70-71.
32 Cf. fin. 2, 18-20; 34; 39 und 41, auch schon 1, 37-39 (frg. IV A 183, 184,
185 und 182 Giannantoni).
33 Cf. fin. 2, 34 und 2, 19 (= frg. IV A 184 und 183 Giannantoni).

216

CARL JOACHIM CLASSEN

urn die jeweiligen Grundsatze geht, nicht urn die Einzelheiten der
verschiedenen Systeme34.
Auch die Frage, ob eine These von dem alteren oder dem jiingeren Aristipp oder von allen Kyrenaikern vertreten wurde, stellt
Cicero nicht. Ihm geniigt es, eine ethische Position anfiihren zu
konnen, deren Verfechter die voluptas-simplex, sola, summa-als
summum bonum, als finis ansahen, expers honestatis, quae maxima
dulcedine sensum movet, und entsprechend dolorals summum malum35,
und eine weitere, nach der-im Gegensatz zu anderen-nicht
jedes Unheil Betriibnis aufkommen laBt, sondern nur unvermutetes und nicht erwartetes3 6 , ferner eine erkenntnistheoretische,
nach der der innere Tastsinn als einziges Wahrheitskriterium
anzusehen ist, und eine 'theologische', nach der die Existenz der
Gotter geleugnet wird und fur die es ihm nicht so wichtig ist, sie
mit den iibrigen Anschauungen der Kyrenaiker zu verkniipfen,
als zwei Reprasentanten zu nennen, neben Diagoras auch Theodoros37.
Fassen wir zusammen. Wie andere Romer, friihere, gleichzeitige und spatere, kennt Cicero einige Anekdoten iiber Aristipp und
dessen Anhanger, die deren Lebensklugheit und Unabhangigkeit
illustrieren3 8 AuBerdem ist er mit einigen philosophischen
Grundsatzen Aristipps und der Kyrenaiker vertraut, vor allem mit
ethischen, auf die er sich in der Regel bezieht, urn gegen sie und
Epikur oder mit ihrer Hilfe gegen Epikur zu polemisieren oder urn
Meinungen und Anschauungen anderer Philosophen durch
Vergleich oder Kontrast deutlicher herauszuarbeiten. Den Versuch, alle Lehren der Kyrenaiker zusammen zu erfassen und als
34 Das gilt auch fiir die kurze Bemerkung off. 3, 116 (= frg. IV A 189
Giannantoni).
35 Simplex: fin. 2, 34; sola: fin. 2, 41; summa: fin. 2, 41; summum bonum: fin. 2,
19; finis: ac. pr. 2, 131; expers honestatis: fin. 2, 35; quae maxima dulcedine sensum
movet:fin. 2, 39 (cf. auchfin. 1, 39 undfin. 2, 18); dolor: Tusc: 2, 15 (= frg. IV A
184; 185; 183; 178; 182 und 204 Giannantoni).
36 Cf. Tusc. 3, 28-31; 52; 75-76 (= frg. IV A 208 Giannantoni).
37 Cf. nat. deor. 1, 2; 1, 63; 117 (= frg. IV H 19 Giannantoni).
38 Von den friiheren ist nur Lucilius zu nennen (frg. 800 Krenke! = frg.
IV A 145 Giannantoni), von den gleichzeitigen nur Vitruv (6, 1, 1 = frg. IV A
50 Giannantoni), von den spateren Horaz (frg. IV A 45; 80; 100 Giannantoni), Valerius Maxim us, Seneca, Apuleius, Celli us, Porphyrio, Ausonius und
Ps.-Caesius Bassus (= frg. IV A 47; 65; 104; 49; 45; 80; 100; 81 Giannantoni); zu
Aristipps Lebensart s. Tertullian apol. 46, 16 (= frg. IV A 54 Giannantoni);
Anekdoten aus Ciceros Werken geben Laktanz und (zu Theodoros) Valerius
Maximus und Seneca wieder, aus Gellius (zu Aristipp) Augustin: frg. IV A 83
und 95, H 8 und A 49 Giannantoni.

ARISTIPP UND SEINE ANHANGER IN ROM

217

philosophisches System zu verstehen ( oder gar zu widerlegen),

unternimmt Cicero nicht. Dafiir hatte er mehrere Griinde. Zum
einen ist nicht sicher, wie weit die Kyrenaiker ein umfassendes
philosophisches Lehrgebaude entwickelt hatten 39 , zum anderen
kannte Cicero offenbar keine Gesamtdarstellung, sondern nur Einzelheiten, und zwar aus der Polemik anderer Schulen, auBerdem
vielleicht auch aus Handbiichern, in denen vergleichbare und
entgegengesetzte Lehrmeinungen zusammengestellt waren; vor
allem aber gab es, wie er selbst gelegentlich bemerkt, in seiner
Zeit keine Kyrenaiker mehr (off. 3, 116 = frg. IV A 189 Giannantoni); vielmehr waren Epikur und dessen Anhanger an ihre Stelle
getreten und verbreiteten ihre vielfach sehr ahnlichen Anschauungen mit groBem Erfolg (Tusc. 4, 6) 4 0. Gegen sie glaubte Cicero
sich mit seinen philosophischen Schriften wenden und sie teilweise auch mit Hilfe der Kyrenaiker bekampfen zu miissen.
Auch spater zeigen die Romer wenig Interesse an der Philosophie
der Kyrenaiker, mogen einzelne auch dieses oder jenes Detail aus
Ciceros Darstellungen reproduzieren 41 Nur bei Seneca findet sich
etwas Neues, eine Untergliederung der Ethik, die er wohl einem
der an solchen Dingen interessierten Handbiicher entnimmt,
ohne auf sie einzugehen (epist. 89, 12, s. Anm. 39). Wenn Horaz
dichtet

39 Vgl. Kleitomachos frg. 10 Mette; Sotion frg. 7 Wehrli und Sen. epist.
89, 12 (= frg. IV A 172 und 168 Giannantoni). AuBerdem ist, wie Doring
(1988) 69 mit Recht betont, anzunehmen, daB die Lehre der 'eigentlichen'
Kyrenaiker (vor Annikeris, Hegesias und Theodoros) erst allmiihlich ihre
endgiiltige Form in der Generation des jiingeren Aristipp fand. Entsprechend wird sie in den einzelnen Phasen ihrer Entwicklung in verschiedener Weise von ihren Gegnern angegriffen worden sein (zu Aristoteles s.
Doring ( 1988) 69 Anm. 146), was die spiitere Uberlieferung, soweit sie sich
auf polemische AuBerungen stiitzte, gepriigt haben diirfte.
4 ! Wenn Epikur und seine Anhiinger auch oft in die Niihe von Aristipp
und den Kyrenaikern geriickt werden, so ist nicht zu vergessen, daB Epikur
und seine Schiiler immer wieder gegen sie polemisieren, was hier nicht im
einzelnen belegt zu werden braucht.
4 1 Vgl. Val. Max. 8, 9 ext. 3 aus Cic. Tusc. 1, 83 (= frg. IV F 5 und 3 Ginannantoni), ferner Quint. inst. 12, 23-24; Min. Fe I. 8, 1-2; Lact. inst. 3, 7, 7; 3, 8,
6-10; 7, 7, 11; epit. 28, 3; 34, 7; 63, 1; ira 9, 7; 10, 47; Arnob. nat. 4, 29; Aug. c.
acad. 3, 11, 26; c. Julian. 4, 14, 72; civ. 8, 3; 18, 41; Eucher. epist. ad Val. p. 724
Migne (= frg. IV A 190; IV H 21; IV A 191; 193; 194; 195; 192; IV H 20; H 22;
IV A 210; 188; I H 13b; 13c; IV A 196 Giannantoni); s. dazu auch die von mir
angeregte, nicht verOffentlichte Magisterarbeit von B. Ch. Weber, Die Rezeption dcr Lehren des Aristipp und dcr Kyrenaikcr durch die ROmer, Gottingen 1993.

218

CARL JOACHIM CLASSEN

nunc agilis flo et mersor civilibus undis

virtutis verae custos rigidusque satelles,
nunc in Aristippi furtim praecepta relabor
et mihi res, non me rebus subiungere conor ( epist. 1, 1, 16-19),

so verrat er nicht Kenntnis der Philosophie der Kyrenaiker (praecepta), sondern illustriert seine eigene Unabhangigkeit, die sich
am Aristipp der Anekdoten orientiert und es ihm erlaubt, ebenso
mit dessen Worten iiber Lais zu spielen (18-19) wie auch einmal
eine 'stoische' Position einzunehmen (16-17)42.
Die lebensbejahenden Gedanken und Lehren Aristipps und der
Kyrenaiker, von Epikur verdrangt und zur Bedeutungslosigkeit
verurteilt, noch ehe die griechische Philosophie nach Rom kam,
haben dort nie selbstandig FuB fassen konnen. Sie haben den
Romern allein dazu gedient, die Lehren anderer Schulen klarer
zu beschreiben oder die Philosophie zu bekampfen, der sie selbst
hatten weichen miissen.
Gottingen

4 2 Die Debatte zwischen Traina (1991), Gigante (1993) und wiederum

Traina (1994) soli hier nicht fortgefiihrt werden. Die Gefahr einer 'sopravvalutazione' Aristipps ist nicht zu verkennen.

ARISTIPP UND SEINE ANHANGER IN ROM

219

BIBLIOGRAPHIE
Fragmentsammlungen auBer denen von v. Arnim, Hense, Jacoby, KasselAustin, Muller, Pfeiffer und v. Straaten:
Angeli, A. (1981) 'I frammenti di Idomeneo di Lampsaco',Cronache Ercolanesi
11: 41-101.
DiMarco, M. (1989) Timone di Fliunte. Silli (Roma).
Doring, K. (1972) Die Megariker (Amsterdam).
Giannantoni, G. (1958) I Cirenaici (Firenze).
- - ( 1990) Socratis et Socraticorum Reliquiae I-IV (Napoli).
Giannattasio Andria, R. ( 1989) I .frammenti delle 'Successioni dei .filoso.fi' (Napoli).
Gigante, M. (1983) 'Frammenti di Ippoboto', in Omaggio a Piero Treves, a cura
di A. Mastrocinque (Padova) 151-193.
Kindstrand, J.F. ( 1976) Bion of Burysthenes (Uppsala).
Krenkel, W. (1970) Lucilius. Satiren 1-11 (Leiden).
Marx, F. (1904-1905) C. Lucilii Carminum Reliquiae 1-11 (Leipzig).
Manne bach, E. (1961) Aristippi et Cyrenaicorum Fragmenta (Lei den).
Mette, H. J. (1985) 'Kleitomachos von Karthago', Lustrum 27: 142-148.
Wehrli, F. (1969) Die Schute des Aristoteles IX: Phainias von Eresos. Chamaileon.
Praxiphanes (Basel).
- - (1978) Die Schute des Aristoteles, Supplement II: Sotion (Basel).
Die Abkurzung frg. ist von mir weitgehend auch dort verwandt worden, wo
genau genommen nur Nachrichten, also testimonia, vorliegen.
Aufsatze und Monographien:

J. (1986) Ansatze. Beitrage zum Verstandnis der frii,hgriechischen Philosophie (Wurzburg).

- - (1989) 'L'aA.iJOeta di Protagora', Rivista di Filoso.fia 70: 163-188.
Cronert, W. ( 1906) Kolotes und Menedemos (Leipzig).
Doring, K. (1988) Der Sokratesschuler Aristipp und die Kyrenaiker (Stuttgart).
Erler, M. (1994) 'Epikur', in Die Philosophie der Antike Band 4. Die Hellenistische
Philosophie, hgg. v. H. Flashar (Basel) 29-490.
Gigante, M. (1993) 'Quel che Aristippo non aveva detto', La Parola del Passato
48: 267-280.
Giusta, M. ( 1963) 'Passi dossografici di morale epicurea nel X libro di
Diogene Laerzio', Atti della Accademia delle Scienze di Torino. Classe di
Scienze Morali, Storiche e Filologiche 97: 120-174.
Gorier, W. (1974) Untersuchungen zu Ciceros Philosophie (Heidelberg).
Martini, E. (1905) 'Diokles', Paulys Realencyclopiidie der classischen Altertumswissenschft, hgg. v. G. Wissowa 5 (Stuttgart) 798-801.
Mejer,J. (1978) Diogenes Laertius and His Hellenistic Background (Wiesbaden).
Traina, A. (1991) 'Orazio e Aristippo. Le epistole e l'arte di convivere', Rivista
di Filologia e di Istruzione classica 119: 285-305.
--(1994) 'In Aristippi praecepta relabor', Eikasmos 5: 243-246.
Classen, C.

LUCRETIUS ON THE ORIGIN AND DEVELOPMENT OF

POUTICAL UFE
(DE RERUM NATURA 5.1105-1160)

P. H.

ScHRIJVERS

Lucretius' sketch of the origin and development of political life,

which forms part of his general history of the beginnings of
human life and civilisation, gives rise to the following two
questions on which this brief article, dedicated to a good friend and
highly esteemed colleague, will focus: (1) Is his account a purely
theoretical reconstruction or is it based on historical evidence in
spite of its anonymous character? (2) Is Lucretius, as an orthodox
follower of Epicurus, exclusively following the doctrine of his
Master, or does the Roman poet belong to the category of polyhistores who make use of all relevant sources and doctrines to create
their own, new, synthesis?
A quick and global view on the paragraph 5.1105-1160 of De
Rerum Natura leads to two rather evident impressions: Lucretius
presents a series of 'political' constitutions which are more or less
characterized by the position of power of certain individuals or
social groups, and this list is drawn up as a process of historical
evolution in which each form of government grows out of another.
I shall start to identify the different individuals and groups whose
changing positions were the cause of the political evolution. I use
the word 'political' in the ancient, etymological sense of 'belonging to' or 'concerning the polis', the city, because the foundation of
cities (urbes) and also of citadels (arces) is stated expressly by
Lucretius at the beginning of his political story ( 5.1108 condere
coeperunt urbis arcemque locare) . 1

At the start he mentions the superiority of certain individuals

who excelled in intelligence and inventiveness. These anonymous inventors who introduced all sorts of novelties, are called
'kings' (1109 reges) .2 In the beginning their rule seems to have
1 In the same way, Lucretius mentions the cities at the beginning of his
next paragraph on religion (5.1162); it is the origin of civil religion he is
dealing with.
2 For the topical character of this list of qualities and activities of the so-

LUCRETIUS ON THE BEGINNINGS OF POLITICAL LIFE

221

been fair because they parcelled out fields to each man and
divided flocks among them, according to their inborn qualities:
beauty, strength, intelligence (5.1110-1111). In accordance with the
ancient, traditional definition of OtKatO<JUVTJ their measures may be
considered an act of distributive justice.3 In spite of the fair
character of their rule these so-called kings did not feel secure;
they founded citadels as strongholds and refuges for themselves
(5.1109 praesidium reges ipsi sibi perfugiumque). Moreover, at a later
moment, this kingship became arrogant (1137 sceptra superba) and
in the end the kings were dreaded overmuch by the population
(1140 nimis ante metutum). Ancient political thought and historical
writing have their stereotyped descriptions: being frightened and
frightening others at the same time is what characterizes the
tyrant. 4 In the view of Lucretius these kingships, good at the start,
changed gradually into tyrannies (second form of government
implied by his text).
In his description of the good kings Lucretius mentions a
category of men distinguished by their inborn qualities, by their
aretai. Here the poet seems to define the class of aristocrats. Mter the
invention of property and the discovery of gold the privileged
position of real aristocracy was taken over by 'the richer men'
(5.1115 divitioris), the plutocrats, and-according to traditional
definitions in the political thought of the Ancients-wealth is the
characteristic feature and basis of oligarchy. 5
These oligarchs (the wealthier people are in my opinion subject
in line 1120: at claros homines voluerunt se atque potentis) wished to be
famous and powerful and started to contest the supremacy of the
called kings compare Polybius 6.7 (and the comm. by Walbank): ... JtOtouvtat
I.I.E'tCx ta\ita 'tlJV atpEOtV 'tCOV ap;(OV'tCIJV Kat jJacnAECIJV ... Kata 'tCx~ tijc; yvW/111<; Kat 'WV
A-ortO/lOV ozmpopac;... !lEV ouv Jtaf..<nov veritpaoKov tai~ pamM:iat~ ol. Kpt6vtt~ ...
'fOirOV<; 'tE Otaq>epovta~ OXVP01J/1EVOI Kat rezxft;ovrec; Kat xropav ICa'falC'fW/lEVOl, cf.
5.1107-1110 and Th. Cole, 'The Sources and Composition of Polybius VI',
Historia 13 (1964) 450-451.
3 Cf. Polybius on the good king (6.6.11) Otavti.I.T)'ttKO~ ... tou Kat'a~iav EKaotot~
and Aristotle w 1250a12 OtKatOOUVT) OE EO'ttV apE'tlJ vuxft~ OtavqtT)'ttKTJ 'tOU
Kat'a~iav, SW3.262 ElttOtTti.I.TI UltOVE1.1.1]'tt~Clj rft~ a~{a~ EKacmp
4 Cf. Polybius 5.11.6 on the tyrant: rvpavvov !lEV yap epyov EO'tt 'tO KaKro~
Jtotouvta tip q~6pcp oroJto/;ttv aKouoirov ./1zaov11evov ~eai 111aovvm toil~ imotattol.l.fvou~.
and Cicero, De rep. 2.45 (rex ille) cum metueret ipse poenam sceleris sui summam,
metui se volebat.
5 Cp. Aristotle, Pol. 2.8.5, 1273a; 6.8, 1294al0, Polybius 6.8.5, Plato, Rep.
8.551a, and in general H.-0. Weber, Die Bedeutung und Bewertung der Pleonexie
von Homer his Isokrates (diss. Bonn 1967).

to

222

P.H. SCHRIJVERS

kings. Envy and ambition caused internal struggles between these

richer men who longed to rule the world with royal power, and
the kings who wanted to keep and to defend their position at the top
(line 1130 quam regere imperio res velle et regna tenertf> does not refer in
a rather pleonastic way to one single category of ambitious men
but to the two different categories of opponents). These internal
struggles caused the fall of the kings (old and new ones all
together) and the break-down of kingship as a regular institution
(1136-1137 ergo regibus occisis subversa iacebatlpristina maiestas soliorum et sceptra superba). The situation (and the power) passed to the
utmost dregs (jaex) of disorder ( turbas) and general violence (or
rule passed to the lowest dregs of the mob).' At last, mankind worn
out by leading a life of violence submitted itself to the constitutional
state with its legislation and magistrates introduced by some
anonymous legislators (5.1143-1147). The expression magistratum
creare ('the appointment of magistrates by the people') shows that in
Lucretius' view the political evolution finally led to a democratic
republic.
Let us summarize: one finds in Lucretius' political history six
categories of people who successively had supreme power or tried
to obtain it, and therefore six forms of government succeeding
each other: kingship-tyranny I aristocracy-oligarchy I ochlocracy-democracy. The conception of political evolution through
six forms of government (three principal forms-kingship, aristocracy, democracy-degenerating resp. into tyranny, oligarchy,
ochlocracy) was rather well-known in Antiquity and has been
called the J.lE'ta~oA.it 1tOAt'tetrov ('change of constitutions'). The idea
was developed and systematized e.g. by the Greek historian
Polybius in his famous book 6 and borrowed by Cicero in his De Re
Publica. Modern studies of the sources of Polybius establish that the
Greek historian made a new synthesis from certain ideas one
finds already-in the course of the 5th century-in the works of
6 Bailey's translation of regna tenere ('to sway kingdoms') is misleading
and pleonastic; velle regna tenere describes in my opinion the aspirations of the
orifinal kings 'to keep their kingdoms'.
5.1140-1141 res itaque ad summamfaecem turbasque redibat,/ imperium sibi cum
ac summatum quisque petebat. There is a problem of interpretation here, concerning the more abstract or concrete sense of the words res ('situation' or
'reign'), turbas ('disorder' or 'mob') and redibat ('passed to' or 'returned'-to
the anarchic situation of primitive times) in these lines (cf. the commentaries of Bailey and Costa ad loc.). However this may be, in my opinion the
Latin words faex and turba allude to the technical term 'ochlocracy'.

LUCRETIUS ON THE BEGINNINGS OF POLITICAL LIFE

223

Herodotus, Thucydides, Xenophon and the Sophists, and in the 4th

century in the political works of Plato, Aristotle and the Peripatetic
scholar Dicaearchus of Messene (one also finds traces of this
theory in works and fragments belonging to the Stoic tradition) .s
Actually, my paraphrase of Lucretius' text is not based on the
commentaries of Lucretius by Bailey and Costa9-who give us too
exclusively the rather slightly attested, Epicurean background of
Lucretius' account-but on the commentary of Polybius VI by
Walbank. It appears that the concept of J.I.E'ta~oA.'h 7tOAt'tEtrov I just
mentioned, was rather popular in the time of Lucretius. We may
suppose that the intellectual aristocracy of late-republican Romewho formed the reading public of the De Rerum Natura 10-were
acquainted with this political theory. Therefore, this reading public
was able to recognize Lucretius' political story as a new version of a
well-known doctrine, presented this time by an Epicurean author.
The originality of Lucretius lies in this new arrangement and
application of traditional ideas, e.g. in the reversed order of
ochlocracy and democracy, the latter growing out of the former
one and not in the opposite direction: democracy turning into
ochlocracy, as Polybius and Cicero affirm. According to the
judgment of Cicero, Lucretius is multis luminibus ingenii, multae tamen
artis, that means in our context: multae artis politicae.
Before analysing Lucretius' view on the origin of democracy I
must make a short digression on the topic of physical beauty
which the poet mentioned as one of the natural qualities of the first
privileged aristocrats (cf. 5.1111, 1114, 1116). The second part of
book 5 of the De Rerum Natura dealing with the origin and rise of
mankind and human civilisation belongs to a definite class of
ancient texts we might call 'archeologies', in the ancient sense of
the word apxawA.oy{a: explanatory stories about 'ta apxata I 'ta
naA.aui, 'reconstructions of prehistoric developments'. So we have
the Archeology by Thucydides (the first 20 chapters of book 1), the
archeologies written by the Greek historians Dicaearchus of
Messene and Diodorus Siculus (at the beginning of their histories),

8 Cf. Th. Cole, arl. cit., H. Ryffel, Metabole Politeioon, Der Wandel der Staatsverfassungen (Bern 1949), G. J.D. Aalders, Die Theorie der gemischten Verfassung
im Alterlum (Amsterdam 1968).
9 C. Bailey, Titi Lucreti Cari De Rerum Natura Libri Sex (Oxford 1947, repr.
1963), vol. 3.1498ff., C.D.N. Costa, Lucretius De Rerum Natura V (Oxford 1984).
10 Cf. M.R. Gale, Myth and Poetry in Lucretius (Cambridge 1994) 89-90.

224

P .H. SCHRIJVERS

the 6th book of Polybius, and we find similar reconstructions of the

shadowy past in philosophical works (Plato's myth of Protagoras,
the 3th book of his Laws), Aristotle's archeology at the beginning of
his Politics: for both historians and philosophers in different ways
the temporal apxfl reflects the archetypal essence.
In their descriptions of more or less primitive times which
represented a level of civilisation less advanced than their own, all
these Greek authors could not ignore the oldest written documents
from their own past: the Homeric epics, and also the Hesiodic
poems (the Erga kai Hemerai containing a very popular version of
human evolution or rather of human deterioration: the story about
the sequence of golden, silver, bronze and iron age). Homer was
for the Ancients in all senses the first historian and also the first
ethnographer.
For his reconstruction of prehistory Thucydides used more than
once the Homeric epics; one finds references to the story of the
Cyclopes-the first antisocials in human history-in serious,
politico-philosophical discussions by Plato and Aristotle; 11 in his
analysis of kingship, in book 3 of his Politics (3.5.14,1285 b 3),
Aristotle had no choice but to analyse the heroic kingship of the
Homeric basileis; Dicaearchus explicitly discussed the historical
value of mythical stories transmitted by the poets; 12 probably, the
idea of a primitive monarchy, based on brute force, as starting
point in Polybius' archeology, has its roots in the same story about
the Cyclopes.l3
Against the background of the influence exercised by Homer
on many a Greek archeology and also on ancient ethnographical
descriptions (cf. Strabo) it is not surprising that physical beauty,
mentioned by Lucretius as a distinctive quality, was also pointed
out as such in hellenistic commentaries on Homer and in ancient
ethnographical treatises.l4
11 Cf. Plato, Leges 3.680b, Aristotle, Pol. 1252b19, and in general P. H.
Schrijvers, 'Intertextualite et Polemique dans le De Rerum Natura (V 925-1010),
Lucrece vs. Dicearque de Messene', Philologus 138 (1994) 295 and note 18.
12 Cf. P. H. Schrijvers, art. cit., passim.
13 Cf. Polybius, 6.5.7 and Aalders, op.cit. 100.
14 Cf. Ps. Plut. De vita et poesi Homeri 136: J..UXAtcrta ... 'tlJV 'ti\c; apE'ti\c; OO~<XV
cruvicrt1Jcrtv, EV otc; tov 'AxtAAE<X1tOtEl ou JlOVOV avOpEtOV a'JJ..i:J. l((lt K'alltcnov to dooc;,
Onesicritus (FGrH 134 F 21) ~acrtA.ea 'tE yap tov KaA.A.tcr1:0v aipEicr9at <P1Jcrtv
'Ov1JcriK:pt'toc; (EV OE 'tU Ka9ai~ K<XtVOta'tOV lO'tOpEt't<Xt 'tO m:pi 'tOV K'allov~ on 'tt)liitat
Ot<X(j)Epov'troc; ... ), Pomponius Mela 3.86 mos est cui potissimum pareant specie ac viribus legere. Beauty as a human value has been discussed also by Aristotle in his

LUCRETIUS ON THE BEGINNINGS OF POLITICAL LIFE

225

In my opinion one finds other traces of Homeric influence on

the Archeology of Lucretius. The situation described in the lines
5.1107-1110 (the foundation of cities and the parcelling out of fields)
is repeated by the poet in the summary which concludes his
prehistory (5.1440-1447):

1445

lam validis saepti degebant turribus aevom,

et divisa colebatur discretaque tellus,
iam <maris> velivolis florebat navibus pontus,
auxilia ac socios iam pacta foedere habebant,
carminibus cum res gestas coepere poetae
tradere; nee multo prius sunt elementa reperta.
Propterea quid sit prius actum respicere aetas
nostra nequit, nisi qua ratio vestigia monstrat.

Actually, the existence of the Homeric epics presupposes the

existence of the phenomena mentioned in the lines 1440ff. (cities,
division of lands, navigation etc.). The final summary implies also
a further development of the first two stages of social life (line 1025
mentions the first pact of cooperation and mutual defense, line
1155 the first pact of submission resulting in general peace:
communia foedera pacis) 15 : at the end a pact (treaty) of submission (to
one chief, and when we fill in the anonymous picture, it will be to
king Agamemnon) and of cooperation between cities has been
concluded ( 1143 iam pacto foedere). The lines 5.324-327 show that
Lucretius considered the Theban War and the Fall of Troy as
historical events which occurred at the beginning of the world, the
earth and the human kind:
Praeterea si nulla fuit genitalis origo
terrarum et caeli semperque aeterna fuere,
cur supera bellum Thebanum et funera Troiae
non alias alii quoque res cecinere poetae?
The final summary alludes to the content of the Iliad and in a very
sophisticated way the allusion is reinforced by the stylistic
formulation of line 1442: the use of the epic epithet velivolus and the
collocation mari 's pontus (I am following here Housman's emendation which has been defended forcefully by Timpanaro) 16 which
Politics (cf. 1282b14) and Ethics. The ancient discussion seems to be linked also
to the well-known locution Ka.AoKaya.66c;.
15 Cf. for the terminology J. Kaerst, 'Die Entstehung der Vertragstheorie
im Altertum', Zeitschrift fur Politik 2 (1909) 505fT.
16 S. Timpanaro, 'Lucrezio V 1442 (e I 314, e V 1203)', Rivista di Cultura
Classica e Medioevale 19 (1977) 730fT. (cf. Housman, Classical Papers II 436-438 =

226

P .H. SCHRIJVERS

translates the Homeric expression iiA.Oc; 7tOV'toc;.17 Here the Homeric

style echoes the event. As a matter of fact one finds other allusions
to the Homeric epics in Lucretius' description of the first primitive
period of mankind (5.948-949 templa Nympharum-Od. 13.103 av'tpov
... ipov VUJ.Lcplirov), the antisocial, 'Cyclopic' way of life (5.958ff.-Od.
9.108) 18 and also, in my opinion, in the paragraph we are talking
about (lines 1125-1128, esp. the locution in Tartara taetra) 19:
et tamen e summo, quasi fulmen, deicit ictos
invidia interdum contemptim in Tartara taetra;20
invidia quoniam, ceu fulmine, summa vaporant
plerumque et quae sunt aliis magis edita cumque;
In accordance with his regular practice Lucretius' allusion to poetic
myth (cf. the explicit reference to the Myth of the Giants in 5.117ff.
and the harangue against Jupiter's thunderbolt, 6.421-422) is
followed immediately by a reference to the same phenomenon
presented now as a natural and even common event (5.1127-1128).
I should like to conclude my digression on Lucretius and
Homer by pointing out that in his treatise entitled On the Good King
according to Homer-which has been dated 59-55 B.C., so before the
death of Lucretius21-the Epicurean philosopher-poet Philodemus
likewise mentions physical beauty as a distinctive trait of the
Homeric heroes and also their mutual envy as a source of internal
troubles.22 I conclude for the moment that this first part of Lucretius'
story about political evolution has in some lines a definitely
Homeric colouring.

journal of Philology 25 (1897) 243-245).

17 The prosody of line 1442 seems rather primitive but may reflect, in a
so~histicated way, the primitive stage of poetry described.
8 Cf. P. H. Schrijvers, art. cit. (n.ll) 288-304.
19 These lines have not been analysed sufficiently by M.R. Gale, op.cit. 188-

189.

20 Cf. Horner, Iliad 8.13 i\ JllV UOOv piwro e~ Tapmpov ~ep6evm (cf. Hom.H.
Mere. 256, 374), Lucretius, 3.966 nee quisquam in bar at h rum nee Tartar a
deditur atra, and fliad 8.14 nJAe JlaA ', fiX! f3a8U11:0V lJTCO x8ov~ ecm f3 Ep E8 p 0 V.
21 I am following at this point the dating proposed by 0. Murray and
followed by T. Dorandi, Filodemo, Il huon re seeondo Omero (ed., trans!., cornrn,

1982).
coli. 38 [tov 0' 'A)XtAAEil Kilt KlXAAUJ'tOV, coli. 29 Kilt to ~1]AO'tll1tOV [0') llltElVIll
Oe'i ... (cf. Dorandi ad loe.)." in line 5.1120 the Latin word potentes may be a
translation of the Greek term ouvacrtllt (cf. also GJ.D. Aalders, 'The Political
Faith of Dernocritus', Mnemosyne IV 3 (1950) 304 on the Greek terms Ouvacrtllt
Na~oli

OUVIl<JtEtll).

LUCRETIUS ON THE BEGINNINGS OF POLITICAL LIFE

227

We turn now to the lines 5.1141-1160. As Lucretius tells us, the

human beings, worn out with leading a life of violence, lay faint
from their feuds and therefore gave in all the more easily, even of
their own will, to strict ordinances and laws (5.1145ff). The
institution of official, criminal justice, surveyed by a chosen
magistrate, limited the excesses of bloody vendettas and personal
revenge (cf. 1146 ex inimicitiis, 1149 ulcisa). I should like to point out
that, as Albrecht Dihle has shown in his monograph entitled Die
Goldene Regel (dealing with popular morality) ,23 Greeks and
Romans were well aware of the primitive and/ or exotic character
of the lex talionis; they generally considered the institution of
regular, criminal justice as a real progress in the history of human
civilisation. In the ethical and rhetorical thought of the Ancients
the impulsive feeling of anger is linked, as in Lucretius' text (11481149 ex ira ... ulcisci), to the act or urge of avenging an outrage, often
in a rather excessive way.2 4 Plato formulates in his Protagoras (324
b) a twofold opposition between revenge ('ttJ.uopia) and punishment
( K6A.acn~): revenge is considered an irrational and even bestial
action, punishment is or ought to be a rational and thoughtful act;
revenge is concentrated on the injustice suffered in the past,
punishment has a preventive effect for the future.25 In my opinion,
one finds this twofold opposition again in the lines of Lucretius (cf.
1151 metus poenarum) although he adds that the aspect of revenge is
not completely absent in official, criminal justice (cf. 1153
'violence and hurt for the most part fall on the head of him from
whom they had their rise').
We may establish that the number of strictly Epicurean ideas in
Lucretius' story of political evolution is rather limited. With
reference to the lines 1117-22 our commentaries duly quote some
parallels with the opinions of Epicurus. But we must take into
account that the depraving effects of avidity, of money and envy
(7tAEOVE~ia, q>8ov6~) are very common themes in ancient literature
23 A. Dihle, Die Goldene Regel (Gottingen 1962) 14, note 1 and ch. 5
'Prinzipielle Ueberwindung des Vergeltungsdenkens'; cf. also the studies by
R. Hirzel, Die Talion (Philologus, Suppl.Bd. 11, 1907-1910, 407-482) and Themis,
Dike und Verwandtes, Ein Beitrag zur Geschichte der Rechtsidee bei den Griechen
(Leipzig 1907).
24 Cf. Aristotle, Rhet. 2.2.1378a31 "Ecrtm oit opril opE~u; !LEta A:llltTt<; T:!J.lmp{a;.
25 OUOEt<; yap KONX~El toil.; aOtKOUVtll<; ltpo.; toutcp tov vouv exmv Kill tOUtOU EvEKil,
Ott TJOtKTtO'EV' OO'tt<; lilt OlcrltEp !htpfov al..oy{crtm.; T:lJ.liDPEl'T:al. b OE !lEta A.Oyou EltlXEtp&v
K'OAa~ElV ... tOU !iElloVtO<; xaptv, cf. Seneca, De ira 1.16.21 nemo pru.dens punit quia
peccatum est, sed ne peccetur.

228

P.H. SCHRIJVERS

and moral philosophy. At the end of Lucretius' story one finds

again the Epicurean notion of 'treaty' (cruv8f)Jcrt, 1155 foedera), qualifying in general the institution of law, civil justice and magistrates. From line 1154 onward Lucretius repeats Epicurus' opinion
according to which it is not easy for one who by his act breaks the
common pact of peace, to lead a calm and quiet life.2 6 Probably,
under the influence of the reproaches made by the Stoics-they
accused the Epicureans of committing or approving crimes as long
as they were sure to escape criminal justice2'-Lucretius reinforces
Epicurus' opinion (as far as we know from the fragmentary
tradition) by referring to a common experience which he had
mentioned already in his explanation of dreams: men often betray
themselves by talking in their dreams (4.1018-1019).
The lines 1156ff. mention the notice of gods and men which a
criminal might escape (etsi fallit enim divum genus humanumque... ).
This reference to the gods, coming from an Epicurean poet, need
not surprise us;2 8 Lucretius' argument has a narrative form in
which a change of viewpoint (in modern terminology: of the socalled focalisation) is quite normal. Of course it is the criminal
who hopes the gods will ignore him. By means of this reference
Lucretius succeeds in connecting the paragraph on political
change to the next one dealing with the origin of belief in gods. In
this new argument topics from his political story return: e.g. the
fear of the gods inspired in proud kings by their bad conscience
(1122-1125), and,-as Lucretius explains at the opening of the
paragraph on religion-men saw in their dreams the glorious
shapes of beings, who-like the first human aristocrats-possessed
noble beauty and ample strength. I should like to point out that this
implied similarity between gods and aristocrats ('godlike men')
has been mentioned explicitly by Philodemus in his treatise On the
Good King according to Homer (cf. the paraphrase of the columns 3739 given by the most recent editor T. Dorandi). I may add that,
syntactically and semantically, line 1156 etsi Jallit ... is not an
everyday expression in Latin-as Costa claims in his comment-,
but has a definite Greek flavour: the use of fallit with a personal
Cf. Bailey p. 1504, Costa p. 129.
27 Cf. the discussion by P.A. Van der Waerdt, 'The Justice of the Epicurean
Wise Man', Classical Quarterly 37 (1987) 409ff.
2S As it did Bailey (see his comment on this line). Lucretius' line may
allude to Hesiod, Op. 265-268 (well known in Antiquity), cf. 268 OUOE A.ft8Et on
the all-knowing and all-seeing God.
26

LUCRETIUS ON THE BEGINNINGS OF POLITICAL LIFE

229

subject and with the gods as object, in the sense of 'being ignored,
escape', is quite unusual in Latin literature and has been built on
the regular use of the Greek verb A.avSavnv ('touc; 8w6c; I 'tOV 8e6v).
By means of this analysis of the paragraph concerning the
origins and the evolution of political life I hope to have shown how
much Lucretius has drawn from an ensemble of common ideas
which were rather widespread in later hellenistic times among
Greek and Roman intellectuals who belonged officially to different philosophical schools. Like Philodemus, Lucretius tried to integrate Epicureanism in the paideia, in the general civilisation of his
times. Leaving aside the relationship between Lucretius and Philodemus, I am not able to point out the direct sources of Lucretius'
argument. Therefore, I used expressions like 'the ensemble of common ideas' or 'ancient thought in general' .29 Methodologically, it
is very important to make a sharp distinction between sources/
ideas attested before or only after Epicurus' lifetime,3 although the
very fragmentary state of the transmission of Epicurus' works
makes the problem of Lucretius' factual dependency on Epicurus
rather intractable. In any case, these direct of indirect sources were
Greek.
Although Lucretius' reconstruction of political change is wholly
anonymous, without any precise chronology or localisation,-and
so without any explicit trace of ethnocentrism. But this impression
is misleading. His theoretical analysis has been inspired by real,
factual history which certainly is not Roman. The change from
kingship to democratic republic by the way of a chaotic, ochlocratic interval is without any parallel in Roman history. Moreover,
Lucretius would have been blatantly anachronistic if, in his reconstruction of the origins of political life, he had referred to Roman
history, the development of which is clearly later in time than
Greek history. Of course, some lines that contain a more generalised moral message about jealousy and wealth may have been
applied to the Roman, late-republican situation (the poet said himself that at a more general level history repeats itself, 5. 1135 nee

29 Cf. the terms 'allgemeines Bildungsgut/ Kulturgut' in W. Spoerri's

monograph Spiithellenistische Berichte uber Welt, Kultur und Cotter (Schweiz.
Beitr. zur Altertumswiss. 9, Basel 1959).
30 Cf. my discussion in 'Philosophie et Paraphrase, Lucrece et les
Sceptiques', in La Langue Latine, langue de la philosophie (Coli. Ecole Fran~:aise de
Rome 161, Rome 1992) 125-140.

230

P .H. SCHRIJVERS

magis id nunc est neque crit mox quam fuit ante) ,31 but chronologically
this history did not start in Rome. Content and style of the first part
of Lucretius' story have more than once a Greek flavour as if the
period of the Homeric kings and their mutual envy hides itself
under the surface of the text. Subsequently, aristocrats, oligarchs,
great legislators who bring the internal troubles and vendettas to a
close, in general tend to evoke events in Greek history: the
legislations introduced by a Solon of Athens or a Dracon, the
bloody revenge, subject of many a Greek tragedy about the House
of Atreus. The cities mentioned in the beginning seem to be Greek
poleis, not Roman urbes which in fact did not yet exist in these
earlier times.
In spite of many a reference I made to the Greek historian
Polybius I do not think that he was a primary source for Lucretius;
this Greek historian combined the doctrine of political change
with the concept of a cyclic return (anakyklosis) which is not
emphasized in Lucretius' story (it is only mentioned in 5.1135). I
think the underlying presence of Greek history and the theory of
J..LE'ta~oA.i't 7tOAt'tet&v point to another Greek scholar, a Peripatetic
who influenced also Lucretius' description of the savage and antisocial life of primitive man (as I tried to prove in my recent article
in Philologus) ,32 a scholar who, in Roman, republican times, was
known as the author of a rather famous Archeology: the first chapters of his Life of Greece; this Greek scholar did make use of poetical
testimonies, had a reputation as a historical critic of the Homeric
epics and was rather well-known for his treatise entitled Tripolitikos
dealing with the three principal forms of government: in short,
once again I am persuaded that Lucretius' story might be influenced-directly or indirectly-by Dicearchus of Messene.33

Leiden
31 In a more general and allusive way, 'Lucretius' model for such a
constitutional development in some ways resembles that of Rome' (Costa, p.
128), and 'superbus was the regular Latin epithet for the tyrant; cf. Tarquinius
Superbus' (Bailey ad 5.1137); for more allusions to the contemporaneous,
Roman situation see D.P. Fowler, 'Lucretius and Politics', in M. Griffin - J.
Barnes (ed.), Philosophia Togata, Essays on Philosophy and Roman Society (Oxford
1989) 120-150.
32 Cited above in note 11.
33 Cf. the collection of testimonies by F.Wehrli (Die Schule des Aristoteles,
Heft 1, Dikaiarchos), the subdivisions of which obscure somewhat the unity of
Dicaearchus' philosophical and scientific activities.

PHILOSOPHICAL INNOVATION IN LUCRETIUS?

H. B.

GOTTSCHALK

There are various ways in which a philosophical or scientific

system-! am using this word loosely, to denote any body of
beliefs felt by its adherents to be connected and interdependentcan be modified. Individual doctrines or theorems can be altered
or abandoned in order to meet criticism or to assimilate new data
or problems which may have arisen after the system was first
formulated; such changes are obvious and relatively easy to
explain. Or there may have been less palpable changes of attitude
which leave the official, written doctrines unaltered but make their
adherents approach them in new ways; these are manifested as
changes in the way in which the old doctrines are stated, changes
of emphasis and balance and, where they concern ethical beliefs,
may lead to changes of behaviour. Most of the changes undergone
by Greek philosophy in the Roman world are of the second kind,
and most of this paper will be concerned with them. Epicureans, in
particular, were reluctant to criticise or tamper with the doctrinal
legacy contained in Epicurus' writings, and Lucretius wanted
nothing more than to be a faithful mouthpiece of his master's
gospel (DRN 3.3). But first I want to discuss one purely doctrinal
modification, partly in order to illustrate the distinction I have just
drawn, but also because it may shed some light on the way in
which Epicurus' teaching was taken up by his followers and on
the kind of sources Lucretius may have used.
In the course of his description of the soul ( 3.230 ff.), Lucretius
tells us that it is composed of four substances: aura or ventus (282),
terms which represent the Greek 7tVEUJ.UX and should be translated
as 'breath' (so M.F. Smith) rather than 'wind'; vapor or calor ('heat',
'tO 8epJ.L6v); aer; and the mysterious fourth, unnamed substance
which is finer than any of the other three and the source of those
movements giving rise to sensation and thought. These four
substances combine to form a single, unified soul, but each retains
its own power within the whole, and at 3.282 ff. the first three are
said to produce different emotional states (7ta81,): heat, anger; breath

232

H. B. GOTTSCHALK

or wind, fear; and air, placidity. The same fundamental doctrine is

found in the doxographical tradition represented for us by Aetius
(ps-Plutarch, Stobaeus and Theodoretus) and in Plutarch's Ad.
Colotem. 1 We can safely say that it was accepted as standard by
Epicureans from at least the first century BC.
However, when we turn to Epicurus' own writings, we find
something different. In his Letter to Herodotus (63), he says that the
soul is composed not of four but of three substances: it is a 'fine body
... most like wind or breath (1tVEUj.La) with a certain admixture of
heat , in some respects like the one, in some like the other; and
there is a third component far exceeding even these in fineness .. .'
This last is clearly the same as Lucretius' fourth substance, but
instead of three there are only two 'earthly' substances, pneuma and
the 'hot stuff. Together these add up to something very like the
Stoic concept of pneuma or, for that matter, the Holy Spirit of the
Pentecost miracle which manifested itself first as the sound of
rushing wind and then as tongues of fire. These two qualities were
regularly associated with the 'life-giving spirit' in the late classical
and Hellenistic periods.
Comparing this with Lucretius' account, we can see that he has
gained his extra component by treating 'air' as something different
from pneuma, although he defines pneuma (here meaning 'wind')
as 'air set in motion' in the section of his work dealing with meteorology (6.685). The question now is, what caused him to make this
change. While the view that pneuma is the carrier of sensation and
other vital functions was widely accepted in the Hellenistic era,
the precise nature of this substance remained an open question.
Aristotle's remarks on this topic were unclear and those of Epicurus studiously vague, as we have seen. Later researchers, medical
men even more than philosophers, tried to solve this problem. It
seemed fairly obvious that the material basis of the vital pneuma
must be the air we take in when we breathe, but that the two were
not identical; the air had to undergo some kind or processing or
'concoction' to turn it into pneuma. One anonymous third-century
writer, a Peripatetic with a strong interest in medicine, tried to
describe the difference by saying that the outer air is 'tame' (1tpauc;)
and becomes pneuma after being taken into the organism.2 Others
I Aetius 4.3.11, DG 388; Plut. Adv.Col. 1118e = Usener, Epicurea 314-5.
2 Ps.-Arist.Spir. 483a33 1:0 yap oA.ov [sc. 7tVEUJ.Ul ] o{Hc a~p, aA./Jx. O"Uil~llAAOilEVOV
'tl 7tp0<; 'tllll't'I]V 'tltv OUV!llltV b a~p. b6 ~ro ~yap 7tpllU<;, E!11tEptATJ<p9d<; OE 7tVEU!lll.

PHILOSOPHICAL INNOVATION IN LUCRETIUS?

233

believed that there might be several kinds of pneuma, of different

degrees of purity and vitality, in the same organism. It is against
this background that we must see Lucretius' innovation. Probably it
originated as an attempt to update Epicurus' teaching by an
adherent with medical interests, perhaps a professional physician;
since the change does not affect any fundamental philosophical
issue, it was not seen as a threat to the system, and it may even
have been sanctioned by Epicurus himself. As for Lucretius, there
is no need to suppose that he turned to a different source for this
part of his poem from the one he followed elsewhere; 3 he could
have used a copy of some work of Epicurus with annotations like
those found in the extant version of Epicurus' Letters. There is only
one feature of Lucretius' account still needing to be explained: he
connects heat with aer, not with pneuma. This is not found in the
doxographical parallels and runs counter to the views of all other
Hellenistic writers on the subject I know, including Epicurus. I
suspect that here Lucretius simply made a mistake.
The other kind of change I mentioned is much harder to pin
down, but every reader can feel it immediately; change not of
doctrine, but of the way the doctrine is understood and of the
responses it arouses. They are not the result of philosophical debate,
but of differences of temperament and circumstance, and often
throw more light on the character of individual thinkers than on
the philosophies they profess or the arguments used to underpin it.
Changes of this kind affect Lucretius' reading of every part of
Epicurus' philosophy. In an interesting article published last year,
K. Summers has shown that Lucretius' attitude to the religion of his
day was much more hostile than that of Epicurus. 4 But it is a fact
that their teaching about the gods and their mode of existence is
identical. The difference lies in their responses to that dogma and
must be due to a difference of temperament or perhaps to differences between the religious practices of the societies in which
they lived. Yet it remains true that Lucretius' attitude was different
from his master's, and on one occasion at least this bias led Lucretius to make a philosophical blunder which Epicurus avoided. Both
begin their account of the natural world by laying down that
nothing comes out of nothing (Ep.Herod. 38, DRN 1.150). This is a
Heinze and Bailey (Commentary p. 1025) do.
Summers, 'Lucretius and the Epicurean tradition of piety', Classical
Philology90 (1995) 32-57.
3 As
4 K.

234

H. B. GOTTSCHALK

fundamental principle which, until the advent of quantum mechanics, was regarded as the basis of all science and philosophy.
Epicurus states it without embellishment and tries to establish its
truth by induction-if it were not true, anything could spring from
anything. He then goes on to establish that the only things which
exist in their own right are matter and void, and all other existing
things are derivatives or accidents of them. Lucretius clearly
means to reproduce Epicurus' argument, but vitiates it by adding
one word to his premiss: nullam rem e nilo gigni divinitus umquam. In
this form, his premiss loses its universality and can no longer be
inferred from the inductive evidence he adduces; for God is not
'nothing', and the fact that particular things are always seen to
come from other particular things, does not prove that there is not a
deity (instead of another particular thing) at the beginning of the
chain. Moreover, since his avowed purpose is to disprove divine
creation ( 1.159), the presence of this word in his premiss makes
the argument circular. Lucretius' eagerness to give his demonstration an extra anti-religious twist has betrayed him into formulating
his initial postulate in a way that renders it invalid.
Lucretius' view of Roman religion could be regarded as an
intensification of Epicurus' attitude. In other ways, the psychological differences between them are more marked, and some of
them touch the very roots of their philosophy. For example,
Lucretius describes the contemplation of nature as thrilling and a
source of god-like pleasure:5
His ibi me rebus quaedam divina voluptas
percipit atque horror, quod sic natura tua vi
tam manifesta patens ex omni parte retecta est.

And he pictured Epicurus as a heroic figure, a victorious fighter for

the truth who burst out of the confines of the world to slay the
dragon Religio by the force and vigour of his mind:6
Ergo vivida vis animi pervicit, et extra
processit longe Jlammantia moenia mundi ...
quare religio pedibus subjecta vicissim
obteritur, nos exaequat victoria caelo.
5 3.28. The expression has a Greek precedent, as well as the Latin ones
adduced by Bailey and the other commentators: Pindar Nem. 1.56 ecrta o
6aJ.1J3et oump6pcp tep1tvcp te fltxSei~. of Arnphytrio on seeing the serpents killed by
the infant Herakles.
6 1.72 ff. Note the use of vis in both places.

PHILOSOPHICAL INNOVATION IN LUCRETIUS?

235

But this is far from the impression we get from Epicurus' own
writings. He holds that the study of nature would be superfluous if
we could banish our fears without it (KD 11); in the Letter to
Herodotus he insists that its only purpose is to obtain a calm life
(EyyaJ.:rtvf~etV, 37) and warns his readers to confine themselves to
the study of general principles and not to delve too deeply into
details, as this might cause them amazement (8aJ.tBo~) if they met
problems they cold not solve (79). For Epicurus, it seems, horror
was not compatible with pleasure. Thus we are faced with two
opposed attitudes to the purpose of natural philosophy: on one side,
a zestful pursuit of a vision which will reveal the truth about the
universe and free men from the terrors of superstition, on the
other, a purely defensive search for 'scientific' support for a rather
timid ideology. One reason, perhaps, why Lucretius' poem is great
literature and Epicurus' works are not. Some of Lucretius' most
memorable lines are those which state the terrifying aspects of his
vision in the most dramatic terms:
Principio mare ac tnras caelumque tune:
quorum naturam triplicem, tria corpora, Memmi,
una dies dabit exitio multosque per annos
sustentata ruet moles et machina mundi. (5.91 ff.)

Contrast this with the anodyne, matter-of-fact statement in the Letter

to Herodotus:
And in addition to what we have already said, we must believe that
worlds, and indeed every limited compound body which continuously exhibits a similar appearance to the things we see, were
created from the infinite, and that all such things, greater and less
alike, were separated off from individual agglomerations of matter;
and that all are again dissolved, some more quickly, some more
slowly, some suffering from one set of causes, others from another.
(73, tr. Bailey)

This is the kind of thing we might expect from a government

spokesman after some embarrassing disaster.
We can also observe a difference between the attitudes of Lucretius and Epicurus to public life. Epicurus taught that the tranquillity
his philosophy aimed at could only be achieved by retiring from
the world and living quietly in a community of friends (KD 14).
This doctrine was hotly debated in the Hellenistic era; most Stoics
took the opposite view and there were many professed Epicureans,
Greeks as well as Romans, who did not follow Epicurus' precept.

236

H. B. GOTTSCHALK

Lucretius did, and in several places he warns his readers against

seeking salvation in political or military power: this will not, he
emphasises, free men from their fears of death or the gods, but
rather induce new fears and cut off the possibility of attaining the
peace of mind they need for happiness. 7 Nevertheless, he seems to
have had some reservations. In his prayer to Venus (1.42 ff.) he
asks for peace because neither he nor Memmius could be happy
doing philosophy at a time when their country seemed to need
their active service; this in an age when even the Stoic Athenodorus recommended that those who found the times too hard
should retire from political involvement.S The attitude implied by
Lucretius' prayer is typically Roman; it has a parallel in a remark
of Cicero's, that the poet Archias came to Italy at a time when the
arts were cultivated in Rome 'because the Commonwealth was at
peace'. 9 While Lucretius regularly preaches against greed and
luxury, he nowhere urges those of his contemporaries who felt
they had a duty to participate in public life, to abstain from doing
so . 10 His philosophy could not altogether prevail against the
instinct of a Roman aristocrat. No doubt he would have justified his
own abstention from politics by the argument, used by Athenodorus and later Cicero (Div. 2.1), that his writing was the greatest
benefit he could confer on his fellow-Romans.
Apart from this, Lucretius' teaching on ethics, natural philosophy and, in so far as he deals with it, epistemology, is completely
orthodox; he had neither the desire nor the ability to make any
original contributions here. There are un-Epicurean features, notably a tendency to personifY nature manifested in such expressions
as natura creatix or natura daedala rerum or even tanta stat praedita culpa
7 2.37 ff., 5.1125 ff. I am grateful to Monica Gale for drawing my attention
to the second passage, which I omitted in the first version of this paper, but
would interpret it in a different way from hers; see n.10 below.
8 Athenodorus ap. Sen. Tranq. An. 3. See further M. Pohlenz, Die Stoa,
(Gottingen 1948) 1.242, 2.124 f.; idem, Kleine Schriften (Gottingen 1956) 1.287 ff.
9 Cic. Pro Archia 3.5: Erat Italia tum plena Graecarum artium ac disciplinarum,
studiaque haec . . . et hie Romae propter tranquillitatem rei publicae non neglegebantur.
10 Monica Gale has suggested that 5.1129 ff. contains such a recommendation, but here, as in bk. 2, Lucretius only warns against ambition because of
the invidia it arouses and recommends the life of a private citizen as a means
of attaining tranquillity: ut satius multo iam sit parere quietum quam regere imperio
... (note the emphatic position of quietum at the end of the line and the
clause). This is compatible with the position adopted at 1.41 ff., that some may
feel an obligation to sacrifice their immediate peace of mind in order to
benefit their community by public service.

PHILOSOPHICAL INNOVATION IN LUCRETIUS?

237

[natura]. But these do not lead him to depart from a strictly mechanistic account of causation; they do not signify a modification of
Epicurus' doctrine but are part of the poetical form Lucretius has
given to his exposition. However, this brings us to the greatest
departure from his master's teaching. In the 'ancient quarrel
between poetry and philosophy' Epicurus was as uncompromising
as Plato. He warns his followers against the snares of poetry, with
its tales of a world controlled by jealous deities and of a dire
existence after death, and tells them to avoid all 'culture' . 11 But
Lucretius, author of the fullest exposition of Epicurean philosophy
that has come down to us, made a poem of it, a poem, moreover,
which is no dry piece of versified argument like the poem of
Parmenides, but used all the devices of high poetry and openly
acknowledged its debt to Empedocles, a myth-maker if ever there
was one. The paradox is pointed up by a passage where Lucretius
warns Memmius not to be put off by vatum terriloquis dictis (1.102 f.).
It has become customary to translate vatum here by 'bards' or even
'priests', 12 but this is mistaken. There are several passages where
Epicurus advises his followers not to be put off by fears arising from
11u8ot; 13 the writers of 'myths' were poets, and when Lucretius goes
on to give examples of the 'terrifying sayings' against which he
wants to warn Memmius, he refers to accounts of the underworld
given by Ennius (1.115 ff.). The term vates is ambiguous, and
Lucretius may have chosen it in order to give an anti-religious
twist to Epicurus' warning; it would then be a parallel to the insertion of divinitus in line 150, with less disastrous consequences. 14 But
the immediate reference is to the stories of poetry, and the paradox
of a poet warning his readers against poetry is something we
should embrace, not fudge. Lucretius did not regard himself as a
11 Ap. Diog. Laert. 10.6 = fr. 163 Usener. natoda includes all kinds of nonphilosophical literary activity.
12 'Bards': Bailey; 'seers': Munro; 'priests': Diels, M.F. Smith. But Knebel
(18312) and Binder (1868) have Dichter.
13 Ep. Her. 81, KD 12. Note that the Greeks mostly speak of J.I.U8ot in the
plural and never use the singular as a collective noun to denote a privileged
group of fictions endowed with some kind of mystical collective significance.
To speak of Lucretius', or any other ancient writer's, use of or attitude to
'myth' (or der Mythos) in this sense, is at best anachronistic and could be
seriously misleading.
14 Another reason for his choice may have been that vates could have a
contemptuous meaning in Lucretius' time (see Munro and Bailey ad loc.). But
this would not fit well with his immediately citing Ennius, in most honorific terms.

238

H. B. GOTTSCHALK

poet of quite the same kind as the rest: when staking his claim to
poetic fame, his first boast was that he 'taught about great issues and
freed minds from the bonds of religion', his second that he 'wrote
clear poetry about an obscure subject'. The poetic quality of his
work, musaeo contingens cuncta lepore, comes last and is justified by
the famous image of the honey on the cup of bitter medicirte. 15
This raises the question of the relationship between the form
and content of his work. There was a time when it was thought
that he began by wanting to write a poem, looked for a subject and
chose Epicurus' philosophy; some nineteenth-century writers,
including Macaulay and Mommsen, even suggested that his
choice was a mistake and complain that he wasted so much poetic
talent on an inherently dull subject, while Munro devotes a whole
page of rhetoric to demonstrating that Epicureanism is a better
subject for poetry than any of the alternative systems available to
Lucretius. 16 Of course this is arrant nonsense. If one thing is clear
from his poem, it is that Lucretius decided to write about Epicureanism because he saw it as the key to the salvation of mankind. His
choice of poetry as his vehicle, in an age when prose had long
been the normal medium for serious philosophical writing, must
mean that, in his own situation, it seemed to offer some palpable
advantage over prose. Our task is to determine what this advantage
may have been.
Lucretius' purpose was to write in a way that would attract the
audience he was aiming at, the narrow, highly sophisticated
circle of the Roman upper class in which he and Memmius, the
dedicatee of his poem, moved. Now prose treatises on philosophy
or other instructional subjects seem to have had a low literary
status, rather like textbooks or what the trade calls 'academic books'
to-day, and according to Cicero, the only Latin prose works on
philosophy produced up to that time, by the Epicureans Catius and
Amafinius, were too shoddily written to appeal to sophisticated
readers, although he admits that they had a wide popular following.17 Those philosophers who aspired to be men of letters wrote
dialogues or open addresses to named individuals, which could be
15 1.93 ff, = 4.6 ff. Clarity (cr!X<JlllVElll) was the one quality on which Epicurus
insisted in his Rhetoric (ap. Diog. Laert. 1013).
16 Th. Mommsen, Romische Geschichte (Berlin 18827) 3.595 f. H.AJ. Munro,
in the introduction to vol. II of his edition (London 19004 ) 5 ff.
1 7 Cic. Ad Jam. 15.16, 19; Tusc. 4.3.6, Acad. 1.3.6. But Quintilian (Inst.
10.1.124) describes Catius as levis quidem sed non iniucundus auctar.

PHILOSOPHICAL INNOVATION IN LUCRETIUS?

239

taken as 'partial dialogues'. Didactic poetry, on the other hand, was

a flourishing genre with a long and honourable history; in Rome,
it had been practised by eminent poets and versifiers from Ennius
to Cicero. Lucretius could have written a dialogue or series of
dialogues, like Cicero a few years later, but they were not trying to
do the same thing. Each of Cicero's works is centred on a problem
or set of connected problems, and Cicero's purpose was to confront
the views about it of all the important schools; this was something
for which the dialogue form was eminently suitable, but it involved a good deal of selection and re-arrangement of his material,
and some distortion was inevitable. Lucretius wanted to give an
account of one system of philosophy, as complete and accurate as
he could make it. The Epicurean writings he followed were very
much in the textbook tradition, consisting of systematic, point-bypoint exposition with very few literary graces. As material for an
urbane dialogue they were unpromising, but since the basic structure of a didactic poem is straight narrative, they could fit into that
almost paragraph by paragraph, as a comparison between the DRN
and Epicurus' letters shows. Their monotony could be relieved by
alternating didactic with descriptive passages, the use of illustrative
similes and variations in stylistic levels, devices which would
have seemed merely precious in a prose work.
In one respect at least, Epicurus' work lent itself particularly well
to this treatment. Similes played a prominent part in epic and
didactic poetry from Homer onwards, but Empedocles had put
them to new use.1 8 Instead of treating them as mere decorations, he
used them as analogies to illustrate and explain phenomena which
could not be directly observed. Aristotle and his followers had
doubts about the probative value of this method, but Epicurus
sanctioned it.19 Lucretius exploited this opening to the full. He
expanded Epicurus' analogies into splendid word-pictures to provide relief for the more technical passages, but they are integrated
into his argument, not otiose distractions. Thus he transforms
Epicurus' exposition while remaining true to his intentions.
Naturally the form of Lucretius' epic had some effect on its content. Aristotle's complaint that 'nature' seemed to mean little more
for Democritus than 'what always happens', can be transferred to
18 Cf. 0. Regenbogen, 'Eine Forschungsmethode antiker Naturwisenschaft', in Kleine Schriften, ed. F. Dirlmeier (Miinchen 1961) 156 ff.
19 Ap. Diog. Laert. 10.32, etc.

240

H. B. GOTTSCHALK

Epicurus without much injustice, but Lucretius personifies her as

'creator' and 'maker of things', as we have seen. Empedocles
would seem to have influenced his outlook as well as his poetic
technique. But even Epicurus, in an expansive moment, was able to
'thank blessed nature for making the necessities of life easy to
obtain' .2 In spite of this, nature never became an independent
force, but remained an epiphenomenon of the mechanical interaction of atoms.
To conclude, the DRN contains one substantive but minor
modification of Epicurus' original teaching, almost certainly not
the work of Lucretius himself; occasional mistakes which we can
attribute, with a high degree of probability, to the poet; and an extra
anti-religious twist which did not entail any doctrinal change.
Lucretius' poetical vision transcends the mechanical character of
his system and almost brings it alive, while his hero-worship of
Epicurus introduces a personal element into his poem not found in
Epicurus' treatises. But this is something that goes beyond his
philosophy.2I
University of Leeds

Fr. 469 Usener.

A version of this paper was read at a meeting of the Leeds International Latin Seminar, on the theme 'Roman poetry and the philosophy of
its time', February, 1996, and I am grateful to the participants, especially
Monica Gale, for their comments. Needless to say, I am solely responsible for
the views and any errors it contains.
20
21

PART FIVE

IMPERIAL PHILOSOPHY

L' OIKEIOSIS SOCIALE CHEZ EPICTETE

BRAD INWOOD

Il y a peu de philologues qui ont fait autant que Jaap Mansfeld pour
traverser les frontieres nationales et linguistiques du monde savant. En
hommage a cet esprit d 'internationalisme, j 'ose offrir la petite discussion qui
suit, convenable (je l'espere) cette occasion et par son theme et par le fait
que celle-ci represente la premiere fois que j'ai pu offrir une communication
dans une langue itrangere. I

Le monde socratique peut etre per~u comme etant un monde dur et

egoiste. En vertu de la prescription du 'soin de l'ame', le soi est au
centre de l'univers moral de chacun, ce qui rend problematique
nos relations avec les autres. Ceci est clair dans un passage du
Gorgias qui deconcerte souvent ceux qui l'aborde pour la premiere
fois. A 458a, Socrate dit qu' il vaut mieux etre refute soi-meme que
refuter quelqu' un d'autre dans la mesure oii 'il est plus avantageux
pour un homme d'etre delivre du plus grand des maux que d'en
delivrer autrui'. Quel signe plus clair que l'ethique socratique
accorde la plus grande importance au soin de soi-meme? Et pourtant c'est Socrate lui-meme qui a defendu l'idee-jusqu' ala mort,
en fait-qu' il vaut mieux souffir une injustice qu' en infliger une.
Quel signe plus clair pourrait-on trouver a l' effet que Socrate prend
comme point de depart un respect fondamental pour autrui?
Voila un probleme bien connu, qui se presente a maintes
reprises dans la tradition socratique. En effet, il est possible que ce
probleme aide a expliquer la mode actuelle parmi les philosophes
pour les etudes de la qnA.ia chez Platon et Aristote. De toute fa~on, il
y a dans le stoicisme une tension entre les vertus orientees vers soimeme qui sont fondamentales dans toute ethique eudemoniste, et
les vertus orientees vers autrui qui jouent un role significatif dans
les theories antiques de la moralite. Cette tension devient plus aigue
avec la doctrine stoicienne de l' oikeiosis. Un ecart peut toujours se
former entre la conception de l'homme en tant qu' animal qui part
1 Je dois remercier Carlos Levy pour !'encouragement qu' il m' a donne et
Jean Baillargeon, qui a traduit le texte anglais. Partout j'emploie les traductions 'Bude' des textes anciens, tout en modifiant legerement celles de
Souilhe pour Epictete.

244

BRAD INWOOD

d'une tendance a la preservation de soi, et la conception de

l'homme en tant qu' animal social. Dans plusieurs contextes on
peut fermer l' ecart, et pour certains auteurs ce probleme est peripherique ou meme n'existe pas. Mais a mon avis, ceci n'est pas le
cas pour Epictete. 11 insiste categoriquement sur la nature irreductiblement sociale de l'homme. Ceci est vrai, meme si on voit
regulierement chez lui une certaine durete et un manque de consideration envers les autres etres humains, leur bien-etre et leurs
interets. La tendance de se soucier premierement de soi-meme que
Socrate manifeste dans le Gorgias se retrouve de maniere prononcee
a travers les entretiens compiles par Arrian, et pourtant Epictete ne
cesse jamais de proclamer la fraternite des hommes. 11 est evident
que la nature des liens sociaux est ala fois interessante et profondement problematique pour Epictete. La solution aux problemes
souleves par la nature des liens sociaux est seulement implicite
dans les Entretiens, et cependant je crois qu' on peut l'articuler de
sorte a montrer a la fois que sa 'theorie sociale' s'accorde avec le
stoicisme plus ancien, et que Epictete apporte sa propre contribution
a ces problemes. Epictete accepte avec enthousiasme la theorie
stoicienne traditionnelle des valeurs, et approuve sans reserve la
cosmologie theologique stoicienne. Ce sont la les deux elements de
base de sa theorie sociale.
Je vais commencer par donner un apeq;u de !'evidence qui
montre que les relations sociales presentaient beaucoup d'interet
pour Epictete. 11 ecrit a maintes reprises que 'tO 1ttcr't6v et 'tO aiOilJlov
sont indispensables pour notre bien-etre, et ces deux traits de
caractere sont essentiellement sociaux: c'est seulement a l'interieur
d'une matrice de relations et de conventions sociales que l'on peut
comprendre ces traits de caracteres. On trouve une illustration de la
dimension sociale de 'tO mcr't6v et 'tO aiOilJlOV dans Entretien 1.5.3-5.
Une condition pathologique que Epictete appelle 'petrification' ou
'necrose' est consideree en rapport avec les deux aspects de la personne, le corps et l'ame; eta l'interieur de l'ame, une autre division
est consideree, so it la division entre !'intellect ('tO VOTj'ttKOV) et le
sens de la honte ('tO ev'tpE1tnK6v). Epictete emploie la strategie
socratique/ platonicienne (qui consiste a exploiter le contraste entre
le corps et l'ame) de sorte a produire un bon resultat rhetorique.
Epictete met ensuite en contraste notre fonction purement intellectuelle (en vertu de laquelle no us pouvons com prendre ou suivre
un argument) avec notre sens social: nous pensons que la faiblesse

L' OIKEIOSIS SOCIALE CHEZ EPICTETE

245

intellectuelle est un signe de misere (KalCOO~ EXEtV), mais nous

considerons aussi Ia 'necrose' sociale, ou l'insensibilite, comme un
indice de force de caractere. II n'y a pas de doute que le contexte de
ce passage est polemique, mais il est quand meme certain que les
relations sociales sont presentees comme etant tout aussi importantes pour le bien-etre humain que Ia sante physique ou intellectuelle. Ces termes ('to mcr't6v et 'tO ai8i1J,wv) reviennent souvent.2 En
particulier, ils figurent de fa~on saisissante dans 2.4.1, ou Epictete
maintient que l'homme est fidele par nature, que Ia 7ttcrn<; est un
trait caracteristique des etres humains. De Ia meme fa~on, dans
2.10.22-3 il maintient categoriquement que nos sens de Ia honte et
du respect, notre sens de Ia fidelite, notre tendance affective,3 notre
disposition a aider les autres et notre esprit de retenue sont tous des
traits naturels, et que Ia faiblesse de ces traits constituent une
penalite et un mal. Nos relations sociales ont une importance qui
gouverne meme nos habitudes d'hygiene personnelle (4.11.1418)-c'est en vertu de nos liens sociaux fondamentaux que Ia vie
cynique n' est pas Ia meilleure vie pour to us les etres humains,
mais seulement pour une minorite d'entre eux. Nos instincts
sociaux sont, comme d'autres stoiciens l'ont affirme, une partie de
notre nature sans laquelle nous ne pouvons realiser notre nature et
atteindre le bonheur.
Epictete presente regulierement des arguments contre ce qu' il
considere comme les implications anti-sociales de Ia pensee epicurienne. Dans 2.20, 3. 7, 1.23 et ailleurs Epicure devient Ia cible de
Ia langue hargneuse de Epictete, simplement parce que sa theorie
est incapable d'appuyer ou d'expliquer !'importance evidente des
relations sociales qui sont au coeur du bien-etre humain. Un
troisieme indicateur de Ia tres grande importance naturelle des
relations sociales est l'emploi que fait Epictete des crxcret<;. Ces
relations-avec les dieux, avec ses parents, avec ses freres et soeurs,
ses enfants, et ses concitoyens-constituent Ia matiere du deuxieme
2 Mais il faut noter 1.4.18-20 ou Epictete dit qu' un resultat du retrait des
choses externelles est Ia preservation de to 1tt<Jt6v et to uioi111ov. Si Epic tete
n'utilise pas 'choses externelles' dans un sens tres etroit ici, ce passage met en
relief Ia tension entre les relations sociales positives et le retrait en soimeme.
3 to <JtEpKttKOV, probablement Ia meme chose que Ia qnA.ocrtopyia de 1.11, ou
Epictete donne un argument assez long a l'appui de !'idee que Ia qnA.ocrtopyia
parentale est un de nos dons naturels. II est a noter aussi qu' a 3.7.27 Ia tendance humaine de rougir est offerte comme preuve que le sens de to aioi111ov
est nature! pour les etres humains.

