LOGIC AND THE IMPERIAL STOA

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

]. MANSFELD, D.T. RUNIA

]. C. M. VAN WINDEN

VOLUME LXXV

JONATHAN BARNES

LOGIC AND THE IMPERIAL STOA

 LOGIC AND
THE IMPERIAL STOA

BY

JONATHAN BARNES

BRILL
LEIDEN NEW YORK KOLN
1997
This book is printed on acid-free paper.

Library of Congress Cataloging-in-Publication Data

Barnes, Jonathan.
Logic and the imperial Stoa I by Jonathan Barnes.
p. em.- (Philosophia antiqua ; v. 75)
Includes bibliographical references and index.
ISBN 9004108289 (alk. paper)
I. Logic, Ancient. 2. Stoics. 3. Seneca, Lucius Annaeus, ca. 4
B.C.-65 A.D.-Contributions in logic. 4. Epictetus-Contributions
in logic. 5. Marcus Aurelius, Emperor of Rome, 121-18Q-
-Contributions in logic. I. Tide. II. Series.
BC3l.B37 1997
160'. 937-dc21 97-10322
CIP

Die Deutsche Bibliothek - CIP-Einheitsaufnahme

Barnes, Jonathan:
Logic and the imperial stoa I by Jonathan Barnes. - Leiden ; New
York ; Koln : Brill, 1997
(Philosophia antiqua ; Vol. 75)
ISBN 9Q-04-l 0828-9 Gewebe

ISSN 0079-1687
ISBN 90 04 10828 9

Copyright 1997 by Koninklijke Brill, Leiden, 7he Netherlands

All rights rescrved.

No part of this publication may he reproduced,
translated, stored in a retrieval system, or transmitted in any form
or by any TIUans, electronic, TIUcho.nical, plwtocopying,
recording or otherwise, without prior written permission
.from the publisher.

PRINTED IN THE NETHERLANDS

FOR ANTHONY KENNY
 CONTENTS

Preface IX

Chapter One The Decline of Logic ..................................................... .

Chapter Two Seneca ............................................................................ 12

Chapter Three Epictetus ..... .... .. .. .. .. ... .. ... .... .. .. ... .. .. .. .. .. .. .. .. .. .. ..... ..... .. .. 24
A: Logic in the Discourses ... .... ... .. .. .. ... .. .. ....... .. .. .. .. .. .. .................. 24
B: The fashion for logic ................................................................ 33
C: Against logic? .......................................................................... 38
D: Exegesis ................................................................................... 43
E: For logic ................................................................................... 55
F: The place of logic ..................................................................... 62
G: The syllabus ............................................................................. 71
H: Analysis .. .. ..... .. .. .. .. .. .. ... .. .. .. .. ... .. .. .. ..... .. .. .. ......... .. .. .. .. .. .. .. .... .. .. . 77
I: Hypothetical arguments ..... ........... ..... .. ...... ..... .... .. .. .. .. .. .. ......... 85
J: Changing arguments ................................................................ 99

Chapter Four Conclusion..................................................................... 126

Appendix Epictetus, diss I vii .............................................................. 129

Bibliography .......................................................................................... 147

Indexes
Passages............................................................................................. 155
Persons .............................................................................................. 159
Topics ............................................................................................... 162
 PREFACE

This piece began life as a seminar paper. It became a couple of lectures.

The lectures were worked into an article. The article grew, as articles grow.
And it turned into a monograph. A word of excuse might be sought.

First, the title: Logic and the imperial Stoa. Chronologically, the argument
is determined by the overlapping careers of three Stoics: Seneca, who died
in his sixties in 65 A.D.; Epictetus, who was a boy when Seneca slashed his
wrists, and who perhaps lived on into the 130s; and Marcus Aurelius, born
in 121 and ruler of the Empire from 161 to 180.
The same three Stoics determine the content of the monograph
-Marcus introduces the comedy; Seneca features in the second act; and
Epictetus is the hero. Three other figures might have expected to star:
Sextus Empiricus, who flourished in the middle of the second century;
Galen, born in 129 and dead by about 210; and Alexander of Aphrodisias,
whose jloruit is conventionally set at the beginning of the third century. For
all three of these men wrote at length on logic: they were competent
logicians themselves, and they are (so chance has arranged it) sources of the
first importance for our present knowledge of Stoic logic. Nonetheless,
none of the three plays more than a cameo role here; for none of them was a
Stoic-and it was my aim not to set down all that is known and all that can
decently be surmised about Stoic logic under the early empire, but rather to
look at imperial Stoic logic through the eyes of imperial Stoic authors and
in the pages of imperial Stoic texts.
The monograph is offered as a contribution to the ancient history of
logic. I confess that its more logical parts contain little in the way of history
and that its more historical parts say little about the stuff and matter of
logic. The history is in any event external rather than internal; that is to say,
it concerns less the development and career of logic itself than the attitudes
which imperial Stoics took to the subject. I should like to write an internal
history of Stoic logic; but that cannot be done on the basis of the texts with
which this monograph is concerned-if indeed it can be done at all.
X PREFACE

The subject, I fear, will excite scepticism or aversion. Scepticism in

those of my near colleagues who share a passion for long dead logicians.
Aversion in a larger population-for to many, logic seems as arid and
forbidding as the Gobi desert.
To the sceptics I avow that I once was a sceptic myself: the subject
deserves a footnote, certainly; a short article, perhaps; but a monograph? a
whole book? There is too little material, and the material-such as it is-is
gossamer. Well, I took the thing up as a wager. Whether or not the bet was
won, I persuaded myself that the subject was, after all, good for a hundred
pages of print. And I am sanguine enough to hope that some at least of my
fellow fans will be induced to allow that the material has more substance to
it than an initial glance may suggest.
As for the larger population, I have done what I can to make the book
unforbidding. (And unarid-for what author does not hope, and secretly
expect, that his book will seduce more readers than the Bible or the best of
Mills and Boon?) The logical issues which I discuss are none of them
esoteric; and I have discussed them in plain prose, spuming the aid of the
symbolism of modem mathematical logic. (The Stoics themselves, after all,
did not develop any symbolism: like the Peripatetics, they used Greek-a
semi-technical and semi-formalized Greek-and any thoughts which a
modem interpretation ascribes to them should in principle be capable of
comprehensible expression in more or less ordinary Greek, and therefore in
more or less ordinary English.) Some people, to be sure, have no taste for
logic at all; and for them I can do nothing. But if you are not disgusted by
the very idea of a syllogism, then you will find nothing in this book to fox
you. (Save, perhaps, the several errors which I have made and overlooked.)
I add that the body of the text presupposes only the most minimal
background knowledge. I expect readers to know that Chrysippus came
after Aristotle, A.D. after B.C. Any other, more recherche, historical
information which is needed to understand the argument is (in principle, at
least) supplied in the text. Again, although some footnotes assume a
knowledge of Greek (and of Latin), the text itself does not: everything
foreign is done into English.
All this has made the book longer than it would otherwise have been.
Moreover, I have been prodigal in quoting ancient texts. Many readers-if I
may generalize from my own idle habits-rarely look a reference up (and
when they do so, they are often surprised or disappointed). If a text matters
to the argument, then it is best to cite it. And if this pads out the discussion,
then let there be padding.
 PREFACE XI

The paper from which the monograph began was given some years ago to
an Oxford seminar. There were the usual sharp comments at the time; and
afterwards Miriam Griffm sent me a sheaf of notes which corrected various
errors and supplemented various lacunae. The lectures were given in
Florence in 1995, under the auspices of the Erasmus programme. I am
grateful to my Florentine audience for their intelligent questions and
pertinent remarks. And I am particularly grateful to Antonina Alberti, who
organised the affair.
I have accumulated several other debts, most of which (such is the way
with debts) I no longer remember. But I am sure that I have stolen several
things from two papers by Susanne Bobzien; and I know that I am indebted
to two Genevan colleagues, Maddalena Bonelli and Ben Morison, who
discussed certain crucial texts with me, producing a number of new
suggestions and preserving me from a number of old blunders. In addition,
Robert Dobbin, who is preparing a commentary on the first book of
Epictetus' Discourses, was obliging enough to read a draft of the whole
piece; and Gudrun Tausch-Pebody gave me valuable bibliographical
support at a late stage in the production. The editors of Philosophia Antiqua
helped me to remove several errors from the penultimate version; Paul
Slomkowski generously detected some fifty errors in what I had taken to be
the fmal version; Job Lisman, of E.J.Brill, guided me with patience and
understanding through the tedious process of preparing camera ready copy.
Finally, I thank Sylvie Germain, mistress of the Genevan Philosophy
Library, who does not blench at my most exotic requests.

JB
Geneva
January 1997
 CHAPTER ONE

THE DECLINE OF LOGIC

Marcus Aurelius did not much care for logic. He thanked Rusticus
-Quintus Iunius Rusticus, statesman and Stoic-that he had 'not been
diverted into an enthusiasm for sophistry or into writing books about
theorems or delivering protreptic little arguments' (I vii 2); and he thanked
the gods themselves that 'when I longed for philosophy I did not fall in with
a sophist or sit down and write books 1 or analyse syllogisms-or spend my
time on meteorology' (I xvii 22). After all, he reflected, you will no more
fmd happiness in a deduction than in riches or reputation (VIII i 5).
Now Marcus was a philosopher, and a Stoic philosopher. 2 Philosophy
-and in particular Stoic philosophy-divided itself into the three
disciplines of logic, physics, and ethics. 3 But Marcus would love neither
physics nor logic, he would be neither q>UcrtK:6c; nor OtaAEK:'ttK:6c; (VII lxvii
3): ethics was his oyster.
He had, it seems, once felt the allure of the syllogism. At any rate, when
his tutor Fronto feared that the young man was abandoning rhetoric for
philosophy, it was logic in particular which seemed to be the fascination
and the danger (eloq iv 5 [p.149 van den Hout2]). Fronto did his best: after
all, he asked, 'is there anything you retain from those logicians of yours? Is
there anything you are pleased that you retain?' (eloq ii 17 [p.144 van den
Hout2 ]). Logic is a villainous study, done in villainous Greek( 13 [p.141]);
and Fronto called as witness his friend Dionysius, a rhetorician who had
once composed, against the logicians, a witty fable on the vine and the

I Reading ',;oi) ouyypacpEtv' (Reiske ): ',;ob<; ouyypacpEt<;' cod d.

2 On Marcus' Stoicism see e.g. Brunt [1974]; Rist [1982]. - References to modern
literature are normally given by author's name, date of publication, and page number: full
details in the Bibliography. For editions, commentaries, and translations a name and an 'ad
loc' (explicit or implicit) are usually enough (where such references also carry a date, it is
preceded by an asterisk -and the work should be sought in sections (i) and (ii) of the
Bibliography). In references to ancient authors titles are abbreviated in a fashion which I hope
is standard-or at least readily intelligible. When only a single work by an author survives, I
generally dispense with a title.
3 A sample of texts in Hiilser [*1987], frags 14-26; see, recently, Ierodiakonou [1993];
Dorrie and Baltes [* 1996], pp.205-231. - The origins of the tripartition were disputed in
antiquity; but the thing became a commonplace: philosophiae Ires partes esse dixerunt et
maximi et plurimi auctores (Seneca, ep lxxxix 9-the chief exceptions were the Peripatetics
[see below, p.20]).
2 CHAPTER ONE

holm-oak. 4 Pronto's fears were not idle, and he lost his pupil to philosophy;
but neither perhaps were his urgings altogether in vain-Marcus kissed
goodbye to logic. 5
And in this respect at least-or so it is often supposed-Marcus was
representative of his epoch. By the early imperial period, Stoic philosophy
had contracted into ethics: physics and logic were forgotten or ignored.
Philosophy, after all, was the science of how to live, the ars vitae; 6 and a
philosopher was expected to spoon out practical advice. Half a century or
more before Marcus' reign, Dio Chrysostom, the celebrated orator and
bellettrist, once found himself taken for a philosopher. 7 The reasons were
wholly external: destitute and in exile, he looked and behaved like a tramp.
Convinced that sartorial neglect was a sign of spiritual elevation, 8

many people came up and asked me what seemed to me good or bad, and I
was obliged to reflect on these matters in order to be able to answer their
questions. Then they urged me to stand up in public and speak; and so it
became necessary for me to speak about what is appropriate for men and what
seemed to me likely to benefit them. (oral xiii 12)

That is to say, if people took you for a philosopher, they took you for a
moral guru. 9
If philosophers were taken for gurus, that was because philosophy was
taken for ethics. Moreover, the ethics for which it was taken was not a
particularly theoretical thing: it preferred protreptic to argument, it pre-
ferred exhortation to analysis-it was a high-brow moralizing. So, above
all, in the Stoa: in the Meditations of Marcus, in the Discourses of
Epictetus, in the letters and essays of Seneca. Three Stoics of very different
background and status and character; three bodies of writing very different
in style and in nature and in purpose; and in all three a common

4 e/oq v 4-5 [p.l52 van den Hout2]; on Dionysius see also ad Caes IIi 3 [p.l7 van den
Hout2].
5 But in Book I of the Meditations, his record of spiritual debts, Marcus' brief sentence on
Fronto (I xi), makes no mention of any intellectual matter at all.
6 E.g. Sextus, M X! 200; Plutarch, quaest conviv 613B; Clement, paed II xxv 3; further
references and discussion in Lakmann [1995], p.26. -The main thesis of Hadot [1995] has it
that philosophy, in the ancient world, was always 'above all a way of life' (p.l9): true, I
suspect, only if the phrase 'way of life' is construed in the most generous (or vacuous) fashion
possible.
7 Dio had studied philosophy under Musonius Rufus (Fronto, e/oq i 4 [p.l35 van den
Hout2]) -whom he later attacked (Synesius, Dio 37). The story of his subsequent
'conversion' to philosophy is a fable (see Moles [1978]); and his intellectual place is in the
literary and rhetorical camp, not among the philosophers. A convenient summary in Desideri
[1994].
8 A common enough conviction, to judge from Epictetus, diss IV viii 5-8.
9 Nothing new about that: see, for an early example, Empedocles, B 112.8-12 Diels-
Kranz.
 THE DECLINE OF LOGIC 3

determination to moralize-a century and more of preaching, which

different tastes will judge inspiring or dispiriting, profound or pretentious.
Such a view of imperial philosophy, and in particular of imperial Stoicism
is, I take it, a commonplace. Ancient historians debate the philosophical
opposition to the bad emperors and discuss the philosophical companions of
the good. They generally and tacitly suppose that philosophy was more or
less equivalent to ethics. After all, if it is philosophical ideas which
encourage you to oppose a regime or a ruler, surely those ideas will derive
from ethical or political philosophy? And if a philosopher is to sustain a
statesman with useful admonition and salutary advice, then, once more, it is
the ethical part of philosophy which will count. To be sure, it does not
follow that a philosopher who loathed or loved an emperor had little time
for logic himself; but it would not be surprising if logic was at best one of
his background interests.
As for historians of logic, they do not dwell long on the names of Seneca
and Epictetus and Marcus. A few texts of these authors are cited in the more
comprehensive handbooks 10-and any treatment of Stoic logic must quote
the celebrated passage of Epictetus on the Master Argument of Diodorus
Cronus. 11 But such texts appear in the guise of 'fragments'; that is to say,
they are adduced not to celebrate the logical opinions of their authors but to
plump out a history of earlier Stoic logic.
Indeed, Stoic logic is often presented precisely as the logic of
Chrysippus-or at any rate, as the logic of the Hellenistic Stoa. And for two
diverse and excellent reasons. First, it is undeniable that Chrysippus was the
greatest of the Stoic logicians; that he not only determined the programme
for Stoic logic but also carried out a major part of it himself. In antiquity he
was the Stoic logician par excellence-it was Chrysippus' logic which the
gods would use had they any need for logic. 12 The surviving catalogue of
Chrysippus' logical works gives us some rough idea of the scope and detail
of his logical interests; 13 and the surviving testimonia to Stoic logic show
very little which cannot be accommodated within those interests. Stoic logic
was not identical with Chrysippean logic; but Chrysippus surely stood to
Stoic logic as Aristotle stood to Peripatetic syllogistic.
Secondly, the state of our sources does not allow us to say much about
any development which Stoic logic may have undergone. There are, it is
true, a few texts which explicitly ascribe arguments or theses to later
logicians; and of the numerous texts which speak generally of 'the Stoics',

10 Thus Hiilser [* 1987] cites one text from Marcus, 18 from Seneca, 25 from Epictetus
(by way of comparison: about I 00 from Cicero).
11 See below, pp.76-77.
12 Diogenes Laertius, VII 180; cf e.g. Cicero, orat xxxii 115; Dionysius of Halicamassus,
comp verb 31; Clement, strom VII xvi 101; see e.g. Frede [1974a], pp.26-29.
13 Diogenes Laertius, VII 189-202; see Barnes [1996].
4 CHAPTER ONE

some at least may plausibly be imagined to refer to the state of the art after
the death of Chrysippus. But we are not yet in a position to write an
interesting history of Stoic logic; and a chronicler of the subject will
reasonably prefer a systematic to an historical account-and will
reasonably pretend that the system belongs, for the most part, to Chrysippus
himself.
I have no quarrel with those ancient historians who ignore physics and
logic when it comes to the machinations of empire; nor shall I insinuate that
it was reflection on the Master Argument which ultimately led Thrasea
Paetus to court an honourable death, or that Seneca whispered syllogisms
into Nero's young ears. But it is one thing to allow an historian of political
life to concentrate his gaze upon the ethical part of philosophy: it is another
thing to concede that the ethical part was the only part which was of any
account in the social and intellectual dramas of the time.
I have no quarrel with those expositors of Stoic logic who think
primarily-or even solely-of the Old Stoa; nor shall I suggest that Seneca
regenerated the study of hypothetical arguments or that Epictetus produced
a novel theory of logical analysis. But it is one thing to allow that the best
exposition we can now give of Stoic logic is an account of the logic of
Chrysippus, and to agree that Chrysippus was responsible for by far the
greater part of the subject: it is another thing to concede that after the
second century B.C. the subject was abandoned, or neglected, or regarded
as a fossil.
'It is one thing-it is another thing': the historian of the Empire and the
historian of logic will no doubt accept the distinctions on the theoretical
plane. Why should they accept anything more? Does 'another thing' mark
more than an unreal possibility?
Well, perhaps the empire knew a few Stoic logicians? Sergius Plautus,
whom Quintilian singled out as the best Latin author who had written on
Stoicism, certainly touched on logical matters insofar as he interested
himself in terminology. 14 Athenodorus of Tarsus, who tutored the young
Augustus (and who is thus an honorary member of the Empire), wrote
something on Aristotle's CategoriesY The Egyptian Chaeremon, who
tutored the emperor Nero, wrote something on conjunctions-and

14 The Stoic Plautus ofQunitilian Xi 124 is surely the same as the Plautus of !I xiv 2 (and
of III vi 23}, and hence is the linguistic innovator referred to at VIII iii 33-where Detlefsen's
'Sergio Plauto' is therefore the right reading ('Sergio Flavo' cod d). And Qunitilian 's Plautus
is surely the Sergius who, according to Apuleius, int i, 190.4-5 Moreschini, used the word
'effatum' for propositions or orationes pronuntiabiles. Sergius Plautus is generally dated to
the first century A.D.; but in truth all our evidence is consistent with a date in the first century
B.C.
15 Entitled 'npoc; tac; "AptcrtotA.ouc; KatT]yopiac;' (Simplicius, in Cat 62.25); see
Goulet [1989d ].
 THE DECLINE OF LOGIC 5

syntactical matters are close to logical matters. 16 Lucius Annaeus Comutus,

friend of the poets Lucan and Persius and author of a tedious little book on
Stoic theology, also wrote-like Chaeremon-on grammar; and in add-
ition-like Athenodorus-he produced a work on Aristotle's Categories. 11
Perhaps a few other names might be promoted. 18 After all, the early
imperial period was thick with philosophers, and in particular with Stoic
philosophers. We know a little about some of these Stoics-about Attalus
and Euphrates and Hierocles and Apollonius of Chalcedon and Cleomedes.
But most of them are no more than names. And there were undoubtedly
many others who are no longer even shadows of names. It is perfectly
possible that one or another of these lost Stoics had a passion for logic. 19

16 See Apollonius Dyscolus, con} 248.1-13. On Chaeremon see Frede [1989]; Goulet
[1994b].
17 See e.g. Porphyry, in Cat 86.23-24; Simplicius, in Cat 62.28 (Comutus wrote 1tpoc;
'A911v6oropov Kat 'AptcrtotH.,v).- Note also POxy 3694: a work llEpt EKtrov, of which only
the title survives. On Comutus see Most [1989]; Gonzalez [1994].
18 Aristocles ofLampsacus (below, p.73), Herophilus (below, p.73 n.l99)?
19 Apart from Arrian, and the Stoic poets (Lucan and Persius and Silius Italicus), and the
Stoic 'opposition'-Thrasea Paetus, Rubellius Plautus, Barea Soranus [see below, p.41],
Q.Paconius Agrippinus, Iunius Rusticus, Helvidius Priscus, Herennius Senecio, Q.Iunius
Arulenus Rusticus, Iunius Mauricus, Plautius Lateranus [?-see Epictetus, diss I i 19]-for
none of whom is it plausible to imagine logical investigations, there might be mentioned:
Seneca's fellow-student Claranus (ep lxvi I); Iulius Canus, vir in primis magnus (tranq xiv 1-
10); Metronax (ep lxxvi 4); Serapio (ep xl 2); Chaeremon's pupil Dionysius (Suda, s.v.
t.wvilcrtoc; 'AA.E~avOpEilc;); Comutus' pupils Claudius Agathinus [?] and Petronius
Aristocrates (vit Pers 5); Plutarch's friends, the poet Sarapion (def orac 402F, with SEG
XVIII 225) and Themistocles (quaest conviv 626E) and Philip (quaest conviv 710B);
Musonius' pupils, Dio and Timocrates and Athenodotus (Fronto, eloq I iv [p.135 van den
Hout2]), and Artemidorus (Pliny, ep III xi 5), and Pollio (Suda, s.v. nroA.irov), and Lucius,
who made his master's discourses public (Stobaeus, eel II xv 46); Timocrates' pupil Lesbonax
(Lucian, sa/tat 69 [305]); Decianus the friend of Martial (I 8), and Fronto (XIV I 06);
Coeranus (Tacitus, ann XIV lix 2) and P.Egnatius Celer (Tacitus, hist IV 10, 40); various
teachers of Marcus Aurelius----Cinna Catulus, Claudius Maximus, Sextus of Chaeronea (HA:
vMarci iii 1-2), perhaps Diognetus and Bacchius and Tandaris and Marcianus (I vi-no
reason to identify this Bacchi us with the Platonist recorded in SIG3 868B), and also Basil ides
(Jerome, chron ad Olymp 232); Philopator (Galen, an mor V 41 Kiihn; Nemesius, nat hom
xxxv [p.291 Matthaei]); Nestor ([Lucian], macrob 21 [223]); Antibios and Eubios of Ascalon
(Stephanus of Byzantium, s.v. 'AcrKaA.rov); Heliodorus (scholiast to Juvenal, I 33); Theo of
Alexandria (Suda, s.v. Elrov 'AA.E~avOpEilc;); Annius, Athenaeus, Herminus, Lysimachus,
Medius, Musonius, Phoebio, Themistocles (Longinus, apud Porphyry, vit Plot 20); Trypho
(Porphyry, vit Plot 17); Callietes (Porphyry, frag 408 Smith = Eusebius, PE X iii I); the
Christian Pantaenus (Eusebius, HE V x); and, from inscriptions, the Stoic Otaooxot
T.Coponius Maximus, Aurelius Heraclides Eupyrides, and Julius Zosimianus (IG IF 3571,
3801, !0046a, 11551--on otaooxm see e.g. Smith [1996], pp.l25-127); L.Petilius Propas
(Inschriften von Olympus no. 453); Theoxenus (IG IF 10046a); T.Ciaudius Alexander and
C.Tutilius Hostilianus (CIL VI 9784, 9785); Cn.Artorius Apollonius (SEG XVII 467);
Meleager and Aelius Aelianus (Bulletin epigraphique 1958, no. 84); T.Avianius Bassus
Polyaenus (lnschriften von Prusa ad 0/ympum, no. 18). See e.g. Zeller [1909], III i, pp.606-
608, 711-717; Tod [1957]; Pohlenz [1970], I pp.276-290; II pp.l42-148; Hahn [1989],
pp.l22-131.
6 CHAPTER ONE

So indeed it is. But the historian will observe, and rightly, that for none
of them is a passion for logic actually attested. And history (as a few
historians continue to believe) is a matter not of interesting possibilities but
of documented facts.
A side glance at the other philosophical schools is appropriate. For if the
other schools had their logicians, they were not numerous and apparently
they made little splash. The Peripatetics continued to discuss Aristotle's
syllogistic. 20 Alexander of Aphrodisias, whose commentary on Aristotle's
Prior Analytics survives, had predecessors whose works are lost but who
certainly did not neglect logical issues. Alexander himself was no logical
ninny21 -nor was he averse to controversy; 22 moreover, he was a public
figure who held an imperial chair at Athens and dedicated one of his works
to his imperial patrons. 23 Nonetheless, it is clear that he was a philosopher's
philosopher-and clear, too, that logic was not his chief concern.
As for the Platonists, the Didaskalikos of Alcinous contains a few pages
on logic. 24 But they are desultory enough-and Alcinous spends more space
on etymology than he does on any serious logical topic. There is also
Apuleius-if indeed the surviving de Interpretatione is his. 25 But the work
is no more than an elementary introduction. The Epicureans? They had
never been enthusiastic logicians, 26 and there is no reason to think that the
imperial Epicureans discovered a hidden interest in the subject: the vast
inscription which Diogenes set up at Oenoanda in the 120s contained a
section on physics and a section on ethics-nothing on logicY The
Pyrrhonian sceptics? They were few, and of little significance. And
although Sextus Empiricus busied himself about logic, 28 he can hardly be
said to have promoted the subject.
The eclectic Galen also demands mention. Marcus knew and admired
him. 29 His was a prominent profile, and his works were widely read in his
lifetime. And he was a gifted logician-the third logician of the ancient
world after Aristotle and Chrysippus. Yet it may be doubted if he was read
for his logic. Indeed, he himself complained, with frequency and acerbity,

2 On later Peripatetic logic see Lee [1984]; cf e.g. Barnes [1986].

2! See e.g. Barnes et a[ [* 1991]; Flannery [1995]. - On Alexander in general see Sharples
[1987]; Goulet and Aouad [1989].
2 2 On the controversy with Galen see e.g. Rescher and Marmura [*1966]; Zimmermann
[1976]; Nutton [*1979], p.l89; Tieleman [1996].
23 Seefat 164.14-16.
24 didask 5-6 [pp.l56-160 Hermann]; see Dillon [*1993], pp.72-76; Gtiransson [1995],
pp.114-118.
25 On Apuleius and logic see especially Sullivan [1967]; cf Hijmans [1987].
26 See below, pp.9-10.
27 See Smith [*1993].
28 See below, p.l27.
29 At least, Galen says that he did: 'first among physicians, unique among philosophers'
(praecogn XIV 660 Kuhn; cfPhiloponus, aet mund 599.24-26).
 THE DECLINE OF LOGIC 7

that his contemporaries neither knew nor loved the subject: his fellow
doctors would not touch it; his fellow philosophers had not bothered to
study it; and as for the general public, to most of them logic seemed a
thoroughly trivial business-you might as well spend your time drilling
holes in rnillet-seeds. 30
If the imperial Stoics cared little for the logical part of philosophy, they
were not idiosyncratic in their attitude. The age, it seems, was not in the
logical vein.
Nor, of course, need this astonish. Philosophy, in general, has often
enough been despised. Under the Empire there were outbreaks of hostility
towards philosophers, in various quarters and at various times; 31 and no
doubt many Romans were content to recite the verse ofEnnius: 'It's enough
if a few do philosophy-in general, it's not very pleasing [philosophandum
est paucis; nam omnino haud placet]' .32 As for logic in particular, it has
rarely been a popular pastime; and those few laymen who like it treat it
rather as a divertissement than as an object of weight and moment.
Even philosophers have been known to disown the subject. The branch
of charlatanry which the English gallantly call 'continental philosophy'
knows nothing of it; and -that most English of philosophers, John Locke,
saw no profit in syllogistic-any washerwoman could reason as
competently as Aristotle. In the ancient world such attitudes were not
unknown. The Cynics-so we are told-rejected logic. The Cyrenaics
rejected logic. The Epicureans rejected logic. 33 Even some Stoics rejected
logic: Aristo of Chios, Zeno's wayward pupil, allegedly had no truck with
the subject; 34 and Zeno himself-our evidence is meagre-once said
something uncomplimentary about it. 35 Logic was not always loved in
antiquity: Seneca, Epictetus, and Marcus were, it seems, heirs to a tradition.

30 See praecogn XIV 605 Kuhn.- On Galen as a logician see Barnes [1991]; [1993a];
[1993b]; Hillser [1992].
31 For expulsions of philosophers see Epictetus, diss IV viii; Athenaeus, 609D-612F;
Tacitus, ann XV 71 (Nero); Pliny, ep III xi 1-3, with Sherwin-White, ad foe; Dio Cassius,
LXVI xiii 1-2 (Vespasian); Dio Cassius, LXVII xiii 1-3 (Domitian). - In general see e.g.
Friedlander [1920], III, pp.249-266; Brunt [1994], pp.39-40.
32 frag 28 Jocelyn (from the Andromacha): cited by Gellius, V v 9, thrice alluded to by
Cicero -a locus classicus (Tacitus, Agricola iv 3).- Compare Epictetus, diss I xxii 18: 'an
old man, his hair grey and his fingers dripping with gold rings, will come along and say: 'My
boy, of course you need a bit of philosophy-but you need a brain too [liEt f.l.EV Kat
q>tl..ocroq>EtV, lid li Kat eyKeq>al..ov EXEtV ]".
33 See e.g. Diogenes Laertius, VI I 03 (Cynics); Seneca, ep lxxxix II (Epicurei ...
rationalem [= AO"ftK!lv] removerunt); 12 (Cyrenaici naturalia cum rationalibus
sustulerunt).
34 See e.g. Seneca, ep lxxxix 13; Sextus, M VII 12; Diogenes Laertius, VII 160 (cf
Ioppolo [1980], pp.63-69). - Note that Marcus had read Aristo, and liked him: Fronto, ad
Caes IV xiii 2-3 [pp.67-68 van den Hout2].
35 'Zeno corilpared the arts of the logicians to correct measures which measure not grain
or anything useful but chaff and dung' (Stobaeus, eel II ii 12). See below, pp.67-68.
8 CHAPTER ONE

A little more precision is desirable. For crude sentences of the form 'x
rejects logic' are unsatisfying, on two counts.
First, 'rejection' took more forms than one. You might dismiss logic
-as a pseudo-subject (like astrology), or as a subject incapable of scientific
study (like rhetoric), or as a subject unworthy of serious attention (like
athletics). Or you might rather disparage logic-as a subject of little
importance, or as at best a distraction from matters of substance. 36 Different
attitudes-but with the same consequence: a rational man will not devote
his time to logic.
In principle, you might 'reject logic' in a different sense: you might
abjure rational argument altogether. And I should perhaps therefore insist
-something which is banal in the extreme but which has been missed by
some commentators-that it is one thing to decline to study the subject of
logic, and quite another to decline to produce arguments. No ancient
philosopher is accused of abjuring reason; and even if the Epicureans, say,
in some fashion 'rejected logic', that did not stop them from producing
arguments for their doctrines, like any other decent group of philosophers.
As Locke waggishly put it: 'God has not been so sparing to Men as to make
them barely two-legged creatures, and left it to Aristotle to make them
Rational'. You may-it is a brutal fact-argue, and argue well, without
ever having opened a text-book on logic.
Secondly, and more importantly, the word 'logic' requires a gloss.
Hitherto I have chucked the term around: henceforth I shall use it in what I
take to be its standard modern sense. This sense is doubtless vague. But
anyone who has a smattering of English has a firm enough hold on the
word's meaning.
Historians of philosophy often use the word 'logic' to name one of the
three parts into which the ancients customarily divided philosophy. Here the
word takes leave of its modern sense and simply serves as a transliteration
of the Greek word 'A.oytKf]'. It thus picks out, quite generally, the study of
whatever falls within the province of reason or A.6yoc;. 'Logic', in this sense,
was customarily divided by the Stoics into rhetoric and 'dialectic' _37
Rhetoric has, thank God, little to do with what a modern logician thinks of
as his patch; 38 and, if only for this reason, A.oytKTJ is not to be identified
with logic. 'Dialectic' brings us closer to logic. But 'otaAEK'ttKTJ' is a
broader term than the modern 'logic'. For dialectic itself was standardly
split into two parts, one of which studied 'signifiers' and the other 'things

36 And an ancient report that 'X rejects Y' may sometimes, I suspect, merely reflect the
fact that X apparently devoted no attention toY.
37 For the several 'divisions' ofA.oyuci] see Hiilser [*1987], pp.LXXVII-XC.
38 But ancient theorists, from Aristotle onward, usually found the need to discuss various
logical matters in their treatises on rhetoric.
 THE DECLINE OF LOGIC 9

signified', and the former part included, for example, the study of sound
and voice-psychological and physio-logical matters which, by our lights,
are no concern of a logician. As for the study of 'things signified', that
included (at least, in some versions of the division) the study of impressions
or cpavw.criat-that is to say, it occupied itself with many of the issues
which we should now incline to classify under the rubric of epistemology.
But in addition, and perhaps primarily, the study of 'things signified'
treated what the Stoics called AEK'ta or 'sayables'; and among sayables are
'assertibles' or a~Hn~a'ta. A sayable is what we can say by uttering some
linguistic expression; and an assertible is what we can assert by uttering a
special sort of linguistic expression. Thus by uttering the expression
'Chrysippus was a Stoic logician' I may-if all the circumstances are right
-assert something, namely that Chrysippus was a Stoic logician. And that
Chrysippus was a Stoic logician is an assertible. 39 Now assertibles are
quintessentially logical items-they are the primary bearers of truth and
falsehood, and the immediate components of normal arguments. It is thus in
the study of Stoic sayables, and in particular in the study of assertibles, that
we fmd most of what we now think of as pertaining to logic.
When an ancient philosopher-or an ancient layman-is said to have
rejected 'logic' what exactly is he supposed to have been rejecting? Our
sources provide no uniform answer to this question. Frequently enough,
they suggest that what is being rejected is A.oytKt,, the part of philosophy.
Thus the Cynics 'determine to abolish the logical and the physical area, like
Aristo of Chios, and to attend to the ethical area alone' (Diogenes Laertius.
VI 103). As for Aristo, 'he does away with the physical and the logical area,
saying that the one is beyond us and the other nothing to us' (Diogenes
Laertius, VII 160). And among the Cyrenaics, 'Meleager . . . and
Clitomachus ... say that they think that the physical and the dialectical parts
are useless' (Diogenes Laertius, II 92). But it is hard to think that any
philosopher could have been prepared to abandon everything which
traditionally sheltered under the generous umbrella of A.oytKt,; and it is not
implausible to discount the reports as jejune summaries or polemical
exaggerations.
In some cases, at least, it is positively plausible. Thus Sextus remarks
that 'some say that Epicurus rejects logical theory' (M VII 15); but he
immediately adds that others denied this. Seneca observes that the
Epicureans found that they could not do without certain parts of A.oytKt,,
which they treated as an annexe to physics; and both Seneca and Cicero

39 See further below, pp.99-102.

10 CHAPTER ONE

give compact accounts of those aspects ofA.oytKlJ which the Epicureans did
in fact cultivate. 40 A similar story may be told of the Cyrenaics. 41
Sometimes our sources suggest more specifically that the rejected object
was dialectic; but even dialectic includes a number of items which no
philosopher is likely to dismiss-and which the Epicureans certainly did
not dismiss. Should we conclude that, often at least, what the ancient
opponents of logic were rejecting was precisely logic, that it is what we
mean by 'logic' which was being ousted from the philosophical nest? Even
this conclusion is, in general, too strong. For again, some Epicureans
certainly took some interest in some parts oflogic.
Then what was rejected? Many of the ancient enemies of 'logic'
evidently had a particular loathing for the logical puzzles and conundrums
which were a celebrated-or notorious-part of ancient logical studies;
and in some cases at least I suspect that a 'rejection of logic' amounts to no
more than an unwillingness to discuss the enigmas of the Liar or to shovel
away at the heap of the Sorites. If such was your attitude, then you could
reject logic while retaining an interest in the greater part of the subject
-you could 'reject logic' without rejecting logic.
In short, the 'tradition' in which Marcus and Epictetus and Seneca have
seemed to stand is a wraithlike thing. Many philosophers had certainly
dismissed from their thoughts certain subjects which belong to what we
think of as logic; many philosophers had no doubt disparaged some parts at
least of the subject. But it may be doubted whether logic as a whole was
often rejected; and if the imperial Stoics did indeed abandon the whole
thing, they were doing more than sauntering along a trite and traditional
path.

However that may be, a common view has it that ethics exhausted or
engrossed the philosophy of the imperial Stoics. The view is part of a
broader panorama, which may be vaguely described as follows. In the
heyday of Aristotle, ethics formed an important part of philosophy; but it
was not a dominant part of the subject, and the Nicomachean Ethics is no
more vital an ingredient in Aristotle's system of thought than is, say, the
Physics or the Analytics-or, come to that, the Generation of Animals.
After Aristotle's death, things changed; and in the Hellenistic schools-in
the Garden of Epicurus and under the Porch of the Stoics-ethics became
the end-all of philosophical reflection. Physics and logic were studied, to be
sure; but they were subordinated to morals, they were treated only insofar

40 Seneca, ep lxxxix II; Cicero,jin I xix 63; cf I vii 22; Diogenes Laertius, X 30.- The
subtlety and detail with which some Epicureans treated the subject is shown by Philodemus'
On Signs, on which see de Lacy [*1978]; Barnes [1988].
41 See e.g. Seneca, ep lxxxix 12; Sextus, MVII II.
42 See below, pp.l4-15; 76-77.
 THE DECLINE OF LOGIC 11

as they might serve an ethical end, and the aim and purpose of philosophy
was to show a man how best to lead his life. (The idea was not offered as a
novelty: on the contrary, it was purportedly a return to the true philosophy
of Socrates.)
Then, as the millennium turned (not that the poor pagans knew it), ethics
became not merely an end-all but a be-all, not merely the heart or centre of
philosophical endeavour but its sum and body. To be sure, it was not
universally so: Platonists, for example, sustained an apparently independent
interest in metaphysics; academic Peripatetics kept some of Aristotle's
unethical ideas alive; amateur philosophers, like Plutarch or Galen,
cultivated an interest, whether antiquarian or progressive, in various aspects
of physics and of logic. But by and large-and above all in the dominant
Stoic school- ethics was emperor, and sole emperor.
Panoramas may delight: metaphorical panoramas bewitch more often
than they instruct. And the particular panorama which I have just described
is, I think, almost wholly lacking in verisimilitude-to the extent that such a
vague picture can be compared with reality at all. But that is not my
concern: only part of the picture here interests me-the part which portrays
ethics as engrossing or exhausting the philosophy of the imperial Stoics.
The picture shows logic dismissed or disparaged; it intimates that the
imperial Stoics forwent the ins and outs of those delicious operations in
which Chrysippus had spent the better part of his philosophical energies.
I do not say that this picture is mere fantasy. On the contrary, it is based
on evidence which seems sturdy enough; and it is intrinsically plausible.
But I think that it is demonstrably false.
I shall say no more about Marcus, who is a hopeless case. 43 The next
Chapter will scan, fairly rapidly, a few passages in Seneca. Chapter 3 will
turn to Epictetus.

43 I do not know what to make of VIII xiii ('always and in the case of every
impression-if you can--<pucrw)..oyEiv, rux9o)..oyEiv, lhaAE1CtEi>Ecr9at'); but here Marcus
certainly appears to encourage some sort of logical activity. Note also III xi I ('always make a
definition or characterization [opov i1 {moypacpi]v] of whatever impression strikes you'). For
physics one might also cite e.g. VIII Iii; X ix 13; xxxi 5. (And see below, p.34 n.47.)
 CHAPTER TWO

SENECA

At first blush, the case of Seneca seems straightforward. To be sure, he had

himself studied some Iogie-no-one could study philosophy without
studying some logic. But he did not write anything expressly on the subject,
and most of his interminable pages make no reference at all to logical
matters. Moreover-or so at first it seems-in those few passages where he
does refer to logic his attitude is uncompromisingly hostile. 1
A few samples. In ep xlv he prosecutes his 'case against the logicians
[lis cum dialecticis]' ( 13). These men ignore what is necessary
[necessaria], and confme themselves to what is otiose [supervacua]: their
study is mere 'quibbling and sophistical argumentation [cavillatio et
captiosae disputationes ]' ( 5). Seneca is speaking of logicians in general
-not merely of his own degenerate contemporaries or of silly school
disputations. Indeed, he specifically alludes to the great men, the magni viri,
and their books; and it is surely Chrysippus and his followers who are
meant. Admittedly, he is prepared to allow, in ep xlv, that such superfluous
triflings are, for the most part, harmless; but he finds it difficult to see how
they might provide even a modest portion of pleasure( 4).
Elsewhere he is less tolerant. Lucilius, to whom the letters are formally
addressed, is chided for wasting his time on childish futilities, pueriles
ineptiae. Philosophy is supposed to deal with the grand questions, to
reconcile us to the misery of life and the emptiness of death, to console us
and sustain us and strengthen us. The trivialities of logic contribute nothing
to these ends: they serve no philosophical purpose at all; and what is worse,
they distract us from serious pursuits (ep xlviii 6-7). Or again: life is too
short for logic (ep xlix 5-6); the shadow of the morrow darkens us;

I ' ... although Seneca accepts the traditional division of philosophy into ethics, physics,
and logic ... , his interest was by no means catholic, and the only division of which he even
intended to give a complete account was the pars mora/is. For the rhetoric and dialectic
included in the pars rationa/is he had nothing but contempt, and even the epistemology is
simply accepted without argument' (Griffin, [1992], p.l75-the 'complete account' of ethics
is the lost fiber de moribus: fragments in the Teubner Seneca, III, pp.462-467). This
judgement, by the most eminent scholar of Seneca's philosophical ideas, will, I guess, find
general acceptance. And this may explain why so little-as far as I know-has been written
on Seneca and logic: I may mention Leeman [1953] and Trillitzsch [1962]; Gould [1965] and
Hachmann [1995], pp.238-262, discuss the role of reason (ratio, A.6yo<;) in Seneca's thought,
but not his attitude to logic.
 SENECA 13

philosophers must steer the leaking ship of humanity through the tempests
of fear and pain; logic is mere trickery and cleverness, subdola ...
calliditas; it is an impertinent and a dishonourable diversion( 12).
Elsewhere Lucilius is urged to abandon philosophical word-games, the
ludus literarius philosophorum (ep lxxi 6--how little the insults have
changed). He must forget logical fripperies, which have no power against
the skeleton with the sickle (ep lxxxii 8): 'in my judgement, that whole
class of things should be rooted out' ( 19). Or again, let him stop playing
games, let him flee logical disputations (cii 20), let him abandon his
melancholy futilities, his tristes ineptiae (cxiii 26). 2
The condemnations are sharp, and they seem to be consistent. Logic is a
worthless enterprise, offering nothing to the serious philosopher. Worse,
logic is a dangerous enterprise-if only because it diverts attention from the
deep problems of philosophy and the deep problems of life. All this is
pretty rhetorical. But then it is the product of a rhetorical virtuoso; and if
Seneca applies his paints with a generous trowel, that is not to say that his
pictures are mere exercises de style.

Now whatever these texts may insinuate about Seneca's own attitude to
logic, they surely shout loud about the attitude of Lucilius. In particular,
they show that Lucilius had a passionate interest in the subject. Just as
Fronto later feared for Marcus, so Seneca fears for Lucilius. For you do not
inveigh against a practice in which nobody engages; and you do not urge a
man to forgo something unless he is already involved in it. Lucilius, then,
was spending his time on logic.
Marcus was surely a special case, and Fronto's letters to him are
personal documents: hence the fact that the young Marcus was tempted by
logic does not in itself suggest any general thesis about the state of Roman
youth. It is different with Lucilius and Seneca; for although Lucilius too
was, no doubt, a real enough individual, Seneca's letters to him are not real
letters-and despite the occasional personal touches (genuine or fictitious)
Seneca is evidently concerned to address a public audience. 3 That is to say,
Seneca is offering public protreptic in the guise of personal advice, and
sauce for Lucilius is sauce for the rest of us. Hence if Seneca 'fears for
Lucilius', he fears for the state of Roman youth; and if he worries that
'Lucilius is spending his time on logic', he worries that young Romans are
spending their time on logic. That is to say, Seneca detected-or else
purported to detect-an unhealthy appetite for logic among his younger
contemporaries. It would be imprudent to conclude that all young Romans

2 See also e.g. ep cviii 12; cxvii 25, 30.

3 On Lucilius see Griffin [1992], pp.347-353, 416-419 (with 519); and, for a survey of
positions, Mazzoli [1989], pp.l853-1855.
14 CHAPTER TwO

were in fact swapping syllogisms and gorging themselves on the lotus fruits
of logic. But it seems to me reasonable to infer that the youth of Rome was
in fact devoting a certain amount of time to the subject.
Seneca's implicit testimony thus refutes the view that logic held little
attraction for the Romans of the early empire: on the contrary, logic-and
not ethics-was, at any rate for some youthful Stoics, the fairest part of
philosophy. Whether or not Seneca's apparent hostility to logic was the
conventional attitude of the day, it went against the preferred practice of the
bright young things.

Nonetheless, Seneca's own attitude seems clear; for surely the texts I have
cited prove that he, at least, was a sworn enemy of logic? The matter is
perhaps a little more intricate than at first it appears.
For it must be asked what exactly Seneca was attacking; and the texts
which I have thus far cited do not suggest that he was attacking logic itself.
Rather, they seem to indicate a different and a far more specific target: the
study of logical puzzles. 4 It is the sophisms and the conundrums, the
crocptcrf.Lata and the a1topa-items on which the old Stoics had lavished or
squandered so much time-which were seducing Lucilius and attracting
Seneca's puritanical displeasure. Thus in ep xlv he refers specifically to the
Homed Man and to the Liar ( 8, 10); in ep xlviii he refers to a silly
paradox which confounds mice with 'mice', the rodents with the noun (mus
syllaba est ... : 6); in ep xlix the Homed Man returns( 8). 5 Most telling is
ep cxi. Seneca explains that there is no decent Latin word for the Greek
'cr6cptcrf1a' (Cicero's 'cavillatio' is the least bad); 6 for the Romans have no
use for the thing and hence no need for the name ( 1). They have no use
for the thing because it is of no practical advantage, ad vitam nihil proficit
( 2). At best the study of sophisms is a game( 4). Thus

I would not forbid you to have dealings with them-but only when you want
to do nothing at all. Yet they do have one very bad effect: they create a sort of
delight in themselves, and they capture and detain the mind which is charmed
by an appearance of subtlety. ( 5)7

Seneca, then-so it might be suggested- , is not urging us to abstain

from logic tout court: he is urging us to abjure a petty interest in piffling
puzzles. And his attitude-so it might be added-is not wholly absurd. For

4See above, p.l 0.

5For these puzzles see below, p.76.
6The word occurs e.g. in Plautus in the sense of 'badinage': Seneca presumably implies
that Cicero was the first to use it in a logical sense. In fact it is not found in any of Cicero's
surviving works.
7 See below, p.36.
 SENECA 15

although the study of logical conundrums is in itself far from trivial, not all
the ancient puzzles are equally serious; nor, when they are serious, is it
immediately evident why they are serious-or even that they are serious.
Discussion of the Liar is easily mistaken for idle chatter; and it will readily
degenerate into infantile word-play. Moreover, it will be allowed that an
accurate study of the Homed Man has rarely eased the agonies of death.
Now the suggestion that Seneca is out to scotch the study of logical
puzzles is attractive; and it does, I think, fit the texts which I have thus far
adduced; but it seems to founder on three further passages from the letters
to Lucilius. I have in mind ep lxxxii 8-9 and lxxxiii 9-11, where Seneca
deals with the celebrated syllogisms of Zeno, 8 and ep lxxxv, which dis-
cusses similar Stoic arguments. An example:

No evil is glorious.
Death is glorious.

Therefore: death is not an evil.

Another:

Whatever is evil harms.

Whatever harms makes worse.
Pain and poverty do not make worse.

Therefore: pain and poverty are not evil.

These and similar syllogisms, put out by the founder of the Stoa, were
roundly criticized by his contemporaries and stubbornly defended by his
followers: Alexinus wrote refutations of them; Aristo of Cl;lios (the Stoic
who 'rejected logic' 9) attacked Alexinus and defended Zeno's inferences;
Diogenes of Babylon refurbished the old arguments; later on, Posidonius
discussed some of them in considerable detail; they are alluded to in
Cicero's de natura deorum, where they embarrass the Stoic Balbus. In
short, the Stoics held the things near their hearts; and they were evidently
still in vogue in Seneca's time. 10
Cicero, no Stoic, derided the Zenonian syllogisms: these little
arguments, ratiunculae, 11 which press us to the conclusion that pain is not

s See also ben VII iv 7-8.

9 Above, p.9.
10 For references and discussion see Schofield [1983].
11 See below, p.56 n.l35.
16 CHAPTER TWO

an evil are futilities, ineptiae. 12 The Stoic Seneca echoes Cicero: the
arguments are Greek ineptiae, tedious trifles; and Seneca confesses that
such things never delight him. Now Zeno's syllogisms are not sophisms,
nor are they conundrums. Yet Seneca refers to them in exactly the tone of
exasperated disdain which he displays towards the Liar and the Homed
Man. No doubt the logical puzzles were, to Seneca's way of seeing things, a
prime example of logical twaddle. But his condemnation of logic did not
limit itself to the puzzles: it stretched to the serious syllogisms of Zeno.
Thus the attractive suggestion-that Seneca aims only at the puzzles
-loses its appeal; for if not only the puzzles but also the syllogisms are
dismissed-and dismissed in the same fashion and phraseology-, then it
surely becomes plausible to suspect that Seneca's warnings do, after all,
apply to logic as a whole.
Plausible-but no more; for at least three distinct reasons. First-and
rather obviously-, in attacking Zeno's syllogisms Seneca is not attacking
logic or any part of logic; for the syllogisms do not fall within the subject of
dialectic at all. They were not presented by Zeno and they are not adduced
by Seneca as items in or elements of the study of logic. Rather, they belong
to moral philosophy, they are putative contributions to practical ethics. In
objecting to them Seneca is not objecting to logical studies: he is objecting
to a certain way of approaching certain questions in ethics.
Secondly-and equally evidently- , Seneca does not object to Zeno's
syllogisms as pieces of logic: he does not criticize their form or structure
from a logical point of view; nor-more pertinently-does he pretend or
insinuate that the use of logic and the formal presentation of arguments is in
general out of place in moral protreptic. His objection is quite other: he
objects to the syllogisms because, in his view, they have no persuasive
force.

'No evil is glorious. Death is glorious. Therefore death is not an evil'.

-You've done it: I'm free of fear. After that, I shan't hesitate to stretch out
my neck ... 13 Good God, it is not easy to say who is the more futile: someone
who thinks that with such an argument he extinguishes the fear of death, or
someone who tries to refute the argument as though it had any bearing on
reality. (ep Ixxxii 9)

12 Tusc II xii 29; cf. xii 42; fin IV iii 7; and, more generally, fin IV xviii 48-49. See also
e.g. Lucian, Jup trag 51 [699]: "If there are altars, there are gods. But there are altars.
Therefore there are gods'. What do you think of that?-Let me finish laughing and I'll tell
you'.
13 Philosophy should enable you to stick your neck out (in the literal sense): Epictetus,
diss I i 18-19.
 SENECA 17

Armed with such feeble weapons you have no chance against death: 'I am
ashamed to join the battle for gods and men armed with a gimlet [subula
armatum]' (ep lxxxv 1).
Seneca's objection to the Zenonian syllogisms seems to be this: when
the syllogisms are proposed and handled in a certain manner, they are
psychologically ineffective. If, faced by Gerontius in articulo mortis, you
glibly recite a few of Zeno's syllogisms, you will do him little good
-certainly, you will not persuade him that death is not an evil. It takes the
angel of the agony all his skill to do that-and the angel is no logician.
And Seneca's point, once again, seems eminently plausible: Zeno's
syllogisms will no more persuade you that pain is not an evil or that death
is not to be feared than Anselm's ontological argument will persuade you
that God exists. Of course, it is another question whether Seneca's point is
an objection against Zeno. Certainly, Anselm never intended or imagined
that his argument would convert the heathen. And it may be guessed that
Zeno was not naive enough to suppose that his syllogisms would convert
the world to Stoicism and rationality. 14 But then Seneca does not mean to
object to Zeno (or at least, he says nothing explicitly about Zeno ): he
objects to those people-whoever they may be-who apparently imagine
that Zenonian syllogisms will work psychological miracles.
The third of the three reasons is the most significant. It is based on ep
lxxxvii 11-41. This long text deals, in some detail, with a sequence of
syllogisms on the vanity of riches. In form and structure these arguments
are indistinguishable from the syllogisms which were scorned in ep lxxxii
and lxxxiii. In ep lxxxvii Seneca apparently sees nothing amiss with the
things. At the end of the letter, to be sure, he turns to the question of the
persuasive force of the arguments:

Let us suppose ourselves summoned to an assembly. A law for the abolition of

wealth is proposed. Shall we persuade or dissuade by these arguments? Shall
we bring it about that the Roman people demands and applauds poverty, the
base and cause of its empire, and fears wealth?( 41)

Of course not: that is to say, the syllogisms are of no use in a political

debate. But whoever thought that they were? And why ever think that if
they are not politically effective, then they are of no value at all? It would
be absurd to take such a view, and I see no indication in the text that Seneca
was tempted by it. The syllogisms are introduced, neutrally enough, as 'a
few arguments of our school which bear on virtue' ( 11). Seneca discusses

14 Note Clement, strom II xx 125.1: 'Zeno well said of the Indians that he would rather
see one Indian roasted than learn all the proofs about pain'.
18 CHAPTER TWO

them seriously and he defends them against subtle Peripatetic objections. 15

And anyone who takes this view of the syllogisms of ep lxxxvii can hardly
fail to take the same view of the similar syllogisms which Seneca discusses
in the other letters.
But this third reason, far from illuminating the issue, may seem only to
darken it; for does it not now appear that Seneca is not only hostile to
Zenonian syllogisms, but inconsistently hostile-that he mocks some of
them and patronises others? How is this inconsistency to be accounted for?
Well, perhaps Seneca was simply inconsistent? (He would not have been
the first philosopher of the breed.) Perhaps we should try to explain away
the inconsistency in terms of an intellectual development? (Other philo-
sophers have changed their minds. Perhaps Seneca started out with a
distaste for logical argumentations, and later reconciled himself to them? 16)
But I incline to a different resolution of the difficulty-I do not think that
there is, at bottom, any inconsistency to explain or to explain away. 17
Rather, Seneca's consistent attitude is something like this: in them-
selves, the Zenonian syllogisms are neutral. What matters is what you do
with them. Most would-be philosophers who toy with them produce
ineptitudes; and if the things are ponderously recited, or flashily paraded,
they are mere frivolities-they cannot subserve the ethical ends which are
their only raison d'etre. But if they are taken in a different way-as concise
expressions of philosophical insight, or as starting-points for discussion and
illustration and development, or as texts on which to hang a protreptic
sermon-, then they are serious things. 18 And in Seneca's hands they will,
no doubt, be turned to serious moral use, they will help the philanthropic
philosopher in his philanthropic task. 19

If I am right in this, the texts on Zeno's syllogisms do not scotch what I

have called the attractive suggestion-the suggestion that Seneca's

15 Trillitzsch [1962), p.51, calls attention to 'the noteworthy fact that Seneca takes the
Stoic inferences which he sets out at the start <in ep lxxxv> as subjects to be proved and
defended, and not as proofs themselves'. A curious remark, and a false antithesis: in
defending Zeno's syllogisms, Seneca is precisely claiming that they are proofs.
16 So e.g. Leeman [1953]: at first (see e.g. ep lxxxii) Seneca had a hearty dislike for all
things logical; then he came to realize that his project for a major treatment of moral
philosophy required a discussion of at least those parts of logic which the great Stoics had
taken to be pertinent to ethics; and in the end he became more or less resigned to the study of
what he had earlier regarded as Greek futilities.
17 I do not mean to deny that Seneca's attitude to logic may have changed in various
ways-! have no opinion on the issue. I maintain only that no developmental hypothesis is
needed in order to account for an apparent inconsistency in Seneca's thought.
18 Cf Cicero, Tusc III x 22; and-with a similar reference to apparently footling exercises
in logic-Taurus, apud Gellius, VII xiii 7.
19 See e.g. ben V xii 3-7; cons/ sap v 3-4; viii 1-2-texts in which comparable syllogisms
are integrated into Seneca's general exposition. On Seneca's (admittedly infrequent) use of
Stoic argument forms see Trillitzsch [1962), pp.46-59.
 SENECA 19

objections to logic modestly limit themselves to the study of sophisms and

conundrums. Nonetheless, I do not think that the attractive suggestion is
entirely correct-for it concedes too much to those who wish to make
Seneca into an enemy of logic. In a certain sense these texts do not manifest
any opposition to logic at all. It is not logic itself which arouses Seneca's
scorn and his ire: it is what he calls cavillatio, logic-chopping. The import
of his urgings is not: Don't do logic. It is: Don't quibble.
This interpretation is indirectly confirmed by ep cxiii 26. Here too we
fmd tristes ineptiae. But the futilities against which Seneca here inveighs in
his familiar fashion are not matters of logic. The question at issue is this: Is
courage an animal? This, I suppose, is a matter of metaphysics-at any rate,
it is not a matter of logic. Seneca spurns discussion of this metaphysical
topic in exactly the terms in which he scorns logic-chopping. He spurns it
because it is idle, futile, a waste of time and a distraction. In short, Seneca
has a general animus against what he regards as intellectual futility: there
are many exercises on which you may waste your mental energies; some of
them, but not all of them, as a matter of fact fall within the sphere of logic.
But that matter of fact is strictly irrelevant to Seneca's censure: his
objection to logical puzzles and to a certain way with Zeno's syllogisms is
not: It is logic-and therefore idle. It is rather, and simply: It is idle. 20
This conclusion, in itself unremarkable, is supported by a few items of
positive evidence which show that Seneca was ready enough to accept logic
when it was properly approached and properly used. Thus in ep lxxxix,
where Seneca sets out to divide philosophy into its traditional parts, he
remarks that some philosophers have rejected logic; 21 but he does not think
to praise these heterodox thinkers. He then offers a brief account of the
content of the logical part of philosophy ( 17). And he adds, speaking of
A.oytKi] but thinking specifically of logic: 'I do not forbid you to read these
things-provided that, whatever you read, you at once bring it to bear on
ethics' ( 18). Again, at ep cii 4 he observes that 'certain logical matters are
themselves involved with ethics [quaedam ... moralibus rationalia inmixta
sunt]'; and he affirms roundly that 'I have dealt with that part <of logic>
which is correct and pertinent to ethics' ( 5).
In sum, Seneca evinces a consequent attitude to logical study; and an
attitude which need neither surprise nor disconcert. Logic is not in itself to
be despised or dismissed: on the contrary, parts of it are useful-perhaps
indispensable-even to the philosopher who is convinced that moral advice
and moral consolation are the aims of his art. Trifling with logic-that is to

20 Compare Seneca's views on the aries liberales: ep lxxxviii-note, e.g. the supervacua
of Didymus ( 37) or Aristarchi ineptiae ( 39); cf brev vit viii 2 (Graecorum morbus); or his
views on the solidly Stoic pastime of giving allegorical interpretations of myths (ben I iv: on
which see Trillitzsch [1962], pp.l 09-11 0).
21 See above, pp.7-8.
20 CHAPTER Two

say, indulging in logical studies which have no ethical purpose or

ambition-is odious and dishonourable, a practice unworthy of a Stoic. But
it is not odious because it is logical trifling: it is odious because it is logical
trifling.

