Alexander of

Aphrodisias
On Aristotle Prior
Analytics 1.1-7

Translated by
Jonathan Barnes, Susanne Bozien,
Kevin Flannery S. J.,
Katerina lerodiakonou

B L O O M S B U R Y
LONDON NEW DELHI NEW YORK SYDNEY
 Bloomsbury Academic
An imprint of Bloomsbury Publishing Pic

50 Bedford Square 1385 Broadway

London NewYork
WC1B3DP NY 10018
UK USA

www.bloomsbury.com

Bloomsbury is a registered trade mark of Bloomsbury Publishing Pic

First published in 1991 by Gerald Duckworth & Co. Ltd.

Paperback edition first published 2014

Jonathan Barnes, Susanne Bobzien, Kevin Flannery and Katerina lerodiakonou, 1991

The translators have asserted their right under the Copyright, Designs and
Patents Act, 1988, to be identified as Author of this work.

All rights reserved. No part of this publication may be reproduced or transmitted in

any form or by any means, electronic or mechanical, including photocopying,
recording, or any information storage or retrieval system, without prior
permission in writing from the publishers.

No responsibility for loss caused to any individual or organization acting on or

refraining from action as a result of the material in this publication can be
accepted by Bloomsbury Academic or the author.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

ISBN HB: 978-0-7156-2347-3

PB: 978-1-7809-3453-2
ePDF: 978-1-7809-3454-9

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress.

Acknowledgements
The present translations have been made possible by generous and imaginative funding
from the following resources: the National Endowment for the Humanities, Division of
Research Programs, an independent federal agency of the USA; the LeverhulmeTrust;
the British Academy; the Jowett Copyright Trustees; the Royal Society (UK); Centra
Internazionale A. Beltrame di Storia della Spazio e delTempo (Padua); Mario Mignucci;
Liverpool University.The editor is grateful to Ian Crystal and Paul Opperman for their help
in preparing the volumes for press.

Typeset by Derek Doyle & Associates, Mold, Clwyd.

Printed and bound in Great Britain
 Contents

Preface vii

Introduction 1
Alexander of Aphrodisias 1
Commentaries 4
Alexander, On the Prior Analytics 7
The translation 14
Logical terminology 17

Translation 35
1. Preface 41
2. Introductory explanations 53
3. Conversions 83
4. The First Figure 101
5. The Second Figure 138
6. The Third Figure 166
7. Further reflections 184

Appendix 1. Translation of Prior Analytics 1.1-7 198

Appendix 2. Aristotelian Syllogistic 208
Appendix 3. Textual Notes 216
Texts and Abbreviations 219
Bibliography 224
Indexes
English-Greek Glossary 230
Greek-English Index 234
Index of Passages 241
Index of Persons 248
Index of Subjects 250
 Preface

This book is the work of a collaborative partnership. Each partner is

equally responsible for everything in it. The work was done in a
series of weekly meetings held in Balliol during 1988 and 1989. We
were from time to time joined by Antonina Alberti, Barbara Cassin,
Jim Hankinson, Anna Maria loppolo, James Irvine, Mario Mignucci
and Voula Tsouna, to each of whom we owe valuable suggestions.
And we are indebted to Ian Mueller for a sheaf of written comments.

Oxford JB
August 1990 SB
KF
KI
 Introduction

Alexander of Aphrodisias
Alexander addressed his essay On Fate to the Emperors Septimius
Severus and Caracalla. The address dates the work between the
years 198 and 209 AD.1 Alexander explains that in the essay he will
present and discuss Aristotle's views on fate and human freedom;
for

I am the champion of his philosophy, having been nominated a

teacher of it on your recommendation. (Fat. 164,14-15)

The Emperors had thus helped Alexander to a Chair in

Aristotelian Philosophy. It is a plausible guess - but no more - that
this was one of the four philosophy chairs which had been
established at Athens some twenty-five years earlier by Marcus
Aurelius.2 (If the guess is lucky, then Alexander will have taught at
Athens.3)
Alexander, we may reasonably infer, was an eminent Aristotelian
who nourished at the beginning of the third century AD. This
conclusion is consistent with the other evidence, scanty and indirect,
which we have about his life and activity - for apart from the
passage in On Fate we hear nothing detailed and datable about his
career.4 We do not even know for sure where he came from: he was
1
See Fat. 164,1-3. Caracalla was made Augustus on 28 January 198; in 209 Geta
became Augustus to make an Imperial trio. It is reasonable - but not perhaps
absolutely mandatory - to date On Fate to the period between the elevation of
Caracalla and the elevation of Geta. See further Thillet, pp. LXXV-LXXIX.
2
The year was 176: see e.g. Cassius Dio, LXXH xxxi 3; cf. e.g. Oliver, pp. 80-5;
Lynch, pp. 169-77; Glucker, pp. 146-50.
3
Arabic sources actually state that Alexander taught at Athens; but they derive
from a passage in Galen (Anat. Admin. II 218K) which refers to Alexander of
Damascus - and there is no reason to identify the Damascene with the Aphrodisian
(see e.g. Thillet, p. XLI). At in Metapfi. 415,29-31, Alexander refers, in an illustrative
example, to 'this statue - say, of Aristotle - which stands in Athens'; but this hardly
implies that Alexander was lecturing in Athens at the time. Note that there is
nothing to be said for the old view that Alexander was the official Peripatetic
scholarch at Athens: see Lynch, p. 214. On Alexander's 'school' see Sharpies (1990).
4
For discussion and bibliography see Thillet, pp. VI-LXXIII; Sharpies (1987), pp.
1177-8.

1
2 Introduction
Alexander 'of Aphrodisias', but more than one town went by that
name and it is at best a reasonable conjecture that he was born in
the rich and flourishing Aphrodisias in Caria.5 As for his personal
features and foibles, we learn nothing - unless we may believe that
he was 'lean and pale'.6
His teachers included Sosigenes and Herminus - whom he
criticises at some length in his commentary on the Prior Analytics.1
We know little about these scholars. Their pupil emerged as a
learned philosopher; but we cannot tell to what extent and in what
ways the pupil's ideas were shaped and formed by his teachers.
His writings contain critical and polemical passages, many of
them directed against the Stoics.8 It is easy to suppose that he was a
controversialist, engaged in living dispute with his contemporaries.
There is indeed evidence that he broke a lance or two with Galen,
his senior by some years and a leading intellectual figure of the day.9
Yet it may be suspected that some, at least, of Alexander's attacks
were bookish - delivered in the calm of the study or the lecture-room
against dead or hypothetical opponents.10
Bookish he certainly was. A dozen major works survive, in
addition to several collections of shorter pieces. And much more has
been lost than survives. Although Alexander did not rival the
productive energy of Galen, his collected works would have occupied
an impressive amount of shelf space.11 Moreover, they had some
influence:12 on his contemporaries (such as Galen), on the later
Greek philosophers (on Plotinus13 and on the Neoplatonic
commentators), and also on Arabic philosophy.14
5
On Aphrodisias see e.g. Reynolds. The name 'Alexander' is found on Aphrodisian
inscriptions (e.g. Monumenta Asiae Minoris Antigua VIII 414, 574); but it is an
extremely common name.
6
See [Alexander], in Metaph. 531,25-6 (cf. 532,7-19) -but this part of in Metaph. is
not from Alexander's own feather.
7
See Moraux (1973/84), II, pp. 335-60 (Sosigenes) and pp. 361-98 (Herminus). On
the controversial figure of Aristotle of Mytilene see Moraux (1973/84), II, pp. 399-425;
Accattino; Goulet, pp. 411-12; and esp. Thillet, pp. XI-XIX.
8
See Todd (1976); Sharpies (1987), p. 1178.
9
For the evidence see esp. Thillet, pp. XXXII-XLIX; cf. von Muller, pp. 424-5;
Pines; Zimmermann, p. Ixxxi n. 2; Nutton (1984), pp. 318-24; Nutton (1987), pp.
45-51; Sharpies (1987), p. 1179.
10
The evidence for living and contemporary debate in Alexander's school is sifted
by Sharpies (1990), pp. 90-2,104-10.
11
For a general characterization and a full catalogue see Sharpies (1987), pp.
1179-99; cf. Thillet, pp. LII-LXXIII; Goulet, pp. 128-39.
12
See, in general, Cranz, pp. 79-82; Sharpies (1987), pp. 1220-4.
13
See Porphyry, Vit. Plot. 14; and the Index Fontium to the Henry-Schwyzer
edition of Plotinus. Note that Eusebius cites Alexander (Fat.) at PE VI ix.
14
See esp. Maroth. Many of Alexander's writings were translated into Arabic, and
some have survived only in their Arabic versions: see Badawi (1987), pp. 109-14;
Strohmaier; Sharpies (1987), pp. 1187-8, 1192-4. For Alexander's influence on
mediaeval thought see Ebbesen (1981b), (1982), (1990).
 Introduction 3
15
In philosophy, he was thoroughly Aristotelian. Indeed, his
philosophical activity can be roughly characterised as an attempt to
organise and present a comprehensive, consistent and contemporary
Aristotelianism. This being so, it is unsurprising to discover that
much of his writing took the form of commentaries on Aristotle's
works; and although there is a substantial body of material, always in
Aristotelian vein, which is less closely tied to the words and works of
the Master, the Aristotelian commentaries constitute the core of his
philosophical achievement.
He wrote commentaries on all the constituent works of Aristotle's
Organon or logical corpus: the commentaries on the Categories, On
Interpretation, Prior Analytics 2, Posterior Analytics,16 and the
Sophistical Refutations11 are lost. But in addition to the work on
Prior Analytics 1, we possess a substantial commentary on the
Topics.18 There were also independent books on logical matters: an
essay on the conversion of propositions,19 and an essay on utterances,
both preserved in Arabic versions;20 several brief discussions on
questions to do with the concept of possibility, found among the
so-called Quaestiones and closely related to Alexander's interest in
determinism;21 a lost essay on hypothetical syllogistic;22 a one-
volume work on syllogisms with modally mixed premisses, and some
Scholia Logica or Logical Notes, both referred to in the commentary
on the Prior Analytics and both lost.23
The chronology of Alexander's writings cannot be established.24
The only absolute date is given by the opening of On Fate - and even
that is not precise. For the internal or relative chronology, we must
rely on occasional cross-references. Thus we know that the essay on
conversions, and perhaps the commentary on the Topics, were

15
For a sketch see Donini, pp. 220-48; Sharpies (1987), pp. 1199-220; the account
in Zeller, III i, pp. 817-31, is still worth reading.
16
For fragments, testimonies and discussion see Moraux (1979).
17
The commentary preserved under his name and published in CAG II 3 is a much
later work: see Ebbesen (1981a), I, pp. 242-4; Sharpies (1987), p. 1184.
18
Edited by Wallies in CAG II2.
19
Arabic text in Badawi (1971); referred to below by 'Alexander, Conu.'.
20
See Sharpies (1987), p. 1193. An Arabic source refers to a second, shorter,
commentary on Are. Pr.: see Badawi (1987), p. 113.
21
See Sharpies (1982), (1983).
22
Known to Avicenna: see Mardth, pp. 7,139.
23
For the peri mixeon see in An. Pr. 125,30; 127,16; 207,35; 213,26; 238,37; 249,38
(cf. Philoponus, in An. Pr. 126,20-3); for the Scholia see 250,2 (cf. Sharpies (1987),
p. 1196). Zeller, III i, p. 820 n. 2, asserts that the reference to the Scholia Logica at in
An. Pr. 250,2 is a later gloss; but he offers no reason for this opinion. Perhaps there
was also an essay on affirmations: see our note to in An. Pr. 36,6 below. Note also the
indeterminate cross-references at in An. Pr. 188,16; 191,17; 284,17; 390,9; and the
prolepticreferences atinAn. Pr. 70,20; 110,20; 193,21; 328,6.
24
See Sharpies (1987), p. 1181.
4 Introduction
written after the commentary on the Prior Analytics;25 and we know
that the commentary on the Prior Analytics had itself been preceded
by two shorter writings on logical matters. Such results are meagre
and unsatisfying. But in any case, chronological questions are
probably of little significance.26 For the commentaries presumably
represent the substance of Alexander's lectures on Aristotle;27 the
lectures were presumably delivered year in year out; and the passing
years presumably brought revisions of various sorts. In such circum-
stances it makes little sense to speak of a relative chronology: in a
way, we might rather suppose that Alexander wrote his commen-
taries concurrently, and perhaps over a decade or more.