246

BRAD INWOOD

champ d'etudes (3.2.4). De tels crx;ecret~ sont sous Ia supervision de

Zeus (3.11.6); elles sont indispensables pour Ia bonne conduite
(4.6.26, cf. 4.12.16-17); on lit dans le Enchiridion que nos Ka8i)Kov'ta
sont 'mesures' (7tap<XJ.1E'tpEt'tat) par nos relations (Ench. 30). Dans
3.7.25. ss., Epictete enumere les principales responsabilites d'un
homme: l'activite civile, le mariage, elever des enfants, Ia piete,
prendre soin de ses parents, et ainsi de suite. L'Entretien 1.2 est
encore plus clair sur ce sujet: a maintes resprises dans ce texte
Epictete met l'accent sur l'idee que notre place dans un contexte
social aide a determiner le role et les devoirs que chacun de nous
doit remplir. Nous ne pourrions meme pas concevoir ce que c'est
que de faire des decisions morales serieuses si nous ne connaissions pas notre place dans la societe.
De toute evidence done, Epictete attribue une tres grande valeur
aux liens que les etres humains entretiennent en societe. Ces liens
sont naturels, surveilles par Zeus, indispensables au bien-etre
humain, et constituent la cle pour la comprehension de nos
Ka8flKov'ta. Tout comme son maitre Musonius, Epictete semble se
saucier sincerement de Ia securite et de l'integrite de !'organisation
sociale-quoiqu' il ne va peut-etre pas aussi loin que le consulaire
Ciceron. Cependant !'attitude d'Epictete n'est pas aussi simple que
cela, et son attitude envers les liens sociaux elle non plus n'est pas
aussi simple. Prenons quelques moments pour considerer les
tendances de sa pensee qui vont a l'encontre de Ia valorisation de la
dimension sociale des etres humains.
Examinons d'abord Entretien 1.1. En soulignant le fait que nous
maitrisons seulement notre vie mentale, que nous controlons nos
impressions, il se tourne dans les sections 14 ss. vers les erreurs que
nous faisons tous dans le cours de notre vie. Notre vie mentale est a
proprement parler la seule chose dont nous sommes capables de
prendre soin, et a laquelle nous pouvons nous consacrer. Neanmoins nous essayons avec insistance de nous occuper de plusieurs
choses, et par consequent nous nous donnons trop de soucis. Nous
prenons un souci pour notre corps, nos possessions materielles, notre frere, notre ami, notre enfant et notre esclave. Nos interets sont
si nombreux qu' ils nous accablent. lei, comme dans 4.1 (surtout 65
ss.), Epictete souligne que pour etre libre de tracas et pour atteindre
la tranquillite, il faut se detacher de tout sauf notre vie men tale. La
famille, les amis, les relations sociales, et meme nos propres corps
sont dans la meme categorie que nos possessions materielles: nous

L'OIKEIOSIS SOCIALE CHEZ EPICTETE

247

nous en occupons au peril de notre bonheur. Comme il le dit dans

4.1.100, les forces cosmiques ont le controle de mes possessions,
mes meubles, mes enfants, et mon epouse. C' est une erreur que de
se soucier excessivement de ces choses-la, non seulement parce
que ~a menace notre bonheur personnel, mais aussi parce qu' un
tel souci ne correspond pas a Ia volonte de Zeus. Le fondement des
ruifhl est }'interet que I' on prend pour les choses externes, y compris
un autre etre humain; et une vie tourmentee par les passions ne
peut pas etre une vie heureuse. Epictete exprime cette idee dans
3.24.58-9 en disant qu' il n'y a aucun avantage a etre affectueux
envers notre famille si cela nous cause de Ia misere. Nous devons
meme aimer nos enfants avec un certain recul, de sorte a etre
indifferent a leurs souffrance eta leur mort (3.24.84-89).
II y a un probleme semblable du point de vue de l'enfant. Je
pense ici au chaptre 30 de I' Enchiridion. Epictete insiste beaucoup
dans ce chapitre sur l'idee que nous avons une affiliation naturelle
( oh:eirocrt~) avec un pere, et non pas avec un bon pere. Notre relation avec lui-comme c'est le cas avec un frere-est naturelle et
basee fermement sur Ia oikeiosis qui nous definit en tant que
membres de l'espece humaine. Les relations sociales que nous
avons avec notre pere, notre frere, notre voisin, notre concitoyen,
ou notre officier superieur sont des relations reelles-et dans ce
passage il y a une suggestion a l'effet que nous partons de nos
relations les plus intimes pour developper nos autres relationsmais elles sont limitees par les frontieres de notre caractere moral.
Nous ne traitons pas notre semblable d'une maniere qui pourrait
compromettre notre propre bonheur. Pour Epictete, nous ne devons
rien faire, meme pour le benefice de nos relations les plus intimes,
qui puisse compromettre le benefice moral. La ~A.a~TJ socratique est
le seul mal veritable.
Epictete est done un veritable eudemoniste socratique. Socrate
pensait que c'etait un plus grand bien d'etre libere du mal que de
liberer autrui; de Ia meme veine, Epictete recommande en toute
franchise que chacun de nous devrait 'jurer de se tenir soi-meme
en plus haute estime que n'importe qui' (1.14.17). Epictete soutient,
tout comme Aristote et Platon, que cet interet personnel ne correspond pas a ce que no us appelons I' egoism e. Mais on ne peut pas
nier la preference pour le soi; voir 3.3.5:
Voila pourquoi on prefere le bien a tous les liens du sang. Mon pere
n'est rien pour moi, mais seulement le bien. 'Peux-tu etre aussi
dur?' La nature m'a fait ainsi. Telle est Ia monnaie que dieu m'a

248

BRAD INWOOD

donnee. Aussi,le bien vient-il a s'opposerau beau etjuste, tout s'en

va, pere, frere, patrie, et tout le reste. Mais quoi? vaisje, moi, negliger mon propre bien pour que tu en jouisses, et vais-je te ceder la
place? En echange de quoi? 'Je suis ton pere.' Mais non un bien ....
Et~a serait une erreur de penser qu' il s'agit d'un bien transcendant
et irnpersonel qui contraint nos obligations sociales; en effet, on lit
dans 3.4.10: 'nul ne rn'est plus cher que rnoi-rnerne'.
Une des illustrations les plus vivides de la determination de
Epictete a rninirniser !'importance des relations sociales pour le
bonheur hurnain se trouve dans un passage ou il recornrnande un
modele d'entrainernernent mental con~u pour nous aider a atteindre le bon degre de detachernent: dans 4.1.111-2 il nous conseille
de 'pratiquer ceci du rnatinjusqu' au soir: ...
Commence par les plus petites choses, par les plus fragiles, un pot,
une coupe, puis, poursuis de Ia sorte jusqu' a une tunique, a un cabot,
a un vieux cheval, a un bout de champ; de Ia, passe a toi-meme, a
ton corps, aux membres [ou parties] de ton corps, a tes enfants, a ta
femme, a tes freres. Regarde bien de toutes parts pour tout rejeter
loin de toi; purifie tes jugements pour que rien de ce qui ne
t'appartient pas ne s'attache a toi, ne fasse corps avec toi, ne te cause
de Ia souffrance, si on vient a te l'arracher. 4

Ceci constitue une inversion frappante des notions que nous

connaissons du passage farnilier de Hierocle, 5 qui preconise la
strategic opposee, selon laquelle nous devrions renforcer nos liens
avec les autres en allant de l'exterieur vers l'interieur, en apportant
chaque niveau de relations sociales plus proche du centre, c'est-adire notre arne et notre centre moral. L' acncrtcrt<; de Epictete preconise le contraire: nous ne devons pas accorder plus d'irnportance a
ceux qui nous sont les plus chers que nous n'accordons a un
marceau de vaisselle ou un vieux manteau. Nous ne trouverons
une rneilleure illustration de la profonde indifference de ce
Stoicien austere face a l' ensemble de nos liens sociaux.
En tant qu'etreshurnains, nous sornrnes inevitablernent dechires.
D'une part, notre tendance fondarnentale est de poursuivre notre
propre bien, ce qui est nous est reelernent avantageux. Nos relations
avec les choses exterieures nous rendent vulnerables, et done on ne
peut pas perrnettre qu' elles deviennent trop irnportantes. Si nous
faisons l'erreur d'accorder trop de valeur aux chases externes, y
cornpris ces etres hurnains qui nous sont les plus intirnes et dont
On peut comparer !'expression plus breve de cette idee dans Enchiridion 3.
Dans Stobee 4.671-673 (=texte 57G dans A.A. Long et D.N. Sedley, The
Hellenistic Philosophers, Cambridge 1987).
4

L' OIKEIOSIS SOCIALE CHEZ EPICTETE

249

les vies sont les plus rapprochees de Ia notre, alors nous courtisons
Ia misere. 6 D'autre part, nos relations humaines constituent une
partie vitale de ce que nous sommes en tant qu' etres humains.
C' est difficile, sin on impossible, de decouvrir ce que no us sommes
et ce que nous devons faire si nous n'avons pas une conception
adequate de nos liens naturels et indissolubles avec les autres. Les
autres etres humains ne sont pas seulement les objets de nos
affections; ils sont aussi des elements constitutifs et indispensables
de notre vie morale. II est difficile d'imaginer une vie caracterisee
par Ia sorte d'indifference aux autres preconisee par Epictete.
Je ne veux pas suggerer que Epictete n'offre aucune resolution a
Ia tension dans ses travaux que je vi ens d' esquisser si brievement. II
n'en est pas ainsi, et je veux me pencher sur une tentative de
reconcilier ces tendances opposees de Ia pensee d'Epictete.7 Mais il
est important, je crois, de partir de !'idee qu' il y a bel et bien une
tension ici, notamment parce que nous devrons eviter d'etre seduits
trop facilement par l'attrait de l'optimisme stoicien. II serait facile
de s'endormir dans une fausse securite, et de presumer qu' il est
tout simplement raisonnable d'avoir beaucoup d'interet et meme
d'inquietude pour les autres. Epictete montre que tel n'est pas le cas.
II soutient qu' un trop grand souci pour les autres est une source de
misere, et que notre interet rationnel pour notre propre bonheur
peut parfois prescrire une cruelle indifference aux autres etres
humains. ('Quel mal y a-ti'l a dire tout bas, en embrassant ton
enfant 'demain tu mourras" 3.24.88). Dans sa theorie quelque
chose d'aussi simple que !'affection d'un pere envers son fils peut
presenter un probleme theorique profond, une theorie qui montre
que nous devons souvent choisir entre nos engagements envers les
autres et notre tendance vers le bonheur.
Les moyens disponibles a Epictete pour reconcilier ces deux tendances ne lui appartiennent pas tout seul. En effet, plus nous regardons les choses de pres, plus nous verrons une continuite avec le
stoicisme ancien. Considerons d'abord Ia contribution apportee par
Ia theorie stoicienne de Ia valeur. Cette theorie est basee sur Ia distinction cardinale entre deux sortes de valeur qui sont considerees
6 Cf. B. Inwood, 'Why do fools fall in love?' Bulletin of the Institute of Classical Studies (London) 1996, et M. Nussbaum The Therapy of Desire (Princeton
1994) ch. 10.
7 Cf. A. Bonhoffer, Die Ethik des Stoikers Epiktet (Stuttgart 1894) 4-5, 158-9.
j'ose offrir dans cette discussion-ci une solution complementaire aux remarques de BonhOffer, qui son tres a propos mais trop breves.

250

BRAD INWOOD

comme etant radicalement incommensurables. D'abord, il y a ce

que nous serions peut-etre tentes d'appeler la valeur 'morale'
(quoiqu' il est important de noter que les stoiciens n'utilisaient pas
cette etiquette pour designer la categorie, ni meme une etiquette
que l'on peut traduire en ces termes). Les seuls membres de cette
categorie sont les choses bonnes et mauvaises, soit la vertu, le vice,
et ce qui 'participe' ala vertu ou au vice. Puis, il yale domaine des
choses indifferentes, des choses qui ont une sorte de valeur, mais
pas une sorte de valeur qui soit de quelque fa~on ou a quelque degre
pertinente pour le bien-etre ou la misere de !'agent. Quelques
instances typiques de ces valeurs sont la vie et la mort, la sante et la
maladie, le plaisir et la douleur, une belle apparence et la laideur,
la force physique et la faiblesse, la richesse et la pauvrete, une
bonne et une mauvaise reputation, une grande ou une petite naissance, les talents naturels, les habiletes, meme le progres moral et
son contraire (D.L. 7.106.), etc. C'est certainement le genre de
valeurs auxquelles la majorite de l'auditoire non-specialise d'Epictete accordent de !'importance. Un stoicien tel Epictete peut considerer les objets de ces soucis humains ordinaires comme rien de
plus que des preferences: les neologismes inventes pour ce genre
de valeur sont 'prHerees' et 'de-preferees' (7tpo1'\y~a., a7to7tpOT\YJ.lEVa.).
Aucune de ces choses n'importent au bonheur humain, peu
importe a quel point notre attachement a ces valeurs est nature!.
II y a aussi, et c'est bien connu, une autre categorie de chases, les
choses qui sont 'absolument indifferentes', les choses qui n'eveillent jamais nos desirs. Par exemple, que le nombre de cheveux sur
notre tete soit pair ou impair, ou la position precise d'un de nos
doigts lorsqu' il est etendu (D.L. 7.104-105). Mais la puissance
inquietante et paradoxale de l'axiologie stoicienne devient evidente
lorsque l'on reflechit sur le fait que ces bagatelles sont dans la
meme categorie generale que la vie et la mort: etant indifferentes
elles sont, quant a leur valeur, plus proches de la sante et de la
richesse que la sante et la richesse ne le sont du bien. On trouve ici
une dichotomie radicale au niveau des valeurs entre la categorie
des vertus et des vices et le reste des choses auxquelles une personne normale accorde raisonnablement de !'importance. Ceci est
signe que I' ethique stoicienne est et do it etre prise comme etant une
theorie radicale et subversive. L'etrangete des idees de Epictete n'a
rien de nouveau. Quand il affirme qu' il n'y a aucun mal a se dire
a soi-meme (et dire a son enfant) que !'enfant pourrait mourir

L' OIKEIOSIS SOCIALE CHEZ EPICTETE

251

demain, c'est choquant, mais pas surprenant.s II n'y a aucun mal

ici parce que le seul mal veritable est le vice et que la mort de qui
que ce soit est une affaire indifferente. Nous pouvons 'de-prHerer'
cet evenement, mais si nous le considerons comme une chose
mauvaise nous compromettons notre tranquillite. Quand nous
sommes invites a commencer avec les bagatelles de la vie
( 4.1.111), un pot, une coupe, une tunique, et a les considerer
comme n'ayant aucune importance, et de mettre ensuite dans cette
categoric un animal familier, notre cheval, ou un bout de champ,
Epictete nous encourage alors a se rappeler que meme les choses
preferees comme celles-la sont indifferentes, qu' elles sont dans la
meme categoric que le nombre de cheveux sur notre tete. Et quand
il passe aux choses indifferentes prHerees qui sont plus proches de
nous, c'est-a-dire nos propres corps et les vies des membres les plus
chers de notre famille, encore une fois il nous rappelle simplement que les seules valeurs qui peuvent avoir un effet sur notre bonheur sont les valeurs 'morales', soit la vertu et le vice. C'est pour
cette raison que nos passions peuvent etre animees par des erreurs
concernant les valeurs-c'est parce que nous pouvons croire que
d'autres etres humains sont indispensables a notre bonheur que
nous eprouvons le chagrin ou la peur, le desir ou le plaisir.
Epictete illustre cette idee avec beaucoup d'insistance dans l'
Entretien 2.17, avec le cas de Medee. Dans les sections 18 ss. Epictete
fait reference a sa croyance erronee que Jason est indispensable a
son bonheur. A partir de cette croyance mal fondee son raisonnement etait neanmoins correct: Medee avait 'une representation
exacte de ce qu' est pour quelqu' un l'echec de ses desirs' (2.17.19).
Done elle tua ses propres enfants. Epictete vient tout pres de faire
l'eloge de Medee pour Ia grandeur de son arne: comme un
philosophe, elle a agi conformement a ses croyances-mais une
de ces croyances etait mal fondee. La morale de cette histoire nous
donne un frisson-la vie d'un de nos enfants ne fait pas le poids si
on la compare avec la poursuite de ce que l'on considere comme
etant notre 'bien'' ce que nous pensons etre indispensable a notre
propre bonheur.
Mais on peut se poser la question suivante: la vertu et le vice ne
doivent-ils pas etre concretises dans les etres humains? N'est-ce pas
cela que les stoiciens veulent dire lorsqu' ils parlent de 'ce qui
8

Voir 3.24.88.

252

BRAD INWOOD

participe ala vertu'? La vertu et la sagesse d'un autre etre humain

ne sont-elles pas dignes que l'on y porte un interet authentique?
Bien sur qu' elles le sont. Mais selon les stoiciens, les membres de
notre famille et nos concitoyens ne sont pas vertueux, et nous ne le
sommes pas non plus. Seule une personne sage et vertueuse
possede la valeur la plus haute, et alors ~eulement aux yeux d'un
autre sage. 9 Or quelle est !'importance pour nous des chances qu'
ont les autres d'atteindre la vertu? N'y a-t-il pas une valeur 'morale'
speciale dans le potentiel pour la vertu, et dans le progres vers la
vertu? Encore une fois, la reponse est dure. Le progres moral se
trouve dans la categoric des choses indifferentes. Le fait que votre
vie ou votre epouse est menacee n'est pas plus serieux parce que
vous ou elle est plus pres de la vertu que quelqu' autre victime. Seule
une personne sage aurait une plus grande importance, et alors r;:a
serait seulement la vertu de la personne qui aurait de !'importance,
pas sa vie.
Cependant cette far;:on de voir le probleme ne le resoud pas, mais
ne fait que le pousser plus loin en arriere dans l'histoire de 1' ecole
stoicienne, ce qui nous rappelle que le probleme des relations
sociales n'est pas apparu plus tard dans l'histoire de l'ecole, au
moment ou elle s'est tournee vers des affaires plus pratiques. Je fais
allusion, bien sur, a la fausse idee que la theorie sociale est apparue
comme un sujet digne d'interet au moment de la fondation du soidisant Portique Moyen de Panaetius, un stoicisme qui s'interessait a
la moralite pratique et qui par la se distinguait de l'ancienne ecole.
A mon avis il n'y a jamais eu une telle 'fondation', et aucune
discontinuite dans la pensee ethique stoicienne. 10 On peut done
s'attendre a trouver dans la pensee stoicienne ancienne des signes
precurseurs des problemes et tensions que l'on trouve chez Epictete.
Et je pense que no us en trouvons effectivement.
La theorie stoicienne de oikeiosis pose qu' il y a deux aspects, ou
meme deux formes, de cet instinct primordial.ll D'abord, il y a une
Voir 2.22.1-3 et Ia discussion ci-dessous.
Cf. M. Schofield, The Stoic Idea of the City (Cambridge 1991); Schofield
demontre une pareille continuite-et developpement-dans l'histoire de Ia
philosophie politique stoicienne.
11 Voir A.A. Long et D.N. Sedley, The Hellenistic Philosophers (Cambridge
1987) 1.350-4 etJ. Annas, The Marality of Happiness (Oxford 1993) 265. Travaux
fondamentaux sur oikeiosis: M. Pohlenz, 'Grundfragen der stoischen Philosophie', Abhandlungen der Giittingen Gesellschaft, phil.-hist. Klasse 3.26 (1940) 1-122;
S.G. Pembroke, 'Oikeiosis', Problems in Stoicism ed. A.A. Long (Londres 1971)
114-149; H. Gorgemanns, 'Oikeiosis in Arius Didymus' (et le commentaire de
9
10

L' OIKEIOSIS SOCIALE CHEZ EPICTETE

253

affiliation a soi-meme et a la preservation de soi (la soi-disant

oikeiosis personnelle) qui est la base de notre poursuite du bien et de
l'avantageux, ces valeurs dont la possession assure le bonheur, qui
est le but ou telos dominant de la vie humaine. Et il y a aussi une
seconde affiliation naturelle, et celle-ci nous lie avec les autres etres
humains-cette relation naturelle que tous les stoiciens consideraient comme !'explication ultime de la nature intrinsequement
sociale de l'homme. Qu' il puisse y avoir un conflit entre ces deux
affiliations dans des circonstances particulieres est clairement
possible-et les stoiciens etaient conscients de cette possibilite. II
semble que Chrysippe pensait que ce conflit ne pouvait pas etre
ultime et insoluble, et que la reconciliation des deux affiliations
fondamentales se trouve dans la relation entre nous-memes et
notre progeniture. Comme je l'ai soutenu il y a quelques annees, 12
il parta probablement de nos relations en tant que parents a nos
enfants. lis sont (selon Plutarque, St. Rep. 12) des 'parties' de nousmemes; si nous sommes con~us par nature pour s'occuper de nousmemes et de nos parties constitutives, alors nos enfants-ttant des
parties de nous-memes-nous sont egalements chers.
Cette relation biologique est evidemment naturelle, etsi c'est bien
le cas, alors la nature providentielle dont nous sommes les produits
devrait garantir qu' il n'y aura pas de conflit entre notre amour de
soi et l'amour de nos enfants. Cette affiliation a nos enfants devient
alors le fondement de notre souci general et impartial pour !'ensemble des hommes. A l'epoque de Hierocle (col. 9 de 1' Ethike stoichei6sis) il y avait meme un terme courant pour cette relation: notre
'affiliation de famille' (auyyEvtri] oiKdroat<;) est Ia base de notre
affection pour les autres-c'est atEpKnKil, Ia base pour qnA.oatopy{a.
B. Inwood), On Stoic and Peripatetic Ethics ed. W.W. Fortenbaugh (New Brunswick NJ et Londres 1983) 165-201; G. Striker, 'The Role of oikeiosis in Stoic
ethics', Oxford Studies in Ancient Philosophy 1 (1983) 145-167; B. Inwood, 'Hierocles: theory and argument in the second century A.D.', Oxford Studies in Ancient
Philosophy 2 (1984) 151-184; J. Brunschwig, 'The cradle argument in Epicureanism and Stoicism', The norms of nature, edd. G. Striker et M. Schofield
(Cambridge et Paris 1986) 113-144. Recemment J. Annas (The Morality of
Happiness 262-276 et 'Aristotle's political theory in the Hellenistic period',
justice and Generosity, edd. A. Laks et M. Schofield [Cambridge 1995] 74-94),
propose une liaison etroite entre I' oikeiosis et le concept post-Kantien de
'impartiality'. Je ne suis pas convaincu; voir mon compte-rendu, dans Ancient
Philosophy 15 (1995) 647-665.
12 'The Two Forms of oikeit5sis in Arius and the Stoa', W.W. Fortenbaugh,
( ed.), On Stoic and Peripatetic Ethics: the work of Arius Didymus (New Brunswick
NJ et Londres 1983) 190-201.

254

BRAD INWOOD

La reconciliation envisagee entre les affiliations personnelle et

sociale dans le stoicisme, done, met l'accent sur la relation entre
l'enfant et le parent, et en particulier sur l'idee que l'enfant est (a
l'origine au mains) une partie du parent. Une telle relation ne
garantit pas automatiquement et sans premisses additionnelles
l'elimination du conflit que la theorie requiert. Apres tout, il y a
une possibilite reelle qu' une de mes parties sera en conflit profond
avec moi-meme pris comme un tout: un organe cancereux
menace le tout et doit etre excise. Pourquoi done cette reconciliation
est-elle censee reussir dans le stoicisme ancien? Rien ne nous le
dit directement-il y a en effet tres peu d'evidence directe sur
ceci-mais le texte central de Diogene Laerce (D.L. 7.85 ss.) nous
donne un indice valable. Ce texte presente un argument sur la
nature de 1' oikeiosis personnelle qui est base sur une assomption
fondamentale concernant une Nature creatrice dont les operations
sont contraintes par les canons de plausibilite rationnelle: t;:a ne
serait pas plausible (eh:o<;) que la nature produise une creature et la
laisse se debrouiller toute seule, en lui donnant ou bien aucune
tendance a la preservation de soi, ou bien des tendances autodestructives. De maniere semblable, Ciceron nous dit qu' il serait
irrationnel que la nature decide de prevoir la reproduction de
l' espece, pour ensuite abandonner les enfants sans pourvoir a leurs
besoins (Fin. 3.62):
Or il serait forcement contradictoire que Ia nature vouh'it Ia
reproduction des etres et ne fit rien pour que les etres reproduits
fussent I' objet d'une affection. Meme chez les animaux se peut tres
bien remarquer cette action de Ia nature: a voir Ia peine gu' ils se
donnent pour mettre au monde et pour elever leurs petits, II semble
que l'on entende Ia voix de Ia nature elle-meme. (tr.Jules Martha)

Evidemment l'affection parentale est le moyen par lequella Nature

a assure la survie de l' espece. Au vingtieme siecle no us sommes
plus enclins a considerer la rationalite de ces arrangements
comme un aspect a posteriori de la biologie evolutionniste, mais
dans le monde ancien le modele de l'artisan dominait. Ainsi les
stoiciens pouvait plausiblement offrir la rationalite de la Nature (ou
Zeus) comme une explication authentique de certaines caracteristiques favorisant l'adaptation et la survie de l'espece, telles que la
tendance a la preservation de soi et le soin de nos enfants.
Epictete, comme les anciens stoiciens, traite la convergence de
la oikeiosis personnelle et des liens sociaux comme une partie du
plan rationnel de Zeus. Dans 1' Entretien 1.9.11 il souligne l' egoisme

L'OIKEIOSIS SOCIALE CHEZ EPICTETE

255

fondamental de Ia nature humaine. L'animal humain, selon

Epictete, fait tout pour son propre bien. Mais il avance aussi l'idee
qu' il est necessaire que cet interet personnel soit satisfait de sorte a
rendre service aux autres. Et le fait que ceci est une dimension
inevitable de Ia vie humaine est assure par le fait que Zeus est
contraint de Ia meme maniere-sa nature, etant rationnelle, n'est
pas en principe differente de Ia notre.
Ce n'est pas de l'egoisme, car telle est Ia nature de l'animal: il fait
tout pour lui-meme. Mais le soleil aussi fait tout pour lui, et du reste
Zeus lui aussi. Quand Zeus, par example, veut etre 'le dieu qui
repand Ia pluie' et 'le dieu qui produit les fruits' et 'le pere des
homme et des dieux', tu vois qu' il ne peut realiser ces actes et
meriter ces appellations sans etre utile au bien commun. D'une
fa~on generale, il a dispose de telle sorte Ia nature de l'animal
raisonnable qu' elle ne puisse obtenir aucun bien particulier sans
contribuer a l'utilite commune. Ainsi n'est-il plus antisocial de tout
faire pour soi-meme. Qu' attends-tu, en effet? Qu' on renonce a soimeme eta son propre interet? Et comment l'affiliation a soi (iJ 7tpoc;
aim1 oiKdrootc;} serait-elle encore un seul et meme principe pour
tousles etres humains? (1.19.11-15) 13
II faudrait noter que le role patemel de Zeus est lie a l'exigence qu' il
realise un equilibre entre les motivations egoistes et desinteressees;
il en va de meme pour son role de bienfaiteur (Zeus qui amene Ia
pluie). Mais ceci n'est pas surprenant, puisque le fils de Zeus,
Hercule, etait un autre bienfaiteur de Ia race humaine, et dans
1.6.30-36 il est clair que l'epanouissement de soi de Hercule passe
par le service aux autres. Pour Epictete, ceci est le propre d'un
animal rationnel.
II est encore plus important de souligner Ia derniere question
rhetorique. Ceci est Ia seule fa~on qu' il puisse y avoir un principe
unique et coherent pour tous, et ce principe doit etre fl7tpoc; aiYta
oilcdrootc;, une affiliation a soi-meme. La tendance a Ia preservation
de soi est a Ia base de nos relations sociales, mais les deux sont liees
inextricablement. Avant de finir je vais explorer ce rapport, en vue
de montrer pourquoi Ia reconciliation du souci oriente vers soimeme et du souci oriente vers les autres doit etre fonde sur Ia
primaute du souci pour soi-meme. II deviendra evident, je l'espere,
que Ia theorie d'Epictete rend explicite quelque chose qui se trouve
cache dans les textes fragmentes du stoicisme ancien.
j'ai souligne que Epictete considere Ia reconciliation de I' oikeiosis
personnelle et de I' oikeiosis sociale comme une exigence ration13

j'ai modifiee legerement Ia traduction de Souilhe ici.

256

BRAD INWOOD

nelle, sans laquelle le monde ne pourrait etre consider: comme un

tout intelligible. Et j'ai insiste sur I' idee que le modele de !'artisan
qui deploie une prevoyance intelligente est indispensable au sens
stoicien de !'organisation et de }'administration cosmique (OtaJCoaJ..lllOt<;). Une comparaison interessante vient a !'esprit si on reflechit
sur le mythe du Protagoras de Platon. Le manque de prevoyance
d'Epimethee a presque laisse l'humanite dans la situation que les
stoiciens consideraient comme implausible, dans la mesure ou
elle est irrationnelle. Quand les moyens disponibles pour la preservation de soi furent epuises, les besoins de l'humanite n'avaient pas
ete satisfaits. L' organisation sociale fut la premiere tentative pour
combler cette lacune, mais il s'avera qu' il manquait aux hommes
les ressources necessaires pour Ia vie sociale-ce qui mena presque
a Ia disparition de l'espece. Seule !'intervention de Zeus, qui donna
aux hommes la justice et le respect devant les autres (atOm<;) pouvait
les sauver. Ce mythe, raconte par Protagore, montre jusqu' a quel
point les instincts sociaux (y compris !'esprit de honte devant les
autres-comme 'tO aiOf\J..lOV chez Epictete) sont essentiels pour la
survie de l'humanite. Traduit dans le langage du naturalisme
stoicien, ce mythe illustre Ia necessite rationnelle des liens
sociaux. Dans un mythe on peut imaginer que Epimethee est
incompetant, qu' il permet que l'humanite risque sa disparition; on
peut imaginer que le Zeus de Protagore choisit volontairement s'il va
sauver les hommes ou non; mais on ne peut parler ainsi du Zeus
stoicien-sa nature (comme Ia nature des dieux de Platon) est
reellement contrainte par la rationalite. Puisque l'on ne peut concevoir que Ia Nature permette !'existence d'une espece incapable de
survivance, son caractere rationnel peut servir d'explication reelle.
Tout ceci confirme davantage, je pense, que la situation difficile
ou se trouve Epictete est Ia meme que celle ou se trouvait toute la tradition Socratique. 11 doit considerer que les liens sociaux constituent
une composante essentielle de notre nature, ancree aussi profondement dans nos natures que notre tendance ala preservation de soi
et !'amour de soi. Et cependant, tout comme ses predecesseurs, il
doit egalement considerer que la tendance individuelle pour 1' eudaimonia est !'interet dominant, plus important que les liens interpersonnels les plus forts, ceux que nous avons avec nos enfants. Qe
sais que j'ai beaucoup mis !'accent sur ces relations, presque a
!'exclusion des autres liens sociaux. C'est parce que les stoiciens
consideraient que les relations familiales etaient fondamentales

L' OIKEIOSIS SOCIALE CHEZ EPICTETE

257

pour l'ensemble des relations sociales; on peut facilement faire une

critique a fortiori de la coherence de toute la pensee sociale stoicienne si leurs idees sur les relations entre parents et enfants sont
mal fondees. On voit !'importance de ces relations pour Epictete par
le fait qu' illes invoque souvent, et aussi par la fa~on qu' il passe de
ces relations a des themes moraux plus generaux; je pense, par
exemple, a I' Entretien 1.11 De l'affection familia{e-..-(ptl..ocr'topyia).
11 reste a se demander si Epictete a fait du progres par rapport au
statu quo dans le stoicisme, et s'il a apporte une contribution
personnelle a la theorie qui sous-tend la these stoicienne sur les
liens sociaux. Je crois que oui, et que sa contribution a la pensee
stoicienne vient du modele qu' il developpa a partir des relations
entre parents et enfants.
Reflechissons un moment sur le modele des relations entre parent et enfant propose par les stoiciens anciens. Nous nous soucions
de nous-memes d'abord et avant tout, et nous nous soucions de nos
enfants parce qu' ils sont des parties de nous. Nos parties nous sont
certainement indispensables, mais elles ne sont pas nous-en tant
que parties elles sont non-pas-autres que nous, comme diraient les
stoiciens (voir Sextus M 9.336) .14 Certaines parties sont indispensables pour nous-memes en tant que totalites (nos poumons, par
exemple), tandis que des autres parties ne le sont pas (nos pieds, par
exemple). Certes, il se peut que l'on doive enlever une partie afin
que le tout survive. Epictete fait reference dans 2.6.10 au fait que
Chrysippe reconnaissait explicitement le role subordonne des
parties comme le pied, et il approuve sa doctrine. Notre pied, etant
une partie, a moins d'importance qu' une personne, qui est le tout
pertinent; le pied se salierait volontiers-ou accepterait d'etre
estropie, suivant la suggestion (non publiee) de A.A. Long 15 de
1tllpoucr8at pour 1tTlAoucr8at-si les interets du tout l'exigeaient. 11 en
va de meme pour les etres humains qui font parties de la totalite
cosmique (voir aussi 2.10); si c'est necessaire, un agent rationnel
sacrifiera au tout n'importe lequel de ses indifferents prHeres. Et Ia
14 Voir J. Barnes, 'Bits and Pieces' dans Matter and Metaphysics (edd. J.
Barnes et M. Mignucci, Naples 1988) 223-294, et 'Partial Wholes' (dans Ethi_cs,
Politics and Human Nature, edd. E.F. Paul, F.D. Miller Jr., et J. Paul, Oxford et
Cambridge MA 1991, 1-23). Barnes commence 'Bits and Pieces' avec une
reconnaisance de !'importance de Ia relation entre Ia partie et le tout dans
l'ethique stoicienne-du debut jusqu' a Ia fin de I'histoire de I' ecole.
15 Une suggestion de Long avancee a son seminaire sur I'ethique
stoicienne a Berkeley au printemps de 1982, mais qu' il n' a pas inclu dans
The Hellenistic Philosophers (Cambridge 1987).

258

BRAD INWOOD

vie de n'importe qui, meme celle d'un membre intime de notre

famille, ne vaut pas plus que ~a.
Nos enfants sont des parties, au moins dans le sens que c'est en
tant que parties de nous-memes qu' ils revendiquent le plus
fortement notre affection. Si l'on peut dans certains cas se passer
d'une de nos parties, on peut se passer encore plus d'un enfant. Et si
I' on peut se passer d'un enfant, on peut se passer encore plus d'un
autre etre humain! Et pourtant, l'enfant est une partie de nousmemes; il n'est pas 'autre' que nous. Je pense que cette perspective
sur nos vies nous permet de bien comprendre Ia tension entre
I' oikeiosis personnelle et I' oikeiosis sociale. L'idee que nos enfants
sont des parties explique pourquoi nous nous soucions tellement
d'eux, pourquoi notre interet pour nos enfants est dans Ia plupart des
cas et dans Ia plupart des circonstances egal a notre interet
personnel. Mais cette idee explique aussi pourquoi, en derniere
analyse, le moi vient en premier. Mes enfants, apres tout, ne sont
pas des 'autres moi'. (11 est interessant de noter que seuls les
hommes sages peuvent etre vraiment des amis et done des 'autres
moi' l'un pour l'autre, de sorte qu' un homme sage a une raison de
considerer qu' une autre personne a une importance egale a Ia
sienne, et alors seulement si cette personne-la est sage elle aussid'ou Ia rarete du dilemme confronte par les deux sages sur Ia
planche, mentionne par Ciceron, De Officiis, 3.89-90)
La conception des enfants pris comme des parties realise exactement ce que l'on veut. Elle explique a Ia fois pourquoi ils sont si
importants et pourquoi ils peuvent etre relegues au second rang en
cas de conflit. Et si nos relations avec les autres etres humains
(notre oikeiosis sociale) sont derivees et formees a partir de nos
relations avec nos enfants, alors ces relations aussi s'inscrivent
bien dans ce modele. On peut presumer que Epictete a emprunte ce
modele a ces predecesseurs. Sa propre contribution a Ia conception
de ce probleme se trouve, je pense, dans son developpement de
l'idee des relations entre parent et enfant. Plus que Ia plupart des
stoiciens dont nous connaissons les textes, Epictete considere que
Zeus est une personne, un pere pour l'espece humaine. L'idee
homerique du Zeus 'pere des dieux et des hommes' fait partie de
l'arriere-plan; comme plusieurs stoiciens Epictete retient et justifie
les aspects de Ia tradition poetique qui conviennent a sa theorie.
Epictete, comme Cleanthe avant lui (dans I' Hymne a Zeus) et Marc
Aurele plus tard (qui a lui-meme ete influence par Epictete),

L' OIKEIOSIS SOCIALE CHEZ EPICTETE

259

soulignait le caractere personnel de nos relations avec Zeus. Zeus

est souvent un 'tu'-un Du Buberien, un autre individu dont le
rapport avec nous est celui d'un parent.
Nous, les etres humains, sommes des parties de Ia totalite
cosmique, des parties ainsi que des allies de Zeus, comme le dit
Seneque (Ep. 92.30). Ce statut que nous avons comme parties du
Cosmos s'accorde bien avec le statut que nous avons comme
enfants de Zeus. Plus on met l'accent sur le fait que nous sommes
les enfants de Zeus, plus il est nature} de penser que notre relation
avec nos semblables est une relation fraternelle, une relation avec
nos freres (et soeurs). Dans ce modele, les autres etres humains
deviennent nos freres (et soeurs). La notion que tousles hommes
sont freres a son origine dans Ia pensee stoicienne, puisque nous
pouvons etre freres seulement si nous avons un parent en commun; les images cosmologiques stoiciennes presentent Ia source Ia
plus riche de cette imagerie de Ia relation entre parent et enfant.
Epictete trouve qu' il est nature} de mettre l'acccent sur Ia conception cosmologique ou Zeus est un parent quand il se penche sur
l'attitude humaine erronee face a Ia perte d'un etre bien-aime, c'esta-dire le chagrin. Dans I' Entretien important 3.24 le souci de dieu
pour les humains est paternel (3.24.3) lorsqu' il rend notre bonheur
independant des choses qui sont hors de notre controle:
A cette fin, il leur a donne des moyens dont certains appartiennent
en propre a chacun et les autres lui sont etrangers. Tout ce qui peut
etre entr(;lve, enleve, violente, ne nous appartient pas en propre; tout
ce qui est libre d'entraves nous appartient en propre. Quant ala vraie
nature du bien et du mal, dieu, ansi qu' il convenait a celui qui
veille sur nous et nous protege a la maniere d'un pere, nous l'a
donnee en propre.
II faut penser a Ia cosmologie stoicienne (3.24.10-11); Epictete dit,
ecoutons les philosophes:
ils [disent] que cet univers n'est qu' une seule cite, que la substance
dont il est forme est unique et qu' il doit y avoir une revolution
periodique ou les choses se cedent mutuellement la place, ou les
unes se dissolvent tandis que d'autres viennent au monde, ou les
unes demeurent au meme lieu tandis que d'autres se mettent en
mouvement. Et tout est rempli d'amis: d'abord de dieux, puis
egalement d'hommes que la nature a rapproches les uns des autres:
les uns doivent vivre ensemble, tandis que d'autres doivent se
quitter, se complaisant parmi ceux qui vivent avec eux et sans
s'attrister de voir s'eloigner les autres.
Comme le montrent les carrieres d' Ulysse et de Hercule, les gens
se deplacent, ils meurent. Pourquoi n'est-ce pas une infortune? En
partie parce que c'est un etat de choses naturel et inevitable. Mais ce

260

BRAD INWOOD

que Epictete dit au sujet de Hercule est plus important. 11 etait

constamment guide par le savoir:
en effet, que nul homme n'est orphelin, mais que tous ont un pere,
qui, sans jamais cesser, prend soin d'eux. Car ce n'etaient pas de
simples mots qu' il avait recueillis, quand il entendait dire que Zeus
est le pere des hommes, lui qui, en verite, le regardait comme son
pere et l'appelait ainsi et, le regard fixe sur lui, accomplissait toutes
ses actions. C'est pourquoi il a pu vivre heureux partout. Mais il est
impossible d'associer le bonheur et le regret des choses absentes.
Car l'etre heureux doit recevoir tout ce qu' il desire; il doit
ressembler a un homme repu: il ne doit eprouver ni Ia soif ni Ia
faim. 'Ulysse, pourtant, regrettait son epouse et, assis sur un rocher,
il pleura1t.' Et toi, tu ajoutes foi a tout ce que dit Homere et a ses
fables? S'il est bien vrai qu' Ulysse pleurait, qu' i:tait-il de plus qu' un
infortune? Or, que! homme de bien peut etre infortune? En verite,
I'univers est bien mal gouverne si Zeus ne veille pas a ce que ses
propres concitoyens soient heureux comme lui. Non, il n'est pas
permis et il est impie de penser cela, mais si Ulysse pleurait et se
lamentait, il n'i:tait pas un homme bon. Qui, en effet, peut etre bon,
s'il ignore qui il est? Et qui le sait, s'il a oublie que tout ce qui vient a
!'existence est pi:rissable et que, pour les hommes, vivre ensemble
n'est pas toujours possible? Quelle sera Ia conclusion? C'est que
desirer !'impossible est le fait d'un esclave, d'un sot; c'est agir
comme un etranger, qui combat contre dieu avec les seules armes
qui soient a sa disposition, ses propresjugements (3.24.15-21)

Mais ceder au chagrin, c'est se battre contre dieu:

Par consequent [du fait que le chagrin d'autrui n'est pas a moi, mais
seulement le mien] je le ferai cesser a tout prix, car cela depend de
moi; pour celui des autres, j'essaierai, suivant mes forces, mais je
n 'essaierai pas a tout prix; sin on, je combattrai contres les dieux, je
me mettrai en opposition avec Zeus, je serai en contradiction avec
lui pour le gouvernement de l'univers. Et Ia peine de cette lutte
contre dieu et de cette desobeissance, ce ne sont pas les enfants de
mes enfants qui Ia paieront, mais moi-meme, de jour et de nuit,
quand mes reves me feront tressauter, quand je vivrai dans le
trouble, tremblant a Ia moindre nouvelle, avec Ia paix de mon arne
(mon a7tli8wx.) suspendue aux lettres des autres (3.24.23-24).

Ce n'est pas un accident si dans cet Entretien Epictete considere

ensuite la relation de Socrate avec ses enfants (3.24.60 ss.)-il les
aimait, malgre son indifference notoire face a leur destin dans ce
monde (voir la fin de l'Apologie). 11 se souvient qu' il doit aimer ses
fils 'comme le fait un homme libre, gardant a !'esprit qu' il doit etre
un ami des dieux d'abord et avant tout'. L'amitie divine, ou la qnA.ia.
par rapport a ce qui est purement rationnel, a la priorite sur nos
amours et affiliations humaines. 11 y a d'autres entretiens ou ces
themes se retrouvent de la meme fa~on-par exemple 1.9, ou l'on
contraste la conscience de la parent: divine avec les relations
sociales humaines. Notre citoyennete humaine dans une petite

L'OIKEIOSIS SOCIALE CHEZ EPICTETE

261

ville est basee sur nos liens de parente avec nos semblables, et notre
citoyennete dans Ia cosmopolis est basee de Ia meme fa~,;on sur
notre parente avec dieu. Le divin est le paradigme pour Ia vie
terrestre; nos relations cosmiques soot a Ia fois le modele et Ia cause
de nos relations humaines-et ce qui est plus important, le divin
nous aide a comprendre que !'experience humaine est un systeme
rationnel et coherent.
En soulignant que Zeus remplit le role du pere commun de tous
les hommes, Epictete augmente et renforce les fondements de Ia
theorie stoicienne des relations humaines. Nos liens avec les
autres hommes ne soot pas, comme les autres stoiciens le pensaient, seulement une extension de }'affection parentale aux autres
etres humains par le biais d'une analogie faible et diluee (comme
le suggere les cercles en expansion de Hierocle); cette fa~,;on de
promouvoir Ia fraternite des hommes fut attaquee par le commentateur anonyme du Theetete, et cette critique nous fait penser aux
commentaires d'Aristote sur les relations familiales artificielles
dans Ia Ripublique de Platon. Plutot, cette relation est ancree dans
une application nouvelle de Ia metaphore des relations familiales,
dans laquelle Ia figure principale est Zeus dans le role du pere. Ce
modele aide aussi a expliquer nos relations sociales aux autres
hommes, et rend compte et de leur caractere naturel et de leur
place secondaire. Le lien le plus fort dans n'importe quelle famille
ancienne (d'un point de vue ideologique) est le lien entre pere et
fils; les relations entre freres (et soeurs) doivent leur existence
meme a Ia primaute de Ia relation paternelle. La primaute de notre
relation avec Zeus le pere est un symbole, pour ainsi dire, de Ia primaute de Ia coherence rationnelle dans nos vies. Et tout comme Ia
primaute de Ia relation parent-enfant met les liens entre freres (et
soeurs) en perspective, Ia primaute de Ia coherence rationnelle
dans nos vies determine que toutes les autres relations-crxaet~
sont secondaires. Ces themes sont tres importants dans le discours
que je vais explorer en guise de conclusion, I' Entretien 2.22 intitutle
De l'amitie.
Epictete prend comme point de depart un argument raisonnable
(eiK6-rco~) modele sur celui que Chrysippe a employe pour fonder Ia
rationalite de notre inclination a Ia preservation de soi dans D.L.
7.85 ss. Nous aimons (cptAe\v) ce que nous prenons au serieux; ceci
est raisonnable, puisqu' il serait absurde de prendre au serieux ce
qui est mauvais pour nous ou ce qui ne presente aucun interet pour

262

BRAD INWOOD

nous. Ainsi nous prenons au serieux seulement ce que nous

considerons bon, et si nous prenons quelque chose au serieux, il
s'ensuit que nous l'aimons. Par consequent, seule une personne qui
connait le bien peut reellement savoir comment etre un ami. Seul
un homme sage peut done etre un ami ou eprouver de l'amitie.
Cette position strictement rationnelle est ensuite contestee par
l'interlocuteur imaginaire d'Epictete, qui dit 'Ecoutez, je ne suis pas
un homme sage, et pourtantj'aime mon enfant'. Epictete doit, dans
ce qui suit, montrer que l'amour d'un fou n'est pas vraiment de
l'amour. Mais il commence par isoler la cause de la folie de son
interlocuteur. C'est qu' il se trompe au sujet de la valeur des choses:
il s'agit d'une erreur axiologique. Le resultat de cette erreur est la
passion et la misere, l'instabilite et l'inconsistance, un panorama
toujours changeant de reactions aux evenements de la vie. Quand
l'interlocuteur reconnait son erreur, Epictete lui demande de
reconnaitre les mauvaises consequences de cette inconsistance sur
l'amitie. Quand notre attention se deplace de l'amour de notre ami
a l'amour de l'argent, notre attitude change. La comparaison avec
les chiens est pertinente:
Eh quoi! As-tu jamais vu des petits chiens se caresser et jouer entre
eux, ce qui te faisait dire: 'II n'y a rien de plus amical'? Mais pour
voir ce qu' est l'amitie,jette entre eux un morceau de viande et tu le
sauras. Jette aussi entre toi et ton fils un bout de champ, et tu sauras
combien ton fils est impatient de t'enterrer et combien tu souhaites,
toi, Ia mort de ton fils. Et tu diras au contraire: 'Que! enfant j'ai
eleve! II y a longtemps qu' il m'enterre.' (2.22.9-10)
Viennent ensuite d'autres exemples d'amour instable, qui proviennent des tragedies greques, I' Alcestis et les Phoenissae en particulier.
Quand n'importe quelle chose qui revet de l'importance pour nous
affaiblit le lien entre deux amis, alors ce qu' ils valorisent vraiment
devient evident, et l'amitie meurt. Les seules amities reelles sont
celles partagees par des gens qui savent la valeur reelle des choses,
puisque meme les relations entre freres et soeurs et entre parent et
enfant sont vulnerables. Le paradoxe stoicien selon lequel seuls les
sages peuvent etre des amis (voir D.L. 7. 124) ret,;:oit ici un sens
concret. 16
16 On pourrait comparer Ia fa.;:on dont Seni:que rend ces paradoxes reels
dans le De Beneficiis (voir mon 'Politics and Paradox' dans justice and
Generosity, edd. M. Schofield etA. Laks, Cambridge 1995) ou Ia strategie tri:s
differente de Cichon dans le Paradoxa Stoicorum (voir W. Englert, 'Bringing
Philosophy to the Light: Cicero's Paradoxa Stoicorum' in The Poetics of Therapy
ed. M. Nussbaum (Edmonton 1990 = Apeiron 23 no.4).

L' OIKEIOSIS SOCIALE CHEZ EPICTETE

263

Selon Epictete, ceci est tout a fait naturel:

En regie generale-ne vous faites pas d'illusion-tout etre vivant n'a
rien qui lui soit plus cher que son propre interet. Des lors, quoi que
ce soit qui lui paraisse y faire obstacle, s'agirait-il d'un frere, ou d'un
pere, ou d'un enfant, ou d'un etre aime, ou d'un amant, ille bait, il
le rejette, il le maudit. Car il n'est rien que, par nature, il aime
autant que son propre interet. C'est la son frere, ses parents, sa patrie,
son dieu. (2.22.15-6)
L'histoire revele que meme les dieux ne sont pas exempts de la
colere d'un etre humain qui pense que ses interets ont ete leses.
Ainsi une fausse conception de mes interets reels detruit meme les
liens humains les plus forts (18-21). Les grandes guerres du passe
sont basees sur de telles erreurs (20-23), et done l' on ne peut pas
penser que meme les plus sinceres et enthousiastes declarations
d'amitie sont secures. En effet, l'on ne peut compter sur les
relations humaines si les agents en question sont des fous sans
vertu (25). Une evaluation de la force de n'importe quellien social
ne peut etre faite a partir de criteres traditionnels (26-29); elle do it
passer par un examen des valeurs morales profondes des hommes
en question. Savent-ils quels sont leurs interets reels? Si oui, alors il
se peut fort bien qu' ils soient des amis authentiques, et qu' ils aient
la fidelite et le respect (1tianc;, aioroc;) qui constituent le fondement
des relations humaines (30). Il n'y a pas d'autre base solide pour les
liens sociaux authentiques, pas meme les liens qui semblent avoir
dure le plus longtemps et etre les plus forts.
Epictete termine en decrivant la fausse opinion sur la valeur des
chases externes:
Telle fut Ia force sauvage, telle fut Ia force destructrice de l'amitie,
celle qui ne permit plus a une femme d'etre une epouse, a une
mere d'etre vraiment mere. Et que celui d'entre vous qui a sincerementa coeur ou de temoigner lui-meme son amitie a quelqu' un,
ou de se gagner l'amitie d'un autre, extirpe ces jugements, qu' il les
haisse, qu' illes banisse de son arne a lui. De Ia sorte, il n'aura point
tout d'abord a s'adresser a lui-meme de reproches, a Iutter contre luimeme, a se repentir, a se tourmenter; de plus, il n'entrera pas en
lutte avec son prochain, absolument pas avec celui qui lui ressemble, et vis-a-vis de celui qui lui est dissemblable, il sera tolerant,
condescendant, doux, indulgent, comme envers un ignorant,
comme envers un homme qui se fourvoie en matiere capitale; il ne
sera dur pour personne, car il comprendra parfaitement le mot de
Platon: 'C'est toujours malgre elle qu' une arne est privee de Ia
verite.' Dans le cas contraire, vous agirez en tout comme agissent
les amis: vous boirez ensemble, vous vivrez sous le meme toit, vous
naviguerez de conserve et vous pourrez avoir les memes parents.
Oui, et les serpents aussi. Mais amis? vous ne le serez pas plus qu' ils
ne le sont tant que vous conservez ces jugements sauvages, ces
jugements pervers. (2.22.33-37).

264

BRAD INWOOD

Et c'est ici que nous devons terminer, avec le diagnostic final

d'Epictete sur la nature des liens sociaux. Dans sa theorie, les liens
sociaux doivent etre secondaires, ils ne peuvent etre au meme
niveau de valeur que 1' oikeiosis fondamentale que chacun a envers
soi-meme. Et ceci est du au fait qu' une appreciation adequate de
notre propre benefice est essentielle a la stabilite morale de
n'importe quel animal rationnel. Etant ce que nous sommes, tels
que Zeus nous a crees, il nous est impossible de poursuivre un autre
but fondamental; notre relation naturelle avec notre propre bienetre doit etre le point de depart de tous nos engagements. C'est
pourquoi il dit que seule 1' oikeiosis personnelle peut etre le principe
unique de la vie morale; si nous mettons les relations sociales au
meme niveau, nous ne pouvons plus maintenir notre equilibre.
11 s'avere done inevitable dans la theorie stoicienne que l'amitie
reelle, la qnA.(a stable, doive s'obtenir au prix du detachement
affectif. A fin de montrer de la loyaute et du respect reels envers
nos semblables-meme pour les membres de notre famille-nous
devons egalement cesser de s'en faire s'ils nous quittent, s'ils
rivalisent avec nous, s'ils nous trahissent, ou s'ils meurent. En effet,
la confusion axiologique est la source de la haine autant que de
l'amour, et c'est aussi l'adversaire de la stabilite rationelle.
Ainsi nous vivons dans un monde cruel, ou les relations
humaines ne peuvent jamais etre ideales-nous pouvons avoir la
consistance et !'affection reciproque, mais pas en meme temps que
la passion profonde. Ou bien il est possible d'avoir des liens
passionnels et devoues, mais seulement au prix d'une troublante
instabilite. On peut se demander comment un Zeus bienveillant
pourrait creer un monde comme celui-la. Le Zeus tout-puissant ne
peut-il pas faire mieux que ~a pour ses enfants?
Je termine avec deux pensees. (1) 11 se peut que Zeus n'ait pas
voulu faire mieux. Dans les affaires affectives, comme pour tout le
reste, nos gouts sont differents et il est possible que notre monde est
reellement celui que Zeus voulait creer. Mais (2) Zeus se trouve probablement sous les memes contraintes que cet autre artisan divin,
le Demiurge de Platon. Comme lui, et comme tous les artisans, le
Zeus stoicien doit travailler avec les materiaux qu' il a sous la main,
et il n'est pas vraiment tout-puissant. Et nous, les etres humains,
no us devons no us debrouiller avec le meilleur des mondes possibles.
National Humanities Center, North Carolina
et Universite de Toronto

THE HUNT FOR GALEN'S SHADOW

Alexander of Aphrodisias, De anima
94.7-100.17 Bruns Reconsidered
TEUN TIELEMAN*

1. Mter long years of comparative neglect, Galen of Pergamon

(129-c. 210 CE) is being re-discovered as a philosopher in his own
right. This re-appraisal is to be applauded, not least because he was
taken seriously as a philosopher in his own day and beyond. 1 In
fact, our earliest extant testimonies-a mere handful-concern his
influence in regard to philosophical, not medical matters. Most
notable are those from (or relating to) the important Peripatetic
philosopher Alexander of Aphrodisias, who was Galen's junior by
about a generation.2 These testimonies have encouraged a comparison between their surviving writings from a doctrinal point of
view. Thus Donini has argued convincingly that Alexander at De
anima 24.18-263 engages in an (implicit) polemical dialogue with
Galen (notably his Quod animi mores) on the view of the soul as a
harmony of the body.4
The question of Galen's presence also arises in regard to the
closing section of the De anima (94.7-100.17), which is concerned
with the question where in our body the centre of command, or, as
it was called, the 'regent part' (flYEJlOV11c6v) of the soul, is located. 5
Alexander presents a battery of arguments in favour of the Aristotelian view that the regent part resides in the heart rather than the
I regard it as an honour and a privilege to be able to dedicate this article
to my esteemed teacher on his sixtieth birthday.
1 See Nutton (1984); cf. also Temkin (1973) 51ff., Frede (1981) 66.
2 Alexander dedicated his De Jato to the emperors Septimius Severus and
Caracalla (164.3 ff. Br.), who exercised joint rule between 198 and 209 CE.
These are the only firm dates we have with regard to his life and career; cf.
Todd (1976) 1. It is a fair assumption that he was born in the 150's or 160's.
Galen was born in 129 and presumably died some time after 210: see Nutton
(1984) 323; id., (1987) 46 ff.
3 References give page- and line-number of the standard edition by I.
Bruns in the GAG-series (Suppl. II, pars I, Berlin 1887).
4 Donini (1974) 148ff.; id., (1982) 22tH!, 232; cf. Mansfeld (1990) 3108n.217.
5 On this long-standing controversy in ancient philosophy and medicine
see Mansfeld ( 1990), Tieleman ( 1996) xxxiiiff.

266

TEUN TIELEMAN

brain. A few decades earlier, 6 Galen had devoted no less than four
books (viz. 1-111 and VI) of his On the doctrines of Hippocrates and Plato
(hereafter PHP) 7 to the same issue. Drawing on his dialectical
skills and extensive anatomical knowledge, he delivered a
thundering vindication of the Platonic tripartition and trilocation.
That is to say, he established the brain as the seat of Plato's rational
part (commonly identified with the regent part8 ), refuting in the
process the main Stoic and Peripatetic arguments in favour of the
cardiocentric theory.
Naturally enough, it has been suggested that Alexander wrote
the last section of De an. in conscious opposition to Galen.9 Some
scholars have even commented unfavourably on the strength of
what is considered Alexander's response to Galen's challenge. 10 In
view of what is otherwise known of Alexander's reactions to
Galen's workll and of the impact of PHP,12 it is quite plausible that
Alexander had the argument of PHP forced upon his attention.
6 PHP books I-VI were written during Galen's first stay in Rome, i.e.
between 162 and 166 CE, see De Lacy (1978) 46-48. Alexander's De anima is
usually dated to the final period of his career; see Moraux (1978) 304-5;
Donini (1982) 231; cf. supra, n. 2.
7 References give book-, chapter- and paragraph-numbers in Ph. De Lacy's
edition (CMG V 4,1,2, 3 vols. Berlin 1978-82)
8 Cf. Alex. De an. 98.24., Gal., PHP II 5.81.
9 Todd (1977); Moraux (1978) 293; Donini (1982) 232; Mansfeld (1990) 3109
n. 223.
10 Moraux (1978) 293; Donini (1982) 231.
11 Alexander, In Top. 549.24 Wallies mentions Galen alongside Plato and
Aristotle as an authority whose views count as voo~a suitable for dialectical
discussion. He composed two tracts against Galen's (lost) On the First Mover
and On Possibility respectively. The former has been preserved in an Arabic
abstract; see Rescher-Marm urra ( 1965). One of the Arabic mss. contains a
notice saying that Galen wrote to Herminus a letter in which he criticized
Aristotle's theory of motion; see Pines (1961) 23, Rescher and Marmurra
(1965) 57f. Herminus was Alexander's teacher, see Simp!. De caelo 430.32ff.
Heioerg. He may be identical to the Peripatetic philosopher and pupil of
Aspasius, whom the young Galen heard lecture in Pergamum, Aff. Dig. 8.4,
p.28.15f. De Boer. See further Moraux (1984) 361ff. On Alexander's affiliations
with (proximate) contemporaries cf. also Sharples (1990). The identification
of Alexander with Alexander of Damascus (De praecogn. 5.9-15 Nutton; De A nat.
Admin. 2.216.6-8 K.)-first made in Arabic sources-is today generally rejected; see Todd (1976) 4ff. esp. 6 n.29, Sharples (1987) 1179 with nn. 18-21.
l2 See esp. PHP VII 1.1-4: books I-VI circulated among Peripatetics, Stoics
and physicians and, if we may believe Galen, had already convinced some of
these opponents. On the Peripatetic section of his audience see also II 3.23. In
fact, books I-VI were written at the request of a Peripatetic, viz. Flavius
Boethus, see Lib. Prop. 1, SM II p.96.19ff. Muller; cf. also De praecogn. 2.24, 5.9,
5.19 with Nutton ad loc.

GALEN AND ALEXANDER OF APHRODISIAS

267

This claim is not invalidated by the fact that Alexander does not
refer explicitly to Galen in the De an. Both Alexander and Galen
follow the common practice of avoiding names of contemporary
opponents when dealing with their arguments.l 3 For example
Alexander directed his tract on the theory of motion against Galen
without mentioning him by name.14
But it is one thing to assume that Galen looms in the background of Alexander's argument; it is quite another to substantiate
and illustrate this assumption on the basis of textual evidence. In
fact, a negative conclusion has been reached by Accattino in the
most recent and fullest available discussion on the matter. 15 This
discussion takes the form of an extensive critique of three parallels
adduced by Todd.I6 These parallels, Accattino argues, fail to establish the thesis that Alexander uses arguments or ideas of Galen. 17
However, the removal of Todd's trio of parallels does not settle the
issue satisfactorily. What we need is a systematic treatment involving more material from Galen. Oddly enough, neither Todd nor
Accattino pays much attention to what might be called Galen's
formal contribution to the controversy over the seat of the regent
part-the PHP.
In the following pages, I will reconsider the relation between
the closing section of the De an. and what is to be found in Galen's
writings, most notably the PHP. I shall argue that Alexander does
respond to specific arguments he had read in Galen. Meanwhile
Accattino's critique may stand as a reminder of the strict criteria
needed for assessing 'parallels' in terms of historical relations.
Mere resemblances are not good enough. Many current arguments and ideas used in the controversy over the regent part
were traditional and hence not confined to Galen and Alexander.
13 Tieleman (1996) 67ff., Sharples (1987) 1178f. The exegetical style of later
ancient philosophy entailed a heavy concentration on the ipsissima verba of
the past masters, notably Plato, Aristotle and Chrysippus. PHP books I-VI,
with their extensive quotations, are a case in point. Likewise Alexander
presents his De an. as an explication of Aristotle's doctrine (2.4-9). In much
philosophical polemicizing of the period, contemporary opponents are thus
involved without being explicitly or separately refuted.
14 See Rescher and Marmurra (1965), esp. 59.
15 Accattino ( 1987).
16 Todd (1977), esp. 117f., 121-3. See infra, nn. 33, 55.
17 In fact Todd (1977) 123 does not even claim that Alexander had read
Galen but merely propounds an indirect kind of relation, i.e. one mediated
by other authors.

268

TEUN TIELEMAN

Medical insights may have reached Alexander by a different

route. The case for direct influence should therefore rest on
specific features of Alexander's argument which are most easily
explicable by reference to Galen. As will be shown, this question
involves not only individual arguments and passages but also
Alexander's overall strategy .
My argument is structured as follows. I shall begin by presenting some observations on the overall design and strategy of Alexander's demonstration ( 2). Next I shall compare his arguments
concerned with the nutritive and other faculties of the soul with
the relevant passages in Galen ( 3). This is followed by a few observations on Alexander's rebuttal of two encephalocentric arguments at the end of his demonstration ( 4). Finally I shall draw
some conclusions with special reference to Alexander's dialectical
procedures ( 5)
2. Alexander turns to the question of the seat of the dominating
part (Tt'YEJ..LOVtKOV) of the soul (94.7-11) at the end of his work-a
placement which attests to its importance:
Where is the regent part of the soul, i.e. in which part of the body ?
And does the soul reside in a substrate ( {moKEtJ..LEVcp) that is numerically one, having differentiations only as to powers (ouvaJ..LEt~) and
conceptually, or are the powers of the soul also separated from one
another by location?

This scheme of options is common ground between Alexander,

Galen and such Platonists as Porphyry and Iamblichus. As such, it
provides a suitable point of departure.1B
Alexander goes on to posit the general principle that in all
things in which one thing is 'more perfect' (uA.no'tepov) and
another 'less perfect' (<heA.tcr'tepov), the less perfect becomes perfect
through a certain addition; hence where is the imperfect thing,
there must also be the 'more perfect' one (94.11-13; cf. 28.14ff.).
Alexander applies this to the psychic functions, starting with the
least perfect among them, viz. the nutritive (8pE1t'ttKTt) power. In an
ascending series, each of the following functions is inferred from
the presence of the preceding one. Thus the perceptive faculty,
Alexander argues, cannot be located separately from the nutritive
one (96.19-23). 19 And in the same section he argues that since the
18 Gal. PHP VI 2.5, Porphyry Fr. 253 Smith; Iamb!. De an. ap. Stob. Eel. pp.
368.12-369.4 W. Cf. also Alex. De an. 94.1-3.
19 Cf. Arist. Iuv. 3, 469a8ff.

GALEN AND ALEXANDER OF APHRODISIAS

269

heart is the source of taste and touch, the other senses must reside
there too (96.25-8) .2 The series culminates in the theoretical
intellect (o 8ecop11'tuco~ vou~). being the perfection of the soul (99.6).
This scheme constitutes so to speak the skeleton of the demonstration as a whole. It is fleshed out by additional arguments establishing the location of each of the main functions independently
of the others. This concatenation of functions corresponds to the
order in which the individual powers are treated in the main body
of the work (27.1-92.11, recapitulated at 92.12-94.6): Alexander
successively deals with the nutritive (94.17-96.1 0), the perceptive
(96.10-97.15), the appetitive-cum-motive (97.15-99.1) and cognitive
(99.1-10) powers. Each of these main powers is further subdivided,
and the resulting sub-species are linked by inferences as well. We
must note in particular his ploy of proceeding by small transitions,
including those from one main function to another. Thus at 98.2999.2 we have the following sequence: where is the appetitive
(opEK'ttKOv), there is also the volitional (~ouA.rtnK6v); and where is
the volitional, there is the deliberative and rational (~oUAEU'ttlCOV 'tE
K<Xt A.oytcr't11COV) faculty.
The ordering of the main functions is loosely modelled on
Aristotle's De an. Books II and III. As to the steps made from one
function to another it is also worth comparing individual passages
such as Arist. De an. B 2, 413b21-421 and the hierarchy of five functions, ibid. 3, 414a31-33. But the prominence given by Alexander to
this inferential sequence is no doubt due to the influence of the
(originally Sceptical) little-by-little argument or s6rites.22
Here Alexander was anticipated by Chrysippus' demonstration
of the cardiocentric view in his De an., from which Galen has preserved extensive quotations in PHP books II and III. In his account
of what he regards as common opinion, Chrysippus started from
the perception of certain passions, notably inflamed anger (6pYJ1),
arising in the heart. But then:

20

Cf. Arist. luv. 3, 469al3-17.

21 'In insects which have been cut in two each of the segments possesses

both sensation and local movement; and if sensation, necessarily also imagination and appetition; for where there is sensation, there is also pleasure
and pain; and where these, necessarily also desire' (Oxford transl.).
22 On the sorites in general see Barnes (1982); Burnyeat (1982). For Chrysippus' use of it see SW 2.665; cf. 1003, 1005, Cic. ND 2.164-166. Cf. Burnyeat
(1982) 337f.

270

TEUN TIELEMAN

given the fact that inflamed anger arises there, it is also reasonable
that the other desires are there, and indeed the rest of the passions
and the deliberations and whatever resembles these (PHP III 2.5 =
SW2.890; cf. Alex. De an. 97.1-3).23
Galen tells us (III 2.4) that Chrysippus used this mode of reasoning
throughout his demonstration, thus indicating, ones presumes, that
the Stoic developed this common pattern of thought further in the
light of considerations of a more philosophical and scientific
kind.2 4 The argument exemplifies Chrysippus' use of the sorites for
constructive purposes, just as he appropriated the Sceptical technique of diaphonia.2 5 The use made of sorites by Chrysippus and
Alexander in this particular context is an effective means of
countering the Platonic tripartition.26
3. Let us now take a closer look at those of Alexander's arguments
which invite comparison with the relevant parts of PHP. Given the
relation posited between things of greater or lesser perfection (see
above) it is of prime importance for Alexander to establish the
presence in the heart of the lowest or nutritive faculty: 'where the
nutritive power is, there are also the more perfect ones' (94.17).
Obviously, the nutritive faculty, being at the bottom of the scale,
cannot be inferred from the presence of another one. Thus he
devotes considerably more space to the principle of nourishment
than to any of the otherfaculties (94.11-96.10).
Alexander begins by pointing to the need for a large supply of
hot and wet material as indispensable for life and hence for the
nutritive soul, affirming that the heart meets this requirement
(94.17-20; cf. 96.5-8) .27 This passage (as well as 39.21-40.3, to which
he refers) is closely similar to an argument paraphrased and
criticized by Galen at PHP II 8.30ff.2 8 For chronological reasons
Galen cannot be attacking Alexander.29 We are dealing with a
traditional Peripatetic (and Stoic) argument reiterated by Alexander, who gives no sign of taking note of Galen's objections.
23 Also quoted at ibid. III 5.2 (SW2.89), II 7.11 (SW 2.887).