Seneca does not applaud the position which, according to some, was
represented by the Cynics and the Cyrenaics and the Epicureans: he does
not reject the logical part of philosophy, nor does he reject that part of the
logical part of philosophy which we call logic. He applauds a different
posture.
Some philosophers-Alexander of Aphrodisias and Galen the most
eminent among them-maintained, against the general opinion, that logic
was not itself a part of philosophy but rather its instrument; that is to say,
they urged that logic should be studied insofar as it is useful and not as an
object of intrinsic worth and interest; they advocated what might be called
logical utilitarianism. 22 Thus Alexander, taking his cue from the first
sentence of Aristotle's Analytics, maintained that the aim of syllogistic is
the production of (mooEii;Et<; or scientific proofs, and that consequently a
philosopher should study only those logical forms which are of use-of
conceivable use-for the formalization of scientific arguments. Galen took
the same line. The scientist needs logic in order to make certain discoveries
and in order to organize and confirm what he has discovered, so that logic
is an indispensable instrument of research and exposition. But beyond that,
logic has no function; and abstract propositions or argument schemata
which have no potential application to actual scientific practice are useless
and otiose-they should find no place in logical studies.
I do not know who first elaborated the theory of logical utilitarianism, or
who first derived it from Aristotle. 23 A number of texts, mostly late and
interdependent, report a debate between those who maintained that logic
was a part of philosophy (and therefore an object of study in its own right)
and those who maintained that logic was an instrument or opyavov of
philosophy (and therefore of purely utilitarian value). 24 To what extent the
debate was an historical reality and to what an expositor's fiction is unclear.
But the Peripatetics were placed in the utilitarian camp, and the Stoics were
presented as their adversaries.
In general, this assessment of the Stoic attitude to logic is surely correct:
at any rate, from Chrysippus onwards, Stoic logicians supposed, explicitly
or implicitly, that numerous logical issues which have no discernible
'utility' were perfectly proper items of philosophical curiosity. They did

22On which see Barnes [1993a].

23See Barnes eta/ [*1991], p.41 n.l.
24 Texts in Hii1ser [*1987], frags 27-32a; see Lee [1984], pp.44-54; Barnes eta/ [*1991],
pp.41-43.
 SENECA 21

not, of course, deny that logic offered important aid in ethics and in
physics; but quite apart from that, it was a respectable object of
philosophical attention in its own right. No doubt there were heterodox
Stoics. Philodemus remarks that

dialectic is an art, and yet it does not effect anything in itself if it is not
allied to ethical and physical arguments, as some of the Stoics maintained.
(rhet I, vi 10-18 = p.19 Longo Auricchio)

And we may recall those celebrated Stoic analogies: logic stands to the
other parts of philosophy as the walls of an orchard shelter the fruit, as the
shell of an egg contains the white and the yolk, as bones and sinews support
and sustain an animal's body. 25 None of this strictly amounts to a logical
utilitarianism; but it is easy to see how such ideas might lead to a utilitarian
position.
Nevertheless, it is generally true that logical utilitarianism was perceived
as an unStoic attitude. It is also a philistine attitude. Seneca, I think, was a
logical utilitarian. In assuming the utilitarian attitude he shows himself an
unorthodox Stoic,26 and a philistine. But-unlike Marcus -he does not
detest the lovely art.

It may seem that Seneca's utilitarianism was of a different stripe from the
utilitarianism of Galen and Alexander: for them, logic served the needs of
physics; for him, it served the needs of ethics. There is a generic identity
and a specific diversity. But this contrast is misleading in one way, and
perhaps in two.
First, if Seneca supposes that the study of logic should subserve a moral
end, he does not thereby suppose that it should be the servant of the part of
philosophy which is called ethics. Ethics is the study of the good, the bad,
and the indifferent: logic, if I understand Seneca aright, has its value not
insofar as it helps us to study the good but rather insofar as it helps us to
become good. For Galen and Alexander, logic is an instrument to be used in
physics or the study of nature. Seneca does not hold, with a pleasing
symmetry, that logic is an instrument to be used in ethics or the study of
conduct-he holds that it is an instrument to be used in conduct itself.
Secondly, something should be said about Seneca's attitude to physics.
He had studied physics from his youth (nat quaest VI iv 2), and he took his
studies seriously-witness the Natural Questions (which dates from the last

25Seee.g. Sextus,MVII 17-19.

26 Of course, no-one maintains that Seneca was an 'orthodox Stoic'-least of all Seneca
himself (e.g. ep ii 5; viii 8; xii I I); see e.g. Rist [I 989]-who does not refer to logic.
22 CHAPTER Two

years of his life)_27 Several passages in this work seem at first blush to show
that in physics, as in logic, he was a utilitarian. Take, say, VI xxxii 1:

it is better to become more powerful [fortior] than more learned; but the one
does not come about without the other-for strength [robur] comes to the
mind only from the good arts, from the contemplation of nature.

The text does not explicitly say that the sole reason for contemplating
nature is to strengthen your moral fibre; but it readily suggests such an
interpretation. Physics is worth doing because it promotes morality. 28
Other passages appear to tell in the opposite sense. Thus in the preface to
the first book of nat quaest Seneca explicitly says that the part of philo-
sophy 'which pertains to the gods [quae ad deos pertinet]' is far grander
than the part which pertains to men-and it is clear from the context that
the former part is physics and the latter ethics (I praef 1). Thus Seneca sets
physics above ethics; and it might thence be inferred that the study of
physics cannot possess a purely instrumental value. But the inference would
be fallacious. Seneca certainly held that physics is a nobler study than
ethics; and unremarkably-for the view, which in any case had the
authority of Chrysippus behind it, 29 was little more than a platitude. 30 One
study is nobler than another if it addresses a nobler object; and the object of
physics is divine. The platitude does not entail that the study of physics is
more important than correct ethical practice; nor, in particular, does it entail
that the study of physics has an intrinsic value, a value which is not purely
instrumental. In principle, at least, the grandeur of physics might lie
precisely in its grand contribution to morality.
Yet it would be a mistake to ascribe a physical utilitarianism to Seneca,
and his attitude to physics must be distinguished from his attitude to logic.
True, he held that the point and purpose of physics is moral. But that is not
to say that the study of physics has point and purpose only insofar as it
leads to or facilitates moral practice. For the point of doing physics may be
moral not insofar as it conduces to some further moral activity but rather
insofar as it is itself a form of moral activity. And Seneca took physics to be
moral in the latter way: he thought that the study of the natural world was in
itself an essential part of the moral life.
The study of physics does not-or does not merely-help you to be
good and happy: studying physics is a part of what it is to be good and
happy; doing physics is itself a form of virtuous and felicitous activity-it
is part of the repertoire of the Good Man, of the Sage. Indeed, Seneca's

27 On nat quaest see e.g. Codofier [1989].

28 Cf e.g. nat quaest II lix 2; ep xxxi 8; lxv 16-24; ad Helv viii 6.
29 See e.g. Plutarch, stoic repugn I 035AB.
30 See e.g. Aspasius, in EN 1.2-2.6.
 SENECA 23

enthusiasm for physics leads him to claim that virtue itself is splendid
precisely because it 'prepares the mind for the contemplation of heavenly
things and makes it worthy to consort with god' (nat quaest I praef 6). 31
Logical study may enable you to lead a good life. Scientific study is a part,
and the supreme part, of the good life. '0 what a wretched thing is a man
who does not raise himself above things human [o quam contempta res est
homo nisi supra humana surrexerit]' ( 5).
This conclusion is no Senecan idiosyncrasy. On the contrary, it was a
commonplace of ancient thought. It goes back to Plato and Aristotle. It may
be ascribed, in one form or another, to many Hellenistic philosophers. It is
expressed in numerous imperial texts. The commonplace depends, at
bottom, on theology. For, first, ethics and physics are each given a
theological turn: ethics, insofar as 'assimilation to God', OJ.!Oi.c.omc; 9(!), is
conceived of as the ultimate end of human activity, or at least as an ideal to
be striven towards; physics, insofar as the study of nature is properly
speaking the study of 'things divine', of the universe and its parts and
pieces. Now, secondly, it is plain that the gods fmd felicity in con-
templation-that is to say, in physics. Hence we shall become god-like
insofar as we study physics.
Logic remains an instrument-and, it may be added, an instrument for
which the gods themselves have no use. It is an instrument by using which
we may come to lead better lives; and insofar as leading a good life requires
us to study physics, it is eminently plausible to suppose-what in any event
must seem a banal enough truth-that logic is an instrument for physics.
The difference between Seneca on the one hand and Alexander and Galen
on the other is a matter of emphasis and of style rather than of substance.

31 Cf e.g. nat quaest I praef 11-17; III praef 18; ep lxv 21; ex 9; cxvii 19; ad Helv xx 1-2.
 CHAPTER THREE

EPICTETUS

A: Logic in the Discourses

Marcus thanked Rusticus 'that I read the records [U1tOfl vi]f!ata] of

Epictetus, of which Rusticus lent me his own copy' (I vii 8). Marcus knew
Epictetus thoroughly and borrowed generously from him for his
Meditations. 1 And this is scarcely odd because Epictetus seems, at first
glance-and again at second- , to be as devoted a moralizer as the
emperor. The Discourses 2 contain just under a hundred conversazioni: not
one of them turns on physics, 3 and only a handful turn on logic. 4
It is true that half of the Discourses is lost. 5 But the Encheiridion is a sort
of chrestomathy put together from the Discourses, and nothing in it
suggests that in their general character the lost books were any different
from those which have survived. 6 Certainly, it would be bizarre to postulate
that we have been deprived of a sustained Epictetan discussion of physics

I See Rutherford [1989], pp.225-255; cf Stanton [1968].- Hadot [1978] argues that the
whole structure of the Meditations is based upon the Epictetan theory of the three 't01tot-see
below, p.34-35.
2 The English 'Discourses' is not altogether appropriate; but it is traditional, and
innocuous. The ancient sources offer 't.tmpt~ai' (Simplicius, Photius); 't.tai..E~Et~'
(Gellius-which he latins as 'Dissertationes'); ''Y1tOJ.!ViJJ.!a'ta' (Marcus; cf Arrian, ad Ge/l =
Epictetus, diss praef 2).
3 But see frag I = Stobaeus, eel II i 31 (below, pp.25-27); frag VIII = Stobaeus, eel IV
xliv 60.
4 All general works on Epictetus say something about his attitude to logic. The best pages
I know are in Bonhiiffer (1894], pp.J22-127. In addition to the various items mentioned in
later footnotes, see Long [1978], pp.l19-121; and I should perhaps refer to de Lacy
(1943]-which, despite its title, does not have much to say on logic; and Xenakis
(1968]-which offers a thin survey spiced with a few howlers.
5 Photius knew eight books (bib/ 58, 17b 17-20 Bekker)--and he knew of twelve books of
'OJ.!tl..tat, also put together by Arrian (cf Schenk! [*1916], pp.XXXIII-XLVII-one fragment
survives: frag XI = Stobaeus, eel IV xxxiii 28); Gellius, XIX i 14-21, cites from the fifth book
of the Discourses(= frag IX).
6 Arrian wrote ench 'having collected from the remarks of Epictetus those items which are
most pertinent and most indispensable in philosophy and most capable of moving the soul, as
he himself wrote in the letter to Massalenus. . .. The same items, in pretty well the same
words, are reported in Arrian's Discourses of Epictetus' (Simplicius, in ench praef 5-11
Hadot). We may therefore take it that everything in ench which has no close parallel in the
surviving Discourses-by my reckoning something between a third and a half of the
work-had a close parallel in the lost books.
 EPICTETUS 25

or of a set of discourses on the logical part of philosophy. 7 It is true, too,

that the Discourses do not rehearse the teaching which went on in
Epictetus' school: they rehearse-they purport to rehearse 8-his informal
discussions and his impromptu expositions. But they also purport to give us
a picture of what the man himself was like, of what constituted the soul of
his philosophy.
And the Discourses show us, indisputably, a man who talked largely in
the ethical part of philosophy. Moreover, his talk was for the most part
practical rather than theoretical, and severely practical at that. He offered
the world a pin-striped cynicism, Diogenes without the barrel; 9 ai:J.d if there
is something obscurely admirable about his dogged idealism, it is difficult
to avoid the thought that the attitudes which he recommended are both
humanly impossible and morally disgusting.
However that may be, the Discourses do contain a substantial number of
references-mostly in the form of asides or obiter dicta-to logical issues.
Many of these references are-or at least appear to be-disparaging; some
of them are-or at least appear to be-scathing. It is easy to conclude -and
it is a conclusion often drawn-that Epictetus hated logic, that he did not
work at it himself, and that he did not urge his pupils, formal or informal, to
work at it. In short, Epictetus 'rejected logic'.
The same has seemed to be true of physics. A fragment, conserved by
Stobaeus, appears decisive.

Why, he says [<p11cri], should I care whether existing things are compounded
from atoms or from indivisibles or from fire and earth? Isn't it enough to learn
the essence of good and bad and the measures of desires and disinclinations
and also of impulses and aversions, to run our lives using these as rules-and
to forget about those things which are beyond us? Perhaps they cannot be
known by the human mind; and even if you were to suppose that it is perfectly
possible to know them, what is the advantage of such knowledge? Shouldn't
we say that those people worry themselves [npayj..lata EXEtv] to no purpose

7 But it is a shame that the parts of the Discourses from which ench 52 was drawn are
lost: see below, pp.38-39.
8 The author of the Discourses is Arrian, not Epictetus (though I shall follow the harmless
custom of referring to them as though they had been written by Epictetus himself). On the
disputed question of the relationship between Arrian's written words and Epictetus' spoken
discourses see e.g. Stadter [1980], pp.26-31; Radt [1990]; Hadot [*1996], pp.l52-160. On
Arrian, in the philosophical vein, see Brunt [1977]; Follet [1989b].
9 See esp III xxii, nepl. K1.lVtO"!.!OU, with Billerbeck (*1978]; IV viii 30-33-for the pin-
stripes see IV xi. - Here and hereafter all merely numerical references are to the Discourses
unless the context indicates otherwise.
26 CHAPTER THREE

who insist that this is a necessary part of what a philosopher must say? (frag I
=Stobaeus, eel II i 31 ) 10

Physics is beyond us. 11 Probably we cannot understand it. Even if we can, it

is of no use to us. Let it go hang.
Epictetus' writings are not easy to interpretY For the most part they are
dialogues or quasi-dialogues, in which Epictetus quizzes an interlocutor
(real or imaginary) or else talks to himself. It is not always evident whether
it is the interlocutor or Epictetus who is speaking, or whether Epictetus is
setting himself a question or answering one. Moreover, he frequently uses
irony, and it is often difficult to be sure, at first reading, whether a passage
is ironical or not. In short, it is easy enough to get the sense of a passage
upside down. 13
And so it is with fragment I, which continues as follows. (The end of the
text in Stobaeus is corrupt; but the essential message is not in doubt.)

-Now you don't think that the Delphic advice 'Know Thyself is
superfluous?
-Certainly not, he says [<p11cri].
-Then what does it mean? If you urged a member of a chorus to know
himself, wouldn't he obey the recommendation by concerning himself with
his fellow singers and with being in tune with them?
-Yes [<p11cri].
-And a sailor or a soldier?
<-Similarly.> 14
-Then do you think that men have been made to live as solitary animals or
rather with a view to some sort of community?
<-With a view to some sort of community.> 15
-And by what have they been so made?
-By nature.
-Then we need not bother ourselves [noJ.:unpaylJ.OVEtv] about nature, what it
is and how it governs the world and whether it exists or not? 16

IO For the text of Epictetus I follow Schenk! [*! 916]-any significant deviations are
noted. I have constantly consulted: Schweighiiuser [*1799]; Long [* 1885]; Oldfather
[*1925]; Souilhe and Jagu [*1975]; Laurenti [*1989]; and Carter/Hard [*1995].
II A view commonly associated with Socrates (e.g. Xenophon, mem I i II; Sextus, M VII
8; Eusebius, PE XV lxii 10-but note e.g. Cicero, Rep I x 16); and therefore not implausibly
ascribable to Epictetus. (Note also Aristo: Diogenes Laertius, VII 160, quoted above, p.9.)
12 Quite apart from the fact that the text is sometimes gravely corrupt.
13 The best account of the style of the Discourses is still that of Hirzel [1895], II pp.247-
250.
14 It seems likely that a brief reply has dropped from the text here.
15 This reply, too, is missing in the MSS: it was added by Heeren. (For 'KotvrovtK6c;' see
esp IV xi I; cfe.g. I xxiii I; II x 14.)
16 For the (curious) question whether nature exists or not see Sextus, PH I 98.
 EPICTETUS 27

The sequence of questions belongs to Epictetus and not to his interlocutor.

Hence the '<p11cri' in the answers refers to the interlocutor. Hence-working
backwards-the initial '<p11cri' does not mean 'Epictetus says': 17 it is the
'<p'llcrt' with which Epictetus frequently introduces an anonymous (and no
doubt feigned) interlocutor. 18
The fmal remark in the fragment thus certainly comes from Epictetus'
own mouth. Equally certainly, it cannot be construed as a straightforward
statement to the effect that physics is not for us-to present such a
statement as an inference from the preceding concessions would be
grotesque. (And, in style as well as in content, thoroughly unEpictetan.) If
the remark is nevertheless printed as a statement, then it must be construed
in a heavily ironical fashion, so that it imports exactly the opposite of what
it says. Better, punctuate it with a question-mark and make it an ironically
rhetorical question. 19
A young man asks why, given that the moral life is the goal of
philosophical study, he should care about the nature of the universe, why he
should con physics as well as ethics. And Epictetus proceeds to show him
why he should. The fragment proves the opposite of what it has been taken
to prove. It proves that Epictetus took physics to be a necessary part of
philosophical study, not that he took it to be of no account. 20 Like the Old
Masters of the Stoa-and like Seneca-, Epictetus was a devotee of the
physical part of philosophy. It remains true that the surviving books of the
Discourses contain no sustained discussion of physics; and the references to
physics, though numerous enough, 21 are usually of a very general nature
and almost invariably set in a moral context. But this fact is hardly
astonishing, and it certainly implies no disdain for the subject on Epictetus'
part.

There are, as I have said, dozens of allusions in the Discourses to matters

which fall within the logical part of philosophy-and in particular within
logic. Epictetus himself had been trained in the subject by Musonius Rufus,
who once memorably castigated him for an error in his logical exercises. 22
And he learned his logic thoroughly. One sign of this is the casual use of

17 Pace Long, Oldfather, Laurenti, Carter/Hard.

18 See e.g. I iii 9; iv 1; vi 6. So too 'inquit' or 'inquis' in Seneca. The ancient theorists
recognized the phenomenon, which they included under the head of 'personification' or
1tpocrro1to1totia: e.g. Quintilian, IX ii 37; Fronto, eloq ii 14 [pp.141-142 van den Hout2],
ascribing the practice to Chrysippus-whose rhetorical and stylistic panache he strangely
praises. See Griffin [1992], p.414.
19 So Schenk!, after Richards.
2o At II xiv 26 there is a close parallel.
21 E.g. I vi 19; x 10; xvi 15-18; xvii 13-19; xx 16; II xxiv 25-28; III xiii; !Vi 104-105; iii
12; ench 49; frag III = Stobaeus, eel IV xliv 66.
22 See I vii 32, quoted below, p.134.
28 CHAPTER THREE

logical terminology which we fmd throughout the Discourses. Thus

Epictetus will refer to O"UVTlJlJlEVa and to C>tEsEUYJlEVa-to conditionals and
disjunctions (I xxix 51; II ix 8); to O"UJl1tE1tAYJlEVa or conjunctions (I xxvi
14; II ix 823 ); to tpo1ttJCa or compound propositions (I xxix 40); to
arguments which are valid or invalid, O"UVUlCttlCOt or clO"UValCtOt (see ench
44);24 to the analysis of universal or JCa8oA.tJCa propositions (II xx 2-3). 25 At
IV viii 9 he accuses his opponents of 'supposing' (8vw;) that someone is a
philosopher, of 'taking' or 'assuming' (A.a~6vw;) that he is a dis-graceful
character, and of 'concluding' (1tayoum) that philosophy is a waste of
time.
The list could easily be lengthened. The items in it come from logic:
they are all technical or semi-technical terms. Epictetus uses them without
warning; nor does he call any particular attention to them or hint that they
might be thought somehow inapposite. Nothing perplexing there-anyone
trained in logic will occasionally allow logical vocabulary to slip into his
ordinary speech. But the vocabulary indicates that Epictetus' early studies
of logic had made some mark on him, a mark which he took no pains to
disguise.
One passage appears to indicate that Epictetus was himself aware of this
verbal tic-and that he half apologized for it. In II xii he discusses Socrates'
argumentative technique: Socrates was a skilled dialectician, and he was
peculiarly concerned to arrive at correct defmitions of things. But

23 'cru~m:JtAq~u~vov', the standard Stoic term for 'conjunction', appears four times in
Epictetus. At II ix 8, Epictetus remarks that a cru~1t1tAq~E:vov is 'saved' or 'preserved'
(crc!JI,;e'tat) if it is 'conjoined from truths [1; aA119tOv crU~1t1tAEX9at]': i.e. a
crU~1t1tAy~E:vov is true if and only if each of its components is true. Here, then,
0"1l1.1.1tE1tAY1.1.EVa are certainly conjunctions. They are paired with Otel,;euy~E:va or
disjunctions. At I xxvi 14 the context gives nothing away; but there is no reason not to take
'crU~1tE1tAEY1.1.EVov' again to mean 'conjunction'. The other two occurrences are in ench-
36 and 42. Here Simplicius takes the word to mean not 'conjunction' but 'conditional': his
illustrative examples are conditionals, and he gives the explicit gloss ''to crU~1tE1tAYI.I.EVov
ij'tot crUVll~~E:vov' (in ench LIV 25 Hadot). 42 pre- sumably derives from I xxix 51. There
Epictetus refers explicitly and unmistakeably to conditionals or crUVlli.I.~Eva. Simplicius
perhaps reasoned that, in constructing 42, Arrian must have intended to retain the sense of I
xxix 51, so that in 42 'crU~1tE1tAEy~E:vov' must mean 'conditional'-the change of word
imports no change of sense. For 36, where cru~1tE1tAEy~E:va are paired with OtE/,;EUYI.I.EVa,
the corresponding part of diss is lost: perhaps there too Arrian changed 'cruv11~~E:vov' to
'crU1.1.1tE1tAEYI.I.EVov'; or perhaps Simplicius thought that the word should be given the same
sense in 36 as it bore in 42. In any event, if Simplicius' comment on these passages is
right, then they contain the only surviving examples of 'crU1.1.1tE1tAEYI.I.EVov' used in the sense
of 'conditional'. (But note Boethius, hyp syll I ii 5, where 'coniunctio' is used to mean
'conditional'.) In that case, perhaps ench preserves the traces of a heterodox Stoic
terminology. Or perhaps Arrian was merely muddled.
24 See below, pp.l40-141.- No parallel to this in the extant Discourses. (But note IIi 3:
'tCx UO"UVa!C'ta cruvayoV'tEc;.)
25 See below, pp.30-31.
 EPICTETUS 29

he did not say 'Define envy for me', and then, when it had been defined,
reply: 'A bad definition-for the definiens is not extensionally equivalent to
the definiendum [OU UVt<XKOAO'OSEt tql KE<paAatroliEt tO optKOV]',
'thereby using technical terms which laymen find tiresome and difficult to
follow-terms which we can't renounce. (II xii 9-10)

'Socrates did not need to load his talk with technical terminology; but-or
so Epictetus seems to confess-I fmd that I cannot avoid it myself.'
This interpretation is, alas, mistaken. II xii is entitled 11Epl. -cou
BtaA.yEcr9at, and in it Epictetus contrasts the sympathy and skill of
Socrates with the bludgeoning practices of his own contemporaries. In the
first sentence of the discourse he speaks in the third person of 'our people',
oi iJJlE'tEpot. In the second sentence 'our people' become 'us', and the first
person plural is maintained for the rest of the discourse. But it is evident
that Epictetus is criticizing his fellow Stoics for their penchant for jargon:
he is not beating his own breast; and the first person plural indicates a polite
complicity rather than an honest confession. (When I say 'We philosophers
no longer write tolerable English' I do not intend to criticize myself. On the
contrary.)
Nonetheless, confessedly or not, logical terms are pervasive in the
Discourses. And Epictetus is also pleased to make logical analogies to life.
Thus in logical exercises we cannot choose which hypothesis we are to
defend: you may not say 'Give me a different compound proposition'-it is
your task to argue from whatever premiss you have been given. 26 And so it
is with the 'hypothesis' on which you must base your life. 27 Or again, if you
judge a true conditional to be false, then it is you who are condenmed, not
the conditional: and in the same way if the court wrongly convicts you of
impiety, it is the court, and not you, which stands condenmed. 28 Or again
(and less enlighteningly), a conjunction or a disjunction is 'saved' if things
actually stand as the word 'conjunction' or 'disjunction' announces or
'promises ' 29 that they stand; and a man is saved if he is actually what the
term 'man' promises that he is. 30

26 See below, p.95.

27 I xxix 39-41; cf I xii 17; II v II. For 'im69ecru;' in the sense of 'role in life' see e.g.
Marcus, VIII i 2; X xxxi 5; XI vii.
28 I xxix 50-51; a comparable analogy at ench 42, with conjunctions replacing
conditionals (above, n.23).
29 EmxyyeA.ia: cf e.g. II ix I; x 4; III xxiii 10; IV viii 6-the verb 'E1tayyAA.ecr9at' in
roughly the sense of 'mean' is common in Stoic and non-Stoic contexts: e.g. Diogenes
Laertius, VII 71, 72. (See also below, p.139.)
30 II ix 7-9.- A further such analogy at ench 36; and perhaps another at IV viii 12 ('as
Zeno says, you must first learn ta to\l A6you crtotxeia')-but this is a notoriously puzzling
text: see e.g. Hiilser [*1987], pp.82-83.
30 CHAPTER THREE

I do not pretend that Epictetus had enjoyed the logical training of a

Galen, or that he had absorbed logic in the way in which Galen had
absorbed it. Casual references to logic are not as copious in his writings as
they are in the medical treatises of Galen. Nor, of course, did he write
works on logic, as Galen was to do. Nonetheless, he learned logic; and it
left a mark on his style. 31

It is natural to wonder how good Epictetus himself was at the subject. The
issue is hard to determine: the student error to which he confesses at I vii 32
is of no account; and the Discourses are not marked by any great density of
argumentation. Of course, Epictetus frequently offers reasons in favour of
his contentions-and he frequently offers reasons against contentions which
he rejects. But these reasons rarely call for the sort of argumentative
complexity which might offer a test of logical competence. (No criticism
here: it is precisely what is to be expected in discussions of this general
sort.) Passages which seem to shed some light on Epictetus' logical prowess
are rare. I may here mention three of themY
At II xx, the discourse directed against the Epicureans and the Academic
sceptics, Epictetus remarks that

the greatest evidence that something is clear [vapy~] is that we find that
anyone who opposes it must necessarily make use of it. For example, if
someone were to oppose the claim that something universal [Ka9oA.tK6v] is
true, it is clear that he would have to make the contrary assertion: Nothing
universal is true. Idiot-not even that. For what is that but 'If something is
universal it is false'? (II xx 2-3)

Epictetus here tacitly presupposes the standard Stoic way of formulating

universal propositions. 33 'All As are B' is remodelled as 'If something is A,
it is B'; 'No A is B' is remodelled as 'If something is A, it is not B'. And he
argues as follows: if you oppose the claim that some universal proposition
is true, you thereby affirm that no universal proposition is true. That is to
say, you affirm that:

If something is a universal proposition, it is not true.

But what you thereby affirm is itself a universal proposition. Hence it

cannot be true. (Suppose that it were true. Then any universal proposition
would be false. Hence, being itself a universal proposition, it would be

31 'How thoroughly <logic> was studied is apparent on nearly every page': Hijmans
[1959], p.39. An exaggeration-but not wild.
32 On a fourth text, II xxv, see below, pp.59-60.
33 See Sextus, M X! 8-11; cf Sedley [1985].
 EPICTETUS 31

false. Thus the supposition that it is true leads to a contradiction. Hence it is

not true.)
II xx does not set the argument out as cleanly as a modern logician
might desire-but the Discourses are not, by their nature, expressed in the
most precise and pedantic fashion possible. In any event, the argument
which Epictetus offers is a good one; and it is not without a certain subtlety.
It is an example, and a perfect example, of a self-refutation or 1tEpt'tpom1;34
and it turns on the exotic truth that anything which implies its own negation
is itself false. So we shall conclude that Epictetus was reasonably expert in
logic? Perhaps. But I doubt if we can put much weight on the text; for I
imagine that the example was a standard example, and Epictetus may have
lifted it from a handbook.
At the beginning of I viii, where Epictetus is concerned with rhetoric
rather than with logic, 35 we fmd-in the standard editions and trans-
lations-the following argument:

If you have borrowed and not returned, you owe me the money.
But it is not the case that you have borrowed and not returned.

Therefore: you do not owe me the money.

The argument is invalid-and pretty scandalously so. Epictetus apparently

argues according to the schema:

IfP, then Q.
NotP.

Therefore: not Q.

The schema is not valid. Nor, so far as I know, did any Stoic ever suggest
-crassly or with subtlety-that it was valid. Epictetus has erred.
But in fact it is not Epictetus who has erred: it is his editors and
translators. In I viii 2 Epictetus is remarking on the rhetorical practice of
collecting groups of equivalent or synonymous expressions, groups of
icroouvajlOUV'ta. 36 The 'syllogism' which I have just cited is not offered as

34 On which see Burnyeat [1976]. Neither 'mpt'tpoml' nor the verb 'mpt'tpbtEtv' is
found in Epictetus.
35 See below, p.58.
3 6 For the practice in a Stoic context see e.g. Fronto, eloq ii 19 [p.114 van den Hout2].
Note that Galen recommends the study of i.croouva~ouv'ta in a logical context (inst log xvii
5), and that he had written an essay llEpt 'tcDV i.croouva~oucrrov 7tpo'tacrEwv (lib prop XIX 43
Kiihn). For the sense of 'i.croouva~dv' see Barnes [1993a], p.46 n.64; for the importance of
such items in connection with what the Stoics called {l7tocru/../..oytcr'ttKoi. arguments see ibid,
pp.45-47.
32 CHAPTER THREE

a syllogism at all: it is offered as an example of a pair of equivalent

expressions; and the Greek should be translated thus:

'If you have borrowed and not repaid, you owe me the money'-'lt is not the
case that you have borrowed and not repaid and yet do not owe me the
money'. 37

This is not a fallacious argument since it is not an argument at all. It is a

pair of sentences of the forms:

If P and Q, then R

and

Not ((P and Q) and not R).

Given a certain interpretation of the word 'if', the two sentences are indeed
logically equivalent to one another: 38 Epictetus commits no error. Indeed,
the example which he manipulates is relatively complex, and his
manipulation betrays a certain sophistication. I do not infer that Epictetus
was a competent logician-merely that he was more competent than most
of his modern commentators.
The third passage comes at IV i 61 (the context is immaterial). There
Epictetus argues as follows:

We conceive that whatever has power to confer the greatest benefit is divine,
and then we incorrectly [K<n::ro~ subjoin: 'This man has power to confer the
greatest benefit'. Necessarily what results from these things is incorrectly
inferred [avayKTJ K<lt 'tO YEVOJ.LEVOV E~ <lU'tOOV E1tEVEX9ijvm K<lKro<;].

Now it is natural enough to suppose that we 'incorrectly subjoin' the second

premiss precisely insofar as that premiss is false. It is then easy to suppose
that we 'incorrectly infer' the conclusion precisely insofar as the conclusion
is false. Hence Epictetus is maintaining that if you have a false proposition

37 Ei oaveicrro Kat J.Ll'l U7tEOroKa~. Oq>EtAEl~ J.lOl to &.pyupwv'ouxt Eoaveicrro Kat OUK
&.1tEoroKa~ oil J.Ll'lv ocpEil..n~ J.lOl to &.pyupwv.- Schweighauser [*1799], II p.l23, emends
the text in order to ensure that the 'conclusion' is clearly marked; Schenk! mispunctuates; and
the passage is mistranslated by Oldfather (who also fails to see that 'toihov tov tp67tov' is
adverbial-'toiltov' is not 'this syllogism'), and by Souilhe, Laurenti and Carter/Hard.
38 Construe 'if' in the 'Philonian' sense (below, p.l 03), so that a conditional is false when
and only when its antecedent is true and its consequent false. Then 'If A, then B' will be false
when and only when 'A' is true and '8' is false; hence when and only when 'A' and 'Not 8'
are both true; hence when and only when 'A and not 8' is true; hence when and only when
'Not (A and not B)' is false.
 EPICTETUS 33

among the premisses of an argument, then the conclusion of the argument

will necessarily be false. And this is an elementary howler: as Aristotle was
the frrst to observe, you may validly infer a true conclusion from premisses
some or all of which are false. 39
But I do not think that we need ascribe this howler--or any other
howler-to Epictetus. No doubt what makes it incorrect to subjoin the
premiss in question is the fact that it is false-but 'incorrectly subjoin' does
not therefore mean 'subjoin something false'. Hence we have no reason to
think that 'incorrectly infer' must mean 'infer a false conclusion'. An
inference is, I suppose, an operation which takes us from a set of beliefs to a
further belief; 40 and one chief point of drawing an inference is to increase
our stock of true beliefs. Thus we might reasonably call an inference
'correct' insofar as it takes us reliably to a true belief; and an inference will
be incorrect insofar as it does not take us reliably to a true belief. In that
case, an inference from premisses some of which are false will indeed be an
incorrect inference. An argument with false premisses may be valid (so that
if its premisses were all true, then its conclusion would also be true); and an
argument with false premisses may have a true conclusion. But such an
argument is not a 'true' argument, as the Stoics put it; 41 and if you use it to
infer a conclusion, then you have no guarantee that the conclusion will be
true.
It may be said that even if IV i 61 does not contain a howler, it at least
gives the impression of containing one; and that is not the mark of a good
logician. Perhaps that is just. I do not want to claim that Epictetus was a
wizard at deduction. But although I do not fmd in the Discourses-and do
not expect to fmd-any positive evidence that Epictetus was a clever
logician, I do not find any negative evidence that he was a bad logician.

B: The fashion for logic

Galen bewails the fact that his contemporaries do not give a hoot for logic.
Seneca does not. Indeed, the fact that he inveighs against logic-chopping
suggests-as I have already argued-that logic-chopping was a popular
hobby in his time. Epictetus also inveighs against logic-and again the
consequence is easily drawn. But with Epictetus we do not need to rely on
our own uncertain inferences: Epictetus himself roundly states what Seneca
merely suggests.

39 See Patzig [1968], pp.l96-203. - You might perhaps take 'incorrectly infer' to mean
'invalidly infer': then Epictetus' howler would be even more grotesque-he would be
implicitly supposing that you cannot make valid inferences from false premisses.
40 See below, p.l23.
41 See below, p.l02.
34 CHAPTER THREE

Contemporary philosophers [oi vuv <ptA.6cro<pot] abandon the first area

['t07toc;] of the subject, and also the second, and they devote themselves to the
third: changing arguments, arguments which conclude by way of
questioning, 42 hypothetical arguments, lying arguments. 43 (III ii 6)

Philosophy has three 'areas' or t61tot, the third of which includes-and is

perhaps exhausted by-logic. And Epictetus' contemporaries give them-
selves wholly to this third area-to logic.
The three t61tot here are not the three traditional parts of philosophy, 44
and Epictetus is not complaining that his contemporaries devote them-
selves exclusively to A.oytKil, to the logical part of philosophy. Rather, they
are 'the three areas in which anyone who is to be noble and good [KaAO<;
Kat <'xya86<;], must train himself' (III ii 1): the first area concerns desires
and disinclinations-the management of the passions; the second concerns
impulses and aversions-the management of the appetite; and the third
concerns assent-the management of the intellect.
The three t61tot are referred to elsewhere by Epictetus. 45 Some scholars
hold that the triad is an original invention of Epictetus' own; that it con-
stitutes a system or a theory; and that it had some historical influence,
notably on Marcus. Now it is true that Marcus does allude once or twice to
the t61tat; 46 but I am not persuaded that the system made any great mark on
him, still less that it pervades the MeditationsY Indeed, I confess that words
such as 'system' and 'theory' seem to me overblown in this connection; and
the three-fold distinction seems little more than a con-venient expository
device. The bare facts are these: in a dozen or so places Epictetus makes a
distinction among three different areas in which you should train yourself,
and he marks a hierarchy among them. This dis-tinction may indeed be
original to him-although he does not himself claim, or even suggest, such
originality. 48 He uses the distinction as though it would be familiar to his
audience, or at least obvious to them; but he does not harp on it. Elsewhere

42 On this curious phrase see below, pp.136-137.

43 I.e. versions of the Liar paradox-not 'lying' (Long), nor 'false reasonings' (Souilhe).
-Oldfather's translation contains a number of errors in the last sentence of this extract.
44 Pace Bonhtiffer [ 1890], pp.18-28; Hadot [1978], pp.69-70. But the pages in BonhOffer
and Hadot nevertheless contain the two best discussions of Epictetus' "t01tOt-see also
Hijmans [1959], pp.64-68; Hadot [*1996], pp.149-151.
45 See I iv 11-12; xvii 22-24; III ii 1-4; cfi xxi 2; II xvii 15-16, 31-33; III xii 13; xxvi 14;
IV i 68-75; iv 13; x 13.
46 Or at any rate, he hints at them in a few passages (e.g. VIII vii I; IX vii}-although he
does not use the word '"t61to\;' save at IX xxxvii, where he is paraphrasing Epictetus (=
Epictetus, frag XXVII).
47 Pace Hadot [1978]. Note that, were Hadot right, then-given his (mistaken)
identification of the three "t01tot with the three parts of philosophy- , it would follow that
Marcus did after all apply himself to logic and physics as well as to ethics.
4 8 Note that Seneca ascribes a comparable (five-fold) distinction to the Cyrenaics (ep
lxxxix 12); and that he himself offers a similar three-fold distinction( 14).
 EPICTETUS 35

he is content to introduce a different triad of 't01tOt. 49 And in general, he is

ready enough-like other ancient philo-sophers-to use the word ''t61toc;'
for an area of thought without having any particular cartography in mind. 50
However that may be, contemporary philosophers spend all their time on
the third, logical, area; and what is more (or so Epictetus' choice of
illustrative examples seems to imply) they devote their time to logical
problems which have no evident links with moral philosophy-let alone
with moral practice. The phrase 'oi vuv qnJ.6crocpot' presumably embraces
teachers of philosophy. And Epictetus assures us that the pupils aped their
teachers. The showy young logician is a figure in Epictetus' Rogues
Gallery, where his portrait hangs alongside those of the time-serving
politician and the smug paterfamilias. For it was not merely that the young
sucked up logic in school: they took it home with them, and they took it out
to dinner-parties.
Flash Harries frequent the philosophers with the specific aim of learning
logic. 51 They bone up on the subject in order that they may display at a
banquet their subtle knowledge of hypothetical syllogistic-and thereby
win the admiration of the senator sitting next to them (I xxvi 9). 52 The
sophisms and conundrums were particularly eligible objects for ostentation:

And as for the Master Argument, what have J53 got to say about it? If I am a
vain man, I thoroughly amaze the guests at a banquet by enumerating the men
who have discussed it: 'Chrysippus wrote splendidly about it in the first book
of On the Possible. Cleanthes wrote a special work on the subject, and so did
Archedemus. Antipater wrote about it too--not only in On the Possible but
also separately in On the Master. Haven't you read the work?'-'No.'-'Oh,
do read it'. (II xix 8-1 0)

Party bores bored with logic.

Plutarch had 'often' found it necessary to reproach guests who 'dragged
the Master Argument into drinking parties'. 54 Aulus Gellius tells a story of a
tiresome young man, a self-professed Stoic, who entertained the dinner-
guests of Herodes Atticus by explaining to them how he alone had solved
all the Stoic a1topa-the Master Argument and the Sorites 'and other
riddles of that sort [alios que id genus griphos ]' (I ii 3-4). Such behaviour

49 See ench 52 (below, pp.38-39): it is plain that this is a different triad; plain too that it
does not correspond to the three parts of philosophy.
50 E.g. I xxviii 29; IIi 3; xix 6; xxiii 31; IV ii I; iv 13.
51 See II xviii 3, 34; xxi 17; III xxvi 13.
52 Cf e.g. Plutarch, prof virt 78EF, who aptly cites Plato, Rep 5398, on the dangers of
allowing young men to sharpen their claws on logic.
53 'I' here does not designate Epictetus: see below, pp.44-45.
54 quaest conviv 615A-from the first question of the work, entitled 'ti ott q>tAOCIOq>etv
1tapa 7t6tov'; cfe.g.profvirt 80A; Lucian, Gallus I [718-719].
36 CHAPTER THREE

would not have pleased the Old Stoics, if Persaeus is to be trusted:

according to him, 'if logicians gathered for a drink and then discussed
syllogisms, you would judge that they were acting inappropriately to the
occasion' (Athenaeus, 607B). 55 Nor did it please Herodes Atticus: he
instructed a servant to fetch a book from his library, which he then read
-in reproof-to the obnoxious youth. The book contained the Discourses
of 'that admirable old man' Epictetus. And Gellius cites II xix 12-17, from
the very discourse which I have just quoted. Of course, it is all too pat, and I
fear that Gellius embroidered-if he did not simply invent56-his little
anecdote. Nonetheless, ben trovato.
It is easy to see how logic came to charm. After all, we cannot study
anything at all without some minimal reflection on language and some
minimal employment oflogic. It is precisely for this reason that

some people are captured and detained by these things-one by language

[I.E~tc;], 57 another by syllogisms, another by changing arguments, another by
some other lodging-house of this sort. And they remain there and rot, as
though they were among the Sirens. (II xxiii 41)

The same thought, without the Sirens, has already been cited from Seneca. 58
Aulus Gellius again echoes Epictetus. He remarks on the curious nature
of logical study: at first it seems rough and uninviting and useless; but then,
as you make progress in it, it comes to seem sweeter-indeed,

there follows a sort of insatiable lust for learning; and if you do not set a limit
to it there will be no small danger that, like many others, you too will grow
old in those whirls and eddies of dialectic, as though you were on the Sirens'
rocks. (XVI viii 16-17)

55 But the Platonist Taurus encouraged the discussion of logical issues at his
ungastronomic symposia: Gellius, VII xiii. (On philosophical symposia see Tecusan [1993].)
56 So, apparently, Holford-Strevens [1988], pp.48-49 (cf. p.206, for Gellius and
Epictetus). Observe that Gellius himself seems fond enough of logical puzzles: V x; xi; IX
xvi; XVI ii; XVII 9; xiii.- On Gellius and philosophy see also Goulet [1989e ].
57 On Epictetus' attitude to rhetoric see below, p.58.- The same complaint-that students
of philosophy would get side-tracked by an interest in style and rhetoric-is found in e.g.
Seneca, ep cviii 6; Plutarch, aud rat 42E. Note that philosophical works were read in the
rhetorical schools (e.g. Dionysius of Halicamassus, imil iv 4 [pp.21 0-211 Usener]; Quintilian,
Xi 35-36; XII ii 1-10; Tacitus, dial xxxi 5-7; Dio Chrysostom, xviii 13; cf. e.g. Stemplinger
[1912], pp.110-114); and that Plato in particular-although he had his detractors (e.g. Lucian,
rhet praec 19)-was generally regarded as a model of good style (e.g. Plutarch, prof vir/ 79D;
Lucian, lexiph 22; Gellius, I ix 10; XVII xx 6; [Longinus], xiii 3, with Russell ad Zoe; cf
Wa1sdorff [1927]; de Lacy [1974]; Trapp [1990]; Dorrie and Baltes [*1990], pp.369-403;
Brunt [1994], p.37). - And let me here paraphrase a later text: 'Plato is now f:v xepcr1 tiiiv
omcouvtrov Etvat cptA.oA.6yrov !lOVOV, while many people read Epictetus in the hope of
improving their lives' (Origen, c Cels VI 2).
58 ep cxi 5: above, p.14.
 EPICTETUS 37

I suppose that Gellius took his Sirens from Epictetus-though the image is
banal enough in itself. 59
And there is also this text from Fronto:

So you may see the orator despised and held in no honour, while the logicians
[dialecticos] are esteemed and cultivated inasmuch as in their arguments there
is always something obscure and twisted-and so it comes about that a pupil
always sticks to his master and serves him, bound and held, as it were, in
permanent chains. (eloq iv 10 [p.l50 van den Hout2])

The Siren-song seduced, and it detained.

Other texts might readily be added. Thus it is worth remarking that
Lucian's numerous jibes against the Stoics frequently accuse them of logic-
chopping.60 And there is a choice passage from Fronto's contemporary,
Maximus ofTyre:

If someone says that this is what philosophy is-verbs and nouns, arts of
arguments, refutations and confutations and sophisms, and hours spent on all
these things-then it is not hard to find a teacher. You will find every place
stuffed with such sophists: the thing is in plentiful supply and appears
everywhere. I dare say that there are more teachers than pupils of this sort of
philosophy. But if these things are a small portion of philosophy-a portion
which it is disgraceful not to know but not splendid to know-then let us
avoid disgrace and know these things; but let us not preen ourselves on them.
(diss i 8)

That logic excited and enticed was a commonplace.

Thus if we consider what Epictetus-and his colleagues-actually say
and imply about contemporary intellectual life, we see a picture quite
different from the one painted by the common view of imperial Stoicism.
Epictetus' contemporaries were not blinkered moralists, ignoring the logical
studies which the Old Stoa had patronized. On the contrary, philo-
sophy-Stoic philosophy-had virtually reduced itself to logic. The philo-
sophers themselves think only of their syllogisms. Their young pupils are
schooled in logical technique and can fmd no better conversation at the
dirmer-table than hypothetical inferences and the Master Argument.
No doubt Epictetus is exaggerating. No doubt there is a basis of truth
behind the exaggeration. Logic, not ethics (and not physics), was the

59 See e.g. Cicero, fin V xviii 49; Plutarch, aud poet 15D-16A; Sextus, M I 41-43; and
e.g. Euripides, Androm 936.
60 See e.g. dial mort i 2 [332]; symp 23 [435]; vit auct 22-23 [562-563]; Hermot 81 [824-
826]. See Helm [1902], pp.266-278.
38 CHAPTER THREE

fashionable part of philosophy. When Epictetus fulminates-or seems to

fulminate-against logic, he is not 'typical of his age', an age which
despised logic and cared only to keep its ethics warm. On the contrary, he
was crying in the wilderness and swimming against the tide: he was not
keeping up with the Joneses, nor were the Joneses keeping up with him.

C: Against logic?

The common view that ethics dominated imperial Stoicism is false:

Epictetus confirms what Seneca implies-that matters were quite otherwise.
But the common view may draw in its horns. 'Perhaps after all (the revised
and hornless view goes )-perhaps after all, hoi polloi were logically
minded; perhaps logic was indeed a popular pastime. But the great and
serious philosophers of the age were thorough-going moralists: they at least
had no time for logic. At any rate, Seneca and Epictetus, who chiefly
represent imperial Stoic philosophy for us, were no logicians.' I have
already urged that this is an inadequate characterisation of Seneca's attitude
towards logic. Is it any more adequate with regard to Epictetus?
Certainly, Epictetus was hostile to the fashionable interest in logic.
Certainly, his texts contain numerous criticisms of logic-or at least of
logic-addicts. And it will be best to start by looking at one or two of these
hostile passages.
The Sirens of logic, according to Epictetus, detain us in the wrong -r61to<;
or area of philosophy. The end of the Encheiridion distinguishes three areas
or -r61tot:

The first and most necessary area in philosophy concerns the use of theorems
-for example, not to tell lies. 61 The second concerns proofs-for example,
why is it that we should not tell lies? The third confirms and articulates the
proofs themselves-for example, why is this a proof? (What is a proof? what
is implication? conflict? truth? falsity?) Now the third area is necessary
because of the second, and the second because of the first. And the first is the
most necessary and the area in which we should take our rest. But as for us, 62
we do the opposite: we spend our time in the third area and all our efforts are
concerned with it. We completely neglect the first area. Hence we tell lies-
and are ready to show how it is proved that we should not tell lies. (ench 52)63

61 'ljiEUOEcr9at': Hadot [*1996], p.l50 n.22, prefers to take the verb here in its larger
sense of 'make mistakes'.
62 The first person plural is not a piece of self-incrimination: it is, again, the plural of
polite complicity (see above, p.29).
63 Nothing in the extant Discourses corresponds to this important text: presumably it
paraphrases something from the lost books.
 EPICTETUS 39

Why do we languish in the third area? Why do the Sirens of logic enchant
us? Their song is seductive because we are both indolent and vain.
We are indolent. We boast about our logical powers: 'Bring me any
syllogism and I'll tell you about it. I'm skilled enough there. But in life it is
all the reverse-ignorance and inexperience, alla8ia Kat a1tEtpia' (II iii 4-
5}.64 For logic is a lot easier than life:

If the woman is willing, and if she nods to me and beckons me, and if she
fondles me and snuggles up to me-if I then hold aloof and conquer, that is a
sophism above the Liar and above the Sorites-you're entitled to be proud
about that, not about propounding the Master Argument. (II xviii 18)

There are no temptations in logic which we must struggle to resist. Or at

any rate, the temptations of the Liar are somewhat easier to resist than the
blandishments ofLais. 65
We are vain. And if we study logic,

it will not merely be an additional distraction from more necessary matters,

but no ordinary occasion of pride and pretentiousness. . . . By what device
could you now persuade a young man who excelled in these things that he
should not become an appendage to them but they to him? He proudly
tramples all our reasons underfoot and walks about puffed up and unable to
bear it if you tap his shoulder and remind him of what he has lost and where
he has strayed. (I viii 6, 9-1 0)

Logic feeds our vanity; and it is a self-sustaining vanity-for who is harder

to persuade than a soi-disant master of persuasive argument? 66
Idle and arrogant, seduced by the Sirens, we linger in the third area. But
that, surely, cannot be all that there is to philosophy?

Is that why you left home? Is that why you wanted to meet someone who
might benefit you? What benefit? So that you might analyse syllogisms more
readily and explore the hypothetical arguments? Is that the reason why you
abandoned your brother, your country, your friends, your relatives-so that
you could learn all that and then return home? (III xxiv 78) 67

64 On this passage see below, pp.83-85.

65 For which see Sextus, M !X 153.
66 Cf II xx 37, on the invincible stubbornness of Academic and Epicurean dogmatists. On
the correct attitude for the expert logician to take see II vi 4.
67 On the notion of leaving house and horne to seek a teacher see I iv 22; III xxi 8; xxiii
32; Dio Chrysostorn, xii 10; Matthew, xix 29.
40 CHAPTER THREE

Epictetus' questions are rhetorical. Had a student taken them seriously, he

might well have answered them with a cheery affirmative: 'Yes-that is
precisely why I left home to study at Nicopolis'. Nor, when he returned
home, would his parents always have been distressed-the rich father of
I xxii 18 gives this advice to his son: 'Learn syllogisms from the philo-
sophers-but as to what you ought to do, you yourself know that better than
the philosophers do'.
For my part, I sympathize with the student and with his father. But
Epictetus is of another mind. If that is what you think, he grumbles, then
'go home and sit in a corner: 68 weave your syllogisms and propound them
to your friends'. 69 You go to your philosophy school in order to learn logic;
and you will learn nothing more. But it is not merely that you do logic
instead of doing other things. The third area is not, after all, just another
part of the wood: logic actually bears on conduct, so that what you learn in
logic you should apply in life. And what happens? You learn to prove that
you should not tell lies-and you lie through your teeth. Logic lovers are
not like people who read novels rather than works of history; they are not
even like people who read works of history rather than enter political life:
they are like people who read medical text-books-and then, when they fall
ill, never think of applying what they have read. 70
This is a particular case of a general complaint, common enough in any
school and in any age. 71 Students of philosophy learn to talk, and they do
not learn to act; they 'take up the teachings of the philosophers only to talk
about them [JlEXPt Myou JlOVov]' (title of II xix). 'They philosophize
without acting-only as far as talking [avu 'tOU 1tpantv, JlEXPt 'tOU
A.yHv].' 72 They are 'lions at school, foxes abroad'. 73 The pupils are only
imitating their masters. For the adulterer who pleads in his defence 'But I'm
a scholar-my mind is on Archedemus [ai..Aa cptMA.oy6c; EtJlt Kat
'ApXEOllJlOV vo&]' is probably a teacher of philosophy-no doubt a
logician, and a Stoic logician (II iv 10-11 ).