Commentaries
Numerous ancient commentaries on Aristotle have survived. And
many more have perished. Alexander's commentaries are, by
general consent, among the best: modern scholars praise him - and
so did some of the ancients.28 But he was not the first in the field: on
the contrary, he was writing in a tradition already some centuries
old.
Aristotle's immediate successors, Theophrastus and Eudemus,
wrote on the same subjects as their master had done; and, to some
extent at least, they conceived of their task as correcting, expanding
and completing Aristotle's work.29 Their works can no doubt be

26
See Conu. p. 60; in Top. 7,11. (But note that cross-references in the
commentaries are primarily concerned with the order in which the readers or the
audience will read or hear the works - they do not necessarily provide evidence for
the order of composition.)
26
The same may be said for the chronology of some of Aristotle's writings: see e.g.
Barnes (1981).
27
But note the doubts expressed by Sharpies (1990), p. 97.
28
See e.g. Simplicius, in Phys. 80,15-16; 795,33-5; in DA 52,27-8. It is regularly
asserted that Alexander was later known simply and honorifically as 'The
Commentator'. The passages adduced to support this claim include: Simplicius, in
Phys. 707,33; 1170,13; 1176,32; Philoponus, in An. Pr. 126,21; Olympiodorus, in
Meteor. 263,21. In all these texts Alexander is indeed referred to by phrases such as
'the commentator' or 'Aristotle's commentator'; but these phrases are not honorific
titles - they are ordinary referring expressions. (If, at the end of a book review, you
read The author deserves our thanks', you will rightly take this for praise - but you
will not think that the author has been honoured as The Author par excellence.)
Alexander is also said to have been called 'the second Aristotle'; but the two relevant
texts ('Ellas' [i.e. David], in Cat. 128,13; Syrianus, in Metaph. 100,6) are peculiarly
difficult (and 'Elias' is certainly corrupt): for opposing views see Moraux (1973/84), II,
p. 401; Thillet, pp. XX-XXXI. Note, too, that Alexander was often criticised (e.g.
Plutarch of Athens, apud Philoponus, in DA. 21,20-3; Philoponus, in An. Post.
111,31-112,6) and sometimes accused of bias (e.g. 'Elias', in Cat. 123,4-8).
29
See Boethius' comment on Theophrastus at in Int. ed. sec. 12,3-16; cf. Barnes
(1985a) for an attempt to reconstruct one such Theophrastean supplement.
 Introduction 5
30
called interpretations or exegeses of Aristotle, in a loose sense; but
there is no reason to believe that they wrote commentaries on his
writings;31 and it seems unlikely that any Aristotelian work
received a commentary before the renascence of Aristotelian studies
in the late first century BC.32 By then the tradition of literary
commentaries, which had been consolidated in Alexandria in the
third century BC,33 was well established, and the hupomnema or
commentary was a recognised genre.34 There were already
commentaries on scientific texts - namely, on the medical writings
in the Hippocratic corpus.35 Perhaps, too, there were commentaries
on philosophical texts, and in particular on Platonic texts. According
to Proclus, Grantor was 'the first interpreter of Plato';36 and whether
or not Proclus means that Grantor wrote commentaries on Plato,37 it
is likely that such commentaries were written during Grantor's
life-time - for two papyri dating from the third century BC preserve
small fragments of what seems to have been a commentary on the
Phaedo.38 The earliest philosophical commentary of which we can
form any general impression is the anonymous commentary on the
Theaetetus, partly preserved on papyrus.39 The papyrus itself is
dated to the first century AD, but the text it contains may go back to
30
Compare the way in which later Epicureans interpreted the works of their
master: see esp. Sedley; Puglia.
31
Pace Geffcken, pp. 406-7: Geffcken's study is a useful survey of the antecedents
of philosophical commentaries; but he fails to distinguish essays in interpretation
from commentaries proper.
32
On this renascence see most recently Moraux (1973/84), I, pp. 1-94; Gottschalk
(1987), pp. 1083-97.
33
See e.g. Pfeiffer, pp. 210-33; Fraser, pp. 447-79; Turner, pp. 112-24; and esp. del
Fabbro, who gives lucid accounts of the forms and varieties of early commentaries
(pp. 70-2; 93-5), of commentatorial style (pp. 97-100), and of the sorts of topics which a
commentary standardly covered (pp. 106-23).
34
For the history of the term see e.g. Bomer.
35
See Fraser, pp. 363-7. Our MSS of Galen ascribe a commentary on Hippocrates'
Aphorisms to Herophilus (in Hipp. Aph. XVIIIA 186-7K); but the text should be
emended to ascribe the commentary to Herophilus' pupil, Bacchius: see now von
Staden, pp. 75-6.
36
See in Tim. I 75,30-76,10 (cf. II 276,31-277,14).
37
See most recently Dorrie, pp. 328-38. The use of the word exegesis does not in
itself imply a commentary: thus we should not suppose that the pleistoi ... hosoi
exegentai the book of Heraclitus, and who included the shadowy Antisthenes (not,
pace Geffcken, p. 399, the celebrated Antisthenes but the Herakleiteios mentioned at
Diogenes Laertius, VI19) and Heraclides of Pontus (Diogenes Laertius, IX 15), wrote
commentaries (even though Heraclides' work was entitled Herakleitou exegeseis:
Diogenes Laertius, V 88).
38
Namely PMon 91 [Carlini (1986), pp. 10-14] and PHeid 28 [Carlini (1978), pp.
201-9]; see also Carlini (1975).
39
See Diels-Schubart; Praechter (1909b) (a new text is being prepared for the
Corpus dei Papiri Filosofici). Note that the anonymus had also written commentaries
on the Timaeus (XXXV 10-12), on the Phaedo (XLVIII 7-11), and on the Symposium
(LXX10-12).
6 Introduction
40
the first century BC. There are scattered references to other
commentatorial activity on Platonic texts.41
The earliest surviving commentary on Aristotle is Aspasius'
commentary on the Nicomachean Ethics.42 Aspasius was a gener-
ation senior to Alexander. Earlier Peripatetics had written earlier
commentaries: Andronicus, Boethus and Ariston in the first century
BC; Sotion, Apollonius, Achaicus and Alexander of Aegae in the first
century AD.43 Of these scholars, Boethus and Ariston are known to
have written about syllogistic;44 and Boethus had indeed proposed
substantial modifications to Aristotle's theories.45 But it is not
explicitly said that they wrote commentaries on the Prior Analytics.
Sotion wrote a commentary on the Topics; and a papyrus preserves
part of a commentary on the Topics dating from the first century AD.46
Contemporary with Aspasius were Adrastus, who was perhaps
Alexander's compatriot,47 and Alexander's teachers, of whom Hermi-
nus almost certainly had written a commentary on the Prior
Analytics.48 In addition, thinkers outside the Peripatetic school had
addressed themselves to Aristotelian logic; and the two earliest
logical handbooks to have survived Galen's Introduction to

40
So Tarrant; but see Glucker (1989).
41
See e.g. Diogenes Laertius, III 65-6; POxy 1609 (a reference to a commentary on
the Timaeus); Suetonius, apud Eustathius, in Od. I 107; Porphyry, Vit. Plot. 14;
Proclus, in Rep. II 96. On Albinus' commentaries see Diels-Schubart, pp.
XXVIII-XXX; on Adrastus' commentary on the Timaeus see Moraux (1973/84), II, pp.
296-313; on Posidonius' alleged commentary on the Timaeus see Kidd, pp. 338-40 (on
PGen 203 see Burkert (1987)). In general see Gudeman, cols. 687-91; Untersteiner,
pp. 205-22; Dillon, pp. 54-7; Westerink (1976), I, pp. 7-19.
42
Edited by Heylbut in CAG XIX1 (see Mercken; cf. Goulet, pp. 635-6).
43
See esp. Moraux (1973/84), I, pp. 97-142 (Andronicus), 143-80 (Boethus), 181-96
(Ariston); II, pp. 211-14 (Sotion), 216-17 (Apollonius), 217-21 (Achaicus), 222-5
(Alexander of Aegae); also Moraux (1986b); Gottschalk (1987), pp. 1097-121 (who also
refers to Platonists, such as Eudorus, and to Stoics, such as Athenodorus, who
discussed but did not compose commentaries upon Aristotle's logical ideas).
44
For Ariston see esp. Mariotti, pp. 59-74.
46
Notably in his thesis that all categorical syllogisms are 'perfect': Themistius,
Max. 190-1; Ammonius, in An. Pr. 31,11-32,7; scholiast to Aristotle, 156b43-157b9
(see our note to 23,14 below). Note too Boethus' claim that hypothetical syllogistic is
prior to categorical: Galen, Inst. Log. vii 2.
46
PFayum 3: see Moraux (1973/84), I, pp. 215-16.
47
For his presence in the surviving commentary on Books II-V of EN (edited by
Heylbut in CAG XX) see Moraux (1973/84), II, pp. 323-30; Gottschalk (1987), p. 1155;
Mercken; Goulet, pp. 56-7. Note too the fragmentary commentary on GC preserved in
'Ocellus Lucanus', 20-35: the date of this work is uncertain, but it may be as early as
the first century BC (see Moraux (1973/84), II, p. 606; (1986b), pp. 133-4).
48
For commentaries on An. Pr. see also Ebbesen (1981b); Lee, pp. 9-12, 36, 128.
(POxy 3320 may possibly come from an early commentary on An. Pr.) For early
commentaries on Cat. see Galen, Lib. Prop. XIX 42K; Simplicius, in Cat. 1,3-2,29;
159,23-33 (cf. Moraux (1973/84), II, pp. 519-27; Gottschalk (1987), pp. 1101-12); for
commentaries on de Int. see Zimmermann, pp. Ixxx-xcii.
 Introduction 1
49 50
Dialectic and the On Interpretation ascribed to Apuleius both
contain accounts of Peripatetic syllogistic.51
Alexander thus had at his disposal a wealth of scholarly work on
Aristotelian logic in general and on the Prior Analytics in particular;
and at least some of this work had been expressed in the form of
commentaries. Alexander rarely names his predecessors.52 But it is
clear from his own commentary that he was thoroughly familiar with
their work. It is clear, too, that the tradition had produced interpreta-
tions of considerable detail and ingenuity. And it is clear that the
tradition was not uniform or univocal: Alexander reports deviant
interpretations; and he also reports the views of logicians who wished
to modify or even reject some parts of Aristotle's logical system.
Alexander's commentary was influential in its turn. Two later
commentaries on the Prior Analytics survive: one, truncated, by
Ammonius,53 and the other by Ammonius' pupil, John Philoponus.54
Ammonius and Philoponus were intimate with Alexander's work,
and each draws upon it. Indeed, there are substantial passages in
Philoponus' commentary which paraphrase, with or without acknow-
ledgment, the corresponding passages in Alexander's work. The two
later commentators were Platonists, and they did not share
Alexander's general philosophical stance. But their differences from
Alexander in matters of logic were relatively trifling; and their
commentaries provide us with interesting, and sometimes illuminat-
ing, parallels to Alexander's discussion.55
Alexander, On the Prior Analytics
The later commentators were conservative creatures. They worked
to a standard form; and they borrowed shamelessly - or perhaps
49
Note also Galen's commentaries on An. Pr., An. Post, and Cat.: Lib. Prop. XIX
47K (see Moraux (1973/84), II, pp. 688-91).
50
For the latest surveys of the long-standing controversy over the authorship of
this work see Hijmans, pp. 408-11; J.-M. Flamand, in Goulet, pp. 298-317.
51
Note also the brief description in Albinus, Didasc. 158-64H.
52
In the first part of in An. Pr. Herminus is the only commentator named (72,27;
89,34; 91,21); but note the several anonymous references to exegetes and critics of
Aristotle: 20,31; 23,3; 65,18; 68,21; 75,11; 81,17; 95,29; 100,17. Sometimes other
evidence allows us to name these people (see e.g. our note to 95,29; and compare in
An. Pr. 125,3-6 with [Ammonius], in An. Pr. 39,31).
53
Edited by Wallies in CAGIV 6.
54
Edited by Wallies in CAG XIII 2: the commentary was taken from Ammonius'
seminars on An. Pr. There is also an Armenian version of David's commentary on An.
Pr., and a fragment of the commentary by Elias (see Westerink (1961)). We perhaps
have references to lost commentaries by Porphyry (see Ammonius, in An. Pr. 31,15),
lamblichus (Philoponus, in An. Pr. 26,5), Themistius ([Ammonius], in An. Pr. 39,2),
Syrianus ([Ammonius], in An. Pr. 39,2), Eutocius (Elias, in An. Pr. 134 Westerink),
Proclus ([Ammonius], in An. Pr. 43,30-1), Marinus (Leo Magentinus: Ebbesen
(1981b), p. 10). For Byzantine commentaries see Benakis.
55
On the substantial issues see Lee, passim. We draw attention to the more
interesting parallels in our notes to the translation.
8 Introduction
reverentially - from their predecessors.56 Moreover, they had
reflected upon the obligations and duties of a commentator, and they
had established certain canons of procedure.57 Earlier scholars had
also thought about the matter. Thus Galen:

Before the detailed interpretation, you should learn in general what

the function of any interpretation is, namely to render clear whatever
is unclear in the work. To prove that what is written is true, or to
refute it as false, or to defend it against sophistical criticisms - these
are not part of interpretation, but they have become standard tasks
for pretty well everyone who writes a commentary. Now there is
absolutely no reason why an interpreter should not touch lightly on
these matters; but a thorough examination of the author's doctrines
falls outside the boundary of interpretation, (in Hipp. Fract. XVIIIB
318-19K)

No doubt such elementary ideas had presented themselves to

Alexander; and if the later canons had not been fixed by his time,
most of their prescriptions are in fact observed by him - he will, for
example, discuss such preliminary issues as the purpose of the work
before him, its utility, the significance of its title. In any case, the
requirements for a commentary - and especially for a commentary
on a logical text - are relatively obvious and uncontroversial; and
Alexander's works no doubt followed, more or less closely, an
established pattern.
The commentary on the Prior Analytics begins with a Preface, on
the nature of the subject-matter and on the name and scope of
Aristotle's treatise. Then follows the commentary proper, divided
into sections by 'lemmata' or citations from Aristotle's text.58 The
intensity of comment is uneven: some Aristotelian passages receive
minute scrutiny, others are ignored. And the unevennesses do not
invariably correspond to the differing degrees of difficulty which a
modern reader will find in the text: what an ancient commentator
considered interesting is not always the same as what a modern
reader finds perplexing. The nature of the comments is diverse:
66
For the later commentatorial tradition see e.g. Praechter (1909a), pp. 526-33;
Richard; Sorabji (1990a); Westerink (1990).
57
The canons were codified by Proclus (see 'Elias', in Cat. 107,24-5; cf. esp.
Westerink (1990)); but they can be traced back to the third century (see I. Hadot) and
perhaps even earlier (see e.g. Moraux (1986b), pp. 134-5). For their description of the
role of the commentator see e.g. Simplicius, in Cat. 7,23-32; 'Elias', in Cat.
122,25-123,11.
58
On lemmata in ancient commentaries see Turner, pp. 114-15; del Fabbro, p. 78.
Ross, p. 91, says that 'it is agreed among scholars that the lemmata were written not
by the commentators but by copyists'. If Ross means that the commentators supplied
no lemmata at all, he is wrong. If he means that the lemmata in our MSS never
represent the lemmata written by the commentators themselves, he is at best half
right (cf. Lamberz, pp. 7-15): see below, p. 17.
 Introduction 9
59
there are textual notes; there are explications of points of
language; there is much explanatory paraphrase of Aristotle's
argument; there are short disquisitions interpolated into the
commentary, where Aristotle has hinted at an issue without
discussing it or where other philosophers have ventured dubious
opinions.
Usually Alexander makes his own judgment plain. Occasionally
he sets out the different possibilities without indicating a
preference. From time to time he refers explicitly to the views of
other scholars - other Peripatetics or members of other schools - but
almost invariably in order to criticize and refute their views. On the
other hand, he rarely offers a substantive criticism of Aristotle
himself:60 he often finds Aristotle obscure and in need of explanation
(why else would he be writing a commentary?61); but, like any good
commentator, he implicitly adopts a 'Principle of Charity' and
strives to find an interpretation of the text which will place it on the
side of the truth.62
For whom was Alexander's commentary written? What is its style
and tone? What is its value, as a commentary and as a contribution
to logical studies?
The commentary is not a literary production, written for a large
public. Rather, as we have said, it may be supposed to represent the
lectures which Alexander gave to his students, perhaps in his
capacity as Professor of Aristotelian Philosophy. Lectures are
usually given more than once, and the incorporation of later
revisions will often leave the text disjointed and uneven. In
Alexander's commentary there are some odd transitions and some
strange non sequiturs.63 They are no doubt to be explained by the
nature of the work. An early comment was later judged to be
inadequate in some fashion; further notes were written and added
to the text; but the new coat showed its seams, and the original
59
See Moraux (1986b), pp. 134-7; note Galen's remarks on this subject at in Hipp.
Off. XVIIIB 630-2K.
60
The only serious exception to this in the part we have translated is at 116,29-35;
but note that Alexander's criticism here must be read in the light of his claim that
Aristotle, in the passage in question, intends to train or exercise his readers
(gumnasion: 113,10). In later parts of the commentary there are more substantial
criticisms, notably in connexion with 'mixed' modal syllogisms; see e.g. in An. Pr.
213,11-27; 214,12-18; 236,8-11; 238,22-38; 240,32-241,9; 249,25-32; 270,6-8.
61
Note that the question 'Why is Aristotle obscure?' was one of the preliminary
topics prescribed by the Procline canon (above, n. 57).
62
No doubt some commentators strove too hard: the Procline canon gravely insists
that 'the interpreter must not force and strain and announce that everything said by
the old author whom he is interpreting is true. He must always repeat: "The man is
my friend, but so too is truth; and while both are friends, truth is the more so"'
CElias", in Cat. 122,32-123,1): cf. Taran. Alexander sometimes forced and strained:
note his remark at DA 2,4-9.
63
See e.g. 116,36 note; 118,10 note.
10 Introduction
material was neither cut out nor refashioned to make the text cohere.
The lecture audience was expected to have a text of Aristotle's Prior
Analytics to hand, and to be prepared to give it a close reading, with
meticulous attention both to dubious exegetical points and to
substantive matters of logical theory. Students would already have
studied the Categories and On Interpretation. It is presupposed that
they have a general competence in Peripatetic philosophy and in its
terminology. (Thus Alexander can use, without explanation, such
technical or semi-technical terms as hule and tropos.64) It is also
presupposed that they have an elementary logical training.65
As for the style, you may call it, unkindly, professorial.66 Alexander
makes no attempt to write 'good Greek', nor does he enliven his
argument with stylistic tricks and tropes. (Contrast the writings of
his contemporary Galen: Galen's works, even the most technical, are
composed in a rich and flowing style, and they are full of rhetorical
embellishments and polemical tours de force. Contrast, too, the later
Greek commentators who are usually fluent and sometimes elegant.)
Alexander's vocabulary is limited. He avoids variation. He is frequent-
ly repetitive. There are no light touches and no jokes.
His syntax sometimes slips. More notably, his sentence construc-
tion is at the same time lax and tortuous. Sentences meander across
the page, flowing languidly toward a finite verb and sometimes
seeping away before they find one. 'Anacolouthon' is the polite - or the
technical - term for such things; but Alexander's anacoloutha are not
conscious rhetorical tricks, designed to heighten the emotional tone.
Alexander's subject is logic, and it might be said in his defence that
logic does not lend itself to simple syntax or to flowery prose.
Logicians must be spare, and they must be explicit: we should expect
long sentences and we should not expect pleasing embellishments.
But Alexander's style is no model for the logician: it achieves
explicitness only at the cost of a tedious opacity. The long sentences
are difficult to construe. The language hangs like a dense fog over the
logical terrain.67 Compare Alexander with another of his contem-
poraries, Sextus Empiricus. Sextus' plain prose possesses a limpidity
and a precision even when he is discussing matters of considerable
logical complexity.
If the style of Alexander's commentary does not commend itself to
the candid reader, what of the substance? As a commentary on
64
See 6,17; 26,15.
65
Alexander will sometimes refer to a separate monograph for a fuller discussion of
technical issues: see esp. in An. Pr. 238,22-38; 249,15-250,2.
66
On commentatorial style see esp. del Fabbro, pp. 97-100, who justly describes it
as 'dry and devoid of any stylistic embellishment' (p. 97). Del Fabbro thinks that the
style reflects the serious and sober aim of the commentator (p. 100); Turner (p. 113)
supposes rather that it reflects the oral origin of the commentaries.
67
A prize example at 72,26-74,7 - which is, syntactically, a single sentence.
 Introduction 11
Aristotle, the work is not without merit. Aristotle's Prior Analytics
is densely written, particularly in its early formal chapters. The text
is sometimes allusive, often obscure. Alexander has an eye for the
difficulties, and he will often make them plain. His paraphrastic
comments are especially useful in this respect; for their aim is
simply to state expansively what Aristotle had stated in rude and
crabbed form. Sometimes Alexander's interpretations are ques-
tionable; and on occasion they seem contrived and implausible. But
all in all readers of Aristotle will gain from consulting Alexander -
and they will gain more from him than they will gain from many a
later commentary.
In addition, Alexander will often elaborate on issues towards
which Aristotle merely gestures. These little essays, which
sometimes draw on later Peripatetic theories and sometimes involve
polemic against the enemies of Aristotelianism, enrich the content
of the commentary. They afford historical information which we
cannot get, or cannot get so well, from other sources.68 They also
collectively delineate Aristotelian syllogistic at the midpoint of its
historical development.69 The conception of logic which manifests
itself in Alexander's pages is in some respects different from
Aristotle's - and different from our own. Consider, for example,
Alexander's interest in the 'generation' of the syllogistic figures;70 or
his concern with the proper 'ranking" of the figures, and of the
syllogisms within the figures.71 Such issues will strike modern
logicians as utterly alien to logic: for Alexander, they have the same
interest and the same standing as such 'genuinely' logical topics as
the reduction or analysis of syllogisms. No doubt Alexander was
wrong, his conception of logic muddled and inadequate. But it was a
conception of considerable historical influence and importance; and
that in itself gives the commentary a peculiar interest.
Alexander was not an original logician. There is no evidence that
he made any substantial logical discoveries or inventions of his own.
There may have been a few terminological innovations;72 some of
the arguments which he deploys against various of Aristotle's critics
he may have excogitated himself; and he may have contributed
68
See e.g. the material on 'unmethodically concluding' arguments at 21,10-22,30;
or the account of Herminus' views on major and minor terms at 72,17-75,34.
69
On this aspect of the commentary see esp. Lee. Lee's monograph contains the
best and most substantial modern treatment of Alexander's logic, but it does not
purport to be comprehensive. Of older treatments, Prantl (see pp. 620-6) offers
nothing of value, and Volait is mostly jejune.
70
See 47,27 note.
71
See 47,22 note; 51,8 note.
72
Thus the words suzugia and sumploke, in their logical senses, meet us for the
first time in Alexander's commentary. But Alexander gives no hint that he is
neologizing, and we may well suppose that he is using an established vocabulary.
12 Introduction
something to the method of 'reducing' all syllogisms to the two first
syllogisms of the first figure.73 However that may be, it is plain that
he is essentially a presenter, not a producer, of logical ideas. (Here,
again, he contrasts with Galen, who made - and was acutely
conscious of having made - no small contribution to logical
theory.74) That in itself is no disgrace. But it must be confessed that
Alexander's presentation of logical ideas is less than perfect. Not
only, as we have said, is his style at odds with the needs of his
subject: there is also evidence that he sometimes failed to
understand what he presented. One example of his failure is worth
elaborating; for it runs through the whole commentary and it may
puzzle an unwary reader.75
In his syllogistic, Aristotle was concerned to show not only that
certain pairs of propositions yield a syllogistic conclusion but also
that certain other pairs do not. He proves that pairs of propositions
are 'non-syllogistic' by providing counterexamples. Thus he shows
that no syllogistic conclusion follows from a pair of propositions of
the form
A holds of no B, B holds of no C
in the following way.76
First, it will be proved that you cannot infer a conclusion of the
form
A does not hold of some C.
Take the terms: science, line, medicine. No line is a science; no
medicine is a line; but every medicine is a science. (The example,
which is Aristotle's own, may sound strange: read 'medicine' as
'branch of medicine'.) Hence three propositions of the forms
A holds of no B, B holds of no C, A holds of every C
may all be true together. Hence a pair of propositions of the form
A holds of no B, B holds of no C
do not yield a conclusion of the form
A does not hold of some C.
73
See 115,18-116,36: the state of the text suggests (but does not prove) that this
passage was a later addition to Alexander's original text, and it may be the case that
it represents a later discovery by Alexander himself. But equally, of course,
Alexander may have read about the method later in a work by some other logician.
There is in general much in Alexander which we find in no earlier source - but we
possess virtually no relevant earlier sources. The text of the commentary rarely gives
us any reason to suspect originality; and since Alexander never claims originality, it
seems probable that he rarely is original.
74
See e.g. Barnes (1985b).
75
See Lukasiewicz, pp. 67-8; Patzig, pp. 169-72; Barnes (1990a), III 7. The
misunderstanding is not proprietary to Alexander: it can be paralleled in most of the
ancient commentators (see e.g. Apuleius, Int. 186,5; 194,23; Philoponus, in An. Pr.
34,7-10; 74,30-75,15 and very often) and in some of the moderns; it seems to
appear from time to time in Aristotle himself (e.g. An. Pr. 38a29-31, bl8-20; 39b3-6).
76
See An. Pr. 26a9-12 (Alexander's comments are at ire An. Pr. 57,7-18).
 Introduction 13
Another triad of terms shows in the same way that a pair of
propositions of the form in question will not yield a conclusion of the
form
A holds of some C.
Now, since the pair does not yield
A does not hold of some C,
it does not yield
A holds of no C;
and since it does not yield
A holds of some C,
it does not yield
A holds of every C.
(For 'A holds of every C' entails 'A holds of some C'. Hence if the pair
yielded the former, it would also yield the latter. Similarly for 'A
holds of no C' and 'A does not hold of some C'.)
Hence no syllogistic proposition predicating A of C can be deduced
from the pair in question. Hence the pair is 'non-syllogistic'.
Alexander frequently discusses this method of proving that a pair
of propositions is non-syllogistic. He always misunderstands it.77 He
takes it that - in thefirstexample we cited from Aristotle -
Every medicine is a science
follows from the pair of propositions
No line is a science, No medicine is a line.
Thus he takes it that
A does not hold of some C
does not always follow from
A holds of no B, B holds of no C
because
A holds of every C
sometimes follows. This is Alexander's own view of the logical
structure of Aristotle's method; and he supposes that he is simply
following Aristotle's view.
It is clear that something has gone seriously wrong. Alexander is
in effect confusing the relation of being compatible with with the
relation of following from. The method of counterexample in fact
employs the relation of being compatible with: that is the relation
which the method requires, and it is the relation which Aristotle's
examples actually exhibit. Alexander speaks instead of the relation
of following from: the method does not and cannot employ this
relation; and Aristotle's examples evidently do not exhibit it.
It is hard to have much faith in a logician who can make such an
elementary error; and it must be allowed that Alexander was not a
77
He may seem to get it right at in An. Pr. 101,14-16 and 328,10-30; but in these
passages it seems reasonable to think that he has succeeded by mistake.
14 Introduction
first-rate logician. Nonetheless, the second-rate are often worth
reading. Even Alexander's blunders have their interest: logic is not
an easy subject; it is salutary to see how readily bad mistakes are
made; and one way of avoiding error yourself is to observe the errors
of others.78 Moreover, the errors in Alexander's commentary are
balanced by a decent weight of truth. And in any event, the
commentary has (as we have indicated) an unparalleled value as an
historical document; for we possess no other contemporary logical
text of comparable length and detail.