24 PHP III 5.43-44 (= SW 2.899); cf. III 3.7 with Tieleman (1996) 249 ff.
25 For the diaphtSnia see esp. Chrys. ap. Gal, PHP III 1.10-15 (SW 2.885)
with Mansfeld (1990) 3167ff. See also, infra, n. 61.
2 6 On Plato as Chrysippus' main butt see Tieleman (1996) 140f.
27 PA f 4,665b, Iuv. 4, 469b7ff.
28 See Tieleman (1996) 67ff.
29 Supra, nn. 2, 6.

GALEN AND ALEXANDER OF APHRODISIAS

271

Alexander next (94.21-6) draws an analogy between plants and

animals with respect to growth and nourishment. In plants these
physiological processes are directed from a single principle located
between roots and stem. In animals too the nutritive and generative
principles are identical. In them, Alexander briskly concludes, it
is the heart which is the starting point of growth and nourishment.
In other words, the nutritive power resides in the heart. Again,
there is a parallel in Galen. At PHP VI 3.10-26 he draws what is
unmistakably the same analogy. Although the passage of Alexander is much more concise than Galen's rather elaborate version,
their arguments run clearly parallel. Of course, Galen argues that
the liver, not the heart, is the analogue of the spot between the stem
and roots of plants. The procedure of arguing from analogy, or
similarity, can be traced back to the Aristotelian Topics.30 But this
elaborate application to the anatomy of plants and animals seems to
be original with Galen.3 1 Alexander simply reverses the argument
into one in favour of the heart.
The plant analogy remains in force, though less prominently,
in Alexander's subsequent observations, which concern the
structure of the blood-vessels (94.26-95.6).32 The blood is the ultimate
form of nutriment, and the source of the vessels is the heart. But
Alexander qualifies the conclusion that the heart is, or contains,
the principle of nourishment by assigning to the liver a preparatory role in the digestive process. It is this view of the functions of
the two organs which Galen refutes at some length at PHP VI 4.
Alexander however does not seem to take this refutation into
account. 33
We should also note that Alexander goes on to point to the fact
that the blood-vessels that grow from the heart are thicker near the
source and ramifY into ever thinner outshoots the farther they are
removed from this source (95.4-6) .34 Galen continues his plant
analogy (VI 3.13ff., cf. 6.43-44) in exactly the same way: the roots

Top. A 13,

105a20ff.
But Galen may owe something to the analogy between animals and
plants drawn by Aristotle at Iuv. 2, 468a13ff., where the nutritive part is associated with the middle part of the body; cf. also ibid. 3, 468b17ff.
32 Cf. the phrase ai <pAEPE~ ... El<: . 'tft~ Kapoia~ ... cb~ EK 't!XU'tll<; 'tTJV apxiJv
exouam 1:ft~ PA.aa1:~aeroc;. Alexander, like Aristotle, does not differentiate
between arteries and veins, speaking of <pA.ePe~ ('vessels') only.
33 On this parallel cf. Todd (1977) 118, Accattino (1987) 455fT.
34 Cf. also Arist. PA f 5, 668b36ff.
30

31

272

TEUN TIELEMAN

of plants are structurally similar to the veins sprouting from the

liver. So once again we may conclude that Alexander uses an
argument of the same form to arrive at an opposite conclusion.
But this is not the end of our list. Alexander's next argument
(95.6-8) resembles the one following in Galen (VI 6.20ff.), viz. the
nutritive principle must be in the organ which is the first to be
articulated in the embryo. Alexander claims that it is the heart for
which this is observed (<paivecr8at) to hold true, whereas Galen
argues that the observation of an embryo of a goat in utero confirms
that the liver is the source of the veins. Alexander also points to the
fact that the heart is the last organ to stop functioning. The idea that
the heart is the first to move and the last to stop (so that its activity
coincides with life itself) belonged to the traditional repertory of
cardiocentric arguments. It is cited and criticized by Galen at II
8.23-24. Here, however, it serves to establish the location in the
heart of the soul tout court. I assume that Alexander restricted the
scope of this argument to the nutritive power because, as we have
noticed, he is in need of arguments for the presence in the heart of
this power in particular.
Alexander, De an. 94.17ff and Galen, PHP VI 3-7 run to a large
extent parallel both as to the individual arguments used and as to
the order in which they are presented. The relation between the
two texts is characterized by the reversal of Galen's arguments.
That is to say, Alexander argues that the indications adduced by
Galen apply to the heart, not the liver (cf. 95.1-2). The term
'reversal' is all the more appropriate since Galen too constantly
opposes the heart and the liver as competing alternatives. 35
But there is more to be said. In Galen the distinctive properties of
the liver and heart serve as signs (<Jl'IJ!Eta, 'tEKJ!tlpta) of the essential
nature of these organs.36 This particular way of combining philosophical methodology-i.e. a basically Platonist method of divisioncum-definition-with anatomical lore is peculiar to Galen, notably
the argument of PHP.37
Alexander's battery of arguments concerned with the principle
35 This is of course because he directs his argument against the
Erasistrateans and Peripatetics, see e.g. VI 6.3ff. On the influence of
Aristotelian physiology on Erasistratus cf. Nat. fac. III 4, p.l65.7ff. Helmreich.
36 See Tieleman (1996) 55ff., 60ff.
3 7 See esp. the logical and methodological sections in Middle Platonist
handbooks, notably Akin. Did. c.5 and Clement, Strom. VII, esp. 7.17-21 on
the role of properties as signs.

GALEN AND ALEXANDER OF APHRODISIAS

273

of nourishment is not yet exhausted. He continues with a few arguments (95.12-96.15) which are used by other authors to establish the
location of the regent part. As far as Galenic parallels are concerned, we now have to switch from PHP VI to II. At II 4 Galen
takes issue with a set of Stoic and Peripatetic arguments 'from position' (II 4.6ff., cf. Arist. PAr 4). According to Alexander, the heart
must contain the principle of life since it occupies the midmost
position in the body, which is safest and least vulnerable to ailments (95.12-16; cf. Arist. PAr 4, 665b25ff.). 3 8 Galen does not discuss this particular train of thought, though he speaks broadly of
various possible arguments from position (II 4.12ff.). Alexander
points to the chest (8ropa~) as the proper body of animal on which
the organs have, as it were, grown; the common psychic faculty
must be in the common bodily part.39 Galen, for his part, refers to
an (otherwise unknown) argument which appeals to the middle
position of the heart in the chest (II 4.14). 40 Galen refers to the
originally Aristotelian 41 argument that the heart lies in the middle
of the body for the sake of an even distribution of the powers ( ouvaJlEOOV) sent out from the heart (II 4.12). Alexander uses this argument at 95.19-24, but restricts its scope to the distribution of nutriment and the nutritive principle for the reasons indicated above. 42
At 96.10ff. Alexander proceeds to the second faculty of the series,
viz. the perceptive (aicr8T)nlCTJ) faculty. He argues that the nutritive
and the perceptive powers must necessarily be located in one and
the same place, or else animals would have two souls instead of
one (and by the same token lose their individuality). 4 " For Galen
38

This point is also made with regard to the perceptive faculty, 98.20ff.

39 This could reflect Arist. Iuv. 2, 468al3ff., where Aristotle says that the

nutritive principle resides in the middle part of the body: some animals
(wasps, bees) can be seen to survive for some time without heads or food
receptacle (viz. the upper and lower parts of their structure). Cf. Alex. De an.
100.8ff. on which see infra p. 13f.
40 This may however reflect Aristotle's observation that in animals the
place of the heart is in the middle of the breast (whereas in man it inclines
slightly to the left), PAr 4, 666b7-9.
4 1 Arist. PAr 4, 666bl8-20, 666al4-15; MA 9, 702bl2-703a3; cf. Simp!. In
Arist. De caelo p. 514.11-3 Heiberg.
42 But cf. Arist. Iuv. 4, 469a24ff. where the same argument is also used
with regard to the nutritive-cum-perceptive function only.
43 Cf. 99.5 ff. where Alexander says that this implication is not
inescapable in the case of the theoretical intellect: we may accept that it exists
separately from the other functions without having to assume two soulswhich is impossible. The fact that he takes into account a counter-argument is
typical of his method, see infra, p. 278.

274

TEUN TIELEMAN

the doctrine that the psychic powers determine the shape of the
bodily organs is related to-and justified by-a scala naturae
expressed in terms of the three Platonic parts of the soul: plants
possess only the desiderative part, 44 non-rational animals have the
desiderative and the spirited part, human have the rational on top
of the other two. 45 Here the non-rational parts of the soul coincide
with the kind of soul possessed by non-rational animals. This scala
naturae thus brings out the status of the non-rational (i.e. the
desiderative-cum-spirited) as a distinct part of the soul. 46 This
emphasis is typical of Galen, who, in PHP and elsewhere,
describes the parts or forms ( etOT)) of the soul as distinct in kind
(yevn) and being (ouaia). On occasion he even speaks-in a way
which is unwarranted by the Platonic text-of a plurality of souls. 47
So here too Galen would seem to be the most likely target of
Alexander's critique.4B
44 Cf. Pl. Tim. 77b with Gal., In Tim. 11.24-27 Schroder. In the context
Galen connects this point to his view that the desiderative part is different in
essence (oucria) from the rational part; pace those who posit one essence with
three powers ( OUVclJ.1Et<;) Plato speaks of forms or parts. This may be directed
against Platonists like Severus who were swayed by Peripatetic criticism to
abandon soul-partition in order to save the unity (and immortality) of the
soul; cf. Deuse (1983) 104ff.
45 PHP V 6.37-9, where Galen ascribes this Platonizing chain of being to
Posidonius (F144 E-K). Posidonius, for his part, conceived of the soul in terms
of powers (OUVclJ.1Et<;) not parts, let alone separate souls; like Aristotle and most
Stoics he assigned the powers (i.e. the regent part) to the heart. See PHP V
4.3; 7.9-10; 7.50; VI 2.5 (Posid. Frs. 142, 144, 145, 146 E-K).
46 See supra, n. 18 with text thereto. Similarly, Alex. De an. 27.8ff. (who of
course speaks of OUVclJ.1Et<;), Nemes. Nat. hom. 16, 73.3-6 Morani.
4 7 PHPV 4.3, 7.9-10; 7.50; VI 2.5, In Hipp. Epid. p. 272.22-273.1 WenkebachPfaff. For a plurality of souls distinct in essence see esp. De meth. med. X,
635.6ff. K., In Plat. Tim. Fr. II 1.87 Schroeder. Cf. also De moribus, p. xvi
Walzer, where Galen presents it as one of the options available in the debate
concerned with the structure of the soul. On his conception of the soul against
the backdrop of contemporary Platonism cf. the concise discussion of Deuse
( 1983) lOOff.
48 It might be supposed that Alexander's target was the Pythagoreanizing
Platonist Numenius of Apamea, who is known to have distinguished between
a rational and a non-rational soul: Numen. ap. Porph. ap. Stob. Eel. I 49.25a
(350.25-351.1 Wachsmuth) = Numen. fr. 44 Des Places, Porph. fr. 253 Smith
(272.18 ff.). Here Numenius is mentioned as one among a plurality of authorities, who remain anonymous but may include Galen; cf. Nemesius referred
to supra, n. 46. Numenius' floruit is generally dated to around 150 CE or the
later 2nd century CE; Frede (1987) 1038 f. So as far as chronology is concerned
he might have been Alexander's butt. But on closer thought this seems far
less likely. The cosmic perspective motivating Numenius' doctrine is something quite different from the anatomical and physiological aspects highlighted by both Galen and Alexander. On Numenius's conception of the soul

GALEN AND ALEXANDER OF APHRODISIAS

275

But can we be reasonably confident about this anti-Galenic

purport? Is Alexander not barring a purely theoretical implication?
We would prejudge the issue if we were to argue that Alexander
knew this implication as being distinctive of Galen. Clearly the
anti-Galenic motivation of this passage has to gain. plausibility at
least to some extent from the indications to the same effect provided
by other passages. But if we add this passage to our list, it can be
seen to fit into the general pattern exhibited by Alexander's way of
dealing with Galenic views and arguments. As such, it presents
another instance of a straightforward reversal of Galen's position.
The point is well-chosen. Galen's emphasis on the spatial separation of the parts of the soul gave rise to serious problems as to their
coherence and interactions on the physiological level-a problem
he never adequately faced. 49
Finally I wish to call attention to 98.7-9:
One could learn that the regent part of the soul is in the heart also
from the structure (KataoKeuijc;) of the organ itself, which it has
received from Nature, which does nothing without reason.

Again, the colouring of this passage immediately recalls Galen.

What we have here is Galen's 'strong' teleology encapsulated in
his oft-repeated adage 'Nature does nothing without reason' .5
Considered in this light, distinctive properties of organs have to
pertain to their essence or function. This teleological perspective is
associated by Galen with the form, or structure, of organs in
particular. 5 1 In PHP it is a distinctive property of the heart that it is
connected with all the arteries, just as the brain is the centre of the
nervous system (PHP I 7). The liver is designated as the structural
centre of the veins with the same purport (PHPVI 5).
Of course, in a by now familiar way, Alexander draws a

see Deuse (1983) 62ff., esp. 79ff., Frede (1987) 1070 ff. At this stage, moreover,
he was not yet a factor on the philosophical stage; see the evidence surveyed by
Frede (1987) 1034 ff. Alexander, at any rate, nowhere refers to him explicitly.
49 Cf. Mansfeld (1991) 138ff.
50 I do not of course wish to imply that this dictum is confined to Galen,
cf. e.g. Arist. De an. 9, 432b21, 12 434a31, PA B 13, 658a5, f 1, 661b24; Alex. Fat.
178.12, De an. 27.9, 13, 28.5. The theme of Nature's design in shaping the
or~ans of the body is also prominent in Alexander's preface, 2.10-25.
1 Cf. also De usu parlium X 5.6, vol. 2, p. 75 Helmreich, Nat. fac. III 8,
p. 227.19ff., Helmreich. Similarly, Alex. De an. 2.19-20, in a context strongly
reminiscent of Galen's De usu partium, a work which, if we may believe
Galen, was popular among Aristotelians, see Lib. Prop. 2, p.100.18-23 Miiller.

276

TEUN TIELEMAN

conclusion opposite to that of Galen, arguing that the properties of

the heart indicate that it is the seat of the regent part. Most notable is
his reference to 'nerves' (veupa) on the heart, following Aristotle,
PAr 4, 666b14-16 (98.15; cf. 76.14ff.). It was Galen who had devoted
much effort in subverting this supposed observation, arguing that
what Aristotle saw were nerve-like tissues rather than nerves and
that the presence of nerves (if granted) is not peculiar to the heart
anyway (PHPI 8.3-15; 10.6-10; II 8.26-28).
4. Having established the heart as the seat of the regent part,
Alexander turns to the arguments advanced in favour of the brain
(99.30-100.17). On the whole these, he says, are 'empty'. Nonetheless, he singles out for treatment two of them which 'lay claim
to a certain degree of plausibility' (100.1-2). The first appeals to
clinical experience: mental disease is cured by tending to the
head, just as mental afflictions result from disease or damage of
the same organ. Alexander for his part points to the principle that
all parts are connected through sympathy, which he backs up with
an appeal to the rational design of the body: ' ... since every organ is
beautifully adapted (scil. to the other paris)' (100.3-4). In other
words, the idea of sympathy demolishes those arguments which
infer the presence of psychic functions from certain physical
effects or affections.52 Thus it can also be used against the stock
cardiocentric argument-used by Alexander himself (97.1-3)which appeals to the physical effects attendant on certain passions,
notably the pulsation of the heart in anger or fear. 53 Galen too
invokes the principle of sympathy when confronted with this
argument (PHP II 7). On the other hand he himself appeals to the
effect of afflictions of the head. 54 In addition, we should once again
note the distinctively Galenic colouring of much of Alexander's
subject-matter. 55 So it seems that Alexander again turns the tables
52 The history of this rejoinder can be traced back to a treatise as early as
the Hippocratic On the sacred disease (later 5th cent. BCE), ch. 17; cf. 14.
53 Chrysippus, arguing in his On the soul in favour of the cardiocentric
position, is sensitive to the counter-argument from sympathy too: see PHP III
5.41ff. with Tieleman (1996) 252ff.
54 See further the part played by sympatheia at Loc. aff III 5, with Mansfeld
(1990) 3110.

55 Cf. also Loc. aff. III 5, 8.158.1ff. K., Meth. med. XIII 21, 10.928.2-932.17 K.
with Mansfeld (1990) 3141ff., 3110 n. 225. Todd (1977) 122f. suggests that Alexander's argument is based on his misunderstanding of Caus. Symp. 7, 1.137
K., but this seems less likely; cf. Accattino (1987) 467f. Likewise Todd (1977)

GALEN AND ALEXANDER OF APHRODISIAS

277

on Galen. His point about sympathy produces a stalemate between

a pair of evenly matched (counter-)arguments.
A similar strategy is followed with regard to experimental observations of animals which stay alive for some time after their hearts
have been removed, thus warranting, it seems, the inference that
the heart cannot be the seat of the soul. Observations of this type,
Alexander points out, can be adduced in support of either side of
the debate (100.9: JCOtV<Xt 7tpo~ aJ.up6'tepa56). Cases of animals (e.g. the
tortoise and the chameleon) that can dispense with a heart for
some time are balanced by others (not specified) which survive, or
at least do not die instantly from, decapitation.57 Alexander adds
that both the heart and brain when removed often continue to exercise their proper functions too (100.9-13). In other words, insofar as
these observations are concerned, the stalemate between the opposing views persists.
Galen had adduced an observation of the same kind. Sacrificial
bulls whose heart lies already on the altar move and bellow in
panic until they die through loss of blood (PHP II 4.45-9). He adds
that the excised heart continues pulsating, thus indicating its
proper function-independent of the brain-in accordance with
the Platonic tripartition (PHP II 4.48) .58 Likewise he establishes the
mutual independence of the heart and the brain through experiments whereby the vessels (veins, arteries, nerves) connecting the
two organs are intercepted (PHP II 6, III 6). This point is echoed by
Alexander, who stresses that either organ, when separated from
the rest of the body, often continues to exercise its proper functions.
In the case of the brain, he may be thinking of the continuation of
the perceptive powers in certain kinds of animals. However this
may be, Alexander urges against those who point to the continued
121f. assumes that the same passage has inspired De an. 98.20-24 (on so-called
'cardiac disease'). But Alexander's point about not allowing oneself to be
misled by terminology has a closer parallel at PHP II 8.3-17 (on mpoaA.yia).
56 Cf. PHP IV 1.5, where Galen accuses Chrysippus of 'taking both sides'
(e7tCXIHPO'tepi~rov) in the debate over the number of parts of the soul. On the
affinities between the methodologies of Galen and Alexander, cf. also infra,
n. 61.
57 Similarly Cic. ND II 24 (SW 1.513), Tert. De an.15.6; cf. Mansfeld (1990)
3lllf.
58 Galen associates the pulse - as a non-rational motion independent of
the brain and essentially different from voluntary motion, with anger, fear
and sexual arousal as characteristic of the Platonic spirited part, cf. e.g. PHP
III 1.30ff., VIII 1.23; cf. II 7.17ff.

278

TEUN TIELEMAN

activity of the brain that the same holds good of the heart. If
intended as a rebuttal of Galen, this is beside the point, for the
latter-in accordance with the Platonic trilocation-assigns to
either organ its own distinctive motion. So the issue is not that of
function vs. loss of function. But if Alexander fails to meet Galen's
argument, that does not mean that he did not respond to it. It is
noteworthy that Alexander says that each organ, when removed
from the body, retains its 'proper activities (oiled a<; evepyda<;,
100.12f.)'. This very point is stressed by Galen in the context of the
same kind of observations, where it supports his anatomically
based Platonic tripartition (PHPII 4.49; II 6.10).59
Alexander's claim that arguments of this type tell in favour of
neither alternative recalls his view of the purpose of a dialectical
investigation in search of the truth in an Aristotelian manner. The
ability to discern the strong points-and argue oneself-in favour
of one of two opposed positions is integral to his concept of
dialectic.6 This made him hospitable to the Sceptical technique of
diaphonia. But the Aristotelian dialectician cannot of course rest
content with a stalemate of conflicting views and arguments, i.e.
he tries to balance the arguments pro and con in order to determine which of two tenets is to be accepted, or at least preferred. Jaap
Mansfeld has established and analysed Alexander's use of this
technique on the basis of his De fato. 61 For our purposes it is
interesting to note that it is part of this procedure that Alexander is
interested not only in arguments or facts supporting either of the
views at issue, but also in points to be scored against both of them.
Those of the two positions which appears more vulnerable to these
objections, and hence more complicated or muddled, is the one to
be discarded in favour of the other. Indeed, the thesis to be accepted
becomes clearer insofar as the alternative can be shown to be
muddled. 62 Hence the exposure of the invalidity of both the
encephalocentric and cardiocentric arguments from sympathy at
59 Cf. also PHP I 6.10, III 6.5, V 3.5, VI 6.44.

In Top. 27.7-29.16 Wallies; In Met. 173.5-174.7 Hayduck.

The procedure of starting from an existing disagreement with a view to
eventually accepting one of the available options can be traced back to Chrysippus; see supra, n. 25. Mansfeld (1987) 185 f. points to PHP V 4.1, where Galen
argues that one should suspend judgement whenever opposed views are of
more or less equal credibility, but that if one of them turns out to be far more
credible than the other, we have to assent to it as being true.
62 See esp. Fat. 164.21-165.8 Br. with Mansfeld (1987) 184 f.
60
61

GALEN AND ALEXANDER OF APHRODISIAS

279

100.2-13. But, as we have noticed, the preceding argument too

bristles with rebuttals of what can be found in Galen's work.

5. What are we to conclude from the preceding survey of resemblances and differences ? Many of the arguments used by Alexander had started their career in Aristotle's PAr 3-4 and Iuv. 2-4 and
found their way into the traditional debate concerned with the seat
of the soul, where they served, often in a somewhat modernized
form, on the Peripatetic side. Alexander uses some of these without
adapting them to meet Galen's objections. It would however be
rash to assume that he did not know, or chose to ignore, Galen's
critique and simply drew on a repertoire of traditional Peripatetic
arguments. Not only would such a dialogue des sourds be implausible in the light of his knowledge of Galen's work. 63 Alexander has
been shown to use specifically Galenic notions and arguments
and to do so in a way which can best be explained on the
assumption that he had indeed read the Galenic passages I have
been referring to.64
Large parts of Alexander's argument read like an inverted mirror image of Galen's demonstration in PHP books II and VI. The
reason is that Alexander applies the technique of turning Galen's
arguments and concepts against him, just as he elsewhere appropriated Stoic concepts and arguments to anti-Stoic ends. 65 This is
the familiar technique of reversal (1tepnp01t1l) in its wider sense of
taking a statement of one's opponent and turning it against him. 66
The opposition of two alternative views (or arguments) is also at
issue in the so-called Placita tradition as reconstructed by Diels in
his Doxographi Graeci. 67 Jaap Mansfeld has demonstrated Galen's
knowledge of this tradition on the basis of De locis affectis III 5. 68
63 See supra, n. 11.
64 See supra, p. 3.
65 Cf. Todd (1976) 27f., Moraux (1984) 335-6, 359, Sharples (1987) 1178 with

further references.
66 The wider sense gained currency under the influence of the rhetorical
tradition; cf. e.g. Arist. Rh. B 23, 1398a3-4. The technical term 7tEpttpo7t~ stems
from dialectic, where it designates self-refutation in the narrower sense of
refuting a thesis from its own contents. One of the most famous instances in
ancient philosophical literature occurs in Pl. Th. 17la-b, where Protagoras'
Man is the Measure-thesis is turned against itself. See further Burnyeat
(1976), esp. 65.
67 Berlin 1879; several reprints. A new study devoted to the so-called doxographic tradition is being prepared by Jaap Mansfeld and David T. Runia.
6'!! Mansfeld (1990) 3141ff.

280

TEUN TIELEMAN

Here Galen presents a list of issues-including that of the location

of the regent part-which runs closely parallel to the relevant
sections in the Placita tradition as to both content and wording.
Interestingly enough, Galen says that the many different statements about the regent part are the subject of dialectical disputation,
thus affording a rare glimpse of the Sitz im Leben of doxographic
schemas. 69 In addition, we must note the question-type familiar
from the Placita tradition, which opposes two alternatives ('whether
P or not-P' or 'whether X or Y). This recalls the so-called 9crt~
('theoretical problem', 'general question') as a concept of ancient
dialectic and rhetoric.7 Indeed, Clement of Alexandria (150-c. 215
CE), Strom. 8.14.4 mentions the question of the location of the regent
part ('The regent part: is it in the heart or in the head?') as a typical
instance of a 9crt~. A 9crt~ is marked by its stress on the opposition
between two views at issue. Both of them are presented, though one
is eventually to be preferred. Clearly, this suits Alexander's concept
of dialectic. 71
It may not be too far-fetched to compare the ~tK'tUaKa ('HuntingNet Arguments') by Dionysius of Aegae, in which fifty medical
theses were both defended and refuted in a way reminiscent of the
much earlier ~tcrcrot A.6yot ('Double Arguments,' DK 90).7 2 Unfortunately, only the theses have been preserved (Photius, Bibl. Codd. 211
and 185).73 But these certainly attest to the gambit of opposing and
reversing medical tenets. No. 49 is: on
7tEpt Kapotav 'tO
OtaVOTJ'ttKOV, UAAU 7tEpt KEcpaATJV Kat O'tt ava7t<X.AtV. Other theses (viz.
nos. 36-50) are also similar or identical to those discussed by Galen
in PHP and Alexander in the final section of the De an. Thus nos.
41 and 42, concerned with the question whether or not the heart or
the liver is the source of the veins, correspond to PHPVI 3-6 and De

ou

Loc. aff. 8.157.17-18 K

On the relation of doxography to the 9Ecrt~ see Mansfeld (1990) 3193ff.
See further Throm (1932); Runia (1981) 116f.
71 For Alexander's view on the 9Eat~ see In Top. 78.21-83.2; cf. also 27.12ff.,
82.19ff. Wallies. At 76.1ff. (in the section dealing with the wider concept of
7tp6~1..1]J.UX) we have three physical issues familiar from the Placita tradition,
which he of course knew: 'whether the world is eternal or not'; 'whether the
world is spherical or not', 'whether the soul is immortal or not'; see Aetius,
Plac. II 4, II 2, IV 6.
72 Photius' preface indicates pairs of argument and counter-argument, see
p.336.1-6 Deichgraber. 'Antilogies' of this type were subsumed under the
9Eat~, see Alex. In Top. 27 .17f. Wallies with Throm ( 1932) 170f.
73 Also printed, with a brief introduction, by Deichgraber (1930) 335-340.
69

70

GALEN AND ALEXANDER OF APHRODISIAS

281

an. 94.25ff. 74 No. 48 dealt with the question whether or not the heart
is the source of the nerves (cf. PHP I 8, 10; De an. 98.15).75 The
formulation in terms of polarities recalls the theses in the doxographic texts. The phrase Kat on ava1taA.w ('and in reverse', 'and contrariwise') may indicate the reversal of argumentation exemplified
by the (lost) pairs of arguments.
According to Photius, Dionysius' booklet was 'quite useful for
dialectical exercise and the knowledge of some doctrines belonging to medical theory' (Cod. 185). It has been assumed that the pairs
of arguments were designed to induce suspension of judgement. 76
But we have to be cautious in drawing precise conclusions about
historical affiliations. The exact dates and identity of the author of
the ~uctua.Ka must remain uncertain (though the issues listed seem
to presuppose Galen's work).77 But it is tempting to assume that the
lost arguments reflected the same kind of method as is followed by
Alexander in the closing section of his De an.
To conclude. We have found a number of parallels which relate
Alexander's argument to Galen's in the various ways I have been
arguing. In the light of this evidence we may, in Alexander's own
fashion, tip the scales in favour of the thesis that De an. 94.7-100.17
is primarily designed as a response to Galen. Particularly striking
are Alexander's use of the sorites and his technique of inverting
Galen's arguments. Thus Alexander succeeds in exposing a few
weak (and blind) spots of Galen's position, especially those relating
to the spatial separation between the 'parts' of the soul. The
cardiocentric view still had a future lying ahead.*
Utrecht University
74

See supra, p. 27lf.

75 See supra, p. 276.

7 6 For this reason the author has been identified with a Dionysius who
was a Methodist physician of the 1st cent. CE. Alternatively, he has been
described as an Empiricist and Sceptical doctor. But his identity must remain
uncertain. Eustathius, In ll. 1192 (= vol. 4, p. 355.10-13 Van der Valk) aligns
the author of the AtJC't\HXJCa with the dialecticians and the Sceptics; cf.
Deichgriiber (1930) 336.
77 E.g. No. 15 at Cod. 185 'that heat does not vary in quality' clearly states
the view of Lycus of Macedonia as discussed by Galen, Adv. Lye. c. 3, pp. 8.1514.4 Wenkebach. Lycus belonged to the generation of Galen's teachers; cf. De
anat. admin. 2.449f., 470 K
I should like to thank Keimpe Algra, Han Baltussen and David Runia
for their comments on an earlier version. Research for this article was made
possible by a fellowship of the Royal Dutch Academy of Arts and Sciences.

282

TEUN TIELEMAN

BIBLIOGRAPHY

Accattino, P. (1987) 'Ematopoiesi, malatti cardiaca e disturbi mentali in

Galeno e in Alessandro di Afrodisia', Hermes 115: 454-473.
Barnes, J. (1982) 'Medicine, experience and logic', in J. Barnes et al. (eds.)
Science and Speculation. Studies in Hellenistic Theory and Practice (Cambridge
1982) 24-68.
Burnyeat, M.F. (1976) 'Protagoras and Self-Refutation in Later Greek Philosophy', Philosophical Review 85: 1, 44-69.
- - , (1982), 'Gods and Heaps', in M. Schofield and M.C. Nussbaum (eds.),
Language and Logos. Studies in ancient Greek philosophy presented to
G.E.L. Owen (Cambridge).
Deichgriiber, K. (1930) Die Griechische Empirikerschule. Sammlung der Fragmente
und Darstellung der Lehre (Berlin).
Deuse, W. (1983) Untersuchungen zur mittelplatonischen und neuplatonischen Seelenlehre, Mainz.
De Lacy, Ph. (ed.) (1978-84) Galen: On the Doctrines of Hippocrates and Plato,
Corpus Medicorum Graecorum V, 4, 1, 2 (Berlin).
Donini, P., (1974) Tre studi sull'Aristotelismo nel II sec. d. C., Torino.
- - , ( 1982), Le scuole, ['anima, l'impero: la filosofia antica da Antioco a Plotino,
Torino.
Frede, M. (1981) 'On Galen's Epistemology', in V. Nutton (ed.), Galen:
Problems and Prospects (Cambridge) 65-86; repr. in id., Essays in Ancient
Philosophy (Oxford 1987) 279-98.
- - , (1987) 'Numenius', ANRWII 36.2, 1034-75.
Mansfeld, J., (1988) 'Diaphonia: the Argument of Alexander De Jato Chs. 1-2',
Phronesis 33: 181-207.
- - , ( 1990) 'Doxography and Dialectic: The Sitz im Leben of the Placita', ANRW
II 36.4, 3056-3229.
.
- - , (1991) 'The Idea of the Will in Chrysippus, Posidonius, and Galen,' in
JJ. Cleary (ed.), Boston Area Colloquia on Ancient Philosophy series, Vol. V,
107-45.
Moraux, P. (1978) 'Le "De anima" dans Ia tradition grecque. Quelques aspects
de !'interpretation du traite de Theophraste a Themistius', in G.E.R.
Lloyd and G.E.L. Owen (eds.) Aristotle on Mind and the Senses (Cambridge)
281-324.
- - , (1984) Der Aristotelismus bei den Griechen. 2. Band. Der Aristotelismus im
I. und II. Jh. n. Chr. (Berlin/New York).
Nutton, V. ( 1984) 'Galen in the eyes of his contemporaries', Bulletin of the
History of Medicine 58 (Baltimore) 315-24; repr. as Study nr. III in id., From
Democedes to Harvey: Studies in the History of Medicine. Variorum Reprints
(London 1988).
- - , (1987) 'Galen's Philosophical Testament: "On my own opinions" ', in J.
Wiesner (ed.), Aristoteles. Werk und Wirkung (Berlin-New York) 2.27-51.
Pines, S. (1961) 'Omne quod movetur necesse est ab aliquo moveri: A Refutation of Galen by Alexander of Aphrodisias on the Theory of Motion', Isis
52: 21-54.
Rescher, N. and Marmura, M.E. (1965) The Refutation by Alexander of Aphrodisias of Galen's Treatise on the Theory of Motion. Translated from the
Medieval Arabic Version, with an Introduction, Notes, and an Edition of
the Arabic Text. Islamic Research Institute, Islamabad, Pakistan.
Runia, D.T. (1981) 'Philo's De aeternitate mundi: The problem of its Interpretation', Vig. Chr. 35: 105-151; repr. in Exegesis and Philosophy: Studies on
Philo of Alexandria (London 1981).

GALEN AND ALEXANDER OF APHRODISIAS

283

Sharples, R.W. (1987) 'Alexander of Aphrodisias: Scholasticism and

Innovation', in W. Haase (ed.), Aufstieg und Niedergang der R.Omischen Welt
II 36.2 (Berlin-New York) 1177-1243.
- - , (1990) 'The school of Alexander ?', in R.R.K. Sorabji, Aristotle Transformed; the Ancient Commentatt:rrs and their Influence (London), 83-111.
Temkin, 0. (1973) Galenism. Rise and Decline of a Medical Philosophy (Ithaca/
London).
Throm, H. ( 1932) Die Thesis. Ein Beitrag zu ihrer Geschichte und Entstehung
(Paderborn).
Tieleman, T. (1996) Galen and Chrysippus on the Soul. Argument and Refutation in
the De placitis Books II-III. Philosophia Anti qua 64 (Lei den).
Todd, R.B. (1976) Alexander on Stoic Physics (Leiden).
- - , (1977) 'Galenic Medical Ideas in the Greek Aristotelian Commentators',
Symbolae Osloenses 52: 117-34.

DOTI NATURAL!, ABITUDINI E CARATTERE NEL

DE FATO DI ALESSANDRO
PIERLUIGI DoNINI

Di Alessandro di Afrodisia lo studioso a cui e dedicato questo

volume ebbe occasione di occuparsi pili di una volta; tuttavia lo
scritto De Jato e fra le opere del commentatore fino a oggi l'unica
cuijaap Mansfeld abbia dedicato un intero saggio. Non ripetero qui
quale importanza abbia quel lavoro in vista dell'interpretazione
complessiva dello scritto; prendero invece in considerazione un
piccolo suggerimento, a mio giudizio estremamente fecondo, che
Mansfeld diede in quelle pagine. Egli prometteva allora di ritornare successivamente sulla questione: il che non essendo finora
avvenuto, spero di fare cosa non sgradita al destinatario del presente volume se mi servo dello spunto da lui fornito non certo con
la pretesa di indovinare le sue intenzioni, rna nella speranza di far
progredire la discussione e di attenuare intanto una notevole difficolta di interpretazione che a suo tempo io stesso avevo sollevato.l
Occorre in primo luogo ricordare i termini del problema, che
concerne la possibilita di riconoscere ad Alessandro e al suo trattato
una sostanziale coerenza di pensiero fra le due parti in cui lo scritto
si articola (cioe i capitoli 1-62 contenenti la dottrina del fato che
Alessandro considera aristotelica e la parte rimanente con la sua
lunga polemica contro il determinismo). L'obiezione che molti
anni addietro3 avevo mosso ad Alessandro era che trattando nel
cap.6 il carattere degli uomini come risultante dalla 'naturale costituzione'4 dell'anima (e quindi come portatore del destino, almeno
1 II saggio di Mansfeld cui mi riferisco e 'Diaphonia : the Argument of
Alexander De Fato Chs.1-2', Phronesis 33 (1988) 181-207. Come Ia tesi illustrata
in que! lavoro contribuisca a chiarire Ia natura e le finalita del De Jato ho
ar~omentato nel saggio citato qui sotto (n. 23) 5043 n. 54.
0 meglio, a rigore, 3-6 dato che Mansfeld, art. cit., ha persuasivamente
mostrato che entrambi i capitoli 1-2 servono come proemio.
3 Nel libro Tre studi sull'aristotelismo nel II secolo d.C. (Torino 1974) 164-165 e
171-173.
4 De Jato 6, p. 170,17.20. Se Alessandro puo parlare di una costituzione
naturale (physike kataskeue) dell'anima, mi pare sicuro che egli si riferisce
implicitamente alia sua dottrina dell'anima come risultante dalla
mescolanza degli elementi (per cui cfr. De anima p. 2,25-11,13).

DOTI NATURAL!, ABITUDINI E CARATTERE NEL DE FATO

285

nelle intenzioni di Alessandro ormai dimostratosi identico alla

natura) 5 egli contraddiceva insieme Aristotele e se stesso: il suo
maestro, in quanto il commentatore attribuiva unilateralmente alle
doti naturali il peso determinante nella formazione del carattere
ignorando il fatto che nelle Etiche 6 Aristotele aveva spiegato che il
carattere e il risultato delle abitudini contratte fin dalla prima
fanciullezza nel corso dell'educazione e in generate dei processi di
socializzazione; e poi anche se stesso, in quanto nelle pagine successive del trattato (e specialmente nei capitoli 27-29) Alessandro
avrebbe molto piu correttamente riprodotto la teoria aristotelica
ridimensionando, o meglio addiritura tacendo l'apporto delle doti
naturali in vista della formazione del carattere morale e facendo
derivare l' ethos dalle abitudini contratte dopo la nascita. Non e poi
necessario riferire qui anche tutti gli interventi successivi nella
discussione, dato che sia la difesa di Alessandro tentata da qualche
studioso, 7 sia i rilievi che a mia volta mossi a questi tentativi ritenendo di confermare le mie precedenti obiezioni8 mi sembrano
oggi non conclusivi e forse diventano completamente superflui
qualora si raccolga e si sviluppi il suggerimento espresso nell'articolo di Mansfeld. 11 quale fu appunto l'unico a notare 9 che gli
esempi di cui Alessandro si serve nel cap.6 (pp. 170,21-171,4) al
fine di provare l'identita del fato con il carattere naturale sono tutti
quanti 'esempi di Kcn:fat nel senso di Aristotele EN 11-V'.
5 Questo e appunto il risultato della discussione del De Jato nei capitoli 3,
p. 166,15-6, p. 170, 9. Non dico nulla in questa sede dell'altra maggiore
difficolta del cap.6, il rapporto fra Ia natura di cui Alessandro parla nelle
pagine ora citate e 'I' oikeia physis di ciascuno' (individuo o tipo?) di cui parla
nel seguito della pagina 170.
6 EN II, 1-4; cfr. EE II 1 (1220a 13 sgg.) - 2.
7 Mi riferisco soprattutto a P. Thillet e all'introduzione alia sua edizione
del De Jato, Alexandre d'Aphrodise, Traite du destin (Paris 1984) CXV-CXXIma a dire i1 vero io _non avevo affatto detto, come apparirebbe dal sunto di
Thillet a p. CXVI, che Ia natura 'determine totalement chaque individu'
secondo Ia teoria esposta nel cap. 6. Avevo spiegato a p. 164 del libro citato
sopra (n. 3) che Ia causalita naturale secondo Alessandro e operante solo per
Ia maggior parte dei casi e che i1 caso di Socrate, di cui Alessandro parla alia
fine del capitolo, p. 171,11-16, e appunto una delle eccezioni ammesse dalla
regola del 'per lo piu'. Ho cosi l'impressione che Ia critica che mi muove
A. Zierl, in Alexander von Aphrodisias, Ueber das Schicksal (Berlin 1995) 160,
dipenda piuttosto dalla lettura delle pagine di Thillet che da quella del mio
libro.
8 Nel saggio 'II De fato di Alessandro. Questioni di coerenza', in AuJstieg
und Niedergang der riimischen Welt, herausg. von W. Haase und H. Temporini,
II. 36.2 (Berlin-New York 1987) 1246-1247.
9 Art. cit. (n. 1) 182 n. 4 sub fin.

286

PIERLUIGI DONINI

Questa preziosa osservazione ha soltanto bisogno di essere difesa

e precisata per quanto riguarda il tipo del Kaptepuc6c; che compare
in 170,25 sgg.; 10 esso infatti, cioe il tipo dell'uomo resistente o dotato
di fermezza, non e di per se presentato come un carattere vizioso
nei libri dell' EN dedicati all'analisi delle virtu etiche e anzi
compare nella tabella delle virtu in EE II 3, 1221a 9 come una
medieta buona fra gli estremi della mollezza e della KaKonaSna
(che e l'attitudine a sobbarcarsi ogni fatica, pena o strapazzo). Se
invece, come appare da 170,26, Alessandro ha trasferito sulla
Kaptepia precisamente quell'attitudine al JCaJConaSei.v che per I'EE
caratterizzava un estremo vizioso, sembra chiaro che egli ha inteso
farne appunto una qualita non positiva; e lo spunto per una simile
valutazione puo essergli venuto da EN VII (= EE VI) 7, 1150a 36
dove egli trovava che per Aristotele Ia continenza (eyKpcl'teta), che e
gia di per se non identica alia virtu (secondo ENVII 1, 1145a 36) e
di questa e certamente meno apprezzabile, era comunque a sua
volta ancora preferibile alia Kaptepia. Era facile concludere che in
questo modo per Aristotele Ia Kaptepia veniva a trovarsi a due gradi
di distanza dalla virtu etica. II carattere eccessivo che Alessandro, o
quanto meno Ia sua scuola, attribuiva alKapteptK6c; e del resto con
an cora maggior chiarezza fatto rilevare dall' opuscolo sui fa to conservato al termine della Mantissa. 11 Di conseguenza, siamo effettivamente autorizzati a interpretare tutti gli esempi di Alessandro a
p.170,21-171,4 come riferiti a caratteri inferiori alia virtu etica e
genericamente identificabili con il vizio.12
Si aggiunga ora a tutto cio Ia considerazione che nel cap.6 del
De Jato l'unico esempio di un carattere che fuoriesce dal condizionamento naturale e 'grazie all'esercizio fondato sulla filo
sofia' risulta alia fine 'migliore della sua natura' (171,11-16) e
quello di Socrate: dunque il caso raro e positivo di un individuo
10 II suo nome e integrato da Bruns a 170,25 sulla base del riscontro fornito dall'opuscolo sui fato in Mantissa 185,29. L'integrazione si puo considerare
sicura (curiosamente invece Thillet, p.10,25 della sua edizione, preferisce un
plurale).
II P.185,29 imepJ3oA.fJ. Sulla questione d'autenticiti dell'opuscolo non e qui
necessario pronunciarsi .
12 Si noti per di piu che nella proposizione di 171,4-7, che e evidentemente destinata a commentare tutta l'esemplificazione precedente includendo
dunque anche il caso della karteria, Alessandro parla di 'mali presenti'
riferendosi a quelle situazioni in cui ognuno dei tipi illustrati sarebbe meritatamente caduto. Questo conferma il loro carattere comune, negativo perche o
identico o assimilabile al vizio.

DOTI NATURAL!, ABITUDINI E CARATTERE NEL DE FATO

287

eccezionale 13 il quale smentisce Ia regola del 'per lo pili' che

vincola invece Ia maggior parte dell'umanita al destino gia segnato
nella costituzione naturale dell'anima. La situazione presupposta da
Alessandro nel cap. 6 e pertanto chiara: l'idea fondamentale e che
il destino e identico alia natura interpretata nel senso, perfettamente ammissibile per un aristotelico, 14 di una causalita che opera
solo per Ia maggior parte dei casi e non di necessita, ammettendo
dunque eccezioni. In quanto e un agente naturale, 'per lo pili' il
carattere iscritto nella costituzione naturale dell'anima determina
Ia qualita morale e il destino degli uomini; rna questo condizionamento ha soltanto esiti moralmente deteriori: i caratteri condizionati dalla costituzione naturale (cioe quelli della maggior parte
degli uomini) sono caratteri moralmente inferiori alia virtu. Le
eccezioni previste dalla statistica aristotelica sono invece quelle di
chi e capace di migliorare Ia propria natura grazie all'educazione
ricevuta-o anche, eventualmente, grazie a un'autoeducazione
filosofica, come sembra probabile che si debba pensare relativamente a Socrate. Si registra allora immediatamente una prima,
benche ancora solo parziale concordanza fra il cap. 6 e il 27, dato
che anche secondo quest'ultimo testo l'esistenza di uomini virtuosi
e cosa estremamente rara:
infatti non vediamo che in possesso delle virtu sono tutti, oppure la
maggior parte degli uomini (il che sarebbe il segno delle cose
che accadono conformemente a natura); rna bisogna con ten tarsi
di trovarne anche uno solo, il quale mediante I' esercizio 1 5 e
l3 II che P. Thillet op. cit. (n. 7) CXVII n. 4 sembra voler mettere in
dubbio. Confesso di non comprendere bene le sue argomentazioni e osservo a
mia volta che, se Ia costituzione naturale determina 'per lo piil' Ia sorte degli
uomini (come Alessandro piil volte ripete nelle pp. 170-171), ci devono per
forza di logica essere anche dei casi che sfuggono alia regola del 'per lo piil' e
allora mi sembra del tutto legittimo considerare questi casi come 'eccezioni'
(cioe eccezioni alia regola, alia norma, alia maggioranza statisticamente
rilevata). Che Socrate sia precisamente una di queste eccezioni non mi pare
che possa essere materia di dubbio: egli e appunto riuscito a essere 'migliore
della sua natura' (171,16). Del res to, se non si considerasse eccezionale il suo
caso, si perderebbe anche un importante motivo di concordanza con il cap. 27
(si veda il seguito della discussione sopra nel testo).
14 Cfr. p.es. Aristot. Rhet. I 10, 1369a 35-b 2; Phys. II 8, 198b 35 ; An.pr. I 3,
25b 14 (altri passi in Bonitz, Index aristotelicus 836a 43-48). Quanto ad Alessandro, che Ia natura ammetta eccezioni e ripetutamente segnalato in De Jato 6:
cfr. specialmente 169,28-170,9; 170,14-16 e forse anche 16-18, se para nella
linea 17 deve essere interpretato come vuole R.W. Sharples, Alexander of Aphrodisias on Fate (London 1983) 47 e 129.
15 acrKTJO"tc;, come nel caso di Socrate in 6, p. 171,16. Alia P 198,23 leggo e
traduco <lva.yKa.'iov seguendo il ms. Marciano, Bruns e Orelli. Sharples

288

PIERLUIGI DONINI

l'insegnamento mostra Ia naturale superiorita dell'uomo sugli altri

animali, dato che da se aggiunge quel che e necessario e che
mancava alia nostra natura ( 198,19-23).

L'accordo con il cap. 6 e qui davvero notevole: prevalenza statistica dei caratteri deteriori; rarita estrema delle persone virtuose; la
virtu raggiunta mediante 1' esercizio e 1' educazione; l'incapacita
della physis a conferire da sola la virtu-su tutti questi punti i due
testi implicano o sostengono le stesse posizioni, per di piu sempre
all'interno della generale nozione aristotelica della natura come
agente che opera 'sempre o per lo piu' . C' e tuttavia un aspetto della
questione per il quale la coincidenza delle tesi espresse da Alessandro nei due passi non e totale. Nel cap. 6 la possibilita di migliorarsi
rispetto a quanto comporterebbe la natura appare si fuori della
normalita statistica (e infatti quanto rimane al di fuori del 'per lo
piu', che opera nel senso di vincolare la maggior parte dell'umanita al suo carattere naturale); rna non e cosi drasticamente
limitata come nel cap. 27 a un solo individuo: Socrate e certamente
un caso eccezionale, rna nel cap. 6 non e detto che sia anche
l'unico. Come spiegare questa difformita di posizioni?
Ora la tesi dell'estrema rarita degli uomini virtuosi puo essere
fatta derivare anche da Aristotele, e vero;I6 rna la formulazione di
Alessandro in De jato 27, p. 198,20 sembra davvero eccessiva per un
aristotelico e di fatto viene a coincidere con la tesi stoica. 11 risultato
e paradossale: il medesimo autore che si sforza lungo tutto il trattato
di distinguere la filosofia aristotelica da ogni possibile versione del
determinismo enuncia in quel passo una proposizione che altrove
rimprovera aspramente agli avversari stoici di aver sostenuto;I7 e la
enuncia senza che nemmeno ve ne sia la stretta necessita: non gli
bastava forse dire che gli uomini per bene non essendo tali per
natura-perche allora sarebbero buoni tutti quanti gli uomini, o
almeno la loro maggioranza-il loro numero si restringe a un
piccolo gruppo? Perche mai dire, invece, 'a uno solo'? Puo ben
darsi, dunque, che sugli avversari di quella filosofia la pressione

(seguito poi da Zierl) accetta invece l'avverbio O.vayKairo~ di tradizione indiretta. Si introdurrebbe cosi una nozione di necessita assoluta che mi sembra
quanto di piil sgradito ed estraneo all'aristotelism9 ci possa essere per Alessandro. 'AvayKaiov si intende invece facilmente nel senso di Metaph. b. 5,
1015a 22-26: e 'que! che e necessario perche il bene si realizzi'.
l6 Come osserva Sharples, op.cit. (n. 14) 162.
17 Cfr. De Jato 28, p. 199,16 sg.

DOTI NATURAL!, ABITUDINI E CARATTERE NEL DE FATO

289

delle tesi stoiche fosse ancora fra il secondo e il terzo secolo tale, da
imporre come ovvie persino tesi che a una riflessione critica attenta
avrebbero dovuto essere rifiutate; rna e anche vero che nel complesso del De Jato Ia concezione di Alessandro non e comunque
pienamente identificabile con quella stoica. Puo essere assimilata
alia posizione stoica Ia tesi che ci siano, o che siano esistiti, soltanto
uno o due uomini per bene; rna certamente Alessandro non accetta
mai il corollario stoico che ne consegue a proposito della rimanente umanita-che si tratterebbe cioe di una massa di stolti, folli e
malvagi tutti ugualmente immersi nell'errore e nel vizio. Ben lontano da queste convinzioni, Alessandro le censura anzi esplicitamente in una pagina del De fato. 18 La sua convinzione 19 sembra
dunque piuttosto essere che l'umanita comune si trovi in una
condizione che e certamente ancora lontana dalla virtu, rna si
colloca tuttavia in una posizione in qualche modo mediana fra la
virtu e il vizio completo. Quel che allora occorre ammettere e che,
almeno in vista dell'argomento del cap.6, Alessandro ha deciso di
non tenere conto di alcuna differenza fra i caratteri che risultano
comunque inferiori alia virtu: tutti risultano assimilati dal fatto di
essere legati al carattere naturalmente costituitosi nell'anima.
Prima di proseguire nell'analisi del nostro problema vorrei
segnalare un aspetto interessante dell'esemplificazione di Alessandro a p. 171, perche esso puo avere qualche importanza in relazione
alia (ancora irrisolta) questione del rapporto fra il trattato maggiore
e l'opuscolo sui fato incluso nella Mantissa. Nel suo articolo Mansfeld segnalo20 come 'a flagrant contradiction of major importance'
il fatto che nella pagina del trattato l'esempio di Socrate serve a
illustrare Ia rilevante eccezione che si puo introdurre al dominio
del fato sulle vite umane, mentre nell'opuscolo (p. 185,14 sg.) lo
stesso Socrate (in compagnia di Callia) e presentato come esempio
della regola generale che il fato governa gli individui. E apparentemente questa contraddizione dovrebbe allora corroborare il sospetto21 che l'opuscolo non sia opera di Alessandro, rna di un suo
scolaro. Dopo aver accettato per qualche tempo l'osservazione di
Mansfeld sono ora giunto alia conclusione che essa e infondata e

18

Cap. 28, p. 199,18-22.

19 Come si potrebbe anche indurre da altre pagine della sua opera. Si veda
in frroposito il mio saggio citato sopra (n. 8) 1254-1256.
0 182 nota 4, sub fin.
21 Espresso da Mansfeld nella stessa pagina gia citata qui sopra, n. 20.

290

PIERLUIGI DONINI

che 1' uso del nome di Socrate nei due esempi non puo servire a
discriminare in nulla le due diverse opere in cui lo leggiamo.
Questo perche, come mostra precisamente il fatto che esso compaia
abbinato con il nome di Callia, nell'opuscolo il nome di Socrate
deve essere letto, secondo l'uso aristotelico ben noto,22 senza alcun
preciso riferimento al personaggio storico: e soltanto un modo di
riferirsi a persone reali, rna la cui individuazione precisa e irrilevante: e l' equivalente, insomma, dei nostri Tizio, Caio o Sempronio.
Diversa sarebbe stata l'importanza del caso se l'opuscolo avesse
citato Socrate in riferimento o alla stessa situazione in cui compare
nel trattato (il giudizio espresso su lui da Zopiro), o comunque a un
episodio preciso della vita del personaggio storico di quel nome.
Siamo forse ora in grado di capire il ragionamento di Alessandro benche esso rimanga largamente implicito e possiamo finalmente rendere conto in modo coerente tanto delle affermazioni del
cap. 6, quanto di quelle dei capitoli 27-29. E' verissimo, infatti, che
in questa serie di capitoli, apparentemente contraddicendo il 6, si
riconduce la formazione del carattere all'educazione, all'esercizio,
alle abitudini e non pili alla natura; rna si doveva aggiungere (cosa
che finora nessuno ha fatto) che in quei capitoli Alessandro intende parlare soprattutto e, anzi, quasi esclusivamente della formazione dei caratteri virtuosi, cioe del modo in cui si diventa cpp6vtJlot :
solo pochissimi e brevissimi sono gli accenni alia formazione dei
caratteri viziosi (e di questi ci occuperemo in un secondo tempo). Se
dunque il cap. 6--quando sia stato debitamente interpretato l'esempio di Socrate-concorda con 27-29 quanto all'ammissione che una
personalita virtuosa, formata come tale dalle abitudini contratte con
l'esercizio e l'educazione, e realmente un'occorrenza molto rara e
rappresenta un'eccezione alla regola che dice che Ia maggioranza
degli uomini e inferiore alla misura della virtu, tutto il discorso di
Alessandro puo assumere una sua coerenza anche se lascia
implicite troppe cose e sorvola inoltre su difficolta tutt'altro che
marginali. II significato complessivo potrebbe essere questo.
II fato e davvero identico alla natura, come vuole il cap. 6, e cioe
e identico negli uomini al destino gia iscritto nella costituzione
naturale della loro anima. Ma questo dominio del fato non si estende alla totalita degli uomini: come prevede Ia dottrina di Aristotele,
la natura puo funzionare come agente decisivo soltanto per una
22 Cfr. Bonitz, Index aristotelicus, 359b 56 e 74lb 45.

DOTI NATURAL!, ABITUDINI E CARATTERE NEL DE FATO

291

maggioranza di casi. Dunque, soltanto per la maggior parte dei suoi

rappresentanti individuali l'umanita non andra al di la del carattere iscritto nella costituzione naturale (e fatale) dell'anima: le cui
potenzialita devono pen) essere state pensate da Alessandro come di
per se unicamente negative. Questa e infatti l'inevitabile illazione
che dobbiamo trarre dal cap. 6. Ma come pretendere allora senza
contraddirsi che sia anche vero (come sostiene il cap. 27, p. 198,29)
che 'i caratteri degli uomini diventano di questa o quella qualita in
forza delle differenti abitudini'? Bisogna semplicemente ammettere che Alessandro abbia sottinteso una premessa di questo tipo: se
non si contraggono buone abitudini, e inevitabile contrarre invece
quelle contrarie e cattive. Se dunque c'e mancanza totale di buona
educazione, o se l'educazione impartita risulti male impostata e
carente, Ia maggioranza dell'umanita non assumed affatto quelle
buone abitudini che potrebbero correggere le potenzialita negative
della costituzione naturale: contrarra invece abitudini cattive; rna a
loro volta queste non faranno altro che confermare e irrigidire
quanto era gia iscritto nel patrimonio naturale della costituzione
psichica degli individui. In tal modo non risultera falso ne dire
(come accade nel cap. 6) che il carattere e negli uomini fissato
dalla costituzione psichica assegnata dalla natura: questo e infatti
vero per Ia maggioranza dell'umanita, per Ia quale il risultato finale dei processi di assuefazione e purtroppo perfettamente conforme
ai presupposti naturali; ne dire (come invece accade nel cap. 27)
che i caratteri umani divengono tali o talaltri in conseguenza delle
differenti abitudini contratte: quest'affermazione e infatti vera
incondizionatamente per quanto riguarda le (poche) persone
virtuose e i phronimoi, che divengono tali soltanto grazie aile buone
abitudini assunte dall'esercizio e dall'educazione; e rimane vera in
modo qualificato riguardo alia moltitudine dei non-virtuosi, nei
quali le cattive abitudini contratte per difetto di educazione portano
aile conseguenze estreme il patrimonio naturale del carattere.
Giunti a questo punto del nostro ragionamento e facile tuttavia
indovinare un'obiezione. Ammesso che si elimini in tal modo Ia
contraddizione che si era creduto di scoprire fra le due parti del
trattato di Alessandro e che tanto nel cap. 6 quanto in quelli successivi (27-29) il commentatore si attenga costantemente all'idea che il
carattere e gia naturalmente determinato nella maggioranza degli
uomini che rimane inferiore al livello della virtu, mentre e prodotto dalle buone abitudini in coloro che riescono a diventare persone

292

PIERLUIGI DONINI

per bene-ammesso dunque tutto questa, che ha a fare pen) una

simile concezione con Aristotele? Al quale sembra ovviamente
impossibile attribuire una siffatta bipartizione e distinzione dei modi in cui gli uomini svilupperebbero il loro carattere; impossibile,
in particolare, attribuire una simile accentuazione della rilevanza
delle doti naturali e per di pili soltanto in vista della sviluppo di un
carattere moralmente deteriore. Certamente l'obiezione sarebbe
ragionevole, ne io intendo dire ora che la teoria di Alessandro
corrisponde perfettamente a quella di Aristotele. Ma credo di dover
aggiungere ancora, nonostante tutto, un ulteriore argomento a
discarico di Alessandro: che cioe egli poteva in buona fede esser
convinto di aver offerto un'interpretazione e una sistemazione
adeguata e coerente del punto di vista di Aristotele.
A proposito del contributo che le doti naturali potrebbero dare
alla formazione del compiuto carattere degli uomini il pensiero di
Aristotele e infatti abbastanza oscillante2 3 e non chiaramente
definito; per modo che Alessandro aveva ragione di trovarsi in
difficolta davanti alle diverse proposizioni aristoteliche relative al
problema e, secondo la sua radicata tendenza, doveva allora cercare
di proporne un'interpretazione unificata e tale da giustificare il
maggior numero possibile dei testi aristotelici che trattavano della
questione. L'aspetto che forse a noi sembrera il pili arrischiato di
tutti nell'interpretazione di Alessandro-vale a dire il giudizio cosi
radicalmente negativo dato a proposito delle potenzialita naturalmente presenti nell'anima-poteva dunque essere da lui giustificato quan to me no fondandosi sul testo di EN VI (= EE V) 13, 1144b
4-9 Secondo cui ciascun tratto del carattere e gia naturalmente
presente negli uomini, rna tuttavia 'disgiunto dal senno appare
manifestamente dannoso', tanto da poter essere paragonato da
Aristotele alle qualita degli animali, oltre che dei bambini. Posto
che il conseguimento del 'senno'24 richiesto per trasformare gli
2 3 Mi sia consentito a questo proposito rinviare al mio libro Ethos. Aristotele
e il determinismo (Alessandria 1989) cap. 5. Per Ia tendenza di Alessandro a
offrire interpretazioni capaci di armonizzare e far concordare fra loro tutte le
tesi e le proposizioni aristoteliche devo rinviare al mio saggio 'Testi e commenti, manuali e insegnamento: Ia forma sistematica e i metodi della filosofia in eta postellenistica', in Aufstieg und Niedergang der riimischen Welt,
herausg. von W. Haase und H. Temporini, II.36.7 (Berlin-New York 1994)
5041. II De anima offre abbondanza di esempi in proposito: il commento di P.
Accattino-P.L. Donini a questo trattato (che e attualmente in corso di stampa)
ne segnala un gran numero.
24 Nous in VI 13,1144b 9.

DOTI NATURAL!, ABITUDINI E CARATTERE NEL DE FATO

293

'abiti naturali' in vere e proprie virtu dovrebbe essere unicamente il

risultato dei processi educativi correttamente impostati-questa e
appunto Ia lezione generale che si dovrebbe ricavare dall'ENAlessandro aveva qualche buona ragione per indurre che l'assenza
di quei processi non poteva far altro che condurre all'acquisizione
di cattive abitudini; e aveva anche ragione, oltre a cio, di pensare
che queste cattive abitudini non sarebbero state altro che l'inevitabile (naturale e anche, secondo lui, fatale) sviluppo di quanto era
gia naturalmente iscritto nella costituzione originaria dell'anima:
se appunto egli aveva posto come fondamento della sua costruzione
Ia dichiarazione di ENVI 13 a proposito del naturale fondamento
di ogni tratto del carattere individuale, rna anche della negativita
di qualsiasi aspetto di tale carattere che rimanesse dissociato dalla
corretta razionalita. E infine, tenuto conto delle lamentele di Aristotele a proposito della scarsa cura che le citta greche dedicavano alia
formazione dei cittadini25 (una situazione che difficilmente poteva
apparirgli migliorata ai tempi suoi), aveva ragione anche di concluderne che per Ia maggior parte degli individui sarebbe stato
fatale il rimaner vincolati al carattere gia determinato dalla natura;
mentre le eccezioni previste dalla regola di quel fatale 'per lo pili'
gli permettevano di soddisfare d'altra parte l'esigenza aristotelica
tante volte chiaramente espressa, che cioe Ia virtu morale ripetesse
Ia sua origine dalle abitudini e dall' educazione. La concezione che
Alessandro si era fatta del carattere e dei modi della formazione di
esso potrebbe dunque davvero essere rimasta Ia stessa per tutto il
trattato ed essere ricondotta all'esigenza di trovare una sistemazione
coerente26 peri differenti giudizi dati da Aristotele a proposito delle
doti naturali. Nelle pagine in cui doveva enunciare una dottrina
'aristotelica' del fato gli risultava utile limitare Ia propria considerazione alia maggioranza dell'umanita che non andava oltre
il suo carattere naturale; nelle pagine in cui doveva parlare

25 EN X 9, 1180a 24-30.

26 Un esempio significativo: in EN II 1, 1103a 18 sgg. Aristotele nega che

le virtu etiche siano 'per natura' (<puaEt) e non fa alcun cenno all'esistenza di
particolari doti innate del carattere, ne negative, ne positive; osserva perc) che
le virtu nascono in noi 'che siamo naturalmente preparati a riceverle'
(7tE<pUKOO't oe~aaOat) e siamo poi portati a compimento dall'abitudine. Alessandro riecheggia da vicino questo passo a 198,3-6 parlando di un'attitudine e di
una potenza ricettiva delle virtu che Ia natura ci avrebbe concesso, rna (ovviamente) non ne tiene alcun conto nel cap. 6, dove invece poteva valersi degli
accenni agli 'abiti naturali' reperibili nel VI libro dell'EN.

294

PIERLUIGI DONINI

specificamente del carattere virtuoso poteva invece prendere in

considerazione I' eccezione costituita dalla minoranza delle persone
per bene e rendere cosi giustizia alia teoria aristotelica che derivava
il carattere dalle abitudini.
Non credo che si avrebbe ragione di obiettare alia spiegazione
che ho qui proposto l'eccessiva quantita dei passaggi che essa
integra o supplisce nel ragionamento di Alessandro (cioe, che
Alessandro non dice mai esplicitamente che l'assenza delle buone
abitudini comporterebbe automaticamente che se ne contraggano
di cattive; o che non segnala mai esplicitamente di aver inteso
rendere ragione di questa o di quel passo aristotelico concernente le
doti naturali ) . Si deve tenere con to di due fatti-e di entrambi
insieme. In primo luogo, il De Jato non e un commentario ad alcun
testa aristotelico, rna uno scritto in cui Alessandro riassume ed
espone le proprie vedute relativamente a un'importante questione
largamente dibattuta nella filosofia dal IV secolo in poi; che alcuni
passaggi del ragionamento e gran parte dei riferimenti27 al testa
aristotelico restino impliciti e percio semplicemente quanto ci si
deve attendere da un'opera che non puo raggiungere il livello di
tecnicita che e invece tipico dei commentari.28 In Secondo luogo, se
e vero che il De Jato non e un commentario ad alcuna opera di
Aristotele, non meno vero e tuttavia che Ia sua composizione
presuppone certamente il lavoro di esegesi tipicamente svolto da
Alessandro nei suoi commentari: qualsiasi pagina dell'opera e il
risultato evidente29 di una puntuale riflessione su qualche testo del
corpus aristotelico. E se anche non fosse sicuro che Alessandro
compose un commentario all'EN (come invece giustamente insiste
a ricordarci Mansfeld), 30 una lettura nemmeno troppo approfondita
27 Salvo casi episodici come quelli di 166,23 o di 167,20.

2B Si confronti con quello del De Jato il caso del trattato De anima che puo
essere messo in relazione al precedente commentario allo scritto di Aristotele dallo stesso titolo (in proposito, si veda l'introduzione di AccattinoDonini al De anima nell'edizione ricordata sopra - n. 23 ) . Anche senza
arrivare a parlare di carattere 'semipopolare' del De Jato (per il quale cfr.
Thillet, op. cit., - in n. 7 - CXXV) e pur ammettendo con Sharples, op. cit.,
(n. 14) 18 che il libro contiene anche parti di elevata tecnicita, rimane il
fatto che in esso Alessandro non si impegna direttamente mai nell'esegesi
di testi aristotelici che siano richiamati esplicitamente e analizzati parola per
parola, come invece accade nei commentari.
29 II commento di Sharples basta a suggerirlo. Sui metodo di lavoro di
Alessandro mi sia consentito di rinviare ancora una volta al commento di
Accattino-Donini al De anima di prossima pubblicazione.
30 Art.cit. (n. 1) 182 n. 4, cfr. anche R.W. Sharples, 'Alexander of Aphro-

DOTI NATURAL!, ABITUDINI E CARATTERE NEL DE FATO

295

del suo trattato sul destino dovrebbe bastare a suggerirci che egli
doveva aver lungamente e seriamente meditato sulle pagine di
almeno quella fra le tre etiche del Corpus (certamente, inoltre, su
un'EN che includeva gia i libri comuni).3I
Ma e infine doveroso che aggiungiamo a questa difesa della
coerenza di Alessandro anche l'ammissione che l'interpretazione
ora proposta non in grado di eliminare tutte le difficolta. Sull'origine del vizio i capitoli 27-29 del De Jato hanno, come si e detto,
qualche raro accenno che non
sempre possibile far concordare
con la tesi generale che ho creduto di attribuire ad Alessandro.
Una spiegazione e possibile trovare per l'implicito accenno di
p. 198,29;32 rna l'esordio del capitolo 27 rimane sotto molti aspetti
problematico:

coloro che sono in possesso delle virtu furono a se stessi causa

dell'acquisizione della virtu in quanto, invece33 di trascurarlo,
scelsero quel che e meglio; e press'a poco cosi (7tapa7tA110i~) coloro
che si trovano nei vizi (197, 6-8).
L'idea di un'originaria opzione (EAOjlEVot!) a favore della virtu e
forse difficilmente riconducibile ad Aristotele e alla sua etica; 34
tuttavia essa appare alquanto piu attendibile in Alessandro, special-

disias: Scholasticism and Innovation', in Aufstieg und Niedergang der romischen

Welt 11.36.2, herausg.von W. Haase und H. Temporini (Berlin-New York
1987) 1186. Che Alessandro avesse una conoscenza anche di EE e MM non era
dimostrabile secondo F. Dirlmeier, in Aristoteles, Magna Moralia (Berlin
1958) 103-104, rna si veda ora il commento di Accattino-Donini a proposito di
De an. p. 89,1-7.
3! La conoscenza, quanto meno, del II e del III Iibro dell' EN risulta chiara
dal cap. 27 (si veda anche qui sopra, n. 26) e quella delle definizioni e
dell'analisi delle virtu dianoetiche nel libro VI risulta indubitabile p.es. da
passi come De an. 66,1&-20. 80, 24-81,9.
32 Sopra, p. 291.
33 Tutte le edizioni posteriori a Bruns accettano qui Ia congettura avd.
34 Quanto di piu vicino si puo trovare a quest'idea e probabilmente Ia
(problematica e limitativa) dichiarazione di EN III 5, 1114b 22-24 che 'noi
uomini siamo in qualche modo corresponsabili' dei nostri abiti morali-un
passo moho discusso, su cui si vedano da ultimo D. Frede, 'Necessity, Chance,
and 'What happens for the most part' in Aristotle's Poetics', in Essays on
Aristotle's Poetics, ed. A. Oksenberg Rorty (Princeton 1992) 203, inoltre il mio
Iibro citato sopra (n. 23), 102-103. Molto piu plausibile sarebbe invece stato per
Alessandro sostenere Ia tesi piu modesta della semplice 'volontariecl' degli
abiti etici: cfr. R. Sorabji, Necessity, Cause and Blame. Perspectives on Aristotle's
Theory (London 1980) 26&-69; inoltre il mio saggio 'Volontariecl di vizio e
virtu (Aristot. Eth. Nic. III 1-7) ', in Etica, politica, retorica. Studi su Aristotele e la
sua presenza nell'etti moderna, a cura di E. Berti e L.M. Napolitano Valditara
(L'Aquila 1989) 3-21.