68 Cf I xxix 54-below, p.56.

69 II xiii 26; cf. e.g. III xxiv 81.
70 The third comparison is from Simplicius, in ench LXVII 29-32 Hadot.
71 Brunt [1979], p.496 n.28, remarks that Epictetus, II xix 20, 'suggests that most persons
with a dilettante interest in philosophy were Epicureans'. What Epictetus says is this:
Observe yourselves in your actions and you will know what sect you belong to--most
of you will find that you are Epicureans, a few of you that you are Peripatetics (but
Peri-patetics of the flabby kind).
This is not a sociological comment on the distribution of dilettantes among the Schools: it is a
typically caustic criticism of soi-disant Stoics. (See below, n.88.)
72 frag X = Gellius XVII xix I (citing Favorinus who cites Epictetus).
7 3 IV v 37; see also, e.g. IIi 30-34; x 29-30; xviii 26; II xvi I; III v 15-17; vii 18; xx 18;
xxi 10; xxiv 16, 38-40; xxvi 13, 16-20 IV i 138; vi 12; cfe.g. Musonius, frag 14 [p.76.14-17
Hense] = Stobaeus, eel IV xxiia 20; Simplicius, in ench praef 12-14 Hadot.
 EPICTETUS 41

There are two rather different complaints here: a complaint that philo-
sophers do not let their philosophy impinge upon their actions, and a
complaint that philosophers act against their philosophical convictions. The
first of the two complaints accuses philosophers of intellectual 'com-
partmentalism': they hold their philosophical beliefs in a separate part of
their minds and do not imagine that such beliefs might have any causal
intercourse with the rest of their lives. Such men might, in principle, act in
accordance with their philosophical beliefs: the charge against them is not
that they act against their philosophy; it is rather that, whatever they may
do, they do not do it because of their philosophy. The second complaint
accuses the philosophers of hypocrisy-and usually of immorality. They
preach one set of beliefs and they act on another; they urge us not to tell
untruths, and they lie with the best.
Epictetus does not distinguish between these two complaints; and in
effect he supposes that if you only philosophize 'as far as talking', then you
are bound to show yourself 'a fox abroad'. No doubt there were many foxes
abroad, then as now; no doubt there were self-styled Stoics who were
thoroughly obnoxious creatures. One such case is celebrated. In 66 A.D.
Publius Egnatius Celer bore witness against Barea Soranus and his daughter
Servilia-to both of whom he had taught Stoic philosophy. He lied in
court-and they were condemned to death. Three years later Musonius
Rufus charged Celer with his crime; and despite the fact that the Cynic
philosopher Demetrius spoke in his defence, Celer was condemned by
Domitian. It is not an uplifting story. 74
History apart, the immoral philosopher had long been a stock figure: 'I
hate men who are rotten in their actions and philosophical in their remarks'
said Pacuvius75-and he is echoed in a hundred texts. 76 Lucian is
particularly keen on the matter. His Symposium is entirely devoted to it.
Zeus' tirade against the philosophers at lear 29-32 makes the same point in
pithier fashion. And there are a dozen other Lucianic passages. It is equally
a theme in Seneca/7 who remarks that the objection 'You say one thing and
you live another [aliter loqueris aliter vivis]' had been brought against Plato
and Epicurus and Zeno the Stoic (vit beat xviii 1), and against Aristotle and
Democritus and Socrates (xxvii 5). And the accusers are delighted with
their accusations-for if the great moralists were themselves dogs beneath
the skin, why should we not all trot along the primrose path (xix 1-3)? The

74 See Tacitus, ann XVI 32; hist IV 10; 40: Moles [1983].
75 odi homines ignava opera et philosopha sententia (cited by Gellius, XIII viii 4).
76 See e.g. Antigonus of Carystus on Pyrrho (6 <ptAoc; EAtyev roc; oil fitlY ltotf]aat
OUf1<prova -cote; Mymc;: Eusebius, PE XIV xviii 26); Nepos in a letter to Cicero (frag 39
Marshall = Lactantius, inst div III xv I 0); Seneca, ep xx I; xxix 4-8; Gellius, X xxii (citing
Plato); Athenaeus, 565DF (on Stoics, citing Antigonus ofCarystus)-other texts in Festugiere
[1960], pp.140-142.
77 See esp vit beat xviii-xxii.
42 CHAPTER THREE

accusations, Seneca insists, are false (see ot vi 5); but they were commonly
repeated and no doubt commonly believed. 78
Philosophy was the art of life. The abstract question 'Can you live your
philosophy?' was posed often and seriously-and not only in connexion
with the Pyrrhonian sceptics. 79 But the more personal question was more
insidious: Do you live your philosophy? Some philosophers did indeed live
their philosophy-or were deemed to do so. Thus the decree in which the
Athenians honoured Zeno the Stoic states that 'he made an example of his
own life, which was consistent with every word which he spoke' (Diogenes
Laertius, VII 10). Similarly, on a less exalted level, Nepos praises Atticus:
'He had such a firm grasp on the doctrines of the chief philosophers that he
used them to manage his life and not for ostentation' (Att xvii 3). And
again-a delectable example-consider Galen's self-congratulations: 'I am
a good doctor and a good man-for (unlike my colleagues) I do what I
recommend others to do' (ord lib prop XIX 53 Kiihn). 80
Concordet sermo cum vita, let your words and your deeds be in harmony
(Seneca, ep lxxv 4): the injunction was a commonplace; and so too was the
claim that few philosophers heeded it. 81 The call to apply your
philosophical doctrines to your life is a special case of the injunction; and
the demand that logical expertise be used in practical life is a special case of
the call.

78 On 'philosophical invective' see e.g. Sedley [1976]; Owen [1983].

79 See e.g. Aristocles, apud Eusebius, PE XIV xviii 26, for the complaint against the
Pyrrhonians; Plutarch, non posse (for the Epicureans); and Colotes' polemic (llept 'toil on
Ka'ta 'tCx 'tiilv iii..Arov qnA.ocr6cprov II6'Yf.LU'ta ooo l;;ilv i:crn), known from Plutarch, adv Col
(for pretty well everyone else).
80 I should also cite Simplicius' eulogy ofEpictetus:

This is the man who says 'Lameness restricts the legs, not the will
[npoaipecrtc;]-unless the will itself wishes it to': he based his words on his life, and
did not strive-as most of us do--to say whatever might be deemed to earn praise. (in
ench XV 45-49 Hadot)

And note the anecdote about Plutarch and his slave: Gellius, I xxvi 6-7.
81 Cf Friedlander [1920], III, pp.260-265; Chemiss [* 1976], pp.412-413; Holford-
Strevens [1988], p.l92 n.2; Decleva Caizzi [1993], pp.316-323; Mansfeld [1994], pp.183-
191.
 EPICTETUS 43

D: Exegesis

And so we do nothing but talk. But there is worse than that; for the talk
itself is second-hand, mere bookishness. We philosophers love our books-
they take the place of life in our lives, and we will not be parted from
them. 82 Asked what moral progress I have made, I reply: 'Take up the
volume On Impulse [TIEpt OPJlil<;] and see how well I've read it' (I iv 14)-
as though virtue were simply a matter of knowing my Chrysippus ( 7), as
though I had left home to acquire book-learning( 22). 83
Or again, every moment of our lives we are struck by a hundred
impressions, some of them cognitive or Kataft.:rptttKai and others not.
What ought we to do? Evidently, we ought to attempt to discriminate
among these impressions, to determine which are genuinely cognitive and
which false or feeble. What do we do? We pick up a book entitled On
Cognition, IlEpt Kataft.:t']\j/Ero<;, and start to read from it.

What is the reason for this? It's because we have never read anything and we
have never written anything in order that in our actions we might take the
impressions which strike us and use them according to nature. Rather, we stop
when 84 we have learned what is said and can explain it [~rm1cra.cr9m] to
other people, when we can analyse the syllogisms and explore the
hypothetical arguments. (IV iv 13-14)

We talk-but we talk about what we have read. No wonder we never

proceed beyond talk to action. Here is one of Epictetus' most common
laments: we read our philosophical texts, we learn what others have written,
we produce an ~flYTJcrtc;-an interpretation, a commentary, an explication
de texte. And then we sit back, the task accomplished. 85
We are set to read the works of Chrysippus. They are difficult. In order
to understand them, we reasonably seek out a commentary or an interpreter
-and then we repeat what we have read, or perhaps even essay an
explanation ourselves. Yet why read Chrysippus in the ftrst place? Well,
Chrysippus too is an interpreter or commentator; but what he interprets and
explains is Nature. And that is surely the only reason why we should read
him: not in order to be able to interpret his writings, but in order- thanks
to his writings-to be able to interpret nature. 86 Equally, it may be assumed,

82 See esp. IV iv.

83 CfiViv 16-17.
84 Adding 'tv ti(l' (Richards) before 'f.La9tiv'.
85 See e.g. II vii 3-4; IV iii 12; and for other remarks on interpreters: I iv 22; II ix 14-19;
xvi 34; xxi 10; III xxi 7; IV iv 14; ench 49.
86 I xvii 13-19; cfvi 19 (we are here to be spectators and interpreters of God's words); II
vii 3; IV iii 12.- Note also Galen, us part III 16 Kiihn: Aristotle was lletv6tat<li; ... tEXVllV
cp'i>crtro<; tl;mf]cracr9at. On such (standard) uses of 'tl;,yeicr9at' see Mansfeld [1994],
44 CHAPTER THREE

we may quite properly read and interpret Chrysippus' works on ethics. But
with what aim should we do so? Not in order to become experts in
Chrysippean exegesis, but in order to become experts, thanks to
Chrysippus, in the art of life: in order to become good men.
Yet we stop at Chrysippus and forget about Nature and the Good. What
could be more foolish? What less philosophical? Progress, 1tpoKo1ti], is not
made by devouring heaps of Chrysippean writings (I iv 5-9). As for
commentaries, we may write them by the score: it will do us no good. A
book sells for a shilling-a commentator is worth no more( 14-16). 87 If
our desires are not in accord with nature, and 'if we do not ensure that
correct beliefs are active, then we shall be no more than interpreters of other
men's opinions' (II ix 14); and the words we utter will come from our lips
and not from our hearts ( 17). Indeed, if commentary is the end and
stopping-point, we might as well sit down and study the texts of Epicurus;
for the mere conning of texts, even of the best Stoic texts, will not tum us
into Stoics( 18-19). 88
I shall cite a further passage on this score, since it has been often
misunderstood. It comes from the discourse on the Master Argument.
Having explained what the Argument is, Epictetus remarks:

Now if someone asks me 'But which of the premisses do you accept?', I will
answer that I don't know, but89 that I have learned the following account:
'Diodorus accepted these, and Panthoides (I think) and Cleanthes those, and
Chrysippus the others.'-'And you?'-'Oh, I am not made for this sort of
thing, for testing my own impressions and comparing what people say and
forming an opinion ofmy own in the area [1:61tov]'. (II xix 5-6)

The passage has been treated as another of the texts in which Epictetus
takes a swipe at logic. For surely the fmal response is intended to distance
Epictetus from the Master Argument: he is not cut out to think about logical
puzzles; he-we infer-looks to another and higher philosophical goal.9

pp.l53-154. (But there is no reason to infer that the verb '~TrYEicr8at' is ambiguous: rather,
there are (at least) two different sorts of object on which we may practise exegesis-we may
try to explain and understand what happens and we may try to explain and understand what
someone has said or written.)
87 Cfe.g. I xvii 13-19 (with 29); II xvii 34-35; III ii 13; xxi 7 (with ench 49).
88 Cf. II xix 20-24; and on pseudo-Stoics see also II vii 18; xxiv 38 (cf Gellius, I ii
6}-above, n.71.
89 Placing a comma rather than a colon after 'otlia': 'but that ... ' is part of the imagined
answer to the imagined question.
90 So e.g. Schweighiiuser [*1799], II pp.523-524; Bonhiiffer [1890], p.7 (citing also I
xxvii 15 and II xviii 18 to show that Epictetus 'confesses his ignorance with regard to the
more difficult problems of logic and epistemology'); Bonhiiffer [1894], p.l23; Hijmans
[1959], p.39 (who, however, strangely suggests that Epictetus is telling 'a fib'); Hershbell
[1993], p.l43.
 EPICTETUS 45

The general sense of the text is exactly the opposite. 91 Epictetus is

speaking with heavy irony-or else, which comes to the same thing, he is
speaking in inverted commas and with the voice of an imaginary
interlocutor. 92 For it is Epictetus' settled and repeated view that he, and all
men, are obliged to judge their own impressions: the one thing which the
gods have left in our power is the use of our impressions, and the one area
in which our duty lies is their management. 93 'Oh, I am not made for this
sort of thing': on the contrary, you are made precisely for this sort of thing.
Far from confessing to a dilettante interest in the doxography on the Master
Argument and denying any need to produce a judgement of his own on the
substantial issues which the Argument raises, Epictetus insists- how could
the point have been rnissed?-that we must judge the Argument for
ourselves; and he deplores mere doxography. If all I know is what Diodorus
and the rest maintained, then 'as far as that is concerned, I am no different
from a philologist [ypajljlattK6~]' who comments learnedly on the poems
of Homer (II xix 7). 94
If we refuse to judge for ourselves, even in arcane matters of logic, if we
remain interpreters and commentators of the judgements of other men, we
cease to play our allotted role in the universal comedy.
Epictetus was neither unique nor original in complaining that men had
become bookworms. Marcus later flogged himself with the same whip:
'Give up this thirst for books, lest you die grumbling 95 ' (II iii 3). And
Alexander of Aphrodisias remarked, naturally enough, that 'today most
philosophical gatherings [cruvoucriat]' are based on books (in Top 27.12-
14). Earlier, Plutarch had reported it as a common opinion-an opinion
which he himself vehemently and virtuously rejected-that a philosopher is
'a man who argues from his professorial chair and gives his lectures on
books' (sen ger 796D). And Seneca had earlier developed the idea in terms
very similar to those used by Epictetus (ep xxxiii 7-10). It is the idea
underlying Seneca's celebrated aphorism that philosophy has become
philology: quae philosophia fuit facta philologia est (ep cviii 23):

91 If you are determined that Epictetus declines to interest himself in the Master
Argument, then (like, e.g. Carter/Hard) you must follow Upton and emend the text ('ovotv
yt:yova' for 'ovo yeyova'}-see Schweighiiuser [*1799], II p.523 (who rejects the
emendation and still misunderstands the text).
92 Compare I xxvii 15: of Academic and Pyrrhonian arguments Epictetus says that 'for
my part, I've no time for these things, and I can't plead in favour of common sense'. The last
thing which Epictetus intends us to infer from his remarks is that he has no time for such
things and cannot plead in favour of common sense.
93 See below, p.61.
94 For 'philologists' in Epictetus see also III ii 13; on the terms 'ypaJ.!J.!CI.'ttK6~' and
'cptA6A.oyo~ see e.g. Pfeiffer [1968], pp.156-159; Pepin [1992].
95 yoyyill;;rov: the verb is used by Epictetus at I xxix 54 (below, p.56), to which Marcus
perhaps alludes; and cf!V i 79.
46 CHAPTER THREE

philosophers are no different from scholars (philologus aut grammaticus)

who ask learned and futile questions about Ennius or Vergil ( 35):

everything they say and spout to a listening throng is foreign to them [aliena]:
'Plato said this, Zeno said that, Chrysippus and Posidonius said the other
thing'. ( 38)96

Later, Galen, a great writer of books and a voluminous commentator,

was aware of the existence of such gibes; for in one of his pharmacological
works he affirms-am I wrong in detecting a defensive tone?-that

no doubt what most of the craftsmen say is true enough-that it is by no

means the same thing to learn from the living voice (1tapa ~OOO"TJ<; <pcovfi<;] and
to read something out of a book. But if you are industrious and naturally
intelligent, you may get no small benefit from reading a book which is
knowledgeably written ... (camp med lac XII 894 KUhn)

And so Galen himself will continue to write. And to write.

The generic objurgation, that to the reading of many books there is no end,
took a variety of specific forms. Galen's craftsmen contrasted book-
learning with learning 7tapa l;;fficr'll<; <provfj<;, from the living voice of a
master; and the notion that word of mouth-if possible, word from the
horse's mouth-is generally better than a written record is expressed in
numerous texts, Greek and Latin, pagan and Christian. 97 Connoisseurs of
ancient philosophy will think first of Plato's Phaedrus, the fans et origo of
the notion (in particular 275A). I have some sympathy with the craftsmen
-try learning anything from a computer manual. But little with Plato; and
none for the generalised version of the thing-in historical studies at least,
written records (however liable to error) are preferable to the stuttering
word of mouth. However that may be, I do not fmd any particular respect
for 'the living voice' in Epictetus.
A somewhat different form of the complaint insists that genuine learning
comes from within the learner: the book-learner is thus set in contrast not
with the apprentice but with the autodidact. 'I searched myself [Mtl;;'llO"Ufl'llV

96 In Athenaeus, the learned Myrtilus is universally admired, save by Cynulcus, who rails
against his empty polymathy. Myrtilus retorts: 'How right I am to hate all you philosophers
who hate scholars [J.ltcroq>tAoMyou~ ovta~]' (6100). But this is jocular-and Cynulcus has
just trumped him on a point of scholarship.
97 E.g. Cicero, ad Att II xii 2; Galen, alim fac VI 480 Kiihn (cf ven sec XI 194 Kiihn);
Porphyry, quaesl Hom 434; Papias, apud Eusebius, HE III xxxix 4 (cf HE V x 4). For Latin
viva vox see e.g. Seneca, ep vi 5-6; Quintilian, II ii 8; Pliny, ep II 3; Gellius, XIV ii I (with an
allusion, I suppose, to Cicero, leg III i 2); Irenaeus, adv haer III ii I. See Karpp [1964] (with
references to Christian passages); Alexander [1990]; Mansfeld [1994], pp.l22-126.
 EPICTETUS 47

EJlErout6v] ', said Heraclitus-and many mistook him (perhaps

intentionally) to be urging first-hand research rather than the taking of
nourishment from other men's spoons. 98 There are intriguing issues here.
But I have not noticed them raised by Epictetus-after all, he earned his
crust by plying the spoon.
Different again is the contrast between book-learning and practical
experience: it is a poor helmsman who will 'steer by the book', EK ~t~A.iou
KU~Epvav. 99 Books may be useful, even for helmsmen; but they must be
supplemented-and may be supplanted-by practice or a<rKTicrt<;. It is by
playing the piano that we learn to play the piano, as Aristotle put it (EN
1103a30-b2)-and he applied the moral to morals. The importance of
practice is a theme of Galen's medical writings; 100 it is a commonplace of
Stoic ethics-and it is a constant refrain in Epictetus' Discourses. 101 And
Epictetus also insists that we become, and remain, pianists by playing the
piano, moral heroes by acting heroically. 102

The philosophers enjoin us not to be satisfied with learning but in addition to

train and to practise. For we are long accustomed to doing the contrary of
what we should do and to putting to use beliefs contrary to the correct beliefs.

98 See B 101 Diels-Kranz; and for the misconstrual e.g. Diogenes Laertius, IX 5; Dio
Chrysostom, xxxviii 2.
99 See e.g. Galen, lib prop XIX 33 Kiihn; temp XI 797 Kiihn; cf antid XIV 6 Kiihn. See
Alexander [1993], p.40 n.39. (The English idiom, 'do it by the book', has a different
implication.)
wo E.g. opt med I 55-61 Kiihn; an morb V 15-26 Kiihn.
101 E.g. I i 25-32; ii 32; xxv 31; IIi 29-31; ix 8-18; xvi 1-23; xviii; III iii 14-19; viii; x 7;
xii (with the title 'llEpt amc:i]crE~'); xxiv 84-118; IV i 111-113; frag XVI= Stobaeus, eel III
xxix 84: see esp Hijmans [1959]. On iimc:11crt~ in Musonius note frag 6 = Stobaeus, eel III
xxix 78 (from a llEpt amc:i]crEm~); cf van Geytenbeek [1962], pp.42-49; Laurenti [1989],
pp.2113-2120. For Seneca e.g. ep xvi 2 (cotidiana meditatio); lxxx 2-4; cfNewman [1989];
Hachmann [1995], pp.257-262. For the Old Stoa note e.g. a llEpt amc:i]crEm~ for Herillus
(Diogenes Laertius, 'VII 166), and one for Dionysius (VII 167). The notion is especially
associated with the Cynics: Diogenes Laertius, VI 70-71; cf Goulet-Caze [1986].- The matter
has been much discussed, often in French: see recently Hadot [1995], pp.276-333. Hadot
explains that an 'exercise spirituel' is 'a practice designed to effect a radical change of being'
(p.271), a change which he dsecribes in terms too deep for me. The exercises may assume a
wide variety of forms-a hermit flagellating himself in the desert and a student assiduously
taking notes on a course of lectures on Aristotle's biology may each be indulging in an
exercise spirituel-; and exercises of this sort are to be discovered throughout Greek
philosophy, at least since the time of Socrates.- Well, I cannot deny that 'radical changes of
being' may be in the wind in a number of ancient philosophical texts. But---or so it seems to
me-the notion of intellectual iicrK'Ilcrt~, of 'mental gymnastics', is at bottom a pretty down-
to-earth sort of thing; and in most ancient texts iicrK'Ilcrt~ aims at nothing so high-falutin' as a
change of being. After all, the idea of training or practice is hardly esoteric or religious (or
even remarkable): it is a piece of ordinary, robust, common sense that if you want to ride a
bike, then you should get pedalling-and that if you want to master logic, then you should do
the exercises at the end of the Chapter.
102 See II ix 10-12; xviii 1-7.
48 CHAPTER THREE

Hence if we do not put the right beliefs to use, we shall be nothing but
interpreters of other men's doctrines. (II ix 13-14)
'You read books-you do not listen to the masters'. 'You read books-
you do not learn for yourself. 'You read books-you do not seek practical
experience'. But the specific form of the complaint which concerns me here
is different again. For Epictetus, in his remarks on exegesis, is concerned
with book-learning in the sense of learning about books. He is worried that
we are too intent on 'metascience', too interested in learning what others
have learned. And although I have nothing against exegesis myself-after
all, that is largely how I earn my own crust-, it is difficult not to feel a
twinge of sympathy.
Yet Epictetus can scarcely have been astonished by the state of affairs
which he lamented; for the educational system, in which he himself was-
so far as we can tell-a reasonably orthodox participant, might have been
designed to produce little interpreters. 103 A carpenter does not say 'Come
and hear me talking about carpentry' -he builds a house and demonstrates
his art (II xxi 4). But what does a teacher of philosophy do? He cries

'Come and listen to me giving my little commentaries [crx6A.ta] ... I'll explain
Chrysippus to you like nobody else, I'll give you the purest account of what
he says-and perhaps I'll add a dash of Anti pater and Archedemus.' ( 6-7)

A teacher of philosophy would give his students a text to read-usually, if

he were a Stoic philosopher, a classical text by one of the Old Stoics. The
students would first read it by themselves (I iv 9). In school they would read
it to the master (I x 8). The master would then 'explain' it; that is to say, he
would offer an explication de texte or commentary 104-or the pupils might
'explain' it to the master (I iv 15-16). 105
Fronto offers us a vignette of the practice-which happens to concern
schooling in Stoic logic:

So you read a book to a philosopher. You listen in silence while the master
explains it. You nod that you've understood. The others read-and as for you,
you usually doze off. You hear 'What is P? what Q? ["ti "to nponov; "ti "tO

103 On school practice see e.g. Plutarch, aud rat-and Bruns [1897]; Hijmans [1959],
pp.41-48; Clarke [1971], pp.55-1 08; Goulet-Caze (1982]; Mansfeld [1994], pp.l93-194;
Lakmann [1995], pp.216-220.
104 Cf e.g. I iv 15; vi II; 19; II xvii 13-19; xxi 10-11; III xxi 8; ench 49. Note that there is
no question in any of these passages of written commentaries: it is always a matter of oral
commentary-or rather, and less ponderously, the 'exegesis' to which Epictetus so frequently
refers is simply a schoolmaster's taking his boys through a tricky text. (For the distinction
between written and oral commentary see Galen, in Hipp fract XVIIIB 321 Kiihn.)
105 Perhaps in a sort of oral examination? See the later and stylized version of such a
thing in Porphyry's catechistic commentary on Aristotle's Categories.
 EPICTETUS 49

OEtl'tEpov;]' again and again. The windows are wide open, and someone
hammers out 'If it is day, it is light [Ei tlJlEpa. EO''tt, q>&<; crnv]'. And then off
you go without a care in the world. (eloq v 4 [p.151 van den Hout2]) 106

Fronto is witty and wicked; but there is no reason to doubt that his
description is, au fond, true to the facts.
The exegetical practice was not peculiar to the Stoic schools. The
Peripatetics went in for it, 107 and so did the Platonists. 108 And it was
followed in schools of rhetoric 109 and of grammar 110 as well as in schools of
philosophy. Moreover, what happened in the schoolroom happerted also in
serious symposia, which would begin with the reading of a text and
continue with an informal 'commentary'. 111 And it happened when
philosophers met to discuss philosophy. 112 Thus when Thaumasius attended
a discussion chez Plotinus, he was surprised-and not a little vexed-to
fmd that he was not regaled by an explication de texte. 113
In short, philosophy in the imperial period was, in principle and
primarily, a matter of exegesis, of explication, of commentary. The thesis is
widely acknowledged; 114 and it is, I think, indisputably true. But the truth
demands a couple of glosses which, taken together, make it somewhat less
excitingly true than certain scholars seem inclined to imagine.
The first gloss concerns the notion of exegesis; for the word 'exegesis' is
generous and covers a range of distinguishable practices. At one end of the
range come formal commentaries on classical texts-Aspasius'
commentary on Aristotle's Nicomachean Ethics, which probably dates from
around the end of the first century A.D., is perhaps the earliest surviving
example of this honourable genre. Then there are partial or selective
commentaries-thus Galen wrote a sequence of comments on the medical
passages in Plato's Timaeus. Or again, an exegete might discuss a problem
raised by a classical text rather than a portion of a classical text-

106 Cf Lucian, Hermot 82 [827-828].

107 E.g. Galen, lib prop XIX 42 Kiihn (a pupil reads Aristotle's Cat with a teacher).
108 E.g. in the school of Taurus: Gellius, XVII xx; XIX vi.
109 E.g. Gellius, I iv 8.
IIO The practice of reading followed by commentary would already have been familiar to
philosophy students from their days in the grammar school: on the method of praelectio see
e.g. Quintilian, I viii 13-17, II v 4, and the sample in Priscian, partit [GL III 459-515]-a
schoolmaster's lesson on verses from the Aeneid. - On a higher level see e.g. Gellius, I xxi 4
(Gellius reads Hyginus' commentary on Virgil to Favorinus).
III See e.g. Gellius, II xxii; III xix.
112 See e.g. Plutarch, adv Col1107F; Porphyry, vii Plot 14, 18.
113 Porphyry, vii Plot 13; see Goulet-Caze [1982], pp.268-269; Lim [1993].
114 For some discussion see e.g. Schreckenberg (1966]; Hadot [1987]; Donini [1992];
Dorrie and Baltes [*1993], pp.162-171; Barnes [1993c]; Mansfeld (1994]; Donini [1994],
[1996]; Sedley [1997]. - Note that there were serious reflexions on the nature and aims of a
commentary: e.g. Galen, in Hipp fract XVIIIB 318-323 Kiihn (which is an epitome of Galen's
lost monograph nept ~rm'lcreroc;); see e.g. Erler [1991]; Barnes [1992].
50 CHAPTER THREE

Alexander's Quaestiones provide an illustrative instance. And, at the other

end of the range, a philosopher in tackling some philosophical issue might
take a classical exposition as his starting point or a classical text as his foil
and reference; or, in the course of investigating a philosophical issue, he
might offer comments and glosses upon pertinent passages in such classical
texts-in these ways much of Plotinus is exegetical. The notion of exegesis
is elastic: only if it is stretched, and stretched pretty taut, is it true to say that
imperial philosophy was primarily exegetical.
In the case of the Peripatetics, much philosophizing was indeed done in
the guise of formal commentary on Aristotelian texts: 115 thus a great part of
Alexander's auvre consists of line by line commentaries. But Alexander
also wrote numerous exegetical works which are not in the form of
commentaries (and his commentaries themselves sometimes allow place to
substantial essays). In the case of the Platonists, we hear of fewer formal
commentaries (and none has survived in its entirety); 116 but it is clear that
behind most Platonist texts there lies a Platonic passage or a Platonic
doctrine-and its interpretation. In the other two schools of the period
formal commentaries appear to have been a rarity: among the Epicureans,
the scholarly, philological exegesis of classical texts had long been an
established practice, 117 and there is no reason to suppose that it ceased; but I
do not know of any reference to a formal commentary on an Epicurean text.
For the Stoics, too, there is little evidence for written commentaries-
although here, too, it is clear that there was much exegesis. 118 Perhaps this
paucity of information is merely a matter of chance-after all, most ancient
commentaries (like most ancient texts) are lost; and most of them, we must
surely imagine, are lost without trace. 119 Or perhaps the Stoics-and the
Epicureans-did not in fact go in for formal commentaries? 120 Whatever the
truth of this matter may be, it is clear-from the words of Seneca and of
Epictetus-that Stoic philosophizing was thoroughly exegetical in spirit;
and it seems to me entirely reasonable to imagine that the same was true of
the imperial Epicureans.

115 See Barnes eta/ [*1991], pp.4-14.

116 For the evidence see Dorrie and Baltes [*1993], pp.l71-226; on the rules for Platonic
exegesis see Diogenes Laertius, III 65. -The anonymous commentary on the Theaetetus, the
first pages of which are preserved on papyrus, has been re-edited by Sedley [*1995].
117 See the fragments of Demetrius of Laconia in Puglia [*1988]; cf e.g. Sedley [1989];
Roselli [1989]; Erler [1993].
118 See below, p.73.
119 Lucian wrote a spoof set of Heraclitean texts. He submitted it to a learned
scholar-who duly wrote a commentary on it. The text and commentary, alas, are lost (but I
suspect that fragments are to be found in Diels-Kranz). We know of the episode because it
was alluded to by Galen in his commentary on Hippocrates' Epidemics, the pertinent part of
which survives only in an Arabic translation: see Strohmaier [1976].
12o In which case we might wonder why: for some speculations see Donini [1994],
pp.5089-5093.
 EPICTETUS 51

Imperial philosophy, whatever school it was taught in, was-primarily

and for the most part-exegetical in its approach to the subject. But it was
exegetical in a broad and relaxed sense of the word.
The second gloss is this: philosophical exegesis was no imperial
novelty. 121 In the flrst century B.C. we hear of Crassus reading the Gorgias
with his Academic master Charmadas; 122 and we hear of Epicureans, at
about the same time, reading Anaxagoras and Empedocles. 123 But there is
no need to hunt down recherche examples. Philosophers, after all, had
always been exegetes, and textual interpretation, in one form or another,
had always been a serious mode of philosophizing. The flrst part of Plato's
Parmenides provides an early and a classic case. Or recall Aristotle's
critical reflections on his predecessors in the flrst book of the Metaphysics;
or-in more elaborate form-the critical analyses which the Politics
devotes to themes from the Republic and the Laws. Or think of
Theophrastus' 'doxographical' works, which were both expository and
critical. And if the early Peripatos produced many things of this sort, it was
not foreign to the early Stoa either: Persaeus wrote on Plato's Laws
(Diogenes Laertius, VII 36); Aristo wrote on Zeno and on Cleanthes-and
on Alexinus (VII 163); Cleanthes wrote on Zeno and on Herillus, and also
on Democritus-and four books entitled Explications of Heraclitus,
'Hpad.t:i:tou E~'llriJcrw; (VII 175); Sphaerus wrote on Heraclitus, and on
the Eretrians (VII 178). 124
And so it has always been. After all, the major part of most
contemporary books and articles on philosophy is devoted to a discussion
of X's views about Y's thoughts on the subject of Z.
Was there anything new, in this respect, in the imperial period? The
growlings of Seneca and Epictetus insinuate that the exegetical tum was an

121 Nor, of course, was it a pagan prerogative-it was also a familiar part of the Judaeo-
Christian tradition. An interesting Jewish example is supplied by Aristobulus, a Jewish
Peripatetic who wrote a commentary on the laws of Moses and dedicated it to two of the
Ptolemies (Anatolius, apud Eusebius, HE VII xxxii 16: see Goulet [1989c])---on Jewish
philosophical commentary see e.g. Goulet [1987]; Mansfeld (1988]. As for the Christians,
recall that after Peregrinus had imbibed i] 9etUJletcrti] crO<pia tmv Xptcrttavmv, he became the
complete prophet and priest-'as for books, some he explained and elucidated (E~fi'YEtto Kat
OtEcra<pEt] and many he wrote himself (Lucian, Peregr 11). On Christian commentaries with
a philosophical bent see e.g. Schliublin [1974]; Osborn [1987]; Neuschlifer (1987]; Gamble
[1995], pp.21-28; Simonetti [1996].
122 Cicero, de orat I xi 47.
123 Philodemus, llpoc; toile;[ frag 116 Angeli; cf Sextus, M X 19 (on Epicurus'
schooldays).
124 For later Stoic commentaries note e.g. Boethus on Aratus (Geminus, xvii 48);
Posidonius 'explained the Timaeus' (Sextus, MVII 93) even if he did not write a commentary
on the work (see Kidd [*1988], II pp.337-343), and he also had something to say on the
Phaedrus (Hermias, in Phdr 114) and on the Parmenides (Proclus, in Parm VI 25); Panaetius,
<ptA.oltA.atrov Kett <ptA.aptcrtotEAT]Cj (ind stoic !xi 2-3 Dorandi), had scholarly interests which
will surely have involved some Platonic exegesis.
52 CHAPTER THREE

unhealthy and recent phenomenon. But I incline to doubt it. At most, we

may speak of a change of emphasis, of a growing interest in exegetical
matters, or of an increased tendency to give philosophizing a textual basis.
Such changes and developments are, at the best of times, difficult to
document. Given the state of our evidence, it would be rash to insist that
they mark and characterize the imperial period. 125
However that may be, if exegesis is such a pillar of philosophy, why
shake it? What is wrong with exegesis? In practice, no doubt, much. Great
fleas have lesser fleas; and whether or not the articles in today's journals
advance the careers of their authors (which, to be sure, is their primary
reason for existence), they assuredly do not advance the subject. But other
and grander objections have been raised. In particular, modem scholars
sometimes suggest that exegesis is the enemy of originality, that a philo-
sophy based on commentary must be dull and conservative and tralaticious.
Now in some ancient quarters the exegetical habit was accompanied by a
self-conscious and principled disavowal of originality:

These theories are not new, nor do they date from today: they were expounded
long ago, in a cryptic fashion; and the theories of today are interpretations of
them, showing these doctrines to be old by the testimony of Plato's own
writings. (Plotinus, enn Vi 8, 10-14)

Plotinus is-or pretends that he is-nothing more than an interpreter of

Plato. 126 Why? The pretence is based on a bizarre idea. When we are
puzzled by a philosophical problem,

we think it enough if, when questioned, we can say what the men of old
thought, content to avoid any further inquiry. For we must think that some of
the old and blessed philosophers have found the truth-and it is our task to
consider which of them were particularly successful and how we may gain an
understanding about the matter. (enn III vii I, I 0-16)

125 Hadot [ 1995] speaks of a 'radical distinction' between the method of philosophical
teaching in vogue in the imperial period (commentary on texts) and the pre-imperial method
(dialectical discussion of problems): p.l65. He appeals to Alexander, in Top 27.13; but
Alexander is not reporting pre-imperial practice: he is invoking, hazily enough, the Good Old
Days of Plato and Aristotle, when there were as yet no books to be the object of
commentaries. In fact, Hadot himself acknowledges that 'the literary genre of the
philosophical commentary is very old' (p.232); and the 'radical change', which is now dated
to the first century B.C., consists in the fact that thenceforth 'the very teaching of philosophy
essentially took the form of textual commentary'. The radical change was not so radical after
all.
126 See also Longinus, apud Porphyry, vit Plot 20.
 EPICTETUS 53

The truth has already been found and expressed: we need only discover the
ancient volumes, dust them down, and read the hidden truth.
Plotinus' view of the history of philosophy is absurd. (It is so absurd that
it is tempting to doubt his sincerity. Perhaps it was a pose? or a deliberately
grotesque exaggeration? Yet there is no view so daft that some philosopher
has not adopted it.) However that may be, what of Plotinus' disavowal of
originality? and what of the general idea that exegesis is the enemy of
originality? Here is an argument in its favour: an exegete will, in principle,
produce sentences of the form 'x said that P'; such a sentence will be a
good piece of exegesis if and only if x in fact said that P; hence if the
exegete, in invoking the sentence 'P', offers something novel to the world,
his exegesis is bad. It does not follow that exegesis is the enemy of
originality. It does follow that good exegesis is incompatible with
originality.
But this argument is a mere pedantry: in principle, its conclusion is
inescapable; in the rough and tumble of practice, it has no value at all. For
no exegete, good or bad, limits himself to sentences of the form 'x said that
P'. And even if the sentences of the form 'x said that P' cannot, in a good
exegesis, convey original thoughts, all the other sentences in the exegetical
work may perfectly well do so.
In truth, the idea that exegesis and originality are at odds is false-
evidently and outrageously false. Exegesis may be adventurous stuff; that is
to say, what is passed off by its author as humble exegesis may strike a
reader as bold and original thought, and what its author genuinely takes to
be humble exegesis may in truth be bold and original thought. What, after
all, is more original than the philosophy of Plotinus, which professes to be
no more than an exegesis of ancient truths and which is in fact solidly
founded on a serious study of Platonic literature? 127 No doubt bold exegesis
is usually bad exegesis; and I suppose that, as a general rule, the more
originality a piece of exegesis displays, the further it is from the truth.
Moreover, if you start out from the Plotinian thesis that the truth is to be
found in Plato's texts, then your chances of producing a dispassionate and
rigorously scholarly interpretation are not very high. 128 Certainly, many of
the things which Plotinus claims to be Platonic are not Platonic at all.
It is sometimes maintained that such worries about originality have
nothing to do with antiquity; that, on the contrary, an ancient author was
more likely to lament than to applaud originality. '0, the novelty of it'-
the expostulation was critical, not adulatory. 'Innovating'-KatVO'tOjlEtV,
VECO'tEpt~Et v-was generally regarded as a bad thing, the intellectual

127 Or what more original than Clement's interpretation of the decalogue (strom VI xvi:
see Osborn [1987], pp.l81-182)? Or than Michael Dummett's interpretation of Frege?
128 'Protected by an authoritative text, the interpreter displays a freedom which amazes
and may alienate' (Erler [1993], p.294).
54 CHAPTER THREE

counterpart to treason or to bloody revolution. 129 Perhaps this is true. But it

is a one-sided truth. For if originality or lcatvotOjlta was a crime, so too
was plagiarism-A.oyoxl..o1tia or, more simply, xl..om1. 130 Seneca, snapping
up an Epicurean trifle, might blithely proclaim that 'what is true is mine
[quod verum est meum est]' (ep xii 11); but his proclamation was evidently
intended to shock. Plotinus, who formally abjured originality, was accused
of plagiarizing Numenius: Amelius and Porphyry, far from welcoming the
charge, rose swiftly to rebut it (Porphyry, vit Plot 17-18}--and Porphyry
himself wrote a piece on the vice of plagiarism. 131 Indeed, there was no
commoner charge against a literary man than the charge of xl..om1. And no
commoner boast than the boast of originality-'the passion for novel
thoughts, that is the chief craze of the day'. 132 The day of the author of On
the Sublime fell probably in the first century A.D. The craze was
permanent.
Katvotojlia-A.oyoxl..o1tia. 'You are innovating'-'You are stealing'.
The fact that both accusations could be brought need neither upset nor
amaze-every author is a Mr.Have-it-both-ways, dyed in the wool of
tradition and also as original as sin; and every competent critic will fire his
arrows in both directions. But the prevalence of both accusations does, I
think, suggest that we should take neither of them too seriously. To be sure,
originality was generally despised. To be sure, originality was generally
prized.
However that may be, Epictetus' objections to exegetes have nothing to
do with matters of originality. He complained, first, that philosophers made
exegesis their intellectual goal: they were content to learn that x said that P;
they did not care to know whether P or not-P. (Contrast Plotinus, for whom
exegesis is the shortest-perhaps the sole-path to the truth about things.
For if we can determine that x said that P, then we have thereby determined
that P-provided, that is, that x is Plato.) And Epictetus complained,
secondly, that his little exegetes did not dream of applying their philosophy
to their lives: content to recognize that x said that P, it never occurred to
them that they might conform their actions to the belief that P. The two
complaints are linked; for insofar as I content myself with the thought that,
according to Chrysippus, you should always declare the defects of any

129 See Sedley [1989]; Erler [1993], pp.296-303; on Greek attitudes to originality more
generally see e.g. Lloyd [1987], pp.S0-1 08.- It is sometimes suggested that the veneration for
authority and the rejection of Katv01:o~ia was an imperial phenomenon: even if the
exegetical practice itself was age-old, there was something new in the imperial period-the
conviction that the old texts contained the truth. But, as Erler [1993] demonstrates, this
conviction can be traced back at least to the second century B.C.
130 See Stemplinger [1912]; Ziegler [1950].
131 frags 408-410 Smith= Eusebius, PE X iii 1-26.
132 'to 7tEpt 'tac; vof]aw; Kat v6a7touliov 7tEpt o Iii] ~aA.ta'ta Kopu~avn&at v oi vuv:
[Longinus], v.
 EPICTETUS 55

article which you are trying to sell, I shall have no reason at all to tell you, a
prospective purchaser, that my desirable residence is soft with dry rot. The
exegetical prefix, 'Chrysippus said that ... ', happily insulates me from any
practical implications which the unprefixed doctrine might uncomfortably
have suggested.
I shall not inquire into the merits of Epictetus' complaints-in truth, we
can hardly hope to tell how far he was playing the honest reporter, how far
the rhetorical grouse. But it is worth saying that, within their context, the
complaints are intelligible-the more so, of course, to anyone who holds
that philosophy is indeed the ars vitae. And it is worth adding that-unlike
many of the jeremiads which we read in the Discourses-the complaints
were not commonplaces: it was, to be sure, a commonplace to claim that
philosophers did not live by their philosophies; but it was no commonplace
to claim that exegesis was a dangerous occupation-on the contrary,
Epictetus' claim implicitly challenged the orthodox philosophical
practice of his age.

E: For logic

Exegesis and Epictetus' reflections upon it bear on imperial philosophy as a

whole: I return to the specific case of logic, and to the texts in which
Epictetus appears to exhibit a hostility to the subject. What is the import of
these texts? what practical conclusion did Epictetus rest on his doleful
reflections? Some passages do indeed suggest an outright rejection of logic.
'Let others practise lawsuits, others problems, others syllogisms: you must
practise how to die, to be imprisoned, to be tortured, to be exiled' (IIi 38).
'Problems' here designates physics, and I suspect that 'lawsuits' stands in
for rhetoric; 133 certainly, 'syllogisms' means logic: thus Epictetus appears to
announce that logic is no business of a philosopher. Or again, if you think
yourself educated because 'you have dealt with syllogisms and changing
arguments', then 'you must unlearn all this if you can and begin from the
beginning, aware that up to now you have not even touched on the matter'
(II xvii 27). Or again:

0 the injustice of the educatedP 34 Is that what you learned here? Why won't
you leave the little arguments [A.oycipta.) 135 about these matters to others, to

133 On rhetoric see below, p.58.

134 I.e. the educated act unjustly in not doing with their education what they should-not
'they are treated unjustly' (Schweighauser [* 1799], II p.342).
135 On A.oyapta (which are not 'fine words', pace Souilhe, Laurenti) see also II xviii 26
('all you know are A.oyapta': cf. II x 29-30; xviii 26; xix 22; III xxiv 78-81). The diminutive
form 'A.oyaptov' is thoroughly in Epictetus' style-compare, say, his penchant for
56 CHAPTER THREE

wretched little men who will sit in a comer 136 and rake in their little fees-or
grumble because no-one offers them anything? Why won't you come forward
and use what you have learned? What is wanted now is not little
arguments-the books of the Stoics are loaded with little arguments. Then
what is wanted? Someone who will use them, who will bear witness to his
words in his actions. (I xxix 54-56)

Socrates, whose life is a paradigm for us in all things 137 and who is the
chief hero of the Discourses, did not want to improve his language and his
logic, his A.EI;dota and his 8Erop11Jl6:tta:

if someone asked him for AE~Ei.lha or 9EOOP11J.!Chta, he sent them to

Protagoras or to Hippias. If someone had come to ask him for vegetables, he
would have sent them to a greengrocer. (III v 17)

Such passages-and others could be assembled-appear to offer handsome

support to what I have called the hornless version of the common view of
imperial logic: surely Epictetus, who urges his pupils to avoid syllogisms
and the lawcourts, to fly Venus and phlebotomy, had no time for logic
himself?
These texts should not be brushed aside as casual grumblings, nor yet
discounted as morsels of rhetorical exaggeration. But in point of fact they
offer no support at all to the common view. It is true that IIi 38 and II xvii
27, taken out of context, may appear to be general prohibitions on logical
study. But, read in their context, they serve a different, and a less universal,
end: like many other apparently general exhortations in the Discourses, they
are in fact directed against particular individuals-or rather against
particular types of individual. Thus in II i Epictetus is speaking in particular
of young men, fresh from school, who fancy themselves as bellettrists. Such
young men love to boast 'See how I can compose dialogues': they are
wrong-rather, their boast should be 'See how I can govern my desires and
my passions' ( 35). 'Let others practise lawsuits .. .': Epictetus does not
mean, quite generally, that rhetoric, physics, and logic are subjects
inappropriate to a philosopher: he means, specifically, that young

'crroJlanov', 'KtT]crilltov', 'vuxaptov', and the like; but it is also the traditional word in this
context: see e.g. Marcus, I vii 2 (cited above, p.l ). ta EK tfi<; 7tOtKtA11<; crtou<; A.oyapia are
found already in the comic poet Theognetus (apud Athenaeus, 1048); and they are still found
in Themistius, oral xxxii 3580. Cicero puts the word into Latin as 'ratiuncu/a' (Tusc II xii
29, above, p.IS; IV xix 43; cf nat deorum III xxix 73) or 'interrogatiuncu/a' !fin IV iii 7).
136 Epictetus alludes to Plato, Gorg 4850 (see Dodds ad foe); for Epictetus' frequent
references and allusions to Plato see Schenk! [* 1916], pp.XCII-XCIV. (And see below,
n.l86.)
137 IV v 2: there are more than 50 references to Socrates in the Discourses, all of them
laudatory-more than twice as many as there are to Diogenes the Cynic.
 EPICTETUS 57

philosophy students should ftrst get their desires in order and learn to live
'in accordance with nature'-lawsuits and the like may come later.
Again, in II xvii Epictetus is aiming his darts at the single-minded
logician, at the man who takes himself to be educated simply because he
has mastered syllogistic. If that is your case, then your best plan is to
unlearn everything and begin afresh. There is no suggestion in the text that
logic in general should be abandoned, or that no-one should ever study the
subject.
And so it is, I think, with Epictetus' other tirades against the study of
logic: they touch not the subject but its foolish or foppish practitioners. The
pertinent distinction is made quite clearly by Epictetus himself. 'If you hear
someone singing badly, you don't say 'Look how musicians sing!'-rather,
you say 'He's no musician'-and so it is with philosophy' (IV viii 8-9). 138
And one text explicitly applies the distinction to logic:

If you ask me now 'Are syllogisms useful?', I will say that they are-and if
you like I will prove how [a1toi>Ei~ro 1t&~].-'So what use have they been to
me?'-My friend, you didn't ask if they were useful to you but if they were
useful in general. Suppose that someone with dysentery asks me if vinegar is
useful: I'll say that it is.-'Useful to me?'-No: first stop the discharge and
let the sores heal. 139 (II xxi 21-22)

Syllogisms are no use to the foolish. It does not follow that they are no use
at all. And Epictetus insists that they are in fact useful-and that he can
prove it.
Even in his own time Epictetus was misunderstood on this point. He
observes that you t:nay speak like Demosthenes and still be unfortunate, that
you may analyse syllogisms like Chrysippus and still be wretched.

When I say these things to some people, they think that I am running down a
concern for speaking or for theorems. But it is not this that I am running
down, but rather the habit of being endlessly concerned with them and placing
your hopes in them. (II xxiii 46)

Logic (the concern for theorems 140) and rhetoric (the concern for speaking)
are in themselves good things, things worth studying. But they must be

13 8 Compare the brief text preserved as PSI II 152 (see Concolino Mancini (1980]). I
hesitate to suggest that the papyrus contains a fragment from the lost part of the Discourses.
139 For medical analogies of this sort see I xxvi 15; II xiv 21-22; xviii 8-11; III xxi 15-22;
xxii 73; xxiii 30-31 ('A philosopher's school is a surgery [i.a'tpEiov]'); frag XIX= Stobaeus,
eel III iv 93. They are, of course, the most common of commonplaces: Pollux, onom IV 39.
See e.g. Voelke (1993].
140 Which are not 'scientific theorems' (Souilhe), as the context shows.
58 CHAPTER THREE

studied appropriately-at the right time, in the right spirit, for the right
ends.
Logic is elsewhere conjoined with rhetoric: both were temptations, and
both seduced the young. Just as Epictetus is usually taken to be an enemy of
logic, so he is usually taken to be an enemy of rhetoric. But he was no more
against rhetoric than he was against logic. 141 II xxiii speaks of rhetoric in
general terms; and Epictetus claims that rhetorical competence, i] <ppacrnKi]
ouvaj..Lt<;, is not a thing of vast importance. But he does not deny it all value:
your ears are not of vast importance, nor are your shoes- but it would be
absurd to conclude that they had no value at all. III xxiii discusses display
speeches; and Epictetus vigorously bludgeons those philosophy lecturers
who preen themselves on their linguistic refmement and draw flattering
crowds to their speechifyings. They confuse philosophy with epideictic
rhetoric. Epictetus himself has little time for stylistic exercises, and he is
sure that they are not at all the same thing as philosophical exercises; but he
does not invite us to infer that epideictic oratory has no value at all-nor
even that philosophers should not care how they write or speak.
I viii is concerned with rhetorical arguments ( 7: i] E1ttXEtprrnKi] Kat
1tt8avoA.oytKi] ouvaj.tt<;), and specifically with enthymemes ( 1-3). 142 I
paraphrase the conversation: 'Philosophers, who deal with syllogisms, are
surely competent to deal with enthymemes too?-Then why don't we train
in them?-You must first get your morals right: rhetoric would only distract
you and puff you up with false pride.-But Plato was surely a
stylist.-Yes, but he was not a stylist qua philosopher, any more than
Hippocrates' style belongs to him qua doctor. After all, because I am lame
you don't think that you must become lame in order to be a philosopher.'
That is to say, style is not an essential or a primary attribute of a
philosopher: nonetheless, it may properly adorn the writings of a polished
thinker.
Rhetoric is often abused, and so is logic. But this must be blamed on the
abusers, not on the abused. In II xxiii Epictetus insists that he does not
disparage the subject of logic. There is no reason to doubt that he is
speaking honestly-here at least there is no tinge of irony in his words.

141 See also Musonius, frag 4 [p.19.8-14 Hense]= Stobaeus, eel II xxxi 123. On Epictetus
and rhetoric see e.g. Brancacci [1985], pp.28-32; Brunt [1994], p.42.
142 On enthyrnemes, here characterized as 'incomplete syllogisms [a'tEA.i]~ cruA.-
A.oytcrJlo~]', see Bumyeat [1994], who discovers the Stoic origins of this (unAristotelian)
notion of an enthyrneme.
 EPICTETUS 59

At II xxvi 21 Epictetus offers to prove that syllogisms are useful: at II xxv

he provides a proof not merely that syllogisms are useful but that 'logic is
necessary' .143 I cite the piece in full:

One of the people present said: 'Persuade me that logic is useful. '-'Do you
want me to prove it to you?' he asked.-'Yes.'-'So I must produce a
probative argument?'-He agreed.-'Then how will you know if I produce a
sophism?'-He said nothing.-'You see,' he said, 'you yourself agree that all
this 144 is necessary, since without it you cannot even learn whether it is
necessary or not.'

The model for this pretty little thing is a celebrated argument which was
published in Aristotle's Protrepticus: 'Either you should study philosophy
or you should not study philosophy; but if you ought not to study
philosophy, then-in order to determine that you ought not to study it-
you must study it: hence you ought to study philosophy' .145 Although
Epictetus' argument is closely related to a certain type of self-refuting
argument or 1tEpt-rpo1ti1, 146 I do not recall having seen anything quite like it
in an earlier Stoic text.
It reads, no doubt, like a sophism. The point of Epictetus' initial remarks
is to show that something like the following thesis must be accepted:

In order to know whether or not it is necessary to study logic, it is necessary to

study logic.

Since Epictetus' interlocutor wants to know whether or not it is necessary to

study logic, he must admit that it is necessary to study logic. One way of
representing Epictetus' reasoning is this:

In order to know whether or not it is necessary to study logic, it is necessary to

study logic.
But you ought to know whether or not it is necessary to study logic.

Therefore: you ought to study logic.

The general form of the argument is this:

143 Under the title 'ltro<; avayKaia 'tCt A.oymi'; compare the title of I xvii (em avayKata
'ta A.oymx). For the titles of individual discourses see Schenk! [*1916], pp.LXXVIII-
LXXXII.
144 'ta'iha are 'ta A.oytKa of the title.
14 5 See Alexander, in Top 149.9-17-other texts are collected in During [*1961], pp.44,
113-114. On the fortune of the argument see most recently, O'Meara [1994].
146 See above, pp.30-31.
60 CHAPTER THREE

Necessarily (ifP, then Q).

It ought to be the case that P.

Therefore: it ought to be the case that Q. 147

And it is a nice question whether such a form is valid or not.

No doubt the considerations which Epictetus advances in favour of the
first premiss of the argument are frail. You might, for example, think it
pertinent to inquire whether all or most trained logicians are rich and happy
rather than poor and miserable; and you might think that you can determine
the truth about this without worrying over the possibility of sophistry and
delusion. In addition, you might fmd something paradoxical -even self-
defeating-in Epictetus' argument; for how, ifl do not yet know any logic,
can I be sure that my logic teacher is not offering me sophistries and
delusions? But whether or not the argument is ultimately any good, its
presence in the Discourses is striking and instructive.
Moreover, there are more important-and more Epictetan-texts than II
xxv. In several passages Epictetus states clearly and positively that we
ought to study logic. After all, our reasoning faculty, iJ A.oyuo)buvaj.tte;, is
'the best of all things and master'; and it alone has Zeus placed within our
power (I i 7). It is master because, alone of the faculties and arts, it
'contemplates both itself and everything else' ( 4)-if it did not
contemplate itself, then there would be an infinite regression of con-
templators (I xvii 1-3). 148 Plainly, then, we should devote ourselves to its
study: Chrysippus and Cleanthes and Zeno, the scholarchs of the Old Stoa,
did so-and so too did Diogenes' teacher Antisthenes, and Socrates. 149 The
appeal to Diogenes (by way of Antisthenes) and to Socrates is not casual: if
Epictetus' two paradigms of human perfection engaged in logical study,
then how can we abjure it? (Chrysippus himself had defended the study of
dialectic against its detractors by invoking the precedent of the Academy
and the Lyceum-and in particular of Socrates. 150)
It might be objected that these programmatic passages in I i and I xvii do
not refer specifically to logic: rather, they refer in general to the reasoning
faculty, and if Epictetus has anything specific in mind then it is surely the
role of reason as the employer of impressions (see I i 5). The message of I i,
in brief, is not 'Study logic', but 'Use your reason'.