The translation
Book 1 of the Prior Analytics divides into three main sections: first,
in 1.1-7, Aristotle introduces his syllogistic theory and expounds the
non-modal syllogisms; then, in 1.8-22, he discusses modal
syllogisms; and thirdly, in 1.23-46, he offers a loosely organized
series of metalogical observations. The first part can sensibly be
read on its own; for, apart from the discussion of modal conversions
in 1.3, it is entirely self-contained. Alexander's commentary also,
and unsurprisingly, divides into three parts. This translation is of
the first part - that is to say, the commentary on 1.1-7. This
accounts for about one third of the total bulk of Alexander's work.
We have endeavoured to do Alexander into plain and intelligible
English. Our readers, we suppose, will be primarily interested in the
history of logic and in Aristotelian syllogistic; and the translation is
designed for such readers. It will be of little interest to anyone
whose main concern is with Alexander's prose style - not that such
people are common.
Our English is sometimes clumsy or ambiguous or obscure. In
principle these failings are deliberate: they reflect failings in
Alexander's Greek, failings which an English reader should be
aware of. We have not, however, attempted to match Alexander's
stylistic inelegancies with analogous inelegancies of our own: mere
bad writing - as opposed to bad thinking - has been effaced. (But we
have no doubt introduced some bad writing of our own.) In
particular, we have consciously striven to break down Alexander's
long and contorted periods into ordinary English sentences. This
betrays Alexander's style, but it does service to his argument. And
Alexander's style merits no loyalty.
We have added a few things of our own. Occasionally we have
interpolated numerals to indicate the division of an argument into
its parts. These numerals have no counterpart in the Greek, and we
have added them only where the train of thought seemed
78
See in An. Pr. 8,19-29.
 Introduction 15
disagreeably obscure without them. The Chapter and Section
headings are also our addition. They too have no counterpart in the
Greek, which is broken only by the Aristotelian lemmata.79 But they
should not be misleading; and they may serve any reader who
wishes to scan through the pages looking for matters of particular
interest - and also any hardened spirit who may determine to read
the book from cover to cover.
As far as vocabulary is concerned, we have not affected a rigorous
purism. Different Greek words are often given the same English
translation if they are synonyms or used synonymously. Conversely,
a single Greek word will get different English renderings if it is
ambiguous - or even if normal English idiom requires or suggests a
variation. (It should be unnecessary to state these elementary
principles of translation; but in recent years a foolish purism, based
on a false conception of fidelity, has become fashionable.) Where
logical terminology is involved, we have, however, been a little more
scrupulous. Here we have allowed ourselves a few barbarisms; for
here it is - or it may be - important to know that Alexander used the
same word in this sentence as in that. The next section of the
Introduction says something about Alexander's logical terminology
and our attempts to translate it. The Greek-English Index and the
English-Greek Glossary offer a more extensive view of our linguistic
decisions.
It will surprise some readers that we have altogether avoided
logical symbolism. There is a temptation to employ modern logical
notation (or at least the standard symbols deployed in modern
treatments of Aristotle's syllogistic) in the translation, or at any rate
in the notes. For symbolism is both concise and precise. But
Alexander himself was no symbolic logician. Like Aristotle, he uses
a technical and semi-technical vocabulary. But he writes in Greek,
and the 'mathematical' aspect of modern logic is entirely foreign to
him (as indeed it is to Aristotle himself).80 Hence (and also for other
reasons) we have dispensed altogether with symbols. We write 'A
holds of every B' rather than 'AaB'. We write 'If P, then Q' rather
than'P-^Q'.
The annotations are mere notes: they do not, and are not intended
to, amount to a commentary (or metacommentary). They are of five
kinds. (1) They provide precise references and cross-references to
ancient texts which Alexander cites or alludes to. Where Alexander
simply remarks "X says' (or 'I have said'), we pedantically supply the
appropriate numbers. Sometimes Alexander's allusions are covert
79
On our treatment of the lemmata see below, p. 17.
80
Alexander follows Aristotle in using letters (A, B, C; M, N, O; P, R, S) to express
syllogistic forms; but these letters are not variables in the modern sense (see note to
53,30).
16 Introduction
(for he knew Aristotle and Plato backwards and his language often
echoes them): we have sometimes made such allusions explicit - but
we cannot pretend to have hunted all of them down. (2) The notes
also record parallel passages in other ancient texts: in Alexander's
other works; in the later commentaries of Ammonius and
Philoponus; and occasionally in other authors. These records, again,
make no claim to completeness; rather, they suggest starting-points
to anyone who wishes to look further into the ancient history of the
issues which Alexander discusses. (3) There are references to the
modern literature. This literature is not in fact vast; for Alexander's
logic has not been a favourite topic among modern scholars. Even so,
there is more than we have looked at, and our notes are selective. (4)
From time to time, Alexander's arguments appear to us to be
obscure and difficult to follow. In such cases we have provided a
paraphrase: some of the paraphrases are no more than tentative
attempts to articulate one possible interpretation of a puzzling text;
and although we have (of course) picked the interpretation which we
think is least bad, we do not suppose that our interpretation is the
only interpretation worth considering. (5) Finally, there are some
textual notes - but they require a fresh paragraph.
The only critical edition of Alexander's commentary on the Prior
Analytics is the text published by Maximilian Wallies in volume II1
of the Berlin series of Commentaria in Aristotelem Graeca. Wallies
relied mainly on three manuscripts (BLM) and on the early Aldine
edition, which he took to represent an inferior manuscript tradition.
There are several manuscripts which Wallies did not consult; and
some of those which he did consult he consulted briskly: a full
critical edition of the commentary would have demanded far more
time than the Berlin Academy allowed its editors.81 And a properly
scholarly edition is yet to be done. (We have looked idly at a few
pages in a late Florentine manuscript;82 this has, unsurprisingly,
brought only one small improvement to Wallies' text - and Wallies'
error was perhaps due to a minor lapse at the proof-reading stage.83)
We have ventured to depart from Wallies' text at some eighty
places. The departures are signalled in notes of the fifth sort.
81
For an enthusiastic account of the origins and the progress of CAG see Usener,
pp. 199-202. Wallies signed the Preface to in An. Pr. on 1 September 1883: the Berlin
Academy had given its formal approval to the CAG project on 9 May 1878.
82
Laurent LXXII11, which Wallies calls C.
83
At 17,4 Wallies prints e en tei entolei e en tei klesei; and his critical apparatus
records no variants. Laurent LXXII 11 has e en tei erotesei after entolei. It is likely
that this is the correct text (for it is hard to imagine a scribe's being moved to
interpolate the words). But we should not infer that the Laurentian MS uniquely
preserves Alexander's text. In fact, the same text is printed in the Aldine edition; and
we suspect that the words in question are to be found in all the MSS - they have been
lost from Wallies' text by some nineteenth-century oversight.
 Introduction 17
(These textual notes are collected in Appendix 3.)
As a matter of fact, Alexander's text seems to be in pretty good
order; and we do not suppose that a new critical edition would often
differ substantially from Wallies. But there is one point where
some uncertainty might be felt. It concerns the lemmata or snippets
of Aristotelian text which punctuate the commentary. There are two
problems here. First, the lemmata run to different lengths in
different manuscripts: it seems likely that different copyists (or their
paymasters) chose to copy more or less of the Aristotelian text; and it
is not clear that we know how extensive the lemmata were in
Alexander's own copy of his commentary. Secondly, the proper
wording of the lemmata is sometimes a matter of doubt. Alexander's
text of the Prior Analytics often differs (usually in trifling particulars)
from more recent texts;84 and copyists would sometimes take their
lemmata from a manuscript of Aristotle rather than from the
manuscript of Alexander which they were reproducing. (This is
proved by the fact that the text of a lemma will sometimes disagree
with the text cited in the course of the commentary.85) Thus we
cannot always be sure that the text which we read in the lemmata
was the text which Alexander himself had before him.
In the translation we have adopted the following policy: in every
case, we have translated all of the lemma which Wallies prints; but
we have often continued the citations beyond the text in Wallies, and
we have occasionally interpolated new lemmata on our own ini-
tiative. All such additions are clearly marked, being enclosed by
angle brackets. They are made, solely for the convenience of the
modern reader, when Alexander's subsequent comments bear speci-
fically on a part of the text which does not appear in the lemma which
Wallies prints.