296

PIERLUIGI DONINI

mente se si tiene conto del modo in cui e illustrato l'esempio di

Socrate nel cap. 6, p. 171,11-16: in quel caso si puo probabilmente
davvero parlare di un'autonoma scelta35 dell'individuo di migliorarsi e di darsi un'educazione filosofica, tanto pili se si trattava per
Socrate prima di tutto di rifiutare le limitazioni che le sue naturali
tendenze gli avrebbero imposte. E non mancano in Alessandro
alcuni altri accenni che fanno pensare che per lui una scelta
fondamentale di vita, una conversione alia virtu o alia filosofia
come frutto di decisione personale fosse diventata una nozione
ovvia e una prospettiva reale: 36 egli avrebbe cosi felicemente risolto
un punto che Aristotele (in un ben diverso quadro storico) aveva
lasciato quanto meno poco chiaro. Ma cio che nel passo di 197,6-8
risulta perfettamente oscuro o, a essere sinceri, completamente
inattendibile e il trattamento riservato all'origine del vizio: che cosa
puo infatti significare la proposizione che 'press'a poco cosi' (cioe
press'a poco come nel caso dell'origine delle virtu: non si puo
intendere altrimenti) andrebbero le cose anche per le persone che
si trovano a essere viziose? Per mantenere un'analogia con il caso
della virtu si deve pensare che in qualche momenta della loro
esistenza vi sia stata da parte di esse (che numericamente devono
per di pili rappresentare la grande maggioranza dell'umanita!) una
consapevole scelta di attenersi al carattere naturale e una rinuncia
formale ad assumere buone abitudini. Ma sarebbe grottesco
immaginarlo e infatti Alessandro non lo dice: dice semplicemente
che le cose andrebbero 'press'a poco allo stesso modo',3 7 lasciando
35 Cfr. spec. 171,13 npoaipEcrt<; K!X'tcbov piov.

36 Cosi sono indotto a pensare anche da 170,24 con l'accenno a 'qualcosa di

pili bello' che si realizzerebbe nel vizioso. La costituzione del testo in questo
passo e estremamente difficile (si veda l'apparato di Sharples), rna forse
qualcosa di meglio di quanto finora si e tentato e ancora possibile suggerirlo.
Veramente impossibile mi pare Ia soluzione di Thillet, nel cui testo bisognerebbe comunque correggere l'accento del verbo a p.10,24 scrivendo eKnecrn .
Sennonche non sembra che EK7tt7t'tCO possa mai avere un significato transitivo e
causativo-suggeritomi dal fatto che Thillet traduce 'l'ait fait sortir de son
vice'-e bisognerebbe dunque dare al verbo il solito suo significato intransitivo: qualcosa di pili bello che realizzandosi nel vizioso 'cade fuori', fuoriesce
dalla malvagita, un' espressione tortuosa e innaturale, dato che il senso che ci
attendiamo e invece precisamente quello assunto dalla traduzione di Thillet.
Questo senso si puo ottenere, certo, con Ia congettura di Diels (accolta infatti
da Sharples) hcrdcrn, che e pero alquanto lantana dalla lezione tradita
EK7tlCJ11't11CJ. Come mi suggerisce oralmente Gianfranco Gianotti, si potrebbe
allora pensare a EK7ttecrn: 'comprima fuori, costringa a uscire' (cfr. Polibio
18.32.3).
37 Oapanl..11criro<; mi pare meglio tradotto da Thillet (a peu pres) che da

DOTI NATURAL!, ABITUDINI E CARATTERE NEL DE FATO

297

dunque intuire che esiste una qualche differenza fra il caso

dell'origine della virtue quello del vizio, rna guardandosi bene dal
precisarla. In realta, nei capitoli 27-29, in cui si ricollegava pili
chiaramente alla tesi aristotelica dell'origine degli abiti morali
(tutti, sia quelli virtuosi, sia quelli viziosi) dai processi di assuefazione, Alessandro era in certo senso costretto a sottolineare l'esistenza
di una fondamentale identita fra le modalita di formazione della
virtu e quelle del vizio; rna doveva anche nello stesso tempo non
smentire in maniera troppo evidente quanto aveva gia esplicitato
nel cap. 6 a proposito dell'origine dei caratteri viziosi in contrapposizione a quello del saggio Socrate: di qui, da questo disagio, viene
probabilmente la formulazione generica e reticente della linea

197,8.
Notevole sembra un altro passaggio del ragionamento di
Alessandro nel cap. 27, in cui c'e un'allusione inequivocabile, rna
anche alquanto problematica, alia teoria del cap. 6. Nel corso di un
argomento inteso a mostrare che le virtu non nascono in nessun
senso per natura, Alessandro osserva (a 197,30-198,3) che
niente di impossibile bisogna richiedere alia natura (essa e infatti Ia
misura del possibile e dell'impossibile; infatti Ia virtu e perfezione e
compimento sommo della natura propria di ciascuno ed e
impossibile che qualcosa che e imperfetto sia nella perfezione, rna
quel che e stato genera to e imperfetto non appena generato).
11 concetto di oh:Eia <pUcrt<;, natura propria, e precisamente quello
che appare centrale nella teoria del fato esposta nel cap. 6;
sennonche, Ia Ia natura propria e unicamente collegata al carattere
deteriore e al destino delle persone non virtuose, tanto da indurci a
pensare che quel che si ha alia nascita come costituzione naturale
dell'anima sia destinato a tradursi in un carattere vizioso ogni volta
che non intervenga l'educazione (o l'autoeducazione) mediante le
buone abitudini. Nei casi eccezionali in cui questo intervento ci sia,
l'opera dell'educazione consisterebbe allora essenzialmente nel
cancellare o radicalmente emendare il carattere naturale (cosi
appunto ci dice anche l'esempio di Socrate) e nessun positivo contributo questo carattere dovrebbe poter dare alla costituzione della
personalita virtuosa. Infatti, coerentemente con queste assunzioni,
implicite o esplicite che siano, lo stesso cap. 27 insiste a presentare
il contributo che Ia natura darebbe alla virtu soltanto nei termini di
Sharples (similarly). Non

e infatti ojlotroc;, ne roaautroc;.

298

PIERLUIGI DONINI

una 'capacita ricettiva'3B della virtu stessa: la .mera attitudine a far

sviluppare in se la virtu grazie aile buone abitudini e all'insegnamento. Ma, perche si possa parlare della stessa virtu come del
compimento e della perfezione (teleiotes) della natura propria di
ciascuno, occorre invece che quella natura propria sia stata pensata
da Alessandro almeno in questo passo del cap. 27 in modo non
esclusivamente negativo, rna quanto meno neutro: intendendo cioe
le doti della costituzione naturale come qualcosa di ambiguo e realmente suscettibile di essere completato anche nel senso migliore.3 9
Di qui si intuisce quanto risultasse difficile per Alessandro la
normalizzazione e la sistemazione delle tesi aristoteliche; e altri
scritti inclusi nelle raccolte delle sue opere minori confermano
che egli stesso, o qualche suo scolaro, deve essersi affaticato piu
volte e senza troppo successo sull'arduo problema delle doti naturali
e della loro valutazione in sede di etica.4o
Non e il caso di infierire ancora su Alessandro insistendo sul
fatto che nemmeno una volta che si siano spiegate cosi, come
appunto egli cerco di fare nel De Jato, le modalita di formazione dei
caratteri virtuosi e viziosi la sua difesa dell'autonomia e del potere
di autodeterminazione da parte dell'uomo diviene realmente piu
forte; infatti, anche concessagli da parte nostra la possibilita dell' opzione originaria in favore del bene e della virtu, rimane comunque
il guaio che il carattere del cpp6vtf.10~ ormai costituitosi come tale
risultera non me no vincolante di quello (considerato espressione
del 'fato', cioe della natura) della moltitudine dei non-saggi. Non

e detto
n. 26), di EN II 1.
Nel senso, dunque, in cui parla delle doti naturali Ia Politica VII 13,
1332b 1-3. Sono pen) evidenti le ragioni per cui Alessandro non poteva
adottare in modo esplicito questa soluzione: a parte il fatto che Ia Politica parla
della presenza innata solo di 'alcune' doti e non di tutte, se Alessandro avesse
ammesso che tali doti erano fondamentalmente ambigue e potevano essere
risolte in meglio o in peggio soltanto dai processi di assuefazione, avrebbe
anche dovuto rinunciare a identificare il fato con una natura a cui non poteva
piu attribuire Ia forza di determinare il corso successivo degli eventi quanto
meno nella maggior parte dei casi. L'intera sua teoria dell'identita tra fato e
natura avrebbe cosi dovuto limitarsi alia sola considerazione del livello della
generazione nelle specie biologiche senza poter discendere aile qualita degli
individui umani (o, tutt'al piu, dei tipi). Per qualche ragione (cfr.anche qui
sotto, nota 41) egli credette di non potersi limitare a questo livello, dove
certamente Ia sua teoria sarebbe potuta risultare piu aristotelica e piu coerente.
40 Cfr. il mio libro del 1974 cit. sopra (n. 3) 165-170.
38

o\Jvcx~ti~ 'tE l((Xl E7tl't1J0ElO't1J~ OEK'tlri] 198,5 - chiaro ricordo, come si

(so~ra,

DOTI NATURAL!, ABITUDINI E CARATTERE NEL DE FATO

299

c'e

nulla da aggiungere, su questo punto, a quanto era gia stato

notato neUe precedenti ricerche sul problema del determinismo in
Aristotele e pili specificamente nel De Jato .41

Universita di Milano

41 Cfr. Ia mia messa a punto del problema nel saggio citato sopra (n. 8)
soprattutto 1249-1258; e da ultimo le chiare osservazioni di D. Frede, art. cit.
(n. 34) 203 a proposito del carattere deterministico che si puo riconoscere
all'etica di Aristotele-Vorrei ancora aggiungere una precisazione. Nei miei
precedenti interventi, soprattutto nel libro citato sopra (n. 2), avevo ricondotto
all'influenza di Galeno Ia teoria esposta net De Jato che assume un'identiti fra
il carattere naturale e il destino. Non vedo contraddizione fra quell'ipotesi e
il suggerimento che do oggi e che mira a ricondurre all'interpretazione
sistematica dei passi aristotelici sulle doti naturali l'origine della concezione
alessandrista. E' infatti perfettamente concepibile che Alessandro, influenzato
da Galeno (di cui comunque conosceva qualcosa), o da una precedente tradizione peripatetica (per cui cfr. Mant. 186,28-31), si fosse convinto dell'importanza del carattere naturale legato alia costituzione dell'anima e che credesse
in buona fede di poter ricostruire Ia medesima concezione anche muovendo
dalle opere di Aristotele, di cui si sentiva comunque tenuto a offrire una
spiegazione che appianasse ogni incoerenza e oscurita. Del resto Galeno
stesso, nello scritto quod animi mores etc. (p. 51,12 sgg. Muller), si appella ad
Aristotele e cerca di includere Ia teoria aristotelica fra i garanti della sua tesi
deterministica. Un caso perfettamente analogo a quello del De Jato e fornito
dalla teoria alessandrista dell'origine dell'anima dalla mescolanza degli
elementi corporei: Alessandro Ia trovava gia presente nella tradizione
peripatetica e in Galeno, rna, d'altra parte, credeva anche di poterla ricavare
da un preciso testo di Aristotele. Riferimenti ai testi e alia letteratura in
Sharples, art. cit. (n. 30) 1203.

ALCINOUS' EPISTEMOLOGY
DAVID SEDLEY

When it came to epistemology, doctrinal Platonists in antiquity

faced a daunting task. They had somehow to integrate into a
single theory at least three very disparate-looking treatments of the
relation between knowledge and opinion:
(1) the Meno's apparently successful account of knowledge as a
species of correct opinion ( doxa);
(2) the Theaetetus' self-acknowledged failure to find out what
species of correct opinion knowledge could possibly be;
(3) the seemingly unbridgeable separation of knowledge from
opinion in both Republic V and the Timaeus.
In some recent work, 1 I have tried to explore how this task of
integration was tackled by three rival groups of Platonists-the
sceptical Academics of the Hellenistic age, who aimed to show
that Plato rejected the possibility of knowledge; the anonymous
commentator on the Theaetetus, who makes the Meno the authoritative Platonic text on knowledge; and the group which he most
strongly opposes, best represented for us today by chapter 4 of
Alcinous' Didaskalikos, an epitome of Plato's thought composed in
the first or second century AD.2 In the present article, in order to
1 G. Bastianini, D. Sedley, edition of Anonymous Commentary on Plato's
Theaetetus, in Corpus dei papiri filosofici greci e latini, vol. III (Florence 1995),
227-562; D. Sedley, 'Three Platonist interpretations of the Theaetetus', in Form
and Argument in Later Plato, ed. C. Gill, M.M. McCabe (forthcoming, Oxford
1996).
2 Citations of the Didaskalikos are from the Bude edition by J. Whittaker
and P. Louis, Alcinoos, Enseignement des doctrines de Platon (Paris 1990), to whose
notes and apparatus I am heavily indebted. Equal thanks are due to the
splendid English translation and commentary by John Dillon, Alcinous, The
Handbook of Platonism (Oxford 1993). Both of these should be consulted for full
details of the Platonic allusions in Alcinous' text, and for many other aspects
not covered in the present article. Other helpful recent work on this chapter
includes R.W. Sharples, 'The criterion of truth in Philo Judaeus, Alcinous
and Alexander of Aphrodisias', in The Criterion of Truth, ed. P.M. Huby and G.
Neal (Liverpool 1989) 231-56; and L.P. Schrenk, 'Faculties of judgment in the
Didaskalikos', Mnemosyne 44 (1991) 347-63. On the date of Did. I suspend judgement for now. It may seem unwise not to acquiesce in the schooled intuitions

ALCINOUS' EPISTEMOLOGY

301

develop a fuller account of how this last party handled the task, I
shall be analysing the Platonic exegesis propounded in Didaskalikos chapter 4.
Their outline position seems to have been as follows. Platonic
episteme, as explained in the Republic and Timaeus, has Ideas as its
objects, while the sensible world is the object of mere doxa. These
two faculties of episteme and doxa were jointly taken to map onto
what in the Hellenistic age had come to be known as the topic 'on
the criterion'-that is, the principles and means of cognition in
general. The Theaetetus was held to deal with only one half of this
topic, albeit the half which had become the central focus of
Hellenistic debates on the criterion: that is, the part which deals
with sensory cognition.3
The Theaetetus fails to find a definition of knowledge, but the
failure is interpreted as a calculated one. The search fails precisely
because what it in fact addresses itself to is the epistemology of the
sensible world, whereas knowledge has the Ideas as its proper
objects. Consequently, while it fails as an account of knowledge,
the Theaetetus succeeds as an investigation of the sensible world's
epistemological structure and of the cognitive faculties which bear
on it. It is to be plundered, not as a radical (and aporetic) reinvestigation of episteme, but as Plato's most fully elaborated account
of sensory cognition.
As for real episteme of the Ideas, it is held by this group of
Platonists that that topic is not tackled in the Theaetetus but in the
immediately succeeding dialogue, the Sophist. Their critic, the
anonymous commentator on the Theaetetus, sums up their view at
the beginning and end of the following passage (2.11-39):
Some of the Platonists have thought that the dialogue was on the
topic of the criterion, in view of the considerable space it also
of Dillon (op. cit.) xii-xiii and P.-L. Donini, 'Testi e commenti, manuali e
insegnamento: Ia forma sistematica e i metodi della filosofia in eta
postellenistica', ANRW II 36.7 (1994) 5027-5100, pp. 5057-8, who plead for a
date in the second century AD on the basis of comparisons with known
Platonists from Plutarch to Numenius. Nevertheless, it seems to me that we
still know far too little about first-century AD Platonism to exclude even so
early a dating. One thing, at least, has ceased to be controversial, and that is
the restoration of the author's name to the transmitted 'Alcinous', in place of
Freudenthal's century-old conjecture 'Albinus'.
3 I leave aside here the question how they managed to accommodate, or
discount, the Meno's definition of knowledge as a species of correct doxa. On
this, see my art. cit. (n. 1).

302

DAVID SEDLEY

devotes to the investigation of this. That is wrong. Rather, the

declared aim is to speak about simple uncompounded knowledge,
and it is for this purpose that he necessarily investigates the
criterion. By 'criterion' in the present context I mean the criterion
through which we judge, as an instrument; for it is necessary to
have that whereby we will judge things; then, whenever this is
accurate, the permanent acceptance of the things which we have
judged properly becomes knowledge.
These people, on the other hand, say that, having made it his
declared aim to investigate knowledge, in the Theaetetus he shows
what its objects are not, while in the Sophist he shows what its
objects are.
On this view, then, the Theaetetus is about the 'criterion'. It tells us
much about sensory cognition, but all that it reveals about real
knowledge is something negative, what its objects are not, namely
sensibles.
Such a view seems to fit closely the epistemological reading
of Plato advocated in Didaskalikos chapter 4, which is itself, significantly, introduced as being 'on the criterion'. 4 The chapter's
opening distinction between the criteria 'through which' and 'by
which' judgement is exercised is inspired primarily by Theaetetus
184-185, where the senses are relegated to the former, purely
instrumental, role, while the judging mind is dignified with the
genuine cognitive agency which the latter role confers on it.5
The chapter's account of episteme reflects, not the Theaetetus, but
the Republic, Timaeus and Phaedrus. Knowledge is of Ideas, not
sensibles. What is especially distinctive of Alcinous' approach is
the elaborate parallelism which he constructs between the intelligible and sensible worlds and, correspondingly, between intellectual and sensory cognition. The inspiration for this parallelism
undoubtedly lies in the Republic and Timaeus.
The nature of the parallelism can be seen in the chart which
follows.

Alcinous, Did. 154.8-9.

However, as Dillon notes (op. cit. [n.2] 61), the most direct source of the
actual expression at 154.15-16, 'tO lli: llt' ou opyavov Kpt'ttKOV ... , is Republic IX,
582d, llt' ou ye llet 6pyavou KpiveaSat (already called a Kpl'tliptov back at 582a).
4

LOGOS

divine

human

154.22-5, cf. 164.1!1-18

154.2!1-5

3 criteria (hence infall-

DOXASTIC LOGOS

ible, 154.21-!l)
tJrrou.gh which we judge,
154.10-21; 'natural',
155.!14-6
4 components of logos?

EPISTEMONIC LOGOS

154.25-!12

154.25-!12

doxai

epistimai haplai

154.40-155.1!1

155.!12-6

(with picturing, and

internal/ external speech,
155.1!1-20)

5 dispositional remnants

of cognitive activities
in 6.
6

.001.1Cft evvota

(etc.)/'memory'
(incarnate, dispositional)

memory
~ {

154.!16-155.1!1

155.20-!14

aisthesis

cognitive activities

(=)

noesis

= archai

(discarnate, active)

154.!12-4

155.20-!12

species of cognitive
activities

primary secondary (of athroismata)

primary

secondary

how the cognitive

activities function

aided by doxastic logos

156.8-9

aided by epistemonic logos

156.5-7

aiding epistemonic logos

156.7-8

objects of cognition

10 the species of 9

aiding doxastic
logos 156.9-10

aistheta

secondary

e.g. whiteness
155.42-156.1

e.g. white things e.g. honey

156.1-2,5
156.!1
(also called
'incidental')

11 the worlds constituted of 10

primary

sensible world
156.1!1-14

athroismata

primary

secondary

Ideas
immanent forms
155.!19, 164.4-5 155.40-1

intelligible world
156.11-1!1

304

DAVID SEDLEY

In lines 1-8 the operations of perception are shown on the left,

those of intellection on the right. The vertically descending lines
are used to divide genera into species, while the ascending arrows
show the causal chains by which reason (logos) is derived from
the primary cognitive activities. Lines 9-11 convey the structure of
the respective worlds on which perception and intellection operate.
Here the vertically descending lines again divide genera into
species. The arrows joining line 10 to line 11 convey the composition of the two worlds.
Following Timaeus 29b, there are two kinds of human logos (line
3), corresponding respectively to the sensible world and the
intelligible world (line 11):
This latter [i.e. human, as opposed to divine, A.Oyo~] is also of two
kinds. One kind concerns intelligibles, the other perceptibles. The
one which concerns intelligibles is 'knowledge' [ E7tt<rt1'11.1TJ] and
'epistemonic reason' [ E7ttO''tTJI.lOVtKO~ A.Oyo~], while the one which
concerns perceptibles is 'doxastic [ oo~acrttKO~] reason' and 'opinion'
[oo~a]. Hence epistemonic reason has firmness and duration,
because it concerns principles which are firm and enduring, while
the persuasive and doxastic has a good deal of 'likelihood' [eiKo~],
because it does not deal with enduring things. (154.25-32)
Doxastic logos is derived from aisthisis, epistemonic logos from
noesis (lines 6-3 of the chart). Our next task is to reconstruct these
parallel processes, by which the two kinds of logos are formed. It
will be convenient to reverse Alcinous' order and begin with the
formation of 'epistemonic' logos:
Intellection [ VOTJO't~] is the activity of the intellect when it contemplates the primary intelligibles. This seems to be of two kinds. One
kind is before the soul came to be in this body, when it was contemplating the intelligibles, while the other is after its introduction
into this body. Of these, the intellection that took place before the
soul came to be in a body was called just that, 'intellection', but once
the soul had come to be in a body, what had then been called
intellection was now called a 'natural conception', being a sort of
intellection stored away in the soul. So when we say that intellection is the starting point of epistemonic reason, we do not mean
what we now call intellection, but the one when the soul was separate from the body, which, as we said, was then called intellection
but is now called a natural conception. The natural conception is
also called by Plato 'simple knowledge' and 'the soul's wingedness', and sometimes 'memory'. It is out of these simple bits of
knowledge that natural and epistemonic reason is constituted,
being present in us by nature. (155.20-36)

ALCINOUS' EPISTEMOLOGY

305

Start on the right-hand side of the chart in line 6 with the cognitive
activity of v611crt<;. This is the discarnate soul's direct apprehension
of the Ideas, as described in the Phaedrus. When the soul is
incarnated, this same apprehension becomes (line 5) a set of
buried memories or innate conceptions. Severally, they are (line
4) simple pieces of knowledge ( bttcr'tfh.t.at a1tA.a'i), but jointly (line
3) they constitute epistemonic reason. This last identification
draws on the common Hellenistic notion of human reason as
constituted by a comprehensive set of ennoiai or conceptions. Thus
our epistemonic or scientific reason is, roughly speaking, an
innate capacity to handle a priori concepts, thanks to our souls' prenatal acquaintance with the Ideas.
An apparent anomaly must be tackled. In the lines just quoted,
Alcinous seems at pains to emphasize that 'intellection' (noesis) is a
term reserved for pre-natal direct acquaintance with the Ideas. Yet
he starts this very same passage by allowing, on the contrary, that
there are two species of noesis, of which one is prenatal contemplation of Ideas while the other in the incarnate soul's latent
memory of those same Ideas. Moreover, later in the chapter he is
ready to use noesis in yet another sense, namely the incarnate
soul's active contemplation of Ideas (156.5-8)-a usage, of course,
with excellent Platonic credentials. It is only within the confines
of the present context, the transition from pre-natal contemplation
to post-natal epistemonic logos, that the restricted usage is imposed.
Why so? The answer will, I think, become clear from the parallel
process by which doxastic logos is formed. Logos serves as a
criterion of truth because it is a stock of fundamental conceptsempirical concepts in the case of doxastic reason, a priori concepts
in the case of epistemonic reason-which furnish a standard
against which reasoning may be judged. In both cases this relies
on the presumption that the criterion in question is not itself the
outcome of any process of reasoning. Only thus are the dangers of
circularity or infinite regress averted-just the same requirement
that had underlain Hellenistic notions of the criterion from
Epicurus (Letter to Herodotus 37-8) onwards. 6 It is for this reason that
6 On the Hellenistic notion of a criterion, see esp. Gisela Striker,
Kpttt\ptov tfjc; aA.f19Eiac;, Nachrichten der Akademie der Wissenschaften in
Gottingen, Phil.-hist. Kl., (Gottingen 1974) 2, 47-110; and 'The problem of
the criterion', in Epistemology, ed. S. Everson, Companions to Ancient
Thought (Cambridge 1990) 143-60.

306

DAVID SEDLEY

the process from pre-natal direct acquaintance with Ideas (line 6) to

the incarnate soul's framework of a priori concepts (line 3) must be
mediated by nothing more than the virtually automatic downloading of the pre-natal knowledge into an innate disposition to
understand. Noesis in the sense of the incarnate soul's intellectual
activity must be allowed to play no part here (even though, when
we move from line 6 to line 7 it will be precisely this kind of noesis
that comes under scrutiny).
The whole picture is shaped by the need for the transition from
the original cognitive act, noesis, to the epistemonic criterion of
truth to run in the closest possible parallel to the transition, on the
left side of the chart, from the original cognitive act, aisthisis, to the
doxastic criterion of truth. It is to this that we must now turn. We
start back in line 6 with the cognitive act of aisthisis (which, unlike
noesis, is essentially incarnate):
Perception is an affection of the soul through the body, whose
primary function is to report from the faculty affected. When there
is engendered in the soul a perceptual imprint, i.e. a perception,
through the sense organs, and then this does not fade through
length of time, but endures and is preserved, its preservation is
called 'memory'. (154.34-40)
Aisthesis, after its occurrence, leaves a memory trace (line 5; the
parallelism to innate knowledge of Ideas, on the right-hand side, is
emphasized by noting that even the same term, 'memory' can be
used for both), 7 and a set of such memory traces leads eventually to
doxastic logos (line 3). Where epistemonic logos was about intelligibles, doxastic logos is said to be about perceptibles. It seems clear
that, whereas your epistemonic logos is your ability to handle a
stock of a priori concepts, your doxastic logos is your capacity to
operate with a set of empirical concepts, derived from repeated
sense-experiences. Alcinous' examples include 'horse' and 'fire'.
There is no reason in principle why there should not be both a
doxastic and an epistemonic conception of the very same thingboth an empirical grasp of what fire is, based on previous experiences though touch and sight, and (as in Timaeus 51b) an intellectual grasp of fire's essence, its Idea. The question is not directly
addressed, but since Alcinous elsewhere has no trouble in admitting Ideas even of sensible qualities like sweet and hot (180.4-5)

7 For 'memory' used of innate ideas, cf. Phdr. 249c5, 250a5.

ALCINOUS' EPISTEMOLOGY

307

there seems every reason to suppose that he would allow both

kinds of conception to co-exist in parallel.B
What is the path from aisthisis to doxastic logos? The text is
inexplicit, and in constructing my diagram I was long hesitant as
to whether memory (line 5) leads to doxastic logos (line 3) directly,
or in partnership with doxa (line 4). The reason for hesitation is as
follows. Doxa is described as the conscious and fallible process by
which we compare a sense-impression to pre-existing memory
traces and identifY an item, either correctly or incorrectly, saying
to ourselves e.g. 'That is a horse':
Opinion is the combination of memory and perception. For when
we first encounter some perceptible, and from it we get a perception, and from that a memory, and later we encounter the same
perceptible again, we connect the pre-existing memory with the
subsequent perception, and say within ourselves 'Socrates!',
'Horse!', 'Fire!', etc. And this is called opinion-our connecting the
pre-existing perception 9 with the newly produced perception. And
when these, on comparison, are in agreement, an opinion turns out
true, but when they differ, false. For if someone who has a memory of Socrates encounters Plato and is led by some resemblance 10
think he is encountering Socrates again, and then, taking the
perception that comes from Plato as coming from Socrates, connects
it with the memory he has of Socrates, the opinion turns out false.
The thing in which the memory and the perception occur Plato
likens to a wax block. (154.40-155.13)
Most of this is, of course, squarely based on the Wax Block model
in Part II of the Theaetetus. But since it is emphasised (in accordance with the spirit of the Theaetetus passage) that doxa can be
wrong as well as right, it is not easy to see how it can in turn
become an adequate basis for doxastic logos. For doxastic logos is a
'criterion' of truth. It is one species of human logos, and human
logos as a genus is explicitly said by Alcinous to be 'infallible so far
as the recognition of things is concerned' (154.23). This perhaps
surprising description is in fact required by the very notion of a
criterion, which must in its nature be something by which the
true is unfailingly distinguished from the false. But how can mere
doxastic logos achieve such reliability? Alcinous, echoing the
8 For such a reading of Plato, cf. Plutarch fr. 215d Sandbach, and Dominic
Scott, Recollection and Experience (Cambridge 1995) 18-23 and ch. 2.
9 I here translate the tranmitted text, tT,v 7tpOtl7tOKEtJ.U~VTJV [sc. aicr9TJcrtv]
(155.4-5). Whittaker and Dillon both adopt Heinsius' supplement <J.!VT\J.!TJV>,
maybe rightly, but since the !lVT\J.!Tl is itself a stored perception the transmitted
reading, if less satisfactory, is not impossible.

308

DAVID SEDLEY

Timaeus, admits that doxastic logos has an inherent instability

because of the unstable nature of its objects (154.31-2). But he still
insists, if I understand him correctly, that it does not make
mistakes in recognizing things (Ka-ta rilv trov 1tpayJ.uitrov yvrom v
ciBuhveucrto~) .10 I assume that he means the following. Your empirical conception of, say, horses, being built on a series of senseperceptions, conveys to you facts about horses which are inherently subject to change-their colours, shapes, movements etc., as
distinct from their totally unchanging essence. Thus in many
ways it is unreliable, and can offer nothing better than the kind of
'likelihood' promised by the Timaeus. But despite this instability, if
you have adequate sensory experience of horses (whether or not
you use it successfully) you have the capacity unfailingly to recognize a horse when you see one. You will lack scientific ('epistemonic') knowledge of what a horse is, but your doxastic grasp of
horses will be entirely adequate for recognitional purposes. This
criteria! reliability, it might be thought, would be more credible if
doxastic logos were the mechanical product of repeated senseexperiences, than if we have to suppose that the passage from senseexperience of horses to a doxastic grasp of 'horse' is mediated by
conscious doxai about horses, some of them mistaken. That would
maintain the parallel with the right-hand side of the diagram,
where, as I have emphasized, no active intellectual judgement is
permitted to mediate the passage from discarnate knowledge to
innate a priori concepts.
On the other hand, it may be that the path to doxastic logos is
after all mediated by doxai, but only by a very simple subset of
them, the minimal recognitional doxai where we do not simply
see horses and fires but also, as in Alcinous' own examples,
identify them as horses and fires.ll Later on we will encounter
evidence that doxa must indeed intervene in this way.
10 I propose this in preference to Dillon's translation 'the other [viz.
human reason] is only free from error when it is engaged in the cognition of
reality' (op. cit. [n. 2] 5, with my emphasis added; on p. 63 he uses the plural
'realities'), presumably meaning that only the epistemonic species of human
reason is infallible. I suspect that the Greek terminology chosen is too loose
to imply such a restriction; there is no sign in Alcinous' usage of this
enriched Neoplatonic sense of 1tpciy~a'ta. For yvroau; in the sense 'recognition', cf. Did. 173.43.
II For Antiochus' similar use of simple recognitional judgements as a
foundation for the criterion of truth, cf. Cicero, Academica II 21, aptly cited
here by Dillon.

ALCINOUS' EPISTEMOLOGY

309

Regardless of these and other details, it seems clear that Alcinous is presenting us with an elaborated cognitive analysis of the
two modes of thought, that of doxa and that of episteme, which the
Republic graphically introduces to us with its parallelisms between
the lower and upper parts of the divided Line and between the
worlds inside and outside the Cave. The next question is where the
epistemology of the Theaetetus fits into this picture.
We have already seen how Part II of the Theaetetus-in particular the Wax Block analogy-is treated by Alcinous as Plato's
canonical account of doxa in the sense specified in the Republic,
fallible empirical cognition. We can now add that Theaetetus Part I
is used as Plato's guide to the epistemological structure of the
sensible world. A distinction is first made (I.e. diagram lines 6-7,
right-hand side) between primary and secondary noesis, distinguished (lines 9-10) by their objects, these being transcendent
Ideas in the case of primary noesis and immanent forms in the
case of secondary noesis:
Therefore, there being both epistemonic and doxastic reason, and
both intellection and perception, there are also the objects of these,
namely intelligibles and perceptibles. And since of intelligibles
some are primary, namely the Ideas, and others secondary, namely enmattered forms which are inseparable from the matter, intellection will also be of two kinds, namely that of primary and that of
secondary intelligibles. (155.36-42)

An analogous set of distinctions is then made for aisthisis (lines 67 and 9-10):
And again, since of perceptibles some are primary, namely
qualities like the colour whiteness, others accidental, like the object
coloured white, and, posterior to these, the aggregate ( &9potcrJ.ta ),
e.g. fire, honey, so too of perception one kind, called 'primary', will
be of primary perceptibles, and the other, called 'secondary', of
secondary perceptibles. (155.42-156.5)
For once Whittaker's excellent apparatus on the Platonic and Aristotelian antecedents is incomplete. He cites Aristotle De anima II 6,
which is certainly the correct antecedent for the terminology of
'accidental' perceptibles, but overlooks the passage's far more
profound dependence on Theaetetus 156d-157c. 12 There Socrates can
12 Whittaker (op. cit. [n. 2] 86 n. 65) does note the occurrence of athroisma
at Tht. 157b9, but does not appear to regard the present passage as alluding to
it. Dillon (op. cit. [n. 2] 70-1) is more attuned to the Theaetetan echoes,
including 157b9. However, he emphasizes 182a-rightly, but at the expense of
the more important 156d-157c.

310

DAVID SEDLEY

be read as making exactly the threefold division of perceptibles

which Alcinous describes. In Socrates' chosen example of perception, that of seeing a white object, the object is said to become
'filled around with whiteness' when seen, but to become itself not
'whiteness' but 'white'. That carefully-worded distinction already
supplies Alcinous' own distinction between the primary perceptible, 'whiteness', and the secondary perceptible, 'the white thing'
('tO AEUKOV) .13
After distinguishing primary and secondary perceptibles,
Alcinous adds a third item, the athroisma or aggregate, exemplified
by honey, which combines flavour, texture, colour etc. in a single
complex item. This too directly echoes the passage at Theaetetus
156d-157c, which ends with the designation of complex objects
like man and stone as athroismata of simple perceptible properties.
(Some have read this sentence in the Theaetetus as the phenomenalist thesis that such objects are nothing more than bundles of
sensible properties. But Alcinous' physics, based on the theory of
elements in the Timaeus, precludes the possibility that he is reducing
physical objects to mere bundles of sense-properties in this way;
his claim is purely about the epistemological status of sensible
substances.)
This tripartite division of perceptibles is a remarkably ambitious
edifice to build on Theaetetus 156-157. What is its purpose? The
answer begins to emerge only in line 11. We need at this stage to
consider the exceptionally off-putting passage which introduces
the materials of lines 7-8 and 11:
Now the primary intelligibles are judged by intellection (VOTI<nc;},
not without epistemonic reason, by means of a kind of comprehension ( nepiATI'Iflc;) and not discourse ( ote~oooc;); the secondary
intelligibles are judged by epistemonic reason, not without intellection. The primary perceptibles and the secondary perceptibles
are judged by perception, not without doxastic reason; the aggregate
is judged by doxastic reason, not without perception. And the intelligible world being a primary intelligible, while the perceptible
world is an aggregate [ &epotOI.Hl], the intelligible world is judged
by intellection 'with the help of reason'-i.e. not without reason-,
while the perceptible world is judged by doxastic reason not
without perception. (156.5-14)
13 Even Alcinous' variant term for the primary perceptible, ltOtO'tll~ or
'quality', is so named for the first time later on in the Theaetetus, at 182a. That
this, in the context of Did. ch. 4, is a specifically Theaetetan usage is confirmed
in ch. 11, where Alcinous uses ltOtO'tll~ more broadly for any quality at all.

ALCINOUS' EPISTEMOLOGY

311

The intelligible world, on the right-hand side, is composed of

Ideas, grasped by primary noesis. 14 However, on the left-hand side,
the sensible world is composed not of primary or secondary
aistheta, but of athroismata. That is, I take it, the sensible world is not
properly grasped as a complex set of colours, flavours etc. (primary
perceptibles); nor is it properly grasped as a complex set of white
things, sweet things etc. (secondary perceptibles). If it were composed merely of either primary or secondary perceptibles, the proper cognitive route to it would be the direct exercise of perception,
albeit (as line 8 reminds us on the left-hand side) with the aid of
doxastic logos. That is, if the sensible world were composed merely
either of colours, flavours etc., or of white things, sweet things, etc.,
we would be left simply to register these with our sense organs,
although doxastic logos would be called in to enable us to identify
such sensibles for what they are-white, sweet etc. That would
make the study of the sensible world a most primitive and unrewarding project.
However, the sensible world is not properly analysed as consisting of these simple sensibles. As line 11 indicates, the sensible
world is composed of objects-mountains, animals, stars etc.each of which is, epistemologically speaking, a complex athroisma
of colours, odours etc. Such objects cannot be grasped by simple
sense-perception, which (as Theaetetus 184-186 demonstrates) is
incapable of synthesizing its own data by placing them in any sort
of relation to each other. Such synthetic judgements are, rather,
argued by Socrates there to be the work of doxa (187a). In other
words, the Theaetetus itself has an epistemology which establishes
that the sensible world is properly studied, not by mere aisthisis,
but by doxa which uses aisthesis as its instrument. And this fits
beautifully with the celebrated epistemological declaration in
Alcinous' favourite dialogue, the Timaeus (28a), that whereas thatwhich-always-is is graspable voiJcret J.l.E'ta A.Oyou, 'by intellection
with the help of reason', that-which-becomes is 00~1J J.l.E't' aicr9iJcrero~
aA.Oyou oo~acr'tov, 'opinable by opinion with the help of irrational
perception'. That is why Alcinous is so careful to state, at 156.9-10,
that the athroisma is judged by 'doxastic logos not without perception'. It is true that Timaeus 28a makes the sensible world graspable,
not by doxastic logos, but doxa itself. But since the Timaeus is explicit
14 That this primary intellection is non-discursive is Alcinous' interpretation of m:.ptATJ1ttov at Tim. 28al-2, represented by his 7tEptAtlljfEt at 156.6.

312

DAVID SEDLEY

soon after (29b) that the intelligible and sensible worlds each have
their own appropriate kind of logos, Alcinous is confident that doxa
at 28a is simply a shorthand for 'doxastic logos'. He has already
carefully prepared the ground for this equivalence by telling us
earlier (154.28-9, p. 304 above) that doxastic logos is sometimes
simply called doxa.
(This realization that doxa, and not bare sense-perception, is
required if we are to have cognitive access to athroismata, can now
be fed back into our earlier inquiry as to whether the passage from
aisthesis (line 6) to doxastic logos (line 3) need be mediated by doxa.
Since doxastic logos deals with complex perceptual items like fire
and man, there should now be no doubt that mere sense-perception
can never be a sufficient foundation for it, and that doxa must be
indispensable to the process of its formation.)
Alcinous' exegetical strategy is hideously complicated. But if we
persevere in tracing it, we can begin to see how thoroughly he and
his fellow-proponents of the interpretation have integrated the
epistemology of Theaetetus Parts I and II into a global reconstruction
of Plato's two-world system. On his view, the Theaetetus may
profess to be an inquiry into knowledge, but in reality it is Plato's
fullest account of perceptual cognition and its objects, demonstrating the parallel structures of the intelligible and sensible worlds.
In this way it analyses the type of reason to which we may appeal
in a cosmological investigation like that conducted in Alcinous'
own favourite dialogue, the Timaeus. Matters of no small importance are at stake:
The aim of physics is to learn what is the nature of the universe;
what sort of animal man is and what place he has in the world;
whether god exercises providence over things in their entirety;
whether this god has others subordinate to him; and what the
relation of human beings is to the gods. (161.3-7)
To reconstruct what cognitive capacity, in Plato's view, entitles us
to reason about such questions is no mean achievement.I5
Cambridge
15 I have benefited from discussion of this material in the Didaskalikos
seminar held at Cambridge in May 1995, and in the Seminaire Leon Robin
at the Sorbonne in December 1995. My thanks to participants in both seminars-including, at the former, the honorand of this volume Jaap Mansfeld,
whose expertise and good will contributed so much to the entire occasion.

'A SIMPLE PHILOSOPHY'

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY
PIETER

W.

VANDER HORST

'Christianity is a simple philosophy.' This is the opening line of

Alexander's llpo~ 'ta~ Mavtxaiou M~a~. 1 When Jaap Mansfeld and I
translated and annotated this neglected treatise in 1973, we paid
some attention to Alexander's way of describing Christianity and
Jesus, but some aspects still deserve a closer look.2 This early antiManichaean work by a Platonist philosopher was probably written
towards the end of the third century CE, in the period when Manichaean missionaries3 were trying to gain a foothold in Egypt, 4 an
attempt that was crowned with success, as the Coptic Manichaica
from the Fayyum and the recent find of numerous Manichaean
papyri in Kellis in the Dakhleh oasis bear witness. It is the radical
dualism of this new religion that was Alexander's main target, but
1 P. W. van der Horst & J. Mansfeld, An Alexandrian Platonist Against Dualism: Alexander of Lycopolis' Treatise 'Critique of the Doctrines of Manichaeus' translated with an introduction and notes (Leiden 1974; originally published as an
article in the short-lived journal Theta-Pi 3 ( 1974) 1-97) 48. The Greek of the
opening line is: 'H Xptanavrov cjltl..oaocjlia cml..fi x:al..et-rat. Our translationwhich I slightly altered in the text above-is based on the assumption that
JCaA.eta9at is here the equivalent of et vat, as is often the case; see LSJ s. v.
JCaA.ero II 2, and C. ]. Ruijgh, 'L'emploi onomastique de KEKAHl:8AI ', in
Miscellanea tragica in honoremj. C. Kamerbeek (Amsterdam 1976) 333-397.
2 Even though A. Villey in his recent commentary has paid more attention to the matter than we did; see his Alexandre de Lycopolis: Contre la doctrine
de Mani (Paris 1985).
3 The first of which was Pappos (2, p. 4 Br.), a close friend of Mani
himself; see S. N. C. Lieu, Manichaeism in the Later Roman Empire and Medieval
China (Tubingen 1992, 2nd ed.) 103, and Villey, Alexandre 109-11.
4 The earliest attempts took place during;Mani's lifetime (216-277), especially in the Lycopolitan area; see G. Stroumsa, 'The Manichaean Challenge
to Egyptian Christianity', in B. A. Pearson & J. E. Goehring ( eds.), The Roots
of Egyptian Christianity (Philadelphia 1986) 307-19. A sure indication of the
movement's presence in Egypt by the final decades of the third century is the
circular letter of Theonas (?), bishop of Alexandria, from ca. 280, in which
he warns Christians against 'the madness of the Manichaeans' (-rij~ J.!OVia~
-rrov Mavtxerov); the text has been preserved in P.Ryl.Gr. 469 (the quote is in
line 30) and is printed in A. Adam, Texte zum Manichiiismus (Berlin 1959, 2nd
ed.) 52-4 (ed. pr. by C. H. Roberts, Catalogue of the Greek and Latin Papyri in the
john Rylands Library, vol. 3, Manchester 1938, 38-46).

314

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY

since he regarded Manichaeism as a Christian sect, 5 he pays some

attention to Christianity as a whole as well. And he does so in a
way that shows some resemblances to, but also marked differences
from, the other- well-known Greek thinkers who attacked Christianity before, during, and after his time, i.e. Celsus, Porphyry,
and Julian.
The fact that Alexander calls Christianity a philosophy is less
remarkable than it might seem at first sight. Not only Christians
themselves called their religion a philosophia, some of their opponents did so as well. 6 As several scholars have demonstrated,7 in the
Hellenistic and Roman period the term philosophia underwent a
semantic development that, as Armstrong put it, 'accounts for the
strong ethical emphasis and, to the modern mind, disconcertingly
close connexion between philosophy and religion which we find
in nearly all the thinkers of the period, in the Greek pagans just as
much as in the adherents of revealed religions.' 8 It should be
added, however, that the emphasis on moral and religious concerns in these 'philosophies' often implied a decrease in attention
to logic (and physics as well, albeit to a lesser degree). And it is
precisely a lack of logical rigour, with its implications for physics,
that Alexander sees as one of the weaknesses of Christianity (and
Manichaeism). He complains that Christianity is 'chiefly devoted
to ethical instruction' (1, p. 3 Brinkmann) 9 but that, as far as clear
and coherent theories and ideas about God are concerned, it gets
5 Rightly so, as we now know from the Cologne Mani Codex, where. it is
explicitly stated that Mani was educated in an Elchesaite milieu; see, e.g.,
Lieu, Manichaeism in the Later Roman Empire 1-85.
6 For references see A. J. Malherbe, 'Not in a Corner': Early Christian
Apologetic in Acts 26:26, The Second Century 5 (1985/86) 197 n. 20.
7 A. M. Malingrey, Philosophia. Etude d'un groupe des mots dans la litterature
grecque des Presocratiques au !Ve siecle apres f.-C. (Paris 1961); R. L. Wilken, The
Christians as the Romans Saw Them (New Haven-London 1984) 72-83; A. Villey,
Alexandre 93-95.
8 A. H. Armstrong, Introduction, The Cambridge History of Later Greek and
Early Medieval Philosophy (Cambridge 1967) 5. Since the dossier on philosophia
consists mainly of pagan and Christian evidence (although Malingrey, Philosophia 87-8, did pay some attention to Philo), I add some Jewish material at
random by pointing out how Josephus speaks about the religious movements
in the Judaism of his time (Essenes, Pharisees, Sadducees, Zealots) as philosophiai, and how he speaks about the Jewish religion in general as 'our
philosophy': Ant. XVIII 9; 11; 23; Bell. II 119; 166; C. Ap. II 47.
9 We still (have to) use Brinkmann's old Teubner edition (Leipzig 1895)
but a new critical text is badly needed. Villey's new chapter division makes
more sense than Brinkmann's from a logical point of view but does not
facilitate reference.

PIETER W. VAN DER HORST

315

no further than enigmatic talk and allusions. Theology requires

logic, but logic is not at all part of the Christian heritage. The
truism that 'the productive cause is the most honourable, the most
important, and the cause of all beings' (1, p. 3 Br.) is apparently all
that Christian philosophy has been able to produce. Even worse is
the fact that also in ethics, the domain that the Christians devote
most of their attention to, they sidestep the difficult issues. They do
not know the important distinctions between various forms of vice
and virtue, of disposition and affection, that Greek philosophy (esp.
the Stoic school) had developed, 10 and consequently are unable to
provide 'the principles according to which each individual virtue
should be acquired' (1, p. 3 Br.). So even their best-ethical instruction-is of a lamentably deficient nature. Nonetheless, Alexander
has to admit, it is effective: many ordinary people who follow these
ethical instructions make great progress in virtue, their characters
receive an imprint of piety, and they are gradually led towards a
desire for the good. So on the very first page of his treatise
Alexander not only debunks Christianity as a philosophy, he also
concedes-and unreservedly at that-that the effects of this simple
ethical instruction are impressive. And he says so because this is
his own experience: K 'tile; 7teipac; ecrn J.L08etv (1, p. 3 Br.). This
gives us an important clue to Alexander's attitude towards
Christianity. It seems to indicate that he knew many (uneducated)
Christians in his own environment whose doctrines he scorned,
but whose way of life he could not but admire. In the following
pages we will see that this ambiguous attitude may be seen at
several turns in his work. II
Now we should bear in mind that late third century Egypt was a
country that witnessed the rise of Christian monasticism. One of
the first Egyptian Desert Fathers (though, as we shall see, not the
very first), the long-lived Antony (ca. 250-355), withdrew in the
early seventies of the third century from the inhabited world to the
desert and the mountains of the Egyptian hinterland. Before the
end of the century many believers followed his example. Their
life was ascetic, often to the extreme, their goal was meditation in
I 0 See M. Giusta, I dossografi di etica, 2 vols. (Turin 1962-67). A similar
critique is voiced by Celsus ap. Origen, CCI 4.
II It should be kept in mind here that Alexander's positive view of Christian ethical instruction stands in contrast to his very negative view of the
Manichaean position that education and punishment should be abolished; see
van der Horst-Mansfeld, An Alexandrian Platonist 44-5.

316

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY

order to become as free as possible from bodily desires and so to

draw closer to God. The Historia monachorum in Aegypto, in spite of its
often hagiographic character, gives us a lively picture of the
impressive zeal and remarkably self-effacing attitude of many of
these early hermits (and of the cenobites not much later).I2 Other
examples could be cited, for instance the even earlier anchorite
Paul of Thebes (reportedly ca. 228-341),13 whom Antony met in the
desert. Antony is mentioned here first because some of his
writings are extant, namely various letters and sermons. 14 What is
striking in these documents is that, although they do contain some
simple theological ideas, the main emphasis is on ethical instruction. So this contemporary of Alexander illustrates exactly what
our philosopher writes about: poor doctrine-often called philosophiaP5-combined with an impressive lifestyle. And this is also
the impression one receives from other documents of the early
monastic movement. There is in these humble Coptic anchorites
an enormous zeal, a total dedication to the search of God, an
extremely strict ethics (often bordering on the bizarre), combined
with a lack of intellectual training.I6 Meetings with people like
Antony and his admirers could not fail to have inspired Alexander's ambiguous attitude. We do not know exactly when Christian
monasticism spread to Lycopolis (modern Assiut), but it is very
probable that our author (could have) got acquainted with this
phenomenon in this city as well. One of the most famous Desert
Fathers was John of Lycopolis who lived there probably only
slightly later than Alexander but who was certainly not the town's
first Christian or first monk.I7
12 See A.:J. Festugiere, Historia monachorum in Aer;ypto (Brussels 1961); N.
Russell, The Lives of the Desert Fathers (London-Oxford 1980); P. W. van der
Horst, Woestijn, begeerte en geloof Het leven van de eerste monniken in Egypte. De
Historia monachorum in Aegypto vertaald en toegelicht (Kampen 1995).
13 See I. S. Kozik, The First Desert Hero: St. Jerome's Vita Pauli (Mount
Vernon 1968); M. Naldini in Dictionnaire encyclopedique du christianisme ancien
(ed. A. di Berardino, Paris 1983) vol. 2, 1951-2.
14 See T. Orlandi in Dictionnaire encyclopedique du christianisme ancien, vol. I,
169, and S. Rubenson, The Letters of St. Antony (Lund 1990; non vidi).
I5 For references see P. Courcelle, 'Verissima philosophia', in Epektasis
(FS[ean Danietou) (Paris 1972) 655-6, nn. 15-19.
I
Several of the documents make clear that many of the Desert Fathers
were illiterates. See R S. Bagnall, Er;ypt in Late Antiquity, Princeton 1993, 246260.
l7 That Alexander's remarks could not be directed against his philosophically trained (near-)contemporaries such as Origen and Pierius ('the
new Origen'), is rightly stressed by Villey, Alexandre 96.

PIETER W. VANDER HORST

317

A short comparison with Galen may be instructive here. In a

lOth century Arabic chronicle by AgapiusiB Galen is reported to
have written the following: 'The people called Christians have built
their doctrine upon enigmatic indications and miracles. But as far
as their actions are concerned, they are not inferior to the genuine
philosophers. For they love continence, keep fasts and prayers, and
avoid unjust actions. Among them there are men who do not
pollute themselves with women.' 19 If this testimony is genuine,2which it probably is since it is independently confirmed by
several other sources-we have here a striking agreement between
Galen and Alexander.2 1 Both of them do not hold the Christian
doctrine in high regard, but they do show appreciation for the way
of life of (some) Christians. And Galen also specifically says that
his high regard for their way of life (not inferior to that of
philosophers!) was primarily inspired by their asceticism and
strict ethical norms, which is exactly what we surmised to be the
case with Alexander.22
It would seem that Alexander not only had knowledge of the
impact of Christian ethics on the lives of simple people, but that he
was also acquainted with Christian theories about history, especially the history of the Christian church in the first 250 years. He
says that 'this simple philosophy has been split up into numerous
factions by its later adherents' and that 'the number of issues has
increased just as in sophistry' ( 1, p. 3 Br.). 23 This was not his own
invention, for in the century before he wrote this, several Christian
writers had developed this theory of the gradual 'heresification
and schismatization' of the church after its initial 'orthodox' start
18 A passage overlooked by R. Walzer, Galen on jews and Christians (Oxford
1949) 56-74, although he does mention several other Oriental authors who
attribute similar statements to Galen (see esp. the quotes at 57 and 65; cf. p. 15).
19 I quote the translation by S. Pines, An Arabic Version of the Testimonium
Flavianum and its Implications (Jerusalem 1971) 74.
20 For the debate see Pines, Arabic Version 73-82.
21 Walzer, Galen 72-3, already noted the parallel.
22 In contrast, most of the other second century critics of Christianity
speak in a very negative sense about Christian morality. See for references
R. J. Hoffmann, Celsus. On the True Doctrine. A Discourse Against the Christians
(Oxford 1987) 12-24.
23 The final words JCa9a7tep ev to'i~ eptO'ttlCOt~ point to a parallel in the
history of philosophy and are reminiscent of the passages from the second
century Middle Platonist Numenius to the effect that the entire history of the
Academy after Plato had been one of strife and lack of agreement on any
essential topic (fragments 1 - 4 Leemans = frgs. 24-27 des Places = Eusebius, PE
XIV 5-8).

318

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY

with an unadulterated gospel. A good example is found in Alexander's earlier compatriot and fellow Platonist, Clement of Alexandria, who says in his Strom. VII xvii 106-7 that the Lord founded the
Catholic Church in the times of Tiberius and that it is evident from
the high antiquity and perfect truth of this church that the later
heresies, from the times of Hadrian and Antoninus onwards, were
new inventions and therefore falsifications of the original truth.24
This myth of Christian origins became immensely influential. In
fact it remained the standard Christian view for some 1500 years
until it was subverted at the end of the 17th century when Gottfried
Arnold published his important Unparteiische Kirchen- und Ketzerhistorie ( 1699-1 700) .
Now one could of course argue that Alexander's observation of
the splitting up of the Christian philosophy into numerous factions
can simply be explained by the fact that in his days Egyptian
Christianity was indeed split up into numerous sects, as the Nag
Hammadi Codices, the Kellis papyri, the Coptic Manichaica, the
other early Coptic literature, and the writings of the Graeco-Egyptian Churchfathers bear witness. All of these testifY to such a great
variety and wide divergence of opinion in Christian Egypt that
any observer could have come to the same conclusion as Alexander.25 This view does not hold, however, for the simple reason that
a wide variety of viewpoints, and even the fact that the various
Christian parties themselves argued that they and only they were
continuing the original unity of doctrine, would not automatically
lead to a conclusion by an outsider to the effect that an original
'simple philosophy' was adulterated only by later (emyEVOjlEVOt is
Alexander's term) depraved adherents. So it would seem that Alexander's view derives from Christian authors or theorists, the more
so since he joins them in accusing the 'dissenters' of 'theoretical
imprecision' (1, p. 4 Br.) which formed the basis of the decline of
24 See also the fragment of Hegesippus (jloruit 3rd quarter of the 2nd cent.)
in Eusebius, Hist. Eccl. IV 22, 4-5, where this author says that the earliest
Church leaders 'used to call the Church a virgin for this reason that she had
not yet been seduced by listening to nonsense,' but soon afterwards this
process of corruption and heresy started. Cf. also the similar statements in
the slightly later Irenaeus, Adv. Hacreses III 4. See S. J. D. Cohen, 'A Virgin
Defiled. Some Rabbinic and Christian Views on the Origin of Heresies',
Union Seminary QJlarterly Review 35 (1980) 1-11, and esp. A. Le Boulluec, La notion
d 'heresie dans la litterature grecque Ile-Ille siecles, vol. 2 (Paris 1985).
25 See W. Bauer, Orthodoxy and Heresy in Earliest Christianity (London 1972)
44-60 (though no longer up to date).

PIETER W. VANDER HORST

319

their ethical instruction and of the inclination to strife of the

common people in the Church, thereby contradicting what he had
earlier said about the low level but high results of the moral
teachings of Christianity.26
In his refutation of Manichaeism, the second part of the book
which forms the bulk of the treatise ( 5-26 Br. = 14-42 Villey),
Alexander begins by saying that he is at a loss how to proceed.27
Like the other Christians, Alexander implies, Manichaeans too
employ no proper methods in their theories and do not care for
proof, an argument that recalls Galen's repeated objection to Christian faith. The fact that they do not use generally accepted rational
arguments makes it wellnigh impossible for him to refute them. It
is interesting to see that one of the arguments of his opponents
seems to have been that they do not need to give proofs because the
basis of their reasoning is 'their old and new Scriptures which
they believe to be divinely inspired' (5, p. 8 Br.). It is most interesting to observe that the term Alexander uses here is 9e01tVEUcr'to<;.
This is a word of very rare occurrence. As far as we know, it first
occurs in the New Testament passage 2 Timothy 3:16 and, apart
from the innumerable instances in the Churchfathers who borrowed the term from 2 Timothy,2B the very few other occurrences
are either from Jewish writings or also based upon this New
Testament passage. It is, therefore, probable that Alexander either
knew that passage itself or had read the term in a Christian or
Manichaean writing that had borrowed it from the New Testament. He does not hide his scorn for this kind of appeal to divine
authority and revelation,29 arguing that under these circumstances
'all normal rules are abandoned' because 'the role attributed by
Greek philosophers to the postulates, namely the underived propositions upon which proofs are based, is represented among these
26 To be sure, it cannot be excluded that Alexander borrowed the idea
from Celsus; see CCIII 10, V 63.
27 He adds the sour remark that this deceitful philosophy has unfortunately been successful in making some of his fellow-students to converts, a
situation comparable to that in Plotinus' school where some of his friends
had adopted gnostic ideas, much to his annoyance (Enn. II 9,10). See on this
remark the observations by Lieu, Manichaeism 152-3.
28 See G. Lampe's Greek Patristic Lexicon s.v.
29 Like Julian. For Julian's repugnance to the notion of revelation see A.
Meredith, 'Porphyry and Julian Against the Christians', ANRW II 23, 2
(Berlin 1980) 1140-1 (1119-49).

320

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY

people by the voice of the prophets' (5, p. 9 Br.). This remark is a

further indication that in his anti-Manichaean struggle he uses
arguments that were originally directed against Christianity in
general. 30 Our conclusion is confirmed by the fact that Alexander
here refers to the Old Testament as well, whereas he knew that
this document was completely rejected by the Manichaeans and
certainly not regarded as 'divinely inspired' by them; on the
contrary, they regarded 'the prophets' as inspired by the devil.31
The same can be seen again in a later passage (10, p. 16 Br.),
where Alexander says that the Manichaeans are even worse than
the mythographers who are responsible for the stories of Uranus'
castration and Kronos' swallowing his own sons, since their
theory of a continual war of matter against God is too crude to
admit of allegorical interpretation and is actually taken literally by
them. Here, too, we have a traditional anti-Christian argument,
already found in Celsus, who not only adduces the same examples
(Uranus, Kronos, etc.) but also says that the Christian writings are
incapable of being interpreted allegorically and are therefore
manifestly very stupid fables (ap. Origen, CCIV 48-50). And when
a little further on Alexander states that the Manichaean theory of
God's anger and his desire for vengeance against matter is
untenable from a philosophical point of view-since 7t09T\ of this
kind 'are absurd even in connection with the good man, let alone
in connection with the Good itself (10, p. 17 Br.)-he again echoes
30 See, inter multos alios, W. Nestle, 'Die Haupteinwande des antiken
Denkens gegen das Christentum', Archiv fur Religionswissenschaft 37 (1941/42)
72 (51-100). In general it can be said that, in spite of the many differences
between Celsus, Porphyry, Julian, and Alexander, what they have in common is a strong resentment against fideism. See also the remarks by A. H.
Armstrong, 'Man in the Cosmos. A study of some differences between pagan
Neoplatonism and Christianity', in Romanitas et christianitas. Studia ]. H.
Waszink oblata (Amsterdam-London 1973) 5-14, and by R. T. Wallis, Neoplatonism (London 1972) 101-105. On Alexander see also G. Stroumsa, 'Titus of
Bostra and Alexander of Lycopolis: A Christian and a Platonic Refutation of
Manichaean Dualism', in R. T. Wallis & J. Bregman (edd.), Neoplatonism and
Gnosticism (New York 1992) 337-349 (340: 'The philosopher is ill at ease
developing dialectical arguments against a protagonist who does not accept
the rules of the game').
31 See Villey, Alexandre 194-5. My interpretation of the passage makes
unnecessary Villey's forced explanation that 'the old scriptures' refers to the
apocrypha and pseudepigrapha in use among the Manichaeans, however true
it may be that these writings enjoyed a great popularity among them; see
J. C. Reeves, Jewish Lore in Manichaean Cosmogony. Studies in the Book of Giants
Traditions (Cincinnati 1992).

PIETER W. VAN DER HORST

321

one of Celsus's anti-Christian arguments to the effect that ascribing

human emotions like anger to God is ridiculous ( ap. Origen, CCIV
71-73), an argument that Origen has a hard time refuting.32
I further wish to draw attention to an interesting aspect of Alexander's polemic on the aeternitas mundi inch. 12 (p. 19 Br.), a view
firmly rejected by the Manichaeans. He quotes Zeno33 to the effect
that the universe will be destroyed by fire, and opposes this by a
quote from 'a subtle thinker' (Theophrastus?, Eudemus?) who is
said to have argued against Zeno: 'What I for one have seen yesterday and a year ago and an even longer time ago, and what I
similarly see today, I do not see to have suffered at all through the
fire of the sun. But in the course of time some little damage should
have been done if we are to believe that the whole universe will
ever be destroyed by fire.' A similar answer, Alexander says, is
valid against the Manichaeans as well, namely: 'But evil has not
in the slightest respect grown less. It hasn't: it was there formerly
in the time of the first man to be born, where brother killed brother, and even now it is still there; wars are the same and desires
are even more complicated. Now it would be reasonable that these
things, if they did not altogether cease, should at least grow less, if
we are to trust that in the long run they will ever cease. As, however, they are the same as formerly, what confidence in future
events can be possible on this basis?' (12, p. 19 Br.). Now, apart from
the fact that, as Mansfeld had already noticed,3 4 Alexander's
predecessor Celsus had already adduced against the Christians the
same argument that 'in the existing world there is no decrease or
increase of evils either in the past or in the present or in the future'
( CC IV 62), two things have to be noticed here. The first is a striking agreement between the argument here and the one we find in
one of the last New Testament documents to be written, namely 2
Peter, ch. 3. In this pseudonymous letter (probably from the early
32 For a concise history of the problem see P. W. van der Horst, 'Philo
Alexandrinus over de toorn Gods', in A. de Jong & A. de Jong (edd.), Kleine
en?:clopedie van de toum (Utrecht 1993) 77-82.
'
3 Honesty demands that we state explicitly that our claim in 1973 to have
discovered here a new and authentic fragment of Zeno overlooked by Hans
von Arnim (van der Horst-Mansfeld, Alexandrian Platonist 74 n. 294) was
invalid. More than 50 years earlier the great Arthur Stanley Pease had
already detected it (Classical Philology 1921, 200). We note that Villey too has
overlooked Pease's contribution.
34 Alexandrian Platonist 75 n. 299.

322

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY

second century) we find the following situation:35 The author tries

to refute opponents, Christian opponents that is, who find it impossible any longer to believe in the second coming (the parousia) of
the Lord, 36 and who say: 'Where is the promise of his coming? For
ever since the fathers37 died, all things continue as they were from
the beginning of creation!' (v. 4). He then goes on to demonstrate
that, despite the apparent unchangeability of all things, the world
will in the end certainly be destroyed by fire and the elements
will be dissolved by it (a Stoic motif}. In his reply there are also a
number of references to the early chapters of the book of Genesis
(esp. to chs. 1 and 6 in vv. 5-7), probably induced by his opponents'
reference to the Patriarchs. This would seem to be not without
significance. The opponents of the NT author use the immutable
nature of the cosmos as an argument against the idea of a final
conflagration (and the second coming of the Lord), and this is
very similar to Alexander's argument. 38 In his counterattack, the
author of 2 Peter refers to the two biblical stories in Gen. 1 and 6, not
unlike AlexandeF's reference to the two biblical stories in Gen. 1
and 4. For-and that is the second aspect to be noticed-his words
about 'the time of the first man to be born, when brother killed
brother' are undeniably a reference to the creation myth in Gen. 12 and to the story of Cain's murder of Abel in Gen. 4. Why should
Alexander use biblical references in his refutation of those who
think the world will be destroyed by fire? Could it be that, because
he regards the Manichaeans as Christians, he tries to beat them
with arguments that were being used within Christian circles in
the debate over the eternity or the destructibility of the world?39
That such a debate was going on in some Christian circles is
apparent not only from the passage in 2 Peter 3, but also, for
instance, from passages such as 1 Clement 23:3 and 2 Clement 11:2,
35 See the discussion in P. W. van der Horst: 'The Elements Will Be Dissolved With Fire.' The Idea of Cosmic Conflagration in Hellenism, Ancient
Judaism and Early Christianity, in my Hellenism-judaism-Christianity: Essays
on Their Interaction (Kampen 1994) 227-251, esp. 243-248. Further R. J.
Bauckham, Jude, 2 Peter (Waco 1983) 300-301.
36 A problem caused by the so-called 'Parousieverzogerung'.
37 I.e., probably, the Jewish Patriarchs (Abraham, Isaac, and Jacob).
38 On Alexander's anti-Stoic stance see Mansfeld in An Alexandrian
Platonist 47.
39 But note that prior to Alexander Celsus too had referred to 'the plot of
brothers against one another' (ap. Origen, CC IV 43 and V 59) as did Julian
after him (C. Galil. 347 A-C = fr. 84 Masaracchia).

PIETER W. VANDER HORST

323

where the authors combat 'the double-minded, who doubt in their

heart, and say: These things we heard of old in the days of our
fathers also, but behold we have grown old and none of these
things has happened to us!' These people doubt God's promises,
among which that of a new world, because they have never seen
any of them come true. Other Christians, who did not doubt God's
promises, still found it difficult to believe in a destruction of the
universe by fire, since as an act of punishment, inspired by anger,
it seemed to be to them not 8eo7tpemic;. 40 So they took recourse to
allegorical interpretation of the biblical texts concerned (e.g.,
Origen). 41 The debate between these diverging Christian groups or
individuals was no doubt waged with constant recourse to the
Scriptures. It cannot be ruled out that this is the background of
Alexander's at first sight rather surprising 42 presentation of an
argument that makes reference to Scriptural passages.43
There is another passage in which Alexander shows his knowledge of the Bible, sc. ch. 24 (p. 36 Br.), where he draws attention to
the fact that there are some parallels to the idea of Christ's selfsacrifice: 'To say in accordance with the doctrine of the Church
that he gave himself up for the remission of our sins has a certain
plausibility in the eyes of the many44 because of historical
parallels: in Greek history where we often read about people giving
themselves up in order to save their cities; 45 also Jewish history,
which prepares the son of Abraham for being sacrificed to God, 46
contains an example of such a tale.' That Alexander knew the
story of Genesis 22 seems beyond doubt at first sight, 47 but we have
to raise the question whether perhaps he knew the story (only or
40 0. Dreyer, Untersuchungen zum Begriff des Gottgeziemenden in der Antike
(Hildesheim-New York 1970).
41 See my essay 'The Elements .. .' 250.
4 2 Surprising also because in ch. 24 (p. 35 Br.) Alexander appears to know
that the Manichaeans rejected the Old Testament.
43 This may have been one of the facts that contributed to the longstanding but erroneous view, inaugurated by Photius (Contra Manichaeos I 11),
that Alexander was the bishop of Lycopolis, which is the reason why his
treatise was also printed in Migne's Patrologia Graeca (vol. 18), published in
1856, even though I. de Beausobre had decisively refuted this view as early as
1734 (Histoire critique de Manichie et du manicheisme, vol. I 236-7).
44 Not of Alexander himself, that is.
45 A parallel drawn already by Celsus in CC I 31.
46 A clear reference to the so-called aqedat Yitschak ('the binding of Isaac')
in Genesis 22.
47 Julian also knew the story, C. Galil. 343C-D = fr. 83 Masaracchia.