147 For 'P' substitute 'You know whether or not it is necessary to study logic'; for 'Q'
'You study logic'.
148 See I xvii 1-3; xx 1-6; II xxiii 5-13.
149 I xvii 12: for Socrates, see above, n.137; there are another four references to
Antisthenes in the Discourses.
150 See Plutarch, stoic repugn 1045E-1046A.
 EPICTETUS 61

Now it is true that, in Epictetus' view, the correct use of impressions, f]

6p9iJ XPflcrt~ 'trov <pav'tacrirov, is the primary task of reason; indeed, in a
certain sense it is the only task of reason and the only task for men. 151 But
there is no real contrast between the study of logic and the rational use of
impressions: in I xvii it is plain that Epictetus is thinking primarily of Stoic
logic; and the close parallel between I xvii and I i makes it certain that the
study of logic is at any rate encompassed by the general notion of the
rational use of impressions.
The decisive text is I vii. 152 The formal aim of the discourse is to show
that, contrary to the opinion of most men, the study of certain apparently
esoteric parts of logic-of hypothetical arguments, of changing arguments
-is useful. In fact, Epictetus' conclusion is more general and more
substantial than this formal aim suggests: he concludes that we ought to
study logic. It may be worth giving a summary of the whole discourse:

The study of changing arguments and the like is in point of fact concerned
with what is appropriate. For in every area we need to know how a good man
can discover the appropriate way to behave. ( 1-2) Hence unless you
suppose that a good man will never engage in argument or will not care how
he behaves in argument, you must allow that logic should be studied. ( 3-4)
Now the point of logic is to enable us to grasp the true, reject the false, and
suspend judgement on the unclear. But it is not enough simply to do this: you
must be able to test arguments and statements so that you can ensure your
success-you need a faculty or skill: getting things right by luck is not
enough. (5-8) Since logic is not only concerned with grasping the truth but
also with accepting what follows from what you have already accepted, you
must further learn under what conditions one thing follows from another-
hence the necessity for logical study and logical practice.( 9-12)
Changing arguments offer a peculiar difficulty; for with them it seems that
you can start from true premisses and proceed by valid methods to a false
conclusion. Hence it is important to survey the varieties and to grasp the
nature of such arguments. ( 13-21) There are also special difficulties with
hypothetical arguments-what hypotheses should you accept, and what
attitude should you adopt to the consequences of those hypotheses which you
have accepted? ( 22-29)

151 See e.g. I iv 4; xii 34; xx 5, 15; xxx 4; II i 4; xix 32; xxii 29; xxiii 42; III i 25; iii I; xvi
15; xxii 21; xxiv 69; IV iv 28; v 23; vi 25, 34; x 13, 26; ench 6; frag IV= Stobaeus, eel II viii
30; frag VIII = Stobaeus, ec/IV xliv 60. - Animals, too, use their impressions; and strictly
speaking what marks off men is not the use of impressions but the self-conscious use of them
(the fact that men 'follow', 1tapaKol..ou9Ei:v, their impressions): I vi 12-21; II viii 4-8; cfii vi
14; IV vii 7, 32. When Epictetus speaks simply of use, we should no doubt take him to mean
self-conscious use.
152 Full text and translation in the Appendix.
62 CHAPTER THREE

Idleness is always with us; but it is no good saying, after a mistake in logic,
that 'it is not as though I have killed my father or burned down the Capitol'.
As Musonius once said to me: an error in logic is an error-we must therefore
work energetically in logic, as in any other matter, in order to do what it is
appropriate to do. ( 30-33)

Logic has turned out to be necessary: 'And has there not arisen among us
study and training in valid arguments and schemata, and has it not proved
necessary?' ( 12). Our conception of the Sage is such that he must be a
logician: 'Will he be able to preserve coherence without some training and
preparation of this sort?' ( 29). And so we too must train in logic: 'Why
are we still lazy and idle and dull? why do we look for excuses for not
working-or even staying awake-and developing our reason?' ( 31 ).
'Using your own impressions at random and in vain and at haphazard, not
following an argument or a proof or a sophism, in general not seeing, in
questioning and answering, what accords with your position and what does
not-is none of this a mistake?' ( 33).
Logic is morally required, it is morally indispensable. We are essentially
rational beings. Our nature thus requires us to develop and to exercise our
rational capacities. We cannot shirk logical argument or refuse to engage in
logical debate. And if we do engage, evidently we must do so with expertise
and not at haphazard. Hence we must acquire the pertinent expertise-we
must learn logic.
Epictetus' argument might be challenged by a determined enemy of
logic. But it is his conclusion which concerns me here. And the conclusion
is plain as pie.

F: The place of logic

It is our moral duty to learn logic. Why and to what end? It is tempting to
ascribe to Epictetus the utilitarianism which I earlier ascribed to Seneca.
After all, Epictetus complains that philosophers do not apply logic in their
lives; the complaint suggests that a chief function of logic is to serve
morality-and thence it is an easy slide to the conclusion that logic is to be
studied only insofar as it may subserve a moral end. Epictetus will then be a
logical utilitarian. 153
No doubt he will have inherited his utilitarianism from his teacher,
Musonius Rufus. In one of the surviving fragments, Musonius maintains

153 So e.g. BonhOffer [1890], p.3: Epictetus has 'a strictly practical interpretation of
philosophy'; p.l9: logic has a purely instrumental function.
 EPICTETUS 63

that women should be allowed to study philosophy. 154 He is then faced with
the telling objection that in such a case women 'will practise arguments and
engage in sophisms and analyse syllogisms when they should be sitting at
home with their embroidery'. His reply to the objection is instructive: his
philosophical women will not neglect their needles. After all, not even men
will devote themselves to logic in a whole-hearted way; rather, 'I say that
when they deal with arguments they should deal with them for the sake of
their actions'. In other words, men-and a fortiori women-will learn logic
only insofar as it may be applied to ethical matters and thus contribute to
moral virtue. Musonius was a utilitarian; and for him logic was the servant
of ethics.
Did Epictetus take this line? Well, one of the central functions of logic,
according to him, is that of providing proofs; and in particular, logic will
enable us to produce proofs in ethics-for example, a proof that we should
not telllies. 155 But logic does not supply proofs in ethics alone (as Musonius
perhaps maintained 156): we may and should also seek proofs in physics.
Thus

if some people can hold this view of things from madness and from habit, 157
can no-one learn by reason and proof [\mo A.6you ... Kat n7tolii~Eoo~ that
god has made everything in the world and the world itself, the whole
ineluctable and perfect, the parts serving the needs of the whole? (IV vii 6) 158

Logic provides proofs in physics too; that is to say, logic serves the needs of
the other parts of philosophy.
Now not every part of standard Stoic logic contributes to the production
of proofs. Indeed-as opponents of Stoic logic frequently observed- , very
much of what the Stoa studied under the heading of logic cannot
conceivably serve the end of scientific proof. Epictetus is a.s aware of the

154 frag 3 Hense= Stobaeus, eel II xxxi 126.

155 Above, pp.38-39.
156 In frag I Hense = Stobaeus, eel II xxxi 125, Musonius acknowledges that it is
necessary for us to use proofs (only the gods, for whom nothing is iili11Aov. have no need for
(moliEt~tc;); but ~Tlliev iiyav---a philosopher should offer his pupils a few clear proofs for
each doctrine, and it is a fault to compile a whole dossier of proofs for a single thesis. The
examples which Musonius gives are similar in their brevity to the Zenonian syllogisms
(above, p.l5; below, pp.71-72); they all concern morals-and perhaps Musonius implies that
all the proofs he will offer fall in the area of ethics. (On Musonius and proof see, with
caution, Laurenti [1989], pp.2106-2113.)
15 7 Perhaps Epictetus is thinking of the Christians-but 'oi ral..tl..ai:ot', which appears in
the MSS, is an evident gloss (as Schweighauser suspected: [*1799], II p.915). Compare the
similar Christian gloss in Marcus, XI iii (on which see Brunt [1979]).
158 See also, for proofs, I xxv 6; III xxvi 16. But the word 'lm61iEt~tc;' in Epictetus does
not always signal the presence of logic: a 'proof may be provided by a style of behaviour
rather than by a formal argument (e.g. I xxiv 8; xxix 33; II xvii 16-18; III x 8; xxii 69; xxiii
14; xxiv 112; IV vii 29).
64 CHAPTER THREE

fact as anyone else. Does he then, like a good utilitarian, abjure the 'useless'
parts of the subject? Not at all. Contrasting philosophers with idle aesthetes,
he remarks:

You, when you have nothing to do, are restless, you go to the theatre, you
wander about. 159 Why shouldn't the philosopher work on his reason? You
have your crystal ware, I my Liar; you have your porcelain, I my Denier. (III
ix 20-21)

Epictetus has his Liar: the paradoxes and puzzles are to him what fine
porcelain 160 is to the aesthete-he collects them, and they divert him in his
empty moments. But his 'empty moments' are not empty, and the philo-
sopher's diversions have a serious end: the paradoxes are used to exercise
and improve his reasoning powers. 'To you everything you possess seems
trivial, to me all my possessions seem grand' ( 22). The paradoxes are not
trifles.
This is not the bluff utilitarianism of Galen. Epictetus does not reject
certain parts of traditional Stoic logic on the grounds that they are useless
-indeed, there is no part of Stoic logic which he expressly rejects. But
perhaps Epictetus supports a refined utilitarianism? He works on items
which Galen regards as useless; but he works on them precisely because, in
his eyes, they are not useless. 161 If they do not directly promote the great
end of logic, which is the establishment or confirmation of physical and
moral doctrine by logical proof, nevertheless they promote it indirectly-
they train the philosophical mind, which may then grasp proofs the more
firmly and produce proofs the more readily. All the parts of logic have a
value-but the value is always instrumental. 'The third area is necessary
because ofthe second'. 162
I do not think that Epictetus is a utilitarian, not even of the most refmed
sort. He says that the puzzles may help us to 'work on' our reason; but there
is no need to give this the gloss which I gave it in the last paragraph. We
have a duty to work on our reason, and the perfection of our reason has a
value of its own: when, in an idle moment-which is not, to be sure,
genuinely idle- , Epictetus reflects on the paradox of the Liar, he is indeed

159 S has 'ava/.:6T)'tE', which the apographs 'correct' to 'ava/.:6E'tE'-accepted by

Schenk!, who interprets it as av-ai..iJE'tE. The word is not otherwise known-and the pun
seems to me unEpictetan. I read 'ai..iJE'tE' (Salmasius): for the word in the sense of 'wander
about aimlessly', 'twiddle your thumbs', see e.g. Dio Chrysostom, vii 79, with Russell ad foe.
Note also the medical sense of the word, already in Hippocrates: e.g. Erotian, s.v. ai..ucrf!OV.
160 'Porcelain' represents 'f!oilpptva'--objects made of felspar, quorum accendit
fragilitas pretium (Seneca, ben VII ix 3): the things were first brought to Rome by Pompey
for his triumph of 61 B.C.: see Pliny, nat hist XXXVII vii-viii 18-22.
161 Cfl vii, title: below, p.l36.
162 ench 52.2: cited above, p.38.
 EPICTETUS 65

training his intellect for future and more serious battles. But he is also doing
more than that: he is doing something which is valuable, and morally
valuable, in and for itself. 163
The passage from the Encheiridion seems more determinedly utilitarian.
But the Encheiridion was put together as a vade mecum, and we should be
wary of discovering in its pages rigorously articulated doctrine. In any
event, I vii tells a different and a more nuanced story. In 1, Epictetus
observes that most people fail to realize that logic is 'concerned with what it
is appropriate to do [1tEpt 1W.911K6vto<;]'. He does not mean, in the
utilitarian vein, that logic must be applied to moral matters. Rather, he
means that logic is itself a moral matter. The wise man must engage in
logic; and when he does so he is morally obliged to argue correctly. There
is no suggestion that the wise man will only engage in argument when
ethical issues are the subject of discussion-the subject matter is irrelevant.
For a mistake in logic is itself a moral error-every mistake is a moral
error.
No doubt logic has value only insofar as it bears upon moral
comportment. But then absolutely anything you may do--and hence
absolutely anything you may do with or to a syllogism-is a piece of moral
comportrnent. 164 Thus Epictetus may take a moralistic view of logic without
taking a utilitarian view. 165 His attitude to logic is not Seneca's attitude to
logic: it is comparable rather to Seneca's attitude to physics. 166
I call this attitude 'moralistic'. You might reply: 'Any rational man must
think that, if he is going to argue at all, he ought to argue correctly. What is
moralistic about that?' But this presupposes a false account of Epictetus'
attitudes. Epictetus thinks, first, that you ought to argue, and that you ought
to study logic. He does not think, merely, that if you argue, then you ought
to argue well. His imperative is categorical, not hypothetical. Secondly, he
thinks that the sense of 'ought' in which you ought to do logic and you
ought to argue correctly is the same sense as the sense in which you ought
to tell the truth and you ought to honour your parents. It is the peculiar view
ofEpictetus---or, more generally, of the Stoics-that you ought, morally, to
get your syllogisms right.

163 Cf Quintilian, I x 5: 'It is not that Homed Arguments and little Crocodiles can make a
man wise, but that he ought not to be mistaken even in trifles'. - But for Epictetus, no
mistake is a trifle: I vii 30-33 (below, p.l34).
164 The pyov of man requires, inter alia, that we be 'undeceivable in affirmations and
suspensions of judgement' (I iv II )-hence the importance of logic in the matter of avoiding
the traps of sophists (I xxvii 6; III vii I). Note also that 'conditionals are
indifferent-judgements about them are not' (II vi I).
165 And in fact the same may, for all we know, be true of Musonius: the surviving
fragments are not extensive enough to determine the issue.
166 But the analogy is not exact: Epictetus does not hold that the study of logic is itself a
part of the good and happy life; rather, the study enables us to engage in a form of intellectual
activity which is part of such a life.
66 CHAPTER THREE

II xxv argues that logic is necessary. It suggests further that logic is also in a
certain sense primary-that if you have not studied logic, you cannot study
anything at all. For any study will presumably proceed by argument, and
you will be safe among arguments only if you have studied logic. What is
suggested by II xxv is stated elsewhere in the Discourses. Thus at II xxiv
13-15 Epictetus explains to a would-be pupil that he cannot teach him
philosophy since he is unable to prove anything to him; and he is unable to
prove anything to him (or even to expose and resolve the contradictions in
current opinions) because he does not yet know what a proof is (or what it
is for one opinion to conflict with another).
The starting-point of philosophy is 'the perception of conflict [atcr9TJcrtc;
flUXTJc;]'. 167 Men's opinions conflict: we must first recognize the fact, then
understand what conflict consists in, and finally discover a criterion with
which we may resolve such conflicts. The message is clear: logic is the first
subject which an aspiring philosopher must master. To be sure, it is not at
all clear why philosophy does or should start from an observation of
conflicting opinions, from ota<provia-why should it not burgeon (as
Aristotle thought) from a natural curiousness? But the view that conflict
provokes philosophy was widespread in antiquity, 168 and it was certainly
Epictetus' view.
Some of the passages in which Epictetus laments that the young do logic
and nothing more than logic suggest that it was common practice in his day
to teach logic first. And one text explicitly confirms the suggestion:

The philosophers exercise us first in theory, where things are easier, and then
lead us to what is harder; for there nothing disinclines us to follow what we
are taught, whereas in matters to do with real life there are many distractions.
(I xxvi 3-cf 14)

By 'theory [8Eropia]' Epictetus means formal logic (see 1). He is not

suggesting that logic is an easier subject to grasp than physics or ethics:
rather, it is easier to teach inasmuch as the teacher has no pre-existing
prejudices or natural inclinations to combat. 169 And although he is here
reporting the practice of his contemporaries, he does so without any hint of
irony or criticism.

II xi 13-18. For f!UXT\ see also I v 3, 8; IIi title; xvii 14; xix 1-4; xxii 1-8; xxvi 1-3.
167
See esp Sextus, PH I 12; and for the importance of Ota<pmvia in Stoic philosophy see
168
Mansfeld [1989], [1990).
169 See above, p.39.
 EPICTETUS 67

Thus Epictetus implicitly takes up a position in an old debate: the debate

about the correct order of the three parts of philosophy. 170 The question at
issue is in principle pedagogical and expository: in what order must or
should pupils imbibe the parts of philosophy? in what order must or should
the parts be set out in a systematic account of philosophy? Some of the
considerations which are pertinent to the issue are of a psychological
nature-so, for example, the considerations mooted in I xxvi. But there are
also pertinent considerations which are philosophical rather than psycho-
logical in character, insofar as they concern the logical and conceptual
relations among the several parts of philosophy. In general, you might
maintain that the study of one subject is prior to the study of another
inasmuch as a grasp of certain truths of the latter logically presupposes a
grasp of certain truths of the former (but not vice versa). In particular, you
might maintain on such grounds that the study of logic is prior to the study
of physics: in order to grasp certain truths of physics you must, say, be able
to construct a certain type of proof, and in order to construct proofs you
must have a grasp of certain truths of logic; on the other hand, nothing in
logical study presupposes a grasp of anything in physics. Hence logic
should be taught-must be taught-before physics (or at least before some
parts of physics).
I spoke of a debate on the question: in fact, pretty well everyone agreed
that logic should come first. To those philosophers who took logic to be an
opyavov or tool of philosophy, the primacy of logic must have seemed
evident; 171 for we need the tools in order to start the job. But of course all
philosophers, whether or not they regarded logic as an essentially
instrumental thing, supposed that it served the other parts of philosophy
insofar as it formulated arguments and proofs; and Epictetus can use an
analogy closely related to the analogy of the tool or instrument: 'Logic is a
power of discriminating and examining other things-and, as it were, of
weighing and measuring them.' 172
Now someone might say that there were more pressing subjects than
logic. But in examining any subject whatever we shall perforce use pro-
positions and arguments-and we shall need to know if the terms we use
are ambiguous and if the arguments we produce are valid.

170 See e.g. Sextus, PH II 13; M VII 20-24; Plutarch, stoic repugn I 035AF (with Chemiss
ad /oc).
171 The view is ascribed explicitly to Andronicus (see e.g. Philoponus, in Cat 5.18-23);
but there is no reason to think that it originated with him (see Barnes [1997]). If it was not
universally shared among later Peripatetics (Boethus wanted to start with physics:
Philoponus, in Cat 5.16-18), it seems to have been generally accepted.
172 I xvii I 0; for the analogy with weights and measures see also I xxviii 28-29; II xi 13-
24. But the thing is a commonplace: see e.g. Zeno, apud Stobaeus, eel II ii 12 (above, p.7
n.35); Aristocles, apud Eusebius, PE XIV xviii 10; Sextus, PH II 15; M VII 31-33; Ptolemy,
crit i 6-7.
68 CHAPTER THREE

That, I think, is why they put logic first-just as we examine the measure
before measuring the grain. For if we do not first determine what a bushel is
and do not first determine what a balance is, how shall we ever be able to
weigh or measure anything? So here if we have not learned-and learned
accurately-the criterion through which other things are learned, shall we be
able to learn accurately any of those other things? How could we? (I xvii 6-9)

If we are to know anything at all, we must first acquire a criterion of

knowledge; and that criterion is provided by logic.
All this seems clear enough. But certain other passages in the Discourses
appear to contradict it. 173 Recall, first, that logic belongs to the third of the
three 'areas' which Epictetus sometimes distinguishes. 174 Now 'the third
area is appropriate to those who are already making some progress' (III ii
5-and see 17). That is to say, logic should be studied only after the other
areas have been mastered-or at least attacked. Why so? Because the
function of logic is to give us 'security in these things, so that even when
we dream or are drunk or are melancholy, an untested impression may not
creep in without our noticing it' ( 5). The task of logic, as Epictetus here
explains it, is to secure our opinions: evidently, then, before we embark on
logic we must have already acquired some opinions which are worth
securing.
The point is made at greater length at III xxvi 14-20, where the analogy
with weighing and measuring recurs with an opposite intention.

Have you ever seen anyone building a coping 175 without putting it on a
wall? 176 What porter stands where there is no door? But you are practising
your skill at proofs-at proving what? You are practising so as not to be
dislodged by sophisms-dislodged from what? Show me first what you are
guarding-what you are measuring and weighing. Then show me the scales
and the bushel. (III xxvi 15-17)

173 Hence Schweighauser [*1799], II p.l99, insists-against the run of the text-that I
xvii 6 does not represent Epictetus' own view.
174 Above, pp.34-35.
175 'tpt"yx6~ (or 8ptyK6~): the word may mean 'wall', 'enclosure'; and it is tempting to
suppose an allusion to the Stoic fancy that logic is like a garden-wall (e.g. Sextus, M VII 17:
above, p.21 ). The text of III xxvi 15, which is uncertain, will then be reconstructed and
interpreted with this image in mind. But in fact Epictetus is plainly alluding to Plato: at Rep
534E (a much cited text: see Soter [1989], pp.342-343) otaA.eK'ttKij is the 8ptyK6~ or coping
set on the propaedeutic J.la8ijJ.la'ta-and hence, pertinently to our text, it is the last discipline
to be studied in Plato's educational programme.
176 Reading 'J.l110EVt 'tEtXtCfl' (Sb): the text is uncertain, the general drift plain.
 EPICTETUS 69

Epictetus means, I take it, that you cannot, as a matter of logic, set a coping
without having built a wall, or hire a porter (9uprop6c;) without setting him
by a door (eupa); similarly, you cannot, as a matter of logic, prove or
defend without having a thesis to prove or defend. The Encheiridion makes
the same point without the similes. The third area of philosophy tells us
what proofs are-hence 'the third area is necessary because of the second,
and the second because of the ftrst; and the ftrst is the most necessary and
the area in which we should take our rest' .177
Thus, according to Epictetus, we must study logic before we study
anything else; and equally according to Epictetus, the study of logic is the
last thing we should undertake in our intellectual careers.
Epictetus apparently contradicts himself. Can the contradiction be
resolved? No text explicitly addresses the issue, and I have not found
anything in the Discourses which delivers a satisfactory resolution.
It has been suggested that we need to distinguish between two levels of
logical study. First, there is the acquisition of an elementary knowledge of
the subject-the sort of thing which modern undergraduates customarily do
(or fail to do) in their ftrst year of philosophy. This sort of logical study,
according to Epictetus, must be acquired before any serious progress can be
made in the rest of philosophy, and this sort of logic is primary. Secondly,
there is the higher study of logic-the sort of thing which graduate students
undertake and which eventually clogs up the professional journals. This sort
of logical study is, as it were, a luxury: it should only be undertaken by
those philosophers who have already established their moral credentials. 178
Epictetus does not himself make this distinction: should he have done
so? Not, I think, in precisely the form I have given to it; for there is no
suggestion in the Discourses that the logical studies taken late and
luxuriously involve serious and esoteric work. In III ii, it is true, the items
which we are urged to postpone are hypothetical arguments and changing
arguments( 17); but III ii and III xxvi also talk of proofs-and most ofthe
proofs with which a moralist or a physicist will be concerned will have a
fairly elementary logical structure.
There is another distinction, related but different, which Epictetus does
make-or rather, which is implicit in his various remarks. It is a dis-tinction
not between two levels of logical study but between two uses of logic.
Roughly speaking, you may regard logic an an instrument of dis-covery and
you may regard logic as a preservative of knowledge. On the one hand,
logic may be used in the gaining of new items of knowledge- either
positively, insofar as it enables you to infer one belief from another, or
negatively, insofar as it enables you to avoid the various snares and

177 ench 52.2: above, p.38.

178 See BonhOffer [1890], p.20; [1894], p.l23.
70 CHAPTER THREE

sophisms which are set along the path to knowledge. On the other hand,
logic may be used to furnish the proofs which are supposed to consolidate
and to systematize the knowledge which you have already acquired. The
standard Aristotelian view of logic makes logic primarily a preservative, 179
and the Old Stoa seems to have regarded logic rather as a method of
discovery; 180 but there is no reason in the world to choose between these
two uses-they are complementary rather than concurrent. Epictetus
evidently had time for each of the two uses of logic. Inasmuch as logic is
useful in learning new truths, you must learn some logic before you embark
on other disciplines-or at least, learn it pari passu with the other
disciplines. Inasmuch as logic is useful in preserving old truths, you will
have use for it after you have acquired knowledge in some other disciplines.
It is thus reasonably clear what should have been Epictetus' position in
the debate over the priority of logic. He should have said that logic must be
acquired as soon as anything is acquired; and he should have added that
certain parts of logic cannot be exercised until other items of knowledge
have been amassed. We can make Epictetus say what he should have said if
we treat as rhetorical exaggeration those passages in which he says that
logic should not be studied until you have acquired some other items of
knowledge: he says that logic must not be done until you have some items
of knowledge on which to exercise it; he means that logic must not be done
unless you are going to acquire some items of knowledge on which to
exercise it.
Did Epictetus 'really mean' this? Plutarch accuses Chrysippus of
precisely the inconsistency which we find in the Discourses. He also
remarks that

in On the Use ofReason Chrysippus writes that, when you take up logic as the
first subject, you should not altogether abstain from the others but rather take
part in them too to the extent that it is offered to you. (stoic repugn I 035E)

For Chrysippus, I suppose, 'Learn logic first' was a slogan, not a doctrine:
the doctrine-if such it may be called-was the more nuanced view
expressed in On the Use of Reason. It is not absurdly charitable to suppose
that Epictetus held a similar view-which was no doubt as orthodox as it is
banal.

179 See Barnes [1969).

180 See Barnes [1980].
 EPICTETUS 71

G: The syllabus

An ancient philosophical education normally included instruction in both

the ancient forms of logic: in Stoic hypothetical syllogistic and in Peri-
patetic categorical syllogistic. (Categorical syllogistic is the theory of
categorical arguments, hypothetical syllogistic the theory of hypothetical
arguments. An argument is categorical if all its component propositions-
its premisses and its conclusion-are categorical propositions; and an
argument is hypothetical if at least one of its components is a hypothetical
proposition. A proposition is hypothetical if it is a compound of two or
more propositions-a conditional, a disjunction, a conjunction. A
proposition is categorical if it is not hypothetical. 181 ) Both syllogistics are
found in Galen's Introduction to Logic, both in Alcinous' Didaskalikos-
and Aulus Gellius manifests a gentlemanly acquaintance with the elements
of each. 182 Some logicians regarded the two logics as in some sense rivals:
thus a Peripatetic or a Stoic might claim that the logic of his own School
was complete and perfect, and that everything in the other system was
either reducible to something in his own or else superfluous. 183 But
generally the two syllogistics were treated as complementary rather than as
concurrent, and a student oflogic would study both. (And even if you were
a logical partisan, you would no doubt have studied your opponent's
arsenal.)
In his discussion of the use of proofs, Musonius offers two illustrative
arguments.

If someone takes the known assumption that everything good is choiceworthy

and further assumes another known item, that some pleasures are not
choiceworthy, we prove that pleasure is not good .... Positing the evident
assumption that everything bad is to be avoided, and positing further the more
evident assumption that many pains are not to be avoided, it is inferred that
pain is not bad. (frag 2 [p.2.12-15, 19-22 Hense]= Stobaeus, eel II xxxi 125)

The arguments are set out with reasonable formality; but they are not
decked in Stoic dress. Rather they are-mildly heterodox-Aristotelian
syllogisms in Baroco. 184 There are parallels to Musonius' examples in the

181 See e.g. Galen, inst log vii 4. - The characterisation I give in the text is rough, but
there is no need here for irritating precision.
182 See Holford-Strevens [1988], p.l69.
183 See e.g. Mueller [1969]; Frede [1974b].
184 See Aristotle, APr 26a36-b I. The schema is:
72 CHAPTER THREE

textbooks; 185 and it seems unlikely that this is an accident. I infer that
Musonius knew his categorical syllogistic-and that he probably taught it in
his school. The conclusion should not surprise.
Now in Epictetus there is no explicit reference to Peripatetic logic; nor
have I noticed any implicit reference-no casual allusion or turn of phrase
which might betray a knowledge of the subject. Perhaps this is due to
chance, the lost portions of the Discourses containing some hints at
Aristotle's logic. Or perhaps Epictetus knew categorical syllogistic and-
for some reason we cannot divine-deliberately hid his knowledge. Or
perhaps Epictetus was simply not acquainted with the Peripatetic form of
logic. 186 However that may be, for our purposes the logic shown up in the
Discourses is Stoic logic.
And in the first place it is the logic of the Old Stoa. 187 Epictetus refers by
name, in logical contexts, to three earlier logicians: to Chrysippus; 188 to
Antipater; and to Archedemus. 189 And to all three together-alongside

B holds of every A.
B does not hold of some C.

Therefore: A does not hold of some C.

The 'mild heterodoxy' consists in the fact that Musonius expresses the two conclusions
without an explicit quantifier.
185 E.g. Galen, inst log vii 6.
186 In all the Discourses there is only one reference-a slighting reference-to the
Peripatetics (II ix 20-22), and I find little reason to think that Epictetus had conned his
Aristotle, whom he never even names. (The importance of 1tpo<xtpcrtc; in Epictetus' moral
philosophy may make modem scholars think of Aristotle: no reason to imagine that ancient
thinkers would have had the same reaction; and even if the remote origin of Epictetus' ideas
in this area is to be found in the Nicomachean Ethics, that work was certainly not his
'source'-see e.g. Dobbin [1991].)- For Epictetus' knowledge of Plato see above, n.l36;
there are also a couple of references to Polemo, and one to Xenocrates. On the sceptical
Academy see I v and II xx; and note the crossing of swords with Favorinus: Favorinus wrote
llpcu; 'E1ttK'tll'tOV (Galen, opt doct I 41 Kiihn), and Galen replied on Epictetus' behalf (lib
prop XIX 44 Kiihn); see also Gellius, XVII xix = frag X. (And there is one allusion to the
Pyrrhonists.) The Epicureans are coupled with the Academics in II xx, and they get a dozen
further references, mostly scathing. In connexion with the Master Argument Epictetus refers
to the 'Megarics' Diodorus Cronus and Panthoides (see above, p.44)---but it is reasonable to
assume that what he knew of them he gleaned from Chrysippus' works. In general, the
Discourses betray little interest in or knowledge of non-Stoic philosophy. But this is, in itself,
anything but remarkable; and it would be rash to infer that the students in Epictetus' school
were ignorant of every philosophy but Stoicism, and equally rash to infer an ignorance (or
even a lack of interest) on Epictetus' own part.
187 In general, Epictetus follows the Old Stoics, as Gellius, XIX i 14 (= Epictetus, frag IX)
remarked: 8 references to Zeno, 9 to Cleanthes, 17 to Chrysippus, 1 to Diogenes of Babylon,
5 each to Archedemus and Antipater; but figures of this sort are at best a crude
indication-the nature of the references shows that Chrysippus towers over the rest. No
reference to Panaetius or to Posidonius. Two references to Euphrates, 6 to Musonius.
188 For Epictetus' knowledge ofChrysippus see Hershbell [1993].
189 I xvii 13; II xvii 34, 40; III xxi 7 (Chrysippus); II xvii 40; III xxi 7 (Antipater); II xvii
40; III xxi 7 (Archedemus ).
 EPICTETUS 73

Cleanthes-in connexion with the Master Argument (II xix 8-10). There is
also a reference to Crinis-and although Epictetus does not refer to him for
his logic, he does refer to him as a logician. 190
The paucity of references is not surprising. The Discourses is not a set of
scholarly essays, and Epictetus had surely read-and had got his pupils to
read-numerous items to which his recorded conversations do not allude. In
any event, who is missing? Zeno-who did little for logic. 191 Diogenes of
Babylon-whose interests (or whose special contributions) lay rather in the
study of language than in logic proper. 192 Apollodorus of Seleucia193 and
Dionysius of Cyrene, 194 both of them pupils of Diogenes. Posidonius-who
certainly interested himself in logic. 195 Otherwise, no-one we can readily
name. 196 But there were indubitably other books whose authors have
disappeared together with their works. For although we should not imagine
that the 'Stoic books' loaded with little arguments to which Epictetus once
alludes 197 were books on or about logic, there will have been many Stoic
books on the subject. Perhaps there were some formal commentaries by
Stoics on classical Stoic texts: at any rate, Aristocles, who wrote a
commentary in four books on Chrysippus' How we speak and think of
things, 198 surely had something to say about logic. 199 And there were
handbooks-introductions or outlines, Eiaayroya\. or U1tO'tU1troaw;. 200

190 III ii 15; cfDiogenes Laertius, VII 71 (alhaAEKttK!l 'tEXVll); see Guerard [1994a].
191 See above, p.7; Rist [1978a]-who, however, ascribes far too much logic to Zeno.
192 See e.g. Guerard [1994b].
193 See Diogenes Laertius, VII 54, 64; Goulet-Caze [1989].
194 See Barnes [1988], pp.93-95; Dorandi [1994].
195 See Kidd [1978]; Barnes [1989].
196 Save, perhaps, Diodotus, Cicero's teacher (see Brutus xc 309); Lucius Aelius Stilo, the
first Latin logician (Gellius, XVI viii 2). -And see above, pp.4-5.
197 I xxix 56--cited above, p.56.
198 Suda, s.v. 'Aptcr'tOKAf\<;: see Follet [1989a], who suggests-without compelling
reason- that this Aristocles is the pupil of Chrysippus mentioned in ind stoic xlvii 7.
199 Other Stoic commentaries on Stoic texts (above, pp.50-51)? The only case I know of
is Theo of Alexandria, who flourished under Augustus and wrote a commentary on
Apollodorus' <I>ucrtoAoytKll EicraycoyT, (Suda, s.v. 8cov 'AAE~avlipEU<;). Zeller [1909],
p.607, guesses that at ind stoic lxxix 2 we should supplement '8]cov 'AAE~avlipEu<;'''but
there are half a dozen other possibilities. I should also mention Herophilus, nEpt cr'tcotKfj<;
OVOIJ.U'tCOV xpT,crEco<;, known from a single passage in Origen (in Psalm pro/ 14-15 Rietz),
which was a work of exegesis in a broad sense of the term, and which may have contained
discussion of some logical jargon. (But the only surviving 6v61J.a'ta are ''tEAo<;' and '9E6<;' .)
Herophilus' dates are unknown-no reason to identify him with the Hellenistic doctor (nor to
emend him out of existence-pace von Staden [*1989], p.584). Since he writes of the Stoics
in the third person he was perhaps not a Stoic himself. (See esp Neuschafer [1987], pp.l46-
155, who suggests that there are other traces of Herophilus in Origen.)- However that may
be, there certainly were non-Stoic commentaries on Stoic texts: Galen wrote at least four
works in at least 16 books on Stoic logic, including commentaries (illtoiJ.VT,IJ.a'ta) on
Chrysippus' :Eui..AoytcrttKT, (lib prop XIX 47 Kiihn).
2oo See e.g. II xvi 34; xvii 40 (where 'cruvaycoya<;' should not be changed to
'cruV'ta~Et<;'); Gellius, XVI viii I; Sextus, M VIII 428. None has survived, whereas for ethics
74 CHAPTER THREE

Some at least of these hypothetical handbooks will, I suppose, have been

written by Epictetus' contemporaries. Are there any allusions in Epictetus
to contemporary research in logic, to original logical work? One text seems
at first glance to be pertinent.

Someone asked him how it was that, though reason is more exercised
nowadays, greater progress was made in the past. He said: In what respect
have we exercised and in what respect was progress then greater? For you will
find that where we now exercise, there we have now made progress.
Nowadays we exercise to analyse syllogisms, and we make progress there.
Then they exercised to maintain their minds in a natural state, and made
progress there. (III vi 1-3)

Epictetus is plainly contrasting the Old Stoa with contemporary Stoicism.

And it is tempting to take him to mean that whereas the Old Stoics made
progress in ethics-that is to say, in the ethical part of philosophy- , the
contemporary school was making progress in logic (and, in particular, in
the analysis of syllogisms).
But temptation must be resisted. The purpose of the passage is not to
contrast the masters of the Old Stoa with the masters of the New; rather, it
is to contrast the philosophical attainments of old Stoic students with the
philosophical attainments of contemporary Stoic students. Epictetus'
interlocutor wonders why, when he and his contemporaries exercise their
minds so strenuously, they make less progress towards virtue and happiness
than did the relatively idle pupils of Chrysippus. Epictetus' mildly ironical
answer is this: 'You do make progress, just as they did, in what you
practise-but all you practise is your logical exercises. ' 201 Perhaps Epictetus
has in mind a bon mot ascribed to Cleanthes:

Someone asked him why, although there were not many philosophers in the
older generation, there were more stars then than now. He replied: 'Then they
attended to things [7tpny).l.a.ta.], now we attend to words [A.6yot]'. (Stobaeus,
eel II ii 16)

The sense is different, the sentiment the same. 202

The passage at III vi 1-3, then, does not advert to original logical work
in the time of Epictetus; nor have I found any other text which carries such
an allusion. But it would be foolish to draw any interesting conclusions

chance has preserved the opening pages of a Stoic textbook: Hierocles, 'H9uo1
crtotXEirocru;-see Bastianini and Long [*1992].
201 Cfl iv 17; II x 30; xvi 4.
202 And the general sentiment-Fings ain't wot they used to be-is an ancient (and a
contemporary) commonplace: e.g. Dio Chrysostom, xlii 13-15 (on philosophy in particular);
vii 89 with Russell ad Joe. We should not be too distressed by the apparent conflict between
this commonplace and the commonplace that philosophers have always been immoral
hypocrites (above, pp.41-42).
 EPICTETUS 75

from this silence. If Epictetus does not affirm, neither does he deny, that
new work was being done by the logicians of his day. Nor should we expect
assertions or implications of this sort in the Discourses.
Again, Epictetus' pupils doubtless followed some sort of 'syllabus' in
logic, as in other parts of philosophy; but Epictetus nowhere describes such
a syllabus: the Discourses do not constitute a handbook, and the last thing
they purport to be is a systematic exposition of the outlines of Stoicism; and
as for the Encheiridion, if it may be called a handbook, it is not a handbook
to Stoicism as a whole-or to logic in particular. Nonetheless we do happen
to learn that Epictetus' pupils engaged in some simple logical exercises (II
xvi 2); that they 'did' syllogisms, and changing arguments (II xiii 21; xvii
27); that they worked through the logical puzzles (III viii 1). And other
texts-Galen's Introduction to Logic, for example-allow us to guess at the
general contents of the school syllabus.
But if Epictetus offers us neither an account of contemporary research in
logic nor yet a description of the orthodox logic of his day, we do find
something else in his pages-for his more or less casual allusions to logic
give an implicit indication of those parts or aspects of the subject which
Ep!ctetus' contemporaries found especially rivetting.
The allusions are mostly brief, and they are repetitive. There are frequent
references to syllogisms, in an entirely general fashion. More specifically,
Epictetus often mentions the 'analysis' of syllogisms. 203 He speaks of
'hypothetical' arguments. 204 He alludes to A.6yot JlE'ta1ti1ttovtEc; or
'changing' arguments. 205 He refers to enthymemes. 206 He talks of the
sophisms and the conundrums, in general and individually. 207 There are
three allusions to arguments 'which conclude by way of questioning [EV t<i)
liprotil0"9at 1tEpatVOVtEc;] '. 208
Most of these allusions contribute little to our understanding of Stoic
logic-and do not even give us much knowledge of Epictetus' views on the
subject. The most interesting of them concern the analysis of syllogisms,
hypothetical arguments, and changing arguments. I shall give a section to
each of these three items. I add here a note on the sophisms and
conundrums.

203 I xxix 34; II iii 4; xxi 17; xxiii 44; III vi 3; xxiv 78; xxvi 18; IV iv 14; vi 12.
204 I vii I, 22-29; xxvi I, 9, 13; II xxi 17; III ii 6, 17; xxiv 78, 80; IV iv 14.
205 I vii I, 13-21; II xiii 21; xvii 27; xxi 17; xxiii 41; III ii 6, 17; xxiv 80; IV vi 15; xii 12.
-At III ii 8 'af.LE't!l1t'troola' presumably puns on the 'f.Lt'ta7tt7t'tOV'tt;' of 6; perhaps there is
a similar pun at III xxvi 13-14.
206 I viii 1-2: above, p.58.
207 II xvi 3; III viii I; xxvi 16.- Note that at I xxvii 2 the word 'oocplof!a'ta' refers not to
logical puzzles but to the sceptical arguments of the Academics and the Pyrrhonians.
208 I vii I; III ii 6; xxi 10: below, pp.136-138.
76 CHAPTER THREE

There are five references to the most celebrated of all logical conundrums,
the Liar. 209 And the Denier, to which Epictetus once alludes, was no doubt a
variant on the Liar. 210 None of these texts explains wherein the paradox
lies-let alone suggests a solution, or even a line of inquiry. There is an
isolated reference, again without any explanation, to the Sorites. 211 One
passage evidently alludes playfully to the Homed Man. 212 The only other
Stoic sophism or conundrum to which I have found an allusion is the
Master Argument-for which, as I have remarked, Epictetus happens to be
our primary source. 213
II xix bears the title 'Against those who take up the ideas of the
philosophers only as far as talking goes'.lt opens abruptly.

The Master Argument seems to have been propounded on this basis: the
following three items are in mutual conflict with one another-that everything
past which is true is necessary, that the impossible does not follow the
possible, that things are possible which neither are true nor will be. ( I)

Epictetus then reports that Diodorus upheld the first two of the three
propositions and therefore denied the third; that some people held on to the
second and third of the propositions and denied the first; and that others
held on to the first and third and denied the second. In 2 he says that
Cleanthes-or perhaps the school of Cleanthes-' seems [801coum ]' to have
taken the second position and that Antipater 'largely supported him'. From
5 we learn that Chrysippus patronised the third position. No-one, it seems,
attempted to conserve all three propositions. At any rate, Epictetus asserts
roundly that 'it is impossible to keep all three because of the mutual
conflict' ( 4). And then he turns to reflect on the appropriate attitude for a
philosopher to take toward the Argument. 214

209 See II xvii 34; xviii 18; xxi 17; III ii 6; ix 21 (above, p.64). Texts on the Liar ('What I
say is false ... ') in Hiilser [*1987], frags 1210-1226. Note that Chrysippus had written at vast
length on the puzzle (Diogenes Laertius, VII 196-197).
21o For the Denier (a7tmpacrKrov) see Diogenes Laertius, VII 44; Clement, strom Vi 11.6.
211 II xviii 18 (above, p.39), where Epictetus uses the alternative name of "Hcruxal;rov'.
Texts on the Sorites ('A single grain of sand does not make a heap; if you add a single grain
to something which is not a heap you do not thereby get a heap ... ') in Hiilser [* 1987], frags
1236-1243; see Barnes [1982].
212 See I xviii 16. For the Homed Man ('What you have not lost you still possess; you
have not lost horns: ... ') see esp Gellius, XVI ii (below, n.285); cf Hiilser [*1987], pp.l760-
1761.
213 It would no doubt be fanciful to detect an allusion to Seneca's mice (above, p.l4) at
III ii 15. With the mice compare Chrysippus' waggon ('What you say comes out of your
mouth; you say a waggon ... '): Diogenes Laertius, VII 187; cf Clement, strom VIII ix 26.5. -
The list of paradoxes in Diogenes Laertius, VII 44, contains three items which do not appear
anywhere in Epictetus: the Nobody, the Veiled Man, the Mower; and Diogenes' list is
incomplete-there were many other puzzles to which Epictetus makes no reference (texts in
Hiilser [*I 987), frags 1199-1257).
214 See above, pp.35, 44-45.
 EPICTETUS 77

The text frustrates scholars. Grateful that Epictetus has at any rate
preserved the Argument from oblivion, they wish that he had formulated its
three constituent propositions a little more rigorously, and that he had
indicated why they were generally taken to form an inconsistent triad. 215 An
intelligible wish-but one which Epictetus had no reason to satisfy. For in
II xix he is not concerned to discuss the Master Argument, nor even to
expound it. The opening paragraphs should be read as though they were in
inverted commas: their function is not to give a formal exposition of the
Argument; rather, it is to mimic an attitude which Epictetus deplores-the
exegetical and doxographical attitude to philosophical issues. 216 A precise
and thoughtful account of the Argument-an account of the sort which
modern scholarship desiderates and which Epictetus himself demanded of
his contemporaries-would have been wholly out of place.
In short, even in II xix Epictetus tells us nothing-save accidentally-
about the paradoxes themselves: no expositions, no solutions. 217 But his
texts do make their contribution to what might be called the sociology of
logic. For they show, first and uncontroversially, that Epictetus' con-
temporaries were fascinated by this particular aspect of Old Stoic logic-
just as the Old Stoics themselves had been. They show, secondly, that
Epictetus himself thought that the puzzles were not trifles-that they were a
proper object of a philosopher's attention, and that the wise man must know
how to deal with them. 218 (Here, at least, his attitude differed from the
attitude of Seneca.) And thirdly, I think that Epictetus realized that, in
general, the puzzles were not mere party tricks; for whether or not he knew,
as Chrysippus had known, that the puzzles-some of them at least-raise
fundamental questions in logic, he never insinuates that they are superficial
rather than profound, tangential rather than central.

H: Analysis

'You analyse syllogisms': the word 'analysis'-my English does no

more than transliterate the Greek-has more than one technical sense. In
general-and vaguely enough-to analyse something is to divide it up or
reduce it to its simple components; and the particular technical senses
which, in a logical text, the word may assume, are determined by the

21 5 On the Argument itself see now Gaskin [1995], with full discussion of earlier
interpretations; Denyer [1996].
216 Above, pp.43-45, 54-55.
21 7 I assume-perhaps optimistically-that there were preferred, and perhaps even new,
solutions going the rounds. We happen to know from Galen that the Sorites was widely
discussed in the imperial period, and that various solutions to it had been canvassed: see
Barnes [1982].
218 See above, p.65.
78 CHAPTER THREE

different ways in which you may speak of 'dividing' something or of

reducing a complex to its simple parts. 219 When Epictetus speaks of
analysing syllogisms, I take him to have a technical notion of analysis in
mind. It is true that the noun 'avaA.ucrt<;' and the verb 'avaA.uEtv' were
also used in less technical ways; and in a few ancient texts it is tempting-
and even plausible-to believe that the phrase 'analyse syllogisms' means
little more than 'do logic'. 220 But Epictetus frequently sets the analysis of
syllogisms on a level with the exploring of hypothetical arguments or the
study of sophisms, so that analysis should be something specific rather than
the general exercise of 'doing logic'. Moreover, the notion should be
reasonably technical; for the contexts in which Epictetus refers to analysis
suggest an exercise which is both fairly difficult-the sort of thing a young
man might pride himself on-and also removed, or at least apparently
removed, from any practical application.
Now in its central and most technical sense the term 'analysis' refers to
the reduction of certain arguments or argument-schemata to other, more
fundamental, arguments or schemata. Analysis, in this sense, is in effect
proof; for to analyse one argument or schema into another is to prove that
the one is valid if the other is valid. 221 A little more precisely: you analyse a
given argument schema into a set of argument schemata insofar as you
demonstrate that if each member of the set is a valid schema, then the given
schema is a valid schema. (Similarly for arguments.)
Thus Peripatetic logicians, following Aristotle, would analyse second-
and third-figure syllogisms into first-figure syllogisms. A trifling example:
take an argument in Cesare, of the second figure. The schema for such an
argument is:

M holds of noN.
M holds of every X.

Therefore: N holds of no X.

Aristotle reasons thus:

Let M be predicated of no N and of every X. Since the negative converts, N

will hold of no M. But M is supposed to hold of every X. Hence N will hold
of no X-this has already been proved. (APr 27a5-9)

219 See e.g. Alexander, in APr 7.12-33; Alcinous, didask 5 [pp.l56-157 Hermann];
Ammonius, in APr 5.10-7.25.
220 So perhaps Marcus, I xvii 22. (But at Musonius, frag 3 [p.l2.9-IO Hense]= Stobaeus,
eel II xxxi 126--above, p.63-l assume that a technical sense is intended.)
221 Note Aristotle's use of 'cm6liEt~u;' in connection with such analyses or reductions
(e.g. APr 27b3; 28a23).
 EPICTETUS 79

That is to say, a little more formally: 222

1 (1) M holds of no N premiss

2 (2) M holds of every X premiss
1 (3) N holds of noM 1, conversion
1' 2 (4) N holds of no X 3, 2 Celarent

Here (1) and (2) are the schematic premisses for an argument in Cesare,
and (4) is the conclusion. The argument shows that, given those premisses,
we may reach the conclusion by the use of exactly two principles of
inference: the conversion principle for universal negative propositions, 223
and the syllogistic form Celarent. 224 Thus if the schemata for Celarent and
for the conversion of universal negatives are valid, then the schema for
Cesare is valid. That is to say, Cesare has been reduced to Celarent and the
conversion principle. 225
Analysis in Aristotelian logic is trivial: the number of syllogistic
schemata which the logic permits is fmite (and small); and it is easy enough
to reduce all valid schemata to schemata of the first figure- indeed, to
Barbara and Celarent. The work was already done by Aristotle himself in
the Prior Analytics. Analysis in Stoic logic is more difficult, if only because
there are infmitely many valid schemata in hypothetical syllogistic so that
we cannot simply survey all the possible forms and proceed to analyse them
seriatim.
Our information about Stoic analysis is scarce and in parts obscure. 226
But it is reasonably clear that classical Stoic logic took the five
222 And in a style which is intended to point up the similarity between ancient analysis
and the sort of proofs which tiro logicians are now invited to construct at the end of each
chapter of their elementary textbooks.
223 The schema for such conversion is this:

A holds of no B.

Therefore: B holds of no A.

224 The schema for Celarent is:

A holds of no B.
B holds of every C.

Therefore: A holds of no C.

225 My version of the reduction sets Celarent and the conversion precisely on a par: hence
Cesare is analysed into two other schemata. Ancient logicians speak of reducing Cesare to
Celarent-conversion is, so to speak, the way by which the analysis proceeds rather than one
of the items to which it proceeds. There are delicate exegetical issues here, and I have
clomped over them-but in the end, the distinction implicit in the ancient texts is of no
logical significance.
226 Texts in Hiilser [* 1987], frags 1160-1189; on Stoic analysis see esp Becker [1957];
Frede [1974a], pp.l67-197; Ierodiakonou [1990]; Mignucci [1993]; Bobzien [1996].
80 CHAPTER THREE

'indemonstrable' forms of Chrysippus as its base, 227 and attempted to

analyse all other syllogisms into syllogisms manifesting these fundamental
forms. Syllogisms in the five basic forms-like the syllogisms in the four
first-figure forms of the Peripatetics-are evidently valid, or at least are
taken to be evidently valid; and they are called 'indemonstrable
[ava1t00Etl('toc;]' inasmuch as they do not need to be demonstrated or
proved. 228 Any other argument needs to be demonstrated; that is to say, it
must be proved to be valid by reduction to one or more of the five
indemonstrables.
Sextus reports a fairly simple example of a Stoic reduction.

The argument in question is compounded from a second and a third

indemonstrable, as can be learned from the analysis-which will become
clearer if we conduct it in terms of the schema, 229 viz:

If both A and 8, then C.

Note.
A.

Therefore: not B.

Since we have a conditional in which the antecedent is a conjunction of A and

8 and the consequent is C, and we also have the opposite of the consequent

227 Texts in Hiilser [*1987], frags 1128-1137. - The five forms may be described as
follows (see e.g. Sextus, PHil 157-158):

(I) From a conditional and its antecedent the consequent follows.

(2) From a conditional and an opposite of its consequent an opposite of its antecedent
follows.
(3) From a negated conjunction and one of its conjuncts an opposite of the remaining
conjunct follows.
(4) From a disjunction and one its disjuncts an opposite of the remaining disjunct
follows.
(5) From a disjunction and an opposite of one its disjuncts the remaining disjunct
follows.

Note: (i) In (3), (4), and (5) it is assumed that conjunctions and disjunctions have precisely
two component conjuncts or disjuncts-it is not clear that this condition was always imposed.
(ii) The indemonstrables are sometimes expressed by way of schemata (see e.g. Sextus, M
VIII 227; Galen, inst log vi 6), so that instead of(!) we have

(I') If A, then B; A: therefore, B.

The standard schemata are not all equivalent to their metalogical correlates, (I )-(5) (see
below, n.309); I prefer the metalogical mode of exposition-but I take no line on what the
official Stoic view of the matter was or on whether there was any official view.
228 For the sense of 'avan6owcto<;' see Sextus, PH II 156; Galen, inst log viii I (cf Frede
[1974a ], pp.l27 -129): the standard English translation, 'indemonstrable', is misleading.
229 Stoic schemata are properly expressed by way of ordinal numerals ('The first', 'the
second', ... : Apuleius, int xiii [p.212 Moreschini]; cf the vignette from Fronto, above, p.49);
but no harm is done here by the use of letters.
 EPICTETUS 81

('Not C'), we shall infer the opposite of the antecedent- 'Therefore: not both
A and B'-by a second indemonstrable. (Indeed this is itself potentially
present in the argument, since we have the premisses which conclude to it, but
it is not expressly stated.) Taking this230 together with the remaining
premiss-A-we shall infer the conclusion 'Therefore: not B' by a third
indemonstrable. (M VIII 235-236)

Sextus' description might be set out a little more briefly and a little more
formally as follows:

I (1) If both A and B, then C premiss

2 (2) Note premiss
3 (3) A premiss
1,2(4) Not both A and B 1, 2 second indemonstrable
1,2,3(5) NotB 3, 4 third indemonstrable

The argument depends solely on the second and the third indemonstrables.
Hence the schema in question is valid if the second and the third
indemonstrables are valid.
In setting out the argument, Sextus adverts-in parentheses-to a
general principle which he had earlier formulated thus:

When we have the premisses which yield a certain conclusion, we have the
conclusion too potentially among them, even if it is not expressly stated.
( 231)

This is, to say the least, vague; nor is it clear what function the principle is
supposed to play in the reduction. Classical Stoic reductions or analyses
apparently made use of certain rules or principles which the Stoics called
'8EIJ.a'ta'. But Sextus' principle is not itself a 8E!J.a, and it is far from clear
how it is related to the 8EIJ.a'ta. What were the 8EIJ.a'ta? how many of them
were there? how precisely did they function in analysis? how did the Stoics
think of analysis? how would they standardly have set out an analysis?
Such questions are controversial. And if the general gist of the matter is not
in doubt, the logical details are all uncertain-and will remain uncertain
until new texts turn up. 231
Yet if the logic is mostly dark, one historical fact is bright enough. Galen
remarks that:

You can now find any number of people who are expertly practised in how
syllogisms with two 232 tropics are analysed, and how indifferently concluding

230 Reading 'oltEp' (Kochalsky) for 'altEp'.

231 A text of Galen's nEpt altoaEi~E~, for example: the Greek is lost (apart from a few
fragments); but substantial portions of the work reportedly survive in an Arabic translation.
232 Some MSS read 'three', and some scholars have liked 'two or three'.
82 CHAPTER THREE

arguments and others of this type are analysed, 233 by using the first and
second etl!a; and with other syllogisms which234 they analyse by the third
and fourth etl!ata. Yet most of these can be analysed more economically in a
different way, as Antipater wrote-and in any case, the whole matter of such
syllogisms is much labour spent on a useless object, as Chrysippus himself
witnesses (inasmuch as nowhere in his writings does he need such syllogisms
in order to prove his doctrines). (PHP V 224 Kiihn)

The text is difficult; and my translation has papered over a few cracks. But
the difficulties are here irrelevant; for the pertinent facts are in the clear.
In Galen's time there were plenty of people who prided themselves on
their skill in analysing certain syllogistic forms by way of the 8j.tata.
Galen himself regards this as a waste of time, for two distinct reasons: first,
Antipater had long ago shown that there were simpler ways of doing the
analyses; secondly, the syllogistic forms in question, being scientifically
useless, are not in any case worth reducing. 235
What in fact was the reason for conducting analyses? Plainly Epictetus,
when he talks of analysing syllogisms, is thinking primarily of school
exercises. Thus he speaks of someone

who has practised analysing syllogisms, and if someone offers him an easy
one [eu:>...utov] he says: 'Give me something more ingeniously constructed so
that I can train myself. (I xxix 34)

Galen too is not thinking of advanced logicians who were engaged in

logical research: his men are going through familiar hoops. But if the
exercises did not demand any great logical imagination or originality, they
were not easy to master and they required some logical talent or expertise. I
suppose that the men who wasted their time on such tricks in Galen's day
were the sons and grandsons of the men who analysed syllogisms in
Epictetus' time.
How did Epictetus' boys tackle their exercises? Epictetus never refers to
the 8j..tata. Now Antipater, according to Galen, had introduced a 'more
economical' way of analysing certain types of syllogism. Neither Galen nor
anyone else tells us what this method was; and all we know is that it was in
some respect briefer than some standard method which, in the particular
cases at issue, involved the use of two 8j..tata. One guess has it that
Antipater invented a way of analysing these syllogisms by the use of a

233 A 'tropic' is a compound assertible, so that 'syllogisms with two tropics' are
syllogisms two of the premisses of which are complex assertibles (see e.g. I xxix 40; Galen,
inst log vii I; Alexander, in APr 20.7-but there are difficulties with the terminology: Frede
[1974a], p.IOI n.25; Barnes et al [*1991], p.69 n.\03). An argument concludes 'indifferently'
if its conclusion is identical with one of its premisses (e.g. Alexander, in Top 10.10-12; cf
Frede [1974a], p.l84).
234 Reading 'e1t' iiA.Aot<; ocrouc;'.
235 On Galen's utilitarianism in logic see above, p.20.
 EPICTETUS 83

single 8Ej.W.. But there is another possibility: perhaps Antipater abandoned

the 8j..La'ta altogether. After all, the analysis which I took from Sextus does
not expressly involve a 8j..La-and yet it contains no logical lacuna. Thus
whatever the standard function of the 8j..Lata may have been, it is clear that
they cannot have been logically essential parts of every Stoic analysis.
Perhaps Antipater saw this, or stressed it? Perhaps he developed techniques
of analysis-such as the technique illustrated by Sextus-which did not
make use of 8j..Lata? And perhaps the fact that Epictetus does not refer to
8j..Lata indicates that he and his students preferred to follow the more
economical methods of Antipater? 236
However that may be (and it is, of course, riotously speculative),
Epictetus does at least offer us some evidence for the nature of logical
studies-of a student's logical studies-in his day. Alas, the evidence of the
Discourses adds little to our understanding of Stoic logic itself. For
Epictetus nowhere conducts an illustrative analysis or indicates how his
students were supposed to attempt their exercises.