Logical terminology86
Alexander had at his service a rich and developed logical

84
See notes to Appendix 1.
85
There is a trivial example of this in the part of in An. Pr, which we have
translated: at Are. Pr. 26al5 our texts of Aristotle read ontos te, and so does the lemma
at in An. Pr. 58,24 (except that the Aldine edition prints tou for te); when Alexander
cites the same line of Aristotle in the body of the commentary, he writes ontos ge
(58,9: the manuscripts havege, the Aldine has nothing - and Wallies prints te). For a
more interesting case compare in An. Pr. 282,14-15 (the lemma, agreeing with our
text of Aristotle) with 283,3 (and 284,20.29), which cites a variant text. Note that the
lemmata in the Aldine edition have often (but not invariably) been written to fit the
current text of Aristotle (e.g. 9,4; 13,27; 23,20 - further references in the notes to
Appendix 1).
86
The Greek-English Index contains references to discussions in this section.
18 Introduction
87
vocabulary. Many of his Greek terms are difficult to translate; and
any translation is certain to break some of the semantic connexions
which link the different Greek terms. In this section we survey
Alexander's logical vocabulary, stating and sometimes explaining
our choice of translation. We do not discuss every word which might
be deemed logical in a broad sense of the term; and we are primarily
interested in Alexander's syllogistic vocabulary.88 The section is
written as a continuous narrative: we begin with a few terms of very
general application; then we turn to arguments and syllogisms;
thirdly, we deal with words connected with propositions and their
interrelations; and finally we discuss the terminology for the parts
and internal structure of propositions.89
Alexander insists that the logician should attend not to words but
to what words mean;90 nonetheless, the objects of his study are, of
course, essentially expressed in words: phone, which we normally
translate 'word' and lexis, which is normally 'expression', pick out
parts of language - words, phrases or sentences. These linguistic
items have a semantic force - they mean something.
Alexander's semantic repertoire includes the following words:
semainein (with semantikos and the associated semeion); deloun
(with delotikos); menuein (and menutikos); endeiknunai. It is
initially tempting to suppose that these different words must have
different senses, and so to establish different English versions for
them. But there is no evidence of any subtle semantic distinctions
among the words, whether in Alexander or in any other Greek
author. (Greek philosophers were scarcely interested in theories of
meaning; Greek grammarians did not develop any semantic
theory.91) Accordingly, we have been content to use 'mean' as a
standard translation for the first three of these four verbs. For the
fourth, endeiknunai, we do reserve a special translation, 'indicate';
but we do so only because the word is used elsewhere in a technical
sense, and an English reader might be interested to see how
Alexander uses it in a non-technical way. (In addition, the four verbs
may have a slightly different meaning in certain contexts. Thus
deloun means 'make delos or clear';92 and 'make clear' may
87
Little work has been done on Alexander's style and language; but see Todd
(1974).
88
For a brief account of syllogistic see Appendix 2.
89
But a desire to group together words from the same stem has occasionally
interfered with this general scheme. 90 See 84,15-19 and note.
91
The grammarian Apollonius Dyscolus has an extensive semantic vocabulary (see
e.g. the opening section of Conj.); but he makes no systematic distinctions among the
words he uses, all of which can usually be translated by the one word 'mean'.
92
'Clear' is always delos: 'obvious' renders enargSs, 'plain' gives saphes, and
'evident' translates phaneros. The Greek words are in fact synonyms. But it is useful
to signal occurrences of phaneros, since the word has a vital role in the account of
 Introduction 19
sometimes carry the sense not of 'mean', i.e. 'make clear in
language', but rather of 'betoken' or 'show'. Similarly, menuein may
be 'mark'; andsemainein is sometimes 'signify'.)
Two different lexeis may be said isodunamein or ison dunasthai
they may be equal in power or capacity to one another. We have
used the English 'equivalent' for this. (The etymology of'equivalent'
is exactly that of isodunamos.) A modern philosopher will ask
whether isodunamein amounts to synonymy (sameness of meaning),
or whether some weaker equivalence is intended.93 The question
cannot be answered: the important distinction between synonymy
and logical equivalence seems not to have been observed by any
ancient logician.
Lexis comes from legein, which is one of several verbs we translate
by 'say'. (Sometimes 'mean' is better, and we have used it.) Legein
also breeds logos, a word often cited as a paradigm of the
untranslatable. It is true that no one English word will pick out the
range of meaning of logos. But that does not make logos a puzzling,
let alone a metaphysically exciting, term. The central idea is
perfectly simple: a logos is what you say. There is no noun in English
which means precisely 'what is said'. We have followed custom and
taken 'account' as our normal rendering. But there are contexts in
which this translation would be pointlessly misleading: first, the
logos in question is sometimes an argument; and secondly, it is
sometimes a mere utterance. In these cases the word logos does not
mean 'argument' or 'utterance'. But clarity and the exigencies of
English demand that we translate logos by 'argument' or by
'utterance'. Sometimes a translator should be content to preserve
reference and let sense go.
Logoi which are arguments are the subject of Aristotle's work. The
kind of argument with which he is concerned in the Prior Analytics
is taken by Alexander to be a species ofpistis.9* Apistis is a way of
warranting or justifying something; and we use 'justification' for
pistis, with 'justify' and 'justified' for the cognate verbs (pisteuein,
pistousthai) and adjective (pistos). (The adjective pistos sometimes
rather has the force of 'justifiable', 'warrantable'. More generally, to
call something pistos is sometimes to say that you are entitled to
believe it. But in the commentary 'justified' will normally do, and we
have stuck by it.)
Arguments yield conclusions. Alexander most often says that an
argument sunagei its conclusion. We translate sunagein by
'perfect' syllogisms.
93
Note that there are also very weak uses of the word: e.g. Philoponus, in An. Pr.
63,12; 98,6 (cf. Ammonius, in Int. 114,22-3; 122,20-26; 246,9; 257,30-258,4; and note
Apuleius,/ret 181,6).
94
See 43,11 note.
20 Introduction
'deduce'. ('Infer' would do as well, but we keep this for another verb,
epipherein, where it is used as a synonym for sunagein.) A sunagoge
is a deducing or, more naturally, a deduction; and we use
'deduction'.
An argument will also perainein, or more often sumperainein,
something. We use 'conclude' for both verbs. And the noun,
sumperasma, is 'conclusion'. The conclusion hepetai or akolouthei
the premisses: we use 'follow (from)' for both verbs - which both
carry the spatial and temporal sense of 'follow' (i.e. 'come after') in
non-logical contexts. Akolouthia is the cognate noun: there is no
good English noun from 'follow', so we use 'implication' for
akolouthia and lose the transparency of its connexion with
akolouthein.
The arguments in which Alexander is interested can be
systematized or subsumed under general rules: there is a methodos
- a method - in their case; and they contrast in some passages with
arguments said amethodosperainein, to conclude unmethodically.
An argument, or an arguer, may establish and prove, or refute
and disprove. 'Disprove' gives diaballein and 'refute' elenkhein?5
(An elenkhos is a refutation.) 'Establish' is kataskeuazein. Trove' is
normally deiknunai - and the associated noun, deixis, comes out as
'proof. In ordinary Greek, deiknunai means, quite generally, 'show';
and in some Alexandrian contexts it is clear that 'show', rather than
the more specific 'prove', is required. You might use 'show'
throughout for deiknunai; but then, since 'show' has no appropriate
noun, the connexion between deixis and deiknunai, self-evident in
Greek, will be lost in the English version. Hence we use both 'prove'
and 'show' for deiknunai, as the context suggests. (In using 'prove'
we are again attending to the reference rather than to the sense of
the Greek verb.) The adjective deiktikos is sometimes 'probative';
but it is more often paraphrased by way of'prove'.96
Apodeiknunai always has the specific sense of'prove' rather than
the generic sense of 'show'. Since we have arrogated 'prove' for
deiknunai, we take 'demonstrate' for apodeiknunai. The translation
is well established in the tradition. 'Demonstration' and 'demon-
strative' are used for the cognates apodeixis and apodeiktikos. An
argument may be demonstrative, in which case it is an argument
which functions as a proof and it stands in contrast to dialectical or
sophistical (dialektikos, sophistikos) arguments.97 Again, Alexander
frequently speaks of demonstrating an argument, i.e. of proving it to
95
But where Greek idiom uses elenkhein English will sometimes demand 'prove'
rather than 'refute': elenkhein to asullogiston is 'to prove to be non-syllogistic'.
96
There is an odd Aristotelian use of the adjective which Alexander explains at
112,13-16 (cf. in An. Pr. 256,11-14).
97
See 7,8 note.
 Introduction 21
be valid. By contrast, certain arguments, which are taken as basic,
are called anapodeiktoi. We translate this by 'indemonstrable'. The
word was used by the Stoics of the basic arguments in their logical
system, and it is generally supposed that the Peripatetics came to
adopt the Stoic terminology. We are told that the Stoics used the
word anapodeiktos in the sense of 'not needing demonstration';98
and it is likely that it should be construed in the same way in
Alexander. But 'indemonstrable' is established as the orthodox
translation in Stoic texts; and in any case there is no convenient
alternative.
The special sort ofpisteis to which the Prior Analytics is devoted
are called sullogismoi. Some translators of Aristotle render the word
as 'deduction' - in the sense in which deductions contrast with
inductions. There is much to be said for the rendering. (Aristotle,
and Alexander after him, contrasts sullogismoi with epagogai or
inductions.99) But there is also one fatal objection: Alexander,
following Aristotle, recognizes the existence of deductive arguments
which are not sullogismoi. This point is not a casual aside: it is
central to Alexander's understanding of sullogismoi.100 Hence we
prefer to align ourselves with an older practice, and to transliterate
rather than translate: sullogismoi are syllogisms.
The noun carries a verb, sullogizesthai. The word 'syllogize' is
ugly, but it exists. Similarly, sullogistikos is 'syllogistic'. But note
that 'syllogistic' in Alexander has two different uses. First, and
generally, it means 'concerned with syllogisms', so that syllogistic is
the science of syllogisms - just as arithmetic is the science of
arithmoi or numbers. (Sullogistike in effect means 'logic', and is
more or less synonymous with logike and dialektike. But note that in
some authors and contexts these three terms are sharply distinct.)
Secondly, Alexander will regularly describe a pair of premisses as
syllogistic: in this sense, to say of something that it is syllogistic is to
say that it can syllogize something, i.e. to say that it will yield a
conclusion by a syllogism.101 'Non-syllogistic' or asullogistos is the
opposite of'syllogistic' in this second use. A synonym of asullogistos
is adokimos, 'unreliable'.