324

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY

also) as expanded in Jewish (or Christian) interpretation, for he

mentions Isaac ('the son of Abraham') as an example of 'people
giving themselves up,' whereas the biblical story portrays Isaac as
just a completely passive object of God's command and his father's
activity. In post-biblical Jewish sources of the Roman period, however, we find the motif that Isaac cheerfully volunteered to be the
sacrifice God wanted to have, and that he did this for the welfare of
his people, the Jews. To quote just one of the many examples,
Targ;um Pseudo-jonathan, an Aramaic (strongly paraphrastic) translation of the Bible from late antiquity (but containing many older
traditions), 48 renders Gen. 22:10 ('Then Abraham put forth his
hand and took the knife to slay his son.') as follows: 'Isaac spoke up
and said to his father: 'Tie me well lest I struggle because of the
anguish of my soul, with the result that a blemish will be found in
your offering, and I will be thrust into the pit of destruction.' The
eyes of Abraham were looking at the eyes of Isaac, and the eyes of
Isaac were looking at the angels on high. Isaac saw them but Abraham did not see them. The angels on high exclaimed: 'Come, see
two unique ones who are in the world: one is slaughtering and one
is being slaughtered; the one who slaughters does not hesitate, and
the one who is being slaughtered stretches forth his neck'. ' 49 In the
first century pseudo-Philonic Libcr Antiquitatum Biblicarum XVIII 5
the soteriological aspect is stressed. Here it is even said that God
chose the people of Israel on account of Isaac's blood (pro sang;uine
eius)! Even though Alexander seems to be aware of the fact that the
biblical story does not describe the actual sacrifice of Isaac ('jewish
history ... prepares the son of Abraham for being sacrificed to God'),
the sheer fact that he mentions this tradition on a par with the
Greek stories of heroes sacrificing themselves in order to save their
cities strongly suggests that he was also aware of the soteriological
interpretation that this story was given in post-biblical Jewish
exegesis. 5 If this is correct, it could mean that Alexander not only
48 See U. Glessmer, Einleitung in die Targume zum Pentateuch (Tiibingen 1995)
181-196.
49 See for this translation and notes M. Maher, Targum Pseudo-Jonathan:
Genesis (The Aramaic Bible, vol. Ib, Edinburgh 1992) 79-80. The Aramaic
original is in Biblia Polyglotta Matritensia, ser. IV: Targum Palaestinense in Pentateuchum, Lib. 1: Genesis, ed. A. Diez Macho (Madrid 1988) 143.
50 Sh. Spiegel, The Last Trial. On the Legends and Lore of the Command to Abraham to Offer Isaac as a Sacrifice: The Akedah (Woodstock 1993 =New York 1967),
offers many examples of this exegesis. The best bibliography on the aqedah to
date is by B. Cozijnsen and P. Wansink at the end of their contribution to

PIETER W. VAN DER HORST

325

knew the Bible, together with various Christian wntmgs and

traditions, but was perhaps also acquainted with Jewish haggadic
material, either oral or written, or with Christian variants of these
interpretations.
This impression is strengthened by ch. 25 (p. 37 Br.) where
Jewish tales are favourably compared with Manichaean ones. In
this case he remarks that 'when the history of the Jews speaks of
the angels who consorted with the daughters of men in order to
have sexual intercourse, this way of telling the story hints at the
nurturing faculties of the soul which comes down hither from
above.' This reference to the famous but enigmatic passage in Gen.
6:1-4 is revealing in a number of respects. Firstly, the biblical text
does not speak about 'angels' but about 'sons of God' who had sexual
intercourse with earthly women. The Septuagint also reads here oi
uiot toil 9eou, although there is a widely attested but certainly
secondary variant dyyeA.ot, a reading supported by Josephus, Ant. I
73, 51 but evidently an interpretation meant to get rid of the
somewhat embarrassing 'sons of God. '52 This interpretation of 'sons
of God' as angels was later adopted also by the Christians, so it is
hard to decide whether Alexander learned this interpretation from
Jewish or from Christian sources (a LXX ms he knew?). Secondly,
he interprets the angels as ai 9pe1ttu.:a1 ouvaJlEt<; tile; 'lfUXfi<; who
come down from above. Though Philo's interpretation of the biblical verse (Gig. 6-18, on Gen. 6:2) is rather different from Alexander's, what they have in common is that the angels are here taken
in an allegorical sense as the souls, or functions of the souls, of
men. It may well be that Alexander is here drawing on a Jewish
Alexandrian tradition of interpretation of Gen. 6:1-4, although it
should be added that this biblical passage has evoked such an
M. van den Berg et al. (edd.), Uit de sjoel geklapt. Christelijke belangstelling voor
joodse traditie (Hilversum 1986) 143-5.
51 Philo has the mixed reading 'angels of God' (Gig. 6), although at
Quaest. Gen. I 92 he remarks that 'Moses sometimes calls the angels 'sons of
God' because they are made incorporeal.' See D. Winston & J. Dillon, Two
Treatises of Philo of Alexandria (Chico 1983) 236: 'It seems as if someone in the
Alexandrian tradition was offended, as well he might be, by the idea of God
having sons, and glosses 'sons' by the less offensive term ayyel..ot.' But the
phenomenon was certainly not restricted to Alexandria, as the Targumim
show (see the references in the next note).
52 See J. W. Wevers, Notes on the Greek Text of Genesis (Atlanta 1993) 75-6;
Ph. S. Alexander, 'The Targumim and Early Exegesis of 'Sons of God' in
Genesis 6', Journal ofJewish Studies 23 (1972) 60-71.

326

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY

enormous amount of speculation and literature among both Jews

and Christians in antiquity that, again, it is hard to decide what
exactly his source was. Nonetheless, we may conclude that he was
acquainted not just with the Bible but also with Jewish and/ or
Christian interpretations of it. 53
It is conspicuous that the biblical passages we have seen Alexander referring to are almost all from the early chapters of Genesis.
This aspect deserves some additional comments. Recently Giancarlo Rinaldi has collected all the references to the Bible in pagan
Graeco-Roman authors of the imperial period.5 4 This useful catalogue contains 715 entries of which no less than 96 pertain to
Genesis. That is to say that some 13.5% of the biblical references in
pagan literature of these centuries is to Genesis, which is quite
disproportionate in view of the fact that the book of Genesis takes up
no more than some 4.4% of the Bible (i.e., OT and NT). That is to
say that Genesis is referred to more than three times as often as
might be expected on average. The first book of the Bible seems to
have been much better known among pagan authors than any
other book of the Old Testament.5 5 When we take a closer look at
the 96 entries for Genesis, we see that no less than 63 of them are
references to what I called the early chapters, i.e. the so-called
'Urgeschichte' in Gen. 1-11.56 These are exactly the chapters that
most of Alexander's references are to: Gen. 1, 4 and 6 (the only
exception being Gen. 22). So Alexander neatly fits into the general
pattern we see elsewhere. The unknown author of the Poimandres,
53 J. Geffcken, Der Ausgang des griechisch-romischen Heidentums (Darmstadt
1963 = Heidelberg 1929) 77, even goes as far as saying that Alexander was
'gleich sovielen Genossen seiner Sekte voll Interesse fiir die Juden.' For
Julian's treatment of Gen. 6:2 see C. Galil. 290B-D = fr. 67 Masaracchia. Does
Alexander refer to the famous myth about the giants in Gen. 6:1-4 because he
knew that the Manichaean Book of the Giants had been inspired, albeit
indirectly, by that passage? See Reeves, jewish Lore 20-1. Perhaps Reeves somewhat too confidently states that it is clear that 'Alexander was acquainted
with both Jewish and Christian literature' (21).
5 4 G. Rinaldi, Biblia Gentium. Primo contributo per un indice delle citazioni, dei
riferimenti e delle allusioni alla bibbia negli autori pagani, greci e latini, di eta
im~eriale (Roma 1989).
5 For comparison's sake: in Rinaldi's list Exodus has 23 entries, Isaiah
16, Daniel 40 (but all of them from Porphyry!); from the NT Matthew has
most entries: 106 (cf.John 65).
56 The number of references to Gen. 1-11 would even have been much
higher if Rinaldi had included the many allusions to these chapters in the
Hermetic tractate Poimandres, which he did not for unclear reasons stated at
p. 52 n. 27.

PIETER W. VANDER HORST

327

who was a somewhat earlier compatriot of Alexander, also draws

mainly upon the chapters of the 'Urgeschichte,'57 as is also done
by some Gnostic writers (see esp. the Apocryphon ofjohn).
Let us finally turn to an interesting and curious passage where
Alexander, after having argued that moral progress is possible,
says the following: 'This was, I believe, correctly understood by
Jesus, and this is why, in order that farmers, carpenters, masons,
and other skilled workers should not be excluded from the good,
he instituted a common circle of all these people together, and
why, by means of simple and easy conversations, he led them
towards an understanding of God and helped them to achieve a
desire for the good' (16, p. 24 Br.). In the first place, it is clear that
Alexander sees in Jesus here (and elsewhere) 'nullement un
redempteur transcendant, mais plutot un professeur de morale
populaire.'58 Rather, Jesus is here depicted as a kind of latter-day
Socrates who also used to have conversations with 5ru.noupyoi. 59
Secondly, the contents of the list of 'farmers, carpenters, masons,
and other skilled workers,' who formed the circle of Jesus'
disciples, is hard to explain, but it may point in the direction of
which kind of people belonged to the Christian communities in
Alexander's own environment. The fact that the final words of this
passage literally echo the words of ch. 1, where he said that he had
seen with his own eyes that Christian ethical instruction, in spite
of its low theoretical level, nevertheless led people by degrees
towards 'a desire for the good,' seems to indicate that it is again his
own experiences with Christians that makes him enumerate here
the sorts of people Jesus was able to attract to his teachings. 60 Here
we can see a marked difference with his predecessor Celsus, who
had said that 'Jesus collected around him ten or eleven infamous
57 See the extensive discussion of this matter by C. H. Dodd, The Bible and

the Greeks (London 1964 = 1935) 99-248.

58 Villey, Alexandre 279. Cf. also Alexander's remark in ch. 24 (p. 36 Br.):
'To have Christ submit to passion to make something manifest is utter
foolishness, when words are sufficient for the teaching and the knowledge of
what is.'
59 Compare Porphyry's remark that Jesus was one of the wise men of the
Hebrews (from De philosophia ex oraculis haurienda, quoted by Augustine, Civ. Dei
XIX 23). Geffcken, Ausgang 77, reckons Alexander among those who 'Porphyrios' Gerechtigkeit gegen Christus selbst noch weiter entwickelt haben.' On
Por~hyry's praise of Jesus see also Wilken, Christians 148-56, 159-60.
6
It is notable that in the third-century Acts of Thomas 2 the apostle
Thomas is also presented as a carpenter. In the Gospels Jesus is himself the
son of a carpenter.

328

ALEXANDER OF LYCOPOLIS ON CHRISTIANITY

men, the most wicked tax-collectors and sailors, and with these he
fled hither and thither, collecting a means of livelihood in a
disgraceful and importunate way' (CCI 61); and, similarly, in CCII
46: 'When he was alive he won over ten sailors and tax-collectors
of the most abominable character.' 61 And especially in CCIII 59-69
Celsus mocks Jesus' and the Christians' love for 'sinners' and tries
to demonstrate that this leads to the church's being nothing but a
band of criminals. 62 Not so Alexander. Of course, his enumeration
of the professions of Jesus' followers implies that they were lower
class people, who for that very reason needed 'simple and easy
conversations' but, contrary to Celsus, he does believe that Jesus
was succesful in bringing these simple people to a higher level
and a genuine 'desire for the good.'63
Even though the subject under consideration deserves more of an
in depth treatment than was possible in this short paper, we may
briefly summarize our provisional findings. Alexander had some
knowledge of the Bible but it was not extensive (especially when
compared with that of his predecessor Celsus and his contemporary and fellow-Platonist Porphyry, not to speak of the later Julian).
It cannot even be excluded that all the biblical references in his
work are based on hearsay, not on his own reading. The (only?)
biblical books he appears to have had some knowledge of were
Genesis and the Gospels (as usual among pagan authors with some
knowledge of the Bible), and perhaps also of 2 Timothy. But ~his
61 The Gospels mention four fishermen ('sailors') and one tax-collector as
Jesus' disciples. In CCI 62 Origen rejoins that a sailor is not the same as a
fisherman and, moreover, that, contrary to the suggestion by Celsus, of the
majority of the disciples we do not at all know what were the trades by
which they earned their living. Yet the designation 'fishermen' for the
apostles (and sometimes for Christians in general) is found more often; see
J. H. Waszink, Tertulliani De anima (Amsterdam 1947) 119. On Celsus' (and
others') uncertainty about the: number of disciples see W. Bauer, 'Das Apostelbild in der altchristlichen Uberlieferung', in E. Hennecke & W. Schneemelcher (edd.), Neutestamentliche Apokryphen, vol. 2 (Tubingen 1971, 3. Au1.)
11-12.
62 CC III 55 makes clear that it is probably also Celsus' own experiences
with Christians in his environment that colours his depiction of Jesus'
followers. Cf. also Lucian, De morte Peregrini 13. See on this topic further W.
Bauer, Das Leben Jesu im Zeitalter der neutestamentlichen Apokryphen (Tubingen
1909, repr. Darmstadt 1967) 426-7.
63 Unlike Alexander, Celsus heaps scorn upon the person of Jesus; for a
list of references see S. Benko, 'Pagan Criticism of Christianity During the
First Two Centuries A.D.', ANRWII 23, 2 (Berlin 1980) 1102.

PIETER W. VANDER HORST

329

knowledge apparently did not inspire him to vehement attacks on

the Bible, in sharp contrast to Celsus, Porphyry, and Julian. He had
some knowledge of Christian, but possibly also of Jewish, exegetical traditions. Even though he found Christian doctrine inferior
from a philosophical point of view, he was positive about the effects
of Christian ethical preaching and he praised Jesus' efforts to bring
common people to 'a desire for the good. ' 64 This might make us
wonder why he was so circumspect and mild in dealing with
Christianity and its founder. Could it have been at least in part
motivated by the fact that some of his own students belonged to the
Christian church?65
Utrecht

64 'Une telle bienveillance n'etait pas courante', says P. de Labriolle, La

reaction pai"enne. Etude sur la polimique antichretienne du Ier au VIe siecle (Paris
1950, lOth ed.) 317.
65 It should be noted that in the three centuries after Alexander Christian
students in the Alexandrian Platonist school were increasingly a common
phenomenon. On the interactions taking place and the compromises
reached between the Alexandrian form of Neoplatonism and Christianity
see L. G. Westerink, Anonymous Prolegomena to Platonic Philosophy (Amsterdam
1962) X-XXV, and R. Sorabji, 'The Ancient Commentators on Aristotle', in
R. Sorabji (ed.), Aristotle Transformed: The Ancient Commentators and Their Influence
(London 1990) 1-30, esp. 10-15. Note the interesting but bold and hypothetical
statement by I. Hadot in the same volume: 'Alexander, the head of the Platonic school of Lycopolis, refutes the Manicbaeans because the Manichaeans
attended his courses' (287).
I owe thanks to the following colleagues for their useful comments on the
first draft of this paper: Keimpe Algra, Jan den Boeft, Roelof van den Broek,
Peter Lautner, David T. Runia (who also kindly checked my English), and
Leonard V. Rutgers.

PART SIX

DIALECTIC AND DOXOGRAPHY

A 'DIALECTICAL' ARGUMENT IN DE ANIMA A 4

On Aristotle's use of topoi in systematic contexts*
HAN BALTUSSEN

1. Introduction
Aristotle's introductory discussions of earlier views on specific
philosophical issues have been studied from different angles. One
fairly recent line of investigation makes use of his handbook on
dialectic, the Topika.I This has proven profitable in many ways, but
has also created a (minor) problem af terminology. Whenever the
introductions are labelled 'dialectical', the meaning of this term is
usually considered as sufficiently clear. However, it hardly ever
receives an explicit definition.2 It seems to take on at least two basic
meanings: 1. Aristotle engages in a critical debate (ou:xA.eyecr8cxt)
and aims at scrutinizing the views; 3 2. Aristotle makes use of
certain techniques (diairesis, aporia, diaphOnia) indicated as useful in
the Topics, which provide the structure for the selection and
presentation of earlier doxai. These are very general ways of
clarifying his approach and are for the most part based on a few
passages in the Topics (mainly the programmatic books A and 8). 4
My objective in this paper is a modest one: (1) I will explore the
possibility of defining the label 'dialectical' more accurately than
has been customary hitherto, 5 and (2) I will argue that the actual
topoi of the central books of the Topics deserve more attention as an
aid for analysing the doxai-discussions. 6 Among several illustrative

* It is my pleasure to offer this paper to Jaap Mansfeld,

vero magistro.
Fundamental paper by Wei! (1951). See Baltussen (1993) ch. 2 for further
literature.
2 His approach in these passages has also been labelled 'doxographical' or
even 'historical'. For some remarks on these (problematic) terms see section 2.
3 For this function of dialectic see Top. 101b4 E~Eta<rnKit and Rhet. 1354a5
1

~eta~etv Kal. imexetv ')Jyyov.

4 On diairesis see Mansfeld (1990) 3062n; on aporia cf. Top. 145blf., on dia-

phiinia (not an Aristotelian term!) see esp. Mansfeld (1990) 3060, 3063, 3092 ff.
5 Mansion (1961) 41 calls his method in Phys. A 'preparation dialectique'
without further clarification; not so Mansfeld ( 1990) 3063 who speaks of
overviews 'for dialectical purposes' or of 'a dialectical discussion (which may
even be purely didactic) .. .' (ibid. 3064) (cf. n. 4).
6 I prefer the expression 'doxaHliscussion' instead of 'doxography' because

334

HAN BALTUSSEN

examples which might be adduced for this purpose,7 Aristotle's

discussion of the views on the soul is very useful one with some
especially interesting aspects. The general features which make
De anima A 2-4 a good example of a dialectical doxai-discussion (in
the sense described above) are well-known and hardly need extensive discussion here. 8 However, some aspects of a more technical
nature may illustrate that the influence of the dialectical method
goes beyond the obvious fact of classification and selection of doxai.
I will call attention to one specific point which to my knowledge
has received little attention so far. 9 In tracing these details in Aristotle's criticisms I shall argue that our understanding of specific
forms of argument may profit from comparison with the Topics.
Thus I also hope to show how we may further qualify the term
'dialectical', while bearing in mind Aristotle's own conception of
dialectic.IO

2. A note on terminology
Since it may not be apparent why Aristotle's approach towards
previous views may be labelled 'dialectical' (rather than 'historical'
or 'doxographical'), some introductory remarks on terminology
are called for. In our view the alternative terms bring in connotations (and analyses) which are either anachronistic or inappropriate, and may actually be misleading in our understanding of
Aristotle's (and Theophrastus') actual procedure.
In modern studies the term 'doxography' appears to mean 'a
description or summary of what philosophers have said', where
it highlights the role of doxa and at the same time draws a distinction
between doxaHiiscussions and doxaH:ollections (cf. below n. 8).
7 Modern studies using the technical elements are e.g. Schickert (1977) 48
ff. (on the Eudemus/ De an. A), Morsink (1982), Beriger (1989) 81 ff. (on
Metaph. M 1-3).
8 The use of endoxa -the mark of applied dialectic- as views endorsed by
'reputable men' (Top. 100b21-23) is more than apparent: apart from adducing
qualified opinions (see esp. 407b4-5, 13, 27) Aristotle adds a positive formulation of a better perspective which itself is backed up by endoxic views ( 408a34b20). Compare 408a10 where he characterizes the discussion as the testing of
views (E'i>E~Etam:o~).
9 Despite many remarks on dialectical aspects Hicks ( 1907) does not
mention it.
10 Irwin's distinction (1988) 19 ff., 116 between 'strong dialectic' and 'pure
dialectic' seems less appropriate, because it does not take its cue from
Aristotle. It has met with strong criticism, see e.g. D.W. Hamlyn, Philosophy
65 (1990) 465-476, R. Wardy, Phronesis 26 (1991) 86-106.

A 'DIALECTICAL' ARGUMENT IN DE ANIMA A

335

particular stress is on the reporting of the doxai.ll It is not hard to see

that this would be a rather narrow description for Aristotle's
approach (who adds criticisms and also uses generally accepted
opinions). Also, the origin of the term is significant. It is thought
that Hermann Diels introduced the terms doxographia/doxographi to
characterize the form in which post-Peripatetic collections of doxai
were cast, while deriving their contents from the material available in Peripatetic writings. 12 What Diels had in mind is illustrated
by the way in which he edited the Aetius text in his Doxographi
Graeci (DC). Diels's intention has been stated as follows:
Diels himself . . . seems originally not to have used doxog;raphia nor
doxog;raphus as an adjective but only doxog;raphi to denote the authors
of books about the physical and metaphysical opinions of ancient
philosophers, as contrasted to authors of biographies (cf. DG 14546),
[ ... ] The main characteristic of all these doxog;raphi is that they
report the philosophical doctrines according to subject in such a
way that the view of each philosopher . . . is given as a separate
entry under a general 'chapter heading' .13
I note that initially the term referred to a (post-Theophrastean)
'genre' of doxai-collections which became common in the Hellenistic period for several purposes. When and where the term
became used to denote 'any report of philosophical doxai', even in
Aristotle and Theophrastus, is not our concern. It will be enough to
point out here that the expression acquired a wider application
which is not completely free of misleading connotations and
should be used with qualification in the case of the Peripatetics.J4
The question of whether Aristotle's approach can be called
'historical' was already debated a few decades ago. 15 No doubt it is
correct to state that Aristotle had great interest in all kinds of transmitted knowledge. But this hardly justifies that we regard Aristotle
and his followers as historians in the modern sense of the word.
The concept of a 'history of philosophy' is fairly recent (18th century) and any comparison with Aristotle's method from that perspective is bound to be anachronistic. Both subject (theories, not events)
Cf. Passmore (1967) 226, Mejer (1978) 81, Rorty (1984).
If Mejer (1978) whom I follow here is correct. See the next note.
Mejer (1978) 8lf. For doxographi see e.g. Diels (1879) title page, 114, 145,
178, 225. Diels may also have used the term in respons to the question of the
Preisfrage of the Academy in Berlin (DG introd., v) 'inquiratur ... qui auctores
... sint' and by analogy with the Latin 'placitorum scriptores' (ibid. p. iii).
14 My own suggestion (1993, 21) is to speak of critical endoxography.
15 See Stevenson (1974) with further literature.
ll
l2
l3

336

HAN BALTUSSEN

and methodology (justification of sources and the aim of being objective) show that he did not self-consciously aim at such an approach.
It could be argued that Aristotle's investigations into the political
constitutions bring him closer to the historian than anything
else.16 Here as in biology his 'investigations' (ia-ropi<X1.) are of a
descriptive character, providing a first collection of the facts.l7 His
conviction that not knowledge of the particular, but of the universal, was essential must have guided him: by supplying himself
with the necessary material he would then be in a position to
perform the inductive reasoning towards the universal. 1B Clearly,
then, ia-rop{a is a term of wide application. 19 It can hardly support
the view that Aristotle was a historian sensu stricto. 2o Considerations
such as these must (and did) lead to the view that Aristotle did not
write (and did not intend to write) history of philosophy.2 1 The
main reason is that Aristotle is very selective in what he thinks is
useful for his own purposes.22
Since it is clear that the terms just discussed are problematic, I
propose to concentrate on the term 'dialectical' as the one which
offers the least problems. For this we may take a look at the Topics,
where we find a connection between arguments and doxai (e.g.
Top. A 14)23 and a statement which presents Aristotle's different
It is said of the works on politics and laws. See the references in n. 22.
Of course they are more concerned with facts than his doxai-discussions.
On historia see the thorough study by Zoepffel (1975) from which I take over
some conclusions.
18 In this respect there is a parallel with the basic principles of dialectic,
see e.g. Top. 105a12-20; 108b10-12. For a more elaborate exposition on the role
of the universal in linking history and Aristotle's general theory of knowled~e see Zoepffel (1975) 17 ff.
9 Zoepffel (1975) 33.
20 Even the famous passage in Poetics A 9, which brings the term historikos
close to what we would interpret as 'historical', cannot alter this conclusion.
Zoepffel (1975: 14f.) argues persuasively that in this passage, where Aristotle
calls poetry 'more philosophical than history, because it says more on the
universal, whereas history deals with the particular', the emphasis of the
comparative form is on poetry; hence it does not follow that (1) it also goes
with the particular thereby 'saving' a bit of the universal for history nor that
(2) poetry is philosophical in an absolute sense.
21 As Barnes has put it (1987, 25): 'These accounts have a historical purpose
and they are written with a philosophical intention; but they are not, properly sfeaking, 'histories of philosophy".
2
Meier (1930), von Fritz (1958), Weil (1965), Huxley (1973) 282 all use
the term 'historical'. More careful views are found in Reinhardt (1954) 82 ff.,
Braun (1973) 16, 20f.
23 Note that even the Greek commentators made a connection between
certain 'doxographic' discussions in Aristotle and the mandate of Top. A 2
!6

17

A 'DIALECTICAL' ARGUMENT IN DE ANIMA A

337

options for the use of dialectic, viz. ( 1) training, (2) everyday discussions, (3a) philosophy (where it may have several functions),
one of which is (3b) the search for archai in all fields of investigation (Top. A 2.10la25-b4).
Admittedly these are very general descriptions, but the inference seems warranted that modes (2), (3a) and (3b), which represent several stages of increasing technical complexity, build onand therefore reflect-the features of (1). And while (2) may have
left its traces mainly in rhetoric, (3a) and (3b) are very likely to be
found in Aristotle's philosophical research. They are different
modes of the same techne and may be regarded as offshoots of the
same stem.2 4 I would prefer to use the term 'dialectical' with reference to these, i.e. Aristotle's own, options. In my view it is against
this general background that we should try to establish to what
extent certain features of his treatment in the doxai-discussions
warrant use of the term 'dialectical'. I will now proceed to deal
briefly with some examples of how such technical features are
applied in a systematic context.

3. A case in point: De an. A 2-4

The following aspects in De an. A 2-4 are indications for the use of
dialectical tactics.
(I) The combination and confrontation of views exemplifies the
use of well-known and accepted views in philosophical investigation; it creates an implicit scale of agreement25 and disagreement26
on certain questions, which helps to make a decision on the
issue. The cases of agreement (related views) and disagreement
(diaphOnia) resemble collections of views and contrary arguments
from the sophistic period.27 In this way the problematic issues and
the arguments with good (endoxic) support can be found.2 8
(II) The aspect of definition and linguistic analysis is worked
(Simplicius, In Phys. 47.22 Diels; id. In De caelo 523.25-27 Heiberg; Alexander,
In Metaph. 174.3 Hayduck; Asclepius, In Metaph. 140.7 Hayduck).
24 In the same sense as Aristotle himself calls rhetoric a 'kind of offshoot'
of dialectic (Rhet. 1356a25, otov ltapaql'llt~ tt).
25 On ~UO'JliX'tiX, 404a3 Atomists and alB Pythagoreans; on O''tOtXEta, 404bll
Em~edocles and bl7 Plato; on KtVTIO't~ 406bl7 Democritus and b26f. Plato.
2 Democritus vs. Anaxagoras - corporeal vs. incorporeal.
27 Cf. Mansfeld (1986) 17, 46. See also De an. 407b4-5 Ka6a1tEp Etro6e 'tE
AEyEa9at Kill ltOAAot~ O"'lVOOKEt; 407b27 OO~IX ... 1tt61XVft ltOAAot~;
28 No doubt these are endoxa of the sophoi (Top. 100b21-24; cf. 104a33-36).

338

HAN BALTUSSEN

out in different ways. For instance, in dealing with the question

whether 'movement' is a characteristic of the soul (406a4 ff.)
Aristotle examines the meaning of the phrase 'being moved'
( 406a4 otx&c; OE lC\VOUJ.lEvOU 7UXV't0~). Of the two possible meaningsbeing moved 'either in virtue of something else', i.e. indirectly, 'or
in its own right', i.e. directly-the first is clarified by common
usage: 'we say that something is moved indirectly when it is in
something' (Kc:x8' -repov OE AE'YOJ.lEV JC'tA.) and an example from
navigation ('as sailors in a ship') .2 9 It concerns essential and
accidental movement (see below (b)). When this distinction has
been stated (otx&~ oil AYOJ!Evou), Aristotle turns to applying this
distinction to the question at issue.
The step from otxro~ OE lC\VOUJ.lEVOU 7tC:XV'tO~ to otx&~ Ofl AEYOJ.lEVOU
marks the transition from descriptive preparation to (formal)
analysis. Here some crucial distinctions, originating in Aristotle's
physical and dialectical theory, direct the argument and conclusions reached.3 They may be summarised as follows:
(a) there are four basic types (species) of movement to be distinguished (406all-12; analysis according to species-JC<l't' e!()TJ, see
especially Top. B 2.109bl4 and cf. Top. A 15.106a9 ff.; A 1.12lbl ff.);
(b) Aristotle opposes essential to accidental movement ( cf. n. 29), and
(c) natural versus forced movement (division according to opposites,
see Top. A 15.106al0f.; Top. B 8.113bl5 ff.; E 6.135b7 ff.);
(d) observed fact leads to a reasonable assumption (E7tet q>a.{veta.t ... ,
euA.oyov ... 406a30-31) but only to prove the unacceptable consequences of the position (the soul leaving and entering the body,
bl-5); 31
(e) accidental movement is then further explored (b5 ff.).

(Ill) Another example of dialectical tactics is the rather elaborate

critique of Empedocles at 408a19 ff., where Aristotle is in the process of treating the question whether soul is a harmonia (a 'hot topic'
in contemporary discussions). He reproduces (in indirect speech)
29 On the use oh:a8' EtEpov instead oh:ata cru11l3El31'\KO~ see Hicks (1907) 241:
'Ka8' EtEpov ... enables A. to include the case of the passenger, who is, strictly
speaking, neither a part nor an accident of the ship [ ... ] his motion ... is as
much conditioned by that of the ship as if he were the ship's mast or its tonna~e'.
0 Hicks

points to some parallels with the Physica: 2.1, 192b13-15, 22; 8.4,
254bl4.
31 Similar terminological distinctions are at the basis of the discussion of
soul as 'harmonia' at 407b27 ff., 'a further theory which has come down to us,
commending itself to many minds as readily as any that is put forward' (esp.
b31 and 408a5). See also section (III).

A 'DIALECTICAL' ARGUMENT IN DE ANIMA A

339

the line of reasoning behind this view-the body consists of opposites, a harmonia is a blending (Kpiic:nc;) or combining (cruv9Ecrtc;) of
opposites, the soul is a harmonia. In his view (407b32 ff.) harmonia
must either mean a ratio (A.oyoc;) between or a combination (cruv9Ecrtc;) of components, but the soul is (a) neither of these (b33) and
((3) this theory cannot explain the cause of motion.
The emphasis on terminology is important, as Aristotle thinks it
is preferable to apply the term harmonia to health or in general to
bodily excellence ( 408al). In the ensuing paragraph ( 408a5 ff.) he
goes even further and distinguishes two senses of harmonia. Much
is now made of the meaning of 'combination' (cruv9Ecrtc;), positing
that 'there are many combinations of the parts, and they combine
in many ways' (allf.). Aristotle asks for clarification as to how the
intellect combines with a part of the body, what the relation is with
the appetites, and denies that the mixture in flesh is identical to that
in bone. 32
The argument(s) concerning part and whole and concerning
'combination' (cruv9Ecrtc;) in Top. Z (which deals specifically with
definition) contain the 'blue-print' so to speak for the tactics in our
passage in the De an.: regarding composition one should state its
method (150b22-26) and what kind is meant (15la22-25). We may
compare:

De an. 408all-1233
~toUai te yap ai c:ruv6roet~ t&v llEpii'>v

Kat Jtoi.J.. ax&s tivos ouv fi Jtro<:

{lltol..a.Peiv tov vouv xplj miv9eow i\
Kat to aicr9TJttKOV Tt opEKtlKOV;
o1.10ieo<; ~ citOJtov Kat to tov i.Jryov til~
lll~Ero~ etvat tTJV 'lfUXTJV oi> yap tOV
ai>tov Exet Myov i] 1.1i~t~ t&v
<JtOlXetiDV Ka6' TIV crapE, Kat Ka9' TIV
ocrtouv.3 4

Top. 150b22; 15la20-5 (OCT)

en

[sc. emcrKoJteiv] ei 1.1iJ etpTJKE tov

tpOJtov tile; c:ruv9creeo<;. [ ... ]
JtaA.tv ei ti]v toutrov miv9ecrtv EtP'IJKE
tO OAoV, OlOV ti\~ 'lfUXtl~ Kat tOll
crro1.1atoc; miv9ecrtv l;cpov, 1tpii'>tov 1.1v
<JKOltelV i llTJ etp'IJKE ltOta <JUV9<Jl<:,
Ka9a1tp i crapKa opt(OUEVOS fi
O<JtOUV tTJV 1tUpoc; Kat yi\c; Kat aEpO~
Etx cruv9ecrtv. oi> yap ci.JtOXPTI to
<JUV9<JlV iJtlV, aAAaKat ltOta tlS
1tpoo~taptcrtov 35

32 The standard distinction Aristotle makes (PA B 1) between homoiomere

and anhomoiomere, that is between materials of which the smallest part is of
the same nature as a larger quantity and those which break up into different
elements.
33 A good parallel is 408bll: ... ci.A.A.oiromv, ltOta Kat Jtro<:. etepoc; E<J'tl i.Jryo~.
34 'There are many combinations of the parts, and they combine in many
ways. Of what is thought or the sensitive or the appetitive faculty the composition? And what is the composition which constitutes each of them? It is
equally absurd to identify the soul with the ratio of the mixture; for the

340

HAN BALTUSSEN

In both passages the difference between the composition of bone

and flesh on the one hand, and that of the elements on the other is
emphasized.
The passage-aimed against Empedocles as we see in 408a19constitutes, as it were, a miniature case of Ota1topf\oat 1tpo~ UJ.lcpO'tpa. Three questions in the form of a dialectical 1tp6PAT\Jl<X (disjunctive question: 1tO'tpov ... il) are stated ( 408a20-24):
(i) Is the soul a ratio or is it something else, which, being distinct,
comes to be in the parts ?
(ii) Is Love the cause of any and every mixture or of a mixture in
the right proportion?
(iii) If the latter, is Love the proportion itself or something other and
distinct from it?
Such then are the difficulties involved in this view. (tr. Hicks,
modified)
We should note how the difficulties are spelled out, though not all
the options are fully explored. Hicks (ad a24, p. 271) rightly points
out that 'the whole criticism tends to pass into a discussion of antinomies similar to those which fill Metaph. B, where arguments pro
and contra are impartially stated and no decision is given ex cathedra'. Aristotle emphasizes the need for characterization in terms of
identity and difference. He elaborates on one point, confronting it
with observed facts (408a24-28):
(a) if the soul is distinct from the proportion [cf. (iii) above], why is
it destroyed together with the essence of the flesh and that of the

other parts of the living being?

(b) Further, if each of. the parts does not have a soul of its own
-unless the soul is the ratio of the mixture- what is it that perishes
when the soul leaves the body?

The whole passage constitutes a kind of diairesis in which disjunctive options are followed through (thus (iii) follows (ii), (a) is
connected to (iii) and (b) to (i)). The question dealing with the
scheme parts-and-whole takes up one of the initial questions of De
an. A 1 and illustrates how the earlier views sometimes fail to
comply to Aristotelian standards. The most interesting point in
mixture of the elements which makes flesh has a different ratio from that
which makes bone.' (tr. Hicks, modified; italics mine).
35 'see if he has failed to state the manner of combination ... First, see
whether he has omitted to state the kind of composition, as (e.g.) in the
definition of flesh or bone as the combination of fire, earth, and air. For it is
not enough to say it is a composition, but you should also go on to define the
kind of combination'.

A 'DIALECTICAL' ARGUMENT IN DE ANIMA A

341

Aristotle's discussion is that he shows that both the acceptance and

the denial of the theory cause serious difficulties.3 6 Thus the
dialectical exploration of this view has shown it should be dismissed-which is in line with the original method stated at the
outset (403a20-25).37
(IV) One other aspect of Aristotle's criticisms deserves to be
mentioned here. Aristotle's most frequent complaint we read concerns inadequacy of formulation (lack of clarity). Thus he may
point out that more distinctions should be made (A 4, 408b32409a20 Ouxq>pew oe'i etc.); or he may complain about the formulation of earlier views. This point is nearly always expressed in the
phrase ou KaAro<;/op9roc; A.yetv. 38 The same formulation occurs in
programmatic passages such as Top. A 2 and De an. A 2. The turn
of phrase iKavroc;-KaA.roc;-op9roc; A.yetv, 'stating things well etc.' is
very frequent in endoxic contexts. 39 A more programmatic illustration of this attitude is the remark in (again) Top. Z, viz. in the
concluding part of ch. 14 [my italics]:
In dealing with any definition, it is a most important elementary
principle to make a shrewd guess in one's mind at the definition of
the object before one or else to take over some happily expressed
definition. For it necessarily follows that, with a model [Cf. Top. Z 12; e 3], as it were, before one, one can see anything that is lacking
which the definition ought to contain and any needless addition,
and thus be the better provided in the points to attack. 40
We may take this dialectical rule as a fundamental instruction,
since the rules on definition clearly belong to the core of dialectic.41 The way in which Aristotle formulates his criticisms in the
course of investigating earlier views (which in his view are part of
36

37

Hicks (1907) 271; Ross (1961) 197.

Otll7t0poUV't~ ... 'tac; 'tOOV 1tpo'tEprov oo~ac; ... Ei OE 'tl uiJ lCllA.i.ik [sc. tipll~ll], 'tOU't'

euA.aBn8rouev. Cf. on this passage Mansfeld (1990) 3202.

38 Cf. Beriger (1989) 84; Hicks ad Zoe.
39 I.e. contexts in which endoxa are prominent. For (ou) iKavroc; in a systematic context see e.g. De caelo 270b12, 29la32, 308a24; Phys. 191b36; HA 513al4;
EN 1136al0; PoL 1335a41. For (ou) KaA.roc; I (ou) op8roc; in a logical context see e.g.
Cat. 8a29-31; Top. 101bl0-13; Soph. EL 175a20, b13, 183bl5.
40 Top. 151 bl7-23: npoc; &nav't~ o 'tOuc; opl0' 11ouc; oulC EMxlO''tov cr'totxlov 'to np(Jc;
EllU'tOV EU<J'tOJ((I)c; opicracr8at 'tO 7tp01CElllEVOV T\ lCCXAroc Eipnuevov opov c:XvcxA.a~E1v.
c:Xva'Y1C1] yap rocrnep npoc; ncxpaouyllcx Serollevov 'to 't' EA.Aeinov J>v npocri\Kev exuv 'tov
bptcrllov Kcxi 'to 7tpoo1CdllVov neptepyroc; JCa8op&.v COO'tE !lft~J,.ov rntJCEtpll!l<i'trov eunopeiv
4! I think this 'rule' goes beyond the level of dialectical exercises. The
reason is that in Top. A. Aristotle states that his main tools for dialectical
analysis are the four predicables (genus, differentia, accidens, definition),
which 'in a way all are concerned with definition' ('tp07tOV 'tlVcl Opt1Ca, 102b34f.).

342

HAN BALTUSSEN

'apparent facts', ta phainomena) and solving the difficulties shows

how the discussion moves quickly from the abstract (or logical)
level to the concrete level and back again, that is to say, between
theoretical considerations and existing views. Such an approach he
no doubt regarded as legitimate, given the nature of dialectic as 'a
technique independent of any specific field of investigation' (Rhet.
1358al 0-35).
The importance of the function of dialectic is further clarified in
the same passage (Rhet. 1358a21-26), when Aristotle connects it to
dialectical moves geared to find starting-points ( archai) within a
philosophical inquiry(= mode 3b mentioned above):
these [general commonplaces] have no special subject-matter, and
will therefore not increase our knowledge of any particular class of
things (genos). On the other hand, the better the selection one
makes of propositions suitable for special commonplaces the nearer
one comes, unconsciously, to setting up a science that is distinct
from dialectic and rhetoric. One may succeed in stating the
required principles, but one's science will be no longer dialectic or
rhetoric, but the science to which the principles thus discovered
belong. (Rev. Oxf. Tr.)

Here Aristotle indicates how the formal moves of rhetoric and

dialectic give a science or field of investigation direction and substance: these can contribute to the attainment of starting-points, thus
setting up a science in a well-founded manner.
4. Conclusion

Our starting-point was the query concerning the use of certain

(modern) labels for Aristotle's discussions of earlier views. Among
the three most frequently used-' historical', doxographical', 'dialectical'-the term 'dialectical' was chosen as the one causing the
least problems. Because the term still needed further qualification,
we opted for an approach in which Aristotle's own position was
studied as found in his work on dialectic, the Topics.
We saw that Aristotle indicates four areas in which dialectic
may be used (Top. 10la25-b4) 42 and suggested that the basic level of
training (YUIJ.VO.ato.) is traceable in all three other areas mentioned. 43 With this in mind it was urged that we exercise caution
in using the term 'dialectical'.
For a new interpretation of this passage see Smith (1993).
Viz. (2) everyday encounters (cf. Rhet. 1354a4f), (3a) philosophy and (3b)
search for archai in all the sciences.
42
43

A 'DIALECTICAL' ARGUMENT IN DE ANIMA A

343

As one possible route towards clarifying the label we have

briefly analysed some passages in De an. A with a view to the statements in the Top. which give dialectic a role in such contexts. This
revealed, among other things, an interesting parallel between the
'rules' from a techne (Top.) and a treatise of systematic nature (De
an.). Not only does it provide counter-evidence to the view that
dialectic was superseded by the Analytics,44 it also gives the label
'dialectical' a fuller meaning by taking more aspects of the dialectical method into account.
I end with a few concluding remarks. The Topics may still be
regarded as a valuable tool for analysis when it comes to interpreting the nature of Aristotle's arguments. The example given above
may seem somewhat tenuous on its own, but it does not stand
alone.45 This is not to say that Aristotle was slavishly following his
own codifications of the handbook, but merely expresses the
conviction that the topoi give us a clue to many of his lines of
reasoning, in particular when it comes to critical evaluations of his
predecessors. My argument has been that we should use the term
'dialectical' with some care, especially in making clear which of
its different aspects is meant. In other words, the more general
features of presentation (the use of endoxa, the opposing of views)
should not be given a too dominant role in defining what dialectic
is at the cost of more technical ones (formal arguments for criticism, technical terms). The first category is important in making
a philosophical point by way of enhancing contrast, the second
helps to establish to what extent certain arguments really represent
the critic's own views or are merely 'logical'. Both aspects of the
method can be found in Aristotle and both deserve attention in
their own right.*
Utrecht

44 On this point see also Baltussen (1993) 16-18.

45 Cf. above n. 7 (not all on doxai-discussions).

I am grateful to K.A. Algra, D.T. Runia and T.L. Tieleman for some
very useful remarks on an earlier version of this paper.

344

HAN BALTUSSEN

BIBLIOGRAPHY

Baltussen, H. (1993) Theophrastus on Theories of Perception. The Argument and

Purpose in the De sensibus (diss. Utrecht).
Barnes, J. (1987) Early Greek Philosophy (Penguin Classics).
Beriger, A. (1989) Die aristotelische Dialektik. Ihre Darstellung in der 'Topik' und
den 'Sophistischen Wiederlegungen' und ihre Anwendung in 'Metaphysik' M /-3.
(Heidelberg).
Braun, L. ( 1973) L 'histoire de l'histoire de la philosophie (Paris).
Fritz, K. von (1958) 'Die Bedeutung des Aristoteles fiir die Geschichtsschreibung', Entretiens Fondation Hardt, Tome iv (Vandroeuvres-Geneve) 83-128.
Hicks, R.D. (1907) Aristotle. De anima (with tr. intr. notes) (Oxford, repr.
Amsterdam, 1965).
Huxley, G. (1973) 'Aristotle as an Antiquary', GRBS 14: 271-286.
Irwin, T.H. (1988) Aristotle's First Principles (Oxford).
Mansfeld, J. (1986) 'Aristotle, Plato, and the Pre-Platonic Doxography and
Chronography' in Cambiano (ed.), Storiografia e Dossografia nella Filosofia
Antica (Torino) 1-59.
- - , ( 1990) 'Doxography and Dialectic. The Sitz im Leben of the 'Placita",
ANRWII 36.4 (1990) 3056-3229.
Mansion, S. (1961) 'Le role de l'exposee et de Ia critique des philosophies
anterieures chez Aristote', in S. Mansion (ed.), Aristote et les problimes de
la methode (Louvain/ Paris)
Meier, M. (1930) 'Aristoteles als Historiker', in: Philosophia Perennis /, Festschrift Josef Geyser (Regensburg) 39-54.
Mejer, J. ( 1978) Diogenes Laertius and his Hellenistic Background (Wiesbaden).
Momigliano, A. (1978) 'Greek Historiography', History and Theory 17: 1-28.
Morsink,J. (1982) 'The Mandate of Topics !.2', Apeiron 16: 102-128.
Passmore, J. ( 1967) 'Historiography of philosophy', in Encyclopedia of Philosophy, ed. P. Edwards (New York-London).
Reinhardt, K. (1954) 'Philosophy and History among the Greeks', Greece &
Rome 1: 82-90.
Rorty, R. ( 1984) 'The Historiography of Philosophy: four genres', in R. Rorty,
J.B. Schneewind & Q. Skinner ( eds), Philosophy in History. Essays on the
Historiography of Philosophy (Cambridge-London-New York) 49-75.
Ross, W. D. (1961) Aristotle. De anima (Oxford).
Schickert, K. ( 1977) Die Form der Wiederlegung beim f'rUhen Aristoteles (Miinchen).
Sichirollo ( 1966) Dialegesthai-Dialektik, von Homer bis Aristoteles (Hildesheim).
Smith, R. (1993) 'Aristotle on the uses of dialectic', Synthese 96: 335-358.
Stevenson, J.G. (1974) 'Aristotle as Historian of Philosophy', Journal of Hellenic
Studies 94: 138-143.
Weil, E. (1951) 'La place de Ia logique dans Ia pense aristotelienne', Revue de
Metaphysique et de Morale 56: 283-315 [= 'The Place of Logic in Aristotle's
Thought', in: J. Barnes e.a. (edd), Articles on Aristotle, 1. Science (London
1975) 88-112]
Weil, R. (1965) 'Philosophie et histoire. La vision de l'histoire chez Aristote',
Entretiens Fondation Hardt, Tome xi (Vandroeuvres-Geneve) 161-189.
Zoepffel, R. ( 1975) Historia und Geschichte bei Aristoteles, Abh. Heidelberger
Akad. der Wiss., Phil.-hist. Kl., Abh.2.

THEOPHRASTUS' DE SENSIBUS ON PLATO

Theophrastus' De sensibus is a unique source of information concerning earlier Greek philosophers' opinions on sense perception
and perceptible properties.2 If his work had not survived, we would
lack details pertinent to understanding Parmenides' obscure lines
on human thought (fr. 16 DK), and we would know very little
I offer this paper as a sixtieth birthday present to my friend Jaap Mansfeld. As the world's greatest expert on the doxography of Greek philosophy,
he will be the first to spot its shortcomings and the first to build on anything
of value it contains.
Much of my material was presented to the Theophrastus conference, which
took place on Lesbos in August 1989. I am grateful for the comments from participants that I received on this occasion, and also to Tom Rosenmeyer, who
helped me to improve this final version. Mter I had submitted it, the editors
recalled my attention to H. Baltussen's chapter, 'Theophrastus' De sensibus and
Plato's Timaeus', which forms part of his doctoral dissertation, Theophrastus on
Theories of Perception. Argument and Purpose in the De sensibus, written under the
direction of Jaap Mansfeld and Hans Gottschalk, and published as vol. VI of
Quaestiones injinitae. Publications of the Department of Philosophy, Utrecht University
(Utrecht 1993). Although there is some overlap in the material we discuss
and in our respective findings, our approaches are so different that I find it
best to let my original text stand for the most part. My purpose is largely
limited to comparing Theophrastus' reports on Plato, as clearly as possible,
with the relevant passages of the Timaeus. Baltussen's study offers much more
in the way of attempts to explain and partly justify Theophrastus' procedures.
2 Assessments of Theophrastus' accuracy in the De sensibus vary considerably. Citing Stratton (n. 3 below), Charles Kahn wrote: 'The performance of
Theophrastus here [i.e. his treatment of Plato], in the sole case where we can
fully control his use of source material, is fine enough to justify a modern
editor's praise of the "high accuracy" of his "dispassionate and marvellously
impartial report", Anaximanderand the Origins of Greek Cosmology (New York 1960)
21. Yet, Statton himself also admitted that Theophrastus, in his criticism of
Plato, 'seems oftener to miss the point' (p. 53). For a very different assessment,
cf. J. McDiarmid, who, in his article 'Theophrastus on the Presocratic Causes',
HSCPh 61 (1953) 133, wrote: 'The fragments considered disclose no evidence
that Theophrastus employed his knowledge of the Presocratics in such a way
as to exercise independent judgment about them ... He has frequently misrepresented his source and has exaggerated the faults present in it. It must be
concluded that, with regard to the Presocratic causes at least, he is a thoroughly biased witness and is even less trustworthy than Aristotle'. To place this
comment in perspective, note J. Mansfeld's comment in his Studies in the Historiography of Greek Philosophy (Assen-Maastricht 1990) 24: 'There is no doxography in the proper sense of the word in Aristotle or Plato, or even in Theophrastus, because the doxai at issue are presented from a systematical point of
view in order to further the discussion of problems of a systematical nature'.

346

A.A. LONG

about the relevant doctrines of Alcmaeon, Anaxagoras and Democritus. Theophrastus' lengthy treatment of Empedocles includes
many points that could not be inferred from the surviving fragments or from other secondary sources. In one instance only,
Plato, we are able to compare Theophrastus' reports with his own
source. Although he does not name the Timaeus, all his information about Plato in the De sensibus, as G.M. Stratton observed, seems
'to be drawn exclusively from' that dialogue.3 Indeed, according to
the same scholar, when he deals in the last part of the work (Sens.
83-91) with Plato's treatment ofperceptibles (aicrOrrra), Theophrastus preserves Plato's order of exposition 'without a single change'.
'This', he continues, 'with the various verbal similarities, makes
one almost see Theophrastus at work with the Timaeus spread
before him'. 4 My purpose in this paper is very simple-to exhibit
what happens when we study Theophrastus' treatment of Plato in
the De sensibus on the basis of Stratton's observation. 5
At the beginning of the work Theophrastus divides the majority
of opinions concerning ai"cr~crt~ into two groups: those that explain
it 'by likeness', and those that explain it 'by the opposite'. As proponents of the first explanation he cites Parmenides, Empedocles and
Plato. In aligning Plato with Empedocles, Theophrastus follows
Aristotle (De an. I 2, 404bl6) whose justification for the claim is
tendentious if not impenetrable. 6 After elaborating on the two types
of explanation, Theophrastus makes the following observation:
As for each particular sense, practically all of them are neglectful,
but Empedocles does try to refer them too to likeness (Sens. 3).
Theophrastus returns to Plato, after a page on Parmenides, with
these words:
3 G.M. Stratton, Theophrastus and the Greek physiological psychology before
Aristotle (London-New York 1917) 159. Stratton's text, which I draw on here,
is substantially that of Diels in his Doxographi Graeci. A new Bude text is bring
prepared by Andre Laks and Glenn Most.
4 Ibid. 203.
5 Baltussen (n. 1 above), p. 129, says: 'We should certainly not imagine
Theophrastus at work' in the way Stratton intimated. I demur. For even if, as
is probable, Theophrastus is writing on the basis of excerpts of the Timaeus,
Baltussen himself assumes (p. 127) that 'Theophrastus has consulted the (complete) Tim. text', and acknowledges the verbatim character of some of his
reports.
6 Aristotle ad Zoe. refers to 'Plato in the Timaeus'. R.D. Hicks, in his great
commentary on the De anima, comments: 'The reference is to the 'lfUXoyovia
in Timaeus 34C sqq.', and 'that like is known by like is the assumption underlying the language of Tim. 47A-C'. I confess to doubt about both proposals.

THEOPHRASTUS' DE SENSIBUS ON PLATO

347

Plato touched on the particular senses to a greater extent, but he did

not in fact speak about all of them, but only about hearing and sight
(Sens. 5).

Theophrastus then gives reports about Plato's account of these two

senses, and concludes with this remark:
Concerning smell, taste and touch he said nothing at all, nor about
whether there are other senses beyond these, but he gives a more
careful treatment of the perceptibles ( Sens. 6).
This is an extraordinary statement because Plato's discussion of
tactile, gustatory and odoriferous properties (Tim. 6le-62c, 65b-67a),
which Theophrastus knows very well (see below), includes an
account of touch, taste and smell. Why, then, has he suppressed
this? (I say 'suppressed' rather than 'ignored' because much later,
when Theophrastus charges Democritus and Plato with contradicting themselves in their accounts of the perceptibles (Sens. 6061), he virtually retracts the statement, claiming that 'Plato refers
practically all the perceptibles to their [perceptible] effects (1ta811)
and to sense perception'!) The answer seems to be that he suppresses Plato's account of touch, taste and smell at this point in his
argument because what Plato says about these perceptual processes
does not fit the explanation in terms of 'likeness'. According to
Plato, these perceptions result from contact with elementary particles that lack the properties we ourselves perceive. In the case of
sight and probably hearing too, Theophrastus leads his readers to
think that Plato's words justify his being regarded as a 'likeness'
theorist. Thus he says that Plato 'makes sight of fire, which is why
he also makes colour a kind of flame coming from bodies .. .'
( Sens. 5). The 'likeness' explanation is also implicit, I think, in his
characterization of Plato's account of hearing, when, going beyond
Plato's text, he says: 'He defines hearing through sound'. The idea
seems to be that, like Empedocles (ibid. 9), Plato took hearing itself
to be a sound within the percipient. In regard to seeing colours,
Theophrastus explicitly and very misleadingly aligns Plato with
Empedocles (Sens. 91). 7 He follows Aristotle in treating Empedocles
as the paradigm for discussing Plato.

7 Cf. Stratton op. cit. (n. 3) 220: 'Theophrastus here does great violence to
Plato's doctrine'. See also D.N. Sedley, 'Empedocles' theory of vision', in
W.W. Fortenbaugh & D. Gutas, Theophrastus. His Psychological, Doxographical and
Scientific Writings (New Brunswick-London 1992) 26-31.

348

A.A. LONG

Theophrastus' repeated claim in De sensibus 5-6 that Plato omitted

any treatment of touch, taste and smell influenced one branch of
the doxographical tradition. Pseudo-Plutarch and Stobaeus confine
their remarks on Plato's perceptual theories to sight and hearing,
omitting any reference to him in their doxographies of smell and
taste. 8 By contrast, Alcinous' summary of Plato in his Didaskalikos
includes a close and lucid paraphrase of most of the relevant parts
of the Timaeus, with no reference to 'likeness' as Plato's generic
mode of explanation nor any hint of the omissions alleged by
Theophrastus. 9 From Alcinous one has the impression that Plato
wrote about the five senses in this order-sight, hearing, smell,
taste and touch.
Before pursuing Theophrastus' treatment of Plato in detail, it
needs to be emphasized that Plato's organization of the relevant
material in the Timaeus is far from straightforward. In summary, it
goes thus:
44d-45b
45b-46c
46e-47c
47c-d
53c-61c
61c-64a

64a-65b
65b-66c
66d-67a
67a-c
67c-68d

construction of the head

construction of the eye and explanation of sight
benefits of sight
benefits of sound and hearing ...
the four regular solids, the four corresponding elements and their species and compounds
tactile properties and their effect on the percipient's body,
in the order: hot cold, hard soft, heavy light, rough
smooth.
relation of perceptions to pleasure and pain
gustatory properties and the effects of elementary
particles on the tongue
smells and the 'power' of the nostrils
'the third aio6rrm:6v (i.e. hearing) and sounds
colours

8 Cf. H. Diels, Doxographi Graeci (Berlin 1879) 4.13.11, Aetius = Ps. Plut. 4.13
901C, Stob. Eel. 1.52.7. Both texts use the expression nA.a.trovtK'iJ auva.uyna. in
summarising Plato's theory of sight.
9 Cf. Alcinous, The Handbook of Platonism, ed. J. Dillon (Oxford 1993), chs.
18-19. Alcinous' reports of the Timaeus are much clearer and more accurate
than those of Theophrastus, especially in the case of sight, though he too
omits Plato's description of transmission back to the eye and the soul. Yet, as
Dillon notes ( 144-5), Alcinous seems to have been influenced directly or
indirectly by the De sensibus. This appears in his treatment of flavours, and
also in the Aristotelianism of his allusion to 'natural place' in his account of
Plato's theory of weight.

THEOPHRASTUS' DE SENSIBUS ON PLATO

349

As this summary shows, Plato gives his account of sight many

pages before he deals with the other senses and their objects
including colour. He returns to sense perception only after he has
elucidated the four kinds of elementary particles and their combinations, and explains his reasons thus:
The various shapes have now been adequately explained with their
figures, combinations and mutual transformations. Now we must
try to clarify the causes that account for their effects (mx.OftJ.Uita).
Our statements must presuppose the existence of sense perception
even though we have not yet expounded the generation of the flesh
and its properties or of the mortal part of the soul. It is impossible to
give an adequate account of these things without the perceptible
effects (7taOftJ.ux.ta aicr9-r]nKa) and vice versa, and impossible to treat
them both together. So we must first hypothesize the one or the
other, and then return later to what we have hypothesized. In order
to speak of the properties corresponding to the kinds (of shapes), let
us first assume the facts to do with body and soul (Tim 6lc-d).
Plato's procedure is complicated, but scarcely confusing. Because
sense perception involves interaction between the elementary particles and the soul-body complex, it requires reference to both of
these. Having not yet completed his account of the latter, Plato
assumes certain facts about it, in order to focus initially on the
perceptible effects. Thus, it appears, what Plato will chiefly explain,
following this methodological passage, is the perceptibles (aicrST\ta
in Theophrastus' terminology) and not the accompanying psychophysical processes.
Theophrastus may be registering this when he says: 'He gives a
more careful account of the perceptibles' (Sens. 6). Reading on (or
remembering), he finds Plato discussing the tactile properties but
without any explicit word on touch (Tim. 6le-63e). Then, after
treating of pleasure and pain, Plato remarks:
This is a more or less complete statement about the (perceptible)
effects common to the body as a whole and of the names given to
the things that produce them. Now we must try, if we can, to speak
about what happens in our specific parts, both the (perceptible)
effects and their causes (Tim. 65b).
Plato's first sentence refers back to his treatment of tactile properties
and touch which, for obvious reasons, he assigns to no 'specific'
bodily part. He then discusses flavours and taste, odours and smell,
sound and hearing, which he calls 'the third perceptible part in us'
(67a, i.e. specific part), and 'our fourth and remaining perceptual
faculty' (67c)-the faculty for perceiving colours.

350

A.A. LONG

Theophrastus had no difficulty in extracting Plato's account of

hearing from this passage, and he incorporated some of Plato's
remarks on colours in what he says about sight. Given Plato's signposting, especially his distinction between touch and 'the specific
parts' and his description of hearing as 'the third perceptible part in
us', Theophrastus' claim that Plato said nothing about touch, taste
and smell is startlingly false. It is also absurd as polemic, because,
as I pointed out, Theophrastus contradicts the claim by his
subsequent comments and reports.
Plato on sight

The crucial features of Plato's account in Tim. 451xl are as follows:

(1) Day-light consists of non-burning fire.
(2) There is a kindred fire within us.
(3) The out-flow of our own fire-light through the eyes in day
time fuses, like by like (oJ.lotov 1tpoc; OJ.lOtoV), with the external light,
and forms a homogeneous visual ray extending from the eye to
the object.
( 4) The interaction of this visual ray with an object results in
movements that are transmitted back into the whole body right
up to the soul, and thereby produce sight.
When Plato later deals with colours (Tim. 67c-68d), he makes these
further points:
(5) Colour is a flame issuing from bodies and consisting of
particles.
(6) The visual ray itself consists of particles.
(7) The flame particles, which are so proportioned to our sight
as to yield perception (O'lfEt O"UJ.lJ.lE'tpa J.lOpta ... 7tpoc; atcr6rtcrtv),
depending on their size and interaction with the visual ray
generate the spectrum of perceived colours.
(8) The interaction of the two sets of particles is transmitted back
to the eye.
What does Theophrastus do with all this? His account runs: 10
Sight he makes of fire. (That is why he also makes colour a kind of
flame coming from bodies, having its parts commensurate with
sight-on the ground that effiuence occurs and they must fit into
10

7) 30.

I adopt Sedley's translation and punctuation of Theophrastus, art. cit. (n.

THEOPHRASTUS' DE SENSIBUS ON PLATO

351

each other.) It passes out of the eye for a distance and coalesces with
the effluence, and that is how we see (Sens. 5).
Theophrastus draws this extremely condensed report from both
contexts of the Timaeus. He begins by alluding to Plato's first point
in (3), but he omits the fact that the eye's fire is light and also that
what it coalesces with is day-light. 11 Plato adduces the principle of
'like by like' to explain this coalescence. Theophrastus, however,
implies that Plato uses the principle for a different purpose-to
explain the relation of sight to colour. He takes Plato's reference to
the commensurability of colour and visual ray particles to involve
an explanation of sight by 'likeness', conflating symmetry of
particles to likeness of substance.
If this is regarded as a relatively minor distortion of Plato's
theory, such charity cannot extend to the rest of Theophrastus'
account. Without Plato's text we would have the impression that he
regarded sight as simply the effect of external interaction between
the visual ray and a colour effluence, without reference to transmission back to the eye and the soul. That impression is hardly
corrected when Theophrastus adds:
One could classify Plato's opinion as intermediate between those
who claim that sight falls upon its object and those who say it is a
case of transmission from visual objects to sight (Sens. 5).
For Plato the coalescence of colour and visual ray particles begins
outside the eye, but seeing occurs as a result of the way that
coalescence changes the visual ray and so affects the eye and
eventually the soul. By omitting all reference to Plato's points ( 4)
and (8), Theophrastus credits him with a misleadingly crude
theory of seeing as the contact between two fires outside the eye.
Plato on sound and hearing

Plato's account of hearing and sound (Tim. 67b) is very brief.

Following his practice of treating sense perception separately from
perceptibles, Theophrastus divides his treatment of Plato's unitary
passage between the two parts of De sensibus. In regard to the first
part he cites Plato almost verbatim:
II Stratton op. cit. (n. 3) 161 comments: 'If Theophrastus had borne in
mind Republic 507C-E, we should not have had his astonishing neglect of
light, which is so important in Plato's theory of vision'. But Plato's emphasis
on light is totally obvious from the Timaeus alone.

352

A.A. LONG

Plato
0~ IJ.EV o.Ov <pCOVTJV ero!J.EV 'tlJV Ot, cittrov
im' aepo~ EYKE<paA.ou 'tE Klli lltiJ.Il'tO~
IJ.EXPt wuxfi~ 7tAt]'Y'lV OtllOtOOIJ.EVllV, 'tlJV
OE im' ll'lhfi~ KlVllOtV, a7to 'tfi~ KE<paA.fi~
~v apx;OIJ.EvllV, 'tEAE'Il'tiDallV OE 7tEpi 'tlJV
'tOU ll1tll't0~ eOpllV, aKO~V.

Theophrastus (Sens. 6)
aKolJV OE ou'x 'tfj~ <provfi~ opt~E'tllt . <pCOVTJV
yap etvat 7tA11YlJV im' aepo~ EyKE<paAo'll
Klli lltiJ.Il'tO~ Ot' cO'tCOV IJ.EXPt wuxfi~. 'tlJV
0' ll1t0 'tllU'tll~ KlVllOtV a1t0 KE<pllAfj~
~x;pt ll1tll't0~ aK~V.

By starting his account with the statement, 'Plato defines hearing through sound', Theophrastus gives the erroneous impression
that Plato takes hearing (the effect of the motion caused by sound
on the percipient's organs and soul) to be itself a sound. Thus the
unwary reader will take Plato's account of hearing to be an
instance of Theophrastus' general characterization of those who
explain sense perception 'by effluence and the transmission of like
to like' (Sens. 1). Apart from this, his report is impeccable and
almost verbatim. Here, in contrast with what he says about Plato's
treatment of sight, Theophrastus makes it clear that the sensory
stimulus has to reach the soul in order to be registered as such.
Why his treatment of the Timaeus should be so tendentious in the
one case and so accurate in the other is a question that is easier to
ask than to answer. For the present, I put it on one side, and
continue the comparison.
Plato's treatment of sound and hearing continues (Tim. 67b) with
a brief statement about how sounds vary in pitch, smoothness and
volume respectively according to the speed, regularity and size of
the motions that produce them. Much later (Tim. 80a-b) he gives a
fuller explanation of harmony and dissonance. Theophrastus
draws briefly on both contexts at Sens. 85, where he is dealing with
perceptibles. First, substituting o for yap, he repeats word for word
the sentence q>rovitv ... J.lEXPt 'lfUXll~ from Sens. 6. Next, he reports
Plato's explanation of differences of pitch, but omits Plato's further
reference to smooth, harsh, loud and soft sounds. Thus he says
nothing about regularity and size as further differentiae of soundproducing movements. He concludes by saying:
Sounds are in concord when the beginning of the slow sound is like
the end of the swift one.
This is a cryptic paraphrase of part of Plato's very difficult account,
which runs:

THEOPHRASTUS' DE SENSIBUS ON PLATO

353

As to the swift and slow sounds that appear high and low, they are
dissonant when the motion they produce in us is dissimilar, and
concordant when that motion is similar. For the slower sounds catch up
with the motions of the earlier and swifter ones, which are already
fading and have reached a similar state to the motion that the
slower sounds, as they approach later, generate in them; when
these slower sounds encounter the others, they do not cause discord
by introducing a different movement but the beginning of a slower
movement which accords with the faster one that is ending, and by contributing similarity they cause a unitary experience in which high and
low are blended.
If Theophrastus had to restrict his paraphrase of this to thirteen
words, he could scarcely have done better. Even so, his words are
so condensed that they do little more than record his favourite
preoccupation in De sensibus--'likeness' as an explanatory factor in
sense perception. In particular, we are left without any understanding of how the two movements come to be 'like' one another,
which is Plato's main point.
Plato on tactile properties (Tim. 61e-64a)

Hot and cold: Theophrastus, Sens. 83:

Plato says that is hot which causes division by the sharpness of its
angles.
This is careless and defective. Plato makes 'sharp' the generic
property of heat, taking this to be observable from heat's divisive
and cutting effects on the body. But he advances the sharpness of
the angles of fire as only one of four imperceptible properties that
account for heat's 'cutting' sensation; the others that he cites are the
thinness of its edges, the smallness of its particles and their speed
of motion (Tim 6le).I2 He does not, as Theophrastus later claims
(87) define heat simply 'by shape' (O"X~J.w:n).
As regards cold, Theophrastus gives a fairly accurate paraphrase
of Plato's account of its production from large liquid particles
compressing small ones (Tim. 62a-b). We then get a revealing
instance of accurate quotation combined with error:

12 Stratton op. cit. (n.3) 204 mistakenly refers to 'the three factors to which
Plato attributes the cutting power of heat', omitting the 'thinness' of the
particles' edges.

354

A.A. LONG

Plato

Theophrastus (Sens. 83)

1:ft 01t ~uxxn Kal. 1:ip cretcrf.!ip 1:ou1:cp 'tpOf.!oc; Kal.

iiiyoc; e1:e~, IJf'llXp6v 1:e 1:0 1t1xOoc; iinav 1:ou1:o
lC<Xt 'tO op&v E<JXEV OVOJ.l<X.

'tft yap J.lcXXU 'tpof.!OV Kat 'tip naOet

p'iyoc; etV<Xl OVOJ.l<X.

Theophrastus sticks very closely to Plato's language, but he

misrepresents Plato in saying that the feeling of cold is called
ptyoc;. Plato reserves this name together with 'tpOf.!oc; for 'shivering'
-'the struggle between the particles'; his name for the agent and
feeling of cold is 'lfUXPOV. More important, Theophrastus fails to
remark that the outcome of the struggle between the particles is
'immobility and congealing' (cruvroatc;). Thus Plato treats the effects
of cold as directly opposite to the cutting caused by heat. If
Theophrastus had attended to this point, he would have had little
grounds for his objecting (Sens. 87) about a lack of congruence
between Plato's treatment of hot as distinct from cold.
Hard and soft: Theophrastus (Sens. 83) reports Plato, Tim 62b6-8
almost verbatim. Next, he cites Plato's explanation of 'the small base'
characteristic of things that yield, but completely omits what
follows-Plato's explanation of non-yielding things (Tim. 62b8-c3).

Heavy and light: Next, Theophrastus (Sens. ibid.) correctly states

Plato's view (Tim. 62c) that there is no absolute 'up' or 'down'. He
continues:
Light is what is easily drawn into the place contrary to its nature,
and heavy what is so moved with difficulty.

This second statement (which completes Theophrastus' account) is

very misleading for two reasons. First, it gives the false impression
that Plato takes heavy and light, in contrast with up and down, to be
objective properties. Secondly, it suggests that, according to Plato,
light and heavy have no natural connexion with up and down.
Here is the most relevant part of Plato's account:
What is light or heavy or below or above in one region will be
found to become or to be the opposite of what has all these characteristics in the opposite region ... Concerning all of these, this one point
has to be grasped-that the progression of each thing in the direction of what is kindred to it makes the moving thing heavy and that
the region to which it moves is below, and vice versa (Tim. 63d-e).

In his subsequent criticism of Plato's doctrine (Sens. 88-9), Theophrastus acknowledges Plato's refusal to define heavy and light

THEOPHRASTUS' DE SENSIBUS ON PLATO

355

a1tAro~, but he adds, very misleadingly, that 'he defines them 'with
respect to what is made of earth'. In fact, Plato no more defines
terrestrial weight than weight in any other region. He takes the
earth as a region by means of which he can explain the principle
of relative weight (Tim. 63c-d). Theophrastus' critique of Plato,
which 'is only repeating Aristotle', as A.E. Taylor says in his commentary on the Timaeus ad loc., gives his readers no opportunity to
grasp the radical divergence between Plato's theory of relative
weight and the Aristotelian theory of objective weight or lightness
as a function of the elements' movement to their natural places.
Rough and smooth: Of these tactile properties Theophrastus says
(Sens. 83):

He passes over them, on the ground that they are sufficiently clear,
and says nothing.