As an appendage to this section I may discuss briefly one pertinent and

puzzling text. The short discourse II iii is formally devoted to a ticklish and
topical question: how should one write letters of recommendation on behalf
of philosophy students? Epictetus reports an anecdote about the Cynic
Diogenes: when a student asked him for a letter of recommendation, he
asked: What is the use of my writing to X that you are a good man? If he
can discriminate the good from the bad, he will himself recognize that you
are good; if he cannot, a thousand letters will not help him. Epictetus
apparently takes this silly remark to convey an important and lamentable
truth. He reflects on it thus:

We need for lives the sort of thing we have for coins, so that I could speak
like the assayer who says: 'Bring me any coin you like and I will appraise it'.
Now in the case of syllogisms I say: 'Bring me any one you like and I will
distinguish the analytical from the non-analytical [tOV avaA.UttKOV tE Kat
J-Lil]' .-How?-I know how to analyse syllogisms. I have the capacity which
anyone must have if he is going to appraise those who succeed in syllogisms.
But in the case of lives what do I do? Sometimes I call one good and at other
times bad. What is the reason for this? The opposite of what is the case with
the syllogisms-ignorance and inexperience. (II iii 3-5)

When it comes to assessing lives-that is to say, to assessing whether a man

has lived or is living a good or a bad life-we lack the discriminative
capacity which any assayer has for coins and which we philosophers have
for syllogisms. The general point of the text is plain; 237 but what exactly

236 See below, pp.J40-141.

237 And cfl xxvi 17.
84 CHAPTER THREE

does Epictetus refer to when he speaks of distinguishing 'the analytical

from the non-analytical'?
'Now in the case of syllogisms I say: 'Bring me any one you like ... ".
We shall, I suppose, naturally take this to mean 'Bring me any syllogism
you like' .238 Hence Epictetus means that he can 'distinguish analytical
syllogisms from non-analytical syllogisms'. 239 But this construal is faced by
two difficulties. First, what is an 'analytical' syllogism? The phrase
elsewhere occurs only in a late commentator on Aristotle; and the technical
sense which it there bears could not have been intended by Epictetus. 240
Perhaps we may guess a meaning for the word? Perhaps an 'analytical'
syllogism is a syllogism which can be analysed? Then Epictetus will be
referring to the capacity to judge which syllogisms can be analysed or
reduced to the indemonstrables and which cannot. 241 But it is far from clear
that the word can be understood in this sense; and it is far from clear that
the sense is pertinent to the context.
Some scholars have scented corruption. 242 Thus instead of
'avaA.unK6~', 'a1toOEtK't6~' might commend itsel:f2 43-at least, it states in
an unpuzzling way the idea which 'avaA.un1<:6~' could only state in a
pointlessly puzzling way. But if emendation is needed, then I prefer
something bolder: perhaps 'cruvaK:nK:6~' or '1tEpavnK6~'. With such a
word, Epictetus will be claiming a capacity to tell which syllogisms are
valid and which are not; and such a claim-unlike a claim about demonstr-
ability-provides the obvious parallel to the claim of the assayer.
But there is a second difficulty. In II iii Epictetus is interested in
capacities for judging people; and in the case of syllogisms, he claims that
he can 'appraise those who succeed in syllogisms' 244-that is to say, he can
tell which of his pupils are decent logicians. Now if the 'analytical' items
are syllogisms and the 'successful' items are students, then we shall have to
ascribe to Epictetus-and ourselves make sense of-the following train of
thought: 'I can tell an analytical syllogism from a non-analytical

238 That is to say we shall read: ' ... <pEpE ov [scilicet ouHoytOilOV] 9i:A.w; .. .'.
239 So e.g. Oldfather.
240 Eustratius, in APst 2.25-27; 3.26-34: an 'analytical' syllogism is one in which the
premisses are 'more familiar to us' and the conclusion 'more familiar by nature'. The phrase,
in other words, is closely tied to a particular thesis within Aristotle's philosophy of science
which finds no echo in Epictetus.
241 Which syllogisms are not analytic in this putative sense? Evidently, at least the
indemonstrables themselves. But perhaps Epictetus is also thinking of the so-called A6yot
all9ooro<; 7tEpatvovtE<;, although they are not strictly speaking ouHoytollot (see Barnes
[1990a], pp.65-83).
242 Schenk! once conjectured 'avaA.utov'; Richards offered 'av<X1tOOEtKttK6v'.
243 See e.g. Alexander, in Met 272.16-17: oi:OEtKtat yap f.v tot<; 'AvaA.uttKot<; tiva ti:
etcrtV cmolietKta !CUt tiva OU.
244 The Greek is 'tiilv 7tEpt ouA.A.oytolloil<; Katop9oilvtrov', and I cannot convince
myself that the participle might be neuter.
 EPJCTETUS 85

syllogism-after all, I can analyse syllogisms-and so I have the capacity

which I need if I am going to distinguish good logic students from bad logic
students'. I do not affirm that no sense can be squeezed out of this- but
you will have to squeeze pretty hard.
Is there an alternative interpretation? Well, if Epictetus is claiming a
capacity to tell who is and who is not a decent logician, then when he says
'I will distinguish the analytical from the non-analytical, 'tOV avaA.U'ttKOV
'tE Kat 1-lil' he must mean 'I will distinguish the man who can analyse [sc
syllogisms] from the man who cannot' .245 This in tum has its difficulties.
First, in the context-and given the parallel with the coins-it might seem
impossible to construe 'Bring me any one' as 'Bring me any man' rather
than as 'Bring me any syllogism'. Secondly, as far as I am aware there is no
other Greek text in which the word 'avaA.unK6~' is used of a person.
There is little to say on the second point; but I suppose that it would not
have been particularly difficult to understand 'avaA.unK6~' as used of a
person-after all 'BtaA.EK'ttK6~' is regularly used of people. In any event,
the word is no harder on this second interpretation than it is on the first.2 46
As to the first point, it is worth noting that the coins were introduced, in the
anecdote about Diogenes, in an exquisitely arch fashion: Diogenes
compared a student who wants to be recommended as a good man to a coin
which might ask to be recommended as a genuine drachma. In the anecdote,
the coin parallels the student. Hence in Epictetus' rumination on the
anecdote, the coin may continue to parallel the student-after 'Bring me
any coin' we may, after all, interpret 'Bring me any one' as 'Bring me
anyone'.
All in all, I incline to the second interpretation. But it is not easy. Nor
does it shed any light on Epictetus' attitude to analysis.

I: Hypothetical arguments

On hypothetical arguments and changing arguments Epictetus provides

more pabulum. The chief text is I vii. 247 I shall reverse Epictetus' order of
discussion and start with hypotheses and hypothetical arguments.
The word '\m6eEcrt~' has a variety of uses in Greek philosophy of the
period; 248 and the term 'hypothetical argument' may describe more than one

245 That is to say we shall read: ' ... cpf:pt ov [scilicet iiv8pro1tov] 9E:A.ttc; .. .'-or rather,
we shall take 'ov' to mean 'whoever'. So Long, Laurenti, Carter/Hard. See Dio Chrysostom,
xii I 0, for a (distant) parallel to the idea.
246 And if we want an emendation to eke out the second interpretation? Why not
'otaA.tK-ctK6c;'?
247 See above, pp.61-62; below, Appendix.
248 See Barnes [1990b], pp.90-95; Freytag [1995], pp.l22-129.
86 CHAPTER THREE

sort of syllogism. In one of its senses, the adjective 'u1t08E'tt1C6<;' contrasts

with 'KU'tllYOptK6<;', and hypothetical arguments contrast with categorical
arguments. 249 In this sense of the word, all Stoic syllogisms are
hypothetical. But when Epictetus speaks of hypothetical arguments in I vii
and elsewhere he is evidently thinking of something more specific-of a
special subclass of Stoic arguments. 250
Chrysippus had written on hypotheses, and he had written at length on
hypothetical arguments. 251 No fragment of these works has survived, and no
Stoic testimonium incontrovertibly derives from them. Nonetheless, it is
certain that Chrysippus did not mean to refer to all non-categorical
syllogisms, certain (in other words) that he was thinking of a special type or
form of Stoic argument-and it will then seem probable that Epictetus is
referring to the same special sort of argument. What special sort? Various
guesses come to mind. 252 One of them-which I find curiously attractive 253
-connects Chrysippus' hypothetical arguments with a special sort of
AEK't6v or sayable: the sort which the Stoics called a 'hypothetical', a
U1t08E'ttKOV. 254
Hypotheticals are sayables saying which we hypothesize something.
They are canonically expressed by sentences of the form 'Let it be
supposed that [U1tOKEtcr8ro] ... '. In uttering a sentence of the form 'Let it be
supposed that P', I say something-and in particular, I hypothesize
something: I hypothesize that P. A hypothesis is then either the act of
making a hypothesis or else the content of the act-the item which I
hypothesize. Hypotheticals in this sense of the word are distinguished from
assertibles or a~tffilla'ta. But although they are not assertibles, and hence
are themselves neither true nor false, they will, I suppose, contain
assertibles. 255 The hypothetical 'Suppose that the earth is spherical' contains
the assertible that the earth is spherical. In uttering the sentence 'Suppose
that the earth is spherical', I will typically hypothesize something, namely

24 9Above, p.71.
250 Pace Schweighauser [* 1799], II p.l 02.
251 Diogenes Laertius, VII 196: the fourth group (ouvta~u;) of the area concerned with
arguments and schemata consists of one work, in three books, On Hypotheses, four works of
or on hypothetical arguments, and one work in one book On Exposition.- On the value of the
titles as evidence for Chrysippus see Barnes [1996).
252 For example, that Chrysippus was dealing with what later logicians called 'wholly
hypothetical' arguments-arguments all the components of which are complex propositions
(on which see Barnes [1983)). Texts in Hiilser [* 1987], frags 1190-1198-but none of the
passages which Hiilser collects refers explicitly to the Stoics.
253 Susanne Bobzien made me aware of the attraction-in a paper due to appear in
Phronesis for 1997 (and to which I am heavily indebted) she offers an account of Stoic
hypothetical reasoning.
254 Texts on hypotheticals in Hiilser [*1987], frags 897-908; see e.g. Schenkeveld [1984).
I should confess that no text explicitly connects Chrysippus with the imo9nK6v.
255 Just as oaths, though not themselves assertibles, contain assertibles: note the phrase
'1tEplElAllJ.lJ.lEVa a~UD[J.lata' at Chrysippus, Aoy Zll't X 9-10.
 EPICTETUS 87

that the earth is spherical. I also and necessarily express the assertible that
the earth is spherical-but I do not assert that the earth is spherical (or
anything else).
Now hypotheticals may and do feature in arguments: 'Suppose that there
is a highest prime number', you may say-and then start out on a reductio
ad impossibile. Hence the Stoics will presumably have made a special study
of hypothetical arguments, distinct from the study of standard syllogisms.
('Presumably': there is no express evidence for this-although we do
happen to know that Chrysippus made a special study of imperative
arguments. 256)
Such speculations fit reasonably well with the Epictetan texts on
hypothetical arguments. To be sure, Epictetus' hypothetical arguments do
not employ-or do not only employ-the canonical form 'Let it be
supposed that ... ' in order to express a hypothetical. Rather, they prefer the
form 'Let it be the case that [cr'tro] ... '; thus:

As we behave in hypothetical arguments, so should we behave in life. 'Let it

be night [Ecr'tco v-6~). '-Let it be so.-'Well then, is it day?'-No; for I
accepted the hypothesis of its being night.-'Let it be the case that you take it
to be night'.-Let it be the case.-'But take it that it is day.'257_That does
not follow from25B the hypothesis. (I xxv 11-12: cf III xxii 68)

Thus we start a hypothetical argument by accepting a hypothesis, stated in

the form 'Let it be the case that ... '.
Some Stoics distinguished between 'Let is be supposed that ... ' and 'Let
it be the case that ... ', which standardly marks what they called an
'expositive' sayable or EJC9EnJC6v. But expositives are found in only two of
the several lists of 'complete sayables' which we possess; and it seems
plausible to think that they were not generally distinguished from
hypotheticals. 259 I imagine that Epictetus regards 'Let it be the case that ... '
as a permissible way of expressing a imo9EnJC6v. (Note that Sextus will use
'Let it be granted [ooocrero] ... ' in the expression of a hypothesis: Mill
17.) In any event, it is hard to see what logical significance the distinction

256 See e.g. Barnes [1984].

257 Reading 'ru.1pa' (Trincavalli) for the transmitted 'vu~ (Oldfather's intricate defence
of the received text-[* 1925], p.160 n.3-is unconvincing).
258 <h:oA.ou9Et: 'be consistent with' (Long, Oldfather) is tempting (cf e.g. II xvi 7 for
'<h:oA.oua6~' in the sense of 'coherent with'); but in the similar context at I xxvi I (below,
p.90), '6.KoA.ou9Etv' must mean 'follow from'.
259 It is true that Chrysippus wrote both nEpt 'il1to9crErov and also nEpt EK9ECJEOJV
(Diogenes Laertius, VII 196), and this perhaps suggests that he had distinguished
hypotheticals from expositives (see Frede [1974a], p.44). But the catalogue contains no
reference to A.Oyot EK9EttKoi.-and the position of the nEpt EK9ECJEOJV in the catalogue (at
the end of the section on hypothetical arguments) perhaps intimates that exposition was
regarded as a special kind of hypothesis. (Perhaps existential hypotheticals were called
expositives?)
88 CHAPTER THREE

between a hypothetical and an expositive might have been thought to

possess.

At this point two questions present themselves. First, how could there be
such a thing as 'hypothetical logic'? And secondly, why should there be
such a thing as 'hypothetical logic'?
Behind the first question there lies a banal truth. Logic is about one
thing's following from another; and, in general, X follows from Y only if,
given that Y is true, X too must be true. But hypotheticals are neither true
nor false. How then can anything follow from a hypothetical? (And how
can a hypothetical follow from anything else?) Moreover, the Stoics defined
an argument or Myor; as a system of assertibles or al;trof!ata; 260
consequently, for the Stoics there can in principle be no hypothetical
arguments-that is to say, no arguments some components of which are
hypotheticals.
Now in fact it is not difficult to explain how, despite lacking a truth-
value, a hypothetical might feature in an argument. Any hypothetical will
be expressed by a sentence such as 'Suppose that P'. In such a sentence, the
component marked by 'P' will express an assertible-something which can
(and must) be either true or false. Hence why not say that something
follows from a hypothetical insofar as it follows from the assertible
contained in the hypothetical? From 'Suppose that P' we may infer 'Q'
provided that, given that P, it follows that Q.
Thus we might imagine that the following schema is a schema for a
hypothetical argument:

Suppose that P.
IfP, then Q.

Therefore: Q.

But the conclusion might give us pause: if we start with a hypothetical,

should we not also end with one and write the conclusion thus:

Therefore: suppose that Q?

Not, I think, according to Epictetus; at least the portion of text which I have
cited from I xxv-

'Let it be night.'-Let it be so.-'Well then, is it day?'-No; for I accepted

the hypothesis of its being night.

260 See below, pp.l24-125.

 EPICTETUS 89

-suggests that the 'conclusion' does not take the form of a hypothetical:
we are being invited to infer that it is day; we are not being invited to make
a second supposition. Hypotheticals may be starting-points of arguments:
they may not be conclusions. And in fact there does seem to be something
odd about an expression of the form 'Therefore: suppose that Q'.
But in that case, what are we to make of the conclusion of the
hypothetical argument? It is presumably an assertible if it is not a
hypothetical; and in setting it down as the conclusion of an argument we
shall presumably be asserting it. But evidently we may not want to assert it
-we may suspect, or even know, that it is false, and neither the wise man
nor the fool wants knowingly to make a false assertion. In any event,
having made our hypothesis we want, so to speak, to stay in the
hypothetical realm. And it seems that the assertible conclusion takes us out
of that green and pleasant land.
The answer to this little puzzle can perhaps be found if we turn to the
second of the two obstacles which seem to block the very existence of
hypothetical logic. I think that we should take seriously the Stoic defmition
of an argument. From that definition it follows that the hypothetical
argument which I have given is not an argument at all: indeed, no
hypothetical arguments are arguments. The phrase 'hypothetical argument'
no more picks out a type of argument than the phrase 'decoy duck' picks
out a species of duck or the phrase 'expectant mother' picks out a subclass
of the mothers of the world. 261 In producing a hypothetical argument you
reason hypothetically. But any argument involved in your reasoning is a
system of assertibles. Such an argument might have the form

P.
IfP, then Q.

Therefore: Q.

In arguing hypothetically, I may produce an argument of this very form-

but I produce it hypothetically; and I mark-or may mark-the fact by
prefixing its first premiss, as I enunciate it, by 'Suppose that ... '. In uttering
the sentence 'Suppose that P', I thereby utter a sentence which expresses an
assertible-but I do not thereby assert anything. And in uttering the
following sentences 'If P, then Q' and 'Q', I again utter sentences which
express assertibles. But again I do not assert anything; for the context of the
utterance-the context marked by the initial 'Suppose that P'-indicates
that this particular utterance of a sentence expressing an assertible is not

261 Note that, according to the Peripatetics, 'wholly hypothetical syllogisms' (above,
n.252) are not syllogisms: e.g. Alexander, in APr 326.12-19.
90 CHAPTER THREE

itself an assertoric utterance. 262 Nor, in uttering the sentence 'Therefore: Q'
do I thereby infer anything.
This observation may make hypothetical logic possible-but surely it
also makes it threadbare? Surely hypothetical logic can be nothing more
than a shadow of the standard logic of assertibles? The only difference
between a piece of hypothetical argumentation and a corresponding piece of
assertoric argumentation will be that in the former, but not in the latter, at
least one of the premisses will be set down in the form 'Suppose that P'.
There cannot be any logical rules peculiar to hypothetical arguments; for
there are no such -arguments as hypothetical arguments. But in that case,
why-how-speak ofhypotheticallogic at all?
Well, Epictetus invokes a few rules of hypothetical argumentation; and it
may be worth seeing what they are before we dismiss the whole subject as a
phantom. Thus

when someone was reading the hypotheticals, he said: 'This too is a

hypothetical rule [v6~-to<;]-accept what follows from the hypothesis.'
(I xxvi I)

A student was 'reading the hypotheticals'-that is to say, I guess, he was

reading a book entitled The Hypotheticals (perhaps even one of Chrysippus'
books )263-and he had ignored or misunderstood one of the rules for
dealing with such arguments, the rule obliging you to accept the
consequences of the hypothesis which you had provisionally adopted.
The hypothetical rule-as Epictetus states it-is not a law of logic; 264
nor is it a rule of inference. It is not a statement that Y follows from X, nor

26 2 Then what is it? If it is neither assertoric nor hypothetical, what else can it be? I can
invent no answer to this question on the basis of the Stoic texts; but I am not inclined to think
that it is a devastatingly difficult question to answer.
263 Bruns [ 1897], pp. 9-10, argued that the student was reading out one of his own essays
(so too Hijmans [1959], p.42 n.5, with further references). But Epictetus, having
corrected-and embarrassed-the student, went on to blame 'the one who had suggested the
reading to him [to\> {llto9EJ.!EVOU aut/il tljv avayvwcnv]'. Presumably an assistant master
(see I xxvi 13; cf iv 9 (?)) had told the student to read something he was not yet able to
follow, and the student duly read it out to Epictetus: what he read was not his own
composition.
264 Hence we do not find in this text an early (the earliest?) occurrence of the notion of a
'law of logic' (a 'law of thought'). But I imagine that our text indicates the origin of the
notion; for the term 'v6J.!oc;' will readily pass from the rule or convention of dialectical
behaviour to the principle or general truth on which the rule is based (see below, n.285). -
One of the Chrysippean works on hypothetical arguments bears the following title: A6yot
U1to9Etllcot Eic; toile; VOJ.!OU<; (Diogenes Laertius, Vll 196). Hadot [1994], p.348, translates
this by 'Arguments hypothetiques du point de vue de leurs lois', and he adduces our Epictetan
text for this use of 'v6J.!oc;'. If Hadot is right, then the Epictetan usage goes back (at least) to
Chrysippus. But his version of 'Eic; toile; v6J.!ouc;' is strained. Rather translate: 'Hypothetical
arguments bearing on laws'. (Tempting to connect it with the celebrated argument ascribed to
Diogenes the Cynic at Diogenes Laertius, VII 72, on which see Schofield [1991], pp.130-
135.)
 EPICTETUS 91

a permission to infer Y from X. Rather, it is a rule of acceptance, a rule

obliging us to accept Y once we have hypothesized X. Hence it seems that
there are, after all, special rules for hypothetical argumentation. But the rule
is based upon a logical connection: we are to accept something if it follows
from the hypothesis. In effect, then, the rule is the hypothetical counterpart
of the first of the five Chrysippean indemonstrables. 265 According to the
indemonstrable, from a conditional assertible and its antecedent the
consequent follows. According to the hypothetical rule, given that if P then
Q, then if we have hypothesized that P we must accept that Q. (We must
accept it, of course, in the context of the hypothetical argumentation-it
would be absurd to imagine that we are saddled with it tout court.)
What other rules are there? I vii gives us the same rule as I xxvi and
adds a mate:

Should we accept what follows from <a hypothesis> and not accept what
conflicts [t<'x J..LClXOJ..LEVa]?-Yes. ( 24)

That is to say, if you have hypothesized that P, then if 'Q' conflicts with
'P'; then you must not accept 'Q'. Ancient logic famously knew different
analyses of conditional assertibles. According to one of those analyses-
which is securely attributed to Chrysippus-'a conditional is sound when
the opposite of its consequent conflicts [j.HXXEtat] with its antecedent'. 266
Hence 'Q' conflicts with 'P' if and only if, if P, then not Q. Hence the two
rules paired in 24 can be put like this:

(I) If you have hypothesized that P, then, given that if P, then Q, you must
accept that Q.
(2) If you have hypothesized that P, then, given that if P, then not Q, you must
not accept that Q.

Perhaps we may think to find a third rule hinted at in 24-25: Do not

accept an impossible hypothesis. 267 (But it is not clear whether Epictetus
presents this as a rule of hypothetical argumentation, or simply as a sound
piece of advice.) It is not in the least difficult to invent other hypothetical
rules-for example, a rule corresponding to the second Chrysippean
indemonstrable:

265 See above, n.227.

266 Sextus, PH II Ill; see e.g. Frede (1974a], pp.82-91.
267 So, explicitly, Taurus, apud Philoponus, aet mundi 188.
92 CHAPTER THREE

(3) If you have hypothesized that P, then, given that if Q then not P, you must
not accept that Q.

But I have not hit upon any other rules in the surviving texts. 268

The thin and frail character of hypothetical logic makes the second of my
two questions the more pressing: why be interested in hypothetical
argumentation at all? Presumably Chrysippus thought that the question had
an answer-otherwise he would hardly have written on the subject. Perhaps
the question may best be approached by asking what the point and purpose
of the hypothetical rules might be; that is to say, what is the aim of
hypothetical argumentation? Is it merely a game-perhaps a serious game,
a game which trains us in logic-or does it also have some more directly
scientific purpose and value? Certain texts strongly suggest that Epictetus
has in mind a sort of game-a logical exercise, similar to the 'dialectical'
exercises which Aristotle describes in his Topics. Thus Epictetus asks: 'Is it
up to you to assume whatever hypothesis you want?' (II xxix 39). The
implied answer is: No. And it is tempting, if not mandatory, to infer that he
is thinking of some form of intellectual gymnastics. 269 But in any case -as
Aristotle's Topics may remind us-an exercise which is sometimes
undertaken in gymnastic spirit may also have a more substantial philo-
sophical or scientific purpose.
A linguistic tum seems to lead us in just such a serious direction. In a
couple of the texts in which he refers to hypothetical arguments, Epictetus
uses the verb 'e<po8Euttv' (III xxiv 78; IV iv 14). Outside logical contexts,
the verb means 'spy out' or 'reconnoitre' (in a military context) or 'chart'
or 'explore' (of a geographer). Such a word no doubt lends itself readily
enough to metaphorical use; and a phrase such as 'explore the hypothetical
arguments' is scarcely puzzling. Yet we might wonder if Epictetus had any
particular reason to use it here.
Now the associated adverb, 'e<po8EU'tuc&c;' makes a celebrated
appearance in Sextus Empiricus. 270 Sextus is explaining what the logicians

268 One of the titles in the Chrysippean catalogue is said to read thus in the MSS: A6yot
\mo9E'tllmt 9EOOP111-HX'trov (Diogenes Laertius, VII 196). It is an odd phrase, and I suspect that
it is corrupt. Perhaps it refers to a work on rules or principles (9Erop{]f!a'ta) concerned with
hypothetical argumentation?
269 You cannot say 'Change the hypothesis' ( 39); you cannot say 'Don't offer me a
tropic of this sort----Qffer me one of that sort', nor can you say 'Don't infer this
conclusion-infer that one' ( 40). Then what can you do? It sounds as though you simply
listen to someone else producing an argument-and putting it down to your account. But that
cannot be right. I take it that X gives Y some hypothesis to defend, the hypothesis that P. The
rules do not allow Y to say 'Give me the hypothesis that Q'. Then X offers Y the
thought-say-that if P then Q. The rules do not allow Y to say 'Offer me the thought that
either P or R'. But Y is not, of course, obliged to accept the thought offered to him.
270 Eight appearances in Sextus (PH II 141-142; M VIII 307 -309)-and never elsewhere,
apparently, in surviving Greek. Nor is the adjective 'EcpoOEU'tL!(oc;' ever found.
 EPICTETUS 93

take to constitute a proof or <'x1t60Etl;tc;. 271 A proof is a special sort of

argument or A.6yoc;. To count as a proof, an argument must be valid, and
'true', 272 and 'probative' (insofar as its premisses present something clear
and its conclusion something unclear). But such properties are not enough;
for some arguments of the sort thus far delineated yield their conclusions
'only in an exploratory fashion [E<pOOEUnK&c;] and not also in a revelatory
fashion [EKKaAU1t'ttK&c;]' (PH II 141 ). A genuine proof 'unveils' or reveals
its conclusion: it proceeds not only E<pOOEUnK&c; but also EKKaAU1t'ttK&c;.
Sextus does not explain the adverb 'e<poOEU'ttK&c;', and neither does anyone
else; but it seems plausible to infer that it was a technical term in
logic-and probably in Stoic logic 273- , and that it was used to dis-
criminate proofs proper from other probative arguments.
It is tempting to connect Sextus' rare adverb with Epictetus' verb.
Epictetus does not refer to people who explore hypothetical arguments in
the sense of making a study of them. Rather, he refers to people who
conduct such arguments; and the verb he chooses indicates that hypo-
thetical arguments, although they do not constitute proofs, nevertheless are
in principle probative arguments-that is to say, arguments which possess a
certain epistemic value.
The suggestion has its attractions; but two reasons incline me to reject it.
The first reason concerns the adverb. Arguments which are only
exploratorily probative are nonetheless arguments, A.6yot; that is to say,
according to Sextus in this context (see PH II 136), their component parts
are assertibles or <'xl;trojla'ta. And there is no hint at all in Sextus that the
distinction between merely probative arguments and proofs runs parallel to
the distinction between hypotheticals and assertibles. What precisely the
Sextan distinction amounts to is obscure and controversial-it appears that
in a genuine proof at least one of the premisses must be derived from a
preconception or 1tp6A.rpjftc;, that it must be a conceptual truth. 274 But
however that may be, there is no whiff of hypotheticals in the text.
Secondly, Epictetus uses the verb '<poOEUEtv' in other texts where there
is no question of probative arguments. Thus at II xxi 21 he urges us to
'explore [e<poOEUEtv]' our impressions; and earlier, in 17, he asks 'Do you
not explore [e<poOEUE'tE] the premisses of the Liar, the hypothetical
arguments?'. In neither of these passages can the verb be associated with

271 PH II 135-143; M VIII 301-314 (cf 411-428): see Brunschwig [1980]; Ebert [1991],
pp.232-279.
272 I.e. its premisses (and hence its conclusion) must be true: see below, p.l 02.
27 3 On the origins of the definition see Brunschwig [1980], who opts for the Stoics (with
subtle discriminations among different versions of the definition); Ebert [1991], pp.287-31 0,
who thinks of 'the Dialecticians', i.e. of Diodorus Cronus and his crew. Whatever the origins
of the thing may have been, it is clear that Sextus regards it as common property of 'the
logicians'-among whom he certainly includes both Stoics and Peripatetics.
274 Note 'npon"-ficp9at', PHil 142.
94 CHAPTER THREE

the technical Sextan adverb; and in the second passage it is connected with
hypothetical arguments. Moreover, the verb is not uncommon outside
Epictetus: the general idea which, in its metaphorical use, it conveys is the
idea of going about something methodically or systematically-thus it can
be used to mean 'survey' or 'go through' a set of considerations, 275 'carry
out' an inquiry or investigation,276 'set out' an argument or proof, 277 and
even 'argue'. 278 In other words, Epictetus' use of 'f:<poOEUEtv' is not
striking and requires no particular explanation-certainly, it will not help us
to understand the nature or the aim of hypothetical argument.
Then let us tum to the main text, I vii 22-29. It divides into two
subsections and discusses two distinguishable issues. Epictetus declares that
these issues arise because 'it is sometimes necessary to postulate some
hypothesis as a sort of base for the argument which follows' ( 22). Why
and in what circumstances is it 'necessary to postulate some hypothesis'?
Epictetus cannot mean-the context excludes the idea-that sometimes the
rules of the game require you to make a hypothesis. Rather, he means that
sometimes you cannot avoid arguing from a hypothesis. When and why
not?
Well, at any point in a serious argument you may say 'Let us suppose
that ... '-sometimes for exploratory reasons, sometimes for a reductio. 279
But this is not the only use for such hypothetical expressions, nor (I think)
is it the use which Epictetus here has in mind. Rather, he is alluding to the
commonplace that 'not everything can be proved'; 280 that is to say, it is a
commonplace that sometimes, when you have proved a theorem in ethics or
in physics, you cannot further prove the premisses on which that proof was
based-rather, you have argued 'from first principles', from items which
cannot themselves be proved.
If you cannot prove first principles, what can you do with them? Well,
according to one view, you must postulate the principles; or, equivalently,
the principles must be laid down as hypotheses. Thus hypothetical
arguments will-sometimes at least-be highly serious things. For every
scientific proof will depend on a hypothesis, either directly (if its premisses
are axioms or first principles) or else indirectly (if its premisses are

275E.g. Sextus, PH I 200, 209.

276See e.g. Sextus, M VJII 222; VI 6.
277 See e.g. Sextus, M VJII 283; Alexander, in Met 166.19.
278 See e.g. Alexander, in Met 125.2-4.- Other words of the family-'7tEptoliEilEtv' (e.g.
JII xv 7; Philodemus, sign xvii 32), 'litE~OiiEilEtv' (e.g. Sextus, PH II 202, 227-228}--are
familiar from other philosophical contexts.
279 For hypotheses in reductio arguments see e.g. Alexander, in APr 24.12-19; cf Barnes
eta/ [*1991], p.77 n.\48.
280 A commonplace-and also a truth? It is true that, in any axiomatised system (in any
'science') there are some propositions which are not proved; it is not evidently true that there
are some propositions which are not proved in any axiomatised system. See Barnes [1993d].
 EPICTETUS 95

themselves derived theorems). And in that case, the axioms of each science
are perhaps most honestly presented by way of sentences of the form 'Let it
be supposed that ... '; and thus arguments involving such axioms are,
trivially, hypothetical arguments.
Given that we must sometimes postulate hypotheses, what difficulties
does the postulant face? Epictetus asks: 'Should we concede all hypotheses
we are given or not all? And if not all, which?'( 23). He does not offer an
explicit answer. I xxix 39-41, which I have already cited, suggests that we
have no choice in our hypotheses: we must accept what we are given and
argue from it as best we can. 281 Now it is plausible enough to suppose that
in a game or exercise in hypothetical argumentation you were obliged to
stick by whatever hypothesis you were given; and the terminology of our
text-in which you are given hypotheses rather than choose them- perhaps
suggests the context of an intellectual game. But this cannot be right: in I vii
Epictetus is talking of the Sage's life, not of the student's preparatory
exercises. The question 'Are all or only some hypotheses to be accepted?' is
evidently not a question about the rules of a game, and Epictetus clearly
implies, first, that we should not accept all hypotheses, and secondly, that it
is not an easy matter to determine which we should accept.
The Pyrrhonian sceptics objected to reasoning based on hypotheses:
such reasoning, they urged, can never accomplish anything, since any
choice of hypothesis must always be arbitrary. Hypotheses are, by
defmition, items which are simply postulated: they are not-they cannot be
-supported by argument or based upon reasons; they are, so the
Pyrrhonists put it, 'bare affirmations'; and there is, in a literal sense,
nothing whatever to be said for any hypothesis. Suppose, then, that you
hypothesize that P, in the hope of drawing some interesting consequences
from it. I will proceed to hypothesize that not P (or that Q, where 'Q' is any
proposition incompatible with 'P'); and I will draw appropriate
consequences from my hypothesis. The results of my reasoning are no
better and no worse than the results of yours; for my hypothesis is no better
and no worse than yours-no hypothesis can be better or worse than any
other hypothesis. Hence no hypothetical argument can show anything. And
if all scientific reasoning is ultimately based on hypotheses, no scientific
reasoning has any probative value. 282
'Should we accept all hypotheses we are offered or not all?'-If we are
offered a choice among P1, P2, ... , Pn, we can give no particular reason for
picking any Pi; for precisely insofar as we are giving a reason, we are no
longer making a hypothesis. Hence, it seems, we should accept either every

281 See above, p.29.

282 Texts: sextus, PH I 173-174; M VIII 369-378; III 1-17; discussed by Barnes [1990b ],
pp.100-112; Freytag [1995], pp.l15-160.
96 CHAPTER THREE

Pi or none. And since we evidently cannot accept every Pi we had better

accept none. That is to say, we had better reject all hypotheses and abjure
hypothetical argumentation.
Whatever the ultimate merit of this sceptical consideration, it possesses
an initial charm; and at the very least it poses a challenge to the advocates
of hypothetical argument. If such argument is to be more than a flexing of
the logical biceps, then we must know which hypotheses it is right to
accept, and why it is right to accept them. I think that this is the issue to
which Epictetus alludes in 23. 283 The allusion implies that the issue was
standardly discussed within the ambit of Stoic logical theory; and perhaps it
insinuates that there were standard answers, or at least standard gestures
towards an answer. But Epictetus offers no hint as to what those answers
were.

The second question which I vii raises about hypothetical argumentation

concerns not its starting point but its development. 284

And once we have conceded a hypothesis, should we stand by the concession

come what may, or should we sometimes renounce it? Should we accept what
follows from it and not accept what conflicts?-Yes.-But someone says: 'I'll
bring it about that, having accepted a hypothesis of something possible, you
are led on to an impossibility.' ( 24-25)

The situation appears to be this. Suppose that it is possible that P; and that
you have accepted, as a hypothesis, that P. Then-by the rules of
hypothetical argumentation-you must accept that Q if, given that P, it
follows that Q. But now someone tells you that you must accept that Q-
even though, in this particular case, it is impossible that Q. You have
followed the rules of hypothetical argumentation-and you are snared.
What are you to do? 285

283 The date of the sceptical argument which I have just rehearsed is not certain; but it
goes back at least to Agrippa (a sceptic of unknown date-but between Aenesidemus and
Sextus: see Caujolle-Zaslawsky [I 989]); and something like it was developed far
earlier-note the reference to Timon at Sextus, M III 2. In any case, it was surely around
before Epictetus taught.
284 Note the Chrysippean titles: Ailcru; tiiiv 'HoilA.ou {llto6e'ttlciiiv P' (Diogenes Laertius,
VII 196); Ailcru; tiiiv 'HoilA.ou U1to6etuciiiv 1tpo<; 'AptcrtoKpovta Kat 'A1toAA.iiv a' (197);
and the spurious Ailcrt<; tiiiv 'AA.e~avopou U1to6EttKiiiv y' (196). Evidently, there
were-Chrysippus thought that there were-difficulties to resolve.
285 Note the parallel case in Aulus Gellius, XVI ii: there is a 'law of logic', lex disciplinae
dialecticae (see above, n.264), which requires a 'Yes' or a 'No' in answer to any question(
I-2). You are asked 'Have you stopped committing adultery?', and if you follow the rule you
are snared ( 4-7). So, generally, with sophisms patterned on the Homed Man ( 8-12:
above, n.212). Hence a rider is normally added to the law of logic: 'But do not answer
captious questions' ( 13).- Gellius is not here referring to hypothetical reasoning, and his
'law' is plainly a rule of procedure for an intellectual game. But the parallel to the
hypothetical snare is evident.
 EPICTETUS 97

In general, 'we must learn how someone, having accepted a hypothesis,

will not be led into an absurdity' (III ii 17). In the particular case here, it
might be thought that the absurdity is pretty easy to avoid: you should
simply reply 'Pull the other one'. For presumably Epictetus is imagining a
student being confused by an unscrupulous sophist-or by a probing
teacher. (So that the second issue concerning hypothetical arguments arises
only in the context of logical gymnastics.) A student must be armed against
such people-he must learn a little elementary modal logic. After all, if it is
possible that P, and if, given that P, it follows that Q, then it is possible that
Q. That is a plain truth of modal logic-a truth known to Aristotle. 286 Hence
if it is possible that P, and if someone tells you that it follows that Q, where
it is impossible that Q, then you can be sure that he is wrong. Either it is in
fact possible that Q (though it may not seem so), or it is after all impossible
that P (despite initial appearances), or else, given that P, it does not follow
that Q. The issue raised in 23-24 is rather easy to deal with.
Yet Epictetus suggests that it is a difficult issue-the Sage must be
prepared to argue with the man who produces impossible rabbits from
possible hats; and this shows that Epictetus, once again, is not thinking of
student gymnastics. Perhaps, then, he has in mind serious masters of
sophistry, reasoners whose fallacies are so subtle and so persuasive that
only a Sage can resist them? The reference to sophisms in 26 supports
such a conjecture. But I doubt if it is correct. For we know-thanks in part
to Epictetus himself-of one philosopher who explicitly maintained that, in
some circumstances, an impossible consequence might be validly drawn
from a possible premiss. The philosopher was Chrysippus, who, in
discussing the Master Argument, expressly denied that 'the impossible does
not follow from the possible'. 287
That is to say, Chrysippus maintained something like the following
thesis: for some values of 'P' and 'Q', all of the following three claims are
true: (1) it is possible that P; (2) given that P, it follows that Q; (3) it is
impossible that Q. One of his illustrations concerns a moribund Dio.
According to Chrysippus, ( 1') it is possible that Dio has died; (2') given that
Dio has died, it follows that this man has died (where 'this man' refers to
Dio ); (3') it is not possible that this man has died. Alexander, observing that
Aristotle had already proved that ( 1)-(3) are mutually inconsistent, remarks
that Chrysippus, ignoring Aristotle's proof,

tried to prove that it is not so by certain examples which are not soundly
constructed. He says that in the conditional

286See APr 34a5-7.

287II xix 5 (above, p.76); and see Hiilser [*1987], frags 992-997; see e.g. Frede (1974a],
pp.115-117; Mignucci [1978]; Bobzien [1986], pp.105-116; Ide [1992]; Gaskin (1995],
pp.297-305.
98 CHAPTER THREE

If Dio has died, this man [oiho~] has died

-which is true when Dio is pointed at-the antecedent
Dio has died
is possible insofar as it is possible that it should sometime be true that Dio has
died; but
This man has died
is impossible. For ifDio is dead, then the assertible
This man has died
perishes [<p9dpEcr9a.t], the item which is pointed to no longer existing. (in
APr 177.26-32)

The sentence 'This man has died' expresses an assertible only if the deictic
phrase 'This man' refers to something. Once Dio is dead, 'This man' no
longer points to anything: it cannot point to Dio, since Dio, being dead, is
no longer a man and thus cannot be pointed to by 'This man'; and it
cannot-in the imagined context-point to anything else. Hence the
sentence expresses no assertible. Hence, if you like, the assertible which it
once expressed has 'perished'.
We cannot suppose that some other sentence might now be appropriate
to express this assertible-what other sentence could possibly be more
appropriate than the sentence 'This man has died'? We cannot suppose that
the assertible continues to possess whatever sort of being assertibles possess
but that it has ceased to be expressible-for there can be no assertibles
which are not in principle expressible. Thus you cannot truly say that this
man has died. The assertible, that this man has died, is an impossible
assertible. And (perhaps) it is impossible that this man has died.
Chrysippus' view is heterodox-by ancient as well as by modem canons
of modal logic. But it was apparently an integral part of his logic, not an
aside or a joke made in the course of a dialectical debate. Epictetus knew
about his heterodox view, and it seems to me plain that he alludes to it in
I vii. The skilful interlocutor in 25 is not an unscrupulous sophist: rather,
he is a Chrysippean logician.
Epictetus implies that this Chrysippean view was still discussed; and he
insinuates that it was taken to raise difficult problems. Alas, he gives no
hint as to how a Sage might set out to solve or to skirt those problems.
 EPICTETUS 99

J: Changing arguments

Changing arguments, A.6yot 11E'ta1tt1t'tOV'tEc;, are mentioned in two titles

in the catalogue of Chrysippus' logical works-but both titles are marked as
spurious. 288 Sextus Empiricus preserves what is generally taken to be an
example of a changing argument (PH II 231, 234). There are no other
surviving references to changing arguments outside the pages of Epictetus.
It is evident from the Discourses that the study of such arguments was a
fixed part of Stoic logic in the imperial period. Yet apart from the
Discourses nothing suggests that the things had any particular importance in
Stoicism or aroused any particular interest among the Stoics.
What are changing arguments? No text offers us a formal definition-or
even a rough characterization. But presumably changing arguments are to
be connected with a more familiar feature of Stoic logic, changing
assertibles or al;trolla'ta 11E'ta1timov'ta.289 Changing assertibles are
assertibles which change their truth-value, assertibles which are now true
and now false. There is nothing inherently scandalous about such an idea.
Those modem philosophers who treat truth as a predicate of sentences are
happy enough with changing truth-values. The sentence 'It is Monday', for
example, changes its truth-value with monotonous regularity: it is true
every Monday, and false on every other day of the week.
Moreover, change of truth-value may be regarded as a special case of a
more general sentential phenomenon. Just as 'It is Monday' has different
truth-values at different times, so 'I am a pipe-smoker' has different truth-
values in different mouths; 'This is a cultural desert' has different truth-
values in different places; and so on. In general, when speaking of the truth-
value of sentences, we shall use such formulae as 'Sentence S is true (or:
false) when produced by utterer U, at time t, in place p, ... '. And it is plain
that one and the same sentence may be true when uttered by U at tin p ... ,
false when uttered by U* at t* in p*, ... (or by U at t* in p, ... , or by U* at t
in p, ... , and so on).
The 'propositions' which feature in most versions of propositional logic
-the Ps and Qs of the standard propositional calculus-do not and cannot
change their truth-value: they are true or false once and for all. (And
everywhere or nowhere, in every mouth or in no mouth, ... ) That is to say,
propositions accept the 'absolute' operators 'It is true that ... ' and 'It is

288 Diogenes Laertius, VII 195-196: 0Epi tiiiv !J.Etmtbttovtoov 1..6yoov 1tpo<; 'A911VcX011V
a' ('lfEUOE1tl ypa<pov), A6yot !J.EtaltllttOVtE<; 1tpo<; ti]v !J.EO"Otllta y' ('lfEUOEltlypaq>av). The
works must have been in the Chrysippean canon by the first century B.C. when Apollonius of
Tyre drew up his Stoic ntva~, from which the catalogue in Diogenes presumably derives (see
Goulet [1989b ); Barnes [1996]). We have no idea why Apollonius athetized the two works,
nor whether his athetization was reasonable.
289 See Alexander, apud Simplicius, in Phys 1299.36-1300.11; Diogenes Laertius, VII 76.
Cf e.g. Frede [1974a), pp.44-48; Bobzien [1986], pp.21-39; Denyer [1988).
100 CHAPTER THREE

false that ... ', and not the 'relativised' predicates ' ... is true when uttered by
U at t in p ... ' and '. . . is false when uttered by U at t in p ... '. So Stoic
assertibles look more like modem sentences than like modem propositions.
But Stoic assertibles are sayables: they are not sentences. They are what
we can assert by uttering a sentence, not what we utter when asserting
something. On the other hand, Stoic assertibles change their truth-value:
they are not propositions. To the modem eye they may therefore seem
monstrous hybrids, half sentence and half proposition. In the ancient world,
they were perfectly familiar denizens of the logical jungle. 290 (And it may
be suspected that they are perfectly familiar today, outside the zoological
gardens of the logicians. 291 ) On Monday I utter the sentence 'It is Monday',
and on Tuesday I utter the very same sentence, 'It is Monday'. On each
occasion I have expressed one and the same Stoic assertible, and I have in
all probability asserted one and the same thing-namely, that it is Monday.
On Monday what I asserted was true, on Tuesday what I asserted -that
very same thing-was false. What I asserted, the assertible which I
expressed, changed its truth-value.
It follows that it is absurd to urge that we should always stick to our
judgements: if I want to think and speak truly about the date, I must change
my mind seven times a week. And one text appears to indicate that
Epictetus saw the point. A friend of his had decided to starve himself to
death. Epictetus visited him:

'I have decided', he said.-Still, what was it that persuaded you? For if you
judged rightly, we'll sit down and work together to get you out of the world.
But if you judged irrationally, change your mind.-'One should stick by one's
decisions.'-What on earth do you mean, my friend?-'Not all decisions, but
those rightly made' .-And if you have recently felt that it is night, then if that
is your opinion do not change it-stick by it and say that one should stick by
one's judgements. 292

Respect for the truth requires us to change our minds: if you have recently
judged, and correctly, that it is night, it is folly to stick by your judgement
-for if you go on believing that it is night, you will be wrong half the time.
Epictetus, it seems, is implicitly rejecting-or at least modifying-a
standard Stoic doctrine, the doctrine that 'unchangingness [UJ.Uo'ta1t1:rocria]
and constancy in judgements is a very great good' (Plutarch, comm not
1061E). But, alas, this is not what Epictetus means. It emerges from 14

290 The classic paper is Hintikka [1967].

291 'The real axiom of Identity is this: What is true in one context is true in another. Or, if
any truth is stated so that a change of events will make it false, then it is not a genuine truth at
all. To most readers this axiom, I have little doubt, will seem a false statement' (F.H.Bradley,
The Principles of Logic (London, 1883), p.l33). I cite the passage for the last sentence, not
for the horrors which precede it.
292 I xv 6-7-1 have changed the standard punctuation.
 EPICTETUS 101

that he is simply urging his friend not to stick by judgements which tum out
to be false. If he recently judged that it was night, then his judgement was
false-and so of course he should be prepared to abandon it, even if it was
'rightly made'. (And 'rightly' here means 'well-groundedly', not
'truly' -perhaps he had dozed off and woke up suddenly in a darkened
room.) The point concerns intellectual pig-headedness, not the subtleties of
changing truth-values.
Nonetheless, it is plain that if it is sometimes the case that P and
sometimes the case that not P, then you should change your judgements as
the assertible changes its truth-value. I assume that the Stoic doctrine of
'unchangingness' must have recognized this fact. I know of no text which
indicates how the Stoics accommodated it.
If assertibles change their truth-value, that is because they contain what
are called 'indexical' elements-elements whose reference depends on
context and circumstance. If I utter the sentence 'It is Monday' and thereby
say that it is Monday, then there is an element in the assertible which
corresponds to the verb 'is' (or rather, to the tense of the verb 'is'): on
Mondays this element picks out Monday, and on Tuesdays the same
element picks out Tuesday. Now verbs-tenses of verbs-are not the only
sentential elements which are in this sense indexical: pronouns ('I', 'you',
'he', ... ), and demonstrative phrases ('This duck', 'That argument', ... ),are
also indexical. If assertibles contain elements which correspond to tenses, it
is easy to imagine that they will also contain elements which correspond to
other indexical items in the sentences which express them. Thus if, say, I
utter the sentence 'I am a pipe-smoker' and you utter the same sentence 'I
am a pipe-smoker', we might suspect that, on the Stoic view, you and I
would say one and the same thing, express one and the same assertible.
And, in virtue of the indexical element in the assertible, what is said may be
true in my mouth and false in yours. In general, if a sentence S is true
relative to U at tin p ... , the Stoic assertible which S expresses (or so we
might suppose) will be liable to vary its truth-value according to the identity
ofU and t and p and ...
But this was not the Stoic view. Rather, the Stoics thought that when I
utter the sentence 'I am a pipe-smoker' I may say that Barnes is a pipe-
smoker, whereas when you utter the same sentence you will say something
different. It is not that we have a single assertible with different truth-values
in different mouths: rather, there are two distinct assertibles which happen
to be expressible by one and the same sentence. It is not easy to give a clear
and precise account of the matter: the texts are meagre, their interpretation
disputable. It is a plausible conjecture that Stoic assertibles contain
indexical elements corresponding to the tenses of verbs and to other
temporally indexical items-'today', 'now', 'a long time ago', 'in the past',
102 CHAPTER THREE

and the like; and that they do not contain indexical elements corresponding
to other linguistic indexicals. But no ancient text says that this is so-let
alone explains why it should have been taken to be so.
However that may be, it is plain that for the Stoics-as for all ancient
philosophers-the primary truth-bearers may change their truth-value. Not
all assertibles, of course, can change their truth-value. Whenever and
wherever anyone utters (in favourable circumstances) the sentence '2 + 2 =
4', what he says is true; and whenever and wherever anyone utters (in
favourable circumstances) the sentence '2 + 2 = 5' what he says is false.
There are also assertibles which can in principle change their truth-value
but which in practice do not do so. If throughout your life you never
experience the sublime pleasure of a pipe of Dunhill Standard, then the
assertible which you may express by uttering the sentence 'I am smoking a
pipe of Dunhill Standard' will always be false-even though, in a better
world, it might sometimes have been true. The Stoic nomenclature,
'al;uhJlata JlEta1tbttovta', has suggested that changing assertibles are
assertibles which actually do change their truth-value rather than assertibles
which in principle can change their truth-value; 293 but it would be
imprudent to attach much weight to this.
Stoic arguments are defmed as 'systems' or ordered sets of assertibles. 294
Hence if an assertible changes its truth-value, we shall imagine that such a
change will have some sort of effect on any argument in which it makes an
appearance. But what precisely are we to understand by a changing
argument?
Well, inasmuch as assertibles are called 'changing' because they change
their truth-value, why not suppose that arguments were called 'changing'
when they change their truth-value? Admittedly, modem logicians fight shy
of applying the word 'true' to arguments; but the Stoics had no such
inhibitions: rather, they said that an argument was true if and only if it was
valid and all its premisses were true. 295 In that case, an argument might
change its truth-value for either of two (non-exclusive) reasons: it might
change in validity, or its premisses might change their truth-value.
How might an argument be now valid and now invalid? Such a change
can readily be explained in terms of change of truth-value; for according to
the Stoics, an argument is valid if and only if the conditional formed from
the conjunction of its premisses as antecedent and its conclusion as
consequent is true. That is to say, an argument of the form 'A1, Az, ... ,An:
therefore B' is valid if and only if the conditional assertible 'If (A 1 and A 2
and ... and An), then B' is true. 296 Then if the corresponding conditional

293 So Bobzien [1986], p.l24 n.l4; contra Frede [1974a], p.44.

294 Texts in HO!ser [*1987], frags 1038-1049.
295 Texts in Hiilser [*1987], frags 1064-1079.
296 Texts in Hiilser [*1987], frags 1058-1073.
 EPICTETUS 103

changes its truth-value, the argument will change its validity: it will be valid
when the conditional is true, invalid when the conditional is false.
Now, as I have said, not all assertibles change--or can change-their
truth-value; and although some conditional assertibles were certainly
supposed to change their truth-value, it might be doubted that the
conditional assertible corresponding to an argument can do so. Whether or
not it can will depend, of course, on the nature of the connective in the
corresponding conditional. There were different accounts of conditional
assertibles current in the ancient world. 297 Consider, say, the assertible
expressed by the sentence

If Barnes is smoking his pipe, Geneva is in France:

if we apply a 'Philonian' account of conditionals, according to which a

conditional is true at a time t if and only if it is not the case that its
antecedent is true at t and its consequent false at t, then this assertible will
be true whenever I happen not to be smoking and false at all other times. In
fact, it changes its truth value about ten times a day.
An example of this sort is ascribed to the Stoics by Alexander: 'If Dio is
alive, Dio will be alive'.