Some syllogisms are teleioi or 'perfect' (in the sense of 'complete',
'lacking nothing1): the term is technical - roughly speaking, it picks
out a class of syllogisms which are basic inasmuch as their validity
is evident. Other syllogisms are ateleis or imperfect. These can,
98
See e.g. Sextus, PHII156; Af VIII 223; Galen, Inst. Log. viii 1; cf. Frede (1974a),
pp. 127-9.
99
And also with 'paradigms',paradeigmata: see 43,11 note.
100
See Barnes (1990a).
101
Note too the technical term huposullogistikos, which we give as 'subsyl-
logisticaP.
22 Introduction
however, be perfected - the verb is teleioun or epitelein. To perfect a
syllogism is in effect to prove it valid: Alexander also uses
apodeiknunai in these contexts, as we have already noted; and like
Aristotle he will speak of analysing or reducing an imperfect
syllogism into or to a perfect syllogism.
'Analyse' translates - or rather transliterates - analuein. There is
a cognate noun, analusis; and an adjective, analutikos, from which
Aristotle's work took its title. Alexander has a short essay on the
notion of analusis.102
'Reduce' represents both anagein and apagein; and 'reduction' is
anagoge or apagoge (the latter noun being by far the more common,
and always occurring in the phrase 'reduction to the impossible'). In
some texts there may be a difference between anagein and apagein;
but we discern none in Alexander and hence do not scruple to use a
single English word.
There are various ways in which syllogisms may be perfected,
analysed, or reduced; each of them involves certain operations - to
which we shall return - on the syllogism or on its component parts.
For a syllogism is a compound item - it is composed of a
conclusion and a pair of premisses. Each of these three components
is a proposition. Where the word 'proposition' occurs in our
translation it represents protasis. (Andproteinein, the cognate verb,
comes out as 'propound'.) A proposition is not a linguistic item nor
yet a psychological event - it is neither a sentence nor an uttering of
a sentence. Rather, it is what you propound when you utter a
sentence of a certain sort. Alexander notes103 that the word protasis
also has a more specific sense: it is used for a premiss of an
argument i.e. it refers specifically to what you propound when you
put forward propositions from which a conclusion is to be deduced.
For this sense of protasis we reserve the word 'premiss'. It is not
always clear when protasis means 'premiss' rather than 'proposi-
tion'; and sometimes our translation depends on the toss of a coin.
But since both 'proposition' and 'premiss' are exclusively used for
protasis, the reader will always know when the Greek text presents
the word protasis .104
Propositions, orprotaseis in the general sense, are true (or false),
alethes (orpseudes). Equivalently, they 'hold' (huparkhein) or fail to
102 103
See 7,11-33 and notes. See 44,19-21.
104
One particular point deserves notice. Alexander frequently speaks of pairs of
protaseis from which no syllogistic conclusion can be deduced. It is natural to suppose
that in such passages protasis means 'proposition' rather than 'premiss'; for it is odd
to speak of premisses which generate no conclusion. Nonetheless, it emerges that in
Alexander's view these protaseis are indeed premisses - premisses which yield now
one conclusion and now another (see above, pp. 12-13). Hence we have supposed that
'premiss' is the correct translation in all these texts.
 Introduction 23
hold. The verb huparkhein is also used with two argument-places,
when the arguments are terms rather than propositions. For A
huparkhei toi B we say 'A holds of B'.105 (A huparkhei toi B, or A
holds of B, just in case A is truly predicated of B. Animal holds of
man: man is an animal.)106 Again, propositions (whether true or
false) are either simple or complex. Alexander occasionally mentions
complex propositions, for which his general name is 'hypothetical
(hupothetikos) proposition'.107 Thus he sometimes refers to
conditionals (sunemmena), which the Peripatetics called 'conti-
nuous' (sunekhe), and to disjunctions (diezeugmena), which the
Peripatetics called 'disjoint' (diairetika). But Aristotelian syllogisms
contain only simple propositions, and we may forget about complex
propositions. From now on, when we refer to propositions we shall
have non-complex propositions in mind.
Alongside protasis, in the general sense, we find apophansis. We
use 'assertion' for apophansis, 'assert' for apophainein, 'assertoric'
for apophantikos. Alexander makes a technical distinction between
& protasis and an apophansis, which our English terms do not well
reflect. But the distinction is explained in the text;108 and in any
case, apophansis is rare in the commentary, while apophainein is
normally used as a mere synonym for legein, 'say".109
Protaseis in the special sense, or premisses, come in pairs. A pair
of propositions is called a suzugia or a sumploke: etymologically, the
words mean 'co-yoking' and 'co-weaving1. We use 'combination' for
the first expression and 'conjunction' for the second.
The word protasis is only one of several devices by dint of which
Alexander refers to the premisses of an argument. Two verbs are
widely used - and each has several compound forms.
First, an arguer will lambanein a proposition. The verb is the
normal Greek word for 'take'; and in many contexts (especially when
it is terms which you must lambanein) this is how we translate it.
But in the case of'taking' a proposition, we need a word to express to
lambanomenon or to lephthen, 'what is taken'; and there is no
suitable noun from 'take'. The right noun for the passive participle
seems to be 'assumption'. (We have also used 'assumption' for the
105
huparkhein in this usage is normally translated as "belong to'; but 'belong to' has
no advantage over 'hold of, and it disguises the connexion with the one-place use of
huparkhein.
106
huparxis is 'existence', and is thus distinguished from hupostasis or
'subsistence': in Alexander there seems to be no difference in sense between the two
words; but the distinction is important in some Stoic contexts. On huparkhousa see
below, p. 30.
107
See esp. in An. Pr. 256,12-14 (cf. 258,24; 261,25-6; in Top. 2,6; Ammonius, in Int.
3,7-15); see e.g. Volait, pp. 24-7.
108
See 10,17 and note.
109
'Statement' at 1,19 represents axioma; elsewhere in Alexander this ambiguous
word is 'axiom'.
24 Introduction
noun lemma and - once - for the abstract noun lepsis, which is
usually paraphrased.) Hence it seems sensible to use 'assume' for
the verb when its object is a proposition or premiss.
Lambanein a proposition is sometimes contrasted with erotan, to
request, a proposition. Dialectical arguers must 'request', i.e. they
may argue from P only if an interlocutor 'grants' (didonai) a request
(erotesis) or gives an affirmative answer to the question 'P?'.
Demonstrative reasoners on the other hand are not thus dependent
on an interlocutor - they should simply 'take', without asking, what
appears to be true. But this contrast between lambanein and erotan
is not always present in Alexander: in particular, when he uses the
word 'assumption' he does not usually have demonstrative
syllogisms specifically in mind.
Of several compounds of lambanein, the most important are
proslambanein, metalambanein and paralambanein. Proslam-
banein is used as a term of art. (We also find proslepsis and
aprosleptos.) It applies specifically to the assumption of the second
premiss of a syllogism: you assume one premiss; and then you
assume in addition, or 'co-assume', a second. We translate
'co-assume'; and proslepsis is 'co-assumption'.
Metalambanein is once used in a technical sense as a synonym for
proslambanein: we there use 're-assume'.110 Normally the word
(which has a cognate noun metalepsis) means 'take ... instead of...':
we sometimes use 'transform' and sometimes paraphrase.
(Paralambanein does not seem to be technical. Alexander uses it
of introducing material - a term or a proposition - into a given
context. And so 'introduce' is apposite enough.)
The second verb is tithenai. Its general meaning is 'put' or 'place'.
It, its cognates, and its compounds are widely used by Alexander in
technical and semi-technical senses. You may tithenai a proposition,
and in particular you may tithenai a proposition as a premiss for an
argument. Here we use 'posit': the verb is unattractive (and its
normal English overtones are inapposite); but nothing preferable
presents itself. We use 'posit', the noun, for to tethen or to tithemenon
(passive participles). The abstract noun thesis, cognate with
tithenai, is normally used in connexion with terms rather than with
propositions. Here too we sometimes use 'posit' - but we also use
'position' and paraphrase.111
In normal Greek, keisthai ('lie') is often used as a perfect passive
of tithenai. For it we use 'be supposed', thus preserving a link with
'posit'. The participle, to keimenon, is frequently applied to what has
been posited or what is supposed - i.e. to a putative premiss; hence
110
See 19,5 and note.
111
On the thesis of terms, and the related notion of taxis ('order' or 'ranking'), see
47,22 note.
 Introduction 25
'supposition'. (The verb also has other, less technical, uses.)
There are five important compounds oftithenai, four of which it
seems best to discuss here, although they are not directly connected
to the concept of a proposition or a premiss.112
Alexander commonly uses paratithenai and parathesis in connex-
ion with the production of specific concrete terms, especially of terms
designed to provide counterexamples to a putative syllogism.113 We
use 'set down' for this verb.
Protithenai and prothesis connote a putting forward; what you
protithes is what you aim at or purpose. For example, Alexander will
explain the prothesis of the Prior Analytics, i.e. what Aristotle
intends or purposes to do in the work. We normally use 'purpose' or
'propose'; but the words are not technical, and we have chosen
different renderings as idiom suggests. There is, however, one
semi-technical use of the verb - or rather of the associated passive
verb and participle, prokeisthai and prokeimenon. In the context of a
syllogism, to prokeimenon is, as it were, what the syllogism or the
syllogizer is aiming at: it is the putative conclusion of the syllogism. If
you say to yourself'I want to find an argument to show X', then X is to
prokeimenon. We have found no simple word for to prokeimenon in
this usage: we generally make do with the clumsy paraphrase 'the
point at issue'.
Hupotithenai and hupokeisthai have three distinct uses. First, to
hupokeimenon is regularly the subject-term of a proposition - it is, as
it were, what 'lies under' or 'is placed under' the predicate. For this
use we keep the traditional word 'subject'; and for the verb: 'be subject
for'. Secondly, the verbs, and also the associated noun hupothesis, are
often used in the context of a particular type of argument, namely a
reductio. In a reductio you make a hupothesis and then show that
something absurd or impossible follows. As a result of this you reject
the hupothesis and thereby establish its negation.114 Here we
transliterate^ with the tradition, to 'hypothesis' and 'hypothesize'.
Thirdly - and relatively rarely - Alexander will use hupokeisthai of
the premisses of an argument. Here it is a synonym ofkeisthai, and
we translate 'suppose', thereby obliterating the distinction between
the compound and the simple verb.
Finally there is ektithenai, ekthesis and ekkeisthai. The words are
not uncommon in a non-technical sense, where they denote the
setting out of a term or a proposition or a problem. And we use 'set
out'. But there is also an important technical usage (but not, strictly
speaking, a technical sense): one of the methods of analysing or