The first part of this sentence is broadly true, the second totally
false. Plato writes (Tim. 63e):
Everyone, I think, who has seen the cause of the sensation of
smooth and rough could explain it to someone else; for the latter is a
combination of hardness and unevenness, and the former is
brought about by a combination of evenness and density.

Plato on pleasure and pain (fim. 64c-d)

Plato
'tO oi] til~ i]oovf!~ IC(Xt l..{l1t1J~ 6>oE OEi OtaV0Etcr8m to 11EV Ttapa qn)crw Kcxi. ~icxwv
yryVOJlEVOV a8p6ov 1tClp' iJJltV 1ta8o~ af..yEtVOV, 'tOo' Ei~ q>Ucrtv amov 1tUAt~ a8p6ov
i]ou, 'tO OE TJPEI!Cl ICClt ICCX'ta CfJlliCpOV avcxicr'tOV, to o JlE't dmetdcx~ ytyv6 11evov aTtcxv
cxicrSTJtov I!EV on 11af..tcrtcx, f..{m1J~ o Kcxi.
i]oovf!~ ou 11Etxov, otov ta 1tepi. ti\v
ovv cxutitv Ttcx8iJ11cxtcx...tcxutn yap to11cxi.
11EV KClt KClU<JEt~ KClt ocrcx iiUcx Ttacrxn
f..u1tCll; ouK E111towucrw, ouo ftoova~ 1taf..w
e1ti. tcxutov aTttoucr1J~ eioo~ .. .~(~ yap to
1tUI11tClV OUK EVt tfl OtClKptcrEt tE cxutil~
KClt <ro)'KptcrEt.

Theophrastus (Sens. 84)

i]ou ot: Kcx\ A.UltTJpov, to JlEv de; cpucrtv
a8p6ov Tta8oc;, tO OE 1tClpa q>Ucrtv KClt ~(~
AU1t1JpOV, ta OE JlECf(X KClt avcxicrSTJtcx ava
f..6yov. 0\0 KClt KClta tO opiXv OUK eivcxt
AU1t1JV ouo' i]oovitv tfl OlClKpicrEt KClt
crU)'KpicrEt.

Theophrastus' report is a clear and accurate paraphrase of Plato

down to the words ~{If AU7tTJp6v. At the corresponding point, Plato
says:

356

A.A. LONG

A gentle and slight process is imperceptible (sc. with respect to

pleasure and pain).
Theophrastus renders this point too, and then writes:
Therefore, in seeing there is no pleasure or pain from the cutting
and combining.
This sentence too is accurate, but scarcely intelligible without allusion to Plato's full treatment of the point. For what he undertakes to
say is that processes such as those involved in sight are supremely
perceptible, but the disturbance to the visual ray that they involve
(cutting etc.) is too slight to cause pleasure or pain.
Plato on flavours (Tim. 65b-66c)

Theophrastus' report (Sens. 84) begins thus:

As for flavours, Plato mentions four species of water in his treatment of water; the juices include wine, verjuice, oil and honey,
and the sensations include the earthy flavour.
Theophrastus' basis for his comment on the juices is the context of
the Timaeus where Plato discusses types of water (58d-60b). At the
end of this passage Plato lists the four types of juice as instances of
'composite water'. Then, at the beginnng of his treatment of taste
and flavours, he says (65c):
First we must explain to the best of our ability what we omitted in
our earlier account of juices-the sensations peculiar to the tongue.
The unwary reader of Theophrastus will form the impression that
Plato's account of the four juices contributes to his explanation and
classification of flavours, and Theophrastus actually says this later
( Sens. 89) in his criticism, but it does not. Plato makes no use of the
types of juices in his treatment of flavours, and when discussing
the juices in the earlier passage the only gustatory sensation he
mentions is the sweetness produced by honey (60b).
As for what Theophrastus calls 'the earthy flavour' ('tov yeroo11
XUf.lOV), this too has no authority in Plato's text. Plato does not cite
earth as a flavour but as a substance whose particles, depending
on their consistency, generate 'astringent (cr'tpucpva) or 'harsh'
(a:u<r'tllpa) tastes' as a result of their interaction with the blood
vessels extending from the tongue to the heart ' (Tim. 65c-d). The
corresponding texts continue thus:

THEOPHRASTUS' DE SENSIBUS ON PLATO

357

Theophrastus (Sens. 84)

Plato
... y{]tv<X ~p111C<X't<Xt111COJ.1EV<X {fi)Vcl'YEt ta q>Af3ta
Kat ci1to~1Jpa\vet, tpax{mpa ~ ovta atpu<pva,
~nov o tpaxuvovta au<rtTJpa <paivemt
ta o tout(l)v te pu1tnKa !Cal. 1t&v to 1tepl.
tlJV yA.&tt<XV cX1t01tAUvovta, 1tEp<X flEV tou
f.I.Etp{ou tOlltO Oprovt<X K<Xt 1tpO<JE1ttA<XJlf3avOflEV<X ro<JtE cXltOtTtKEtv autil~ til~ q>U<JE(l)~
otov fJ t&v A.hpffiv MvaJlt~. lttKpa 1tciv6' out(!)~
wvoJlaatat, ta o imooeeatepa til~ A.ttpwoou~ E~E(l)~ Eltt to f.I.Etpt6v tE tfl pu1j1Et
XPIDflEV<X W..uJCa iiveu lttK~to~ tpaxda~ ...

K<Xt ota tauta auvciyovta tou~

K<Xt <JUyKptVOV't<X, ta flEV
tpaxutepa atpu<pva dvat, ta o
~ttov auatTJpci. to o pu1tnKov t&v
1topffiv Kat ci1toKa9apttKov aAJlupov
to o a~p6opa pu1tnK6v. mate Kal.
EJC'titKEtV, 1tt1Cpov.
XUAOU~

As this comparison shows, Theophrastus gives an accurate paraphrase of Plato's treatment of astringent, harsh, bitter, and salty, but
he reverses Plato's order of presenting the latter pair of flavours and
omits soda, Plato's illustrative example for explaining the difference between them.
Next he reduces Plato's six-line explanation of 'pungent' (Optj.!Ea)
to a single line (ibid.):
Things that are warm and move upward and have a cutting effect
are pungent ( tn 8epJ.latVOJ.l.EV!l K!lt &vw cpepOJ.l.EV!l K!lt Ot!lKptVOVt!l
optJJ.ia).

oe

This telegraphic style is quite inadequate to convey Plato's account:

Things that share in the mouth's warmth and are softened by it,
become inflamed and heat in their turn the thing that heated them,
and rising because of their lightness to the senses of the head, cut

everything on which they impinge (Tim. 65e).

This is a brilliant description of the sensations concomitant on eating spicy food. Theophrastus, having falsely declared that Plato has
nothing to say on taste, makes Plato's report virtually unintelligible.
Next, Theophrastus writes (ibid.):
Things that have a seething effect are sharp (tn

oe K\)K&vta o~w).

In his lengthy account (Tim. 66a-b) Plato uses the word 'seething' to
describe the general effect that (probably) acidic substances have
on the blood vessels of the tongue, an effect he actually calls
'bubbles, boiling and fermentation'. He does not characterise the
gustatory sensation itself as 'sharp', but says 'the cause of these
sensations is called sharp' .13
13

Cf. Aristotle's list of flavours at De sensu 442al9, which corresponds

358

A.A. LONG

Finally, sweet. Theophrastus (ibid.):

Things that in company with the moisture in the tongue cause it to
expand or contract into its nature are sweet.
Again his compression of the original is extreme and partly
opaque. Plato's account goes:
When the composition of the things entering (the mouth) in liquid
form, being naturally akin to the state of the tongue, smoothes and
strokes its roughened parts, and contracts or expands any unnatural
contraction or expansion, and makes everything accord with
nature as far as possible, and any such remedy for the enforced
sensations is pleasant and agreeable, it is called sweet. ( Tim 66c).
Half ofTheophrastus' report is taken up with the obscure expression
cruv 'tU uypO'tT]'tt ~ EV 'tU yA.c.Onn, yet this fails to bring out Plato's point
that the causes of sweet tastes are themselves liquids.

Plato on smells (Tim. 66d-67a)

Plato begins his account of smells by denying that these can be
classified according to species (a point he repeats later), and that
any of the four primary bodies is perceptible to smell. Smells, he
says, occur when air is changing into water, or vice versa. The rest
of his account (with the exception of a final sentence about the
difference between pleasant and unpleasant smells) is taken up in
detail by Theophrastus:

Plato
Eiaiv tE o<rJlal aUJllta<rat JCaJtvoc; ft OJlll(All,
to{mov o to JlEV ~ ciEpoc; de; uorop iov
OJlll(All, too ~ UOatoc; de; &epa JCaJtv6c;
o8Ev AElttOtEpat JlEV uOatoc;, Jtal(UtEpat
of: O<rJlal <rUJllta<ral yEyovacnv &poe;.
~l..ouvtat of: oJt6tav nvoc; avtuppax8vtoc;
ltEpt tTtV avaltVOTJV UYI1 nc; ~i~ tO ltVEUJla
de; aut6v. tOtE yap OO"Jllt JlEV OUOEJlla auvOt118Ettat, to o JtvEuJla t&v oaJl&v EP1lllro8f:v auto JlOVOV EltEtal. ou' o;)v taUta
avroVUJla ta tOUt<oV ltOl!ClAJlata yeyovEV,
OUJC EJC Jtol..l..&v OUOE b.Jtl..&v do&v ovta,
&I..J..Cx oxn t6
iJou JCal. to l..u1t11Pov
aut68t JlOV<o Ota<pavft 1..yEa8ov ...

Theophrastus ( Sens. 85)

tac; o oa11ac; E'Lo11 11v ouJC exEv, &J..J..Cx
tip AU1tl1Pifl Kat i)OEt Ota<pEpEtv. dvat
of: tTtV OO"JllJV uOatoc; JlEV AElttOtEpov'
&poe; of: ltal(UtEpov. O"llJlElOV of: Otl
otav Eltl<ppa~avtoc; ava<rltiDO"lV, UVEU
oaJlftc; to ltvEuJla dapxEtcxt oto
JCa8aJtep KaJtvov Kat OJltXI..llv dvm
'tiDV <r<oJlnt<oV aopatOV. etVal OE
JCaltVOV JlEV JlE'ta~ol..itv E~ uOatoc; de;
&epa, OJlll(AllV of: tlJV E~ &poe; de;
uorop.

exactly to Plato's terms: yl..u!CU, ai..Jlup6v, 1tl1Cp6v, aU<rtllPOV, OplflU, <rtpu<pv6v, o~u.

THEOPHRASTUS' DE SENSIBUS ON PLATO

359

In his treatment of this passage, Theophrastus offers an impeccable

account of Plato's text, accurate, clear and so full that he even
includes the proof that the smell of an object vanishes when
someone inhales through an obstruction. He also adds a detail
Plato omits-the invisibility of the vapour that causes smell. Why
is he so careful and patient here and so hasty and careless elsewhere? The question cannot be answered, but before assessing his
overall treatment, let us conclude the details with his account of
Plato on colours.
Plato on colours (Tim. 67c-69a)
Theophrastus begins his treatment of Plato's fourth 'perceptible' that
affects the body's particular parts by repeating (Sens. 86) almost
verbatim his accurate citation (ibid. 5) of Plato's account of colour,
which he derived from this passage (Tim. 67c6). Omitting Plato's
remarks on the different sizes of colour particles and on
transparency (ibid. 67d3-6), he continues:
What penetrates sight is white, and what compresses it is black,
drawing almost word for word on Tim. 67e5-6. He also alludes with
reasonable clarity to the analogy Plato draws between the different
particles responsible for the perceptions of black/white, hot/cold
and astringent/pungent respectively (Tim. 67d7-e2).
Next, he gives his own truncated description of what Plato calls
'brilliant' (AaJ.Utp6v, Tim. 68b5), 'fiery white' ('to 7tUproOE<; AEUKov),
after which the transmittc:;d text contines: 'and the rest of the
colours from these'. Because Plato's account of the rest requires the
prior specification of red as the fourth primary colour, we should
assume, with Stratton, 14 a short lacuna in which Theophrastus
mentioned this. He presumably omitted Plato's explanations of
orange, purple, and the other derived colours.
So far, so fairly good. Theophrastus concludes his brief report on
colours thus (Sens. 86):
But in what proportions [sc.the rest of the colours are blended], he
says one should not state even if one knew them, since we do not
have a likely or a necessary account of them. Nor is it at all surprising if on experiment the outcome is different, but god has the power
to do this.
14

op.

cit. (n. 3) 212.

360

A.A. LONG

This is a curiously synthetic report of two independent observations by Plato. 15 At Tim. 68b6-8, Plato writes:
Bright mixed with red and white becomes orange. It would make
no sense to state the proportions, even if one knew them, given that
no one could be even moderately capable of stating either their
necessity or the likely account.
Plato then (ibid. 68b8-d2) gives an account of many other colours
that are generated by mixtures, and concludes:
These examples make it more or less clear to what mixtures the
other colours must be assimilated in order to preserve the likely
account. But if anyone investigating should apply a test to the facts,
he would prove his ignorance of the difference between divine and
human nature; for god has the knowledge and the power to blend
the one into many and to resolve the many into one, but no human
being exists now or ever will exist who can do either of these things
(Tim. 68d2-7).
Theophrastus in his second sentence (cited above) alludes to Plato's
warning against experimentation and to the superiority of divine
power, but his brevity leaves his readers guessing as to what 'this'
is that god can do by way of frustrating the temerarious experimenter.
Conclusions
I now summarise the results of this survey. Theophrastus' reports
on Plato's sensory doctrines in the Timaeus range from being
almost verbatim and entirely clear at one extreme to misrepresentation and opacity at the other.
Accurate and clear. hearing, smell, the tactile properties hard and
soft, pleasure and pain, the gustatory properties astringent,
harsh, bitter, salty, the colours black and white
Venial omissions: explanations of four of the types of sound and of
rough and smooth
Damaging omissions: sight
Extreme condensation: sight, concord in sound, no pleasure or pain
in seeing, the gustatory properties pungent, sweet and sharp, his
concluding remarks about Plato on colours

15 Cf. Baltussen (n. 1), pp. 118-19, who comments cogently on other aspects
of Theophrastus' method of excerpting here.

THEOPHRASTUS' DE SENSIBUS ON PLATO

361

Serious misrepresentation:
1. (general) the attribution to Plato of 'likeness' as the explanatory principle, the denial that Plato explains the operations of
touch, taste and smell;
2. (particular) the explanations of sight, hot and cold, heavy and
light, the first part of flavours

In regard to the general misrepresentations, something has

already been said. Theophrastus followed Aristotle in classifying
theories of perception in terms of explanation by 'likeness' or its
opposite. This damages his treatment of Plato's explanation of sight,
but it has little effect on the accuracy of his other reports. The
complexity of Plato's arrangement of topics in the Timaeus may
help to explain Theophrastus' extraordinary claim that Plato has
nothing to say about touch, taste and smell, but since his later
reports show otherwise the claim seems to be largely gratuitous,
motivated by unjustified preconceptions that he subsequently
drops. It is hard to see why, even allowing for his polemical leanings, he failed to state explicitly that Plato's account of these and all
other perceptual processes except sight is premissed on Plato's
theory of the four primary bodies and their particles.
As regards his detailed misrepresentations, the two worst are his
treatment of Plato's accounts of sight, and heavy and light. With
regard to sight, Theophrastus, by his own words, assimilated Plato
to Empedocles (Sens. 91). On heavy and light, his Aristotelianism
probably helped him to misread Plato's radical account of relative
weight. His introduction to Plato's account of flavours is decidedly
skewed, but the text at this point is sufficiently difficult to warrant
suspicion of corruption and lacunae. The misrepresentations in his
account of hot and cold may be due to inadequate note-taking, lapse
of memory, desire for brevity, and a combination of these.
Why he should be so full and accurate in some instances and so
compressed and inaccurate in others is a question to which I have
no satisfactory answer to offer. We need, of course, to take account
of his dialectical methodology, his Aristotelianism, the restricted
focus of his interest in the doctrines he reports.l 6 A case can be
16 For good remarks on all this, cf. H. Baltussen, 'Peripatetic dialectic in
the De sensibus', in F orten baugh & Gu tas op. cit. ( n. 6) 1-19. Baltussen' s approach to understanding the basis and methodology of Theophrastus' critical
comments is much more sophisticated than that of Stratton and other earlier
scholars. Yet he is too kind to his author when he describes Theophrastus'

362

A.A. LONG

made, as Han Baltussen shows (n. 1 above, p. 129), for explaining

Theophrastus' omissions as due to a deliberate choice of 'only
clear-cut statements pertaining to physiology'. Baltussen quite
rightly insists too that we not saddle Theophrastus with our own
standards for accurate treatment of an author's text. Yet, whatever
we say in explanation of Theophrastus' methodology in the De
sensibus, the validity of his reports as an account of Plato's doctrines
remains an issue. If these were consistently inaccurate or tendentious, that would be telling, but they are not. Theophrastus is
simply very uneven in his level of accuracy and clarity. I see no
reason to presume that his starting-point was an epitome of the
Timaeus as distinct from a text virtually identical to our own. He
may at times be relying on memory or on (by our standards)
inadequate excerpts, whether made by himself or by a pupil. But
apart from the cases of sight and weight, I do not think that
extraneous factors, his Aristotelianism or his explanatory purposes
go far to account for the unevenness of his accuracy, clarity and
comprehensiveness.
We do not need Theophrastus for studying Plato. We do need
him for access to Plato's predecessors. If my findings about his
treatment of Plato are extrapolated, they suggest that in every
instance we need to be on our guard in the following ways.
First, his general claims about what a philosopher said or did not
say may be inaccurate either by omission, inappropriate addition,
or by failing to fit the data precisely.
Second, his criticisms of a philosopher should be cited as evidence only in cases where independent corroboration is available.
Third, it should never be assumed without further evidence that
any report Theophrastus gives is comprehensive, completely accurate in formulation or adequate in clarity. Much that he says will
be correct, but no sound criterion exists within his text for determining exactly where he can be trusted.
Berkeley

reports as 'brief, yet very informative' (9) and characterizes his criticism of
Democritus and Plato as 'precise and thorough' (11). This apologetic tendency
is my only objection to his excellent study of Theophrastus' treatment of the
Timaeus in his dissertation (n. I above.)

ADDITIONAL FRAGMENTS
OF ARIUS DIDYMUS ON PHYSICS
DAVID

T.

RUNIA

For more than a century a minor but not insignificant figure in

the study of ancient philosophy has been the doxographer Arius
Didymus. Ever since Meineke and Diels presented their hypothesis,1 it has been customary to attribute the following documents,
all located in Eusebius and Stobaeus, to this obscure personage. 2
1. Eusebius PE 11.23.2-6 fragment on Plato's ideas taken EKtmv
AtOUJlq> llEpt trov apE<JKovtmv llMitmvt auv'tE't!l'YJlEVmv (a very similar text is found in
Alcinous, Didaskalikos 12.1);
2.
PE 15.15
Stoic fragments on cosmo-theology a1to tile;
'E7tttoJ.Lflc; 'ApEiou dtOUJlOU;
3.
PE 15.18-20 Stoic fragments on cosmology and psychology a1to trov 'E7tt'tOJ.100V 'ApEiou AtOUJlOU;
4. Stobaeus Eel. 1
large number of fragments on physics pertaining to Aristotle and the Stoics, but without
any reference to their source; also at 1.12.2a
the same fragment on Plato's ideas that is
found in Eusebius;
5.
Eel. 2.1.17
epistemological fragment beginning with
Xenophanes, entitled AtOUflOU EK tou llEpt
aipecrEmv;
6.
Eel. 2.7.1-4
untitled introductory discussion on ethics,
containing OtatpEaEtc; by Philo of Larissa and
Eudorus and with copious references to
philosophers, esp. Plato and Aristotle ( =
ethical doxography A) 3
7.
Eel. 2.7.5-12 synopsis of Stoic ethics entitled Zflvmvoc; Kat
trov A.omrov ~tmtKOOV 00yJ.1!l't!l7tEpt tou i]0tKOU
1 A. Meineke, 'Zu Stobaeus', Berliner Zeitschrift fur das Gymnasialwesen [ =
Miltzells Zeitschrift] 13 (1859) 563-565; H. Diels, Doxographi Graeci (Berlin 1879,
19764 ) 69-88 (henceforth DC).
2 References to Mras' edition (Berlin 1956) for Eusebius' Praeparatio Evangelica, Wachsmuth-Hense for Stobaeus (1884-1912). I omit two references to a

Didymus in Clement of Alexandria which are of lesser importance and

difficult to integrate in our picture of Arius Didymus. There is also a reference to Didymus' account of Aristotelian doctrine in Priscianus Lydus Sol. ad
Chosr. 42.39-40 Bywater, but no fragmentary material is furnished.
3 The useful labels A, B and C for the three ethical doxographies were introduced by D. E. Hahm, 'The Ethical Doxography of Arius Didymus', ANRW
2.36.4 (Berlin-New York 1990) 2935-3055; it is fast becoming standard.

364

DAVID T. RUNIA

~pouc; -rile;

qnl..ocso<piac; (=Ethical dox. B)

8.

Eel. 2. 7.13-26 synopsis of Peripatetic ethics entitled 'Aptcs-

9.

Flor. 4.39.28

-ro-r/..ouc; Kat 'tiDV Aoutrov 11Ept7tU't'll'ttKIDV 1tEpt

-rrov iJ8tKIDV (= Ethical dox. C)
excerpt on EUOUtJ.LOVia entitled EK -rile; AtOUJlOU
'E7tt'tOJlllc; identical to EeL 2.7.17 (and so furnishing the basis of Meineke's source
theory).

A further aspect of the Meineke-Diels hypothesis that has long

won acceptance was the identification of this Arius Didymus with
the Stoic philosopher and confidant of Augustus, Arius of Alexandria (c. 70 to c. 5 BCE).4 During the last 15 years there has been a
gradual realization that the hypothesis has its shaky aspects, but no
direct challenge was mounted.5 This has now come in the form of
a fresh and incisively written monograph by Tryggve Goransson,
who argues that the identification with the Stoic Arius is unsound,
because the direct references to the doxographer listed above always
use the identificatory name Didymus, but this name is never used
for Arius the court philosopher. 6 We thus lose our chronological
anchor for the doxographer and can only locate him at some time
between the mid first cent. BCE and Eusebius.
The Swedish scholar's point is well taken, even if it may not be
quite as strong as he thinks it is.7 Our concern in this article,
however, is only with a limited section of the uncontested and
contested remains of Arius Didymus' muvre, namely the physical
fragments listed under 3. above. In his Doxographi Graeci Diels identified a number of such fragments and edited them in an edition
4

On this intriguing figure see G. W. Bowersock, Augustus and the Greek

World (Oxford 1965) 33-35.

5 See the evaluatory articles by C. H. Kahn, 'Arius as a Doxographer', in

W. W. Fortenbaugh (ed.), On Stoic and Peripatetic Ethics: the Work of Arius

Didymus (New Brunswick 1983) 3-13; Hahm, art. cit. (n. 3). Both concentrate

primarily on the ethical doxographies.

6 T. Goransson, Albinus, Alcinous, Arius Didymus, Studia Graeca et Latina
Gothoburgensia 61 (Goteborg 1995), esp. 182-226.
7 Effectively what he has done is forcefully remind us of the hypothetical
character of the identification of the doxographer with the Stoic court philosopher. It cannot be taken as a fact that the two are the same person. The problem of the names was already recognized by Diels and Zeller. But this
century scholars simply assumed the identification. For a spirited defence of
the hypothesis see B. Inwood's review of Goransson's book, Bryn Mawr Classical Review 95.12.8 (electronic document). A further important result of his research is that he casts serious doubt on the assumption shared by all scholars
since Diels that Ethical doxography A is by the same author as B and C (who,
at least in the case of C, must be Arius Didymus).

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

365

that has been the point of reference ever since.B Paul Moraux, to
take a prominent example, presents a lengthy analysis of the
Aristotelian physical fragments purely on the basis of Diels'
identifications, the methodology of which is not even discussed. 9
Goransson states somewhat apodictically that the criteria used by
Diels to sift out the Didymean material 'are perhaps not as indisputable as they have been regarded every since',l 0 but makes no
attempt to embark on this investigation himself.
My intention is to reexamine the question of the separation of
the Didymean material on physics in Stobaeus' Eclogai afresh. I
will argue that Diels' criteria need to be refined and more consistently applied. This will lead to the identification of a number of
additional fragments, but also to the removal of a few snippets from
Diels' collection. The investigation may be regarded as a preliminary study for a new edition of these physical fragments, which remains a real desideratum. It is a joy to be able to dedicate this contribution to my mentor, colleague and friend Jaap Mansfeld on his
sixtieth birthday. We have discussed these and related issues on
numerous occasions during our joint research on ancient doxography, and have not always been able to reach complete unanimity. So I am sure that he will scrutinize these pages with an
even sharper eye than usual.
The disputed fragments we are concerned with are all located
in Book I of Stobaeus' Eclogae. This book, which unfortunately only
survives in a truncated version,ll contains a vast amount of doxographical material in the area of physics, but gives no indication
whatsoever as to where it was drawn from. As we all know, Diels,
basing his theory on earlier research but also adding new elements of his own, argued that most of this material was derived
from a compendium entitled Tiept apE<H.:OV't(I)V (De placitis) also
8 DG 449-472. As far as I know, apart from Wachsmuth, on whom more
below, no attention has been given to the question of the identification of
physical fragments of Arius Didymus in Stobaeus, except incidentally, for
example in my own articles cited below in n. 35 & 39.
9 P. Moraux, Der Aristotelismus bei den Griechen von Andronikos von Rhodos his
Alexander von Aphrodisias, Band I: Die Renaissance des Aristotelismus im I. ]h. v.
Chr. (Berlin-New York 1973) 276-305.
10 op. cit. 220. No doubt he is especially thinking of Diels' criterion of a
certain Stoicizing flavour (DG 75); see further below n. 19.
II The prologue is missing; chapters 1-31 are fairly complete, but chapters
32-60 have been considerably abridged by an epitomator, who wrote out only
the Platonic and Aristotelian material.

366

DAVID T. RUNIA

exploited by Theodoret of Cyrrhus, who appears to refer to its

author as the otherwise totally obscure Aetius. 12 An important task
that Diels set himself in his Doxographi Graeci was to separate this
Aetian material from a second important source of material drawn
from the doxographical work of Arius Didymus. The results of
Diels' analysis were incorporated by Wachsmuth in the very
structure of his edition,l3 which therefore cannot be regarded as ari
independent piece of work. But we must give the latter scholar
credit for subjecting the details of the theory to a thorough
examination, which led to a number of minor improvements.
In the light of the sceptical remarks about the Meineke-Diels
hypothesis outlined above, it is worth emphasizing that there can
be no doubt whatsoever that Stobaeus did make use of a work of
Arius Didymus.I 4 One of his excerpts shows a strong verbal resemblance to an Didymean fragment preserved by Eusebius (= fr. 36 in
Diels' collection). The texts are best placed side by side: 15
Eusebius PE 15.18.3 Mras
apEO"KEt yap tOt<; ~tOltKOt<; cptAoO"Ocpot<; tl]v OA:rw oucriav de; 7tUp J.l.Et<l~aAAEtV, oiov de; cr7tEpJ.La, Kat 7taA.tv EK tOUtOU auti]v cl7tOtEAEtcr8at
tl]v OtaKOO"J.l.TJO"tV, o'ia to 7tp6tepov
~v. Kat tolho to OOyJ..La trov a1to tTl<;
aipecrero<; oi 7tp&tot Kat 7tpEcr~uta
tot 7tpocr{JKavto, Z{jvrov tE Kat KA.eav8T]c; K<lt XpUO"t7t1tO<;. tov J..LEV yap
toutou J.La8'1]ti]v Kat Otaooxov tf\<;
crxoA.f\<; Z{Jvrova cpacrtv E7ttcrxE'iv
7tEpt tf\<; EK7tUpcOO"EOl<; trov oA.rov.

Stobaeus Eel. 1.20.1e Wachsmuth

Z{jvrovt Kat KA.Eav8et Kat Xpucrt7t7tcp apEO"KEt tl]v oucriav J.l.Et<l~aAAEtV oiov de; cr7tEpJ.La (de;) to
m>p' K<lt 1tUAtV EK tOUtOU tOtaUtT]V
a7toteA.E'icr8at tl]v OtaKOO"J.l.TJO"tV,
ota7tp6tepov ~v.

llavaitto<; 7tt8avrotepav dvat

voJ.Liset Kat J.LiiA.A.ov &pecrKoucrav
autip tl]v ai:OtOt'l]t<l tOU KOO"J.l.OU ~
tl]v t&v oA.rov de; 1tup J.l.Eta~oMv.

12 For Aetius, Diels' theory and the contributions of his predecessors I

refer the reader to volume 1 of the study being prepared by J. Mansfeld and
myself entitled Aiitiana: the Method and Intellectual Context of a Doxographer. We
show that Diets' theory, though deficient in some details, can withstand the
attack launched against it by A. Lebedev.
13 Berlin 1884, and since then never superseded.
14 Cf. Goransson op. cit. 219-220; note, however, the slight inconsistency
between the statements that Stobaeus may have quoted 'from a parallel doxographical tradition' and that 'he demonstrably excerpted Didymus at least
once'.
15 I have included Heeren's conjecture in the Stobaean text. Pace Diels, it
seems to me justifiable in the light of the repeated Ei<; 1tup in the last line.

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

367

Eusebius cites verbatim, as is his wont, and cites his source (at
15.20.8), whereas Stobaeus leaves his source unnamed and takes
the kind of liberties that one might expect from an anthologist. The
names of the three Stoics are brought to the fore in order to make
the contrast with the other doxai clearer. Only the change from
au-ri'tv to 'totaU'tllV might be thought worrying. It could easily be
interpreted as a Verschlimmbesserung, i.e. Stobaeus cannot imagine
the Stoics maintaining a cycle of identical worlds. But we must be
wary of hyperinterpretation (and how can we be sure that -rotaU'tllV
is not a textual corruption?).
We know, therefore, that Stobaeus (henceforth S) made use of
the compendia of both Aetius (henceforth A) and Arius Didymus
(henceforth AD) in his compilation. Most, though not all, of his
doxographical information on physics can be reduced to these two
sources. 16 Decisive criteria here are comparison with other sources
(notably Ps.Plutarch's abridgement of A, henceforth P 17 ) and considerations of style and content. The macro-structure of the book is
loosely based on A, but the anthologist has seen fit to introduce all
manner of structural changes in his material, involving the
opposed techniques of coalescence and separation, so that the
process of disentangling his original sources has become an immensely complex and often wearisome task. It is thus of vital
importance to determine various criteria that allow the two chief
sources to be separated. In our view the most significant of these are
eight in number. Obviously our list takes as its point of departure
the list of ten criteria which Diels outlined in his rigorous and admirably succint analysis. 18 It would seem better, however, to draw
up a new and revised list, in which his results are incorporated
rather than repeat his list more scholastico and comment on each of
his criteria individually. Unlike Diels we shall exclude any
considerations that are based wholly on content. 19 It needs to be
emphasized that these criteria are generalizations, and that for
16 These matters can in the present context not be discussed in detail. The
reader is referred to the study announced in n. 12.
17 In what follows references are to the Teubner edition of J. Mau, Plutarchi Maralia vol. V fasc. 2 pars 1 (Leipzig 1971).
18
19

DG73-75.

I.e. Diels' last three criteria. Diels argues that AD uses Stoic terminology in describing Aristotelian doctrine, but he may well be influenced by
his conviction that AD is the Stoic Arius. The criterion of a direct use of
Aristotle's Metearologica can only be verified as part of a thorough examination
of the meteorology of A's book III.

368

DAVID T. RUNIA

each one there are exceptions possible. Ideally every lemma found
in our sources should be individually analysed, a tedious and
time-consuming task. In his DC Diels did little more than present
the results of his analyses, so often the reasons for his decisions
need to be divined.
Here, then, is the list of eight criteria for separating A and AD.
1. The lemmata of AD are often longer than those of A, not only
because they enter into more detail, but also because they often
combine more than one topic. For example some of the longest
lemmata in A are found on the subject 11Ept 'tU~Eroc; 'tOU KOCJIJ.O'l> (cf.
P 2.7), e.g. Parmenides at S 1.22.la, Philolaus at ld. But they are not
even half the length of Chrysippus' long expose at S 1.21.5 (=AD fr.
31 Diels) on the nature and structure of the cosmos. Because AD
treats a number of topics together in a continuous expose, his
fragments often have a more fluent and discursive style than the
compact and sometimes crabbed style of A. There are, however,
two further complicating factors. Some of the fragments attributed
by Diels to AD are so short that the above remarks can hardly be
applied. Moreover we have to take into account that S in his
coalescing of various chapters in A also groups together subjects
that A kept apart. This means that the combinations of AD have to
be distinguished from the coalescences of S. As we shall see, this
proves to be a highly tricky business.
2. The standard formula of A's placita is to have the doxa immediately follow the name-label, whereby the verb of assertion is generally understood (occasionally cpTJcr{, EcpTJ or <htEcpf)va'tO is included).
The topic, which is always given in the chapter-title, is sometimes
repeated in the first lemma and thereafter it too is generally understood. Sometimes, however, it is even omitted in the opening
lemma. Such procedures come much less naturally to AD because
he, as far as we can tell, did not have chapter titles2 and also
makes less use of name-labels. Among the fragments of AD in S
four different types of lemmata can be observed: (a) those which
follow the usual style of the Placita, i.e. have the name-label in the
nominative at the beginning; (b) those which have the name-label
only in the genitive, without a direct grammatical relation to the
contents of the lemmata; (c) those which have the name of the
20 The title llEp\. 'tEAoU<; at 2.45.11 may not be original, but added either by S
or a scribe. See also above n. 7 on doubts as to whether Ethical doxography A
in Scan be ascribed to AD.

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

369

philosopher in the nominative but not at the beginning; (d) those

which have the name-label in another case, e.g. accusative in
oratio obliqua or dative as in the text at 1.20.1e quoted above. In the
case of type (a) discrimination is difficult, but remodelling by S
can in some cases be demonstrated, particularly by means of his
use of characteristic introductory phraseology (e.g. verbs of saying
or opining) Type (c) and (d) unambiguously reveal AD because
such practices are never found in A.2 1 Type (b) was regarded by
Diels as furnishing a water-tight criterion, but the evidence in the
mss. is less unambiguous than he thought.22
3. When S arranges the excerpts he has collected, he often has to
add introductory phrases of his own in order to make necessary
connections. For example in 115 Tiept OXllJ. uhcov he wants to quote
Tim. 33b on the sphericity of the cosmos. So he adds the words
(145.9-11): TIA.a:tcov Eq>llOE crcpa.tpOEtOii "COV lCOOJ.lOV U1t<XPXE1V. A.eyeqap
o'ihco~ .v 't4'> TtJ.la.tcp. We may be fairly certain that there was no
such doxa in A.23 In a number of cases (including this example, as
will become clear below), S uses introductory verbs not found in P,
a fact which makes his intervention very probable.
4. In the case of AD, because S is taking excerpts from a continuous
expose, they frequently contain o as connecting particle. In the
ethical doxai there are very few lemmata without some kind of
connective.2 4 In A, on the other hand, the particle is usually only
used directly after the name-label if he wants to make a deliberate
contrast (see exx. at P 1.22, 1.24, 2.3, in all three cases the 2nd
lemma). A third possibility, suggested by the text from 115 just
cited, is that S feels no need for a connecting particle when he
himself inserts introductory connecting phrases.
5. Because the two authors report the views of others, they both
make extensive use of indirect speech. But they reveal opposite
2! If so, there are grounds for suspicion. The text at P 5.20 ad init. is almost
certainly corrupt. The lemma at S 1.50.3 which starts Ka.ta toil~ lleputa.tllttKou~
is attributed to A by Diels DG 394, but may well derive from AD.
22 On this problem see 0. Hense, Art. 'Stobaios', RE 9 (1916) 2565, who
makes clear that this criterion is vulnerable to the interventions of scribes
and editors, and so needs to be handled with care.
23 Cf. P 2.2 886B and S 1.15.6b. If the doxa was in A, then both sources
must have changed the label from Plato and the Stoics to the Stoics only. This
is less likely than that S added the doxa from his own reading, as he often
does in the case of Plato.
24 One ex. at 2.57.18. There are a number of exx. in the first doxography,
e.g. at 39.20, 49.8, 50.11, 52.10, 52.13, 56.8. But see above nn. 7 & 20.

370

DAVID T. RUNIA

tendencies. AD uses indirect speech almost incessantly. His compendium must surely have been a pain to read. Of the forty fragments in Diels' collection only a handful contain material cast in
direct speech. 25 A in contrast often prefers not to linger too long in
indirect speech. Exx. of lemmata in A where he reverts to direct
speech are: P 1.3 at 876F2 and 877F3, P 1.5 at 879B9, P 2.12 at 888C5
(= S 1.23.3), P 2.20 (= S 1.25.3d), etc. The entire (exceptionally) long
passage in P 3.5 at 894B-E is cast in oratio recta until for the briefer
doxai of the Presocratics at the end he returns to oratio obliqua. This
criterion, we note, was not made explicit by Diels.
6. Detailed comparison of the adaptation of A by P and S respectively is often revealing. P tends to abridge by simply deleting whole
lemmata or (less often) by abbreviating longer lemmata. This
means he most often retains the original order in A. Comparison
of S with P can reveal that S has replaced A with material drawn
from AD. For example in S's chapter TIEpt ioErov, parallel toP 1.10,
he changes the order from Plato Aristotle Stoics to Aristotle Plato
Stoics.2 6 This, together with the fact that the contents of the lemmata are quite different, makes clear that he has inserted the fuller
reports of AD (including the passage also found in Eusebius).
7. S tends to begin his chapters with material from A, and group
the excerpts from AD towards the end. This characteristic is hardly surprising in light of the fact that he (and the Placita in general)
rather often begins with Presocratic philosophers, who of course
hardly occur in AD's physical fragments.27 There are also cases,
however, where Diels has postulated fragments from AD right in
the middle of a series of doxai from A (e.g. 1.14.1c, 25.1i). For the
brief Chrysippean lemma at 1.8.40b this must be the case, since, as
Diels acutely saw, the sentence is repeated at 1.8.42 (at 106.6). In
other cases, however, one should be suspicious, as we shall see.
8. Since, to the best of our knowledge, the physical fragments of
AD are confined to material on Aristotle and the Stoics (apart from
the solitary fragment on Plato just referred to above2 8 ), it is these
2 5 E.g. fr. 2, 3, 39.
2 6 The reason for the change is rather obscure. A not very flattering suggestion is that he saw the phrase
AEYOI!EV<X etOl] Kat ta~ ioa~ in the Pythago-

ta

rean lemma (deleted by P), and so decided on a sequence Aristotle (etOl]),

Plato ( iom) .
27 Heraclitus is twice incidentally mentioned in relation to the Stoic doctrine of the soul at Eus. 15.20.2-3 (= fr. 39 Diels).
28 Stobaeus mostly prefers to quote Plato in the original.

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

371

name-labels which must arouse suspicion. It is striking that the

presentation of the views of various members of the Stoic school in
S is very rich, whereas in A as represented in P the divergence in
Stoic doxai is much more limited. It is possible that this was the
result of P's efforts at abridgement, i.e. he replaced the names of
individual Stoics with the generic name-label. On the basis of the
independent evidence in Eusebius we may be certain that AD
repeatedly drew attention to the contributions of individual Stoics. 29
It is on the basis of these eight criteria, combined with thorough
analyses of the relevant chapters of Stobaeus' anthology, that we
must try to identify the physical fragments of Arius Didymus.
Since there is no need to reinvent the wheel, Diels' results should
be taken as our starting-point. His collection, published in DC, consists of 64 lemmata of greatly varying length, presented as 41 fragments.3o As noted above, almost all these identifications were taken
over and correspondingly labelled in Wachsmuth's text. Diels
cautiously stated that his collection was not definitive and might
contain errors, 31 adding that the reader will be warned of difficulties in apparatus to his reconstruction. In a footnote he appended the
remark that he believed that 'scraps from Didymus were mixed in'
and cites six Aetian passages.32 This remark is not very precise,
and might be mistakenly taken to mean that material from Arius
Didymus was excerpted by Aetius.33 What he wants to say is that
in the case of some texts (all of which we shall examine below),
Stobaeus has coalesced together material from both sources in a
single lemma. Starting off from Diels, therefore, we shall successively examine the two possibilities open us: first, that he has failed
to include material in his collection that should be there, and
second, that he has included material that should not be present. 34

29 See the quote cited above at n. 15. Note that I am not basing my argument on any assumption that AD himself was a Stoic.
30 DG 447-472; the fragment appended at 854 may be considered fr. 41.
31 DG 75: neque tantum vereor ne quid falso in syllogen receperim
quantum ne omnia.
32 Ibid. n. 2: frustula ex Didymo admixta credo I 7 31 18 6 23 2 III 1 7 2 3 7 4.
33 I owe this observation to Jaap Mansfeld.
34 It will be understood that for reasons of space an exhaustive interpretation of these texts cannot be presented. The analysis concentrates on aspects
relevant to the question of identification. For example in the case of the first
text I ignore the fact that in P the lemma is attributed to Socrates and Plato.
The reader is encouraged to have the relevant texts at hand.

372

DAVID T. RUNIA

I. Additions to Diels' collection

The following texts in Stobaeus, which include the six of Diels'
footnote, deserve to be considered for inclusion among the physical fragments of AD.
l.11.29b, 37.9-16 W.: Plato. The second half of S's lemma is not found at
P 1. 7, 881E. Diels, noting the Stoicizing flavour of the phrase A.Oyot acrffiJ.uitot, surmised that this section might come from AD. This supposition is
almost certainly wrong. The lemma must be interpreted in relation to
four surrounding lemmata with differing fortunes in the sources:
Xenocrates
only in S
short in P, long inS
Plato
Aristotle
long in P, short inS
Stoics
identical in P and S
Epicurus
identical in P and S
These five lemmata have a formal feature in common, viz. that in
answering the question tic; ecrnv b 8E6c; they list a a hierarchy of gods at
various levels, beginning with the highest god (the same feature also
occurs in the well-known doxographies in Cicero ND 1.25-41 and
Philodemus De pietate printed in two columns by Diels in DC 531-550).
This common feature suggests that the lemmata all come from the same
document, which must be A rather than AD. P commonly abbreviates
longish lemmata in A. It is unusual for S to abbreviate the Aristotelian
lemma; he does this because he will later in the same chapter cite a long
quote from the De mundo, the authenticity of which he does not doubt.
The final part of the Xenocratean lemma has two features which
might suggest derivation from AD. (1) The formula apEO"KEt BE: Kat autcp
is common in AD (twice used in the fragment cited in double columns
at n. 15). It is only found once in A, at 5.29.1, where it is used to introduce
an additional comment, just as happens here. (2) The final remark that
Xenocrates handed a Platonic doctrine down to the Stoics might be
thought more in the style of AD than A (cf. again the fragment cited at
n. 15). Occasionally, however, A does give information about the provenance and appropriation of doctrines, e.g. at 1.3 877D-F on Epicurus and
Democritus. Moreover we have no idea whether AD's treatment of Plato
also took developments in the Academy into account.
2. 113.1b, 138.9-12: Aristotle. Diels DC 64 with great confidence declared
that P was wrong in associating Pythagoras and Aristotle on the question
of causes (P 1.11.3) and that S, who has separate lemmata for both
thinkers, conveys the true picture of the original A. But it is he who is almost certainly wrong. The opening words 'AptcrtotEATic; E!J>TIO"EV provides
a first clue. The verb in this formula is indicative of Stobaean intervention. It is never found in A, but does occur in 5 fragments of AD as
collected by Diels: cf. fr. 20, 21, 26, 34, 35. In all cases but the third it
stands at the beginning of the excerpt, i.e. where the anthologist has to
cover the traces of his excerpting. In the longer ethical passages from AD
in book 2 of S it is not found at all. The use of the formula may thus with
some confidence be taken to be S's work (note also the ex. cited above at n.
23), and so supports the hypothesis that S has inserted a fragment from

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

373

AD. The formulation as found in S is rather clumsy; both Heeren and

Meineke wanted to emend it. If there were originally two lemmata in A,
then there is an awkward dilemma: either Aristotle preceded Pythagoras, in which case one would have expected a contrast between Plato
tptx&c; and Aristotle tEtpax&c;; or Aristotle succeeded Pythagoras, resulting in interference with the (hacpcovia between Pythagoras and the Stoics
on the materiality of causes. We note too that inS the Aristotelian lemma
is followed by three others from AD, of Zeno, Chrysippus, Posidonius;
similar sequences are found in 18.40-42, 11.4-5, 12.1-2, 17.2-4, 25.4-15.
In short both P's and S's arrangements make much more sense if this
fragment is assigned to AD.
3.114.1e, 142.1-7: Chrysippus. Diels placed this in A's equivalent of P
1.16 on account of the subject matter and because it follows fairly soon
after the Aristotelian lemma from the same chapter (certainly from A
since in P). But that lemma is placed there because of coalescence
(joined with 3rd lemma in P's 1.12), and it is followed by a lemma from
AD. The formula Oeiva ecpaaKE is not found in P. It does occur twice in
S, here and at 1.15.6a (see next text). Both the verb and the absence of a
connecting particle suggest Stobaean intervention (see our remarks above
on the 4th criterion). More importantly the lemma seems otiose in P's
(and A's) 1.16, which appears to have a simple systematic structure A B
A-B, with the third, Aristotelian, lemma inserted as a compromise between the other two positions. Note too how the subject treated in the
Chrysippean fragment is much broader than the theme of A's chapter.
These considerations all point to AD as the source.
4. 115.6a, 146.13-21. This group of five brief lemmata is problematic.
Diels included all 5 in his reconstruction of A 1.14. Certainly the short
Zenonian lemma at the end seems misplaced. Diels thought it belonged
to 1.12 but printed it in 1.14 because of its location in S. The formula
Zf]vcov ecpa<JKE once again points to AD or rather S's adaptation of AD. A
difficulty here is that we might have expected the fragment to have been
taken from fr. 23 Diels, which discusses Zeno's views on the motion and
weightlessness of fire, but it is not found there. The association with the
previous Cleanthean lemma which also deals with fire is clumsy l:ut
transparent. But this lemma too is problematic. Its position in the Aetian
chapter is not impossible, but one would sooner expect it after the second
lemma oi a1to llu8ayopou. It is in fact possible that this lemma too comes
from AD, but there simply no way to decide. Note finally that the
Zenonian lemma is repeated in the mss. at 1.15.6e in the middle of a
quite different lemma (at 147.22). This would seem to confirm that it is a
Fremdkorper, but an ascription to AD is, all things considered, too risky.
5.118.1c, 156.15-25: Aristotle. A real puzzle. The entire lemma in S appears to consist of three parts: (i) a quote on the Pythagoreans contained
ev tEtapt<p <l>uatlcilc; aKpoaaECO<; (= lines 8-11); (ii) another quote, this time
not verbatim, but also concerning the Pythagoreans, located ev tip 7tEpt tile;
llu8ayopou cptA.oaocpiac; 7tpmt<p (= 11-15); (iii) further doxai on space and
void introduced K(lt EV aA.A.mc; AEyEt. .. (= 15-25). Diels rightly suspected
some of this material may have come from AD, but printed the entire
passage in A nevertheless. But we can fairly sure that the lemma as a
whole does not belong together. Both (i) and (ii) deal with the extra-

374

DAVID T. RUNIA

cosmic void, whereas (iii) discusses place and void in quite general
terms, without any cosmological reference. Moreover between (ii) and
(iii) the account clearly passes from the views of Aristotle on the
Pythagoreans to those of Aristotle himself (although this is not explicitly
stated). It is thus logical to conclude that there is break between (ii) and
(iii). The style of the third part of the lemma points to AD (note the use of
indirect speech). A further hint is supplied by the fact that S has deleted a
lemma on Aristotle's views on space in P 1.19. Do the first two parts of the
lemma come from A or AD? It is certainly unusual for A to have a
lemma containing two complete references, but there are about 15 texts
in which he does refer to writings with greater or lesser accuracy. 35 The
possibility that S added the learned references is also not so likely
because elsewhere he does not cite the Aristotelian corpus or lost works
(preferring to use the De mundo). The references also do not fit easily into
AD's work, which-as far as we know-did not deal with Pythagorean
philosophy directly. These considerations point to the conclusion that the
first two parts of the lemma come from A and the third part from AD. On
the other hand it should be recognized that the differences between P
and S in the first part cannot be readily explained on this hypothesis. 36
6.119.1, 162.19-163.14: Aristotle. Once again Diels thought there might
be Arian material in this lemma, but nevertheless printed it as part of his
reconstruction of A 1.23. Wachsmuth shows the same vacillation, printing it as A but adding in the apparatus haec fort. rectius Ario tribuas. There
can be no doubt, however, that, had Diels followed his own criteria, he
would have concluded that the entire lemma was from AD. Not only is it
very long, but it replaces a very short lemma in P and occurs at the end
of the sequence. It is most perplexing that Diels should have printed this
piece as part of A. Perhaps he was swayed by the fact that it starts with the
name in the nominative, as is usual in A, and is not preceded by the
name in the genitive. But the ms. P does add 'AptcrtotO. (sic) in the
margin. These genitives, as argued above, are not a reliable criterion.
7. 127.7, 226.21-26: Aristotle. Diels in his reconstruction of A 3.1 (IX; 365)
states that he thinks the second part of the lemma in S is Arian since it
disagrees with P and is clearly excerpted from Meteor. 1.8 346a19ff. He
did not, however, include this section in his collection. This passage is
the only place in the whole chapter where P and S disagree. We note the
use of oratio obliqua, even though the second word of the additional
section is yap. It is to be concluded that S has added an excerpt from his
reading of AD to A.
8. 128.1a, 227.17-21: Aristotle. Again the text for A 3.2 in P and Sis identical except for the additional Aristotelian material. Diels suspected that it
came from AD. Wachsmuth ad loc. disagreed, arguing that, since the
35 I have listed these in my article 'Xenophanes or Theophrastus? an
Aetian Doxographicum on the Sun', in W. W. Fortenbaugh and D. Gutas (edd.),
Theophrastus: his Psychological, Doxographical and Scientific Writings (New Brunswick 1992) 112-140, on p. 122-123.
36 K. A. Algra, 'Posidonius' Conception of the Extra-cosmic Void: the Evidence and the Arguments', Mnemosyne 46 ( 1993) 484-485 tentatively suggests
that it may have been contaminated with material from the Posidonian
lemma in the parallel chapter at P 2.9.

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

375

material conflicts with views expressed in the Meteorologica, it is better to

assume a lacuna and attribute the doxa to Posidonius. From the formal
viewpoint, however, the clues clearly point to AD, not only because of the
remaining parallelism between P and S, but also on account of the
similarity to S 1.29.1, where the extract on Aristotle from AD (= fr. 12
Diels) has a string of subjects introduced with the exactly the same formula twac; o otav plus subjunctive (admittedly it also occurs occasionally in A, e.g. in the Democritean lemma in the same chapter). It is possible that there is a lacuna in the text or that S made a mistake and added a
Stoic fragment from AD to an Aristotelian lemma. It is also possible that
AD had an alternative source for Peripatetic meteorological theory.
9.132, 248.7-11: Aristotle. Once again the fact that two explanations of the
origin of winds are given in the same lemma indicates that S has
coalesced two views, the first from A, the second from AD; cf. Diels rx;
375, accepted by Wachsmuth ad Zoe. But again the fragment is not found
in his collection.
The results so far have been quite remunerative: at least seven fragments should be subtracted from A and added to AD. Six relate to
Aristotle, the seventh to Chrysippus. Diels does his reader a disservice by including them in the reconstruction of A, especially because he creates the impression-most blatantly in the case of the
Aristotelian lemma in his reconstruction of A 1.23-that A contained more long and detailed lemmata than in actual fact was the
case. It also meant that these texts were missing in his fragment
collection of AD.
At this point we should note that four of the above texts (nos. 5, 79) shed valuable light on the technique of coalescence that S used
to weld his doxographical material together. It appears that he not
only merges lemmata from different chapters of A, but also is
prepared to join together in the same lemma material from both A
and AD.37 This means, I believe, that we should be on our guard in
the case of other longish lemmata which Diels attributes wholly to
A. I wish to consider three further texts which form part of Diels'
reconstruction of A 1.25-29. These chapters are highly complex,
and many of Diels' solutions, taken over by Wachsmuth in his
edition of S, are rather questionable.
10.15.15, 79.12-20: Chrysippus. The lemma on EiJ.LapJlEVTI in its entirety
would appear to be to be on the long side for A. From the beginning to
"YEVTI<JEtat P (at 1.29 885B) and S run fairly parallel (though P has
shortened the reference to Chrysippus' works). The text from JlEtaAaJ.L~avet to E7tt~oMc; is probably also Aetian, since it continues the theme of

37 Diels recognized this in a number of texts on Aristotle, though the presentation in Wachsmuth tends to conceal it: cf. 14.lc, 22.lc, 24.lm, 27.7.

376

DAVID T. RUNIA

Chrysippean diversity of terminology. But the following words J.lOtpll<; Be

introduce a break in thought. Moreover the style, with reversion to oratio
obliqua and the sequence of single nouns followed by explanatory
phrases, reflects the more descriptive, cataloguing style of AD.
11.16.17a, 87.23-88.6: Aristotle. Comparison of the sources and examination of Diels' double columns (IX; 324-5) reveal a very complicated
situation here: P's Aristotelian doxa (at 1.30 885C) can be divided into two
parts (let us say K & L). S wishes to begin with the distinction between
and lllrt<lJllltov (cf. the title of his chapter), and so places that part of
A's lemma (= L) first. He thus reverses the order of these two parts,
separates them (by means of a Platonic lemma slightly longer than what
is found in P) and adds sections (let us say M & N) to both. So S's lemma
can be represented as L M-Plato-K N. The lemma in P makes good
sense, as well asS's continuation Kilt tl]v JlEV tUXTIV ... ~ouA.Eueevto<; (= M).
But the passage with which S ends, Kilt \moeecrEt<; ... tEtllYJlEVot<; (N), is
more discursive, It introduces various new elements (e.g. apxlli, ta~t<;,
atll~tll). It also contains an entire sentence on EiJliXpJ.LEv'TI which seems
quite out of place in this chapter. This mixing of themes is characteristic
of AD rather than A. We conclude therefore that this final section (=
87.23-88.6 W.) is more likely to derive from AD than A.
12.16.17c, 89.2-5: Theophrastus. If we wish to attribute the Aristotelian
passage just dealt with to AD, we must also consider whether the Theophrastean lemma on the next page-in between S interposes an excerpt
from Plato's Laws-is also derived from him, for it appears to continue the
theme of causes discussed in the previous Aristotelian passage, i.e. Theophrastus is reported to add to the causes related to EiJ.LilPJlEVTI that of 7tpolltpmt<; missing in Aristotle. A strong argument in favour of attribution to
AD is the verb used, 7tpocrBtllpepot. It recurs in a doxographical context in
AD at Eel. 2.76.1 (i.e. Ethical doxography B), while elsewhere it occurs
only twice in the entire TLG! A lemma referring to Theophrastus would
be unique in the physical fragments of AD that we have. But we note that
he does occur in the Aristotelian-Peripatetic Ethical doxography C (at
2.140.7 W. = FGS&H fr. 449A). A revised attribution to AD would be
consistent with the fact that in the rest of A Theophrastus only occurs as a
reporter of Presocratic views (S 1.25.1b, on which see Runia art. cit. (n. 34),
and 1.26.3). The recent editors ofTheophrastus' fragments have followed
the Dielsian attribution to Aetius (= FGS&H fr. 503). 38 I submit that the
clues point to AD.

mm

The balance of probability, therefore, points to the conclusion that

these three lemmata also derive from AD's compendium.

38 Where I would translate the words ipEpEtat 1) 1tro<; 'it contributes in a

way' rather than 'he inclines in a way', as done by the editors; see W. W.
Fortenbaugh, P. M. Huby, R. W. Sharples, and D. Gutas, Theophrastus of Eresus:
Sources for his Life, Writings, Thought, and Influence, 2 vols., Philosophia Antiqua
54 (Leiden 1992) 2.329.

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

377

II. Subtractions from Diels' collection

We turn now to the second category of texts we have to consider,

namely those that Diels may have wrongly included in his
collection.

1. 'I 1.26, 31.12-14: Chrysippus. Diels included this short lemma of 3

lines on the names of Zeus (= fr. 30) because of a general similarity with
an extract from AD on the Stoics found in Eus. PE 15.15.1-9. There we
read that the cosmos is called Zeus because it is the cause of life. A
similar connection of the accusative form .!l{a with the phrase Ot' autov
1tavta is attributed to Chrysippus at Johannes Lydus Mens. 4. 71 Wunsch
(= SVF 2.1063, cf. DL 7.147). Here the context appears to be theological
rather than cosmological. An argument against Diels' attribution is the
fact that S does not use any further material from AD in this chapter. The
fragment is in direct rather than indirect speech, but this is also the case
for the similar statement in Eusebius. All in all it was wise of Wachsmuth to relegate Diels' attribution to the apparatus. The attribution of such
a short fragment to AD is no more than an educated guess.
The remaining texts whose attribution is questionable are all found
in fr. 33-34, where Diels has collected together various Stoic extracts
on the sun and moon which he thinks S drew from AD rather
than from A. Here his assumption that it is possible to separate the
material from the sources-which we have taken over in this
article-starts to crumble, Some of these fragments are so short
that they are stylistically virtually indistinguishable from the
telegram-like Aetian lemmata coalesced together by S. A more
complete treatment of the question would require that we make full
reconstructions of A's sequence of chapters on the sun and moon.
This I have done elsewhere. 39 In the present context I will largely
confine my remarks to stylistic and terminological considerations.

2. '125.5, 213.13-27: Zeno. I think there can be no doubt that this text is
from AD. In the first place it combines information about on the
substance of sun and moon and other stars which A keeps well separate
(and S does not coalesce together). Secondly it adds information about
kinds of fire that would be difficult to place in A. Thirdly it combines
information about the movement and eclipses of sun and moon, whereas
in A these are dealt with in separate groups of chapters (which S coalesces in his chapters 1.25 and 1.26 respectively).
3. '125.5, 214.1-3: Chrysippus. This lemma is much more difficult precisely because it is so short. It is possible that it represents two lemmata
39 In art. cit. (n. 35); see also 'Xenophanes on the Moon: a Doxographicum
in Aetius', Phronesis 34 ( 1989) 245-269. I hope to return to this problem in a
full reconstruction of Book II of Aetius.

378

DAVID T. RUNIA

from A joined together. Three arguments favour Diels' attribution. (1)

The lemma is placed towards the end of the chapter, and follows two
long and (in my view) indubitably Arian texts. (2) The rare term ava.9ullta.IJ.!l is used for exhalation; this is also the case in the other Chrysippean text to be discussed below, whereas in A the term is ava.GulltWJt<;
(used at least a dozen times, cf. P 2.5 887A, 2.20 890A, 3.1 893A, etc.). (3)
The same combination of substance and shape is found in the
Chrysippean lemma in the next chapter, where it is complemented with
an explanation and definition of the term 'month' which is very difficult
to place in A. None of these arguments are decisive, but Diels should be
given the benefit of the doubt.
4. 126.1i, 219.12-13: Zeno. As argued above (1, text n. 2), the formula
ZiJvrov E<pTlO'EV strongly suggests Stobaean remodelling. Moreover the
excerpt is almost identical with the Zenonian text from AD in the
previous chapter (only the word acrtpov is added). Diels' attribution is
thus justified.
5. 126.li, 219.14-15: Cleanthes. Here the situation is desparate. From the
formal point of view the two brief lemmata are identical with what we
find in P, except that they are joined together. On grounds of content I
have argued elsewhere that the former derives from A. 40 Diels' attribution is probably wrong, though it cannot be proven to be so.
6. 126.11, 219.24-220.2: Chrysippus. As argued under text 3 above, the
clustering of lemmata here points to AD. Diels' intuition should be
followed.

Conclusions
On the basis of our investigation the following results have been
reached.
(1) The criteria that Diels and Wachsmuth used to separate out the
physical fragments of Arius Didymus in Stobaeus Eclogae book I
can be slightly improved. If intelligently and carefully applied,
they are equal to their task.
(2) Particular attention should be given to the fact that the anthologist Stobaeus sometimes joins together material from his two
sources in one and the same lemma.
(3) In the case of very short fragments formal and stylistic criteria
are ineffective. It in fact becomes almost impossible to determine
with any degree of certainty whether the texts should be assigned
to the one doxographical source rather than the other.
( 4) On the basis of the examination carried out above the following
ten fragments should be added to Diels' collection.

40

Ibid. 258.

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

379

a.
b.
c.
d.
e.
f.
g.
h.
i.

5.15, 79.12-20
Chrysippus on the Moipat
6.17a, 87.23-88.6
Aristotle on eiJ.l.CXPJ.l.Evll
6.17c, 89.2-5
Theophrastus on eiJ.l.CXPJ.l.Evn
13.1b, 138.9-12
Aristotle on the four causes
14.1e, 142.1-7
Chrysippus on the divisibility of bodies
18.1c, 156.15-25
Aristotle on place
19.1, 162.19-163.14 Aristotle on motion
27.7, 226.21-26
Aristotle on the milky way
28.1a, 227.17-21
Aristotle on comets
j. 32,248.7-11
Aristotle on winds
Of great interest is the fact that in one of these texts Theophrastus
was mentioned. This confirms that the physical doxography dealt
with the thought not only of Aristotle, but also of his Peripatetic
successors. 4 1
(5) Conversely the following two fragments should be subtracted
from Diels' edition.
a. 1.26, 31.12-14
Chrysippus on the name of Zeus
b. 26.1i, 219.14-15
Cleanthes on the moon.
(6) For the assistance of scholars who wish to make use of this important doxographical source, I append below a full list of the physical fragments of Arius Didymus, taking into account the results
of this article. The sequence is determined by the following criteria: (i) the philosophers or philosophical school dealt with (PlatoPeripatos-Stoa); (ii) the source (Eusebius, abbreviated as E; Stobaeus,
abbreviated asS); (iii) the location in the source (in the editions of
Mras and Wachsmuth respectively). In the list I try to separate individual Stoic philosophers as much as I can, splitting up the fragments in Diels' collection in the process. 4 2 This leads to complications in the case offr. 36, parts ofwhich are found in both sources.
(7) Finally, I remind the reader of the remark, made at the outset
of the article, that a new edition of these fragments is highly desirable.
Leiden
4 1 Cf. the heading given to the lemma at Stob. EeL 1.17.2, 'AptcrtoteA.ouc; Kai
t&v an' auto\i, and the name-label 'AptcrtotEAtKOl at 8.40d (but these may have
been the work of Stobaeus). The title of Ethical doxography Cat 2.116.19-20 is
'AptcrtotEAouc; Kai tiDv Aol!tiDV llEptltatTittKIDV ltEpi tiDV lj9tKIDV.
4 2 There is no need to try to retain Diels' sequence of fragments, since his
ordering is by no means always logical (e.g. why place fr. 16 between 15 and
17?).

380

DAVID T. RUNIA

APPENDIX
THE PHYSICAL FRAGMENTS OF ARIUS DIDYMUS

No. Diets philosopher

source

subject

Academy
1

Plato

E 11.23.3-6, 51.18-52.11
& S 12.2a, 135.20-136.13

Aristotle
Theophrastus
Aristotle
Aristotelians
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
& successors
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle
Aristotle

S 6.17a, 87.23-88.6
S 6.17c, 89.2-5
S 8.40c, 103.10-16
s 8.40d, 103.18-104.5
s 11.4, 132.10-25
s 12.1b, 134.18-135.18
s 13.1b, 138.9-12
S 14.1c, 141.7-22
s 17.2, 152.14-17

unnamed
Cleanthes
Chrysippus
unnamed
Chrysippus
Zeno of Tarsus

E
E
E
E
E
E

Peripatos
2
3
4
5
6
7
8
9
10

6
7
2
3
5
4

11

12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

9
8
10

12
14a
11

13
14b
15
17a
17b
17c
17d
16

S 18.1c, 156.15-25

s 19.1, 162.19-163.14
S 22.1c, 196.5-15
S 24.1m, 204.8-12
s 25.4, 212.13-213.13
s 27.7, 226.21-26
S 28.1a, 227.17-21
s 29.1, 234.5-235.8
s 30.2, 240.13-242.17
s 31.6, 243.23-245.21
s 32, 248.7-11
s 36.2, 249.12-251.27
s 39, 253.25-255.7
s 51.5, 482.11-19
s 52.19, 484.15-21
s 54.3, 492.21-22
s 55.1, 494.6-7
s 56.1, 496.15-23
s 58, 497.15-25

ideas

heimarmene
heimarmene

time
time
archai: matter
archai: form

the four causes

simple bodies
mixture of bodies
place
motion
stars consist of ether
motion of the stars
sun
the milky way
comets
storms etc.
haloes and rainbows
rain and mist
winds
earthquakes
bodies of water
sense-perception
sight and hearing
smell
taste
touch
phantasia and criteria

Stoa
30
31
32
33
34
35

29a
29b
29c

36a
36b
36d

15.15.1-6, 379.12-380.12 cosmo-theology

15.15.7, 380.12-14
hegemonikon of cosmos
15.15.8, 380.15-17
hegemonikon of cosmos
15.18.1, 382.22-24
ekpyrosis
15.18.2, 383.2-7
ekpyrosis and change
15.18.3, 383.11-12
ekpyrosis doubted

ARIUS DIDYMUS' PHYSICAL FRAGMENTS

36

36c

37
38
39

37
39a
39b

40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59

39c
39d

60

61
62
63
64

65
66
67
68

69
70
71
72

26a
26b
26c
26d
21
20a
20b
20c
40
18a
18b
1&
19
38a
38b
28
25
22
23
24
36e
27a
27b
31
33a
33b
34a
34c
35

Zeno
-Cleanthes
-Chrysippus
unnamed
Zeno
Zeno
-Cleanthes
unnamed
unnamed
Chrysippus
Zeno
Apollodorus
Posidonius
Chrysippus
Chrysippus
Zeno
Chrysippus
Posidonius
Zen<r8toics43
Zeno
Chrysippus
Posidonius
Chrysippus
unnamed
Zeno
Cleanthes
Chrysippus
Chrysippus
Chrysippus
Zeno
Apollodorus
Panaetius
Posidonius
Mnesarchus
Chrysippus
Zeno
Chrysippus
Zeno
Chrysippus
Chrysippus

381

E 15.18.3, 383.7-12
& S 20.le, 171.2-7

cosmic ekpyrosis

E 15.19, 383.15-384.5
E 15.20.1, 384.7-13
E 15.20.2-3, 384.15-24

cosmic regeneration
seed and birth
nature of soul

E 15.20.4-5, 385.2-7
E 15.20.6-7, 385.9-15
s 5.15, 79.12-20
S 8.40e, 104.7-11
s 8.42, 105.8-16
s 8.42, 105.17-106.4
s 8.42, 106.5-107.7
s 10.16c, 129.2-130.20
S 11.5a, 132.27-133.5
S 11.5a, 133.6-11
S 11.5c, 133.18-23
s 12.3, 136.21-137.6
S 13.lc, 138.14-22
S 13.lc, 138.23-139.4
S 13.lc, 139.5-8
S 14.le, 142.1-7
s 14.11, 143.24-144.10
s 17.3, 152.19-153.6
s 17.3, 153.7-22
s 17.4, 153.24-155.14
s 18.4d, 161.8-26
s 19.3, 165.15-166.2
s 19.4, 166.4-22
s 19.5, 166.24-167.14
S 20.le, 171.5-7
s 20.7, 177.21-179.5
s 20.7, 179.6-17
s 21.5, 184.8-185.24
s 25.5, 213.15-27
s 25.5, 214.1-2
s 26.li, 219.12-13
s 26.11, 219.24-220.2
s 31.7, 245.23-30

cosmic soul
soul and body
Moirai
time
time
time
time
elements
ousia as matter
first matter
ousia and matter
concepts
cause
cause
cause
divisibility of bodies
three-dimensional body
cosmic cycle
cosmic cycle
kinds of mixture
place
motion
cosmic motion
motion
ekpyrosis rejected
generation-destruction
generation-destruction
description of cosmos
sun-moon-stars
sun
moon
moon
mist etc.

43 Diels' emendation Zi!vrovor; (Kai trov alt'ai>tou) is taken over by Wachsmuth. It is far from certain, but later in the lemma S uses the label oi I:trotKoi
<ptA.60'Q(pOl.

PART SEVEN

HISTORY OF SCHOLARSHIP

THE TWO PLATOS OF VICTORIAN BRITAIN

jOHN GLUCKER

A sixty-page survey of the fortunes of Plato in England (and

Scotland) since 1804 can be, in the nature of things, only a small
map of a vast territory. This is what I attempted to produce, a few
years ago, in a long article. 1 Much of it was an attempt to correct
misconceptions and put things in perspective. I emphasized the
three stages in the rediscovery of Plato in nineteenth-century
England: Thomas Taylor's first English translation of the whole of
Plato (1804) and its influence on some major literary figures; John
Stuart Mill's translations/paraphrases of some dialogues (1834-5)
and George Grote's Plato (1865), representing a searching, openended image; and Benjamin Jowett's translation with introductions of the whole of Plato (1871) as a token of Plato's final 'arrival'
on the British academic scene. It was Jowett's image of Plato, 'the
father of idealism', rather than the 'negative' and searching Plato
of Grote and Mill, which won the day and influenced the whole
course of Platonic studies in Britain and beyond.
'It is interesting to speculate,' I wrote, 'on what would have
happened if J .S. Mill had translated the whole of Plato. '2 Hastening
on with my bird's-eye view, I did not pose the question as fully as I
should have done. It is not only that Mill's translations/paraphrases became more popular than anything he had published
before 1834.3 Grote's Plato, published in 1865 and reprinted in 1867,
was reviewed in most British journals, and two of the reviewsthose in The Westminster Review and in The Edinburgh Review of
1866-were extremely favourable. 4 Yet Grote's book was only
reprinted once again, in 1874, and it exercised little, if any, influence. True, Jowett was an establishment figure, and his part in
Essays and Reviews made him something of a household name. Yet
1 J. Glucker, "Plato in England, the Nineteenth Century and After", in
Utopie und Tradition, Platons Lehre vom Staat in der Moderne, ed. H. Funke
(Wurzburg 1987) 149-210.
2 Glucker, op. cit., 174.
3 Glucker, op.cit., 169.
4 See nn. 59; 60 below and context.