Even if this is true now, inasmuch as it begins with a truth, 'Dio is alive', and
ends in a truth, 'He will be alive', nonetheless there will be a time when, the
further assumption [1tp6crA11'1ftc;], 'But Dio is alive', being true, the
conditional will change to being false; for there will be a time when, 'Dio is
alive' still being true, 'He will be alive' will not be true, and if this is not true
then the conditional as a whole changes and becomes false-it is not always
the case that when 'Dio is alive' is true so too is 'Dio will be alive'; for in that
case Dio would be immortal. (Alexander, apud Simplicius, in Phys 1300.1-8)

Insofar as he refers to a 'further assumption' Alexander suggests that the

conditional in question is being treated as a premiss in an argument. But it
is not the conditional which corresponds to the argument itself.
Moreover, Alexander plainly takes the conditional a la Philo; and it
seems clear that the conditional assertibles which correspond to arguments
are not to be construed a la Philo. Rather, they are to be construed in the
Chrysippean fashion and explained in terms of 'connectedness' or
cruvap'tllcrts; that is to say, 'If P, then Q' is true if and only if the
contradictory of 'Q' 'conflicts' with 'P'. 298 Can any Chrysippean con-
ditionals change their truth-value? We could answer this question if we
knew what 'conflict' was supposed to be; but we do not. 299 And although I

297 Above, p.91: texts in Hiilser [*1987], frags 952-965; see e.g. Frede [1974a], pp.80-93.
298 See above, p.91.
299 For a guess see e.g. Stopper [1983].
104 CHAPTER THREE

incline to think that Chrysippean conditionals probably could in principle

change their truth-value, this is only a guess.
However that might be, it seems to me that the Stoics ought to have
discussed the possibility of an argument's changing its validity. (And for all
we know, they actually did so.)
It is far easier to see how an argument might change its truth-value in the
second way-in virtue of a change in the truth-value of one or more of its
premisses. The text I have cited from Alexander implies just such an
argument-and I guess that Alexander was drawing on a Stoic discussion of
changing arguments. Or consider the following syllogism:

If it's Monday, it's not Tuesday.

It's Monday.

Therefore: it's not Tuesday.

The second premiss of this argument is true on Mondays, false on any other
day of the week. Hence the argument is true on Mondays, false on any other
day of the week, since only on Mondays are all of its premisses true. There
are one or two hints that the Stoics considered such things; 300 and I do not
see why they should not have called them 'changing arguments'.
The conditional corresponding to an argument might, in principle,
change its truth-value. A premiss of an argument might, in principle,
change its truth-value. And there is a third possibility which might as well
be mentioned: one of the component assertibles of an argument might
change. Any argument a premiss of which changes truth-value is an
argument a component of which changes truth-value. But not vice versa; for
the component assertible which changes its truth-value need not be a
premiss of the argument. Consider the following argument:

If either it is Monday or Frege wrote Begriffsschrift, then there was at least

one German logician.
Frege wrote Begriffsschrift.

Therefore: there was at least one German logician.

The assertible expressed by the sentence 'It is Monday' changes its truth-
value. But no premiss in the argument changes its truth value, nor does the
argument itself change its truth-value. Nonetheless, it might seem not
umeasonable to refer to an argument as changing if one of its components
changed truth-value.

300 See the text of Alexander; and note e.g. Sextus, PH II 139 (an argument is said to be
false 'while it is day'); MVIII 418 (similarly).
 EPICTETUS 105

When Epictetus talks of changing arguments in I vii, he has in mind

something rather more specific than any of the characterizations I have thus
far given. In 20 he enjoins us to 'consider premisses [A.fu.tJlCt.'ta.] ... and
their alterations and changes as a result of which, in the very process of
questioning or answering or inferring or whatever, they undergo changes'.
The phrase is not pellucid; but it is clear that the changing arguments to
which Epictetus alludes in I vii change insofar as their premisses change
truth-value and do so in the course of the argument- that is to say, while
the premisses are being propounded, while they are being accepted, while
the conclusion is being drawn, and the like. 301 Changing arguments--or at
least, the changing arguments to which Epictetus draws attention in I
vii-are so-called in virtue of the fact that one or more of their premisses
change truth-value in a particular way. 302
How might an argument change in this particular way? Consider again
my earlier example:

If it's Monday, it's not Tuesday.

It's Monday.

Therefore: it's not Tuesday.

Suppose that I propound the argument very late on Monday evening. I

begin with the first premiss at 11.59, and-speaking in a slow drawl-I
produce the conclusion just after midnight has chimed. The first premiss is
always true. The second premiss was true when I propounded it; but it
became false by the time I arrived at the conclusion. Perhaps, then, this is a
simple example of what Epictetus is here thinking of as a changing
argument?
Such an argument changes in the course of being propounded, and it
does so in virtue of the second of the two ways in which an argument can
change its truth-value-it does so in virtue of a change in the truth-value of
the premisses. This suggests that we might add two other potential sorts of
changing argument to the sort which Epictetus has delineated: first,
arguments which, in the course of being propounded, change their validity;
and secondly, arguments which, in the course of being propounded, suffer a
change in truth-value of one of their component assertibles.
However that may be, it may be wondered whether my sample argument
is really an example of what Epictetus has in mind. And at any rate we can
-and may as well-introduce a distinction here, whether or not it
contributes to the understanding of Epictetus' text. My sample argument

301 On the idea of a change 'in the course of the argument' see below, p.125.
302 Frede [1974a), p.48, takes this to be the defining feature of changing arguments;
contra Bobzien [1986), p.l27 n.27.
106 CHAPTER THREE

sometimes, but not always, changes--or rather, it will change while being
propounded in certain circumstances, but it will not change while being
propounded in all circumstances. Perhaps Epictetus is thinking not of such
contingently changing arguments, but rather of arguments which change in
whatever circumstances they are propounded? We shall shortly see that
there are examples of such things to hand. And so we may introduce a
further kind of changing argument, again with three varieties: first,
arguments which, whenever they are propounded, change their validity;
secondly, arguments which, whenever they are propounded, suffer a change
in the truth-value of one or more of their premisses; and thirdly, arguments
which, whenever they are propounded, suffer a change in the truth-value of
one of their component assertibles.
Let me summarize: we may distinguish arguments which
(I) for some period or other
(a) are valid at one time in that period and invalid at another, or
(b) have at least one premiss which is true at one time in that
period and false at another, or
(c) contain at least one assertible which is true at one time in that
period and false at another;
(II) for some period during which they are being propounded
(a) are valid at one time in that period and invalid at another, or
(b) have at least one premiss which is true at one time in that
period and false at another, or
(c) contain at least one assertible which is true at one time in that
period and false at another;
(III) for any period during which they are being propounded
(a) are valid at one time in that period and invalid at another, or
(b) have at least one premiss which is true at one time in that
period and false at another, or
(c) contain at least one assertible which is true at one time in that
period and false at another.
Epictetus is thinking of arguments which meet condition (II)(b) or perhaps
condition (III)(b). Perhaps the Stoics limited the term 'changing argument'
to arguments which meet one or other of these conditions-or perhaps they
used the term more generously to cover several, or even all, of the
arguments I have characterized.

Why should changing arguments seem puzzling or perturbing? According

to Epictetus

there are cases in which we have soundly granted the premisses and such-and-
such results from them: although it is false, nonetheless it results. (I vii 13)
 EPICTETUS 107

That is to say, there are arguments which are plainly valid, in which all the
premisses are true, and of which the conclusion is false. And there is the
paradox; for a valid argument with true premisses must surely-and
trivially-have a true conclusion. 303
Sextus, as I have said, offers us what is generally taken as an example-
the only surviving ancient example-of a changing argument. 304 He is dis-
cussing sophisms, and he says that, according to the logicians, 305 'a sophism
is a plausible and treacherous argument leading one to accept a conclusion
which is either false or similar to something false or unclear or in some
other way unacceptable' (PH II 229). 306 The four types of conclusion in
effect distinguish four types of sophism, and Sextus next gives an example
of each. The argument which concerns me here illustrates the third type, the
type which leads to an unclear conclusion.
A little later Sextus remarks, opaquely enough, that 'inference
[<'utayoyyfJ] to the unclear, they say, is from the class of things which
change [h: tou yvouc; t&v jlta1tt1tt6vtrov Ecrt\.v]' ( 234). If the last
clause may be paraphrased as 'belongs to the class of changing arguments',
then Sextus does classify our example as a changing argument. But a better
paraphrase reads 'depends on the class of changing items'; and Sextus
probably means to say that the sophistical arguments of the third type turn
on the fact that (at least) one of their component assertibles changes its
truth-value. Thus Sextus does not explicitly offer his argument as an
example of a changing argument. Nonetheless, it is not outrageously
audacious to suppose that, in point of fact, the argument which Sextus
presents was taken to be, or would have been taken to be, a changing
argument. Of course, it is a further question whether the argument actually
is a changing argument and which (if any) of the characterizations I have
given apply to it. In any event, I shall discuss Sextus' argument here.
This is it:

It is not the case both that I have propounded something to you already307 and
that it is not the case that the stars are even in number.

303 Note the parallel (cf I vii 22) with the problematical hypothetical argument: an
argument with possible premisses is valid-and has an impossible conclusion.
304 On the passage see Ebbesen [*1981), I pp.23-24; Ebert [1991), pp.l89-190; Atherton
[1993], pp.424-450.
3 5 oi lltaAEKtucoi: or rather 'some logicians ... (for others say other things)' ( 235). It
is futile to try to put a name to these logicians (pace Ebert [1991), pp.177-182-see Atherton
[1993), p.428 n.28).
36 The origin of the classification is disputed. It is evidently not a formallltaiptcrt;, and
I suppose that it was a makeshift and practical device without any theoretical pretensions.
307 1tpi.Otov: perhaps, more specifically, 'as a first premiss'. -Note, in any case, that the
sentence 'I have propounded something to you already' must be taken to mean 'I have
already offered you a premiss in this argument' -what may have happened in another
argument is irrelevant.
108 CHAPTER THREE

But I have propounded something to you already.

Therefore: the stars are even in number. (PH II 231 )308

The ftrst premiss has a complex structure: it is the negation of a con-

junction, the ftrst conjunct of which is

I have propounded something to you already.

The second conjunct is

It is not the case that the stars are even in number,

that is to say, it is a negated assertible.The form of the argument can thus be

expressed schematically as follows:

Not (both P and not Q).

P.

Therefore: Q.

The argument has the form of a Stoic third indemonstrable. 309 Hence a Stoic
must accept that it is valid-as no doubt it is.
As for the truth-value of the premisses, the commentary which Sextus
ascribes to his logicians is lucid enough:

Since-let us suppose-nothing has yet been propounded, the negation of the

conjunction is true. For the conjunction itself is false in virtue of the fact that
it contains the false component 'I have propounded something to you
already' .310

That is to say,

308 oilxt Kat i]protflKa ti cre 7tpi.ihov Kat oilxt oi acrti:pet; iipttoi Eicrt v i]protflKa o
ti (J 7tpi.iltov oi iipa acrtpt; iipttoi icrt v.
309 Given the metalogical account of the indemonstrables, and given that 'Q' and 'not Q'
are opposites, then Sextus' example is a straightforward case of a third indemonstrable (see
above, n.227). Given an account in terms of schemata, this is not-or not evidently-the
case; for it is not evident that Sextus' example has the precise form:

Not (both A and B).

A.

Therefore: not B.

310 J.Lll0VOt; yap 7tpOflpCJJtflJ.LEVOU KCLta ti]v l.l7t09crtV tO cl7tOq>UttKOV tilt; CJUJ.l7tAOKllt;
aA.fl9Et; yivetat, weuoout; tfit; crUJ.L7tA.oKfit; ol>cr11t; 1tapa to EJ.l7tE7tA.xem weuoot; to
i]protflKa ti cre 7tpi.iltov v ailtft.
 EPICTETUS 109

I have propounded something to you already.

is false-for the argument has only just begun. Hence the conjunction,

I have propounded something to you already and it is not the case that the
stars are even in number,

is false-for any conjunction one conjunct of which is false is itself false.

Hence the negation of this conjunction is true. And the negation of this
conjunction is the ftrst premiss of the argument.

But after the negation of the conjunction has been propounded, the further
assumption ('I have propounded something to you already') becomes true,
since the negation of the conjunction has been propounded before the further
assumption. (PH II 234) 311

That is to say, the truth of the second premiss is guaranteed by the existence
of the ftrst premiss.
Hence we have a valid argument with two true premisses. Hence the
conclusion is true. That is to say, the stars are even in number. Moreover,
we can be sure that the two premisses are true; hence we may infer-and
accept with confidence-that the stars are even in number.
But this is absurd-how could anything about the stars be established by
an argument such as the one Sextus expounds? Moreover, if this argument
establishes that the stars are even in number, then we can establish
absolutely anything we like-including the proposition that the stars are not
even in number. For the content of the proposition whl.ch is represented by
'Q' in the schema plays no part at all in the unfolding of Sextus' argument:
the place of 'Q' may be taken by absolutely any proposition whatever and
the resulting argument will go through exactly as Sextus' argument goes
through.
The conclusion of Sextus' argument is said to be not false but unclear. 312
This is not an accident; for in 231 the argument is expressly adduced to
illustrate the type of sophism which treacherously leads us into the bog of
unclarity. Sextus' logicians apparently hold that any sophism which leads to
an unclear conclusion will depend on changing items; for the remark that

311 f.LEta o to pcot,Silvat to c'xltoqxxtucov t~ crUf.LltAOic~. t~ ltpocrA. TJ'lfEcoc; c'xA.,aouc;

-yEvof.LEVllc; (i]prot,Ka o ti crE 1tpohov) ota to i]pcotijcr6at 1tpo tile; 1tpocrA.i]wEcoc; to
CxltOq>attKOV tile; 0"\lf.lltAOK~, ...
312 The proposition that the stars are even (or odd) in number is a stock example of
something unclear or iio,A.ov (i.e. something 'once and for all' unclear, something which
forever escapes our knowledge), over which a good Stoic must suspend judgement: e.g. I
xxviii 3; II xvi 2; cfSextus, PHil 90, 97.
110 CHAPTER THREE

'inference to the unclear ... is from the class of things which change'
plainly refers to the third type of sophism in general and not just to the
particular example which Sextus has chosen. 313 This opinion is false-and
evidently false. For unclear conclusions can be produced in any number of
sophistical ways. Either Sextus has garbled the view which he ascribes to
his logicians, or else the logicians were a confused bunch.
However that may be, there must be something wrong with the
argument. And-or so perhaps we may suspect-what is wrong with it
must have something to do with the fact that it is some sort of changing
argument. Now it is clear that the change which takes place takes place in
the course of the argument; it is clear, too, that it takes place whenever the
argument is propounded-the change depends not upon the external
circumstances in which the argument is propounded but rather on the
internal content of the component assertibles of the argument. Hence the
example belongs to the third of the categories which I distinguished-it is
an example of type (III)(b) or perhaps (III)( c). And perhaps this is precisely
what accounts for its sophistical character.

Sextus preserves not only the argument itself but also the response which
his logicians gave to it. He does not himself think much of the response-
what the logicians say on these matters, he growls, may 'tickle our ears',
but it is an unnecessary waste of time ( 235). However, he produces no
specific criticism of the response, nor does he suggest anything to
supplement or supplant what the logicians say. This is the response:

After the negation of the conjunction has been propounded, the further
assumption ('I have propounded something to you already') becomes true,
since the negation of the conjunction has been propounded before the further
assumption; so that the proposition of the negation of the conjunction comes
to be false inasmuch as what was false in the conjunction comes to be true.
Hence the conclusion can never be inferred, since the negation of the
conjunction does not hold together with the further assumption. (PH II 234) 314

The response to the argument thus consists of an exegesis and a solution.

First, the exegesis. The logicians claim that 'the negation of the con-
junction comes to be false', i.e. that the first premiss of the argument
changes its truth-value. The second premiss, when it is propounded, is true.

313 But the logicians need not be taken to insinuate that any sophism which depends on
changing items will lead to an unclear conclusion, an insinuation which is certainly false.
314 !!Eta liE tO EpiDt119flvat tO c'utO<pattKOV t~ OUI!ltAOKfj~. tfl~ 1tpooA.'i]ljiEID~
<iA-1190\i~ "'(EVOI!EV~ (i]pWt11KU liE ti OE 1tprotov) lita to i]pmtflo9at 1tp0 t~ 1tpooA.TjljiEID~
to altmpattKov t~ ou~-tltA.oK~. ti toli altO<pattKoli t~ OlJI!ltA.oK~ 1tp6taot~ yi vEtat
ljiEUii~ toli EV til> OU1!1tE1tAE"'f1!EV(!l ljiEilliou~ "'(EVOI!EVOU <iA.119oli~ ro~ 1!1lliE1tOtE liilvao9at
OUVUX9flvat tO OUJ..LltEpUOI!U 1!i] OUVUltCXpXOVtO~ tOU UltOcpattKOV tfl~ OUI!ltAOKfj~ tft
1tpooA.i]'lfEL
 EPICTETUS 111

But the second premiss is the first conjunct of the conjunction negated in
the first premiss. That is to say, it is the same assertible as that conjunct,
given the Stoic account of assertibles. 315 Hence the first conjunct has
changed its truth-value. Hence-<>r so the logicians infer-the first premiss
has changed its truth-value: it was true and it is now false.
But this last inference-as several scholars have justly observed-is
invalid. 'P' was false; and its falsity guaranteed the truth of 'Not (both P
and not Q)', whatever the truth-value of 'Q'. 'P' is now true; but the truth
of 'P' does not in general guarantee the falsity of 'Not (both P and not Q)'.
Rather, given that 'P' is true, then the negated conjunction is false if 'Not
Q' is true, true if 'Not Q' is false. That is to say, the first premiss changes
its truth-value if and only if it is false that Q-if and only if as a matter of
fact the stars are not even in number. 316
The logicians' inference is invalid, and they do not establish that the first
premiss changes its truth-value. Nor, equally, can we show that the first
premiss does not change its truth-value-as long as the number of the stars
remains obscure, we shall not know whether the first premiss changes or
rests. But we surely can tell that the sophism does not depend on any
putative change in the truth-value of the first premiss. For we can see,
evidently enough, that the argument is sophistic, and we cannot tell whether
or not the first premiss changes its truth-value. Thus the invalid inference
which seduced Sextus' logicians led them to a false diagnosis of the
sophism.
It will not help, of course, to change the example-the error does not
derive from the particular choice of illustration which the logicians hit
upon. In any illustrative example, the assertible for which 'Q' stands must
be unclear; and any unclear 'Q' whatsoever will fall foul of the objection I
have just rehearsed.
We might think of refming the logicians' idea. Consider not a single
argument, but a pair of arguments of the Sextan form. In one of the
arguments substitute any unclear statement for 'Q'. In the other substitute
the negation of this same statement (which will also be unclear). Thus take
Sextus' actual argument:

3 15 Note that we cannot replace 'I have already propounded something' by 'This is not the
first premiss of this argument', which may well seem equivalent to it. For the proposed
replacement will not induce any change in truth-value: the sentence 'This is not the first
premiss of this argument' will express two different assertibles in its two different
occurrences in the argument.
316 Suppose, first, that 'P' is true and 'Q' is true. Then 'Not Q' is false. Hence 'Both P and
not Q' is false. Hence 'Not (both P and not Q)' is true. Hence the first premiss has not
changed its truth-value. Suppose, secondly, that 'P' is true and 'Q' false. Then 'Not Q' is true.
Hence 'Both P and not Q' is true. Hence 'Not (both P and not Q)' is false. Hence the first
premiss has changed its truth-value.
112 CHAPTER THREE

A] It is not the case both that I have propounded something to you already
and that it is not the case that the stars are even in number.
But I have propounded something to you already.

Therefore: the stars are even in number.

and couple it with:

B] It is not the case both that I have propounded something to you already
and that the stars are even in number.
But I have propounded something to you already.

Therefore: the stars are not even in number.

Each argument concludes to something unclear. Moreover, if 'Q' is true-

if it is true that the number of the stars is even-then the first premiss of
argument [B] changes its truth-value, and if 'Q' is false-if it is not true
that the number of the stars is even-then the first premiss of argument [A]
changes its truth-value. Since 'Q' is either true or false, either [A] or [B] is
a changing argument. Since we cannot tell whether or not 'Q' is true, we
cannot tell which of the two arguments is a changing argument- but we
can be sure that we have in fact given precisely one example of a changing
argument. Very ingenious-but Sextus' logicians certainly had no such
refinement as this in their minds, and it cannot be adduced to help them out
of the mire.
Or perhaps they are not in the mire at all-perhaps I have been unfair to
their logical prowess? I have been supposing that the logicians took the first
premiss of the argument to change its truth-value. Now although the
premiss cannot be proved to change its truth-value, it can be proved to
undergo a different change-and a change which seems rather pertinent to
the status of the argument as an example of the third type of sophism. For
the first premiss can be proved to change what might be called its cognitive
status: it changes from being clear to being unclear.
Presumably the first conjunct of the conjunction included in the first
premiss, namely

I have not already propounded anything,

is clear or 1tPOOTtA.ov; that is to say, it is either clearly true (when it is true)

or else clearly false (when it is false). Now when the first premiss is
propounded, this conjunct is clearly false. But a conjunction one conjunct
of which is clearly false is itself clearly false; and the negation of something
which is clearly false is itself clearly true. Hence-or so it seems plausible
to conclude-the first premiss itself, when it is propounded, is clearly true.
 EPICTETUS 113

But once the second premiss has been propounded, the truth-value of the
first premiss is no longer clear; for the truth-value of the second conjunct is
unclear, and this infects the whole premiss. (Suppose that it is unclear
whether 'Q' is true. Then it is unclear whether 'Not Q' is true. Hence it is
unclear whether 'Both P and not Q' is true-and hence it is unclear whether
'Not (both P and not Q)' is true.) The argument has changed in this sense:
one of its premisses was once clear and then became unclear. And we might
then imagine that the sophism which the argument presents may perhaps
depend on precisely this change in cognitive status.
Should we then conclude that the logicians who discussed the argument
actually diagnosed not a change in truth-value but a change in cognitive
status-and that Sextus misunderstood or distorted what they had said? And
should we further conclude that Sextus' example-whatever he himself
may have thought-is not and was not presented as a changing argument,
so that no ancient example of a changing argument has been conserved?317 I
confess that I am not myself in the least inclined to draw such conclusions.
Often enough when scholars detect a betise in a 'secondary' author such
as Cicero or Sextus, they incline to pin the error on the secondary author
himself: Chrysippus originally presented a brilliantly clear and concise
argument-which dull old Cicero then fuddled; the logicians originally
wrote with lucidity and exactitude-and silly old Sextus made a fist of it all.
Of course, some primary sources are brilliant and some secondary sources
are dullards-or worse. But in the present case we have no independent
knowledge of the primary source; and as for the secondary source, it is far
from evident to me that Sextus was a dullard. It is reasonable to accuse him
of garbling only if we have some particular grounds, in this or that passage,
for suspecting malicious or careless distortion. In the section on sophisms
there are no such grounds. Sextus' report is pellucid, and he has no axe to
plough or furrow to grind.
Moreover, in the present case we have a special reason not to ascribe an
error to Sextus. For the logicians' resolution of the sophism-to which I am
coming--clearly presupposes that what they diagnosed was a change in
truth-value rather than a change in cognitive status. That is to say, if we
suppose that Sextus garbled the logicians' diagnosis, then we must also
suppose that he garbled their resolution-and garbled it in such a way that it
fits, to a T, the garbled version of the diagnosis. Well, you may suppose
what you will-but this way madness lies. 318

31 7 Unless, of course, we imagine that the ancient category of changing arguments was
broad enough to include arguments the components of which change their cognitive status.
318 'What Sextus should have said, presumably, is that the conjunction becomes iilil]A.ov
... I am reluctant to say whether the fault is his or his source's' (Atherton [1993], p.447 n.56).
But what Sextus should have said is whatever the logicians in fact said: he purports to
114 CHAPTER THREE

I add that a reference to change in cognitive status is not in fact to the

point. As I have said, the content of the proposition which replaces 'Q' in
the sophistical argument is wholly immaterial. Then replace it by something
which is not unclear but rather evidently false; say, by

The number 3 is even.

Then the cognitive status of the first premiss will not change-but the
argument will still be a sophism, and a sophism (or so I suppose) for the
very reason that Sextus' argument is a sophism. Cognitive status, in other
words, is a red herring: truth-value is what matters.
For Sextus' argument does undergo a change in truth-value: one of its
component assertibles changes in truth-value (so that in effect the argument
falls in the class (III)( c)); and in addition one of its premisses changes in
truth-value (so that it also falls in the class (III)(b )). For the assertible
expressed by

I have propounded something to you already

changes its truth-value in the course of the argument-it changes its truth-
value during any period in which the argument is propounded. This
assertible is a component of the argument-it is the first conjunct of the
negated conjunction; and it is also a premiss of the argument-it is the
second premiss. Thus the logicians were quite right to claim that there is a
change of truth-value as the argument is propounded; but they called
attention to this change-for reasons which I cannot fathom-in connexion
with the wrong premiss.
The second premiss is false at the start of the argument-and this is what
ensures that the first premiss is (then) true; and the second premiss is true
when it is propounded. It is precisely this change of truth-value which
allows the argument to proceed regardless of the content of the second
conjunct in the first premiss-and hence regardless of the conclusion. It is
precisely this change in truth-value which accounts for the sophism: not, of
course, for the fact that the sophism concludes to something unclear, but for
the fact that the sophism concludes to anything whatever (and a fortiori to
anything unclear).

I turn now to the resolution which Sextus' logicians offer us. They affirm
that 'the conclusion can never be inferred since the negation of the
conjunction does not hold together with [cruvumipXEt v] the further

rehearse their resolution, not to resolve the sophism himself. And I confess that I can see no
reason for any reluctance.
 EPICTETUS 115

assumption'-that is to say, because the two premisses are never true at the
same time as one another. 'Not (P and not Q)' and 'P' are not true at the
same time. Hence we cannot infer 'Q' from them.
If we are allowed to generalize this resolution, then we shall ftrst arrive
at something like the following principle:

(PI) Given that 'At. A2, , An: therefore B' is a valid argument: then infer
that B only if, for some time t, all the Ais hold at t.

(P1) may be an acceptable principle; but it is less than satisfactory. For it is

a negative principle which offers only a necessary condition-'only if ... '
-for the inference. It tells us, in effect, when we may not draw a
conclusion: it does not tell us when we may draw a conclusion.
Then let the 'only if be strengthened to an 'if and only if, to give the
principle:

(P2) Given that 'At. A2, ... , An: therefore B' is a valid argument: then infer
that B if and only if, for some time t, all the Ais hold at t.

But (P2) is unacceptable-it licenses inferences which we shall want to

avoid. Take again my Monday-Tuesday argument, propounded as midnight
chimes. The two premisses of the argument are sometimes both true-
indeed, for most of the week they are both true. Hence there is some time at
which they are both true. Hence, according to (P2), we may infer the
conclusion. But the conclusion will often be false-hence (P2) allows us to
infer false conclusions from true premisses.
But perhaps neither (P1) nor (P2) answers exactly to Sextus' report.
According to Sextus, the logicians hold that 'the conclusion can never
[oUOE1tO'tE] be inferred' because the premisses are not true at the same time;
and (P1) and (P2) pay no attention to the word 'never'. I take it that the
logicians hold that at no time can you infer the conclusion because at no
time are the premisses all true. And this suggests not (P 1) or (P2) but rather
a principle such as:

(P3) Given that 'At. A2 , ... , A0 : therefore B' is a valid argument: then infer at
t that B if and only if all the Ais hold at t.

This principle is superior to (P2). Whereas (P2) requires only that the
premisses all be true at some one time, (P3) requires that they all be true at
the time of the inference. Hence (P3) deals adequately with my Monday-
Tuesday example. And we might think to refme the principle a little further
in order to leave open the possibility of an argument's changing its validity,
thus:
116 CHAPTER THREE

(P4) Given that 'At> A2 , ... , An: therefore B' is a valid argument at t: then
infer at t that B if and only if all the Ais hold at t.

Unlike (P2), the two principles (P3) and (P4) will never allow us to infer
a false conclusion by way of a valid argument. Suppose that an argument is
valid (or valid at a time t) and that each of its premisses is true at t. Since
the argument is valid at t, then the corresponding conditional- 'If A 1 and
A2 and ... and An, then B'-is true at t. Since each Ai is true at t, the
conjunction 'A1 and A2 and ... and An' will be true at t. Hence at t we have
two truths:

If A 1 and A2 and ... and An, then B

and

and in that case 'B', the conclusion of our inference, is also true at t. Thus
(P3) and (P4) protect us from error.
Perhaps they do so at a price? Perhaps they exclude certain inferences
which we should find acceptable? Examples of at least two different kinds
might be alleged. Consider, first, the following sophisticated case. 319

Socrates is pale.
Anything which is pale at any time in its history is coloured at every time in
its history.

Therefore: Socrates is coloured at every time in his history.

The argument seems to be valid. Suppose that 'Socrates is pale' is true
when we begin to propound the argument but false by the time we draw the
conclusion. (As we argue, Socrates blushes violently.) Then, according to
(P3) or (P4), we may not draw the conclusion. But surely we are warranted
in drawing the conclusion? After all, provided that at some time or other
'Socrates is pale' is true, the conclusion is guaranteed to be true. Other
refmed examples of this sort are easily thought up.
The second type of case is less sophisticated. The two principles outlaw
inferences from false premisses: if any one of the Ais is false at t, then-
according to (P3) and (P4}--I may not infer at t that B. But surely we often,
and legitimately, argue from premisses which are false. Yesterday I said to
myself:

If today is Tuesday, I'm lecturing at 8.00.

319 Which I owe to Ben Morison.

 EPICTETUS 117

Today is Tuesday.

Therefore: I'm lecturing at 8.00.

And I acted on the conclusion. Alas, the second premiss was false-but my
inference, surely, was impeccable?
In this case the falsity of the premiss was unsuspected; but we
sometimes make inferences from premisses which we take to be false or
even know to be false: standard reductio ad impossibile arguments are often
like this. I want to prove that there is no highest prime number. I decide to
proceed by reductio. Thus I start by setting down a premiss which I
strongly suspect to be false, namely

There is a highest prime number.

And I proceed to make inferences on the basis of this premiss.

The refmed example I leave to the reader. As to inferences from false
premisses, I may remark that Frege notoriously maintained-and
frequently-that inferences cannot be drawn from false premisses, or at any
rate cannot be drawn from premisses which are taken to be false. After all,
the point of an inference is to get you from one belief to another-and you
aim to get to true beliefs. That is why a good inference must start from true
premisses. If a logician of Frege's genius could subscribe to such a view,
we need have no qualms in attributing it to Sextus' logicians. Indeed,
something like it has already turned up in Epictetus. 320 What, then, is to be
done with reductio arguments and the like? Frege had an answer to this
question. No doubt we can invent one for Sextus' logicians.
But I must return to Epictetus. What does he say about his changing
argJiments? His remarks, like those of Sextus' logicians, divide into two
parts: there is an exposition of the problem, and then a resolution. First, the
exposition:

Now there are cases in which we have soundly granted the premisses and
such-and-such results from them: although it is false, nonetheless it results.
Then what is it appropriate for me to do? Accept the falsity? How can I? Say:
'I did not soundly concede the premisses I agreed to?' That is not allowed
either. Or: 'It doesn't result by way of what was conceded'? Nor is that
allowed. Then what is to be done in such a case? (I vii 13-16)

I grant, correctly, the two premisses of an argument. The conclusion

follows-and yet it is false. (Epictetus, unlike Sextus' logicians, speaks
expressly of the falsity of the conclusion-he is not concerned with its
cognitive status but with its truth-value.) Then what am I to do? There are

320 Above, p.33.

118 CHAPTER THREE

three straightforward responses-but each of them Epictetus outlaws. First,

I cannot simply grit my teeth and accept the conclusion-for (ex hypothesi)
the conclusion is false. Secondly, I cannot go back on the premisses and
deny that I was ever right to accept them-for (ex hypothesi) I was right to
accept them. Thirdly, I cannot maintain that the argument is invalid, that the
conclusion does not follow from the premisses-for (ex hypothesi) the
argument is valid.
There is the problem. Three possibilities are excluded-and quartum, as
it appears, non datur. But there must be a fourth option. Here is the second
part of the text:

Well, just as the fact that you have borrowed is not enough to show that you
are still in debt-rather, it must be true in addition that you still stand under
the debt and that it has not been dissolved -, so here the fact that we have
granted the premisses is not enough to show that we must concede the
conclusion-we must still stand under the concession of the premisses. Now if
they remain to the end as they were when they were conceded, then it is
absolutely necessary that we still stand under the concession and accept what
follows from the premisses. But if they do not remain so, we need not accept
it. For this conclusion no longer results for us nor does it accord with our
position, since we have renounced the concession of the premisses. ( 16-19)

There is, after all, a fourth option: we may 'renounce' the premisses. That is
to say, while continuing to maintain that we were right to accept the
premisses when we did accept them, we may now refuse to accept them any
longer. Such is Epictetus' solution to the sophism: how good is it?
It has a 'moral' smell to it-a smell produced not only by the putative
parallel with debtors and creditors, but also by the wording of the text.
Epictetus insinuates that he is concerned to tell us what we ought to do,
what morally speaking we ought to do. And after all, the official function of
I vii as a whole is to remind us that logic is concerned with 'what is
appropriate' .321 Logic teaches us how we ought to behave in certain areas-
it teaches us (as a later generation of logicians put it) the laws of thought,
the laws according to which we ought to think. This may seem quaint. It is
quaint. It may seem worse than quaint; for surely the pressing question is
not: How ought I, morally speaking, to react to a changing argument of this
sort? Rather it is: What is the correct way of solving the puzzle raised by
the argument, of resolving the paradox, of disarming the sophism? The
question, in short, is a logical question and not an ethical question. 322

321 Above, pp.61-62.

322 'What ought I to do about the Sorites paradox?'-'Nothing at all: just ignore it'. Good
moral advice, particularly to someone incapable of sustained philosophical reflection. But
hardly a decent logical answer to the question-and when Chrysippus said (or seemed to say)
something like it, Cameades was rightly scornful (Cicero, Luc xxix 93: cf Barnes [1982]).
Why should I not also cite Wittgenstein? 'You might getp.-,p by means ofFrege's system. If
 EPICTETUS 119

Now this is surely true-but it would not have discountenanced

Epictetus; for, or so he evidently holds, what a good man ought, morally
speaking, to do when faced by a logical puzzle is precisely to fmd the
correct logical solution to it. The appropriate behaviour, in the area of logic,
is precisely the behaviour which logic demands.

Will a wise man refuse to engage with this man, and avoid examining the
matter and discussing it with him? Then who else is capable of using his
reason and clever at questioning and answering-and indeed undeceivable
and immune to sophisms? Then will he engage but not care if in matters of
reason he behaves at random and haphazard? How then will he still be the sort
of man we conceive him to be?( 25-27)

The Sage, the Good Man, is a logician-indeed, it was Old Stoic doctrine
that only the Sage is a logician (Diogenes Laertius, VII 83). After all, logic
itself is a virtue (Diogenes Laertius, VII 46). All this may well seem pretty
dubious; and even if you are prepared to swallow it, you may still think that
it imports the wrong sort of consideration into a logical inquiry.
Nonetheless, the result is that-in principle-Epictetus urges us to proceed
in the way in which any decent logician would urge us to proceed.
How, then, does he urge us to proceed? His solution to the puzzle is
advocated by way of an analogy with debt. You lent me 5 a few weeks
ago: do I now owe you 5? Well, the fact that in the past I borrowed 5
from you is not a sufficient condition for my now owing you 5-since I
may, for example, already have paid off the debt, or you may have waived
it, or one of the conditions on which it was made may have failed, .... True
enough-but how does this apply to the changing argument? My originally
accepting the premisses corresponds to my borro~ing the 5. My still being
obliged to admit the premisses corresponds to my still owing you the 5.
My drawing the conclusion corresponds to my paying you the 5 (with
interest?). So-and this is the crucial point of the analogy-just as the fact
that I borrowed from you in the past is not a sufficient condition for my still
being in debt to you, so the fact that I admitted the premisses in the past is
not a sufficient condition for my now being obliged to admit them: after all,
things may in the interim have changed.
And, in the case of the changing argument, things have changed. The
premisses were true when I accepted them. That is what Epictetus means
when he says that I correctly accepted them; and that is why I cannot now
go back on that acceptance and claim that I was wrong then to accept them.
To do so would be to imply something false, namely that the premisses

you can draw any conclusion you like from it, then that, as far as I can see, is all the trouble
you can get into. And I would say, 'Well, then, just don't draw any conclusions from a
contradiction".
120 CHAPTER THREE

were not then true. But the premisses were then true. However, they are no
longer true. And this change-so Epictetus implies-is enough to cancel
the debt. I no longer owe you the conclusion: not because I have already
paid you the conclusion, nor because you no longer insist that I draw it
(these options in any event make little non-metaphorical sense); but because
the circumstances have changed in such a way as to invalidate or unmake
the debt-the premisses have lost their value.
The analogy is charming. But what precisely does it imply? That is to
say, what principle-parallel to principles (Pl) to (P4)-might we extract
from it. Well, Epictetus says that we must repay a debt at t if and only if we
still owe the sum at t. That is to say, we must infer a conclusion at t if and
only if we are obliged to accept the premisses at t; or in other words: the
crucial feature of the premisses is not their truth-value at the time when you
ftrst accepted them but their truth-value at the time of the inference. The
essential point, then, is this: Given that an argument is valid (at t), then at t
draw the conclusion if and only if at t all the premisses are true. And that is
simply an informal way of stating principle (P3)-or perhaps principle (P4).
In short, I incline to ascribe to Epictetus and to Sextus' logicians the
same principle, and hence the same resolution of one of the problems
caused by some changing arguments. And I guess that the principle may
have been a standard piece of imperial Stoic wisdom. I add only that in
Epictetus' text there is no sign of the logical morass into which Sextus'
logicians slipped.

Sextus' argument is evidently conceived of as an event: it is something

which takes time, and its constituent parts-the positing of each premiss,
the drawing of the conclusion-follow each other in a ftxed temporal order.
Epictetus is also evidently thinking of an event-and here of a public event,
of an argument between a pair of interlocutors. You ask me questions. I
respond. You invite me to draw an inference. I accede. The matter takes
time, and its consituent parts, themselves events, are temporally ordered.
Nor does this seem to be an accidental feature of the two discussions; for if
assertibles are to change their truth-value in the course of an argument, then
time must elapse as the argument proceeds- and so the argument, it seems,
must be conceived of as an extended event.
It is sometimes suggested that ancient logic was, in general, conceived
of on a 'dialectical' model: an argument is essentially a sort of debate,
conducted either in foro publico or in foro interno. 323 Debates, and hence
arguments, are complex events. Consider the following passage in which
Sextus argues that arguments do not exist:

323 The classic account is in Kapp [ 1931], pp.l 061-1064.

 EPICTETUS 121

An argument is a compound of assertibles, and compound objects cannot exist

unless the items from which they are composed co-exist with one another (this
is clear from beds and the like). But the parts of an argument do not co-exist
with one another. For when we say the first premiss, neither the second
premiss nor the conclusion yet exists; and when we say the second premiss,
the former premiss no longer exists and the conclusion does not yet exist; and
when we utter the conclusion the premisses no longer subsist. Therefore the
parts of an argument do not co-exist with one another-whence it will be seen
that arguments do not exist. 324

The merits of this reasoning are debatable. Here I cite it for its
presuppositions; for Sextus evidently takes it for granted that arguments are
temporally extended items. And there is no hint that this might be regarded
as a controversial view.
Or again, consider the hoary debate over the correct order of the
premisses in an argument-a debate which, historically speaking,
concentrated on Aristotelian syllogistic and buzzed around the silly issue of
the legitimacy of the 'fourth figure'. 325 The general question is this: Does
the order of the premisses of an argument matter to it, from a logical point
of view?
Ancient texts show some confusion. Thus Apuleius, for example,
remarks that two of the 'additional' syllogisms of the Aristotelian first
figure require a change in the order of the premisses 'so that the first
premiss is negative' (int ix, 205.1-2 Moreschini)-the order of the
premisses, it seems, does make a logical difference. But later Apuleius
rejects the view of Theophrastus, who added a 'second Darapti' to the third
figure, on the grounds that 'it makes no difference which premiss is first
enunciated' (xi, 207.22). And on the second figure Apuleius havers: the
first two syllogisms in it do not differ from one another-except insofar as
'the order of enunciation is varied' (x, 207.2-3): Cesare and Camestres do
not differ from one another--or else perhaps they do. Does the order of the
premisses matter to a Peripatetic syllogism, or is it a thing of indifference?
Apuleius has not made up his mind-and as a result, his account of
categorical syllogistic is muddled and inconsistent.
However that may be, many logicians did regard the order as logically
significant. That is to say, conceiving of a syllogism as a complex event
with temporally ordered parts, they reasonably supposed that an argument
of the form 'AJ, A2: therefore B' was in principle to be distinguished from
an argument with the same component propositions but having the form
'A2, A1: therefore B'; and they further supposed that in principle one such
argument might be valid and the other invalid--or at any rate, that there
might be some logical difference between the two arguments. The modem

324 PH II 144; cfll 109 (on compound assertibles); M VIII 81-84, 135-136.
325 See e.g. Patzig [1968), pp.59-61.
122 CHAPTER THREE

orthodoxy on Aristotelian syllogisms will have none of this: of course, it is

said, the order of the premisses can have no logical importance-'A 1, A2:
therefore B' is valid if and only if' A2, A 1: therefore B' is valid.
But this modem orthodoxy is only a little better than the view it opposes;
for in denying that the order of the premisses in an argument is logically
pertinent, it shares the presupposition of that opposite view-the
presupposition that the premisses of an argument have an order. But the
premisses of an argument have no order at all: they form a set, they do not
form an ordered sequence. To be sure, in expressing the premisses of any
argument, you must express them in one order or another: if you speak
them, you must speak one before another; if you inscribe them, then the
inscriptions will, additionally and trivially, stand in some spatial relation to
one another. Moreover, from a psychological point of view, the order in
which you choose to produce the premisses may make a difference-to
your chances of persuading an interlocutor, for example. But none of this
says anything about the premisses themselves. From the fact that any
uttering of the premisses has a certain temporal order we cannot infer that
the premisses themselves have a temporal order-as well infer from the fact
that uttering is done at a certain pitch and volume that the premisses have a
certain pitch and volume.
Arguments are not in fact temporally extended or temporally ordered
items. Ancient arguments, it seems, were different; or rather, ancient
logicians, it is commonly thought, took a different view of the nature of
arguments. Ancient arguments are datable; and they have temporal careers
-they last, they endure. Their premisses possess an order, a chronological
order. It is sometimes supposed that this marks a general and crucial
difference between ancient and modem logic; for a modem logician does
not take an argument to be an event, and such questions as 'When?' and
'For how long?' do not arise in the timeless realm of his logic. It might
further be suggested that it is precisely the ancient conception of arguments
as events which allows the formulation of certain puzzles-notably the
puzzles raised by changing arguments-which do not and cannot arise
within a modem context. A fmal suggestion: so much the worse for ancient
logic.
In retort to all this it might be observed, first, that the distinction
between the ancients and the modems is at best overblown. After all,
modem logicians and philosophers do still employ a conception of
inference according to which inferences are events. Thus Frege insisted that
an inference is a complex psychological event in which one judgement (the
judgement which expresses the conclusion) is made after, and in
consequence of, a number of other judgements (the judgements which
express the premisses). Fregean judgements are quite explicitly said to be
 EPICTETUS 123

events-they are datable psychological acts. And Fregean inferences are

quite certainly complex events-they are ordered sequences of mental
events. We can intelligently ask 'When did you infer that so-and-so?', just
as we can intelligently ask 'When did you judg.e that so-and-so?'. Nor is
Frege idiosyncratic. These two questions are perfectly ordinary and
intelligible questions.
True enough-but surely inferences, pace Frege, are of no interest to
logic-they are of interest only to biographers and to psychologists. 326
Logic studies the formal relations among assertibles which underwrite any
inferences which may happen to take place, relations which hold
independently of any such mental events. I may judge, at t, that P, and then
infer at t* that Q; and my inference may be valid. The act of judging is an
event, and the act of inferring is an event. But what I judge-the premiss-
is not an event; what I infer-the conclusion-is not an event; and the
relation between the premiss and the conclusion is not an event. Logic is
exclusively concerned with these non-events.
In truth, it matters little whether we allow that a logician may interest
himself in acts of judgement and acts of inference, or whether we prefer a
more austere view which consigns such things to the junk-room of the
psychologist. It matters greatly that we should distinguish between
inferences and arguments. The argument:

If Frege thought that inferences were events, then a great logician thought that
inferences were events.
Frege thought that inferences were events.

Therefore: a great logician thought that inferences were events.

has no date, and its component propositions have no temporal ordering.

Any propounding of that argument-say the propounding which I have just
carried out-has a date; and in any such propounding, some components
will be propounded before others. The argument is one thing, the
propounding another.
Was it really otherwise in antiquity? Did the ancient logicians confuse
arguments and propoundings? Did they falsely think that arguments
actually were, or had to be construed as, events? Well, in presenting and
discussing arguments, the Stoics-like modem logicians-often use the

326 In Frege's idiosyncratic symbolism-his Begriffsschrift-there is a special sign to

mark the fact that a formula is judged true; and all the premisses (and all the conclusions) of
the arguments in Frege's major work, the Grundgesetze, are prefixed by this special sign: all
of them, in other words, purport to record certain datable events in Frege's own mental career.
Most readers of Grundgesetze ignore this curious fact; no reader takes it to be other than
irrelevant to Frege's ends-and mildly embarrassing. But Frege himself, alas, took it
seriously.
124 CHAPTER THREE

language of events, and they often invent an imaginary debate in order to

illustrate a logical point. In particular, they will no doubt often have
presented changing arguments in such dramatic form. But despite this
common and understandable habit, Stoic arguments are quite certainly not
events. Assertibles do indeed have temporal careers; and they may even, in
some cases, come into being and be destroyed. 327 But assertibles are not
events-they do not happen. A Stoic argument is a system or sequence of
assertibles: hence it is not a system or sequence of events; nor therefore is it
a complex event. You cannot decently ask of a Stoic argument, any more
than you can decently ask of a modern argument, when it took place or for
how long it went on.
Let it be added that the premisses of a Stoic argument are not ordered. It
is true that the Stoics sometimes distinguished, in a two-premissed
argument, between a 'governing assumption' or ilYEJlOVtKov 1-:fllllla and a
'further assumption' or 1tp6crA.TJ'lft<;. 328 This terminology implies an
asymmetry, and might perhaps suggest that one of the premisses comes
'after' the other: thus what is further assumed, you might think, must be
assumed after something else. No doubt some philosophers did in fact think
in this way-Sextus, for one. Perhaps the Stoics themselves were not, or not
always, clear on the matter. But it remains true that the Stoic account of the
nature of assertibles and the Stoic definition of an argument demand that
arguments are not events and that therefore the premisses of an argument
stand in no pertinent chronological relation to one another. 329
But if this is so, and arguments are not events, then how can there be
changing arguments? And how, in particular, can an argument change
during the time in which it is propounded? Well, there is no difficulty here
at all-once arguments are distinguished from propoundings. For although
an argument is not an event, the propounding of an argument is an event.
You accept the premisses at a time; you draw the conclusion at a time.
Since the assertibles which compose an argument are true or false not
absolutely but in relation to time, you can always ask whether or not they
are true at the time at which you propound them or infer them.
Consider the following argument:

(1) I have not already propounded a premiss in this argument.

(2) Either I have already propounded a premiss in this argument
or I am in Geneva.

Therefore: I am in Geneva.

327 See above, p.98.

328 See e.g. Diogenes Laertius, VII 76; Barnes eta/ [*1991], p.67 n.93.
329 I may add, by the by, that we should not speak of 'major' and 'minor' premisses in
connection with Stoic arguments: these terms of art have a sense only within Aristotelian
categorical syllogistic.
 EPICTETUS 125

What is the truth-value of premiss (1)? That is to say, what is its truth-value
at t? Well, that depends on t-that is to say, on what happens at t and before
t. Suppose that I express premiss (1) before I express premiss (2): then
premiss (1) is true at the time when I express it. Suppose that I express
premiss (2) first: then premiss (1) is false when I express it.
The argument is a changing argument in that it contains assertibles
which are now true and now false. The truth-value of the premisses at a
time does not depend on their order in the argument; for they have no order
in the argument. The truth-value of the premisses at a time does depend on
the order in which they are propounded; and if they are propounded at all,
then-trivially-they must be propounded in some order or other.
In brief, changing arguments are odd items. But they do not depend on
the view-the muddled view-that arguments are themselves complex
events.
 CHAPTER FOUR

CONCLUSION

Epictetus is a frustrating author-at least for an historian of logic. Like the

Lord whose oracle is at Delphi, he neither states nor hides: he hints.
As far as the technical development of logic is concerned, he hardly
even hints. His texts are-I admit-strictly consistent with the hypothesis
that no late Stoic made any interesting advance in the subject; that Stoic
logic remained the logic which Chrysippus had elaborated and to which
Antipater had put the fmal touches. This hypothesis, I am sure, is false; but
if there is conclusive evidence for its falsity, that evidence is to be looked
for in sources other than Epictetus.
As far as Stoic attitudes to and interest in logic are concerned, Epictetus'
hints are numerous and instructive. His own apparently con-suming interest
in moral precepts and ethical improvement was not characteristic of the
Stoic philosophers of his age. On the contrary, in this respect at least
Epictetus was unorthodox: his contemporaries-Stoic teachers and Stoic
pupils-were obsessed not by ethics but by logic; they gave themselves to
logical matters with a passion, a single-mindedness, and no doubt a
pedantry which galled Epictetus-as it had galled Seneca, and as it has later
galled so many earnest philosophers. Doubtless Epictetus exaggerates.
Nonetheless, it seems to me beyond doubt that logic engrossed men during
this period in the history of philosophy as it has rarely engrossed men in
any other period.
Moreover, the logic which engrossed was not the logic of the elementary
handbooks-that sort of thing, after all, is essentially unengrossing. It was
more subtle, more advanced, more technical. If much of it was based on
exegesis or commentary and hence on books, it was none the worse for that;
for the books included some tough texts-Chrysippus on the Master
Argument was surely no tiro's read. The standard sophisms and
conundrums fascinated Epictetus' contemporaries; and if an interest in such
things is compatible with intellectual frivolity, it may equally show logical
sophistication and seriousness. Moreover, problems of syllogistic analysis
and questions about changing arguments and hypothetical arguments 'and
the like' do not appeal to the logical butterfly.
Sextus Empiricus presents a sceptical attack on Dogmatic philosophy. In
both versions of the attack, in the Outlines of Pyrrhonism and also in
 CONCLUSION 127

Against the Mathematicians VII-XI, he busies himself with all three parts of
Dogmatic philosophy: with logic, with physics, with ethics. In PH logic
gets a book to itself; physics has two thirds of a book; and ethics is tucked
in at the end. In M, Books VII and VIII deal with logic, Books IX and X
deal with physics, and Book XI deals with ethics. The imbalance is striking
-ethics comes a distant third. 1 Not only that: it has frequently been
observed that the content of the ethical sections is 'disappointing'.
Compared to the sections on physics and logic, the sections on ethics are
jejune. The historical material is astoundingly thin, the range of subject
matter is astoundingly limited, the arguments are astoundingly bad-silly or
repetitious or merely dull.
Why should this be so? Perhaps it is a reflection of the Pyrrhonists' own
balance of interests--or of Sextus' personal concerns. After all, you will
expect a sceptic to look harder at logic-which includes what we call
epistemology-than at any other part or aspect of philosophy; and you
might imagine that a sceptic would walk with a gingerly tread along the
paths of ethics-it does not do to preach moral scepticism, or at least it does
not do to preach moral scepticism openly and at length and in practical
detail. Perhaps the imbalance rather reflects a phenomenon common to
many authors and many works of philosophy: as you approach the end of a
long inquiry, you tend to hurry up. Ethics comes at the end, both in PH and
in M, and Sextus no doubt felt weary, his pen-hand cramped. Perhaps,
again, the imbalance reflects the interests of a past age of Dogmatists-the
interests of the Dogmatists whom Sextus' sources confronted rather than the
interests of the Dogmatists who were Sextus' own contemporaries.
Perhaps so. But there is, after all, a far simpler hypothesis: perhaps the
imbalance in Sextus' treatment of the parts of Dogmatic philosophy reflects
the philosophical interests of his age. Ethics was, no doubt, a matter of
moment in practical life, and a matter of concern to preachers and
moralizers. But it was no more than a relatively small, and a relatively
unexciting, part of the great business of philosophy.
However that may be, I fmd it a curiously pleasing reflection that, while
old Epictetus solenmly moralized at his pupils in Nicopolis, the common-
rooms of the cultivated world stank of logic.