112
For the fifth, antitithenai, see below, p. 26.
113
See below, p. 30.
114
See 24,18 note.
26 Introduction
reducing a syllogism employs what Aristotle calls ekthesis.115 We
retain 'set out' for the verbs in this usage. The translation is
adequate. But there is no corresponding noun; hence for ekthesis in
this usage we have adopted the traditional Latinism 'exposition'.
Propositions stand in a variety of relations to one another. If two
propositions are both true, they are said to be true together with one
another (sunaletheuein)116 or to hold at the same time as one
another (sunuparkhein). Again, there is akolouthia, the relation of
following or implication, which we have already mentioned.117 One
proposition may yield another by akolouthia, and it may also
anairein it. X anairei Y when the truth of X determines the falsity of
Y. The verb literally means 'destroy'; but 'destroy' is too dramatic a
metaphor in English, and we prefer 'cancel'. (But when the verb has
a personal subject we use 'reject'; for then 'destruction' is simply
denial.) Similarly for the cognates anairesis, anairetikos; and for the
compound sunanairein. If X destroys Y and Y destroys X, then X and
Y 'fight', makhontai: they conflict, we shall say (and there is a
conflict, makhe, between them). More soberly - and more frequently
- they are said to be opposed, antikeisthai.
We normally translate antitithenai, antikeisthai, antithesis and
antikeimenon by 'oppose', 'be opposed', 'opposition' and 'opposite': X
and Y are opposed when they conflict, i.e. when they cannot both
hold at the same time. There are different varieties of opposition. In
particular, there is a distinction between enantia ('contraries') on
the one hand and items which are opposed antiphatikos
('contradictorily') on the other. X and Y are opposed antiphatikos
just in case X and Y cannot both hold at the same time and also
cannot both fail to hold at the same time. (For example, and most
obviously, a proposition and its explicit negation are contradictorily
opposed.) If X and Y are opposed antiphatikos, then they are or form
an antiphasis.116 We usually translate this word by 'contradictory
pair'; but sometimes the abstract term 'contradiction' is more
appropriate.
In Stoic logic antikeisthai has the specific sense of 'be
contradictorily opposed',119 and sometimes Alexander uses ta
antikeimena, without any qualification, to refer to a contradictory
pair of opposites. Does he use antikeisthai in the Stoic sense in these
passages? Sometimes it is certain that he does, and there we
translate the word by 'contradictories'. But in many passages it is
115
See 32,33 and note.
116
But sunaletheuein is also used non-symmetrically: see 29,18 note.
117
antistrophd or conversion will be discused later, p. 31.
118
See the discussion of antiphasis at Ammonius, in Int. 81,13-84,25 (cf. 77,13-15),
which depends on Aristotle, Int. 17a33.
119
Note also Aristotle, ATI. Pr. 59b8-10.
 Introduction 27
plain that Alexander uses ta antikeimena to refer to contradictories
but unclear whether he uses it in the sense of 'contradictories'; and
in these passages we have preferred 'opposites'. (It is always clear
from the context whether the opposites in question are in fact
contradictories.)
Enantia or contraries cannot be true together but can be false
together; paradigm contraries in the context of Aristotelian
syllogistic are propositions of the form 'Every A is B' and 'No A is B'.
The pair 'Some A is B' and 'Some A is not B' are not contraries: they
may both be true but they cannot both be false. The relation in
which they stand is known as 'subcontrariety' and hupenantios is
rendered by 'subcontrary'.120 Further, 'Some A is B' is 'subaltern'
(hupallelos) to 'Every A is B', and 'Some A is not B' to 'No A is B'.
These various relations are the constituents of the so-called 'square
of opposition'.121
In addition to standing in these relations, the members of a pair of
propositions may also be 'similar in form' or 'dissimilar in form',
homoioskhemon or anomoioskhemon. These are terms of art and
their explanation demands a detour. Every proposition is either
affirmative or negative. 'Affirmative' gives kataphatikos: 'affirm' is
kataphaskein, 'affirmation1 kataphasis. 'An affirmative' is katapha-
tikon. (Aristotle - and hence Alexander - sometimes uses
kategorikos as equivalent to kataphatikos. We render this word by
'predicative', for reasons to be given later.122) 'Negative', 'negation'
and 'a negative' are apophatikos, apophasis, apophatikon. But for
the verbs apophaskein and apophanai 'negate' will not work, and we
use 'deny'. (Aristotle and Alexander sometimes use steretikos as a
synonym for apophatikos. We translate 'privative' the root verb
sterein means 'deprive'; but the intended sense of 'privative' is no
different from that of 'negative'.) The status of a proposition as
affirmative or negative is called by Alexander its quality, poiotes.
Propositions differ in quality, kata to poion, when one is affirmative
and the other negative. If the propositions in a pair differ in quality,
they are anomoioskhemones or dissimilar in form; if they are of the
same quality, they are homoioskhemones or similar in form.123
A proposition has a quantity as well as a quality: propositions
may differ or agree kata to poson, in quantity. For in every
120
Alexander appears to mean that 'Some A is B' is subcontrary to 'Some A is not
B', so that 'subcontrary' means something like 'quasi-contrary'. (The prefix
hupo- often has this semantic force.) Note, however, that Ammonius (in Int. 92,21-24)
supposes that 'Some A is B' is subcontrary to 'Every A is B', and that 'Some A is not B'
is subcontrary to 'No A is B': the particular propositions are 'subcontraries' inasmuch
as they fall under the universals, which are contraries.
121
See e.g. Ammonius, in Int. 91,4-93,18; Boethius, Syll. Cat. 800AC.
122
See below, pp. 30-1.
123
The terms do not seem especially apposite, but they are Aristotle's own.
28 Introduction
proposition, the predicate is predicated (or denied) of all or of some
of its subject; and differences in 'all' and 'some' are differences kata
to poson. A proposition which has the 'all' quantity is called
katholou, the word being used as an undeclinable adjective. In
Greek 'A holds of the whole of B' is 'A holds kath' holou tou B'; the
preposition kata (kath') and the quantitative adjective holos then
coalesce, giving katholou. We translate by 'universal'.
For 'some'-propositions Aristotle and Alexander use epi merous,
kata meros or en merei. For all three expressions, which do not differ
in sense, we use the adjective 'particular'. Corresponding to en merei
for 'some'-propositions, there is a use of en holdi - 'in a whole' - for
'all'-propositions. Now you might expect 'A is en holdi toi B' to mean
'A is in the whole of B', i.e. A is found everywhere in B, or A holds of
every B. But in fact the Greek means exactly the opposite: A is en
holdi toi B if B is true of every A. B is, as it were, a whole for A. To
make this reasonably clear, we have followed the customary - and
awkward - translation, writing 'A is in B as in a whole'.12*
(Meros occurs in other contexts, and is always translated 'part'. In
normal Greek, the diminutive morion is synonymous with meros.
And we have usually translated it too by 'part'. But morion
sometimes means 'subpart' - one of the few places being the opening
pages of Alexander's commentary.125)
Propositions are complex items: they sunkeitai ('are compounded':
suntithenai is 'to compound', sunthesis is given by the noun
'compounding'); they have a sustasis or 'construction' (the verb is
sunistanai, 'to construct'); in them, one element is linked
(sunaptein) or connected (suntattein) to another. (Suntattein gives
the noun suntaxis, from which derives the English 'syntax'; but in
Alexander 'syntax' is not strictly a linguistic notion: it is parts of
propositions, not parts of sentences, which are connected.)
The chief components of a proposition are terms or horoi. But a
proposition may also contain a diorismos, or a sign to mark its
quantity,126 and a tropos, or a sign to mark its modality. A
proposition with no specific indication of quantity, such as 'Pleasure
is good', is called adioristos. The best English for this is
'indeterminate'. Hence we use 'determination' for diorismos (and
'determine' for diorizein). Strictly speaking, there are only two
determinations of propositions, a universal and a particular; but the
124
For a clear account of the locution see Boethius, Syll. Cat. 810B. In Greek, as in
English, it is natural to take part and whole as mutually exclusive things; hence if A
is en holoi tdi B, it might seem to follow that A is not en merei toi B. But in Aristotle's
logic, 'Every A is B' entails 'Some A is B'; hence in this context wholes and parts are
not mutually exclusive. (In the modern jargon, 'part' here does not mean 'proper
part').
125 126
See Barnes (1988), pp. 240-2. Later authors preferprosdiorismos.
 Introduction 29
commentators all speak of four - 'Every', 'Some', 'No' and 'Some ...
not'. Hence the determination in effect gives the quality as well as
the quantity of the proposition. Thus in the determinate proposition,
'Every pleasure is good', 'Every' is or marks the diorismos. (We say
'is or marks'; for it is unclear in Alexander whether a diorismos is a
sign of quantification, i.e. a quantifying particle, and hence part of a
sentence; or whether it is, so to speak, the quantity signified by the
sign.)
Note that the word adioristos is also used in a different context,
involving the truth-conditions of particular propositions. We use
'indeterminate' here too - Alexander himself explains what he
means by this sort of indeterminacy.127
Tropos has a wide non-technical use: 'way', 'method', 'mode' are
variously appropriate, and we have used them as the context
suggests. In addition, there is a technical use of the word: here we
always use 'mode' (some might prefer 'mood'). If you say 'ex anankes,
A is B' then 'ex anankes' is or marks a mode.128 ('is or marks': the
same uncertainty arises.) We always translate ananke as 'necessity'
and ex anankes as 'by necessity'. Anankaios is usually 'necessary'.
Dei also marks necessity: we render it by the word 'must'. And for
the occasional opheilein we use 'should'. (In our part of the
commentary modal issues occupy a minor place; but elsewhere they
are of central importance, and we have therefore been uncharacter-
istically pedantic in translating modal words.)
A second tropos is that of possibility. Here there are familiar
difficulties for the translator. For Aristotle distinguishes different
sorts of possibility, and he also has two different words (or families
of words) to designate possibility, namely dunasthai and endek-
hesthai.129 Dunasthai and dunatos are always translated as 'be
possible' and 'possible' where Alexander is explicitly discussing
modal propositions. (And adunatos is 'impossible'.) Elsewhere we
often use 'can', rather than the cumbersome 'be possible', for
dunasthai. Endekhesthai, in normal Greek, is virtually a synonym
for dunasthai; and in contexts which are certainly non-technical we
use 'be possible' or 'can' for it too. But where Alexander is discussing
modal propositions, it is important to mark the difference between
endekhesthai and dunasthai. Accordingly, we have used 'be
contingent' for endekhesthai. This works well in some places. (For
example, 'A is B endekhomenos' comes out as 'A is contingently B',
127
See 66,1-10 and notes; cf. e.g. Ammonius, in Int. 94,7-24; 118,7-8; Philoponus, in
An. Pr. 82,1-84,11; 98,4-12. For a further type of indeterminacy see Ammonius, in Int.
90,19-20.
128
Hence the modern phrase 'modal logic' - but tropos embraces far more than the
modalities treated in standard modal logic (see Barnes (1991)).
129 There are also other words enkhorei, exesti, esti, and so on.
30 Introduction
i.e. 'A is B, but not by necessity' - and that is exactly the intended
sense.) But whereas in English 'It is contingent that A is B' suggests
that A is indeed B, it is not clear that the Greek carries the same
suggestion - 'endekhetai that A is B' may be true when A is not B. To
this extent our translation is unhappy. But we see no alternative
which is not at least equally bad.
Alexander holds that the 'plain' proposition 'A is B' also has a
modality:130 he has no special term for it, but (following Aristotle) he
refers to such propositions by the participle huparkhon, 'holding'.131
We might say that 'A is B' is taken to say that A is actually B, as
opposed to being necessarily or possibly B. And so we have used
'actual' for huparkhon in this usage. (The translation disguises
connexions which are patent in the Greek; and the expression 'an
actual proposition' is vaguely comic. But, again, alternatives seem
worse.)
Terms constitute the hule or 'matter' of a proposition (and hence,
indirectly, of a syllogism). And they contrast with its eidos or 'form',
which - in effect - is determined by its logical structure, or its
quality, quantity and mode. Aristotle and Alexander regularly use
'dummy letters' - ABC, MNO, PRS - in formulating syllogisms.132
Thus they will represent the form of a proposition by such a formula
as 'A holds of every B'. When he wants to give a genuine proposition,
with genuine terms, Alexander will talk of setting down133 hule.
(Here we translate hule as 'material instance'.) When Aristotle gives
a material instance or concrete term, he uses either a common noun
('man', 'swan') or a mass term ('snow') or an adjective in the neuter
singular ('wild', 'rational'). We translate these straightforwardly
into English. The result is usually barbarous; for in ordinary
English such a formula as 'No wild is man' is ill-formed. Yet it is
clear what those formulae are intended to express; and the Greek
formulae which they represent are also barbarous.
In a simple proposition, one term is predicated of (kategoreisthai)
another; and it is called 'the predicate' (to kategoroumenon) of the
proposition. The other term is the subject (hupokeimenon, as we
have already remarked). The verb kategorein has compounds,
antikategorein, 'counterpredicate', and proskategorein, 'co-
predicate'. There is also an adjective kategorikos: as we have said, it
is used specifically to designate affirmative propositions; but it is
also used generically to designate simple propositions,134 and hence
130
The point later aroused controversy: see Barnes (1991).
131
The participle is usually in the feminine, huparkhousa, agreeing with an
unexpressed protasis.
132
See 53,30 note.
133
paratithenai: above, p. 25.
134
As opposed to complex or hupothetikoi propositions: above, p. 23.
 Introduction 31
to pick out those syllogisms whose component propositions are all
simple.135 We use the adjective 'predicative', to preserve the trans-
parent connexion with kategorein. (Latin authors regularly use
praedicativus.) It is true that 'predicative' does not transparently
designate these items - but then the same holds for kategorikos.136
To express the predicative tie, Aristotle also uses legesthai, 'be said
of, and huparkhein, 'hold of (and especially in schematic formulae,
he often omits a verb altogether). 'B is A', 'A holds of B', 'A is said of B',
'A is predicated of B' - all these are equivalent.137 Note in particular
that 'A is said (or: predicated) of B' does not mean 'Someone says: "A is
B"'.
Subjects and predicates are alike terms: they belong to the same
logical category. (In this they are fundamentally different from the
subjects and predicates of modern 'Pregean' logic.138) Hence they may
exchange places in a proposition. Alexander's technical term for this
exchange is antistrophe, with the verb antistrephein. Antistrophe
may denote either an operation on a proposition, viz. 'conversion', or
else the result obtained by performing the operation, i.e. a converse.
Thus if you operate on 'No B is A' to reach 'No A is B', you perform a
conversion (antistrophe) and you get a converse (antistrophe). It is not
always clear whether antistrophe means 'conversion' or 'converse'.
We have chosen what in each context seems the more plausible option
- in point of fact little turns on the issue.
The verb antistrephein sometimes takes a personal subject: the
logician converts a proposition. More often the subject is a proposi-
tion. Then there are two grammatical constructions: antistrephein +
dative, antistrephein + pros + accusative. There is a difference in
usage here.139 Antistrephein pros marks a symmetrical relation: if X
antistrephei pros Y, then Y antistrephei pros X. We talk here of
'converting with'. Antistrephein + dative, on the other hand, is
non-symmetrical: if X antistrephei toi Y, then Y may or may not
antistrephei toi X. Moreover if X antistrephei toi Y, then X may be
derived from Y by a process of 'prepositional conversion'. Hence we
use 'convert from' for antistrephein + dative. (Prepositional antistro-
phe, as Alexander explains, is simply an interchange of terms which
preserves truth.140)
IBS por ancient explanations of the Aristotelian use of the term see Dexippus, in
Cat. 12,29; Ammonius, in Int. 70,7; 87,13.
136
kategorikos is usually transliterated to 'categorical'; hence Aristotle's logic is
known as 'categorical syllogistic'.
137
Note that, according to Apuleius (Int. 192,30-193,5) the Peripatetics prefer the
expression 'A holds of B', whereas he normally uses 'B is A' (cf. below, 54,25-29;
Albinus, Didasc. 158-9H).
138
On this see esp. Barnes (1983).
139
The difference emerges most clearly at in An. Pr. 392,19-26.
uo
See 29,1-29 and notes.
32 Introduction
Two terms may stand in various relations to one another. If every
B is A, then A 'includes' or 'encompasses' B, periekhei or
perilambanei. (There is a nounperilepsis.) A is 'over', huper, B, and
B is 'under', hupo, A. A may, in addition, hold of things which are not
B. Then A is of wider extension, epi pleon, than B, and it (or some
part of it) 'falls outside' B (piptein ektos). Again, if no A is B, then A
and B are 'disjoined' (apezeugmenon).
Asuzugia or combination of premisses contains three terms in all:
its two constituent propositions 'share', koinonein, a term.
(Koinonia, and the adjective koinos, are also used in this connexion.)
The shared term or the term taken twice - once in each proposition -
is the middle term, meson; and the other two unshared terms are
'extremes', akra. Any further term which may be invoked to show
that the combination is syllogistic is said to be taken 'from outside',
exothen. (And Alexander also speaks of operations, such as
antistrophe, being introduced 'from outside', a locution which is not
illuminating.)
The extremes are distinguished in two ways: first, one of them
may be called 'first' (prdtos) and the other 'last' (eskhatos). Secondly,
and more importantly, one may be called meizon and the other
elation. The standard translations are 'major' and 'minor'; and we
have retained these English words, not without misgiving. Literally
the words mean 'greater' and 'less'. They are comparative adjectives.
A is greater than B if and only if B is less than A; and A is greater
than B if A is more extensive (or holds epi pleon) than B. In English,
'major' and 'minor' do not carry such evident implications of relative
extension; nor, of course, are they genuine comparatives. This
means that our translation suppresses an important idea present in
the Greek. In addition, in some passages Alexander speaks
expressly of A being meizdn than B. In these passages we hedge by
using phrases such as 'A is major in extension with respect to B'.
(The notion of relative extension is important in certain contexts.
Thus it is connected with definitions (horismoi), which standardly
analyse the definiendum into genus (genos) and differentia
(diaphora) and which may do so by the method of 'division'
(diairesis).141)
Finally, the terms in a combination have an order and a position
(taxis, thesis)142 The order and position of the middle term
determines the type or 'figure' (skhema) of a combination.143
141
Hence diairein is normally 'divide'. The compound verb antidiairein is used
when you pick out items at the same level in a division: we use 'co-ordinate' here,
losing the connexion with diairein.
142
On these notions see 47,22 note.
143 For the figures see Appendix 2.
 Introduction 33
Most readers will find some of our translations misguided. There
is nothing much to be done or said about this. All translators are in a
pickle, and logical texts are no easier to deal with than texts in
ethics or metaphysics.
This page intentionally left blank
 Alexander of Aphrodisias
On Aristotle Prior Analytics 1.1-7
Translation
This page intentionally left blank
 Contents