386

JOHN GLUCKER

Grote's stature as a scholar was far higher than that of Jowett: as we

shall soon see, Jowett himself, in his public utterances, treated him
with respect bordering on reverence. Why, then, was Grote's view
of Plato consigned to oblivion, and what made Jowett's view
become so dominant and seminal?
Limitations of space would render it impossible-even if it were
desirable-to give here even a brief sketch of the development of
attitudes to Plato since the eighteenth century, and of British
philosophy in the nineteenth. References to some standard works
will have to do. I shall concentrate on two issues: the nature of the
images of Plato presented by Grote (and Mill) and Jowett; and the
intellectual climate at the time their works were produced.
Grote's image of Plato is best seen in Chapter VI of his Plato,
"Platonic Compositions Generally". 5 It is still extremely fresh and
readable, after all these years; but here we can only indicate some
of its salient points and cite some central passages.
Grote must have been familar with the review of Thomas
Taylor's Plato in The Edinburgh Review of 1809, where the reviewer
regards Plato as a sceptic, citing Cicero, Acad. I, 46. 6 He was also
aware of other ancient views of Plato as a sceptic: on p. 212, he cites
in evidence D.L. 3. 52 and the anonymous Prolegomena. His own
view is that 'Plato is sceptical in some dialogues, dogmatical in
others. And the catalogue of Thrasyllus shows that the sceptical
dialogues (Dialogues of Search or Investigation) are more numerous than the dogmatical (Dialogues of Exposition)-as they are
also, speaking generally, more animated and interesting. '7 One
would, then, expect Grote to devote equal space, in this chapter, to
the discussion of each of these types of dialogue. He does nothing
of the sort. Having accepted, in principle, this division of Thrasyllus, he now examines its detailed application to the dialogues. 8 As
5 G. Grote, Plato and the other Companions of Sokrates (London 1867) vol. 1,
212-280. I shall refer throughout to this second edition, which was available
to me. With very slight changes, it is essentially a reproduction of the first.
6 See n. 34 below and context.
7 I shall only mention Grote's total rejection of the possibility of reconstructing any liypacpa MrJ.lata (215-21; 273-4) and his acceptance of all
dialogues not athetized in antiquity, and all epistles, as genuine (221-6). The
latter, one of the weaker points of his book eagerly seized on by his critics,
was probably arrived at mainly because of the value of the epistles to the
historian of Greece.
8 On pp. 158-164, he had rejected the 'dramatic' division of Thrasyllus,

THE TWO PLATOS OF VICTORIAN BRITAIN

387

against Thrasyllus' list of 22 dogmatic and 14 sceptical dialogues,

Grote argues for his own 'more balanced' list of 19 sceptical and 14
dogmatic ones (236). He concludes from this proportion that Plato
is 'more negative than affirmative-more ingenious in pointing
out difficulties, than successful in solving them'. (236--7)
The discussion of 'philosophy as now understood' which follows (237-8) is crucial for our understanding of the philosophical
climate in 1865. Most people, says Grote, ask 'What positive system,
or positive truths previously unknown, has he established? ... The
philosopher is assumed to speak as one having authority; to have
already made up his mind; and to be prepared to explain what his
mind is ... Above all, an affirmative result is indispensable.' But ' ...
if philosophers are to be estimated by such a scale, he [Plato] will
not stand high up on the list. Even in his expository dialogues, he
cares little about clear proclamation of results, and still less about
the shortest, straightest, and most certain road for affirming them.'
Almost all the rest of the chapter is devoted to the 'Dialogues of
Search', discussing the educational nature of the Socratic EAE'YXO<;;
the genuine search for the truth by a genuinely ignorant Socrates;
the 'negative procedure,' inherited by Plato and the Megarians
from Socrates, and its nature as an all-out war against all false
pretensions to knowledge; the place of Socrates among the few
people in each generation who think for themselves and defy
King Nomos (one of Grote's favourite expressions)-including
most Presocratics; drama and the dicasteries as places where 'two
or more different views of looking at a question' were admitted;
and the price paid by Socrates, regarded by the Athenians as a
sophist, for expressing his doubts and views (238-267). It is only
then, on the last few pages of this long chapter, that Grote has a few
words to say on the 'affirmative' aspect of some dialogues. In my
longer survey, I had room only for a string of passages from these
pages. 9 Here one must quote at far greater length.

into tetralogies, as fanciful, but accepted his 'philosophical' division, into

sceptical and dogmatic dialogues, as significant. His reconstruction is based
on D.L. 3.49-51 and on Alcinous p. 148 Hermann. Jaap Mansfeld, Prolegomena,
Q)lestions to be settled befure the Study of an Authur, ur a Text (Leiden 1994) 74-97,
has argued convincingly that the 'philosophical division' is of Middle Platonic provenance, and is later than Thrasyllus.
9 Glucker, op. cit. (n. 1), 171-2.

388

JOHN GLUCKER

The negative procedure is so conspicuous, and even so preponderant, in the Platonic dialogues, that no historian of philosophy can
omit to notice it. But many of them (like Xenophon in describing
Sokrates) assign to it only a subordinate place and a qualified application: while some (and Schleiermacher especially) represent all
the doubts and difficulties in the negative dialogues as exercises to
call forth the intellectual effort of the reader, preparatory to full and
satisfactory solutions which Plato has given in the dogmatic dialogues at the end. The first half of this hypothesis I accept: the last
half I believe to be unfounded. The doubts and difficulties were
certainly exercises to the mind of Plato himself, and were intended
as exercises to his readers; but he has nowhere provided a key to the
solution of them. Where he propounds positive dogmas, he does not
bring them face to face with objections, nor verify their authority
by showing that they afford satisfactory solutions to the negative
procedure. The two currents of his speculation, the affirmative and
the negative, are distinct and independent of each other. Where the
affirmative is especially present (as in Timaeus), the negative altogether disappears. Timaeus is made to proclaim the most sweeping
theories, not one of which the real Sokrates would have suffered to
pass without abundant cross-examination; but the Platonic Sokrates
hears them with respectful silence and commends afterwards. The
declaration so often made by Sokrates that he is a searcher, not a
teacher-that he feels doubts keenly himself, and can impress
them upon others, but cannot discover any good solutions of
them-this declaration, which is usually considered as irony, is
literally true. The Platonic theory of Objective Ideas separate and
absolute, which the commentators often announce as if it cleared
up all the difficulties-not only clears up none, but introduces fresh
ones belonging to itself. When Plato comes forward to affirm, his
dogmas are altogether a priori: they enunciate preconceptions or
hypotheses, which derive their hold upon his belief, not from any
aptitude for solving the objections which he has raised, but from
deep and solemn sentiment of some kind or other-religious, ethical, aesthetical, poetical, &c.-the worship of numerical symmetry
or exactness, &c. The dogmas are enunciations of some grand
sentiment of the divine, good, just, beautiful, symmetrical, &c.,
which Plato follows out into corollaries. But this is a process in
itself; and while he is performing it, the doubts previously raised
are not called up to be solved, but are forgotten or kept out of sight. It
is therefore a mistake to suppose that Plato ties knots in one dialogue
only with a view to untie them in another; and that the doubts
which he propounded are already solved in his own mind, only
that he defers the announcement of the solution until the embarrassed hearer [sic] has struggled to find it for himself. (270-273)
With persons who complain of prolixity in the dialogue-of
threads which are taken up only to be broken off, devious turns and
"passages which lead to nothing"-of much talk "about it and about
it," without any peremptory decisions from an authorized judgewith such complainants Plato has no sympathy. He feels a strong

THE TWO PLATOS OF VICTORIAN BRITAIN

389

interest in the process of enquiry, in the debate per se: and he

presumes a like interest in his readers. He has no wish to short en
the process, nor to reach the end and dismiss the question as settled.
On the contrary, he claims it as the privilege of philosopical
research, that persons engaged in such discussions are noway tied
to time; that they are not like judicial pleaders, who, with a klepsydra or waterclock to measure the length of each speech, are
under slavish dependence on the feelings of the Dikasts, and are
therefore obliged to keep strictly to the point.IO Whoever desires
accurate training of mind must submit to go through a long and
tiresome circuit. Plato regards the process of enquiry as being in
itself, both a stimulus and a discipline, in which the minds both of
questioner and respondent are implicated and improved, each
being indispensable to the other: he also represents it as a process
carried on under the immediate inspiration of the moment,
without reflection or foreknowledge of the result. 11 (274-276)

These passages constitute Grote's summing-up of the chapter, and

contain his only reference, in this chapter, to the affirmative aspect
of the 'expository' dialogues. A chapter which began with admitting that, in many dialogues, Plato is dogmatic, has ended as one
of the clearest and strongest statements in modern literature about
Plato as 01C1t'tt1CO~ or ~TJ'tTJ'ttlCO~ in the ancient sense of these words.
The dogmas expressed by him are unexamined: in the best case,
they are, to him, euA.oya.
I shall quote one other passage, from Grote's discussion of Plato's
Apology.
I have already said... that I cannot take this view [of Socrates'
refutations as merely an indirect method for inculcating positive
doctrines], either of Sokrates or of Plato. Without doubt, each of
them had affirmative doctrines and convictions, though not both
the same. But the affirmative vein, with both of them, runs in a
channel completely distinct from the negative. The affirmative
theory has its roots aliunde, and is neither generated, nor adapted,
with a view to reconcile the contradictions, or elucidate the obscurities, which the negative Elenchus has exposed. That exposure does
indeed render the embarrassed respondent painfully conscious of
the want of some rational, consistent and adequate theoretical explanation: it farther stimulates him to make efforts of his own for the
supply of that want. But such efforts must be really his own: the
Elenchus gives no further help; it furnishes problems, but no
solutions, nor even any assurance that the problems as presented
admit of affirmative solutions. Whoever expects that such consummate masters of the negative process as Sokrates and Plato, when
10
II

Theaetetus 172d4 ff. (Grote's reference).

Republic 3. 394d (Grote's reference).

390

JOHN GLUCKER

they come to deliver affirmative dogmas of their own, will be kept

under restraint by their own previous Elenchus, and will take care
that their dogmas shall not be vulnerable by the same weapons as
they had employed against others-will be disappointed. They
do not employ the negative tests against themselves. When Sokrates
preaches in the Xenophontic Memorabilia, or the Athenian Stranger in the Platonic Leges, they jump over, or suppose to be already
solved, the difficulties under the pressure of which other disputants
had been previously discredited: they assume all the indefinable
common-places to be clearly understood, and all the inconsistent
elements to be brought into harmony. Thus it is that the negative
cross-examination, and the affirmative dogmatism, are (both in
Sokrates and in Plato) two unconnected operations of thought: the
one does not lead to, or involve, or verify, the other. (292)
The unprejudiced reader who has ever tried the Socratic A.enoc; on
the arguments of Book X of The Laws; or has been puzzled by the
string of 'Platonisms' poured forth by Socrates at Theaetetus 176a-d,
without argument or proof-in a dialogue where almost every
statement made by Theaetetus is mercilessly examined by Socrates; or has ever wondered what might have happened to the numerous dogmatic assertions by Socrates of Republic 11-X, so readily
accepted by Glaucon and Adimantus, had they been put to the test
of Socrates of Book 1-such a reader, whatever his own approach to
the dialogues may be, would feel no small measure of sympathy
for Grote's 'schizophrenic' picture of the dialogues. In his Preface
(p. IX), he writes: ' ... I have interpreted the numerous negative
dialogues in Plato as being really negative and as nothing beyond.
I have not presumed, still less tried to divine, an ulterior affirmative
beyond what the text reveals-neither arcana coelestia like Proklus
and Ficinus, nor any other arcana of terrestrial character.' I shall
leave it to the interested reader to see how consistent Grote
remains, throughout the three volumes of his book, to these
principles. If his approach strikes some 'modern readers' as somewhat excessive and severe, this may be partly, at least, because 'the
modern reader' has been largely conditioned by a very different
conception of the Platonic dialogues: indeed, a conception Grote
himself warned against in his strong words on 'philosophy as now
understood.'
Benjamin Jowett's translation of the whole of Plato was first
published in 1871, only six years after the first edition, and four
years after the second edition, of Grote's Plato. In the Preface to this
first edition, two out of its three pages are devoted to Grote. I shall

THE TWO PLATOS OF VICTORIAN BRITAIN

391

quote most of them, omitting Jowett's disagreement about the Platonic canon, which is beside our main argument:I2
I have derived much assistance from the great work of Mr. Grote,
which contains excellent analyses of the dialogues, and is rich in
original thoughts and observations. I agree with him in rejecting as
futile the attempts of Schleiermacher and others to arrange the
dialogues of Plato into a harmonious whole ...
If Mr. Grote should do me the honour to read any portion of this
work he will probably remark that I have endeavoured to approach
Plato from a point of view which is opposed to his own. The aim of
the Introductions in these volumes has been to represent Plato as the
father of Idealism, who is not to be measured by the standards of
utilitarianism or of any other modern philosophical system. He is
the poet or maker of ideas, satisfying the wants of his own age,
providing the instruments of thought for future generations. He is
no dreamer, but a great philosophical genius struggling with the
unequal conditions of light and knowledge under which he is
living. He may be illustrated by the writings of moderns, but he
must be interpreted by his own, and by his place in the history of
philosophy. We are not concerned to determine what is the
residuum of truth which remains for ourselves. His truth may not
be our truth, and nevertheless may have an extraordinary value
and interest for us ...
It will be seen also that I do not agree with Mr. Grote's view about
the sophists; nor with the low estimate which he has formed of
Plato's Laws; nor with his opinion respecting Plato's doctrine of the
rotation of the earth. But I am not going to lay hands on my father
Parmenides (Soph. 24ld), who will, I hope, forgive me for differing
from him on these points. I cannot close this preface without
expressing my deep respect for his noble and gentle character, and
the great services which he has rendered to Greek literature. 13

So much in public. Mter all, Grote did have an international reputation; he had been made an honorary Doctor of Oxford University
in 1853, and had been Vice-Chancellor of the University of
London since 1862. Uowett was to become Vice-Chancellor of the
University of Oxford only in 1882). In private, Jowett could be far
more explicit. In a letter to Sir Alexander Grant, written on August
21, 1865, he says:
I have been reading Grote with very great interest, but with a good
deal of disagreement ... I think without fancy that there is more to
!2 The Dialogues of Plato, translated into English with Analyses and
Introductions by B. Jowett, Fourth Edition (Oxford 1953) vol.l, xxvii-xxix. This
part of the Preface is .reprinted unchanged from the First Edition.
13 I think it was Heine who said of Victor Cousin: 'M. Cousin is doing
his best to understand German philosophy and interpret it to his countrymen; and I must admit that M. Cousin is a very nice man.'

392

JOHN GLUCKER

Plato than he supposes. I object to a kind of modern rule by which

he judges him, and I don't believe Socrates to be a mere professor of
negation, as he supposes.14
Nine years later, on April 5, 1874, he writes to Lewis Campbell:
I have been reading Grote and Schleiermacher on the Republic ...

Grote is really inexcusable in his matter-of-fact and at the same

time inconsistent manner of reading Plato, never seeing anything
according to its meaning and intention, and defending Plato as
paradoxically as he attacks him. He is always thinking, and
always thinking wrong.15

'Inconsistent manner of reading Plato'-but had not Jowett himself

maintained, against Schleiermacher, that 'there is a common
spirit in the writings of Plato, but not a unity of design in the
whole, not perhaps a perfect unity in any single dialogue [emphasis
mine. JG]. The hypothesis of a general plan which is worked out
in the successive dialogues is an afterthought of the critics who
have attributed a system to writings belonging to an age when
system had not as yet taken possession of philosophy.'? 16
What is of greater importance is that Jowett had accused Grote,
by insinuation, of reading Plato through utilitarian spectacles,
while he, Jowett, did not 'measure him by the standard of any
modern philosophical system'. Yet Plato is, for Jowett, 'the father of
ldealism ... providing the instruments of thought for future generations.' Which 'instruments of thought', and for which generations?
And is 'Idealism', capital I, less of a 'modern philosophical system'
than 'utilitarianism', minuscule u?
Jowett refers us to his Introductions to the various dialogues. The
interested reader will find in them ample material to exemplify
Jowett's approach and compare it with that of Grote. 17 I shall
14 E. Abbott and L. Campbell, The Life and Letters of Benjamin Jowett ...
(London 1897) 1.413.
15 Op. cit. (n. 14) 2.92.
16 B. Jowett, op. cit. (n. 12) xxviii. Did Jowett remember that even the
philosophical concept of <JU<J't'llJ.HX was a Stoic invention?
17 Just two examples for the contrast. In his discussion of Meno, Grote (n. 5)
2.14-23, accepts the aporetic end as a genuine aporia, and he emphasizes (23)
that nothing is said in that dialogue on the Ideas. Jowett (n. 12), 1.258-263,
has a long excursus on the Ideas, which he calls 'the ever-varying expression
of Plato's idealism' (263), and which he finds, at least hinted to, in Meno as
well (260). On Euthyphro, Grote 1.322-330, accepts the aporetic end as a genuine expression of Plato's position at the time. Jowett, ibid. 307-8, has a ready
answer: 'Socrates ... is seeking to realize the harmony of religion and morality, which the great poets Aeschylus, Sophocles and Pindar had unconsciously
anticipated, and which is the universal want of all men ... But when we

THE TWO PLATOS OF VICTORIAN BRITAIN

393

concentrate on the Introduction to one particular dialogue, which

can be regarded as most expressive of Jowett's view of Plato, and
may also throw some light on his cryptic expressions in the
Preface: his s~cond Introduction to Sophist.
In the first edition of Jowett's translation, 1B we have only one
Introduction to each dialogue, followed by the Analysis and the
translation. In the second edition, a second Introduction, following
the analysis, had been added. 19 The second introduction to Sophist
was further enlarged in the third edition by adding a few paragraphs and sentences, but in all essentials, it remains the same.2 I
shall refer to this second Introduction as it stands in the third
edition.
This second Introduction falls into two parts. The first (313-323)
is a comparison of 'the identity of being and not-being' in Sophist
with Hegel's dialectic. The second is a summary-in the mellifluous style so characteristic of Jowett's writings-of Hegel's own
dialectic. Already on the first page of the first Introductionappearing in all three editions-2 1 we are given a foretaste of what
is to come: ' ...The kindred spirit of Hegel seemed to find in the
Sophist the crown and summit of the Platonic philosophy-here is
the place at which Plato most nearly approaches to the Hegelian
identity of being and not being.' Things become clearer in the
Second Introduction:
In Plato we find, as we might expect, the germs of many thoughts
which have been further developed by the genius of Spinoza and
Hegel. But there is a difficulty in separating the germ from the
flower, or in drawing the line which divides ancient from modern
philosophy. Many coincidences which occur in them are unconscious, seeming to show a natural tendency in the human mind
towards certain ideas and forms of thought. And there are many
expect him [Euthyphro] to go on and show that the true service of the gods is
the service of the spirit [!! JG] and co-operation with them in all things true
and good, he stops short; this was the lesson which the soothsayer could not
have been made to understand, and which everyone must learn for himself.
18 The Dialogues of Plato, translated ... by B. Jowett (Oxford 1871) 3.445-473.
19 The Dialogues of Plato etc. (Oxford 1875) 4.401-424. In vol. 1 of this
edition, Preface, xi, we are told: 'The Prefaces to the Dialogues have been
enlarged, and essays on subjects of modern philosophy having an affinity to
the Platonic Dialogues have been introduced.'
2 The Dialogues of Plato etc. (Oxford 1892) 4.313-338 (25 pages, as against 23
in the second edition.) In the fourth edition (n. 12) 351-359, that part of the
second Introduction which compares Plato with Hegel was still reprinted, but
the longer section, devoted to Hegel's own system, was omitted.
21 Op. cit., third edition (n. 20) 283.

394

JOHN GLUCKER

speculations of Plato which would have passed away unheeded,

and their meaning like that of some hieroglyphic, would have
remained undeciphered, unless two thousand years and more
afterwards an interpreter had arisen of a kindred spirit and of the
same intellectual family. For example, in the Sophist Plato begins
with the abstract and goes on to the concrete, not in the lower sense
of returning to outward objects, but to the Hegelian concrete or
unity of abstractions. In the intervening period hardly any importance would have been attached to the question which is so full of
meaning for Plato and Hegel.' 22
In what follows, Jowett points out some differences between the
two, and spells out the similar elements. One of the 'general
aspects of the Hegelian philosophy' is presented thus: 'It is the ideal
philosophy which, in popular phraseology, maintains not matter
but mind to be the truth of things, and this not by a mere crude
substitution of one word for another, but by showing either of them
to be a complement of the other.'23
In such passages, we seem to have the key to those cryptic
expressions in the Preface, about Plato 'the father of Idealismproviding the instruments of thought for future generations.' Here
he is, indeed, 'illustrated by the writings of the moderns'-or of
one of them; but is he, in these passages, really 'interpreted by his
own, and by his place in the history of philosophy'? Jowett does his
best to point out differences-but in true Hegelian fashion, Plato is
taken to be something of a 'prefiguration' of Spinoza and Hegel. To
anticipate possible objections, I should point out another part of the
second Introduction, in which Jowett cites passages from Symposium, Republic I, and Theaetetus, to show that the issue of the unity of
opposites is not unique to Sophist, and concludes: 'And the Idea of
good is the source of knowledge and also of Being, in which all
the stages of sense and knowledge are gathered up and from being
hypotheses become realities. '24 'The rational is the real and the real
the rational'? Clearly, something very similar.
To understand why Jowett's Hegelian Plato won the day in the
1870's, one should take a close-albeit somewhat selective-look at
some of the developments in British philosophy in the nineteenth
century.
When Thomas Taylor published his Neo-Platonic translation of
22 Jowett, op. cit. (n. 19) 316.
23 op. cit., 317.
24 op. cit., 322-3.

THE TWO PLATOS OF VICTORIAN BRITAIN

395

Plato in 1804, his 'Proclean' conception of Plato was already out of

date in academic circles. Taylor's immense influence on some of
the greatest English poets of the period25 was due mainly to the
neglect, and general ignorance, of Plato in England during most
of the eighteenth century.2 6 On the Continent, the Neo-Platonic
conception of Plato, which had been the dominant one since Marsilio Ficino, had already been challenged many times over. The
influential works of 'the father of modern ecclesiastical history',
Johann Lorenz von Mosheim, and of 'the father of the modern
historiography of philosophy', Johann Jacob Brucker, had drawn a
clear line between Plato and his 'eclectic' interpreters in late
antiquity; and in this, they were following in the footsteps of some
distinguished predecessors.27 Mosheim's Plato was the amalgam of
Socratic and Pythagorean views, familiar to us from Cicero's De Re
Publica 1.16 and parallel passages. Brucker's 'Platonic system' was
based chiefly on 'Middle Platonic' surveys.28 What they had in
common was a total rejection of the Neo-Platonic image of Plato.
Both were read by British scholars.29 We do not know where the
largely self-taught Taylor had obtained his view of Plato.3 But
25 Glucker (n. 1) 155-164.
26

op.

cit., 151-5.

27 On the whole of this issue, the best general survey is still E.N.

Tigerstedt, The Decline and Fall of the Neoplatonic Interpretation of Plato (Helsinki
1974). It is too brief, and there are gaps. The same author's Interpreting Plato
(Stockholm 1977) is also too brief, and devotes only two pages to Grote and
none to the Mills or Jowett. The proper and extensive history of the modem
study of Plato is yet to be written. On Mosheim and Brucker, see The Decline
and Fall ... 55-61; 62-3, and notes.
2S Tigerstedt, Decline and Fall ... (n. 27) failed to notice that in his Historia
Critica Philosophiae ... , 2 vols. (Leipzig 1742) 1.669, Brucker says that in his
survey of Plato's system, he will follow Cicero's Academicus Primus (that is,
Varro's account, 19-33), Apuleius, and Alcinous, 'qui, ut omnium optime ad
ordinem nexumque philosophiae Platonicae attendit, ita dux erit nobis in
itinere ... '
2 9 A shortened version of Brucker appeared in English in 1791, in a
translation by William Enfield. The account of Plato appears in vol.l, 199125, and it contains all the main ingredients of the Latin version. James
Mill, in his review of Thomas Taylor of 1804 (see n. 33 and context), is still
apparently unaware of Brucker; but in the 1809 review in The Edinburgh
Review, he already quotes Brucker. See further nn. 45-46 and context.
30 According to the two biographical accounts reprinted in Kathleen
Raine and George Mills Harper (eds.), Thomas Taylor the Platonist (Princeton
1969) 105-132, Taylor came to Plato through reading Aristotle and his commentators, and what made him opt for the Neo-Platonic view was his reading ofProclus' Platonic Theology (112-3 and 125, probably derived from it). The
first of these accounts is almost certainly Taylor's own story (see ibid. 105,

396

JOHN GLUCKER

among the few scholars interested in Plato, this view was already
out of date.
Nor was the climate of philosophical opinion in Britain in 1804
favourable to 'the perennial philosophy' or to metaphysics in
general. Readers of that neglected genius, Thomas Love Peacock,
will remember his many digs at German philosophy, and
especially at 'the profound Kant'. On 27 September 1804, Francis
Horner, lawyer, scholar, one of the founders of The Edinburgh
Review, and a future influential politician, wrote to a Scottish friend
who proposed to translate Kant into English, and attempted to
dissuade him from carrying out his project, since metaphysics
had never been a favourite of the British philosophical reader, and
its present status-at least in the literary circles in London-was
worse than ever before.3I 1804 was obviously a good year for taking
a 'tough-minded', empirical, or even sceptical look at the dialogues
of Plato. Even in 1834-5, Mill's translations/paraphrases of some
dialogues still had a wide appeal.32 Had Mill and Grote published
their work during those years, they may have created a very
different image of the Platonic writings than that which was created by Jowett and others in the 1860's and the 1870's, and became
the most influential one for decades. But in 1804 Mill, however
precocious he was, had still two years to wait for his own birth, and
Grote was still a boy of ten. The only person, during those years,
who was capable of expressing a more empirical view of Plato was
the father of John Mill and the teacher of both, James Mill.
I have found no external evidence that the reviewer of Taylor's
Plato in vol. 14, 1809, of The Edinburgh Review [henceforth ER] was
James Mill. 33 But an earlier review, in The Literary Journal [henceforth Lj] 3 (1804) 449-61; 577-89, is signed, in both instalments,
"M. ", which stands for the founder and editor, Uames] M [ill]. The
style of both reviews is the clear and trenchant style of Mill's
n. *). But is this the whole story?
31 L. Horner, Memoirs and C01Tespondence of Francis Horner, M.P. (Boston and
London 1853) 1.282-3.
32 See n. 3 and context.
3 3 The only full biography of James Mill is still that of Alexander Bain,
James Mill: A Biography (London 1882). Although Bain deals in a number of
places with articles in ER which can be ascribed to Mill, this one is never
mentioned. Nor is it mentioned in two more recent studies of early ER
reviewers-! omit the details for lack of space. Kathleen Raine (n. 30) 535,
ascribes this review to Mill, but gives no reference to any source. Her words
were taken as evidence or near evidence by some more recent scholars.

THE TWO PLATOS OF VICTORIAN BRITAIN

397

criticism of literary and philosophical opponents. The structure of

both reviews is the same. Both begin with a general stock-taking of
the state of the study of Plato in England at the time; both proceed to
examine Taylor's Neo-Platonic approach, and to reject it in favour
of an open, sceptical approach; and both end with criticisms of
Taylor's English and specimens of his mistranslations from the
Greek (although each review takes its examples from different
dialogues.) Both reviewers (LJ 453; ER 199) cite with approval
Cicero's description of Plato as a sc.eptic at A cad. 1.46. 34 Both reviews
(L] 452-3; ER 199) describe the positive views expressed in some of
Plato's dialogues as 'hypotheses', or 'specimens of investigation',35
and emphasize the searching and examining nature of Plato's, and
Socrates', way of doing philosophy. It is more than a mere
probability that both reviews are by the same author.
I have cited elsewhere36 some passages showing that this view of
the 'sceptical' and 'dogmatical' sides of the Platonic dialogues was
shared by Grote and John Mill, and suggested James Mill as their
'archetype'. Let us see how James Mill himself regarded these two
sides of Plato. I shall cite a passage from each review.
34 In 1J 453, the citation begins "cujus (Platonis) in libris". In ER 199, it
is already "In Platonis Libris". Did the ER reviewer copy this citation from
the LJ reviewer and simplify the 'style of quotation'? In the LJ review, the
reviewer also quotes 'Varro's' strangely sceptical description of Socrates at
Acad, 1. 15-16. It appears that, in 1804, Mill had just read this work of Cicero
-or had found this particular passage quoted somewhere-and he regarded
this statement as having an 'authority sufficient to confirm an opinion
respecting the writings of Plato', and as evidence for 'the opinion entertained down to his [Cicero's] time, and entertained by himself, of the
philosophy of Plato' (Lj 453). By 1809, Mill was too busy with other literary
and political schemes (having meanwhile made the acquaintance of Bentham in 1808), and it was easier to quote from his earlier review. John Mill
was taught by his father 'several of the orations of Cicero': Autobiography, in
J.M. Robson et al. (eds.), The Collected Works of john Stuart Mill (Toronto 1981)
1.14/15. John Mill was also familiar with some of Cicero's letters to Atticus
(ibid. 48/49), but there is no indication of any acquaintance with Cicero's
philosophical works.
I have read through most of the four MS volumes of James Mill's Commonplace Book, presented by his son to the London Library. They date from his
years of increasing political involvement and activity in London. Most items
are of a political nature, and none is strictly philosophical. The ancient
works most frequently quoted are Aristotle's Rhetoric and Politics, Plato's
Republic, and Xenophon's Cyropaedia. Cicero's De Natura Deorum is cited once,
on a religious/political issue.
35 One remembers that Grote, in passages quoted above and elsewhere,
also uses these expressions.
36 Glucker (n. 1) 171-4.

398

JOHN GLUCKER

Plato is always understood to deliver his own opmwns in the

person of Socrates. But in almost all the dialogues of Plato, Socrates,
as Cicero remarks, in the quotation above [Acad. 1.15-16, quoted
earlier on the same page], advances no opinion himself, but refutes
the opinions of some other person. It is true he often proposes
hypotheses on the subjects under consideration, and contrasts them
with those of the person he is examining, in order to puzzle him,
and contribute to his refutation. But it scarcely ever appears to be
any part of the intention of Plato, that these hypotheses should be
considered as the opinions of Socrates, or as his own. It would have
been perfectly inconsistent with the character of Socrates to
represent him as advancing these or any other hypotheses as his
own opinions. It always appears to us in the perusal of these
dialogues, that the hypotheses which have been denominated the
Platonic visions, and stated as the foundation of what has been
called the Platonic philosophy, were introduced for the sake of
refutation merely, and that Plato in general adhered to the rule of
his master; played with the theories of others; and advanced
nothing seriously himself but what concerned virtue. If any
authority be sufficient to confirm an opinion respecting the writings of Plato, it is that of Cicero, who expressly affirms that in this
respect the practice of that philosopher was conformable to that of
his great predecesor, "Cujus (Platonis) in libris," says he, " nihil
affirmatur, & in utramque partem multa disseruntur; de omnibus
quaeritur, nihil certe 3 7 dicitur." (Lj 452-3).
One of the most remarkable features of the writings of Plato is, that
he affirms nothing ... In most of the Dialogues of Plato, the object is
to refute the tenets and expose the ignorance of some of those
sophists who travelled about Greece, under pretence of teaching
eloquence and philosophy, and who, in general, filled the minds
of the youth with a spirit of mere logomachy, and with the worst
impressions of right and wrong, with regard both to public and
private life. 38 The ingenuity, the acuteness, the address, the
eloquence with which this delicate and important task is performed, render the perusal of these Dialogues among the most
improving exercises which can engage a juvenile mind. Hardly
any thing, in the way of example at least, can be conceived to
sharpen the faculties; to render acute in discerning, and ingenious
in exposing, fallacies; to engender a love of mental exercise; and to
3 7 All MSS read nihil certi. So do all the editions known to me. I have
checked all eighteenth-century editions available in London libraries, as
well as the 1804 Oxford edition. Nihil cert!J. dicitur would mean 'nothing,
indeed, is said', or 'nothing is really said'. Both are absurd in themselves,
and not what the sentence requires. No sane scholar would offer such a reading as an emendation. It seems clear that the author of the LJ review made a
mistake in copying this sentence, and the author of the ER review copied
from the same mistaken copy-or from the LJ review. Hardly a coincidence.
38 This is a view of the sophists which Mill's pupil, Grote, was later to
reject, both in his History of Greece and in his Plato.

THE TWO PLATOS OF VICTORIAN BRITAIN

399

elevate with the ambition of mental excellence [as do these

dialogues]. 39 In some of the dialogues, as in those with Alcibiades,
the object is to expose some of the false impressions which are apt to
prevail in the mind of men, and to lead to the most dangerous
consequences. 40 In these the skill with which the misapprehension
is analyzed; the variety of ridiculous lights into which it is thrown;
and the power of argument as well as of satire which is employed
to expose it, operate as the stronger sanative. In those of different
description, as in the books concerning Polity and Laws, the
business is to give specimens of investigation, to let in rays of light,
to analyze particular points,and by throwing out queries or hypotheses, to encourage speculation, rather than lay down and establish any system of opinions. Accordingly, Cicero tells us, 'In
Platonis libris nihil affirmatur; et in utramque partem multa
disseruntur; de omnibus quaeritur, nihil certe dicitur.' (ER 119).

We know very little of James Mill's formative years before his

arrival in London in 1802. We do know that, during his period in
Edinburgh, 1794-5, he was an avid reader of Plato, and borrowed
his works again and again from the library.4 1 But we have no
information concerning his source for the sceptical view of Plato
(and no evidence that he was an avid reader-or a reader at all-of
Cicero's philosophical works: see n. 30). His chief teacher of
philosophy in Edinburgh was Dugald Stuart,42 and in his published works, Stuart treats Plato clearly as a philosopher with
positive opinions and doctrines. 43 When he wrote his LJ review,
39 In 1809, John Mill was only three, and was only beginning to read
Plato in Greek under his father's tuition: Autobiography (n. 34) 8/9. When he
was twelve, he studied Plato more properly. His father's opinions on the value
of Plato 'to young students', and the impression made on him at the time, are
narrated in Autobiography 24/5(the 'early draft' is here much shorter than the
full draft), in terms similar to those of our ER passage. But it has a sentence
which was mitigated in the final version: ' ... and I have ever felt myself,
beyond any modern that I know except my father and perhaps beyond even
him, a pupil of Plato, and cast in the mode of his dialectics.'
One should, however, also note Mill's words on the limited usefulness of
'negative philosophers' in his essay on Bentham, Collected Works (n. 34), vol.
10 (1969) 79-80. Comparing that passage with both James and John Mill's
emphasis on the use of Plato for young students, one can hardly reconcile this
position with the words of praise just quoted from the early draft of the
Autobiography.
4 0 Bentham's famous Book of Fallacies springs to mind. It was published
only in 1824, but Mill must have known some of its contents from his many
conversations with Bentham, and he may well have seen part of the
manuscript.
41 Bain, op. cit. (n. 33) 18-19.
42 Bain, op. cit. (n. 33) 13-14.
4 3 See, e.g, Sir William Hamilton (ed.), The Collected Writings of Dugald
Stuart (Edinburgh 1854) 1.48-9, 260-1, 368; 2.368; 4.71, 75, 180, 238.

400

JOHN GLUCKER

Mill was familiar with the translation of six dialogues by Floyer

Sydenham, of whose scholarship he speaks with great respect. 44
But Sydenham, although he follows the Thrasyllan division of the
dialogues into 'sceptical and dogmatical', provides the reader with
a view of Plato's dialogues as forming a systematic whole. 45 The
author of the ER article is also familiar with Brucker, whom he
quotes extensively in disparagement of the Neo-Platonists (ER 1947). But was he familiar with Brucker before 1809? He is not
mentioned in the LJ review. 46 In any case, Brucker's Plato was no
sceptic, but rather a Middle Platonist. 47 One is tempted to think that
Mill had read Orner Talon's Academica of 1550, in which the
author discusses duas disputationis vias et rationis of the Academy,
the negative one, in utramque partem, and the positive probabilis
disputatio, citing as evidence Cicero, who is his major source in the
whole of this work-but also the dialogues of Plato. 4B Be that as it
may, one notes that he held this view of Plato four years before he
knew Bentham. Thus, the sceptical view of Plato which he was to
pass on to John Mill and to Grote was all his own. This is not to
deny that, once he came under the spell of Bentham, he began to
ascribe to Plato some 'Benthamite' features. 49
44
45

L]577.

Dialogues of Plato, three volumes (London 1767, 1769, 1773.) No translator's name on the title-pages, but each dialogue has a dedication to some
public figure, signed by Sydenham. Vol. 1, 5-19, has 'A general View of the
Works of Plato,' where the translator's opinions are made clear. Sydenham's
translation is a rare book today; but the 'General View' was reprinted by the
learned Thomas James Mathhias in his-more accessible-edition of the
works of Thomas Gray, vol. 2 (London 1814) 289-296, as an introduction to
Grats own notes on the dialogues.
4
In 1805, James Mill published in London a translation of Charles
Villers, An Essay on the Spirit and Influence of the Reformation of Luther. On p. 315,
Brucker is mentioned-but in a note by the author.
47 See n. 28 and context.
48 Audomari Talaei, Academica (Paris 1550) 11-15. Talon also cites Quintilian 2.15.26, for the division of Socrates' sermones into O.. qKtucoi and
OOYJ.lllttKoi. On Talon, see Tigerstedt, Decline and Fall (n. 28) 36, and C. B.
Schmidt, Cicero Scepticus, a Study of the Influence of the Academica in the Renaissance
(The Hague 1972) 78-91. It is at least likely that the Calvinist Talon, friend
and apostle of the Calvinist Ramus, would continue to be popular in Calvinist
Scotland even after the decline of Ramism. This should be looked into by
someone more familiar with Scottish philosophy and theology than myself.
See also next note.
49 See nn. 38-39 and context. In his A Fragment on Mackintosh (London
1835) 25, James Mill regards Plato as the originator of Bentham's method of
classification. A similar view is expressed by John Mill in his essay on
Bentham (n. 38), 88; and in his review of Grote, Collected Works (n. 34) 11.405,

THE TWO PLATOS OF VICTORIAN BRITAIN

401

Two reviews were hardly sufficient to exercise a wide influence

on people's view of Plato-especially since, in the first decade of
the century, Plato was still largely ignored in Britain. Although he
still considered applying to the chair of Greek in Glasgow as late as
1818, 50 James Mill became, after 1808, almost entirely absorbed by
political and economic theory and practice. The task of expounding a sceptical Plato to an unsuspecting British public was left to his
two disciples. They, too,were busy men. John Mill published his
first Platonic articles in 1834-5, and even they were part of an
incomplete project.5I Grote did not publish his book on Plato until
1865, 'as a sequel and supplement to my History of Greece'.52 In the
meantime, some momentous developments were taking place in
British philosophy.
Whether the final 'arrival' of Hegelianism in England should
be dated from the publication of James Hutchison Stirling's The
Secret of Hegel in 1865,53 or with Benjamin Jowett's reading and
disseminating of Hegel's philosophy in Oxford after his long visit
to Germany in 1844,5 4 is an issue yet to be determined on an
examination of the available evidence. In the first half of the century, Utilitarianism was the most prominent philosophical doctrine around. By the 1860's, it was clearly being pushed aside by
British Hegelianism, which held the field well into the present
century. 55 But 'idealistic' tendencies began to appear in British
philosophy-mostly under German influences-much before the
1860's. Sir James Mackintosh became familiar with Kant, Fichte
he regards the method of division and classification in dialogues like Sophist
and Statesman as 'an anticipation of Ramus and Bentham'. This, incidentally,
may imply some familiarity with Ramus: why not with Talon as well-and
perhaps, in both cases initiated by James Mill? Thus, by 1835, James Mill can
treat Plato as having positive doctrines-if only in matters of government,
where Socrates' exposition in Republic appears to agree with his own view
(Mackintosh 285-290, citing passages from Rep. III, and speaking of Plato
'taring it down as a universal truth'.)
0 Bain (n. 33) 165-8.
51 Glucker (n. 1) 167-170.
52 Grote, Plato (n. 5) l.iii.
53 See Amelia Hutchison Stirling, james Hutchison Stirling, His Life and Work
(London 1912) 114-5, 119-135.
54 Abbott and Campbell (n. 14) 89-92; Geoffrey Faber, Jowett, A Portrait with
Background (London 1957) 177-183.
55 For a balanced account of these developments and their aftermath, see
Frederick Copleston, A History of Philosophy, vol. 8 (London 1967). For an
analytic philosopher's brief review of British Hegelianism, see G.J. Warnock, English Philosophy since 1900 (London 1958) 1-11.

402

JOHN GLUCKER

and Schelling in the early years of the century, and Sir William
Hamilton's essays in the 1830's show close familiarity with
Continental philosophy in general. Coleridge's, and Carlyle's,
works of the 1820's and 1830's had already made a wide circle of
readers aware of the revolutionary nature of much of the new
German philosophy. About the same time when Jowett began to
acquaint Oxford with Hegel, a group of philosophers in Trinity
College, Dublin, began to turn philosophy in their university in
the new, 'idealistic', direction. In 1837, William Archer Butler was
made the first incumbent of the new chair of Moral Philosophy.
His lectures on ancient philosophy, published many years after
his death,56 must date from the 1830's or the early 1840's: they
could not have been delivered later than 1847, when he was
'engaged in the preparation of some work on faith' .57 In them, he
shows familiarity with Kant, Fichte and Schelling, who are
mentioned by name, and uses some ideas and expressions which
may come from Hegel. Butler attempts to reach a compromise
between 'the Science of Real Experience' and 'the Philosophy of
Induction.' But when he comes to Plato, he sees the Idea of the
Good as the main object of Plato's philosophy. 'The predominant
quality of Platonic philosophy' is for him 'the perfect union of
Absolute Goodness with Absolute Reality. '58
One of Butler's successors in Trinity, Thomas Maguire, Professor of Moral Philosophy from 1880 to his death in 1889, was described by his biographer as 'a thorough idealist in philosophy.'59
In his Essay on the Platonic Idea of 1866, and Lectures on Philosophy of
1885, he fights the battles of idealism against many an 'empiricist'
of the time, and regards Plato and Hegel as the two great
champions of idealism. To go back in time, and to return to
England: in 1841, Professor William Sewell of Oxford published a
book called An Introduction to the Dialogues of Plato. Although Sewell
has some interesting things to say about the dialogues as works of
art, he is sure that Plato expected the intelligent reader to arrive at
the 'grand formularies of his doctrine:' a philosophical system,
with the Idea at its centre and the education and improvement of
5 6 William Archer Butler, Lectures on the History of Ancient Philosophy, ed.
W. H. Thompson, 2 vols. (Cambridge 1856).
57 DNB, 'Butler, William Archer' (by John Dowden).
58 Lectures ... , 1.63.
59 DNB 'Maguire, Thomas' (by E.S. Robertson.)

THE TWO PLATOS OF VICTORIAN BRITAIN

403

man as its goal. In 1845, George Henry Lewes published the first
edition of what was to become a very popular book, A Biographical
History of Philosophy. Lewes was a professed positivist, but he treats
Plato clearly as an idealist, with the Idea of the Good at the centre
of his thought, and with his own psychology, dialectic, theology,
cosmology and ethics.
Against this background, we should be surprised, not at the
failure of Grote's work to exercise any significant influence on the
study of Plato in Britain for the next few generations, but at the fact
that some of the reviews of it were positive and laudatory. But
everyone must have known or guessed that the review in ER 6o
was written by a close associate, and it was soon revealed that the
author was John Mill. As for the favourable review in The Westminster Review 61-what could one expect of the 'Benthamites"
periodical, edited for many years by James Mill? The short and
cursory review in The Gentleman's Magazine and the long and
learned review in The Qy,arterly Review, 62 while full of respect for
Grote's learning, were entirely opposed to his views of Plato. The
QR reviewer is clearly an idealist, and he takes Plato's idealism for
granted. He ascribes to Plato's 'most exact writings ... the necessity
of reconciling the abstract with the concrete, the Ideal with the
Actual,' and 'an approach ... to a new and larger idea of knowledge, not merely as the Universal, in which subjective peculiarities are done away [Schelling?], but as the union of all permanent
relations in the contemplation of the mind [Hegel?].' Jowett
himself could not put it better: indeed, I would not be surprised if it
were revealed that Jowett himself was the reviewer. Needless to
say, the reviewer dismisses Grote's whole conception of Plato's
dialogues as inadequate, and adds: ' ... it remains for those whose
point of view is nearer to Plato's own to vindicate him.' This, of
course, is insidious: 'Plato's own point of view' can only be a
construct, based on someone's reading of the dialogues. But such
expressions show that by 1866, Grote's idealist critics were fighting
from a dominant position. Needless to say, Grote is accused again

60 The Edinburgh Review 123 (1866) 297-364; reprinted in Mill's Dissertations

and Discussions vol.3 (1867) 275-379, and now in Mill's Collected Works (n. 34)
vol. 11 (1978) 377-440.
61 The Westminster Review 28 (1866) 459-482.
62 The Gentleman's Magazine N.S. II (1866) 382-390; The Qy.arterly Review 119
(1866) 109-153.

404

JOHN GLUCKER

and again-sometimes explicitly, often by insinuation-of intruding his Radical or Utilitarian views into the dialogues (a thing
which he does from time to time,63 but not as often as one could
guess from his reviewers.) And, as if it was not bad enough, in the
1860's, to be a Radical-Grote even served as a Radical Member of
Parliament between 1832 and 1841-he was also labelled 'Positivist'
by the Gentleman s Magazine reviewer. Grote was not a Positivist, but
his personal friendship with Comte was well-known, and in the
same year as his Plato-1865-his good friend and fellowUtilitarian John Mill published his Auguste Comte and Positivism.
The reviewer knew his master's crib when he wrote: 'The secularist [i.e. Grote] cannot sympathize with the great thinker who
clings with the tenacity of a drowning man . .. less to the arguments for the soul's immortality which unassisted reason supplies,
than to the belief itself, which so commended itself to his mind
and heart.' 64 This sounds almost like Augustine's 'Christian'
Plato-with his Christianity brought in by insinuation.
John Mill's laudatory review of Grote's Plato was not the only
place where he echoed the sceptical image of Plato which he
inherited from his father. It is now common knowledge that
many of the ideas of On Liberty of 1839 were conceived as part of his
rebellion against Bentham's proposals for a 'totalitarian democracy', in which the majority has full control of the expression of
thought.65 One may, perhaps, offer the suggestion that much of
Chapter II of that work was the result of Mill's reflections on
Socrates, Plato, and even the sceptical Academy. I refer not only to
the famous passage beginning with the words 'Mankind can
hardly be too often reminded that there was once a man called
Socrates. '66 I refer also to the passages beginning 'The beliefs
which we have most warrant for. . .';67 and 'He who knows only his
63 E.g., in his Plato (n. 5) 2.77-89, he takes Socrates' proposal of a 'beatific
calculus' in Protagoras as a central view of Plato himself. This is one of the
issues which his critics jump on, as a proof of Grote's parti pris in general.
6 4 Ibid. 390. Grote also assumed that Plato held such beliefs, often 'from
deep and solemn sentiments,' and even from religious sentiments. See, e.g.,
his Plato (n. 5) 1.271. What he was not prepared to do is to accept such beliefs
as if they were the essence of Plato's philosophy.
65 For two recent works, see Isaiah Berlin, Four Essays on Liberty (Oxford
1969) 172-206, esp. 175-185; Gertrude Himmelfarb, On Liberty and Liberalism, the
Case ofjohn Stuart Mill (New York 1974), esp. 3-22.
66 J.S. Mill, On Liberty in Collected Works (n. 34) vol. 18 (1977) 235.
67 Mill, op. cit. (n. 66) 232.

THE TWO PLATOS OF VICTORIAN BRITAIN

405

own side of the case .. .'68 Such passages are full of sceptical
Academic terms like 'assent' and 'suspension of judgement,' and
advocate the Academic practice in utramque partem disputandi.
But by the 1860's, as we have seen, the game was over. Plato was
studied once again in Oxford -and very soon, in the other
universities, and the idealists were now in control, both of the
academic study of philosophy and of much of public opinion. The
idealist image of Plato became canonical, and has been taken for
granted since, in various shapes and forms, by most students of
Plato and ancient philosophy. Even some analytic philosophers,
finding in some of the Platonic dialogues 'prefigurations' of their
own discussions, felt it necessary to point out that this is not the
way Plato himself regarded what he was doing.69 It was only
recently, in a book on Plato published in 1982, that an analytic
philosopher rediscovered the two sides of Plato with which the
reader of this article should by now be thoroughly familiar. He
calls them, for convenience's sake, Pato and Lato.70 Was Professor
Hare aware of the fact that Grote and Mill had already distinguished between these two sides of Plato more than a century
ago?
In the past, I expressed the view that, if I had to make a choice, I
would rather accept the sceptical Academy's image of the sceptical
Plato than any of the, far more influential, dogmatic images of
him, from those of Xenocrates and Aristotle to the present day.
Meanwhile, a new approach to the Platonic dialogue has made its
appearance: reading the dialogue as an organic unity in which
drama, narrative, characters and philosophical arguments work
together for the purpose of achieving a philosophical discourse in a
concrete, human (albeit larger than life) situation.71 I regard this
relatively new approach to Plato as an important and seminal new
method, which does justice to the Platonic texts on all their aspects.
What made it easy for me to understand this approach and adopt it
68 op. cit., 245.
69 See, e.g., G.E.L. Owen, 'Plato on not-being', in his Logic, Science and

Dialectic, ed. M. Nussbaum (Ithaca, N.Y. 1986), 104-137, The article was first
published in 1970.
70 R.M. Hare, Plato (Oxford 1982), esp. 26-7 and 69 ff.
7l Of a number of works, I shall mention two recent ones: Michael C.
Stokes, Plato's Socratic Conversations ... (London 1986); Ivor Ludlam, Hippias
Major: an Interpretation (Stuttgart 1991). I do not refer to the rather eccentric
use made of the dialogue form by the late Leo Strauss.

406

JOHN GLUCKER

was, among other things, my long preoccupation with the sceptical Academy's approach to Plato. I wonder if such a new and
fruitful way of reading Plato might not have been discovered
earlier, had the British philosophical community adopted Grote's
and Mill's image of the sceptical Plato rather than the dogmatic
image promoted by Jowett and other idealists.
Tel-Aviv University

DREI BRIEFE WILHELM DILTHEYS AN ULRICH VON

WILAMOWITZ-MOELLENDORFF ( 1908-191 0) 1
WILLIAM

M.

CALDER

III

Einleitung
Hochbetagt schrieb Wilhelm Dilthey (1833-1911) die nachfolgenden drei Briefe an Ulrich von Wilamowitz-Moellendorff
(1848-1931).2 Sie weisen eine Reihe von Gemeinsamkeiten auf.
Zum einen handeln sie allesamt von Wissenschaftlern des
kaiserlichen Berlin, namlich von Eduard Zeller, Georg Misch,
Reinhard Kekule von Stradonitz und Friedrich Hiller von
Gaertringen, sowie von einem Projekt der Koniglich PreuBischen
Akademie der Wissenschaften. Zum anderen ist ihnen gemein,
daB die genannten Gelehrten iiberwiegend einerseits der Berliner
Universitat verbunden waren und andererseits der Koniglich
PreuBischen Akademie der Wissenschaften angehorten.
Zu den zu jener Zeit nach Ausweis der Mitgliederverzeichnisse
rund dreiBig ordentlichen Mitgliedern der philosophisch-historischen Klasse der Berliner Akademie zahlten sowohl der Verfasser

1 Ich danke Prof. Dr. Ulrich K. Goldsmith (Universit.iit Colorado) fiir

seine glanzende Abschrift der schwierigen Texte. Dr. Sven Rugullis
(Berlin/Urbana), Oldfather Research Associate (1992/93) hat die deutsche
Version der Einleitung geschrieben und vie! Material iiber Dilthey fiir mich
gefunden. Ich bin ihm dafiir sehr dankbar. Der honorandus, Prof. Dr. Jaap
Mansfeld, hat in einem Brief an mich vom 29. Juli 1980 zum erstenmal
versucht, die Briefe zu datieren, und auf wichtiges exegetisches Material
hingewiesen. Es ist mir eine echte Freude, die editio princeps der Briefe Jaap
Mansfeld zu widmen, der so viel fiir die moderne Wissenschaftsgeschichte,
besonders fiir die Wilamowitzforschung getan hat.
2 Die Behauptung, zwischen Dilthey und Wilamowitz habe kein Kontakt
bestanden, ist falsch. So aber Klaus Oehler, 'Dilthey und die Klassische
Philologie', in: Hellmut Flashar - Karl fried Grunder - Axel Horstmann
(Hg.), Philologie und Hermeneutik im 19. Jahrhundcrt. Zur Geschichte und Methodologie der Geisteswissenschaften (Gottingen 1979) 190. Fiir die Druckerlaubnis
der Briefe ( = Wilamowitz-NachlaB No. 273) danke ich Dr. H. Rohlfing
(Handschriftenabteilung der Niedersachsischen Staats- und Universitatsbibliothek Gottingen).

408

WILLIAM M. CALDER III

als auch der Adressat der drei Briefe. Dilthey3 war 1882 an die
Friedrich-Wilhelm-Universitat berufen und 1887 als ordentliches
Mitglied in die Akademie aufgenommen worden. Wilamowitz,
dessen Schwiegervater Theodor Mommsen der Akademie ein
halbes Jahrhundert lang angehort hatte, kam 1897 nach Berlin
und wurde nach einem gescheiterten Aufnahmeversuch 1899 4
ordentliches Akademiemitglied.5 Zeller hingegen war bereits
1872 aus Heidelberg einem Ruf nach Berlin gefolgt und noch im
gleichen Jahr als ordentliches Mitglied in die Akademie kooptiert
worden. Nach seiner Ubersiedlung nach Stuttgart im Jahre 1894
blieb er ihr als auswartiges Mitglied verbunden. Kekule, unter
dessen Bonner Studenten sich im Sommersemester 1881 der
einstige preuBische Thronanwarter und damalige Kaiser Wilhelm II. befand, 6 wurde von diesem 1889, d.h. ein Jahr nach der
Thronbesteigung, zum Direktor der Sammlung antiker Skulpturen
und Gipsabgiisse bei den Berliner koniglichen Museen sowie zum
Honorarprofessor und spater zum Ordinarius an der Berliner
Universitat ernannt. Im selben Jahr folgte seine Aufnahme in die
Akademie als ordentliches Mitglied7. deren 'unmittelbarer Protector' Wilhelm II. war.B Im dritten Brief wird ferner Friedrich Leo
(1851-1914) genannt, der Wilamowitz' Gottinger Kollege gewesen
war und seit 1906 als korrespondierendes Mitglied der preuBischen Akademie angehorte. In jenem Schreiben wird auBerdem
Wilamowitz' Schwiegersohn Friedrich Hiller von Gaertringen er3 Thomas Kornbichler, 'Wilhelm Dilthey', in: Michael Erbe (Hg.),
Berlinische Lebensbilder. Geisteswissenchaftler (Berlin 1989) 195-208, 201ff.
4 Vgl. Theodor Lorenz (Hg.), Friedrich Paulsen An Autobiography (New York
1938) 300-301.
5 Christa Kirsten (Hg.), Die Altertumswissenschaften an der Berliner Akademie.
Wahlvorschliige zur Aufnahme von Mitgliedern von F. A. Wolf bis zu G. Rodenwaldt
(1799-1932) (Berlin 1985) 117.
6 Hans Schrader, 'Reinhard Kekule von Stradonitz', Bursians Biographisches
Jahrbuch 35 (1913) 1-40, 20.
7 SB Akad.Wiss.Berlin (1912) 616; Schader, a.a.O., 22. Der Wahlvorschlag
wurde von Alexander Conze ( 1831-1917) eingebracht und ist lediglich von
Johannes Vahlen (1830-1911) mitunterzeichnet. Er beginnt mit folgenden
Worten: 'Nachdem es misghickt [sic] ist die gegenwartige vakante Fachstelle
fiir Kunstwissenschaft mit einem Vertreter der Wissenschaft moderner
Kunst zu besetzen, wird man zur Besetzung der Stelle einen Vertreter der
Wissenschaft antiker Kunst suchen wollen. DemgemaB schlagen die Unterzeichneten Herrn Reinhard Kekule yon Stradonitz, den Nachfolger unseres
heimgegangenen Curtius in dessen Amtern an Universitat und Museum, zu
dessen Nachfolger auch in der Akademie vor.' (Kirsten, a.a.O., 113).
8 Adolf von Harnack, Geschichte der Kiiniglich Preujlischen Akademie der
Wissenschaften zu Berlin 1 (Berlin 1900) 1044.

DREI BRIEFE DILTHEYS AN WILAMOWITZ

409

wahnt. Nachdem Wilamowitz 1902 die Leitung der griechischen

Inschriftensammlung, welche die Akademie betreute, von Adolf
Kirchhoff (1826-1908) iibernommen hatte, wirkte Hiller seit 1904
in einer eigens dort fiir ihn geschaffenen Stelle als wissenschaftlicher Mitarbeiter an den lnscriptiones Graecae.9 Die Abhandlungen
und Sitzungsberichte der PreuBischen Akademie schlieBlich erschienen in Kommission bei Georg Reimer, dem Schwiegervater
von Mommsen und dem GroBvater von Wilamowitz' Gemahlin.
Eine Ausnahme in diesem Kreis bildete lediglich Georg Misch.
Als einziger der Genannten war er zeit seines Lebens nicht mit
der Berliner Akademie verbunden. Dafiir hatte er 1905 den Hauptpreis in der 1900 von ihr gestellten Preisaufgabe uber eine Geschichte
der Autobiographie gewonnen. Eingereicht wurden his zum 31.
Dezember 1904 zwei Arbeiten, die beide einen der zwei ausgesetzten Preise erhielten: Mischs Werk gewann den mit 5000 Mark
dotierten Hauptpreis, Kurt Jahn fiir sein unvollstandiges Traktat
den zweiten Preis in Hohe von 2500 Mark.IO
AuBer diesen personellen Verbindungen besteht eine weitere
Gemeinsamkeit der drei Briefe darin, daB sie alle ohne Jahreszahl
datiert sind. Der erste wurde durch eine Kritik von Wilamowitz an
dem Nachruf auf Eduard Zeller veranlaBt, den Dilthey am 5. April
1908 fiir die Wiener Neue Freie Presse verfaBt hatte. Das zweite
Schreiben behandelt Diltheys Sorgen iiber die Verbindung seiner
Tochter Clara mit Georg Misch. Und im letzten Dokument wird
weder die Krankheit noch der Tod Reinhard Kekules erwahnt, der
nach langerem Leiden am 22. Marz 1911 verstarb.II Damit bildet
das Jahr 1910 den terminus ante. Der terminus post fiir dieses
Schreiben ergibt sich aus Diltheys Angabe 'GrunewaldSiemensstr. 37'. Bis zur Ausgabe fiir 1909 benennt der Kiirschner
'Berlin W, Burggrafenstr. 4' als Diltheys Wohnsitz.I2 Das Verzeichnis der Akademiemitglieder fiihrt Dilthey bereits am 1.
Januar 1910 unter der neuen Anschrift.I 3 Damit sind die drei
Briefe in die Zeit von 1908 his 1910 zu datieren.
9 SB Akad.Wiss.Berlin physikal.-mathem. Kl. (1936) 158; Leo Stern
(Hg.), Die Akademie d(r Wissenschaften in der Zeit des lmperialismus, Teil 1,
(Berlin 1975) 161.
lO SB Akad.Wiss.Berlin (1905) 686-689.
II Schrader, a.a.O., 33.
I2 Kurschners Deutscher Literatur-Kalender auf das Jahr 1909, 31. Jahrgang
(Leipzig 1909) 302.
I3" SB Akad.Wiss.Berlin (1910) IX.

410

WILLIAM M. CALDER III

Als letzte Gemeinsamkeit sei erwiihnt, daB Dilthey und Wilamowitz nicht nur durch die Berliner Universitiit und Akademie,
sondern gleichfalls durch die zentrale Person des ersten Briefs
miteinander verbunden waren, niimlich durch Eduard Zeller
(1814-1908). Diltheys Verhiiltnis zu Zeller ist bekannt. 14 Als junger
Privatdozent lernte Wilamowitz ihn durch Theodor Mommsen
kennen und-wie bereits zuvor sein Studienfreund Hermann
Diels (1848-1922) 15-schiitzen. 16 Davon zeugt auch sein Brief, den
er am 7. Dezember 1881 an Zeller richtete:17
gewohnt daB meinen arbeiten nichts als hohn und rancune
gegeniiber laut wird, ist mir nachst dem eigenen gewissen die
einzige beruhigung daB die manner, auf deren urteil ich den
hochsten wert lege, das urteil der tageswelt nicht teilen. es verstand
sich bei dieser arbeit 1Bvon selbst, daB ich an Sie, hochverehrtester
herr geheimrat, an Bernays und Usener als Ieser eigentlich
immer dachte.
Angeregt wurde das erste der nachstehenden Schreiben des
Geheimen Regierungsrats Dilthey, der 1905 seine Lehrtiitigkeit
aufgegeben hatte, durch die Reaktion des Wirklichen Geheimen
Rates Wilamowitz auf den Nachruf, den Dilthey auf den kurz
zuvor verschiedenen Wirklichen Geheimen Rat Eduard Zeller
verOffentlicht hatte.l9 Das Thema war ihm wohlvertraut. Nach
mehreren Rezensionen und Aufsiitzen iiber das Schaffen Zellers
14 Siehe nur Eduard Zeller, Erinnerungen eines Neunzigjiihrigen (Stuttgart
1908). Vgl. Maximilian Braun, William M. Calder III, Dietrich Ehlers,
Philology and Philosophy: The Letters of Hermann Diels to Theodor and Heinrich
GomtJerz (1871-1922) (Hildesheim 1995) 67, 169.
l5' Zu Diets und Zeller siehe Otto Kern, Hermann Diels und Carl Robert. Ein
biographischer Versuch (Leipzig 1927) 60f., 65, 103-107. Ferner Dietrich Ehlers
(Hg.), Hermann Diels- Hermann Usener- Eduard Zeller. Briefwechsel (Berlin 1992)
mit Nachweisen in Bd. 2, 497 (sub nomine Dilthey, Wilhelm).
16 Ulrich von Wilamowitz-Moellendorff, Erinnerungen 1848-1914 (Leipzig
21929) 175, 180.
1 7 Ulrich von Wilamowitz-Moellendorff, 'Selected Correspondence 18691931', hg. von William M. Calder III, Antiqua 23 (Neapel 1983) 301.
IS Ulrich von Wilamowitz-Moellendorff, 'Antigonos von Karystos', Philologische Untersuchungen 4 (Berlin 1881) (Nachdruck Berlin-Zurich 1965).
19 Wilhelm Dilthey, 'Eduard Zeller', Neue Freie Presse, 5. April 1908, 3135. Dieser Beitrag wurde urspriinglich nicht in Diltheys Gesammelte Schriften
aufgenommen, wei! er nach Meinung des Herausgebers des vierten Bandes,
Herman Noh!, gegeniiber Diltheys friiherem, in die Edition aufgenommenem Beitrag 'Aus Eduard Zellers Jugendjahren' inhaltlich nichts Neues
enthalte (578; vgl. Vorwort, VIII). Noh! irrt. 1970 schlieBiich wurde der
Nachruf in dem von Ulrich Herrmann edierten fiinfzehnten Band der
Gesammelten Schriften wiederabgedruckt (267-278).

DREI BRIEFE DILTHEYS AN WILAMOWITZ

411

hatte er bereits 1897, Zellers fiinfzigstem ProfessorenjubiHium zu

Ehren, eine Festrede gehalten, in der er dessen Jugendjahre und
anfangliche intellektuelle Entwicklung beschrieb.2 Als Zeller am
19. Marz 1908 im Alter von 94 Jahren verstarb, hielt sich Dilthey
zur Kur in Tirol auf. Dort konnte er, wie er klagte, den Nachruf
nur mit Hilfe eines Konversationslexikons anfertigen. Ein Exemplar jenes Artikels sandte er sodann an Wilamowitz, woraufhin
dieser wissenschaftsgeschichtliche Ungenauigkeiten monierte.
Wilamowitz meinte, dem Artikel entnehmen zu miissen, daB
Dilthey Ferdinand Christian Baur (1792-1860), 'dem groBen Begriinder der modernen, kritischen Erforschung des Urchristentums'21, und seiner Tiibinger Schule, die in Baurs Schwiegersohn
Eduard Zeller und in David Friedrich Strauss ihre bekanntesten
Vertreter fand, eine 'exakte philologische Methode' zuschrieb.22
Gegen die Unterstellung eines derart 'groblichen Versehens' oder
'Schnitzers', wie Dilthey es formulierte, verwahrte sich der 74jahrige. Er suchte zu ergriinden, wie seinem Kollegen jener
lrrtum unterlaufen sein konnte, und fand sich schlieBlich damit
ab, daB der Altphilologe Wilamowitz ungenau gelesen haben

muBte.
Der zweite Brief handelt iiber die Verlobung der 1877
geborenen Dilthey-Tochter Clara mit Georg Misch (1878-1965),
einem Schuler ihres Vaters, der zu jener Zeit sein Leben als
Privatdozent fiir Philosophie an der Berliner Universitat fristete,
jedoch 1911 in Marburg ein Extraordinariat erhalten und 1916
20 Diese Rede wurde unter dem Titel 'Aus Eduard Zellers Jugendjahren'
in der Deutschen Rundschau (1897, 294-309) veroffentlicht und ist in Diltheys
Gesammelten Schriften, Bd. 4 (Leipzig/Berlin 1921) 433-450 Ieicht greifbar.
2 I Wilhelm Dilthey, 'Emilie Zeller', Nationalzeitung, 17. Mai 1904;
wiederabgedruckt in: ders., Gesammelte Schriften, Bd. 15 (Gottingen 1970) 265f.,
265.
22 Uber Baur und seine Schule vergleiche Eduard Zeller, 'Ferdinand
Christian Baur', ADB 2, 172-179; ders., 'Die Tubinger historische Schule', in:
ders., Vortriige und Abhandlungen 1 (Leipzig 21875) 294-389; ders., 'Ferdinand
Christian Baur', in: ders, Vortriige und Abhandlungen, Bd. 1, 390-479; Ferdinand
Christian Baur, Kirchengeschichte des neunzehnten Jahrhunderts, hg. von Eduard
Zeller (Tubingen 1862) mit Diltheys Rezension in: ders., Gesammelte Schriften,
16 (Gottingen 1972) 449-456; Eduard Zeller, Geschichte der deutschen Philosophie
seit Leibniz (Munchen 21875) 721, 723, 729 mit der Rezension von Dilthey in:
ders., Gesammelte Schriften 17 (Gottingen 1974) 322f. Ferner seien Zellers
Strauss-Editionen erwahnt sowie sein Aufsatz 'StrauB und Renan', in: ders.,
Vortriige und Abhandlungen, 1. 480-555; s. besonders Eduard Zeller, Ausgewiihlte
Briefe von David Friedrich Strauft (Bonn 1895). Zur Tubinger Schule vgl. Horton
Harris, The Tii.bingen School (Oxford 1975).