1 In PH logic gets 259 sections, physics 167, ethics 112; in M logic gets 934 sections,
physics 791, ethics 257. It is true that the figures given here for logic are actually figures for
i..oyuci], and that the figures for logic itself would be substantially lower. It is also true that
the figures for ethics are inflated inasmuch as a substantial portion of each treatment is
devoted to a discussion of the possibility of teaching and learning.
 APPENDIX: Epictetus, Discourses I vii

I vii is the most important Epictetan discourse to bear upon logic. Several
parts of it have been discussed in the body of this monograph, some of them
at length. Nonetheless, given the significance of the text and the fact that
most translators and commentators have misunderstood some of its crucial
sentences and phrases, it seems worthwhile to set out the whole thing as an
appendix.
The text largely follows Schenkl: all differences, save differences of
punctuation, are explicitly noticed in the apparatus criticus (which itself is
wholly derivative). The translation aims at fidelity to the thought of
Epictetus rather than to his style. The commentary is not comprehensive:
rather, it is designed to supplement what I have already said in earlier parts
of the monograph.
 130 APPENDIX

IlEpt tfj~ XPta~troY JlEta1tt1ttOVtCOV Kat

imo8nK&v Kat t&v 6Jloi.cov
[1] iJ 1tpt tOU~ JlEta1tt1ttOVta~ Kat imo8nKOU~ En oE: tCj)
~pcotfjcr8at 1tpai.vovta~ Kat 1tavta~ U1tA&~ toil~ totoutou~ A6you~
5 1tpayJlatda Aaveavt tOU~ 1t0AAOU~ 1tpt Ka8i]KOVtO~ oucra. [2]
STJtOUJlEV yap E1tt 1tUcrTJ~ UATJ~ 1t&~ &.v EUpot 6 KaAO~ Kai ayaeo~ t~V
OtE~OOOV Kai avacrtpocpi]v ti]v EV aut'fi Ka8i]Koucrav. [3] OUKOUV ft
tOUtO AYEtcocrav, on ou cruyKa8i]crt Ei~ EpcOtTJcrtV Kai 6.1t6Kptcrtv 6
cr1touoaio~, ft on cruyKa8i~ ouK E1ttJlEAi]cretat tou JlTJ EiK'fi JlTJO' ro~
10 Etuxev ev epcoti]crn Kat a1toKptcrt 6.vacrtpcpecr8at. [4] Jl T] toutcov
Jl TJOEtEpov 1tpocrOXOJlEVOt~ avayKatOV OJlOAOYEtV on E1ttcrK\jlt v n va
1tOtTJtEOV t&v t01t(I)V tOUtCOV 1tpt ou~ JlUAtcrta crtpcptat EpcOtTJcrt~ Kai
U1t0Kptcrt~;
[5] ti. yap 1tayyUetat tv A6ycp; - t6.ATJ8fl nevat, ta \j/Euofj
15 atpEtV, 1tp0~ ta UOTJAa E1tEXtV. - [6] &p' ouv apKEt tOUtO JlOVOV
Jla8tv; - apKEt, cpTJcriv. - OUKOUV Kai tCj) ~OUAOJlEVql EV xpi]crt
VOJltcrJlato~ JlTJ Ota1tt1tttV apKEt tOUtO UKOUcrat, Ot6n ta~ JlEV
OOKtJlOU~ OpaxJla~ 1tapaoxou, ta~ o' UOOKtJlOU~ U1tOOOKtJlUS; - OUK
apKEl. - [7] ti. ouv 0t tOUtql 1tpocrAa~tV; - ti. yap UAAO ft OUVaJltV
20 OOKtJlacrn Ki]v t Kai OtaKpttt KTJV t&v OOKtJlCOV t Kai aOOKtJlCOV
opaxJl&v; - [8] OUKOUV Kai E1tl A6you OUK apKet to AX8v, UAA'
avayKTJ OOKtJlacrnKOV yevcr8at Kai OtaKptnKOV tOU UATJ80U~ Kai tOU
\j/EUOOU~ Kai tOU aOi]Aou; - avayKTJ. - [9] E1tl tOUtOt~ ti.
1tapayyEAAtat EV A6ycp; - to UKOAOU8ov tot~ Oo8lcrt v U1t0 crou
25 KaAro~ 1tapa0EXOU. - [10] ayE apKet OUV KUVtau8a yv&vat tOUtO; -
OUK apKet. - od oE: Jla8tV 1tro~ ti. nat v UKOAOU8ov yi. Vtat Kai 1tOtE
JlEV EV E:vi UKOAOU8t, 1tOtE oE: 1tAtocrtv KOt v'fi; [ 11] Jl, 1t0t ouv Kai
tOUtO avayq 1tpOcrAa~tV tOY JlEAAOVta EV AOYQl cruvet&~
avacrtpacpi]crecreat Kai aut6v t' U1t00t~t v EKacrta U1t000Vta Kai

3 tCj) (vel EK tou, codd recc)] non exhibet S, om Schenkl :: 6 eupot

(Meibom; Schenkl)] eupoot codd :: 10 Jlil (codd)] ij Schenkl, alii alia :: 15
1tp6~ add Meibom, Schenkl :: 17 Ot6n (vel on, codd recc)] Ota ti. S,
Schenkl :: 18 1tapaoxou ... 6.1toOoKtJlal;e (S supra lin)] 1tapaoxn ...
a1toOoKtJlasn~ S, Schenk} :: 19 toutcp (codd, Schenkl)] toutov
Schweighauser :: 26-27 1tOtE ... 1tOtE ... (Schenkl)] 1tOt ... 1tOt ... codd
:: 29 6.1toOei.~nv EKacrta (codd, Schenkl)] a1tOOt~tv (ed pr) EKacrtcov
Meibom, alii alia
 APPENDIX 131

On the Utility of Changing Arguments, Hypothetical

Arguments, and the lik((

[1] Most men are unaware that the study of changing arguments and
hypothetical arguments-and also of arguments which conclude by way of
questioning, and in general all arguments of this kind-is concerned with
what it is appropriate to do. [2] For in any matter we ask how a good man
may discover the conduct and behaviour which is appropriate to that
matter. [3] Hence they must say either that a virtuous man will not stoop to
questioning and answering or that when he does so he will not care if in
questioning and answering he behaves at random and haphazard. [4] If they
accept neither of these options, then surely they must agree that some
consideration should be given to those areas with which questioning and
answering are especially concerned?
[5] What, then, does reason counsel?-'Affmn the true, reject the false,
suspend judgement over the unclear.'- [6] Then is it enough to have
learned this and nothing more?-'It is enough', he says.-Then if you want
to avoid mistakes when it comes to using coins, is it enough to have heard
this: 'Accept genuine drachmas and reject counterfeit ones?'-It is not
enough.- [7] Then what do you need to acquire in addition to
this?-Surely a capacity for testing and discriminating between genuine and
counterfeit drachmas?- [8] Then what we said is not enough in the case of
reason either: rather, it is necessary to become capable of testing and
discriminating among the true and the false and the unclear.-It is
necessary.- [9] What in addition to this does reason enjoin?-'Accept
what follows from what you have correctly granted.'- [10] Then is it
enough here too to know this?-It is not enough.-But you must learn how
one thing follows others-how one thing sometimes follows one thing and
sometimes several jointly? [11] Then surely this too must be acquired in
addition by anyone who is going to behave intelligently in matters of
reason-if he is going both to defme and prove things himself and to
 132 APPENDIX

'tote; Cl1t00Et1CVUOUCH 1ta.pa.KoA.ou9i)OEtv 11110' um) 'tOOV OO<pt~OI1EV(J)V

ota.1tA.a.v119'1loEo9a.t roc; (UtoOEtKvu6v'trov; [12] ouKouv A.i)A.u9Ev t111tv
1tEpi 't&v ouva.y6v1:rov Myrov 1ea.l. 1:p61trov 1tpa.y11a.'tEia. 1ea.l. yu11 va.o\a.,
lCO.t ava.yKa.ta. 1tE<p11VEV;
5 [13] aA.A.a oi) tonv E<p' rov OE0c01C0.11EV uyt&c; 'ta A.ili111Cl't0. lCO.t
ou11Pa.\ VEt 'tOU'tt 1; a.u1:&v 'lfEUOoc; o ov ouov fl't'tOV ou11Pa.iVEt. [ 14]
'tt ouv 110t Ka.9i)KEt 1totdv; 1tpooOEXE09a.t 'tO 'lfEUooc;; Kat 1t&c; ot6v 'tE;
[15] aA.A.a AtyEtV on OUX uyt&c; 1tO.PEXcOP1100. 'ta cDI10AOY1111EVa.; Kat
11TJV OUOE 'tOU'tO OtOO'ta.t. aA.A.' on ou Ot>11Pa.ivEt Ota 'tOOV
1o 1ta.pa.1CExrop1111tvrov; aA.A.' ouo 1:ou1:o ot.oo'ta.t. [16] 1:1 ouv 1tl. 'tou'trov
1tot11'tEOV; 1111i)1tO'tE roc; OU'IC aplCEt 'tO oa.vEioa.o9a.t 1tpoc; 'tO En O<pEtAEt v
ana OEt 1tpOOEtVa.t lCO.t 'tO E1tt11EVEt v E1tt 'tOU oa.vEiou Kat 11,
OtO.AEAuo9a.t O.U'tO, OU'troc; OU'IC aplCEt 1tpoc; 'tO OEtV 1ta.pa.xropEtV 'tO
1ttcpEpo11 Evov 1:0 oEoro1etvm 1:a A-i11111a.'ta., oEt o' 1tt11tvEtv 1tt. 1:f\c;
15 1ta.pa.xropi)oEroc; a.u't&v; [ 17] Kat oi) 11EVOV'trov 11EV O.U'tOOV Eic; 'tEAO<;
01tOta. 1tO.pExropil911, 1tUOa. av6.yK11 tll1<ic; E1tt 'tf\c; 1ta.pa.xropi)oEroc;
E1tt11EVEt v 1ea.l. 'tO aK6A.ou9ov a.u'totc; 1tpoootxm9a.t <[ 18] 11i) 11Ev6v'trov
o, ou OEt 1tpoooxEo9m> [19] ouo yap t111t:v en ouo Ka.9' t111ac;
ou11Pa.\ vEt 'tou'to 'to E1tt<pEP011Evov, E1tEtOTJ 'tf\c; ounropi)oEroc; 1:&v
20 A-1111116.1:rov a1tEO't1111Ev. [20] OEt ouv 1ea.l. 1:a 'tota.u'ta. 1:&v A-1111116.1:rov
io'topf\oa.t 1ea.l. 'tTJV 'tota.U't11V 11E'ta.poA.i)v 'tE 1ea.l. 11E't6.1t'trootv a.u't&v v
O.U'tTI 'tTI pro'ti)OEt i1 'tft a1t01CptoEt i1 'tql OUAAEAoyio9a.t , n Vt aAA.cp
'tOtoU'tq> A.a.11P6.vov'ta. 'tac; 11E't0.1t'tcOoEtc; acpop11 i)v 1tO.PEXEt 'tote; avoi)'totc;
'tOU 'ta.p6.cmEo9a.t PA.E1t0Uot 'tO aK6A.ou9ov. [21] 'ttvoc; EVE '!CO.; tv' EV 'tql
25 1:61tq> 'tOU'tCfl 11il 1ta.pa 1:0 Ka.9f\Kov 11110' EiKft 11110 ouyKEXWEvroc;
ava.o'tpE<pcD11E9a..
[22] 1ea.l. 'to a.ino 1ti 'tE 1:&v u1to9oErov 1ea.l. 1:&v i>1to9EnK&v Myrov
ava.yKa.tov yap ton v O't' a.i 'tf\oa.i n va. U1t69Eot v c001tEp mpa9pa.v 'tql
1;f\c; Mycp. [23] 1tlioa.v ouv 'tTJV oo9Etoa.v 1ta.pa.xrop111:Eov i1 ou 1tlioa.v;
30 lCO.t Ei OU 1tUOO.V, 'ttva.; [24] 1ta.pa.xropi)oa.vn OE 11EVE'tEOV

2 t111t:v] + t1 codd recc :: 17-18 < ... > (supplevi)] lacunam indicat Schenkl,
alii alia :: 22 ouA.A.EA.oyicr9a.t (S, Schenkl)] an ouAA.oyi~Eo9a.t? :: 23 post
'tOtoU'tq> obelum statuit Schenkl, alii alia :: 24 'ta.p6.oom9a.t] + 11'11 codd,
Schenkl :: 30 'ttva.;] + 1tEpi 1:\.voc; t1 OKE'Iftc;; - 1tEpt Ka.9i)Kov'toc; codd, del
Wolf, Schenk!
 APPENDIX 133

follow others when they offer proofs, and also not to be misled by those
who produce sophisms as though they were proving things. [12] And has
there not arisen among us study and training in valid arguments and
schemata, and has it not proved necessary?
[13] Now there are cases in which we have soundly granted the
premisses and such-and-such results from them: although it is false,
nonetheless it results. [14] Then what is it appropriate for me to do? Accept
the falsity? How can I? [15] Say: 'I did not soundly concede the premisses I
agreed to?' That is not allowed either. Or: 'It doesn't result by way of what
was conceded'? Nor is that allowed. [16] Then what is to be done in such a
case? Well, perhaps just as the fact that you have borrowed is not enough
for you still to be in debt-rather, it must be true in addition that you still
stand under the debt and that it has not been discharged- , so here the fact
that we have granted the premisses is not enough for us to have to concede
the conclusion-we must still stand under the concession of the premisses.
[17] Now if they remain to the end as they were when they were conceded,
then it is absolutely necessary that we still stand under the concession and
accept what follows from the premisses. [ 18] But if they do not remain so,
we need not accept it. [19] For this conclusion no longer results for us nor
does it accord with our position, since we have renounced the concession of
the premisses. [20] So we must also consider premisses of this sort, and
their alterations and changes as a result of which, in the very process of
questioning or answering or inferring or whatever, they undergo changes
and give the foolish an opportunity to get in a state when they see what
follows. [21] Why? So that we do not behave in this area in a way which is
inappropriate or random or confused.
[22] It is the same with hypotheses and hypothetical arguments. It is
sometimes necessary to postulate some hypothesis as a sort of base for the
argument which follows. [23] Then should we concede all hypotheses we
are given or not all? And if not all, which? [24] And once we have
 134 APPENDIX

de; &1tav E1tt tfic; 1tapaxrop1loeroc; 11 EO'ttv O'tE Ct1tOO'tatov; 'tCt o'
CtK6A.ou8a 1tpOOOEK'tEOV Kat 'tCt f.LaXOf.LEVa OU 1tpOOOEK'tEov; - vai. -
[25] aA.A.a AEYEt nc; O'tt 1tOt1loro OE ouvatou OESCtf.LEVOV {)1t6eEOtV E1t'
CtOUVa'tOV Ct1taX8fivat. 1tpoc; 'tOU'tOV OU OUyKa81lOEt 6 cppOVtf.LOc;, aUa
5 cpEUSE'tat tsttaotv Kat KotvoA.oy\.av; [26] Kat ti.c; en &.Uoc; EO'tt Myq>
XPTJO'ttKoc; Kat OEt voc; EV tpro'tllOEt Kat Ct1tOKpi.OEt Kat V'll ~\.a
CtVESa1tU'tTJ't6c; 'tE Kat ao6cptO'toc;; [27] aA.A&. ouyKa81loEt f.LEV, OUK
E1ttO'tpacp1loE'tat OE 'tOU f.LTJ EtlCTI Kat roc; E'tUXEV avaotpEcpE08at EV
A6yq>; Kat 1t&c; E'tt EO'tat 'tOtOUtoc; otov aU'tOV E1ttVOOUf.LEV; [28] aA.A.'
10 UVEU 'tt voc; 'tOtaU'tTJc; YUf.L vao\.ac; Kat 1tapaOKEUfjc; cpuA.a't'tEt v ot6c; t'
tmt to E:sfic;; [29] touto ot::tKvutrooav Kat 1tapA.Kt tO. 8rop1lf.Lata
'taU'ta 1tUVta U't01ta yap ~v Kat avaK6A.ouea 'tTI 1tp0All\jfEt 'tOU
o1touoa\.ou.
[30] ti. E'tt apyot Kat (l<t8Uf!Ot Kat vroepo\. EOf.LEV Kat 1tpocpaotc;
15 ~TJ'tOUf.LEV Ka8' &c; ou 1t0Vll00f.LEV ouo' aypU1tVll00JlEV ESEpya~OJlEVOt 'tOY
aut&v Myov; - [31] &v ouv EV 'tOU'tOtc; 1tAaVTJ8&, Jl, 'tt 'tOY 1tatpa
Ct1tEK'tEtVa; - avopa1tOOOV, 1tOU yap tveaoE 1tat1)p ~v tv' aU'tOV
a1toKtEivnc;; - ti. ouv E1tOi.TJoa; - o J.L6vov ~v Kata tov t61tov
UJlUp'tTJJla, 'tOU'tO ilf.LUP'tTJKac;. [32] E1tEi 'tOt 'tOUt' auto Kat tyro 'Poucpq>
20 d1tov tm ttf.L&vt\. ]lOt on to 1tapaA.t1tOJlEVOV v tv ouAA.oytof.Lc'i'> n vt
oux EUptOKOV. oux otov d, cpTJf.Ll, 'tO KamtroAtOV KCt'tEKauoa 6 o',
avopa1tOOOV, EcpTJ, tveaoE 'tO 1tapaA.Et1tOJlEVOV KamtroA.t6v EO'ttV. [33] ft
tauta Jlova aJlaptilf.Lata ton to Ka1tttroA.wv Ef.L1tpfiom Kat tov
1tatpa Ct1tOK'tEtVat, 'tO o' EiKT\ Kat f.LU'tTJV Kat roc; E'tUXEV xpfio8at 'tate;
25 cpavtaoiatc; tate; autou Kat f.LTJ 1tapaKoA.ou8Eiv Myq> JlTJO' a1tooEist::t
f.LTJOE oocpi.Ojla'tt, JlTJO' a1tA.&c; ~AE1tEtV 'tO Ka8' autov Kat ou Ka8' au'tOV
EV EprotiloEt Kat Ct1tOKpiOEt, 'tOU't(I)V o' ouotv EO'ttv Uf.LUP'tTJJla;

1 1tapaxropfloroc; (Reiske)] 'tTJpfloroc; codd, Schenk! :: 6 tv (add

Meibom)] E1t' add Richards :: 8 EiKT\] +Kat cptA.Ooocpov Sin margine :: 12
1tavta &.to1ta yap (Richards)] 1tavta, &.to1ta codd, Schenk!, alii alia:: 16
aut&v (Salmasius, Schenk!)] aut6v codd :: 18 E1t0lTJOa] E1tOlTJOac; codd,
Schenk! :: 21 oux otov d, cpTJf.Li (Elter)] oux oiov JlEV, cpTJf.L\. S; f.LTJ yap,
EcpTJV codd recc; oux otov f.LEV, cpTJf.Li, Ei Blass, Schenk! :: KatKauoa
(Schenk!)] Ka'tEOKEuaoa S, Ka'tEOKa\jfa vel EVE1tpTJoa codd recc
 APPENDIX 135

conceded a hypothesis, should we stand by the concession come what may,

or should we sometimes renounce it? Should we accept what follows from
it and not accept what conflicts?-Yes.- [25] But someone says: 'I'll
bring it about that, having accepted a hypothesis of something possible,
you are led on to an impossibility.' Will a wise man refuse to stoop to this
man, and avoid examining the matter and discussing it with him? [26] Then
who else is capable of using his reason and clever at questioning and
answering-and indeed undeceivable and immune to sophisms? [27] Then
will he engage but not care if in matters of reason he behaves at random
and haphazard? How then will he still be the sort of man we conceive him
to be? [28] But will he be able to preserve coherence without some training
and preparation of this sort? [29] Let them prove this and all these theorems
are superfluous; for after all they are out of place and do not fit our
preconception of what a virtuous man is.
[30] Why are we still lazy and idle and dull? why do we look for
excuses for not exercising-or even staying awake-and working on our
reason?- [31] Well, ifl go wrong in these matters, it's not as if I've killed
my father.-Slave, your father wasn't there for you to kilL-So what did I
do?-You made the only mistake which you could make in this area. [32]
Actually I myself said the very same thing to Rufus when he criticized me
because I hadn't found the one omitted item in some syllogism. 'Well,' I
said, 'it's not as if I've burned the Capitol down.' 'Slave,' he said, 'here the
omitted item is the Capitol.' [33] Are burning the Capitol and killing your
father the only mistakes? Using your own impressions at random and in
vain and at haphazard, not following an argument or a proof or a sophism,
in general not seeing, in questioning and answering, what accords with
your position and what does not-is none of this a mistake?
136 APPENDIX

COMMENTARY

Title: IIEpt tile; XPEia.c;] 'XpEia' normally means 'use' in the sense of
'utility', not in the sense of 'usage' (which is 'xpilcrtc;': so, e.g., Aspasius, in
EN 96.27-28); and so here (pace Long, Oldfather, Souilhe, Laurenti,
Carter/Hard)-for the question at issue is not 'How do we use such
arguments?' but ~y should we use them?'.

JlE'ta1tt1tt6vtrov] The noun to be supplied or understood is 'A6yrov' (see 1);

hence not 'premisses' (Oldfather-cf Carter/Hard) but 'arguments'. The
participle means 'changing'; and that is how best to translate it. Certainly not
'sophistical' (Long) or 'equivocal' (Oldfather, Souilhe, Laurenti,
Carter/Hard). The mistaken translations in this case betoken a failure to
understand what a changing argument is; and the failure derives ultimately
from Schweighliuser. His enormous note on the phrase ([* 1799], II pp.94-
102) runs through the various solutions which his predecessors had proposed
for the difficulties which they found; and it ends with the lame assertion that
changing arguments are arguments in which the sense of the premisses
changes. There is nothing to be said for any of the early suggestions; and the
only interesting questions which they raise concern their authors.

1: n OE tC\l T]protilcr9at 1tEpaivovtac;] These items are mentioned again at

III ii 6 and xxi 10. At III xxi 10, all the MSS offer a definite article before
'T]protilcr9at'; at III ii 6 none do. I cannot construe the phrase without an
article (for a naked infinitive cannot depend on '1tEpaivovtac;'); hence I print
't<il' here, from the apographs-and Oldfather is right to add it at III ii 6. But
that is a minor problem. The chief question concerns the sense of the phrase.
In none of the three passages does Epictetus offer any gloss on it; and so far
as I know neither the phrase itself nor anything like it occurs in any other
ancient text on logic. What, then, does it mean?
In a certain sense all arguments 'conclude by way of questioning':
premisses are frequently represented as 'questions' (or, better, as responses to
real or imaginary questions). In Academic dialectical jousting one man asked
another a question (a question which could be answered by 'Yes' or 'No');
and his partner then argued from the question, that is to say from his answer
to the question. Hence 'eprotfiv', 'question', comes to mean 'initiate an
argument', 'propound an argument', 'argue'; and it is used very frequently in
this sense in ancient texts on logic. (In Latin the word comes out as
'interrogare'.) But here Epictetus is plainly not referring to arguments in all
their generality: he is inviting us to consider a particular class of arguments, a
class whose members are distinct from but have some affmities with
hypothetical arguments and changing arguments.
 APPENDIX 137

Schweighiiuser [* 1799], II pp.l 03-104, guessed that Epictetus refers to

the Sorites: nothing tells for the suggestion, and something against it-we do
not expect a particular sophism but a class of arguments. Hadot [1978], p.69
n.15, suggests that Epictetus is thinking of dialectical arguments, i.e. of
arguments which take place between two interlocutors, as in Aristotle's
Topics. But in I vii Epictetus plainly envisages changing arguments and
hypothetical arguments as taking place within a dialectical context; and these
arguments are set on a par with 'arguments which conclude by way of
questioning' and are not special cases of such arguments. Hiilser [*1987],
pp.1518-1519, translates 'die Entwicklung giiltiger Argumente': I do not see
how anything like this can be twisted out of the Greek-and I have not the
remotest idea what it might mean. Atherton [1993], pp.433-434, compares
Diogenes Laertius, VII 44, where the word '1tEpa\.vovtE~' occurs in a list of
sophisms; and she glosses the phrase 'tq) ilprotfjcr9at 1tEpa\.vovta~' as 'those
<arguments> one or more of whose constituent propositions change(s) truth-
value as the argument is being posed'. This is at least the sort of thing which
is wanted. But I doubt if the passage in Diogenes Laertius is pertinent (it
lacks the crucial 'tq) ilprotfjcr9at')--and in any case the text is quite certainly
corrupt. Moreover, on this suggestion arguments which conclude by way of
questioning become difficult to distinguish from changing arguments. I have
been tempted to take the word 'ilprotfjcr9at' in its literal sense, and to
imagine that Epictetus alludes to arguments in which the premisses are
genuine questions. We know that Chrysippus did some work on imperative
logic (see Barnes [1984]), and we know that he had some interest in
questions (note a TIEpt prot-rlcrEro~: Diogenes Laertius, VII 191): hence he
could have turned his hand to erotetic logic. But I am reluctant to believe that
the Greek phrase can be paraphrased by 'valid arguments which contain
questions among their premisses'.
The phrase surely ought to refer to arguments which are valid in virtue of
the fact that they have been propounded. How could there be-or even seem
to be-any such arguments? Well, there are assertibles which are true in
virtue of the fact that they are expressed: consider the assertible expressed by
the (stock) sentence 'I am conversing [ycb OtaA.EyoJ.Lat]'. And so an
argument will be valid in virtue of the fact that it has been propounded if and
only if its corresponding conditional is true in virtue of the fact that it has
been expressed (see above, pp.l02-103). It is not difficult to dream up
examples of this sort of thing. Consider the argument:
138 APPENDIX

If any utterance truly says that it is day, then it is day.

This utterance truly says that it is day.

Therefore: it is day

If this argument is not propounded, then its second premiss will not subsist,
since 'This utterance' will refer to nothing (see above, p.98). Hence, or so we
may suppose, the corresponding conditional will not subsist, and a fortiori
will not be true. Thus the corresponding conditional is true only if it is
expressed, so that the argument will be valid only if it is propounded. It is
easy to invent other parallel arguments. But though I suspect that this
suggestion is along the right lines, it is no doubt profitless to develop it
further.

tou~ 'tOtOU'tOU~ A.6you~) Of what kind? (And what are 'the like' in the title?)
Clearly, Epictetus must be thinking of types of arguments which pose the
same sort of difficulties as those which he is going to discuss explicitly. But
nothing encourages me to guess what these things might have been.

1tpay~atei.a] The word sometimes means, concretely, 'treatise' (so Hiilser

[* 1987], p.1517); but equally often it has the abstract sense of 'treatment',
'study'-and so, plainly, here. (As in 12, the only other occurrence of the
word in Epictetus.)

2: E1tt micr11~ UATl~] For this use of 'uA.11' see e.g. I iv 20; vi 34; xv 2; xxvi
2; xxix 41.

3: ouJCouv ft 'tOU'tO A.qttrocrav] I.e. 'those people who do not realize that
the study of changing arguments and the like is concerned with 'tO Ka9fjKov
must accept one of two (unpalatable) options'. Why the options are
unpalatable is indicated later, in 25-27.

epcO'tTlO"tV Kat Cl1t01Cptcrtv) On 'questions' see above, p.136. The phrase

'questions and answers' is standardly used to distinguish dialectic from
rhetoric-e.g. Alexander, in Top 5.7-9.

eixft) See e.g. I xxviii 29-30; II xii 22; III xv 7; PHerc 1020, ii l-20 [=
Hiilser [* 1987), frag 88). The adverb is standardly used to characterize the
way in which a philosopher should not speak or act: e.g. Aristotle, Met
984b17; Alexander, in APr 3.24-30.
 APPENDIX 139

3-4: ava.cn:pcpecrem. Jlll toutrov] Schenk! (who emends the text at the
beginning of 4) places a conuna after ava.crtpE<pE0"9at and a full stop at the
end of 4. The MS text can, I think, be conserved if 4 is punctuated as a
question. (The use of 'Jlil' to introduce a question is conunon enough in
Epictetus: e.g. I vii 31; II xxi 9; III xv 14.) In any case, the general sense of
the paragraph is not in doubt.

4: t&v t61trov toutrov] For this use of 't61to~ see above, pp.34-35.
Laurenti takes 'toutrov' to refer back to the arguments of 1. In that case we
must give a strong sense to the verb 'crtpE<pE'ta.t'-'struggle with'; for it
would be odd to say that logic is especially concerned with those rather
esoteric areas. But the verb most probably has its usual weak sense ('engage
with'); and so we should construe 'toutrov' as the redundant antecedent of
'ou~': the areas in question are (i) the judging of the truth of statements ( 5),
and (ii) the assessment of the validity of inferences( 9).

JlaA.tcrta. crtpE<pE'tat] The superlative adverb introduces, here as in many

texts, an ambiguity: 'logic deals with these areas more than it deals with any
other areas', or 'logic, more than any other discipline, deals with these areas'.
Presumably the latter interpretation is intended-other disciplines
(psychology, rhetoric, ... )at least touch on the areas in question.

5: 'tt yap E1ta.yyA.A.eta.t tv A.Oyq>;] The sentence is doubly uncertain in

sense. First, the verb '1ta.yyA.A.etat' may be taken in more than one way: it
might mean something like 'promise' (e.g. I iv 3; xv 2) and so be connected
with the sort of E1ta.yyeA.\.a. which is tantamount to meaning (above, p.29,
n.29); or it might mean something like 'advise', 'reconunend'. The former
view is intrinsically attractive; but the parallel occurrence in 9 of '1ta.p-
a.yyA.A.eta.t', which presumably means 'conunand' or 'enjoin' (cf e.g. II ix
13), tells in favour of the second view. Secondly, the phrase 'tv A.Oyq>' is
irritatingly vague, and the version I liave given is only one of several options.
(See Schweighauser [*1799], II p.l06.) But here, too, the general sense is not
in doubt.

'tUA119fl nevm ... E1tE:XEtV] I take the infmitives to have imperatival force:
note the imperatives in 9. For the sentiment see e.g. I xvii 1; xxviii 2.

6: Ot6n . . . a1t0001CtJ16.~E] S offers two readings here. The version which

Schenkl adopts gives 'have heard why you accept ... '. This is unclear in
itself, and it provides a poor parallel to the logical case. The other version
makes everything intelligible. The point which Epictetus is making is this: it
is not enough to know that it is imprudent to accept counterfeit coins-you
140 APPENDIX

must also be able to recognize counterfeits when they are passed to you. For
the grammar see e.g. Iii 18 ('on' introducing an imperative in oratio recta)
with e.g. I xi 5 ('~h6n' in the sense of 'on').

SpaxJ.uxc;] For the analogy see e.g. I xx 8-10; II iii (above, p.83).

7: toiltq> 1tpocrA.aptv] For the dative see e.g. Musonius, frag I [p.2.13-14
Hense]= Stobaeus, eel II xxxi 125.

SoKtJ..LacrnKijv tE Kat StaKptnKijv] For the combination of terms see e.g. I

xx 7; for logic as a SuvaJ..Lt<; SoKtJ..LacrnKij see I i 1; for 'StaKptttx6c;' see
e.g. I xvii 10.

8: to AEXStv] Not 'the spoken word' (Oldfather-cf Carter/Hard).

Epictetus means: it is not enough to know that you should accept what is true
and reject what is false-you must also be able to recognize truths and
falsities when they approach you.

SoKtJ..LacrnKov yEvtcrem] Epictetus alludes to the standard Stoic defmition

of logic-see e.g. Diogenes Laertius, VII 62 (with reference to Posidonius);
Sextus, PH II 94; M XI 187. That is to say: you must acquire the capacity to
discriminate among the true, the false, and the unclear; this capacity is in fact
logic: therefore you must acquire logic.

10: 1tOtE J..LEV ... , 1tOtE S ... ] Oldfather prints '1tott ... 1tOt ... ' and
translates '1t6tE ... 1tOtE ... '. - Epictetus means that you must learn in what
cases a conclusion follows from a single premiss and in what cases a
conclusion follows from a set of premisses. He does not indicate that there is
anything untoward or exciting about this. But ancient logicians-including
Chrysippus-generally denied that there were any valid single-premissed
arguments (e.g. Alexander, in APr 17.10-18.6: Barnes et al [1991],
pp.64-66). We hear of only one dissenting voice-that of Antipater (texts in
Hiilser [*1987], frags 1050-1057; see Barnes [1980], p.l75; Burnyeat [1994],
pp.46-48). Then did standard Stoic logic in Epictetus' time on at least one
point prefer Antipater to Chrysippus?
The inference may seem audacious; but it receives some support from a
passage in the Encheiridion.

These arguments are invalid [acruvaKtot]:

I am richer than you.

Therefore: I am better than you.

 APPENDIX 141

I am more eloquent than you.

Therefore: I am better than you.

These, on the other hand, are valid [auvm.:-rucoi]:

I am richer than you.

Therefore: my possessions are better than yours.

I am more eloquent than you.

Therefore: my diction is better than yours.

(ench 44)

(There is no parallel to this in the surviving Discourses: Schenkl's reference

to III xiv 11 is inappropriate.) All four arguments are single-premissed. It is
hardly plausible to suppose that the examples were carelessly written and that
Epictetus really held, after Chrysippus, that a valid argument must have two
(or more) premisses. I suppose that he deliberately followed Antipater. And
of course he was right to do so. (We know nothing of the arguments which
induced Chrysippus to reject single-premissed arguments-the surviving
Peripatetic arguments to the same conclusion are embarrassingly silly.)

11: Kat -rou-ro] I.e. (I suppose): 'then you must also acquire the capacity to
determine when one thing follows from another or from others'. (But Long
takes '-rou-ro' to refer forward to 'Kat au-r6v -r' ... '.)

avaa-rpacpflaea9a.t Kat au-r6v -r'] The 'Kat, after 'avaa-rpacpflaea9at, I

take to be epexegetic: this perhaps makes the odd combination of particles
('Kat . . . "t . . . Kat ... ') more tolerable-but I have found no precise
parallel.

a1tooi~etv Kaa-ra a1to06v-ra] The text has been doubted; but 'a1to-
oioovat' may mean 'defme' (Sextus, M XI 8; Diogenes Laertius, VII 60;
scholiast to Dionysius Thrax, 107.5-the defmition of defmition ascribed to
Chrysippus). The defmitions, I take it, are premisses or principles for proofs:
the Sage must lay down his principles and then conduct proofs from them.

i>1to -r&v aocpt~o~tvrov ... ] Cf below, 26; I xxvii 6; and esp PHerc
~'118'
1020 = Htilser [*1987], frag 88.

12: -rp61trov] '-rp61to~' is the technical Stoic term for an argument schema:
see e.g. Diogenes Laertius, VII 76; Sextus, MVIII 227. And for 'auvayetV'
142 APPENDIX

in the sense of 'be valid' see e.g. Diogenes Laertius, VII 78; Sextus, M VIII
413.

14: -co 'lfEUOo~] I.e. the false conclusion-Epictetus does not mean 'fallacy'
(pace Oldfather).

16: -co oavEicracrSat] For the example see I viii 2 (above, pp.31-32). -
'otaA:UEtv' is the technical term for 'discharge a debt'; I have found no
parallel for 'bttJlEVEt v E1tt', but I suppose that that too was technical.

18: JlTt JlEV6v-ccov 0 ... ] The supplement which Upton introduced from his
codex and which Schweighauser printed in an emended form (see [* 1799], II
pp.113-114) is evidently a learned guess-and a poor one. (On Upton's
'codex'-a miscellany of conjectures from various sources-see Schenkl
[*1916], pp.LVII-LIX.) My supplement purports to give the general sense of
the missing section, not the original words in which the sense was clothed.

19: ouoi: yap ilJltV En ouoi: JCaS' iJJl6.~] '!CaS' i]Jla~ is picked up in 33
by '!CaS' au-c6v' (cf I xxviii 5). Epictetus does not mean 'the conclusion no
longer follows from the premisses'; he means that it no longer 'results for
us', i.e. that it no longer follows from anything which we are obliged to
accept. Oldfather and Carter/Hard make nonsense of the text.

20: icr-copficrat] For 'icr-copficrat' see II xiv 28; III vii 1: the word means
'look at', 'contemplate'-Epictetus does not say that we should survey or
classify or give an account of changing premisses; merely that we should be
on the qui vive for them. He does not tell us how to identify such premisses;
but no doubt sentences which use temporal indexicals in certain ways will be
likely to express pertinent assertibles. Sextus' example contains the word
'already [1tp&-cov]'. It is easy to construct parallel cases with such terms as
'not yet', 'only once', 'for the last time', and so on.
One of the spurious items in the Chrysippean catalogue (above, p.99
n.288) is titled 'A6yot JlE'ta1tt1t'tOV'tE~ 1tpo~ -ci}v JlEcrO'tll'ta'. The phrase
'1tpo~ -ci}v JlECfO'tllta' is puzzling. It is usually connected with a passage in
Diogenes Laertius, VII 57, according to which Antipater introduced the
JlECfO'tll~ as a sixth 'part of speech' in addition to the five which Chrysippus
and Diogenes of Babylon had recognized (proper name, appellative, verb,
connector, article: further texts in Hiilser [* 1987], frags 536-549).
MEcrO'tll'tE~, in Antipater's sense, are then taken to be adverbs; and in the
pseudo-Chrysippean title '-ci}v JlECfO'tll'ta' is translated as 'adverb' (so e.g.
Hadot [1994], p.347-with a gloss: 'plainly adverbs of time are in question')
or 'temporal adverb' (e.g. Hiilser [*1987], p.181-cfBobzien [1986], p.32).
 APPENDIX 143

All very splendid-but in fact there is no evidence at all for identifying

Antipater's J..LEO'O't'll'ttc; with temporal adverbs, and little enough for
identifying them with adverbs. The word for 'adverb' in Dionysius Thrax,
and in all the later Greek grammarians, is 'E1ttPP'IlJ..La', not 'J..LEO'O't'llc;'. (More
precisely, an E1ttPP11Jl<l is closer to a modem adverb----of which it is the
ancestor-than it is to any other modem part of speech.) Dionysius lists 26
varieties ofE1ttPP11Jl<l ( 19 [74.3]), and the second variety is described thus:
'ta oE: J..Ltcr6't'll'toc; <o11A.rowca> otov lCaA.&c; cr6cproc;. This short entry has
puzzled modem scholars, as it puzzled ancient scholars (for whose
implausible guesses see the scholia to Dionysius Thrax, GG I iii, 59.27-60.2;
97.31-98.5; 274.28-275.28; 429.34-430.7); but whatever may lie behind it,
one thing is clear: E1ttppi1J..La'ta of this sort are not temporal adverbs. (They
are perhaps adverbs which terminate in -roc;, or else, more specifically,
adverbs which derive from adjectives and terminate in -roc;.) Scholars have
conjectured that the word 'J..LEO'O't'llc;' was earlier used as a generic term for
'adverb', and that '1tiPP'IlJ..La', a more transparent term, later supplanted it
(see Ax [1991], pp.284-285; Schenkeveld [1993], pp.276-277). The scholia
to Homer's Iliad use the term 'J..LEO'O't'llc;' a dozen or more times to designate a
part of speech. In most of these cases, the part in question is clearly an
adverb of Dionysius' second type (e.g. 0 277a). The only passage I have
noticed which might be thought to contain a designation of adverbs in
general is I 446c: ou A.tyet vtov 1Ca'ta 'tTJV itA.tlCiav O.A.A.O. 1Ca'ta J..Ltcr6't'll'ta
av'tt 'tOU VECOO''tt itProv'ta. But the scholiast may have taken 'VECOO''ti'' an
elongated form of 'vtroc;' (Herodian, GG III i, 536.19-22; III ii, 19.19-21;
464.11-14), to belong to Dionysius' second type. Simplicius reports that oi
1taA.awi used the word '1t't&crtc;' of many different items, including 'tac; vuv
JroAOUJ..LEvac; J..LEO'O't'll'tac; (in Cat 37.12-13-cf208.17-20); and at 388.19-27
he may implicitly ascribe this use of the word 'J..LtcrO't'llc;' to the Stoics. But it
is not clear what sense Simplicius himself gives the word; and all his
examples ('avoptiroc;', 'lCaA.roc;', 'cppoviJ..Lroc;', 'acpp6vroc;') are in fact adverbs
of Dionysius' second type. This is, at best, frail evidence that the word
'J..LtO'O't'llc;' originally designated adverbs in general; and I know of no
stronger evidence.
Nonetheless, it will be demanded, what else could Antipater's J..LEO'O't'llc;
have been if it was not an adverb? Well, a pronoun (which the early Stoics
did not distinguish from a name) or a preposition (which they did not
distinguish from a connector) or a participle (which they did not distinguish
from a verb)-cf scholia to Dionysius Thrax, GG I iii, 356.7-16. Participles,
for example, might readily be thought of as 'means' between nouns and
verbs (cfPlutarch, quaest Platon 1011C: a participle or JlE'tOXTt considered as
a J..Lt'YJ..La pilJ..La'toc; ... !Cat 6v6J..La'toc;; Dionysius Thrax, 15 [60.1-2]: A.tl;tc;
J..Lt'txoucra 'tflc; 'trov P'llJ..LU'trov !Cat 'tflc; 'trov a'trov iot6't'll'toc;; Priscian, inst XI
144 APPENDIX

1: participium ... ex utroque [sc nomen and verbum] nascitur; cf scholia to

Dionysius Tirrax, GG I iii 254.27-32). Or I wonder if Diogenes has not
garbled his material: in the grammarians, 'JlEOO'tllc;' is a standard term for the
middle voice (e.g. Dionysius Tirrax, 13 [48.1-49.3]): perhaps Antipater
distinguished the middle from the active and the passive voices, and his
insight was misconstrued as the discovery of a new part of speech.
In sum, I do no! think that we know what Antipater did-and we are
certainly in no position to interpret the pseudo-Chrysippean titles.

fl'tU1t'tCOcrtV au1:&v] Not 'modified meanings' (Oldfather-after Long)-see

1.

tv au'tft 'tTI tpco'tflcret ... ] I.e. 'a premiss may change [sc. its truth-value]
while you are offering me a potential premiss or while I am accepting it or
while I am inferring something from what I have accepted and so on'. I do
not understand the perfect tense of 'cruA.A.eA.oyicr9at'. The final 'etc' is also
puzzling: what else could Epictetus have in mind? My assenting to the
conclusion, once I have drawn it? My using the conclusion as a premiss in a
further argument?

ij n vt &.A.A.(!) 'tOtoU't(!)] The received text may stand, pace Schenk}, who
obelizes. See Schweighliuser [* 1799], II p.115.

'tap6.crcrecr9at ~A.btoucrt] According to the MS reading, the foolish get upset

when (or because) they do not see what follows. But surely they get in a state
precisely because they do see what follows? They see that something follows
from what they have correctly conceded; and they also see that this thing is
false. Hence their perplexity. (I cannot see that the occurrence of ~AE7tEtv' in
33 helps to defend the received text; nor is it plausible to construe '1:0
a1(6A.oueov' in a loose sense-it picks up '1:0 a1(6A.oueov' in 17.)

22: l(<Xt 1:0 au1:o] For the parallelism between the case of hypotheticals and
the case of changing arguments see above, p.107 n.303; and compare 19
with 24.

ai'tfjcrai nva tm6eecrtv] For hypotheses as 'postulates' see e.g. Sextus, M

III 4; Proclus, in Euc/76.24-77.3.

E7tt~6.9pav] Epictetus is thinking of Plato, rep 511B ('7ti~acrtc;')-cf Symp

211C ('1tava~a9Jl6c;').
 APPENDIX 145

23: 1tEpt ... Ka8f1Kovtoc;] Wolfs deletion, which Schenk! accepts, is hard
to resist. It would, to be sure, be tolerable to have a further reminder that we
are still concerned with to Ka8fjKov (cf 13); but the form which the
reminder takes in the MSS is intolerable. - The train of thought in 23-24
is, to say the least, abrupt. Epictetus offers no answer to the question raised in
23, and he begins 24 with an entirely new question. I suspect that in place
of the phrase which Wolf deleted there originally stood an answer to the
question of 23.

25: 6 q>p6vt)loc;] Not 'a prudent young man' (Carter/Hard). 'q>p6vt)loc;', like
'cr6q>oc;' and 'cr1touoaioc;', is a standard designation for the Sage.

26: OEtvoc; f.v f.protflcrEt] 'oEtv6c;' with a bare dative is hard to parallel; and
Meibom's addition is now supported by PHerc 1020, ii 4.

avE~a1t<X't11toc;] See I iv 11; III v 7; PHerc 1020, i 1-6.

27: E1ttvOOUJlEV] Not 'think' (Oldfather, Laurenti, Carter/Hard): Epictetus

means-as 29 ('1tp6A11'Jftc;') shows-that indifference to logical propriety
is inconsistent with the concept of a wise man.

30: aypu1tvflcrojlEV] Cfe.g. II xxi 19; III xv 11; IV i 176; ix 16; and see
Hijmans [1959], pp.68-70.

~Epyas6JlEVot tov ... Myov] Cf e.g. III ix 20 (above, p.64); xv 13; IV x 13.

32: According to Hirzel [1895], p.246 n.1, this little exchange shows that
Epictetus started life as a Cynic and was converted to Stoicism by Musonius.
I cannot for the life of me see why Hirzel makes this inference.

to 1tapaA.Et1tOJlEVOV v f.v cruA.A.oytcrJltj) n vt oux EUptcrKov] Presumably

Epictetus is referring to a logical exercise: the teacher offers the pupil an
enthymematic argument, and he has to 'fmd the missing premiss'. Compare
the (technical) use of 'to 1tapaA.Et1tOJlEvov' at Alexander, in Top 9.16 (cf
15.4; 32.23); Galen, natfac II 95-96 Kiihn (cf art sang IV 710 Kiihn; PHP V
257 Kiihn, 258 Kiihn, etc); Simplicius, in ench XXXV 527 Hadot; and note
'1tapai.Et'Jitc;' at e.g. Sextus, PHil 150.

to Kam troA.wv KatKaucra] A stock example, according to Upton, apud

Schweighauser [*1799], II p.120. Even if the example is not a topos, we
hardly need see an explicit allusion to the fire of 69 (or 80), pace Millar
[1965].
 BIBLIOGRAPHY
The Bibliography lists, first, the texts and translations of Epictetus which I have used;
secondly, editions of various other ancient texts; and thirdly, the items of modem literature to
which I have referred. It is not a bibliography of work on Epictetus, nor of work on Stoic
logic; but it does contain-to the best of my knowledge--everything pertinent to Epictetus'
attitude to Stoic logic.

(i) Epictetus:

M.Billerbeck [*1978]: Epiktet: vom Kynismus, Philosophia Antiqua 34 (Leiden)

E.Carter [*1995]: Epictetus: the Discourses, the Handbook, Fragments, revised by R.Hard
(London)
R.Laurenti [*1989]: Epitteto: le Diatribe e iframmenti (Rome/Bari)
G.Long [*1885]: The Discourses of Epictetus (London)
W.A.Oldfather [*1925): Epictetus, Loeb Classical Library (Cambridge Mass)
H.Schenkl [*1916]: Epictetus: dissertationes ab Arriano digestae (Leipzig [2nd ed])
J.Schweighliuser [*1799): Epictetae Philosophiae Monumenta (Leipzig, 1799-1800)
J.Souilhe and A.Jagu [*1969]: Epictete: Entretiens (Paris, 1969-1991 [2nd ed])

(ii) Other ancient texts:

I have standardly used the standard editions-OCT, Teubner, Bude, CIAG. Galen is cited
(whenever possible) with references to the Kiihn edition; but more recent editions (CMG,
Teubner, etc) have been consulted where they exist. Note also the following:

A.Angeli [*1988]: Filodemo: Agli amici di scuola (PHerc. 1005), La scuola di Epicuro 7
(Naples)
J.Barnes, S.Bobzien, K.Flannery and K.Ierodiakonou [*1991]: Alexander of Aphrodisias: On
Aristotle's Prior Analytics 1.1-7 (London)
G.Bastianini and A.A. Long [*1992]: 'Hierocles: elementa moralia', in CPF [*1992), 268-451
H.F.Cherniss [*1976]: Plutarch's Moralia XIII ii, Loeb Classical Library (Cambridge Mass)
CPF [*1992, *1995): Corpus dei papiri filosofici greci e Iatini I 1.. , III (Florence)
P.H.de Lacy and E.A.de Lacy [*1978]: Philodemus: On Methods of Inference, La scuola di
Epicuro I (Naples)
J.Dillon [*1993]: Alcinous: the Handbook to Platonism (Oxford)
E.R.Dodds [*1959): Plato: Gorgias (Oxford)
H.Dorrie and M.Baltes [*1990, *1993, *1996]: Der Platonismus in der Antike II, III, IV
(Stuttgart/Bad Cannstatt)
T.Dorandi [*1994]: Filodemo: Storia dei filosofi-La stoiz de Zenone a Panetio (PHerc.
1018), Philosophia Antiqua 60 (Leiden)
I.Diiring [*1961]: Aristotle's Protrepticus, Studia graeca et latina Gothoburgensia 12
(Goteborg)
S.Ebbesen [*1981): Commentaries and Commentators on Aristotle's Sophistici Elenchi,
Corpus Latinum Commentariorum in Aristotelem Graecorum 7 (Leiden)
I.Hadot [*1996]: Simplicius: Commentaire sur le Manuel d'Epictete, Philosophia Antiqua 66
(Leiden)
K.-H.Hiilser [*1987]: Die Fragmente zur Dialektik der Stoiker (Stuttgart/Bad Cannstatt,
1987-1988)
I.G.Kidd [*1972, *1988]: Posidonius, Cambridge Classical Texts and Commentaries 13 and
14 (Cambridge)
F.Longo Auricchio [*1977]: Philodemus: de rhetorica I et II, Ricerche sui papiri ercolanesi 3
(Naples)
148 BIBLIOGRAPHY

V.Nutton [*1979]: Galen: On Prognosis, Corpus Medicorum Graecorum V 8, I (Berlin)

E.Puglia [*1988]: Demetrio Lacone: Aporie testuali ed esegetiche in Epicuro, La scuola di
Epicuro 8 (Naples)
N.Rescher and M.E.Marmura [*1966]: The Refutation by Alexander of Aphrodisias of
Galen's Treatise on the Theory ofMotion (Islamabad)
D.A.Russell [*1964]: 'Longinus ':on the Sublime (Oxford)
D.A.Russell [*1992]: Dio Chrysostom: Orations VII, XII, XXXVI (Cambridge)
D.N.Sedley [*1995]: 'Commentarium in Platonis 'Theaetetum", in CPF [*1995], 227-562
A.N.Sherwin-White [*1966]: The Letters of Pliny (Oxford)
M.F.Smith [*1993]: Diogenes ofOenoanda: the Epicurean Inscription, La scuola di Epicuro
suppt 1 (Naples)
H. von Staden [*1989]: Herophilus: the Art of Medicine in Early Alexandria (Cambridge)

(iii) Modern literature:

A.Alberti (ed.) [1990]: Logica, Mente e Persona (Florence)

L.Aiexander [1990]: 'The Living Voice', in Clines eta/ [1990], 221-247
L.Alexander [1993]: The Preface to Luke's Gospel, Society for New Testament Studies
monograph 78 (Cambridge)
K. A. Algra, P. W. van der Horst, D.T. Runia (edd.) [1996]: Polyhistor: Studies in the History
and Historiography of Ancient Philosophy, Philosophia Anti qua 72 (Leiden)
C.Atherton [1993]: The Stoics on Ambiguity (Cambridge)
W.Ax [1991]: 'Sprache als Gegenstand der alexandrinischen und pergamenischen Philologie',
in Schmitter [1991], 275-301

J.Bames [1969]: 'Aristotle's Theory of Demonstration', Phronesis 14, 123-152

J.Bames [1980]: 'Proof Destroyed', in Schofield eta/ [1980], 161-181
J.Bames [1982]: 'Medicine, Science and Logic', in Barnes eta/ [1982], 24-68
J.Barnes [1983]: 'Terms and Sentences', Proceedings of the British Academy 69, 279-326
J.Barnes [1984]: 'The Logical Investigations of Chrysippus', Jahrbuch des
Wissenschaftskolleg zu Berlin 1984/5, 19-29
J.Barnes [1986]: 'Peripatetic Negations', Oxford Studies in Ancient Philosophy 4, 201-214
J.Barnes [1988]: 'Epicurean Signs', Oxford Studies in Ancient Philosophy suppt, 91-134
J .Barnes [1989]: review of Hulser [*1987], Classical Review 39, 263-264
J.Bames [1990a]: 'Logical Form and Logical Matter', in Alberti [1990], 7-119
J.Barnes [1990b]: The Toils of Scepticism (Cambridge)
J.Barnes [1991]: 'Galen on Logic and Therapy', in Kudlien and Durling [1991], 50-102
J.Barnes [1992]: 'Metacommentary', Oxford Studies in Ancient Philosophy 10, 267-281
J.Barnes [1993a]: 'Galen and the Utility of Logic', in Kollesch and Nickel [1993], 33-52
J.Bames [1993b]: "A Third Sort of Syllogism'-Galen and the Logic of Relations', m
Sharples [1993], 172-194
J.Barnes [1993c]: 'Imperial Plato',Apeiron 26,129-151
J.Bames [1993d]: 'Aristotle's Philosophy of the Sciences', Oxford Studies in Ancient
Philosophy 11, 225-241
J.Bames [1996]: 'The Catalogue ofChrysippus' Logical Works', in Algra et al [1996], 169-
184
J.Barnes [1997]: 'Roman Aristotle', in Barnes and Griffin [1997], 1-69
J.Barnes, M.Schofield and R.Sorabji (edd.) [1975]: Articles on Aristotle I (London]
J.Barnes, J.Brunschwig, M.F.Burnyeat and M.Schofield (edd.) [1982]: Science and
Speculation (Cambridge)
J.Barnes and M.Mignucci (edd.) [1988]: Matter and Metaphysics (Naples)
J.Barnes and M.Griffin (edd.) [1989, 1997]: Philosophia Togata I, II (Oxford)
O.Becker [1957]: 'Ueber die vier 'Themata' der stoischen Logik', in his Zwei
Untersuchungen zur antiken Logik, Klassisch-philologische Studien 17 (Wiesbaden)
27-49
S.Bobzien [1986]: Die stoische Modallogik (WOrzburg)
 BIBLIOGRAPHY 149

S.Bobzien [1996]: 'Stoic Syllogistic', Oxford Studies in Ancient Philosophy 14, 133-192
J.Bollack and A.Laks (edd.) [1976]: Etudes sur /'Epicurisme antique, Cahiers de philoIogie 1
(Lille)
A.Bonhtiffer [1890]: Epictet und die Stoa (Stuttgart)
A.Bonhtiffer [1894]: Die Ethik des Stoikers Epictet (Stuttgart)
G.Boter [1989]: The Textual Tradition of Plato's Republic, Mnemosyne suppt 107 (Leiden)
G.W.Bowersock (ed.) [1974]: Approaches to the Second Sophistic (University Park Penn)
A.Brancacci [1985]: Rhetorike Philosophousa (Naples)
L.Brisson, M.-O.Goulet-Caze, R.Goulet, and D.O'Brien (edd.) [1982, 1992]: Porphyre: Ia Vie
de Plotin I, II (Paris)
!.Bruns [1897]: De schola Epicteti (Kiel)
J.Brunschwig (ed.) [1978]: Les stoi'ciens et leur logique (Paris)
J.Brunschwig [1980]: 'Proof Defined', in Schofield et al [1980], 125-160
P.A.Brunt [1974]: 'Marcus Aurelius in his Meditations', Journal ofRoman Studies 64, 1-20
P.A.Brunt [1977]: 'From Epictetus to Arrian', Athenaeum 55, 19-48
P.A.Brunt [1979]: 'Marcus Aurelius and the Christians', in Deroux [1979], 483-520
P.A.Brunt [1994): 'The Bubble of the Second Sophistic', Bulletin of the Institute of Classical
Studies 41, 25-52
A.Bulloch, E.S.Gruen, A.A.Long and A.Stewart (edd.) [1993]: Images and Ideologies
(Berkeley Cal)
M.F.Burnyeat [1976]: 'Protagoras and Self-refutation in Later Greek Philosophy',
Philosophical Review 85, 44-69
M.F.Burnyeat [1994): 'Enthymeme: Aristotle on the Logic of Persuasion', in Furley and
Nehamas [1994], 3-55

F.Caujolle-Zaslawsky [1989]: 'Agrippa', in Goulet [1989a), 71-72

M.L.Ciarke [1971 ): Higher Education in the Ancient World (London)
D.J.A.Ciines, S.E.Fowl, and S.E.Porter (edd.) [1990]: The Bible in Three Dimensions, Journal
for the Study of the Old Testament suppt 87 (Sheffield)
C.Codofier [1989]: 'La physique de Seneque: ordonnance et structure des 'Naturales
Quaestiones", in Haase [1989], 1779-1822
A.Concolino Mancini [1980]: 'Sui testo filosofico di PSI II 152', in Pintaudi [1980], 51-55

F.Decleva Caizzi [1993]: 'The Porch and the Garden: early Hellenistic images of the
philosophical life', in Bulloch eta/ [1993], 303-329
P.H.de Lacy [1943]: 'The Logical Structure of the Ethics of Epictetus', Classical Philology
38, 112-125
P.H.de Lacy [1974]: 'Plato and the Intellectual Life of the Second Century AD', in
Bowersock [1974), 4-10
N.Denyer [1988]: 'Stoicism and Token Reflexivity', in Barnes and Mignucci [1988], 375-396
N.Denyer [1996]: 'Gaskin on the Master Argument', Archiv for Geschichte der Philosophie
78, 166-180
C.Deroux (ed.) [1979]: Studies in Latin Literature and Roman History I (Brussels)
P.Desideri [1994): 'Dion Cocceianus de Pruse, dit Chrysostome', in Goulet [1994a], 841-856
A.Dietrich (ed.) [1976]: Akten des VII. Kongresses for Arabistik und Islamwissenschaft,
Abhandlungen der Akademie der Wissenschaften in Gottingen, phil.-hist.Klasse 98
(Gottingen)
J.Dillon and A.A.Long (edd.) [1988]: The Question of 'Eclecticism' (Berkeley Cal)
R.Dobbin [1991]: 'fipoai.pecnc; in Epictetus',Ancient Philosophy II, 111-135
K.Doring and T.Ebert (edd.) [1993): Dialektiker und Stoiker, Philosophie der Antike
(Stuttgart)
P.-L.Donini [1992]: 'Plutarco e i metodi dell'esegesi filosofica', in Gallo and Laurenti [1992],
79-96
P.-L.Donini [1994): 'Testi e commenti, manuali e insegnamento: Ia forma sistematica e i
metodi della filosofia in eta postellenistica', in Haase [1994], 5027-5100
P.-L.Donini [1996): 'Alessandro di Afrodisia e i metodi dell'esegesi filosofica', in Moreschini
[1996], 107-129
150 BIBLIOGRAPHY

T.Dorandi [1994]: 'Dionysius de Cyrime', in Goulet [1994a], 865-866

T.Ebert [1991]: Dialektiker und friJhe Stoiker bei Sextus Empiricus, Hypomnemata 95
(Gottingen)
R.H.Epp (ed.) [1985): Spindel Conference 1983: Recovering the Stoics, Southern Journal of
Philosophy 33 supplement (Memphis Tenn)
M.Erler (1991): ''E1tttllliUElv uc:Tiicpetav', Cronache Ercolanesi 21,83-88
M.Erler [1993]: 'Phi/alogia medicans: Wie die Epikureer die Texte ihres Meisters Jasen', in
Kullmann and Althoff [1993), 281-303

A.J.Festugiere [1960]: 'Lieux communs et themes de folk-lore dans l'Hagiographie primitive',

Wiener Studien 73, 123-152
K.L.Fiannery [1995): Ways into the Logic of Alexander of Aphrodisias, Philosophia Antiqua
62 (Leiden)
S.Follet [1989a ): 'Aristocles de Lampsaque', in Goulet [1989a ], 381-382
S.Follet [1989b ]: 'Arrien de Nicomedie', in Goulet [1989a), 597-604
M.Frede [1974a): Die stoische Logik (Gottingen)
M.Frede [1974b): 'Stoic vs. Aristotelian Syllogistic', Archiv fiir Geschichte der Philosophie
56, 1-32 [reprinted in his Essays in Ancient Philosophy (Oxford, 1987), 99-124]
M.Frede [1989): 'Chaeremon der Stoiker', in Haase [1989], 2067-2103
W.Freytag [1995]: Mathematische Grundbegriffe bei Sextus Empiricus, Spudasmata 57
(Hildesheim)
L.Friedllinder [1920): Darstellungen aus der Sittengeschichte Roms (Leipzig [9th ed])
D.J.Furley and A.Nehamas (edd.) [1994]: Aristotle's Rhetoric (Princeton N.J.)

I. Gallo and R.Laurenti (edd.) [1992]: 1 Moralia di Plutarco tra filologia e filosofia, Strumenti
per Ia ricerca plutarchea I (Naples)
H.Y.Gamble [1995): Books and Readers in the Early Church (New Haven Conn)
J.Gaskin [1995): The Sea Battle and the Master Argument, Quellen und Studien zur
Philosophie 40 (Berlin)
T.Goransson [1995): Albinus, Alcinous, Arius Didymus, Studia graeca et latina
gothoburgensia 61 (Goteborg)
P.P.F.Gonzalez [1994): 'Comutus', in Goulet [1994a], 460-473
J.B.Gould [1965): 'Reason in Seneca', Journal for the History of Philosophy 3, 13-25
R.Goulet [1987]: La philosophie de Moise, Histoire des doctrines de l'antiquite classique II
(Paris)
R.Goulet (ed.) [1989a, 1994a]: Dictionnaire des philosophes antiques I, II (Paris)
R.Goulet [1989b]: 'Apollonius de Tyr', in Goulet [1989a], 294
R.Goulet [1989c]: 'Aristoboulos', in Goulet [1989a), 379-380
R.Goulet [1989d): 'Athenodore de Tarse', in Goulet [1989a), 654-657
R.Goulet [1989e): 'Aulu-Gelle', in Goulet [1989a], 675-687
R.Goulet [1994b): 'Chairemon d'Aiexandrie', in Goulet [1994a], 284-286
R.Goulet and M.Aouad [1989]: 'Alexandre d 'Aphrodisie', in Goulet [1989a ), 125-139
M.-O.Goulet-Caze [1982): 'L'arriere-plan scolaire de Ia Vie de Plotin', in Brisson et al
[1982), 229-327
M.-O.Gou1et-Caze [1986]: L 'ascese cynique (Paris)
M.-O.Gou1et-Caze (1989]: 'Apollodore de Seleucie', in Goulet [1989a), 276-278
M.Griffin [1992]: Seneca: a philosopher in politics (Oxford [2nd ed))
C.Guerard [1994a): 'Crinis', in Goulet [1994a), 511
C.Guerard [1994b]: 'Diogene de Seleucie', in Goulet (1994a], 807-810

W.Haase (ed.) [1987, 1987, 1989, 1990, 1992, 1994]: Aufstieg und Niedergang der
romischen Welt II 36.1, 2, 3, 4, 5, 7 (Berlin)
E.Hachmann (1995]: Die Fuhrung des Lesers in Senecas Epistulae Morales, Orbis Antiquus
34 (Munster)
P.Hadot [1978]: 'Une cle des Pensees de Marc Aurele: les trois topoi philosophiques selon
Epictete', Les etudes philosophiques I, 65-83
 BIBLIOGRAPHY 151

P.Hadot [1987]: 'Theologie, exegese, revelation, ecriture dans Ia philosophie grecque', in

Tardieu [1987], 13-34
P.Hadot [1994]: 'Liste commenre des oeuvres de Chrysippe (D.L. VII 189-202)', in Goulet
[1994a], 336-356
P.Hadot [1995]: Qu'est-ce que Ia philosophie antique? (Paris)
J.Hahn [1989]: Der Philosoph und die Gese/lschaft, Heidelberger Althistorische Beitrage und
Epigraphische Studien 7 (Stuttgart)
R.Helm [1902]: 'Lucian und die Philosophenschulen', Neue Jahrbiicher 5, 188-213, 263-278,
351-369
J.P.Hershbell [1993]: 'Epictetus and Chrysippus', Illinois Classical Studies 18, 139-146
B.L.Hijrnans [1959]: Al:KHEil: (Assen)
B.L.Hijmans [1987]: 'Apuleius, philosophus Platonicus', in Haase [1987], 395-475
K.J.J.Hintikka [1967]: 'Time, Truth and Knowledge in Ancient Greek Philosophy', American
Philosophical Quarterly 4, 1-14 [reprinted with additions in his Time and Necessity
(Oxford, 1973), 62-92]
R.Hirzel [1895]: Der Dialog (Leipzig)
L.Holford-Strevens [1988]: Aulus Gellius (London)
K.-H.Hiilser [1992]:'Galen und die Logik', in Haase [1992], 3523-3554

H.A.Ide [1992]: 'Chrysippus' Response to Diodorus' Master Argument', History and

Philosophy of Logic 13, 133-148
K.lerodiakonou [1990]: Analysis in Stoic Logic (diss. London)
K.lerodiakonou [1993]: 'The Stoic Division of Philosophy', Phronesis 38,57-74
A.-M.Ioppolo [1980]: Aristone di Chio e lo stoicismo antico (Naples)

E.Kapp [1931]: 'Syllogistik', Pauly-Wissowas Realencyclopiidie der classischen Altertums-

wissenschaft IVA (Stuttgart), I 046-1067 [reprinted in his Ausgewiihlte Schriften
(Berlin, 1968, 254-277); English translation in Barnes et al [1975], 35-49
H.Karpp [1964]: 'Viva vox', Jahrbuch for Antike und Christentum Erganzungsband I:
Mullus: Festschrift for Theodor Klauser (Munster), 190-198
A.Kessler, T.Ricklin and G.Wurst (edd.) [1994]: Peregrina Curiositas, Novum Testamentum
et Orbis Antiquus 27 (Gottingen)
l.G.Kidd [1978]: 'Posidonius and Logic', in Brunschwig [1978], 273-284
J.Kollesch and D.Nickel (edd.) [1993]: Galen und das hellenistische Erbe, Sudhoffs Archiv
Beihefte 32 (Stuttgart)
F.Kudlien and R.Durling (edd.) [1991]: Galen 's Method of Healing (Lei den)
W.Kullmann and J.Aithoff (edd.) [1993]: Vermittlung und Tradierung von Wissen in der
griechischen Kultur (Tiibingen)

M.-L.Lakmann [1995]: Der Platoniker Tauros in der Darstellung des Aulus Gellius,
Philosophia Antiqua 63 (Leiden)
R.Laurenti [1989]: 'Musonio, maestro di Epitteto', in Haase [1989], 2105-2146
T.-S.Lee [1984]: Die griechische Tradition der aristotelischen Syllogistik in der Spiitantike,
Hypornnernata 79 (Gottingen)
A.D.Leernan [1953]: 'Seneca's Plans for a Work 'Moralis Philosophia' and their Influence on
his Later Epistles', Mnemosyne 6, 307-313
R.Lim [1993]: 'The Auditor Thaumasius in the Vita Plotini', Journal of Hellenic Studies 113,
157-160
G.E.R.Lioyd [1987]: The Revolutions of Wisdom, Sather Classical Lectures 52 (Berkeley Cal)
A.A.Long [1978]: 'Dialectic and the Stoic Sage', in Rist [1978b], 101-124

J.Mansfeld [1988]: 'Philosophy in the Service of Scripture: Philo's Exegetical Strategies', in

Dillon and Long [1988], 70-102 [reprinted in his Studies in Later Greek Philosophy
and Gnosticism (London, 1989), ch 10]
J.Mansfeld [1989]: 'Chrysippus and the Placita', Phronesis 34, 311-342
J.Mansfeld [1990]: 'Doxography and Dialectic', in Haase [1990], 3056-3229
J.Mansfeld [1994]: Prolegomena, Philosophia Antiqua 61 (Leiden)
152 BIBLIOGRAPHY

G.Mazzoli [1989]: 'Le 'Epistulae morales ad Lucilium' di Seneca', in Haase [1989], 1823-
1877
B.F.Meyer and E.P.Saunders (edd.) [1982]: Jewish and Christian Self-definition III (London)
M.Mignucci [1978]: 'Sur Ia logique modale des stolciens', in Brunschwig [1978], 317-346
M.Mignucci [1993]: 'The Stoic themata', in Doring and Ebert [1993], 217-238
F.Millar [1965]: 'Epictetus and the Imperial Court', Journal ofRoman Studies 55, 141-148
J.L.Moles [1978]: 'The Career and Conversion of Dio Chrysostom', Journal of Hellenic
Studies 98, 79-100
J.L.Moles [1983]: "Honestius quam ambitiosius'?-an Exploration of the Cynic's Attitude to
Moral Corruption in his Fellow Men', Journal of Hellenic Studies 103, 103-123
C.Moreschini (ed.) [1996]: Esegesi, parafrasi e compilazione in eta tardoantica (Naples)
G.W.Most [1989]: 'Comutus and Stoic Allegoresis', in Haase [1989], 2014-2065
!.Mueller [1969]: 'Stoic and Peripatetic Logic', Archiv for Geschichte der Philosophie 51,
173-187

B.Neuschlifer [1987]: Origenes als Philologe, Schweizerische Beitrlige zur Altertums-

wissenschaft 18 (Basel)
R.J.Newman [1989]: 'Cotidie meditare: Theory and Practice of the meditatio in Imperial
Stoicism', in Haase [1989], 1474-1517

D.O'Meara [1994]: 'Faut-il philosopher?-le voyage d'une question dans le monde antique',
in Kessler et al [1994], 1-12
E. Osborn [1987]: 'Logique et exegese chez Clement', Cahiers de Biblia Patristica 1:
Lectures anciennes de Ia Bible (Strasbourg), 169-190
G.E.L.Owen [1983]: 'Philosophical Invective', Oxford Studies in Ancient Philosophy I, 1-25
[reprinted in his Logic, Science and Dialectic (London, 1986), 347-364]

G.Patzig [1968]: Aristotle's Theory of the Syllogism (Dordrecht)

J.Pepin [1992]: 'PHILOLOGOS/PHILOSOPHOS', in Brisson et al [1992], 477-501
R.Pfeiffer [1968]: History of Classical Scholarship (Oxford)
R.Pintaudi (ed.) [1980]: Miscellanea Papyrologica, Papyrologica Florentina 7 (Florence)
M.Pohlenz [1970]: Die Stoa (Gottingen [4th ed])

S.L.Radt [1990]: 'Zu Epiktets Diatriben', Mnemosyne 43, 363-373

J.M.Rist [1978a]: 'Zeno and the Origins of Logic', in Brunschwig [1978], 387-400
J.M.Rist (ed.) [1978b]: The Stoics (Berkeley Cal)
J.M.Rist [1982]: 'Are you a Stoic?-the Case of Marcus Aurelius', in Meyer and Saunders
[1982], 23-45
J.M.Rist [1989]: 'Seneca and Stoic Orthodoxy', in Haase [1989], 1993-2012
A.Roselli [1989]: 'Appunti per una storia dell'uso apologetico della filologia: Ia nuova
edizione di Demetrio Lacone (PHerc 1012)', Studi classici e orientali 39, 117-138
D.A.Russell (ed.) [1990] Antonine Literature (Oxford)
R.Rutherford [1989]: The Meditations of Marcus Aurelius (Oxford)

C.Schliublin [1974]: Untersuchungen zu Methode und Herkunft der antiochenischen Exegese

(Cologne)
D.M.Schenkeveld [1984]: 'Stoic and Peripatetic Kinds of Speech Act and the Distinction of
Grammatical Moods', Mnemosyne 37,291-353
D.M.Schenkeveld [1993]: 'Scholarship and Grammar', Entretiens Hardt 40: La philologie
grecque a l 'epoque hel/enistique et romaine, 263-301
P.Schmitter (ed.) [1991]: Sprachtheorien der abendliindischen Antike, Geschichte der
Sprachtheorie 2 (Tiibingen)
M.Schofield [1983]: 'The Syllogisms of Zeno ofCitium', Phronesis 28, 31-58
M.Schofield [1991]: The Stoic Idea of the City (Cambridge)
M.Schofield, M.F.Bumyeat, and J.Bames (edd.) [1980]: Doubt and Dogmatism (Oxford)
H.Schreckenberg [1966]: 'Exegese I', Reallexicon for Antike und Christentum 6 (Stuttgart),
1174-1194
 BIBLIOGRAPHY 153

D.N.Sedley [1976]: 'Epicurus and his Professional Rivals', in BoBack and Laks [1976], 119-
159
D.N.Sedley [1985]: 'Stoic Metaphysics', in Epp [1985], 87-92
D.N.Sedley [1989]: 'Philosophical A11egiance in the Greco-Roman World', in Barnes and
Griffin [1989], 97-119
D.N.Sedley [1997]: 'Plato's auctoritas and the Rebirth of the Commentary Tradition', in
Barnes and Griffin [1997], 110-129
R.W.Sharples [1987]: 'Alexander of Aphrodisias: Scholasticism and Innovation', in Haase
[1987], 1176-1243
R.W.Sharples (ed.) [1993]: Modern Thinkers and Ancient Thinkers (London)
M.Simonetti [1996]: 'Ornelie e comrnentari patristici', in Moreschini [1996], 361-381
M.F.Smith [1996]: 'An Epicurean Priest from Apamea in Syria', Zeitschriftfiir Papyrologie
und Epigraphik 112, 120-130
P.A.Stadter [1980]: Arrian o[Nicomedia (Chapel Hill N.C.)
G.R.Stanton [1968]: 'The Cosmopolitan Ideas of Epictetus and Marcus Aurelius', Phronesis
13, 183-195
E.Stemplinger [1912]: Das Plagiat in der griechischen Literatur (Leipzig)
M.R.Stopper [1983]: 'Schizzi Pirroniani', Phronesis 28, 265-297
G .Strohmaier [ 1976]: 'Uebersehenes zur Biographie Lukians', Philologus 120, 117-122
M.W.Su11ivan [1967]: Apuleian Logic (Amsterdam)

M.Tardieu (ed.) [1987]: Les regles de /'interpretation (Paris)

M.Tecusan [1993]: Symposion and Philosophy (diss. Oxford)
T.Tieleman [1996]: 'The Hunt for Galen's Shadow: Alexander of Aphrodisias, De anima
94.7-100.7 Bruns Reconsidered', in Algra eta/ [1996], 265-283
M.N.Tod [1957]: 'Sidelights on Greek Philosophers', Journal ofHellenic Studies 77, 132-141
M.B.Trapp [1990]: 'Plato's Phaedrus in Second-century Greek Literature', in Russe11 [1990],
141-174
W.Trillitzsch [1962]: Senecas Beweisfiihrung (Berlin)

A.C.van Geytenbeek [1962]: Musonius Rufus and Greek Diatribe (Assen)

A.-J .Voelke [1993]: La philosophie comme therapie de l'iime (Fribourg)

F.Walsdorff [1927]: Die antiken Urteile iiber Platons Stil, Klassisch-philologische Studien 1
(Bonn)

J.Xenakis [1968]: 'Logical Topics in Epictetus', Southern Journal of Philosophy 6, 94-102

E.Ze11er [1909]: Die Philosophie der Griechen III i (Leipzig, 19094 )

K.Ziegler [1950]: 'Plagiat', Pauly-Wissowas Realencyclopiidie der classischen Altertums-
wissenschaft XX (Waldsee ), 1956-1997
F.W.Zimmerrnann [1976]: 'Al-Farabi und die philosophische Kritik an Galen von Alexander
zu Averroes', in Dietrich [1976], 401-414
 INDEXES

Index of Passages

Passages are listed only if they are quoted, paraphrased, or discussed.

Alexander of Aphrodisias 57 142

in APr 177.26-32 97-98 62 140
in Met 272.16-17 84 n.243 71 73n.l90
in Top 27.12-14 45, 52 n.l25 83 119
149.9-17 59 160 9
apud Simplicius 103-104 191 137
Antigonus of Carystus 196 86 n.251, 87 n.259, 90 n.264,
apud Eusebius 41 n.76 92 n.268, 96 n.284, 99 n.288, 142
Apuleius Dionysius Thrax
inti p.l90 4 n.l4 13 143
ix p.205 121-122 15 143
X p.207 121-122 19 143
xi p.207 121-122
Aristotle Ennius
APr26a36-bl 71 n.l84 frag 28 7
27a5-9 78 Epictetus
34a5-7 97 diss I i 4-5 60
Protr 59 i 7 60
EN 1103a30-b2 47 iv 5-9 44
Athenaeus iv 7 43
I 04B 56 n.l35 iv9 48
607B 36 iv II 65 n.164
6100 46n.96 iv 14 43
iv 14-16 44
Boethius iv 15-16 48
hyp sy/1 I ii 5 28 n.23 iv 22 43
vii 61-62, 129-145
Chrysippus vii I 65
Aoy Zll't x 9-10 86 n.255 vii 13 107
Cicero vii 13-16 118
de orat I xi 47 51 vii 16-19 118-119
Luc xxix 93 119 n.322 vii 20-21 105-106
Clement vii 22-29 94-98
strom II xx 125.1 17 n.l4 vii 24-25 91
vii 25-27 119-120
Dio Chrysostom vii 32 27, 30
xiii 12 2 viii 1-2 31-32
Diogenes Laertius viii 1-7 58
II 92 9 viii 6 39
VI 103 9 viii 9-10 39
VII 10 42 X 8 48
44 137 XV 6-7 100-101
46 119 xvii tit 59 n.l43
156 INDEXES

[Epictetus, diss I] xxi 21-22 57

xvii 1-3 60-61 xxiii 41 36
xvii 6-9 67-68 xxiii 46 57-58
xvii 10 67 xxiv 13-15 66
xvii 12 60 XXV 59-60, 66
xvii 13-19 43 xxix 39 92
xxii 18 7 n.32, 40 xxix 39-40 92 n.269
XXV JJ-12 87, 88-89 diss III ii 1 34
xxvi 1 90 ii 5 68
xxvi 3 66-67 ii 6 33-34, 136
xxvi 9 35 ii 8 75 n.205
xxvi 14 28, 28 n.23 ii 17 69,97
xxvii 2 75 n.207 v 17 56
xxvii 15 45 n.92 vi 1-3 74
xxix 34 82 viii 1 75
xxix 39-41 29, 95 ix 20-21 64
xxix 40 28 ix 22 64
xxix 50-5I 29 xxi 10 136
xxix 51 28, 28 n.23 xxiii 58
xxix 54 45 n.95 xxiv 78 39-40, 92
xxix 54-56 55-56, 73 xxvi 14-20 68-69
diss II i 3 28 n.24 xxxiii 30-31 57 n.\39
i 35 56-57 diss IV i 61 32-33
i 38 55, 56-57 iv 13-14 43
iii 3 45 iv 14 92
iii 3-5 83-85 v 37 40
iii 4-5 39 vii 6 63
iv 10-11 40-41 viii 8-9 57
vi I 65 n.\64 viii 9 28
ix 7-9 29 viii 12 29 n.30
ix 8 28, 28 n.23 ench 36 28 n.23
ix 13-14 47 42 28 n.23
ix 14 44 44 28, 140-141
ix 17-19 44 52 35, 38-39, 64, 69
ix 20-22 72 n.\86 frag I 25-27
xi 13-18 66 IX 72 n.\87
xii 9-10 28-29 X 40
xiii 21 75 XXVII 34 n.46
xiii 26 40 Eusebius
xvi 2 75 PE XIV xviii 26 41 n.76
xvii 27 55, 57,75 Eustratius
xvii 40 73 n.200 in APst 2.25-27 84 n.240
xviii 18 39
xviii 26 56 n.135 Fronto
xix tit 40 eloq ii 13 p.\41
xix 1-5 76-77,97 ii 17p.l44
xix 5-6 44-45, 45 n.91 ii 19 p.l14 31 n.36
xix 7 45 iv5p.l49 1
xix 8-10 35,72 iv 10p.l50 37
xix\2-17 36 v4p.l51 48-49
xix 20 40 n.71 v 4-5 p.\52 1-2
XX 72 n.\86
XX 2-3 28, 30 Galen
xxi 4 48 comp med Zoe XII 894 46
xxi 6-7 48 in Hippfract XVIIIB 321 48 n.\04
xxi 17 93 XVIIIB 318-323 49 n.\14
xxi 21 93
 INDEX OF PASSAGES 157

[Galen] PSI II !52 57 n.l38

inst log xvii 5 31 n.36 Philodemus
lib prop XIX 43 31 n.36 ind stoic !xi 2-3 51 n.l24
47 73n.199 7tpo~ 'tOU~[ frag 116 51
ord lib prop XIX 53 42 rhet I vi 10-18 20-21
PHPV224 81-82 Plato
praecogn XIV 660 6 n.29 Phdr275A 46
us part III 16 43 n.86 Rep 534E 68 n.l75
Gellius Plotinus
Iii 3-4 35-36 III vii I, 10-16 52
Vv 9 7 vi 8, 10-14 52
XIII viii 4 41 Plutarch
XVI ii 97 n.285 comm not 1061E 101
XVI viii 16-17 36-37 quaest conviv 615A 35
XVII xix I 40 quaest Plat IOIIC 143
XIX i 14 72 n.l87 sen ger 796D 45
stoic repugn 1035E 70
Heraclitus 1045E-1046A 60-61
frag 101 46-47 Porphyry
vit Plot 13 49
[Longinus] 17-18 54
v 54 Priscian
Lucian instX! I 143
!car 29-32 41
Jup trag 51 16 n.l2 Quintilian
Peregr II 51 n.121 Ix5 65n.163
Symp 41 VIII iii 33 4 n.14

Marcus Aurelius Scholia

I vii 2 I Dionysius Thrax, I iii 356.7-16 143
I vii 8 24 Homer, Iliad I 446c 143
I xvii 22 I Seneca
III xi I II n.43 ben VII ix 3 64 n.l60
VII lxvii 3 I brev vit viii 2 19 n.20
VIIIi5 I ep xii II 54
VIII xiii II n.43 xvi 2 47 n.l01
IX xxxvii 34 n.46 xxxiii7-10 45
Maximus ofTyre xlv4-5 12
i 8 37 8 14
Musonius Rufus 10 14
frag I 63 n.l56, 71 13 12
3 63, 78 n.220 xlviii 6-7 12, 14
8 14
Nepos xlix 5-6 12
Att xvii 3 42 12 13
lxxi 6 13
Origen lxxxii 8-9 13, 15-16
c Cels VI 2 36 n.57 19 13
in Psalmprol14-15 73 n.199 lxxv4 42
lxxxiii 9-11 15
Pacuvius lxxxv 15
apud Gellius 41 lxxxv I 16
Papyri lxxxvii 11-41 17-18
PHerc 307 86 n.255 lxxxviii 37 19 n.20
1005 51 lxxxviii 39 19 n.20
1018 51 n.124 lxxxix 9 I n.3
1427 20-21 II 7 n.33
158 INDEXES

[Seneca,eplxxxix] 139 I 04 n.300

12 7 n.33, 34 n.48 141 92-93
14 34n.48 142 93 n.274
17-18 19 144 121
cii 4-5 9 157-158 80 n.227
20 13 229 107
cviii 23 45-46 231 I 08, II 0
35 46 234 107, 109, 110-115
38 46 235 I 07 n.305
cxi 1-5 14, 36 Simplicius
cxiii 26 13, 19 in Cat 37.12-13 143
nat quaest I praef I 22 in ench praef 5-11 24 n.6
5 23 XV 45-49 42 n.80
6 22-23 LIV 25 28 n.23
VI xxxii I 22 LXVII 29-32 40
at vi 5 42 in Phys 1300.1-8 103-104
vit beat xviii I 41-42 Stobaeus
xix 1-3 42 eel II i 31 25-27
xxvii 5 42 ii 12 7 n.35
Sextus Empiricus ii 16 74
MIII17 87-88 xxxi 125 63 n.156, 71
VII15 9 xxxi 126 63
VIII 231 81 Suda
235-236 80-81 s.v. 'Aptcr'tOICAfj~ 73
418 104 n.300 s.v. etmv 'AAEI;avOpE\J~ 73 n.199
PHI 173-174 95-96
II Ill 91 Theognetus
136 93 apud Athenaeus 55 n.l35
 INDEX OF PERSONS !59

Index of Persons

In the case of authors, see also the Index of Passages.

Agrippa 96 n.283 and logic 3-4, 20, 60-61, 72, 79-80, 86-
Alcinous 6, 71 87,
Alexander of Aphrodisias 6, 20, 21, 23, 45, 91-92, 103-104, 126, 140-141
49 on logical puzzles 76 n.209, 76 n.213,
Alexinus 15,51 77,
Amelius 54 119 n.322
Anaxagoras 51 on Master Argument 35, 72, 76, 97-98
Andronicus 67 n.171 on physics 22
Antipater 35, 48, 72,72 n.l87, 76, 82-83, style of 27 n.18
126, 140-141, 142-144 works of 43-44,86 n.251, 87 n.259,
Antisthenes 60, 60 n.l49 90 n.264, 92 n.268, 96 n.284, 99,
Apollodorus of Seleucia 73 99 n.288, 142
Apollonius ofChalcedon 5 Cicero, M.Tullius 15-16, 56 n.l35, 113
Apollonius ofTyre 99 n.288 Cleanthes 35, 44, 51, 60, 72,72 n.l87, 74,
Apuleius, L. 6, 122 76
Archedemus 35, 41, 48, 72,72 n.187 Clement of Alexandria 53 n.l27
AristoofChios 7,9, 15,26n.ll,51 Cleomedes 5
Aristobulus 51 n.121 Colotes 42 n.79
Aristocles 73, 73 n.l98 Cornutus, L.Annaeus 4-5, 5 n.17
Aristotle Crassus, M.Licinius 51
ethics 10, 23, 47,49 Crinis 72-73
physics 6, 23, 43 n.86
works Demetrius the Cynic 41
Analytics 20,78-79,84,84 n.240, 97 Democritus 42, 51
Categories 4-5, 48 n.l 05, 49 n.l 07 Dio Chrysostom 2
Protrepticus 59 Diodorus Cronus 3, 44,72 n.l86, 76,93
Topics 92, 137 n.273
seealso 42,51,72n.l86 Diodotus 73 n.l96
Arrianus, Flavius 5 n.19, 24 n.6, 25 n.8, 28 Diogenes of Babylon 15, 72 n.l87, 73, 142
n.23 Diogenes the Cynic 25,56 n.l37, 60, 83,
Aspasius 49-50 85,90 n.264
Athenodorus of Tarsus 4 Diogenes ofOenoanda 6
Attalus 5 Dionysius ofCyrene 73
Atticus, T.Pomponius 42 Dionysius the renegade 47 n.IOI
Dionysius the rhetor 1-2
Balbus, L.Cornelius 15
Barea Soranus, Q.Marcius 5 n.19, 41 EgnatiusCeler,P. 5n.l9,41
Boethus (Peripatetic) 67 n.171 Empedocles 2 n.9, 51
Boethus (Stoic) 51 n.124 Ennius 7, 46
Epictetus
Carneades 119 n.322 on changing arguments 99, 100-101,
Chaeremon 4-5 105,
Charmadas 51 107, 118-121, 144
Chrysippus on his contemporaries 29, 33-35, 58, 66,
criticized 12, 70 74
and Epictetus 72 n.l86, 72 n.187 and cynicism 25
exegesis of 46, 48, 54-55, 73, 90
160 INDEXES

[Epictetus) Hippocrates 58
Discourses 24-25, 24 n.2, 24 n.6, 25 Hyginus, C.Julius 49 n.110
n.28, 36, 72 n.186, 75
lost parts of 24 n.6, 28 n.23, 28 n.24, Lucan 5
39 n.63, 57 n.138, 72, 141 Lucian 37, 41,50 n.l19
Encheiridion 24, 65, 75 Lucilius 12-13, 14-15
ethics 25, 27, 35, 55, 65,65 n.l66, Marcus Aurelius
72n.l86, 119-120 and Aristo 7 n.34
eulogized 42 n.l8 and books I, 45
on exegesis 43-45, 48, 52, 54-55 and Epictetus 24, 24 n.1, 34
on hypothetical reasoning 87, 90-92, and Galen 6
92 n.269, 94-98 and logic 1-2, II n.43, 13, 21,34 n.47
irony in 26, 27, 45, 59, 74,77 teachers of 5 n.19, 24
and logic Massalenus 24 n.6
his knowledge of 27-33,44-45,44 Musonius Rufus, C.
n.90, and Egnatius 41
59-60, 145 and Epictetus 27,72 n.187, 145
need for 59-62, 62-65, 65 n.l64 and logic 27,62-63,63 n.156, 65 n.165,
rejection of? 25, 38-40, 44-45, 55-58 71-72,71 n.184
in school syllabus 66-70,72-75 and practice 47 n.101
and logical puzzles 64, 76-77 pupils of 2 n. 7, 5 n.19
originality of? 34-35
and physics 25-27,43-44,43 n.86, 55, 63 Numenius 54
and practice 47-48
and pupils 35,39-40,48,48 n.l04, 66,
82, Pacuvius 41
90,90 n.263 Panaetius 51 n.124, 72 n.l87
on rhetoric 31-32, 55, 58 Panthoides 44, 72 n.186
style of 26, 27,28-29,30,39-40,56-57, Peregrinus 51 n.121
55 n.l35, 66, 73 Persaeus 36, 51
Epicurus 9, 42, 44 Persius 5
Euphrates 5, 72 n.l87 Philo the Megaric 32 n.38, 103-104
Plato
and Epictetus 56 n.136, 58, 68 n.175,
Favorinus 40 n.72, 49 n.l10, 72 n.l86 72 n.186, 144
Fronto, M.Comelius 1-2, 13 exegesis of 46, 49-50, 50 n.116, 51,
51 n.124, 52-53
Galen and logic 35 n.52
and Epictetus 72 n.l86 style of 36 n.57
and exegesis 49,49 n.114, 50 n.ll9, see also 23, 42
73 n.l99 Plautus, Sergi us 4, 4 n.14
and logic 6-7,11, 30, 33, 71, 75, Plotinus 49, 50, 52-53, 54
77 n.217, 81-82,81 n.231 Plutarch 11, 42 n.80, 45, 70
logical utilitarianism of 20, 21, 23, 64, Polemo 72 n.l86
and practice 47 Porphyry 48 n.105, 54
self-congratulations of 6 n.29, 42 Posidonius 15, 46, 51 n.124, 72 n.187, 73,
Gellius, Aulus 24 n.5, 35-37, 36 n.56, 49 140
n.IIO, 71, 96 n.285
Rusticus, Q.lunius 1, 5 n.19, 24
Heraclitus 47,50 n.l19, 51
Herillus 47 n.I01, 51 Seneca, L.Annaeus
Herodes Atticus 35-36 and epistemology 12 n.l
Herophilus 73 n.l99 on ethics 12, 16, 18,21-23
Hierocles 5, 73 n.200 and exegesis 45-46, 51, 52
and hypocrisy 41-42
 INDEX OF PERSONS 161

[Seneca] Socrates 11,26 n.11, 28-29,42,56,56

and liberal arts 19 n.20 n.l37, 60
and logic Sphaerus 51
knowledge of 12 Sti1o, L.Aelius 73 n.196
rejection of? 12-13,18 n.16, 19, 33,
38 Taurus, Cal venus 36 n.55, 49 n.1 08, 91
role of 20-21,23,62,65 n.267
and logical puzzles 14-15, 18,76 n.213, Thaumasius 49
77 .Theo of Alexandria 73 n.l99
'orthodoxy' of 25-26, 26 n.26, 54 Theophrastus 51, 121
physics 21-23, 27 Timon 96 n.283
and practice 47 n.101
and rhetoric 12 n.l Xenocrates 72 n.l86
and syllogisms of Zeno 15-18
Sextus Empiricus 92-93, 113-114, 126-127 Zeno (Stoic) 7, 42, 46, 51, 67 n.l72, 72
Silius !talicus 5 n.19 n.187,73
syllogisms of 15-18, 63 n.156
162 INDEXES

Index of Topics
The Greek alphabet precedes the Roman. Latin words in the Roman list are italicised.

UOT]AO~ II 0 n.312 Kataf. ll'l't~ 43

aKoJ.ou6ei:v 87 n.258 KUtT]yOptKO~ 85-86
aK6J.ou6o~ 87 n.258, 144 KA01t{j 54
aJ.uetv 64n.l59 KOt VffiVtKO~ 26
UjlEtam:roala 75 n.205
see also: change of mind AEKtOV 9
avaJ.uat~ 78 see also: sayables
see also: analysis J.oytK{] 7 n.33
avaf.UttK6~ 84-85 see also: philosophy, tripartition of
avaltOOEtKtO~ 80, 80 n.228 J.oyapta 55,55n.l35
see also: indemonstrables J.oyoKA01tta 54
a~trojla 9
see also: assertibles JlCxAtata 139
an6liEL~t~ 63 n.158, 78 n.221 JlUXECJ6at 91
see also: proof see also: conflict
iinopov 4 JlEOOtT]~ 142-143
aaKT]at~ 47n.l01 jlEtU1tt1ttEtV 75 n.205, 107
see also: practice see also: changing
aauvaKtO~ 28,28 n.24, 140-141
see also: validity v6Jlo~ 90 n.264
see also: logic, laws of
litaf.EKttK6~ 8, II n.43, I 07 n.305
see also: dialectic 1tapaKOAOU6EtV 61 n.J5]
litE/;WYJlEvOV 28, 28 n.23 to 1tapaAEt1tOjlEVOV 145
see also: disjunction napa l;roa11~ cprovfi~ 46
1tEpt tpo1t{] 31, 31 n.34
EiKfl 138 see also: self-refutation
EK6EttK6v 87-88, 87 n.259 npayjlatEia 138
EKKUAU1tttK6~ 93 1tpoaipEm~ 42 n.80, 72 n.l86
E~11YEta9at 43 n.86 np6f.n1J!t<; 93, 145
see also: exegesis 1tpoaf.niJit~ 103, 124
evapy{j~ 30 npoarono1tmta 27 n.l8
enayyEJ.la 29 n.29, 139 ntiiim~ 143
enayEt v 28, 32
ElttppT]jlU 143 a6q>tajla 14, 75 n.207
see also: adverbs see also: sophisms
eprotiiv 136-137 <JUjl1tE1tAYjlEvOV 28, 28 n.23
Eq>OOEUEt v 92-94 see also: conjunctions
eq>oOEUttK6~ 93-94, 93 n.270 auvayEtv 28, 28 n.24, 142
see also: validity
TJYEJlOVtKOV AfjjljlU 124 auvaKttK6~ 28, 141
see also validity
6Ejla 81 auvaptT]at~ I 03-104
see also: analysis see also: conditionals
9pt yK6~ 68 n.l75 CJUVT]jljlEVOV 28, 28 n.23
see also: conditionals
i.aoouvaJlEtv 31, 31 n.36 ac\>I;Etv 28 n.23
i.atopei:v 142
t61to~ 34-35
Ka6fjKOV 65 see also: philosophy, areas of
see also: appropriate tpOlttKOV 28, 81 n.233
Ka90AtKOV 28,30 tp6no~ 142
Kat VOtOjlta 54 see also: argument schemata
 INDEX OF TOPICS 163

u/..11 138 Baroco 71 n.184

i.>7t69ecn<; 29 n.27, 85 Camestres 121
see also: hypotheses Celarent 79, 79 n.224, 79 n.225
U7t09E'tt1CO<; 85-6 Cesare 78-79, 121
see also: hypothetica1s Darapti 121
fourth figure 121
q>9Eipecr9m 98 cavillatio 12, 14, 19
Q>Tlcri 27 change of cognitive status 112-114
change of mind 100-101
xpEia 136 changing arguments 34, 36, 55, 61, 69, 75,
see also: utilitarianism 99-125, 136-137
xpfim<; 61, 136 types of 105-106
changing assertibles 99-101, 103, 104-105,
wei>oecr9at 38, 142 137-138, 144
Christians 46, 51 n.121, 63 n.157
Academics 30,39 n.66, 45 n.92, 72 n.l86, coins, testing of 83-84, 139-140
75 n.207, 136 commentaries 43-44, 48, 48 n.l 05, 49-50,
see also: scepticism 52 n.l25
adverbs, temporal I 02, I 08, I 08 n.307, Ill Christian 51 n.l21
n.315, 142-143 Jewish 51 n.121
analysis 77-78 Stoic 50,51 n.124, 73,73 n.l99
analysis of syllogisms I, 39-40,43,57,63, see also: exegesis
74-75,77-83 conditionals 28, 29, 32, 32 n.38, 38, 65, 80
rules for 81-83 n.227, 91, 103-104, 137-138
anafytical syllogisms 83-84, 84 n.240 conflict 38-39, 91, 96, 104
animals 61 n.l51 conflict of opinion 66
appropriate 61, 65, 118, 144 coniunctio 28 n.23
arguments conjunctions 28,28 n.23, 29, 108-109, 113
definition of 9, 88-89, 102, 121, 125 conundrums: see puzzles
as events 120-125 conversion, syllogistic 79, 79 n.223
with false premisses 32-33, 117 criterion of truth 68
hyposyllogistic 31 n.36 Crocodile 65 n.163
and persuasion 16-17 Cynics 7, 9, 20, 25,47 n.JOI, 145
probative 93 Cyrenaics 7, 9-10, 20, 34 n.48
'by way of questioning' 34, 75, 136-137
single premiss 140-141 death 12-13,15,16-17
true 33, 93, 102 definition II n.43, 28-29, 141
unmethodical 84 n.241 demonstratives 98, 101-102, 137-138
wholly hypothetical 86 n.252 Denier 64, 76,76 n.210
see also: categorical, changing, dialectic, dialectic 8-9, I 0, 12 n.l, 16, 21, 28-29, 68
enthymeme, hypothetical, n.175,85
indemonstrables, games 29, 92, 92 n.269, 94, 95, 97 n.285,
inference,logic, proof, reduction 136-137
argument schemata 49, 78, 80-81,80 'Dio has died' 97-98
n.227, 80 n.229, 108 n.309, 141 disjunctions 28, 29, 80 n.227
ars vitae 2, 14, 55 doxography 45, 77
see also: philosophy
assertibles 9, 86-87,89-90,93, 101-102, enthymeme 58,58 n.142, 75
111 Epicureans 6, 7, 9-10, 10 n.40, 20, 30,39
impossible 98 n.66, 40 n.71, 50,72 n.186
perishing 98, 124 epistemology 9, 12 n.l, 94-96
see also: changing see also: proof
Eretrians 51
books 1, 43, 45-48 ethics 10-11,21-23,34,38,74, 119-120
categorical syllogisms 71-72, 71 n.l84, 78- practical 2, 4, 16, 19-20, 25
79,85-86, 121 see also: logic
Barbara 79
164 INDEXES

exegesis 43-55, 43 n.86, 77 logic

dangers of 52-55 charm of 1-2, 36-38, 58, 77
history of 51-52 dangers of 12-13
nature of 49-51 definition of 140
see also: commentaries and ethics 16, 18, 18 n.l6, 19, 35, 38-39,
exercises spiritue/s 47 n .I 0 I 64, 65, 69,118-119
expositives 87-88, 87 n.259 ofthe gods 3, 23,63 n.l56
hypothetical 88-92
first person plural 29, 38 n.62 laws of 90-91, 90 n.264, 97 n.285
further assumption I 03, 124 moral necessity of 60-62
nature of 140
gods 14 n.12, 22-23,43 n.86, 63 part or instrument? 20, 23, 67
see also: logic and physics 63, 69
good old days 52 n.l25, 74, 74 n.202 primacy of 66-67
grammar school 49,49 n.IIO rejection of 7-10, 12-13, 19-20,25,55
terminology of 4, 4 n.l4, 27-29, 28 n.23,
handbooks 73-74,73 n.200 81 n.233, 124, 141
Homed Man 14,65 n.l63, 76,76 n.212, 97 useless parts of 20, 63-64, 82
n.285 see also: arguments, categorical, dialectic,
hypocrisy 41-42, 74 n.202 hypothetical, Peripatetics
hypotheses 29, 61-62, 85-86, 92 n.269, 'the logicians' 7 n.35, 12, 37, 93 n.273,
94-96 107, 107 n.305
impossible 91
hypothetical arguments 34, 39-40, 43, 61- Master Argument 3, 35-36, 39, 44, 45, 72-
62, 69, 75, 78, 85-98, 86 n.251 73,72 n.l86, 76-77,97
rules for 90-92, 92 n.269, 96-97, 96 medical analogies 57 n.J39
n.285 metaphysics II, 19
hypothetical syllogisms 35, 71, 85-86 Mice 14, 76 n.213
hypotheticals 86-88, 87 n.259 modallogic 97-98

if see: conditionals nature 26, 43-44

immorality of philosophers 40-42 see also: physics
imperatives 87, 137 oaths 86 n.255
impressions 9, 43 originality 52-54
use of 43, 45,60-61,61 n.J51
indemonstrables 79-81, 80 n.227, 84, 84 parts of speech 29 n.30, 142-143
n.241, 91, 108, 108 n.309 Peripatetics 6,11,17-18,40n.71,49,50,
indexicals I 01-102 72 n.l86
ineptiae 12-13, 16, 19 on logic 20, 71-72,78-79, 89 n.261
inference 33, 69, I 09, 123-124 93 n.273, 121-122
from false premisses 32-33, 117 persuasion 16-17
rules of 90-91,115-116,120-121 philosophy
see also: arguments areas of 24 n.l, 34-35, 34 n.46, 38-39,
infinite regress 60 40,
inquit 27 n.l8 44, 64, 68-69, 139
interrogare 136 artoflife 2, 10-11,14,19,42,44,55
interrogatiuncu/a 55 n.135 grand questions of 12-13
hostility to 7
leaving home 39-40 tripartition of I, 8-9, 12 n.l, 19, 34
letters of recommendation 83, 85 34 n.47, 126-127
lex dia/ecticae 96 n.285 analogies for 21
see also: logic order of 67-70,67 n.l71
Liar 14, 34,34 n.43, 39, 64, 76,76 n.211, way of life 2 n.6
94 physics I, II, 21-23,24-27,43-44,55,63,
liberal arts 19 n.20 65,67
living voice 46 plagiarism 54
 INDEX OF TOPICS 165

Platonists 6, II, 49, 50 schools 40-41,48,49,49 n.IIO, 52 n.l25,

porcelain 64, 64 n.J60 58, 66, 68, 72, 82-83, 90, 90 n.263, 95,
postulates 95, 144 145
see also: hypotheses self-evidence 30
practice 47-48, 62, 74 self-refutation 30-31, 59-60
preconception 93, 145 sentences 99-1 00
premisses Sirens 36-37, 39
major and minor 124 n.329 sophisms 14, 16, 18-19,39,63,65 n.l64,
order of 121-122, 124-125 68, 69,75 n.207, 78,107
see also: arguments see also: puzzles
principles 94-95, 141 Sorites 36, 39, 76,76 n.211, 77 n.217, 119
progress 43, 74 n.322
proof 18 n.J5, 20, 38, 59, 63,63 n.156, stars,numberof 109-110, I!On.312
63 n.158, 66,68-69,78,92-94,93 n.273, Stoic books 56, 73-74
141 Stoic logicians 4-5, 72-74
and discovery 20, 69-70 Stoics, imperial 4-5, 5 n.l9
'not everything can be proved' 94-95, lost 5
94 n.280 pseudo 40 n. 71, 44, 44 n.88
propositions 99-100 style 36 n.57, 56, 58
categorical 71 syllogisms see: analysis, categorical,
hypothetical 71 conversion, hypothetical
universal 28, 30-31 symposia 35-36, 36 n.55, 38, 49
puzzles, logical 10, 14-16, 18-19,35-37, synonyms 31-32
39, 63, 64, 75, 76-77
see also: sophisms talk vs action 40-41
Pyrrhonists 6, 42, 45 n.92, 72 n.186, 75 thought, Jaws of 90 n.264, 119
n.207, 95-96, 127 tropics 28, 82 n.233
truth 9, 38,99-100, 100 n.291, 102
questions 136-137
unclear items II 0, II 0 n.312, 111-113
ratiuncula 15, 55 n.l35 utilitarianism, logical 20-21, 57, 62-65, 82
reductio ad impossibile 94, 94 n.279, 117-
118 validity 102-103, 107
reduction of arguments 78-81, 79 n.225 change in 102-103, 116
rhetoric 1-2,12 n.l, 31-32,36 n.57, 37, 49, viva vox 46 n.97
55, 56-58, 138
Waggons 76 n.213
the Sage 22-23, 62, 65, 77, 89, 95, 97, 98, weights and measures 7 n.35, 67-68, 67
119-120,145 n.J72
sayables 9, 86-87, 100 women 63
see also: assertibles
scepticism 6, 30, 42,45 n.92, 61,72 n.l86, young men Jove logic 13-14,35, 35 n.52,
75 n.207, 95-96, 96 n.283, 110 n.312, 127 37-38, 39-40
scholarship, mere 40-41,45-46,46 n.96
 PHILOSOPHIA ANTIQUA
A SERIES OF STUDIES ON ANCIENT PHILOSOPHY

EDITED BY

]. MANSFELD, D.T. RUNIA

ANDJ.C.M. VAN WINDEN

l. VERDENIUS, W.J. and WASZINK,]. H. Aristotle on Coming-to-Be and Passing-

Awqy. Some Comments. Reprint of the 2nd (1966) ed. 1968. ISBN 90 04 01718 6
7. SAFFREY, H. D. Le nepi l(J!Aoao({!{ar; d'Aristote et la thiorie platonicienne des idees
nomhres. 2eme ed. revue et accompagnee du compte-rendu critique par
H. Cherniss. 1971. ISBN 90 04 01720 8
13. NICOLAUS DAMASCENUS. On tlu Philosophy qf Aristotle. Fragments of the
First Five Books, Translated from the Syriac with an Introduction and Com-
mentary by H.J. Drossaart Lulofs. Reprint of the 1st (1965) ed. 1969.
ISBN 90 04 01725 9
14. EDELSTEIN, L. Plato's Seventh Letter. 1966. ISBN 90 04 01726 7
15. PORPHYRIUS. IIpor; MapKeA.A.av. Griechischer Text, herausgegeben, iiber-
setzt, eingeleitet und erklart von W. Potscher. 1969. ISBN 90 04 01727 5
17. GOULD, J. B. 7he Philosophy qf Chrysippus. Reprint 1971. ISBN 90 04 01729 1
18. BOEFT, J. DEN. Calcidius on Fate. His Doctrine and Sources. 1970.
ISBN 90 04 01730 5
19. POTSCHER, W. Strukturprobleme der aristotelischen und tluophrastischen Gottesvor-
stellung. 1970. ISBN 90 04 01731 3
20. BERTIER, J. Mnesithie et Dieuches. 1972. ISBN 90 04 03468 4
21. TIMAIOS LOKROS. Ober die Natur des Kosmos und der See/e. Kommentiert von
M. Baltes. 1972. ISBN 90 04 03344 0
22. GRAESER, A. Plotinus and tlu Stoics. A Preliminary Study. 1972.
ISBN 90 04 03345 9
23. IAMBUCHUS CHALCIDENSIS. In Platonis dialogos commentariorum .fragmenta.
Edited with Translation and Commentary by J.M. Dillon. 1973.
ISBN 90 04 035 78 8
24. TIMAEUS LOCRUS. De natura mundi et animae. Oberlieferung, Testimonia,
Text und Obersetzungvon W. Marg. Editio maior. 1972. ISBN 90 04 03505 2
26. GERSH, S. E. K{vryal<; aK{vryror;. A Study of Spiritual Motion in the Philos-
ophy ofProclus. 1973. ISBN 90 04 03784 5
27. O'MEARA, D. Structures hierarchiques dans la pensee de Plotin. Etude historique et
interpretative. 1975. ISBN 90 04 04372 1
28. TODD, R. B. Alexander qf Aphrodisias on tlu Stoic Physics. A Study of the De
Mixtione with Preliminary Essays, Text, Translation and Commentary. 1976.
ISBN 90 04 04402 7
29. SCHEFFEL, W. Aspekte der platonischen Kosmologie. Untersuchungen zum Dialog
'Timaios'. 1976. ISBN 90 04 04509 0
30. BALTES, M. Die Weltentstehung des platonischen Timaios nach den antiken Interpreten.
Teil l. 1976. ISBN 90 04 04720 4
31. ED LOW, R.B. Galen on lAnguage and Ambiguiry. An English Translation of Galen's
De Captionibus (On Fallacies), With Introduction, Text and Commentary. 1977.
ISBN 90 04 04869 3
34. EPIKTET. Vom Kynismus. Herausgegeben und iibersetzt mit einem Kommen-
tar von M. Billerbeck. 1978. ISBN 90 04 05 770 6
35. BALTES, M. Die Weltentstehung des platonischen Timaios nach den antiken Interpreten.
Teil 2. Proklos. 1979. ISBN 90 04 05799 4
37. O'BRIEN, D. Theories qf Weight in the Ancient World. Four Essays on Democritus,
Plato and Aristotle. A Study in the Development of Ideas 1. Democritus:
Weight and Size. An Exercise in the Reconstruction of Early Greek Philos-
ophy. 1981. ISBN 90 04 06134 7
39. TARAN, L. Speusippus qf Athens. A Critical Study with a Collection of the
Related Texts and Commentary. 1982. ISBN 90 04 06505 9
40. RIST, ]. M. Human Value. A Study in Ancient Philosophical Ethics. 1982.
ISBN 90 04 06757 4
41. O'BRIEN, D. Theories qfWeight in the Ancient World. Four Essays on Democritus,
Plato and Aristotle. A Study in the Development of Ideas 2. Plato: Weight
and Sensation. The Two Theories of the 'Timaeus'.l984. ISBN 90 04 06934 8
44. RUNIA, D. T. Philo qf Alexandria and the Timaeus qf Plato. 1986.
ISBN 90 04 07477 5
45. AUJOULAT, N. Le Neo-Platonisme Alexandrin: Hieroclis d'Alexandrie. Filiations
intellectuelles et spirituelles d'un neo-platonicien du Ve siecle. 1986.
ISBN 90 04 07510 0
46. KAL, V. On Intuition and Discurwe Reason in Aristotle. 1988. ISBN 90 04 08308 I
48. EVANGEUOU, CH. Aristotle's Categories and Porphyry. 1988.
ISBN 90 04 08538 6
49. BUSSANICH, ]. The One and Its Relation to Intellect in Plotinus. A Commentary
on Selected Texts. 1988. ISBN 90 04 08996 9
50. SIMPUCIUS. Commentaire sur les Categories. Traduction commentee sous Ia
direction de I. Hadot. I: Introduction, premiere partie (p. 1-9, 3 Kalb-
fleisch). Traduction de Ph. Hoffmann (avec Ia collaboration d'I. et P. Hadot).
Commentaire et notes a Ia traduction par I. Hadot avec des appendices de P.
Hadot et J.-P. Mahe. 1990. ISBN 90 04 09015 0
51. SIMPUCIUS. Commentaire sur les Categories. Traduction commentee sous Ia
direction de I. Hadot. III: Preambule aux Categories. Commentaire au
premier chapitre des Categories (p. 21-40, 13 Kalbfleisch). Traduction de
Ph. Hoffmann (avec Ia collaboration d'I. Hadot, P. Hadot et C. Luna). Com-
mentaire et notes a Ia traduction par C. Luna. 1990. ISBN 90 04 09016 9
52. MAGEE, J. Boethius on Signification and Mind. 1989. ISBN 90 04 09096 7
53. BOS, E. P. and MEIJER, P. A. (eds.) On Proclus and His Irifluence in Medieval
Philosophy. 1992. ISBN 90 04 09429 6
54. FORTENBAUGH, W. W., et a!. (eds.) Theophrastes qf Eresos. Sources for His
Life, Writings, Thought and Influence. 1992. ISBN 90 04 09440 7 set
55. SHANKMAN, A. Aristotle's De insomniis. A Commentary. ISBN 90 04 09476 8
56. MANSFELD,]. Heresiograpfry in Context. Hippolytos' Elenchos as a Source for
Greek Philosophy. 1992. ISBN 90 04 09616 7
57. O'BRIEN, D. Theodicee plotinienne, theodicee gnostique. 1993. ISBN 90 04 09618 3
58. BAXTER, T. M.S. The Cra!Jlus. Plato's Critique of Naming. 1992.
ISBN 90 04 09597 7
59. DORANDI, T .. (Hrsg.) Theodor Gomper;::. Eine Auswahl herkulanischer kleiner Schrif
ten (1864-1909). 1993. ISBN 90 04 09819 4
60. FILODEMO. Storia deijilosofi. La stoa da Zenone a Pane;::io (PHerc.l018). Edizione,
traduzione e commento a cura di T. Dorandi. 1994. ISBN 90 04 09963 8
61. MANSFELD, ]. Prolegomena. Questions to be Settled Before the Study of an
Author, or a Text. 1994. ISBN 90 04 10084 9
62. FLANNERY, S.J., K.L. Ways into the Logic qf Alexander qf Aphrodisias. 1995.
ISBN 90 04 09998 0
63. LAKMANN, M.-L. Der Platoniker Tauros in der Darstellung des Aulus Gellius. 1995.
ISBN 90 04 I 0096 2
64. SHARPLES, R.W. Theophrastus qf Eresus. Sources for his Life, Writings, Thought
and Influence. Commentary Volume 5. Sources on Biology (Human Physiol-
ogy, Living Creatures, Botany: Texts 328-435). 1995. ISBN 90 04 10174 8
65. ALGRA, K. Concepts qf Space in Greek Thought. 1995. ISBN 90 04 I 0172 I
66. SIMPUCIUS. Commentaire sur le manuel d'Epictete. Introduction et edition critique
de texte grec par Ilsetraut Hadot. 1995. ISBN 90 04 09772 4
67. CLEARY, J.J. Aristotle and Mathematics. Aporetic Method in Cosmology and
Metaphysics. 1995. ISBN 90 04 10159 4
68. TIELEMAN, T. Galen and Chrysippus on the Soul. Argument and Refutation in the
De Placitis Books II-III. 1996. ISBN 90 04 10520 4
69. HAAS, F.A.J. DE. John Philoponus' New Definition of Prime Matter. Aspects of its
Background in Neoplatonism and the Ancient Commentary Tradition. 1997.
ISBN 90 04 10446 1
70. HABETS, A.C.J. A History of the Division of Philllsophy in Antiqui!)'. 1997.
ISBN 90 04 10578 6
71. ANDIA, Y. DE. Henosis. L'Union aDieu chez Denys l'Areopagite. 1996.
ISBN 90 04 10656 1
72. ALGRA, K.A., HORST, P.W. VANDER, and RUNIA, D.T. (eds.) Pof:yhistor.
Studies in the History and Historiography of Ancient Philosophy. Presented to
Jaap Mansfeld on his Sixtieth Birthday. 1996. ISBN 90 04 10417 8
73. MANSFELD,]. and RUNIA, D. T. Aeliana. The Method and Intellectual Con-
text of a Doxographer. Volume 1: The Sources. 1997. ISBN 90 04 10580 8
74. SLOMKOWSKI, P. Aristotle's Topics. 1997. ISBN 90 04 10757 6
75. BARNES, ]. Logic and the Imperial Stoa. 1997. ISBN 90 04 10828 9