412

WILLIAM M. CALDER III

Extraordinarius sowie 1919 in der Nachfolge von Heinrich Maier

(1867-1933) Ordinarius in Gottingen werden sollte. Wilamowitz
war dieser Brauch aus eigener Erfahrung nur allzugut bekannt.
Hatte er nicht selbst 1878 als junger Gelehrter Theodor Mommsens Tochter Marie geheiratet und erlebt, wie spater seine eigene
Tochter Dorothea den Epigraphiker Friedrich Hiller von Gaertringen ehelichte, der im dritten Brief erwahnt ist? Dilthey konnte
mithin bei Wilamowitz nicht nur Verstandnis, sondern Mitgefiihl
erwarten. Auch war der junge Misch fiir den Altphilologen schon
Hingst kein Unbekannter mehr. Man erinnere sich etwa an
Wilamowitz' Besprechung von Mischs erstem Band der Geschichte
der Autobiographitf23, die aus der erwahnten Preisaufgabe hervorging, oder an die lobende Erwahnung, die dem damaligen
Berliner Privatdozenten in Wilamowitz' Schreiben vom 1. August
190824 an die nicht lange zuvor in den Ruhestand getretene 'Graue
Eminenz' des preuBischen Hochschulwesens, Friedrich Althoff
(1839-1908), zuteil wurde, dessen Intervention Wilamowitz seinerzeit seinen Lehrstuhl erst in Gottingen, spater in Berlin nicht
unmaBgeblich zu verdanken hatte.25 Gleichfalls sind fiir die
Folgezeit mehrere Briefe von Wilamowitz an Misch erhalten.2 6
Als eine Akademiesitzung ausfiel, wandte sich Dilthey mit seinem Anliegen schriftlich an Wilamowitz und verfaBte den letzten
Brief der vorliegenden Trias. Ihm kann entnommen werden, wie
ein Werk der Akademie entstand. Nach einer Begutachtung
23 Georg Misch, Geschichte der Autobiographie, Bd. 1: Das Altertum (Leipzig
1907); Wilamowitz, 'Die Autobiographic im Altertum', lnternationale Wochenschriftfor Wissenschaft, Kunst und Technik 1 (1907) 1105-1114; wiederabgedruckt
in: ders., Kleine Schriften 6 (Berlin/Amsterdam 1971) 120-127.
2 4 William M. Calder III I Alexander Kosenina, Berufungspolitik innerhalb
der Altertumswissenschaft im wilhelminischen PreujJen. Die Briefe Ulrich von Wilamowitz-Moellendurff an Friedrich Althoff (1883-1908) (Frankfurt/Main 1989) 161.
25 Vgl. Karl Reinhardt, 'Ulrich von Wilamowitz-Moellendorff', in: Die
grojJen Deutschen 5 (Berlin 1957) 415-421, 417, woes euphemistisch heiBt: 'Der
Universitatsgewaltige Althoff verhalf ihm [Wilamowitz] zu seiner Villa in
Westend.' William M. Calder III, 'Die Rolle Friedrich Althoffs bei den
Berufungen von Ulrich von Wilamowitz-Moellendorff', in: Bernhard vom
Brocke (Hg.), Wissenschaftsgeschichte und Wissenschajtspolitik im Industriezeitalter.
Das 'System Althoff in historischer Perspektive (Hildesheim 1991) 251-266.
26 William M. Calder III und Sven Rugullis, 'Ulrich von WilamowitzMoellendorff on Wilhelm Dilthey. His Letters to Georg Misch (1914-1928),'
Illinois Classical Studies 16 (1992) 337-345 = William M. Calder III, Further
Letters of Ulrich von Wilamowitz-Moellendorff (Hildesheim 1994) 137-145. Vgl.
dazu Jaap Mansfeld, 'Two Letters to G. Misch,' Wilamowitz nach 50 jahren, hgg.
von William M. Calder III, Hellmut Flashar, Theodor Lindken (Darmstadt
1985) 203-206.

DREI BRIEFE DILTREYS AN WILAMOWITZ

413

durch Friedrich Leo und durch den 'Hausarchaologen' Kekule

wurde das Vorhaben, 'archaologische Arbeiten und Gelegenheitsschriften' zu veroffentlichen, dem Akademieangestellten Hiller
von Gaertringen (1864-1947) uber seinen Schwiegervater Wilamowitz angetragen. Hiller akzeptierte und veroffentlichte die ArkadischenForschungen zusammen mit Heinrich Lattermann (1882-1914).
Das Werk erschien 1911 in Kommission bei Reimer.27
II
Briefe

[ 19 April1908]

Hochverehrter Herr College,

Besten Dank fur Ihre freundlichen Worte uber das zu Zellers
Gedachtnis Gesagte.28 Ich glaubte dem Freunde das schuldig zu
sein29, und so habe ich es in sehr schlechtem Gesundheitszustande und ohne alle Bucher, nur mit Hilfe von ein paar Zahlen
aus dem Conversationslexicon niedergeschri[e]ben; selbst meine
friihere Rede die Zellers Jugendjahre behandelte war mir nicht
zur Hand.30 So muB ich nachtraglich urn Nachsicht fiir das
Zugesandte bitten.
27 Friedrich Hiller von Gaertringen/Heinrich Lattermann (Hg.), Arkadische Furschungen (Berlin 1911).
28 Wilhelm Dilthey, 'Eduard Zeller', Neue Freie Presse, 5. April 1908, 3135. Vgl. die Gediichtnisrede, die Hermann Diels auf Zell~r am 2. Juli 1908
in der Akademie vortrug (Abh. Akad.Wiss.Berlin 1908). Uber den Verbleib
der Briefe Wilamowitz an Dilthey ist mir nichts bekannt.
2 9 Uber sein Verhiiltnis zu Zeller iiuBerte sich Dilthey in Briefen an
Paul Yorck von Wartenburg. Ende 1883 klagte er: 'Mit Zeller lebe ich gut
zusammen, doch eigentlicher geistiger Austausch will sich nicht entwickeln.' Ein halbes Jahr spiiter schrieb er: 'Den Gespriichen mit Zeller
fehlt ein belebendes Etwas, ein Hauch, der die Gebeine des Vergangenen auf
der Triimmerstiitte der Geschichte wieder lebendig macht. Noch gestern
saBen wir hier lange auf dem Balkon zusammen. Am meisten lebendig
wurde doch wieder unsre Unterhaltung als wir auf die Tiibinger Schule
kamen: da er einst von dem Genie Baurs personlich benihrt worden ist.'
(Brie.fwechsel zwischen Wilhelm Dilthey und dem Grafen Paul Yurek v. Wartenburg
1877-1897 [Halle 1923] 37, 46).
30 Aus AnlaB des fiinfzigjahrigen Professorenjubilaums Zellers hielt
Wilhelm Dilthey eine Lobrede, die unter dem Titel 'Aus Eduard Zellers
Jugendjahren' in der Deutschen Rundschau 1897, 294-309 publiziert wurde. Sie
ist wiederabgedruckt in: ders., Gesammelte Schriften 4 (Leipzig/Berlin 1921)
433-450.

414

WILLIAM M. CALDER III

Aber berichtigen mochte ich doch Ihre Auffassung meiner

Worte nach welcher ich 'der Tiibinger Schule exacte philologische Methode nachgeriihmt' hatte, und darum erlaube ich mir
diese Zeilen zu senden, urn den Vorwurf eines so groblichen
Versehens iiber einen Gegenstand dem ich in meiner Jugend
und auch spater zeitweise viel Zeit gewidmet habe 3I nicht auf
mir sitzen zu lassen. Ihre Auffassung konnte leicht-wenn ich
meinerseits eine Vermuthung wagen darf-, daraus entstehen,
daB Sie etwa p 1. unten u. p 2. oben in den Satz 'Verbindung einer
exakten philologischen Methode' nicht32 die Worte 'Tiibinger
Schule' im vorhergehenden Satze als Subjekt in Gedanken hiniiberbringen.33 In Wirklichkeit sondere ich gerade seine 'eigenthiimliche Stellung' in der Schule darin von Baur u. den anderen
Freunden ab. Baur ging ausschlieBlich von historischen Apercii' s
aus, unterstiitzt durch religionsgeschichtliche Vergleichungen.
Ich habe das friiher in einem groBeren Aufsatz ausfiihrlicher dargelegt.34 Was Zeller selbst betrifft, so glaube ich, als ich die entsprechenden Aufsatze in seinen Jahrbiichern 35 und die Apostelgeschichte36 durcharbeitete, doch das was ich in meinem friiheren
aus der Rede entstandenen Aufsatz in der Rundschau37 naher
dargelegt und jetzt in dem Nachrufe wieder angedeutet habe,
richtig gesehen zu haben. Wenn man seine Richtung auf 'exakte
philologische Methode' eben in den Schranken der Zeit nimmt.
Der Grundzug seines Verfahrens bleibt ja immer, auch in der
griechischen Philosophiegeschichte die Systematisierung nach
31 Diltheys altphilologische Studien bezeugt Herman Noh!, 'Wilhelm
Dilthey', in: Die groflen Deutschen 4 (Berlin 1957) 193-204, 196. Vgl. Arnold
Bork, Diltheys Auffassung des griechischen Geistes (Berlin 1944).
32 So Jaap Mansfeld. Die Lesung ist nicht sicher.
33 Dilthey hatte geschrieben: 'In dieser Untersuchung Zellers [sc. Zellers
Apostelgeschichte] zeigte sich aufs glanzendste die Eigenttimlichkeit der
historischen Methode, durch welche er in der Ttibinger Schule seine eigenttimliche Stellung behauptet. Sie lag in der Verbindung einer exakten philologischen Methode mit dem groBen Zug der Anschauung einer historischen
Entwicklung, wie er in dem Geiste Baurs war.'
34 Wilhelm Dilthey (Wilhelm Hoffner), 'Ferdinand Christian Baur',
Westermanns Illustrierte Deutsche Monatsschrift (1865) 581-599; wiederabgedruckt
in: ders., Gesammelte Schriften 4 (Leipzig/Berlin 1921) 403-432.
35 Eduard Zeller war der Herausgeber des Archivs fur systematische Philosophie und Soziologie sowie der Theologischen jahrbilcher.
36 Eduard Zeller, Die Apostelgeschichte nach ihrem Inhalt und Ursprung untersucht (Stuttgart 1854).
37 Wilhelm Dilthey, 'Aus Eduard Zellers Jugendjahren', Deutsche Rundschau ( 1897) 294-309.

DREI BRIEFE DILTHEYS AN WILAMOWITZ

415

einem festen Schema, und ich habe, so wie man das in diesem
Momente durfte, darauf hingewiesen, wie 'geniale Leistungen',
die aus einer hochentwickelten Philologie kamen, von ihm theils
verwerthet wurden, theils hat er in der ihm eigenen Continuitat
seiner Arbeit sich mit ihnen abzufinden gewuBt, ohne daB er
doch selbst Philologe in diesem groBen Sinne gewesen ware.3B
Das wird man bei aller Verehrung fiir seine groBe Leistung
zugestehen miissen.
Und nun verzeihen Sie diese Weitlaufigkeit iiber etwas das
wenn Sie meine Zeilen erhalten schon hinter Ihnen liegt. Aber
gerade bei Ihnen mochte ich doch nicht gern im Verdacht eines
solchen Schnitzers stehen.
lch hatte das miindlich gesagt-aber mein Gesundheitszustand
ist Ieider so erbarmlich daB ich an eine Riickkehr nach Berlin
noch lange nicht denken kann.
In herzlicher Verehrung
der Ihrige
[gez.] Wilhelm Dilthey
Meran SchloB Labers
Ostern

[21 Mai 1908]

Hochverehrter Herr College,

Nehmen Sie meinen besten Dank fiir die aufrichtigen Worte die
Sie Ihrem Gliickwunsch zur Verlobung meiner Tochter39 hinzugefiigt haben. Sie begreifen sicher, daB es mir Bediirfniss ist,
einige Worte darauf zu erwidern. Sie wissen wie stark und lebhaft
38 In Wilamowitz' Geschichte der Philologie, die 1921 erstmals erschien, ist
Baur nicht erwahnt. Uber Zeller auBert er sich sehr positv. Die Stelle lautet
im Nachdruck der dritten Auflage (Leipzig 1959) 67: 'Alles tritt in Schatten
vor Eduard Zellers groBer Geschichte der Philosophie. Durch ihn dringt
ein, was die Tiibinger Theologenschule vor den Philologen voraushatte: eine
geistige Bewegung durch die Personen der Trager hindurch zu verfolgen,
also den geschichtlichen Zusammenhang neben dem Herausarbeiten der
einzelnen dogmatischen Systeme.'
39 Diesem Schreiben ging Wilamowitz' Antwortschreiben auf Diltheys
Verlobungsanzeige seiner Tochter Clara mit Georg Misch voraus. Clara
Misch verOffentlichte 1930 Der junge Dilthey. Ein Lebensbild in Briefen und Tagebilchern 1852-1870, das in zweiter Auflage (Stuttgart/Gottingen 1960) erschien.

416

WILLIAM M. CALDER III

ich zu jeder Zeit fiir meinen kiinftigen Schwiegersohn 40 eingetreten bin. Und meine Uberzeugung von seinen hervorragenden
geistigen Gaben und einer seltenen Energie und verzichtenden
AusschlieBlichkeit in der wissenschaftlichen Arbeit war immer
verbunden mit der von seinem Charakter. Daher habe ich, als
seine Anfrage 41 kam, nach eingehendem Gespdi.ch mit meiner
Tochter meine Einwilligung gegeben, aber mir die Bedingung
vorbehalten, daB er ein Extraordinariat oder eine demselben aquivalente Stellung erst erwerbe; Und ich habe meine Uberzeugung,
daB diese Bedingung nach der Lage der Dinge erforderlich sei,
wie sie auf langem aufreibendem Nachdenken beruht, selbstverstandlich auch nicht aufgeben konnen. Meine sehr geliebte und
innigst mit mir verbundene Tochter hielt aus Grunden, die ich
zwar anerkenne, die mich jedoch durchaus nicht iiberzeugen
konnen, ein moglicher Weise langes Warten nicht fiir richtig.
Entstand nun hieraus zwischen uns Beiden eine Differenz, so ist
doch unsere Liebe zu einander unversehrt aus ihr hervorgegangen, meine Achtung fiir Misch ist unverandert, und mein
Herz und meine Fiirsorge sind ganz bei der Verlobung, Ausstattung, Verbindung meiner Tochter, wenn auch mein sehr
bedenklicher Gesundheitszustand mich von Berlin fernhalt. 42 Ich
bin schon seit Jahren sehr leidend, und in diesem Winter hat
mein Gesundheitszustand eine gefahrliche Form angenommen:
zu arbeiten werde ich natiirlich immer wieder versuchen, wie
sehr es mich auch schadigen mag: ich konnte sonst nicht Ieben:
das muB ich besser wissen als die Arzte. II Ich habe keinen Grund
von diesem ganzen Verlauf irgend etwas zu verhehlen, nachdem
gegen meinen Willen durch Unberufene ein thorichtes Gerede
entstanden ist.
Nehmen Sie nochmals fiir
Ihren Brief besten Dank.

40 Georg Misch. Dieser war nach dem Tod Diltheys einer der Herausgeber seiner Gesammelten Schriften, beschiiftigte sich aber auch monographisch
mit seinem Schwiegervater (Lebensphilosophie und Phiinomenologie. Eine Auseinandersetz.ung der Dilthey'schen Richtung mit Heidegger und Husser[ (Leipzig 1930);
Vom Lebens- und Gedankenkreis Wilhelm Diltheys (Frankfurt/Main 194 7). Vgl.
Josef Konig, 'Georg Misch als Philosoph,' Nachrichten der Akad. Wiss.Gott.
Philol.-hist. Klasse 7 (1967) 151-243.
41 Urspriinglich: sein Antrag.
42 Clara Dilthey setzt schlieBiich ihren Willen durch. Die EheschlieBung erfolgte noch 1908 (Wer ist wer? 1 [Berlin 1962] 1027).

DREI BRIEFE DILTHEYS AN WILAMOWITZ

417

In aufrichtiger und treuer Verehrung

der Ihrige
[gez.] Wilhelm Dilthey
Me ran,
Lahers
Siidtirol
d. 2l.sten Mai

Grunewald-Siemensstr. 37

Hochverehrter Herr Kollege.

Beifolgenden Brief Leos hatte ich Ihnen in der Sitzung vorlegen
wollen, da ich nun aber vernehme, daB sie verschoben ist, erlaube
ich mir, denselben zu iibersenden. Die archaologischen Arbeiten
und Gelegenheitsschriften sind wie ich voraussetze an Herrn von
Kekule 43 gesandt. Hat auch dieser sein Urteil abgegeben, dann
hatten Sie vielleicht die Freundlichkeit, die heiden AuBerungen
Ihrem Herrn Schwiegersohn 44 vorzulegen, damit derselbe sich
auBere, ob er geneigt sei, die Herausgabe zu iibernehmen. Man
wiirde dann auch einigermaBen den U mfang abschatzen konnen
und auf Grund davon mit dem Verleger in Verhandlung
eintreten.
In Verehrung ergeben
der Ihrige
[gez.] Wilhelm Dilthey
Berlin/Urbana

43 Reinhard Kekule von Stradonitz (1839-1911). Zur Korrespondenz

zwischen Wilamowitz und Kekule siehe William M. Calder III, 'Ulrich von
Wilamowitz-Moellendorff to Kekule von Stradonitz on Friedrich Gottlieb
Welcker', Studi italiani di filologia classica III 2 (1984) 116-133 =Further Letters,
81-95.
44 Friedrich Freiherr Hiller von Gaertringen (1864-1947): vgl. Hiller von
Gaertringen, 'Erinnerungen und Betrachtungen eines alten Epigraphikers,'
NJAB (1942) 108-112 und G. Klaffenbach, Gnomon 21 (1949) 274-277.

STEPPING INTO THE STREAM

A Bibliography of the Publications of Jaap Mansfeld

(1964-1995)
Note from the editors. This bibliography contains all major publications (books, articles of monograph length, articles, review articles,
reviews) published by Jaap Mansfeld in the period 1964-1995. It
does not include some minor publications or contributions (such as
newspaper reviews) written in Dutch.
1964

[ 1] Die Offenbarung des Parmenides und die menschliche Welt, Philosophical texts and Studies 9 (Assen: Royal Van Gorcum &
Comp.), pp. 285 (simultanously published as Ph.D. Dissertation,
University of Utrecht, under the direction of C. J. de Vogel).
1965

[2] review of F. Adorno, La Filosofia antica I, Mnemosyne 18: 291-3.

[3] review of K Axelos, Heraclite ou la philosophie, Mnemosyne 18:
417-9.
1966

[ 4] 'Parmenides Fr. B 2, 1 ', Rheinisches Museum 109: 95-6.

[5] 'Hermeneutische techniek: Heidegger's methodiek bij de
interpretatie van enige vroeggriekse denkers', Algemeen Nederlands Tijdschrift voor Wijsbegeerte 58: 19-34.
[6] 'A Reply to Prof. Vollenhoven', Philosophia Reformata 31: 124-5.
[7] review of E. Milobenski, Der Neid in der griechischen Philosophie,
Mnemosyne 19: 431-2.
1967

[8] 'Heraclitus on the Psychology and Physiology of Sleep, and

on Rivers', Mnemosyne 20: 1-29.
[9] review of H. Cherniss, Die altere Akademie, Gymnasium 74: 362-4.
[10] review of W. N. A. Klever, Anamnesis en Anagoge, Archiv fur
Geschichte der Philosophie 49: 324-6.
[11] review of L. Tarcin, Parmenides, Mnemosyne 21: 315-9.

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

419

1968
[12] review of J. Bollack, Empedocle I, Mnemosyne 21: 426-8.
[13] review of M. Giusta, I dossografi di etica I, Mnemosyne 21: 436-8.
[14] review of G. F. Parker, A Short Account of Greek Philosophy,
Bibliotheca Orientalis 25: 224-5.
[15] review of Th. Cole, Democritus and the Sources of Greek Anthropology, janus 55: 220-4.
[16] 'Positivistische Hermeneutiek bij Popper', in Handelingen dertigste nederlandse filologencongres (Groningen: Wolters-Noordhoff),
65-7 [Abstract].

1969
[17] review article G. E. R. Lloyd, Polarity and Analogy: Two Types
of Argumentation in Early Greek Thought, Lampas 2: 46-57.
[18] review of F. M. Cleve, The Giants of Pre-Sophistic Greek
Philosophy 1-11, Mnemosyne 22: 436-8.
[19] review of F. Dieterle, Platons Laches und Charmides, Gymnasium
76:93-5.
[20] review of D. J. Furley, Two Studies in the Greek Atomists,
Gymnasium 76: 100-3.

1970
[21] 'cr<pa.tpf\c; EV<X.AtyJCtOV oyKcp', in Akten des XIV. Internationalen
Kongresses fur Philosophie, Wien 2-9 Sept. 1968, Bd. V (Wien:
Herder), 414-9.
[22] review of E. de Strycker, Beknopte geschiedenis van de antieke
filosofie, Mnemosyne 23: 198-9.
[23] 'Nawoord', in Grondslagen van het wetenschappelijk denken,
Voordrachten Studium Generale Rotterdam (Rotterdam:
Universitaire Pers Rotterdam), 141-51.
[24] 'Naschrift' in Dutch translation with commentary of M.
Heidegger, Wat is dat -Filosofie? (Tiel-Utrecht: Lannoo), 60-1.

1971
[25] The Pseudo-Hippocratic Tract IIEPI 'EB.tWMA.J.QN Ch. 1-11 and
Greek Philosophy, Philosophical Texts and Studies 20 (Assen:
Royal Van Gorcum & Comp.), pp. 271.

420

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

1972
[26] 'Ambiguity in Empedocles B 17, 3-5: A Suggestion', Phronesis
17: 17-39.
[27] 'Three Notes on Albinus (1. Problems of Transcendence. 2.
Forms and Qualities. 3. The Descent of Soul)', Theta-Pi 1: 61-80;
repr. in [125].
[28] 'Man the Measure and Sense-Perception', Theta-Pi 1: 128-39.
[29] review of]. Moreau, Le sens du Platonisme, Gymnasium 79: 254-5.
[30] 'Literatuurbericht: Wijsbegeerte in de Oudheid, 1', Algemeen
Nederlands Tijdschrift voor Wijsbegeerte 64: 140-58.
[31] 'Literatuurbericht: Wijsbegeerte in de Oudheid, II', Algemeen
Nederlands Tijdschrift voor Wijsbegeerte 64: 216-33.

1973
[32] Theorie en Empirie. Filosofie en geneeskunst in de voorsokratische
periode. Assen: Van Gorcum, pp. 38 [inaugural lecture Utrecht
University].
[33] 'Notes on Some Passages in Plato's Theaetetus (l. The Objects
of Knowledge. 2. The Mathematical Passage)', in Zetesis. Liber
amicorum E. de Strycker (Antwerpen-Utrecht: De Nederlandse
Boekhandel), 108-14.
[34] 'Moeite met Plato', Lampas 6: 255-66.
[35] review of Th. G. Sinnige, Matter and Infinity in the Presocratic
Schools and Plato, Mnemosyne 26: 64-70.
[36] review of M. Giusta, I Dossografi di Etica II, Mnemosyne 26: 75-6.
[37] review of K. Oehler, Antike Philosophie und byzantinisches
Mittelalter, Gymnasium 80: 324-5.
[38] 'Grieks-wijsgerig commentaar', philosophical comments in
Dutch translation with commentary of M. Heidegger, Brief over
het humanisme (Tiel-Utrecht: Lannoo), 125-9.

1975
[39] (with P. W. van der Horst) An Alexandrian Platonist Against
Dualism: Alexander of Lycopolis' Treatise 'Against the Doctrines of
Manichaeus', translated with an introduction, notes, and index
(Leiden: E. J. Brill), pp. 99; introduction repr. in [ 125]; simultaneously published at Theta-Pi 3: 1-97.
[ 40] 'Alcmaeon: Physikos or Physician?', in J. Mansfeld and L. M.
de Rijk (eds.), Kephalaion. Festschrift C. J de Vogel, Philosophical

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

421

Texts and Studies 23 (Assen: Royal Van Gorcum & Comp.), 2638.
[41] 'Voorsocratici', Lampas8: 318-26.

1976
[ 42] Review of H. Lloyd:Jones, Semonides on Women, Maatstaf 24:
71-3.

1977
[ 43] 106 articles on ancient philosophy, ranging from the very
brief to the quite extensive, in K. Kuypers (ed.), Encyclopedie van
de .filoso.fie (Amsterdam-Brussel: Argus Elsevier); revised and
bibliographically updated versions of contributions to the major
Dutch general encyclopedia, the Grote Winkler Prins, 7th edition
(Amsterdam-Brussel: Elsevier), 1966-1975; these have been
reprinted in the Grote Winkler Prins, 8th entirely revised edition
(Arnsterdam-Brussel: Elsevier), 1979-84.
[44] 'Aristoteles over waar en onwaar', Wijsgerig Perspectiefl8: 18-9.

1978
[45] 'Zeno ofCitium', Mnemosyne3I: 134-78.
[ 46] review of]. Bollack, Empedocle II-III, Mnemosyne 31: 207-9.

1979
[ 47] Heraclitus: Fragmenten, bezorgd, vertaald en toegelicht (Amsterdam: Polak en Van Gennep), pp. 79; 2nd rev. ed. 1987.
[ 48] 'The Chronology of Anaxagoras' Athenian Period and the
Date of his Trial. Pt. I, The Length and Dating of the Athenian
Period', Mnemosyne 32: 39-69; repr. in [134].
[49] 'Providence and the Destruction of the Universe in Early Stoic
Thought', in M. ]. Vermaseren (ed.), Studies in Hellenistic
Religions, EPRO 78 (Leiden: EJ. Brill), 129- 88; repr. in [125].
[50] review of H. A. K. Hunt, The Doctrines of Zeno the Stoic, Mnemosyne 32: 411-4.

1980
[51] 'The Chronology of Anaxagoras' Athenian Period and the
Date of his Trial. Pt. II, The Plot against Pericles and his
Associates', Mnemosyne 33: 17-95.

422

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

[52] 'Theoretical and Empirical Attitudes in Early Greek Scientific

Medicine', in M. Grmek (ed.), Hippocratica (Paris: Editions du
CNRS), 371-90.
[53] 'Anaxagoras' Other World', Phronesis 25: 1-4.
[54] 'A Pseudo-Fragment of Zeno Stoicus', Hermes 108: 255-8; repr.
in [134].
[55] 'Plato and the Method of Hippocrates', Greek Roman &
Byzantine Studies 21: 341-62.
1981

[56] 'Bad World and Demiurge: A 'Gnostic' Motif from Parmenides and Empedocles to Lucretius and Philo', in R. v. d. BroekM. J. Vermaseren (eds.), Studies in Gnosticism and Hellenistic
Religions. Festschrift G. Qyispel, EPRO 91 (Leiden: EJ.Brill), 261314; repr. in [125].
[57] 'Hesiod and Parmenides in Nag Hammadi', Vigiliae
Christianae 35: 174-82; repr. in [125].
[58] 'Protagoras on Epistemological Obstacles and Persons', in G. B.
Kerferd (ed.), The Sophists and their Legacy, Hermes Einzelschriften 44 (Wiesbaden: Steiner), 38-53.
[59] review of W. Kullman, lnterpretationen zur aristotelischen Theorie
der Natunvissenschaft, Bibliotheca Orientalis 38: 219-25.
[60] 'De metafysika van Aristoteles', in C. A. van Peursen -E. J.
Petersma (eds.), Metafysica (Meppel: Boom), 16-48.
[61] 'De maat en de mensen' (oration on the occasion of the dies
natalis of Utrecht University), Utrecht, The University, pp. 21.
[62] 'Pythagoras en het oude Pythagoreisme', Hermeneus 53: 80-93.
1982

[63] 'The Cleanthes Fragment in Cicero, De Natura Deorum II 24',

in J. den Boeft- A. H. M. Kessels (eds.), Actus. Festschrift H. L.
W. Nelson (Utrecht), 203-10; repr. in [125].
[64] 'Digging up a Paradox: A Philological Note on Zeno's
Stadium', Rheinisches Museum 125: 1-24; repr. in [134].
[ 65] review of W. Scheffel, Aspekte der platonischen Kosmologie,
Mnemosyne 35: 168-72.
[66] review of M. Schofield, An Essay on Anaxagoras, Mnemosyne 35:
360-7.
[67] review of R. B. Todd, Alexander of Aphrodisias on Stoic Physics,
Mnemosyne 35: 388-92.

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

423

[68] art. 'Midden-Platonisten', in Grote Winkler Prins, 8th entirely

revised edition, vol. 15 (Amsterdam-Brussel: Elsevier), 355.
[69] contributions on 'Parmenides', 'Heraclitus', 'Aristoteles', in
]. Bor- S. Teppema, (eds.), 25 Eeuwen Filosofie (Meppel: Boom),
29-40, 57-60 [2nd and 3rd editions in 1985 and 1987].
1983

[70] Die Vorsokratiker, I: Milesier, Pythagoreer, Xenophanes, Heraklit,

Parmenides (Stuttgart: Philipp ReclamJun.), pp. 336 [2nd edition
1986].
[71] 'Intuitionism and Formalism: Zeno's Definition of Geometry
in a Fragment of L. Calvenus Taurus', Phronesis 28: 59-74.
[72] 'Apollodorus on Democritus', Hermes 111: 253-8; repr. in [134].
[73] 'Resurrection Added: The interpretatio christiana of a Stoic
Doctrine', Vigiliae Christianae 37: 218-33; repr. in [125].
[74] 'Techne: A New Fragment of Chrysippus', Greek Roman &
Byzantine Studies 21: 1-1 0; repr. in [ 125] .
[75] 'Cratylus 402 a-c: Plato or Hippias?', in L. Rossetti (ed.), Atti
Symposium Heracliteum 1981, I (Roma: Ed. del Ateneo), 43-55; repr.
in [134].
[76] 'Two Heraclitea in Philo Judaeus', ibid., 63-4.
[77] 'The Historical Hippocrates and the Origins of Scientific
Medicine; Comments onJoly', in M. Ruse (ed.), Nature Animated.
Papers of the Jrd International Conference on the History and
Philosophy of Science, Montreal1980, vol. II (Dordrecht: Reidel), 4976.
[78] review of D. E. Hahm, The Origins of Stoic Cosmology,
Mnemosyne 36: 193-7.
[79] review of A. A. Mosshammer, The Chronicle of Eusebius and
Greek Chronographic Tradition, Mnemosyne 36: 202-7; repr. in [134].
[80] review of P. Mastandrea, Un Neoplatonico Latino. Cornelio
Labeone, Mnemosyne 36: 437-41; repr. in [125].
[81] 'Een brief van Wilamowitz over Nietzsche', Maatstaf31: 11524; earlier version of [104].
1984

[82] 'Zeno and Aristotle on Mixture', Mnemosyne 36: 306-12; repr. in

[134].
[83] 'Heraclitus fr. B 63 D.-I{.', Elenchos4: 197-205; repr. in [125].

424

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

[84] 'Mito Scienza Filosofia: Una Questione di Origini', Quaderni di

Storia 10: 43-67.
[85] review of H. Grensemann, Hippokratische Gyniikologie, Gnomon
56: 385-8.
[86] review of M. R. Wright, Empedocles: The Extant Fragments,
Canadian Philosophical Reviews 4: 136-8.
1985

[87] 'Wilamowitz' Ciceronian Philosophy', in W. M. Calder IIIH. Flashar- Th. Lindken (eds.), Wilamowitz nach funftig jahren
(Darmstadt: Wissenschaftliche Buchgesellschaft), 178-221.
[88] 'Some Stoics on the Soul', Mnemosyne 37: 443-5; repr. in [125].
[89] 'Versions of the Nobody', Mnemosyne 37: 445-7.
[90] . 'On Two Fragments of Heraclitus in Clement of Alexandria',
Mnemosyne37: 447-51; repr. in [125].
[91] 'Aristotle and Others on Thales, or the Beginnings of Natural
Philosophy. With Some Remarks on Xenophanes', Mnemosyne
38: 109-29; repr. in [134].
[92] 'Heraclitus, Empedocles, and Others in a Middle Platonist
Cento in Philo of Alexandria', Vigiliae Christianae 39: 131-56; repr.
in [125].
[93] 'Philosophy and Philosophies: A Reply to Leszl and Vegetti',
Qy,aderni di Storia 11: 177-83.
[94] 'Myth, Science, Philosophy: A Question of Origins', in
W. M. Calder III - U. K. Goldsmith - P. B. Kenevan (eds.),
Hypatia. Festschrift H. E. Barnes (Boulder Co.: Colorado Associated
University Press), 45-65; English version of [84]; repr. in [134].
[95] review of D. O'Brien, Pour interpretrrr Empedocle, Mnemosyne 38:
183-8; repr. in [125].
[96] review of H. B. Gottschalk, Heraclides of Pontus, Mnemosyne 38:
202-8.
[97] review of A. M. Ioppolo, Aristone di Chio e lo Stoicismo antico,
Mnemosyne 38: 209-15.
[98] review of S. Stiidele, Die Briefe des Pythagoras und der Pythagoreer, Mnemosyne 38: 215-7.
[99] review of G. Rocca-Serra, Censorinus. Le jour nata~ Mnemosyne
38:234-6.
[100] review of L. J. Alderink, Creation and Salvation in Ancient
Orphism, Mnemosyne 38: 436-8; repr. in [125].
[101] Absolutisme en relativisme, Haarlemse voordrachten XLV

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

425

(Haarlem: Hollandsche Maatschappij der Wetenschappen),

pp.32.
1986

[102] Die Vorsokratiker, II: Zenon, Empedokles, Anaxagoras, Leukipp,

Demokrit (Stuttgart: Philipp Reclam Jun.), pp. 351; vol. I & II
reprinted with some bibliographical additions in 1987; reprinted
parts (on Anaximander, Heraclitus, Anaxagoras) of this edition
have in turn appeared in K.-H. Hartmann (ed.), Lob der Verganglichkeit. Gedichte, Geschehen und Gedanken (Stuttgart: Reclam), 1993,
28-34.
[103] 'Aristotle, Plato, and the Preplatonic Doxography and Chronography', in G. Cambiano (ed.), Storiografia e dossografia nella
filosofia antica (Torino: Tirrenia Stampatori), 1-59; repr. in [134].
[104] 'Diogenes Laertius on Stoic Philosophy', Elenchos 7: 295-382;
repr. in [134].
[105] 'The Wilamowitz-Nietzsche Struggle: Another New Document and Some Further Comments', Nietzsche-Studien 15: 41-58.
[106] review of F. Decleva Caizzi, Pirrone. Testimonianze, Mnemosyne
39:478-80.
[107] review of D. Pesce, La Tavola di Cebete / J. T. Fitzgerald -L. M.
White, The Table of Cebes, Mnemosyne 39: 484-6; repr. in [125].
[108] review of A. Baumgarten, The Phoenician History of Philo of
Byblus. A Commentary, Mnemosyne 39: 486-8.
[109] review of D. Winston - J. Dillon, Two Treatises of Philo of
Alexandria, Mnemosyne 39: 491-3.
[110] 'Heraclitus fr. 111 D.-K.', Algemeen Nederlands Tijdschrift voor
Wijsbegeerte 78: 234-5.
[Ill] 'De Centrale Interfaculteit', in H. W. von der Dunk- W. P.
Heere- A. W. Reinink (eds.), Tussen ivoren toren & grootbedrijf.
De Utrechtse Universiteit, 1936 - 1986 (Maarssen: G. Schwarz),
492-6.
[112] 'In memoriam C. J. de Vogel', Algemeen Nederlands Tijdschrift
voor Wijsbegeerte 78: 293-4.
1987

[113] 'Historical and Philosophical Aspects of Gorgias' 'On What

Is Not', in L. Montoneri- F. Romano (eds.), Gorgia e la sofzstica,
Siculorum Gymnasium N.S. 38: 243-71; repr. in [134].

426

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

[114] 'Theophrastus and the Xenophanes Doxography', Mnemosyne

40: 286-312; repr. in [134].
[115] 'Number Nine (Diog. Laert. IX 87) ', Revue de Philosophie
AncienneS ( publ. in 1988): 235-48; repr. in [134].
[116] 'Plato over de Vrouw', Algemeen Nederlands Tijdschrift voor
Wijsbegeerte 79: 119-20.
1988

[ 11 7] Het leerdicht en de paradoxen. De fragmenten van Parmenides

en Zeno, vertaald met inleiding en commentaar (Kampen: KokAgora), pp. 166.
[118] 'Philosophy in the Service of Scripture: Philo's Exegetical
Strategies', in J. Dillon -A. A. Long (eds.), The Question of
Eclecticism: Studies in Later Greek Philosophy (Berkeley: University
of California Press), 70-102; repr. in [125].
[119] 'Compatible Alternatives: Middle Platonist Theology and the
Xenophanes Reception', in R. van den Broek- T. Baarda- J.
Mansfeld (eds.), Knowledge of God in the Greco-Roman World, EPRO
112 (Leiden: EJ. Brill), 92-117.
[120] 'Diaphonia: The Argument of Alexander De Fato Chs. 1-2',
Phronesis 33: 182-208.
[121] 'De Melissa Xenophane Gorgia: Pyrrhonizing Aristotelianism',
Rheinisches Museum 131: 239-76; repr. in [134].
[122] review of M.-0. Goulet-Caze, L'ascese cynique, Classical Review
N.S. 38: 162-3.
[123] review of S.-T. Teodorsson, Anaxagoras' Theory of Matter,
GOteborg 1982, Mnemosyne41: 395-7.
[124] review of R. W. Sharples, Alexander of Aphrodisias On Fate,
Mnemosyne 41: 416-9.
1989

[ 125] Studies in Later Greek Philosophy and Gnosticism. A selection of

papers and reviews, with preface, index, addenda, and corrections, Collected Studies Series 292 (London: Variorum Revised
Editions and Reprints), pp. x + 324.
[126] 'An Echo of Middle Platonist Theology in Alexander's De
Jato', Vigiliae Christianae 43: 86-91.
[127] 'The Stoic Cardinal Virtues at Diog. Laert. VII 92', Mnemosyne 42: 88-9.

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

427

[128] 'Fiddling the Books: Heraclitus on Pythagoras (DK 22B129)',

inK Boudouris (ed.), Ionian Philosophy (Athens), 229-34; repr. in
[134].
[129] 'Gibt es Spuren von Theophrasts Phys. op. bei Cicero?', in W.
W. Fortenbaugh- P. Steinmetz (eds.), Cicero's Knowledge of the
Peripatos, Rutgers Studies in Classical Humanities Vol. IV (New
Brunswick N.J.-London), 133-58; repr. in [134].
[130] 'Greek Philosophy in the Geschichte des Altertums', Elenchos 10:
23-60.
[131] 'Stoic Definitions of the Good (Diog. Laert. Vll 94) ',
Mnemosyne 42: 487-91.
[132] 'Chrysippus and the Placita', Phronesis 34: 311-42.
[133] review ofJ.-H. Kiihn- U. Fleischer, Index Hippocraticus. Fasc.
1: A-A. Fasc. II: E-K, Mnemosyne 42: 182-6.
1990

[134] Studies in the Historiography of Greek Philosophy. A selection of

papers and one review, with preface, corrections, addenda, and
indices locorum potiorum and nominum (Assen: Van Gorcum
& Comp.), pp. x + 482.
[135] 'Doxography and Dialectic: The Sitz im Leben of the 'Placita",
including index locorum, in W. H. Haase (ed.), Aufstieg und
Niedergang der Romischen Welt II Vol. 36.4 (Berlin-New York: De
Gruyter), 305~3229.
[136] 'Greek Philosophy in the Geschichte des Altertums', in W. M.
Calder III - A. Demandt (eds.), Eduard Meyer: Leben und Leistung
eines Universalhistorikers, Supplements to Mnemosyne 112
(Leiden: EJ. Brill), 255-84; reprint of [130].
[137] 'Physikai doxai i Problemata physika od Aristotela do Aitija (i
kasnije)', Filosofska Istrazivanja 37 (Zagreb), 1051-78. Croatian
transl. of abridged version of [141] (with summaries in German
English French).
[ 138] 'Physikai doxai and Problemata Physika from Aristotle to
Aetius (and Beyond)', Synthesis Philosophica 5 (Zagreb): 469-97
(abridged version of [141] with summaries in German English
French).
1991

[139] 'Two Attributions', Classical Qy,arterly 41: 541-4.

428

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

1992

[140] Heresiography in Contert: Hippolytus' Elenchos as a Source for

Greek Philosophy, Philosophia antiqua Vol. LVI (Leiden: E.J.
Brill), pp. xvii + 391.
[ 141] 'Physikai doxai and Problemata Physika from Aristotle to
Aetius (and Beyond)', in W. W. Fortenbaugh- D. Gutas (eds.),
Theophrastus: His Psychological, Doxographical and Scientific Writings,
Rutgers Studies in the Classical Humanities, Vol. V (New
Brunswick N. ].-London), 63-111.
[142] 'Heraclitus Fr. B 85 D.-K', Mnemosyne45: 9-18.
[143] 'Physikai doxai et Problemata Physika d'Aristote a Aetius (et audela)', in A. Laks (ed.), Doxographie antique, Revue de metaphysique
et de morale97: 327-64; abridged French version of [141].
[ 144] 'IIEPI KOl:MOY. A Note on the History of a Title', Vigiliae
Christianae 46: 391- 411.
[145] 'Enige opmerkingen over de 'mens' als idea en eidos in het
griekse denken', in G. Quispel (ed.), De Hermetische Gnosis in de
loop der eeuwen (Baarn: Tirion), 176-84.
1993

[146] 'The Idea of the Will in Chrysippus, Posidonius, and Galen',

in]. J. Cleary (ed.), Proceedings of the Boston Area Colloquium in
Ancient Philosophy Vol. VII, 107-145 + 153-7.
[147] 'A Theophrastean Excursus on God and Nature and its Aftermath in Hellenistic Philosophy', Phronesis 37: 314-35.
[148] 'Aspects of Epicurean Theology', Mnemosyne 46: 172-210.
[149] 'Physikai Doxai e Problemata Physica da Aristotele ad Aezio (ed
oltre) ', in A. M. Battegazzore (ed.), Dimostrazione, argomentazione
dialettica e argomentazione retorica nel pensiero antico, Genova, pp.311382; Italian trans!. of revised and amplified version of [ 141].
[150] review of Timone di Fliunte: Silli, Mnemosyne 46: 397-400.
[ 151] review of L. Zanoncelli, La manualistica musicale greca,
Mnemosyne 46: 408-9.
[152] review of J. Dillon, The Golden Chain, Vigiliae Christianae 47:
305-9.
[153] review of H. Dorrie (- M. Baltes), Der Platonismus in der Antike.
Bd. 1-11, Mnemosyne 46: 567-72.
[154] review of M. Tardieu, Les paysages reliques, Mmemosyne 46:
572-5.

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

429

1994
[ 155] Prolegomena. Questions to be Settled before the Interpretation of an
Author, or a Text, Philosophia Antiqua Vol. LVIII (Leiden: EJ.
Brill), pp. vii + 252.
[156] 'A Lost Manuscript of Empedocles' Katharmoi', Mnemosyne 47:
79-82.
[157] 'Epicurus Peripateticus', in A. Alberti (ed.), Realta e ragione.
Studi di filosofia antica (Firenze), 29-4 7.
[158] 'The Rhetoric in the Proem of Parmenides', in L. BertelliP.-L. Donini (eds.), Filosofia, Politica, Retorica. Intersezioni possibili
(Milano: Francoangeli 1994) 1-11.
[159] review of M. Frede - G. Patzig, Aristoteles, Metaphysik Z,
Mnemosyne 47: 120-1.
[160] review of Dexippus, On Aristotle, Categories. Translated by John
Dillon, Mnemosyne 47: 287.
[161] review of Alexander of Aphrodisias, Quaestiones 1.1-2.15.
Translated by R. W. Sharples, Mnemosyne 4 7: 427-8.
1995
[162] 'Critical Note: Empedocles and his Interpreters', Phronesis 40:
109-15.
[163] 'Aenesidemus and the Academics', in L. Ayres (ed.), The
Passionate Intellect. Festschrift I. G. Kidd, Rutgers Studies in the
Classical Humanities, Vol. VII (New Brunswick/London:
Transaction Publishers), 235-47.
[164] 'Papias over traditie', Nederlands Theologisch Tijdschrift 49
(1995): 140-53.
[165] 'Insight by Hindsight. Intentional unclarity in Presocratic
Proems', Bulletin Institute Classical Studies 42: 225-32.
Major works In Progress

-A new edition, with philological commentary, of the fragments

of the Early Stoics, to replace the edition by J. von Arnim,
Stoicorum veterum fragmenta, Vols. I-III, Leipzig, Teubner 1903-05.
Work on this project commenced in 1979.
- Aiitiana: The Method and Intellectual Context of a Doxographer
(together with D. T. Runia). Work on this project commenced
in 1988.

430

A BIBLIOGRAPHY OF JAAP MANSFELD'S PUBLICATIONS

Editorial Work

(1970) Grondslagen van het wetenschappelijk denken (Rotterdam:

Universitaire Pers Rotterdam), pp. 151.
(1974) Lampas 8, fasc. 5 (pp. 318-93) on the Presocratics (contributions in English, German, and Dutch).
(1975) [with L. M. de Rijk] Kephalaion. Studies in Greek Philosophy
and its Continuation. Festschrift C. J de Vogel. Philosophical Texts
and Studies 23 (Assen: Royal Van Gorcum & Comp.), pp. 234.
(1988) [with R. van den Broek & T. Baarda] Knowledge of God in the
Greco-Roman World, EPRO 112 (Leiden: E.J. Brill), pp. 290.
(in progress) [withJ. Barnes & M. Schofield], The Cambridge History
of Hellenistic Philosophy (to appear in 1997).
Dissertations supervised

(1983) A. C. J. Habets, Geschiedenis van de indeling van de filosofie in de

oudheid (Utrecht: The University Press), pp. 223.
(1988) K. A. Algra, Concepts of Space in Classical and Hellenistic Greek
Philosophy, (Utrecht: Department of Philosophy), pp. viii + 199.
(1990) M. Desbordes, Introduction a Diogene Laiirce. Preliminaires
critiques a une lecture de l'oeuvre (Utrecht: Department of Philosophy), Vol. 1. pp. xxii + 357; Vol. 2, pp. 335.
(1992) T. L. Tieleman, Galen and Chrysippus: Argument and Refutation
in the De Placitis, Bks. //-Ill, Quaestiones infinitae Vol. 3
(Utrecht: Department of Philosophy), pp. xxx + 287.
(1993) J. N. M. Baltussen, Theophrastus on Perception. Argument and
Purpose in the De Sensibus, Quaestiones infinitae Vol. 6 (Utrecht:
Department of Philosophy), pp. xxii + 306.
Mfiliations

Member of the Hollandsche Maatschappij der Wetenschappen

Corresponding Fellow of the Real Academia de Barcelona
Fellow of the Academia Europaea
Fellow of the Koninklijke Nederlandse Academie van Wetenschappen
Member of the Erasmus network Ancient Philosophy
Member of the supervising committee Symposium Hellenisticum.

INDEX OF ANCIENT AUTHORS

Aeschines of Sphettus 171
Aeschylus 392
Aetius 7, 24,192,232,280,335,348,
366-379
Alcinous 300-312, 348, 395
Alcmaeon 346
Alexander of Aphrodisias 118,
154, 191-193, 195-197, 200-203,
265-281, 284-299
Alexander of Damascus 266
Alexander of Lycopolis 313-329
Alexis 206
Amafinius 238
Ammianus Marcellinus 141
Anacharsis 28
Anaxagoras 3-19, 24, 337, 346
Anaxarchus 27, 28, 169
Anaximander 7, 169
Anaximenes 169
Anaximenes of Lampsacus 157-58
Andronicus of Rhodes 157
Anniceris 206, 217
Antigonus of Carystus 207
Antiochus 170
Antipater 211
Antiphon 61
Antisthenes 169, 209
Antisthenes of Rhodes 207
Antony 315
Apollodorus of Seleucia 170, 380
Apollonius of Tyre 170-172
Apuleius 216, 395
Arcesilaus 29, 169
Archelaus 169
Aristippus 100, 111, 171, 206-219
Aristodes 21, 212
Ariston of Chios 208
Aristotle 7, 13, 49, 55, 77, 89, 93,
100, 102, 105, 109-100, 115-134,
136, 140, 146, 154-158, 177, 186191,192,199,204,221,224,227,
239,243,247,266,267,269,271,
273-276,279,285,286,288,290,
292-296, 299, 309, 333-342, 347,
357, 367, 370, 372-375, 379-380,
397,405
Aristotle the Cyrenaic 207, 217
Arius Didymus 191, 363-381
Arius of Alexandria 364

Asdepius 337
Aspasius 266
Athenodorus of Tarsus 236
Atticus 82, 84
Augustine 194, 216, 404
Aulus Gellius 193, 198, 203-204, 216
Ausonius 216
Basil of Caesarea 86
Bion 169
Bion of Borysthenes 206
Boethus of Sidon 136
Callias 289-290
Callimachus 206
Calliphon 215
Carneades 169, 215
Catius 238
Celsus 315, 319-323, 327-328
Chamaeleon 206
Chrysippus 169-183, 191-193, 197200,203,253,257,261,267,269270,276-278,366,368,373,375,
377-381
Cicero 21, 23, 32-37, 191, 194, 196198,204,208-217,222,223,236,
238-239, 246, 254, 258, 262, 277,
308,386,395,397-398
Cleanthes 171, 258, 366, 378-381
Clement of Alexandria 100, 197198,280,318
Cleomedes 195
Clitomachus 169, 208
Corpus Hermeticum 326
Critolaus 192
Democritus 7, 8, 22, 24, 27, 32, 36,
37,140,169,239,337,346
Diagoras of Melus 211, 216
Dichaearchus 223-224, 230
Diodes 208
Diodes of Cnidus 208
Diodes of Magnesia 175, 178
Diodorus 169
Diodorus Siculus 140
Diogenes Laertius 22, 27-28, 101,
114, 145-165, 169-183, 207, 208,
212,254,386
Diogenes of Babylon 136, 181

432

INDEX OF ANCIENT AUTHORS

Diogenes of Sinope 169, 171

Diogenes of Smyrna 27
Diogenes of Tarsus 207
Diogenian us 192
Dionysius of Aegae 280-281
Dionysius of Halicarnassus 177
Dionysius the Stoic 171
Dionysodorus 28
Empedocles 24, 237, 239, 240, 337,
340,346,361
Ennius 237, 239
Epicrates 110
Epictetus 243-264
Epicurus 99, 206, 207, 211, 213-218,
220, 227-229, 231-240, 245, 372
Epiphanius 22-25
Eubotes the Cyrenaic 207
Eubulides 169
Eucherius 217
Eudemus 321
Eudoxus of Cnidus 100-101, 169
Eusebius of Caesarea 21, 23, 30-32,
208, 363-381
Eustathius 281
Flavius Boethus 266
Galen 73, 172-73, 191, 194, 196-197,
209,265-281,299,317
Gnostica 327
Gorgias 28
Hegesander 207
Hegesias 206, 210, 217
Hegesippus 318
Heraclides Lembus 208
Heraclides Ponticus 77, 169
Heraclides the Pythagorean 208
Herillus 171
Hermesianax 206
Herminus 266
Hermodorus 156
Herodicus 69-71
Herodotus 223
Hesiod 224, 228
Hierocles 248, 253
Hieronymus of Rhodes 215
Hippobotus 208
Hippocrates 61-75
Hippolytus 25
Homer 224, 226, 239
Horatius 216-218
Iamblichus 102, 106, 267

Idomeneus 207
Irenaeus 318
!socrates 157-58
John Chrysostom 151
John Lydus 377
John Philoponus 86-87
Josephus 314, 325
Julian 319-320, 322-323, 326
Lactantius 216
Lacydes 169
Leucippus 7, 169
Lucian 328
Lucilius 208-209, 216
Lucretius 22, 139, 220-230, 231-240
Lycon 169
Lycus 281
Mani 314
Marcus Aurelius 258
Meleager 208
Meleager of Gadara 208
Melissus 24, 169
Menedemus of Eretria 169
Menedemus the Cynic 169
Metrodorus of Chios 21-37
Mnesarchus 381
Monimus 28
Musonius Rufus 246
Myson 169
Nemesius 192, 274
Nessas of Chios 27
New Testament 319, 321-322
Nikolaus 208
Nilus 163
Numenius of Apamea 274, 301, 317
Old Testament 322-326
Onesicritus 224
Origen 328
Ovid 141
Panaetius 136, 171-72, 208, 252,
366,381
Parmenides 24, 237, 345, 368
Pausanias 141
Phainias 206
Philip the Megarian 206, 207
Philo of Alexandria 135-144, 325
Philodemus 22, 25, 170, 224, 228,
229
Philolaus 142, 368
Photius 280-281, 323

INDEX OF ANCIENT AUTHORS

Pindar 234, 392

Plato 5, 13, 24, 41-59, 61-75, 76-96,
99-101, 105-113, 119, 122, 126,
133, 136, 146, 155-58, 169, 171,
177,201,207,221,224,227,237,
243,247,256,264,266,267,274,
279, 300-312, 337, 345-362, 369370,372,380,385-406
Pliny 141
Plotinus 102, 172, 193, 319
Plutarch 84, 181, 191, 198, 212, 224,
232,253,301,307
Polemo 169
Polybius 141, 221-224, 229
Pomponius Mela 224
Porphyrio 216
Porphyry 87, 126, 267, 320, 327-328
Posidonius 144, 274, 374, 375, 381
Proclus 78, 83-89, 92-96, 390, 395
Protagoras 21, 24, 28, 177, 213, 224,
279
Ps.Aristotle 374
Ps.Caesius Bassus 216
Ps.Philo 324
Ps.Plutarch 198, 348, 366
Pyrrho 21
Pythagoras 24
Quintilian 217, 238
Seneca 141,216,217,227,259,262
Severianus of Gabala 151
Severus 274
Sextus Empiricus 23, 28-30, 181,
194,196,212,257
Simplicius 4-9, 12, 15, 203, 273, 337
Socrates 41-59, 169, 243, 244, 260,

433

286-290, 297, 371, 387-390, 392,

395,397-398,401,404
Sophocles 392
Sosicrates 171-72
Sotion 207, 208, 218
Speusippus 77, 82,99-114, 156
Sphaerus 171
Stobaeus 348, 363-381
Stoics 191-205, 274, 366-381
Strabo 140-141,170, 172
Teles 207
Tertullian 216, 277
Theodoret of Cyrrhus 366
Theodorus 206,208,209,211,216
Theophrastus 7, 105, 135-144, 321,
334-335, 345-362, 376, 379-380
Theopompus 207
Thrasyllus 169, 386-387
Thrasymachus of Calcedon 41-59
Thucydides 223, 224
Timon 206
Valerius Maximus 216, 217
Varro 395, 397
Vitruvius 209, 216
Xeniades 28
Xenocrates 77, 82, 156, 372, 405
Xenophanes 24-25, 28, 36
Xenophon 62, 223, 388, 397
Zeno ofCitium 139, 170-171, 181,
213, 321, 366, 373, 377-381
Zeno of Elea 24
Zeno of Tarsus 380
Zopyrus 290

INDEX OF MODERN AUTHORS

Aalders, G. J. D. 223, 224, 226
Abbott, E. 392, 401
Accattino, P. 267, 271, 276, 292,
294-295
Ackrill, J. L. 115-116, 118, 120,
122, 127-129, 132-133
Adam, A. 313
Adam, J. 53-54
Aleshire, S. 66
Alexander, P. S. 325
Alfieri, V. E. 27, 33-34
Algra, K. A. 374
Alt, K. 77
Althoff, F. 412
Annas, J. 43-44, 46-47, 56-57, 252,
253
Anscombe, G. E. M. 186
Appuhn, C. 34
Armstrong, A. H. 314, 320
Armstrong, D. M. 24
Arnim, H. von 42, 135, 137, 139,
176,195,204,321
Arnold, G. 318
Baeumker, C 76
Bagnall, R. S. 316
Bailey, C. 22, 31, 34, 222, 223, 228,
230,233,234,237
Bain, A. 396, 399, 401
Baldassarri, M. 169-70, 173
Baltes, M. 76, 78, 82-83, 85-88, 9294,96
Baltussen, H. 333, 335, 343, 345,
346, 360, 361-362
Barigazzi, A. 35
Barnes,]. 8, 29, 31, 34,176,190,
194,257,269,336
Bastianini, G. 300
Bauckham, R. J. 322
Bauer, W. 318,328
Baur, F. C. 411, 413-415
Beausobre, I. de 323
Benko, S. 328
Bentham, J. 397, 399-401
Beriger, A. 334, 341
Berlin, I. 404
Bickel, E. 91
Binder, W. 237
Bonhoffer, A. 249

Bonjour, L. 24
Bork, A. 414
Boudreaux, P. 148-150, 162-164
Bousquet, J. 66
Bowersock, G. W. 364
Braun, L. 336
Braun, M. 410
Brinkmann, A. 315
Brisson, L. 77, 172
Brochard, V. 31
Brucker,J.J. 395, 400
Bruns, I. 202, 265, 286
Brunschwig, J. 170, 177, 253
Burnet, J. 8, 104
Burnyeat, M. 22, 26, 29, 269, 279
Butler, W. A. 402
Calder III, W. M. 410-412, 417
Cambiano, G. 68
Campbell, L. 392, 401
Cappelletti, A. J. 31, 33
Carlyle, T. 402
Charlton, W. 188
Chenu, M.-D. 76
Cherniss, H. 77, 87, 89
Christ, G. 155
Classen, C. J. 206, 213
Cohen, S.J. D. 318
Cole, T. 221, 223
Coleridge, S. T. 402
Comte, A. 403
Conche, M. 34, 37
Conze, A. 408
Copleston, F. 401
Cornford, F. M. 6
Costa, C. D. N. 223, 228, 230
Courcelle, P. 316
Cousin, V. 391
Cozijnsen, B. 324
Cronert, W. 25, 170, 208
Cross, R. C. 41, 56
Daiber, H. 144
Dal Pra, M. 34
Dancy, R. M. 99
Deichgraber, K. 70, 280-281
DeLacy, P. 22, 266
Denyer, N. 5
Des Places, E. 31

INDEX OF MODERN AUTHORS

Deuse, W. 274-275
Diels, H. 33, 35-36, 237, 279, 335,
348, 363-381, 410
Dies, A. 104, 107, 111
Diez Macho, A. 324
Dihle, A. 227
Dillon, J. M. 300, 302, 307-308,
325,348
Dilthey, W. 407-417
Dirlmeier, F. 295
Dodd, C. H. 327
Donini, P.-L. 265, 266, 284-285,
292,294-295,298,301
Dorandi, T. 169-70, 226, 228
Doring, K. 104, 206, 207, 212, 217
Dorrie, H. 76, 173
Dowden, J. 402
Dreyer, 0. 323
Dryoff, A. 170, 177
Dumont, J. P. 31, 34
During, I. 104
Easterling, H. J. 77
Ebert, T. 90
Eggers Lan, C. 84, 86
Egli, U. 173
Ehlers, D. 410
Enfield, W. 395
Englert, W. 262
Erler, M. 207
Ernout, A. 22, 36-37
Faber, G. 401
Fakhry, M. 76
Festugiere, A.-:J. 316
Fichte, J. G. 401
Ficino, M. 390, 395
Fortenbaugh, W. W. 144, 376
Fowler, D. P. 230
Frankel, H. 3, 7, 9-11, 13
Frede, D. 188, 295, 299
Frede, M. 194, 199, 265, 274-275
Freudenthal, J. 301
Friedlander, P. 33, 104
Friedrich, C. 70
Fritz, von, K. 336
Furley, D. 3, 5, 9, 11-13
Furth, M. 3, 14, 17
Gaiser, C. 156
Gale, M. R. 223, 226, 236
Gauthier, R. A. 104
Geffcken, J. 326-327
Giannantoni, G. 206
Giannattasio Andria, R. 170

435

Gigante, M. 22, 25-26, 169, 218

Gigon, 0. 42, 50, 55-56, 58, 66, 149158
Giusta, M. 207, 315
Glessmer, U. 324
Glucker,J. 385, 387, 395, 397, 401
Goedeckemeyer, A. 34
Goldsmith, U. K. 407
Goransson, T. 364, 366
Gorgemanns, J. 252
Gorier, W. 214, 215
Gosling, J. C. B. 105, 109-110
Gottschalk, H. 345
Graeser, A. 139
Grant, A. 391
Grensemann, H. 68
Grmek, M. 65
Grote, G. 385-406
Gutas, D. 144, 376
Guthrie, W. K. C. 6, 43, 56, 77, 83
Hackforth, R. 76, 83, 86, 105
Hacking, I. 205
Hadot, I. 6, 329
Hadot, P. 169, 173, 176
Hahm, D. E. 170, 363, 364
Halfwassen, J. 99
Hambruch, E. 155
Hamilton, W. 399, 402
Hamlyn, D. W. 334
Hankinson, R. J. 24-25, 194, 197
Harder, R. 136
Hare, R. M. 405
Harnack, A. von 408
Harper, G. M. 395
Heeren, A. H. L. 366, 373
Hegel, G. W. F. 393-394, 401-403
Heine, H. 391
Heinze, R. 233
Heitsch, E. 93
Heitz, E. 152, 158
Hense, 0. 369
Herrmann, U. 410
Hicks, R. D. 153, 334, 338, 341, 346
Hiller von Gaertringen, F. 407409,413
Himmelfarb, G. 404
Hirzel, R. 227
Hocutt, M. 186
Hoerber, R. G. 52
Hoffmann, R. J. 317
Horner, F. 396
Horst, P. W. van der 313, 315, 316,
321, 322-323
Housman, A. E. 225

436

INDEX OF MODERN AUTHORS

Huby, P.M. 144, 376

Huffman, C. A. 142
Hiilser, K.-H. 169-70, 176
Hume, D. 185
Hussey, E. 25
Huxley, G. 336
Ierodiakonou, K. 178, 181
Inwood, B. 3, 14, 17, 249, 253, 262,

Levy, c. 35
Lewes, G. H. 402
Lieu, S. N. C. 313-314, 319
Lloyd, G. E. R. 67
Long, A.A. 175,195,252,257
Long, H. S. 153
Lorenz, T. 408
Louis, P. 300
Ludlam, I. 53, 405

364

Irwin, T. H. 334
Isnardi Parente M. 22, 99
Jahn, K. 409
Jensen, C. 163
Jolif, J. Y. 104
Joly, R. 68, 71
Jori, A. 68
Jowett, B. 56, 385-406
Kaerst, J. 225
Kahn, C. H. 41-42, 48-49, 345, 364
Kant, I. 396, 401-402
Kardaun, M. 117, 123
Kekule von Stradonitz, R. 407-409,
417
Kerferd, G. B. 41, 56
Kern, 0. 410
Kidd, I. G. 143-144
Kindstrand, J. F. 207, 208
Kirchhoff, A. 409
Kirk, G. S. 3, 7
Kirsten, C. 408
Klaffenbach, G. 417
Knebel, K. L. von 237
Knoepfler, D. 146-148
Kornbichler, T. 408
Kosenina, A. 412
Kozik, I. S. 316
Kramer, H. J. 104, 156
Krenkel, W. 209
Labriolle, P. de 329
Laks, A. 346
Lambinus, D. 212
Landmann, M 77
Langerbeck, H. 22-23, 29, 33, 35-36
Lanza, D. 8
Lattermann, H. 413
Le Boulluec, A. 318
Lebeck, A. 48-49
Lebedev, A. 366
Leo,F. 408,413,417
Leon, P. 15, 18
Lesher, J. H. 25

Macaulay, T. 238
Mackie,]. L. 186, 190, 200
Mackintosh, J. 401
Maguire, J. P. 43-44, 56-57
Maguire, T. 402
Maher, M. 324
Maier, H. 412
Malherbe, A. J. 314
Malingrey, A. M. 314
Mannebach, E. 206
MansfeldJ. 3, 10, 13-19, 21, 25, 32,
34,36,50, 68, 71, 77,83,92, 135,
143-144, 146, 157-158, 169, 174,
178,181,265,266,270,275-280,
284-285,289,294,313,315,321322, 333, 337, 341, 345 366, 371,
387,412
Mansion, S. 333
Marmura, M. E. 266, 267
Mathhias, T. J. 400
Matthen, M. 131
Mau, J. 367
McDiarmid, J. B. 137, 345
Meier, M. 336
Meineke, A. 363, 373
Mejer, J. 170, 335
Meredith, A. 319
Merlan, P. 99, 102
Mill, James 396-406
Mill, J. S. 385-386, 396-406
Misch, G. 407, 409, 411, 415-416
Misch, geb. Dilthey, C. 409, 411,
415-416
Mohr, R. D. 77-78, 83-84, 89
Mommsen, T. 238, 408-412
Mondolfo, R. 22, 31
Montanari, E. 115, 126
Moraux, P. 148-164, 266, 279, 365
Morsink, J. 334
Mosheim, J. L. von 395
Most, G. 346
Mras, K. 379
Muller, G. 44
Munro, H. A. J. 237, 238
Mure, G. R. G. 186

INDEX OF MODERN AUTHORS

Murray, 0. 225
Mutschmann, H. 146-164
Nachmanson, E. 177
Naddaf, G. 77, 84-85
Naldini, M. 316
Nestle, W. 320
Nettleship, R. C. 58
Nicholson, P. P. 53, 56-57
Nohl, H. 410, 414
Nussbaum, M. C. 249
Nutton, V. 66, 265
Oehler, K. 407
Orelli, G. 287
Orlandi, T. 316
Ostenfeld, E. 77
Owen, G. E. L. 405
Pasquali, G. 158
Passmore, J. 335
Peacock,T.L. 396
Pease, A. S. 321
Pembroke, S. G. 252
Pepin, J. 77
Philippson, R. 104
Pines, S. 266, 317
Plasberg, 0. 23
Pohlenz, M. 252
Rackham, H. 23, 34
Raine, K. 395
Ramus, P. 400-401
Raven, ]. E. 3, 6
Reeve, C. D. C. 43, 55, 58
Reeves, J. C. 320, 326
Regenbogen, 0. 239
Reid, ]. S. 23
Reimer, G. 409
Reinhardt, K. 336, 412
Renan, E. 411
Rescher, N. 266, 267
Rijk, L. M. de 119, 121-125, 130133
Rinaldi, G. 326
Roberts, C. H. 313
Robertson, E. S. 402
Robin, L. 22, 36-37, 78
Robinson, T. M. 77, 81, 88
Robson, ]. M. 397
Rohlfing, H. 407
Rorty, R. 335
Rose, V. 153-154, 158, 163-164
Ross, W. D. 115, 156, 190, 341
Rossitto, C. 148-164

437

Rubensen, S. 316
Rugullis, S. 407, 412
Ruijgh, C.]. 313
Runia, D. T. 49, 136-137, 141, 279280,366,374,376-378
Russell, B. 190
Russell, N. 316
Rutherford, R B. 41, 43, 53-54
Ryffel, H. 223
Salmon, W. 186, 204
Sandbach, F. H. 194
Scheffel, W. 76, 78, 80, 86
Schelling, F. W. ]. von 402-403
Schickert, K. 334
Schleiermacher, F. 388, 392
Schmidt, C. B. 400
Schofield, M. 3, 7, 9, 16, 104, 252
Schrader, H. 408-409
Schrijvers, P. H. 224, 226, 229
Schwyzer, E. 91
Scott, D. 307
Sedley, D. N. 31, 135, 141, 195, 252,
300-301, 347, 350
Sewell, W. 402
Sharples, R W. 144, 202, 266, 267,
279, 287-288, 294, 296-297, 299,
300,376
Shorey, P. 22
Shrenk, L. P. 300
Sider, D. 4
Sidgwick, H. 56
Simon, H. 76
Skemp,]. 83
Smith, M. F. 231, 237
Smith, R. 342
Smith, W. D. 71
Solmsen, F. 77
Sorabji, R 17, 76, 83, 186, 188, 295,
329
Souilhe,]. 158, 255
Spiegel, S. 324
Spinoza, B. 394
Spoerri, W. 229
Steenberghen, F. van 76
Stern, L. 409
Stevenson,]. G. 335
Stirling, A. H. 401
Stirling, ]. H. 401
Stokes, M. C. 44, 52, 405
Strang, C. 11, 15, 18
Stratton, G. M. 345-347, 351, 353,
359
Straume-Zimmermann, L. 22-23,
29,35

438

INDEX OF MODERN AUTHORS

Strauss, D. F. 411
Strauss, L. 405
Striker, G. 253, 305
Stroumsa, G. 313, 320
Stuart, D. 399
Suchow, F. G. G. 155
Sudhaus, S. 163
Summers, K. 233
Susemihl, F. 155
Sydenham, F. 400
Szlezak, T. 94
Talon, 0. 400
Taran, L. 77-78, 83, 86-89, 99, 102,
104-105
Taylor, A. E. 104, 355
Taylor, T. 385, 386, 394-396, 400
Temkin, 0. 265
Thiilet, P. 285-287, 294, 296
Thompson, W. H. 402
Throm, H. 280
Tieleman, T. L. 178, 265, 270, 272,
276
Tigerstedt, E. N. 395, 400
Timpanaro, S. 225
Todd,R.B. 26~267,271, 276,279
Traina, A. 218
Usener, H. 195, 208
Vahlen, ]. 408
Van der Waerdt, P. 22, 31, 228
Vegetti, M. 71, 73
Venn,J. 185,204
Villers, C. 400
Viiley, A. 313, 314, 316, 320, 327
Vlastos, G. 6-7, 9-11, 19, 42, 77, 84
Vries, G. de 53
Wachsmuth, C. 366, 374-375, 379

Walbank, F. 221
Wallis, R. T. 320
Walzer, R. 317
Wansink, P. 324
Wardy, R. 334
Warnock, G.]. 401
Waszink, ]. H. 328
Weber, B. C. 217
Weber, H.-0. 221
Wehrli, F. 208, 230
Weidemann, H. 115-120, 124, 126130, 132-133
Weil, E. 333
Weil, R. 336
Welcker, F. C. 417
Wendland, P. 163
Westerink, L. G. 329
Wevers, ]. W. 325
White, N. P. 42
White, S. 201
Whittaker,]. 300, 307, 309
Wieland, W. 119
Wiersma, W. 137
Wiggins, D. 189
Wilamowitz-Mollendorf, U.
von 26, 42, 104, 407-417
Wilke, K. 163
Wilken, R. L. 314, 327
Wilpert, P. 155
Winston, D. 325
Woodruff, P. 52
Woozley, A. D. 41, 56
Yorck von Wartenburg, P. 413
Zeller, E. 22, 31, 77, 135, 155, 364,
407-411, 413-415
Zierl, A. 285, 288
Zoepffel, R. 336