1. Preface
1.1 The status of logic 41
1.1.1 Instrument or part of philosophy? 41
1.1.2 Useless parts of logic? 43
1.1.3 The value of logic 46
1.2 The Prior Analytics 48
1.2.1 Contents and title 48
1.2.2 Procedure 50

2. Introductory Explanations
2.1 ThegoaloftheAraaZyics 53
2.2 Propositions 54
2.2.1 The general account 54
2.2.2 Demonstrative and dialectical propositions 57
2.3 Terms 61
2.3.1 Subjects and predicates 61
2.3.2 The copula 61
2.4 Syllogisms 63
2.4.1 The premisses 63
2.4.2 The conclusion 66
2.4.3 Necessitas consequentiae 71
2.4.4 Non-syllogistic arguments 71
2.4.5 'Positing' 75
2.4.6 Perfect syllogisms 75
2.5 Kinds of proposition 77
2.5.1 Universal Propositions 77
2.5.2 Modality 79
2.5.3 Quality and quantity 80
2.5.4 Modes 81

3. Conversions
3.1 Conversions in general 83
3.2 Non-modal conversions 84
3.2.1 General remarks 84
37
38 Translation
3.2.2 Universal negatives 86
3.2.3 Other propositions 90
3.3 Modal conversions 91
3.3.1 Necessary propositions 91
3.3.1.1 Necessary universal negatives 91
3.3.1.2 Other propositions 93
3.3.2 Contingent propositions 95
3.3.2.1 Contingent affirmatives 95
3.3.2.2 Contingent negatives 96

4. The First Figure

4.1 Introductory 101
4.1.1 General remarks 101
4.1.2 Syllogisms and justification 102
4.1.3 Problems 104
4.1.4 Propositions which share two terms 106
4.1.5 Combinations and figures 107
4.2 Why is the first figure first? 108
4.3 First figure syllogisms in general 111
4.3.1 General rules 111
4.3.2 Ranking 113
4.3.3 The number of combinations 113
4.3.4 Aristotle's general comments 115
4.4 Combinations of universal premisses 116
4.4.1 Barbara and Celarent 116
4.4.2 Non-syllogistic combinations 118
4.4.3 Concluding remarks 120
4.5 Combinations of universal and particular premisses 121
4.5.1 Darii and Ferio 121
4.5.2 Non-syllogistic combinations 124
4.5.2.1 Major particular, minor universal 124
4.5.2.2 Minor particular and negative 126
4.5.2.3 The indeterminacy of particulars 130
4.6 Combinations of particular premisses 133
4.7 Additional syllogisms 134

5. The Second Figure

5.1 General remarks 138
5.2 Major and minor terms 140
5.3 Second figure syllogisms 144
5.4 Combinations of universal premisses 147
5.4.1 Cesare and Camestres 147
5.4.2 Non-syllogistic combinations 150
 Translation 39
5.5 Combinations of universal and particular premisses 151
5.5.1 Festino 151
5.5.2 Baroco 152
5.5.3 Non-syllogistic combinations 155
5.5.3.1 Major universal, minor particular 155
5.5.3.1.1 Aristotle's proofs 155
5.5.3.1.2 The view of Herminus 160
5.5.3.2 Minor universal, major particular 162
5.6 Combinations of particular premisses 163
5.7 Concluding remarks 164

6. The Third Figure

6.1 General remarks 166
6.2 Third figure syllogisms 168
6.3 Combinations of universal premisses 171
6.3.1 Darapti 171
6.3.2 Felapton 174
6.3.3 Non-syllogistic combinations 175
6.4 Combinations of universal and particular premisses 176
6.4.1 Disamis and Datisi 176
6.4.2 Bocardo 177
6.4.3 Non-syllogistic combinations 178
6.4.4 Ferison 180
6.4.5 More non-syllogistic combinations 181
6.5 Combinations of particular premisses 182
6.6 Concluding remarks 183

7. Further Reflections
7.1 Syllogisms with non-standard conclusions 184
7.2 Reduction 187
7.2.1 All syllogisms reduced to the first figure 187
7.2.2 All syllogisms reduced to Barbara and Celarent 189
7.3 Final remarks 195
This page intentionally left blank
 1
Preface

1.1 The status of logic

1.1.1 Instrument or part of philosophy?1

Logic2 or syllogistic is the study now before us. Under it fall 1,3
demonstrative, dialectical and examinatory methods, and also
sophistical procedure.3 It is the product of philosophy: some other 5
sciences and arts do indeed use it, but they take it from philosophy,
to which belong its discovery, its construction, and its most
important uses.
Since logic is the product of philosophy, some people think that it
is also a part of philosophy.4 Others, however, say that it is not a
part but an instrument of philosophy.5
1
For other accounts of this dispute see esp. Ammonius, in An. Pr. 8,15-11,21;
Philoponus, in An. Pr. 6,19-9,20 (other texts in Hiilser, pp. 22-39 - to which add Elias,
in An. Pr. 134,4-138,13). We suppose that the dispute which Alexander rehearses
reflects a genuine historical debate; but the date of the debate is uncertain. (See,
perhaps, Seneca, -Ep. 88,21-8 [= Posidonius, F 90 EK], which shows that Posidonius
engaged in an analogous debate about the parts and instruments of philosophy and
which at least indicates that the debate described by Alexander would not have been
unthinkable in the first century BC.) For some account of the nature and importance
of the dispute see e.g. Mueller (1969), p. 184; Moraux (1986a), pp. 268-9; Gottschalk
(1987), p. 1099; Lloyd (1990), pp. 17-21; and esp. Lee, pp. 44-54.
2
1.3-2.2 = FDS 27.
3
For Alexander's explanations of 'demonstrative', 'dialectical', 'examinatory' and
'sophistical' methods see in Top. 2,20-3, 2,23-5, 22,10-14, and 2,25-6. For the species
of syllogisms see below, 7,9 note.
4
This view is generally associated with the Stoics; but it was a commonplace of
Hellenistic and post-Hellenistic thought to maintain that philosophy had three parts,
logic and physics and ethics. (See e.g. Seneca, Ep. 89,9: 'Most, and the most
important, authors say that there are three parts of philosophy ethical, physical,
logical"; Apuleius, Int. 176,1-3: 'The study of wisdom, which we call philosophy, seems
to most people to have three parts ....' cf. P. Hadot.) The Peripatetics, who divided
philosophy into two parts (theoretical, which is roughly equivalent to physics, and
practical, which is roughly equivalent to ethics), were heterodox; and Alexander's
opponents need not have been exclusively Stoics.
5
The 'others' include Aristotle, according to Alexander: in Top. 74,29-75,3 (cf.
94,7-10; and also below, 3,5, which implies that 'the ancients' had called logic an
instrument). No Aristotelian text expressly says that logic is an instrument of

41
 42 Translation
6
10 Now those who say that it is a part were led to their view by the
following consideration. Just as philosophy concerns itself with the
other items which are universally agreed to be parts of philosophy,
making it its business to discover and order and construct them, so
too it concerns itself with the study before us. But although this
study is the product of philosophy, it is not a subpart of either of the
other parts of philosophy - neither of the theoretical nor of the
15 practical part. For its subject matter is different from theirs, and the
purposes of the three are distinct. The theoretical and the practical
parts differ from each other in certain ways, and in virtue of these
differences they are co-ordinate with one another: logic differs from
each of them in the same ways, and therefore it may reasonably be
taken as co-ordinate with them. It differs from them both in
subject-matter (for its subject-matter consists of statements and
20 propositions7) and also in its end and purpose (for its purpose is to
prove that, when propositions are compounded with one another in
2,1 certain ways, something may be deduced by necessity from what is
posited or conceded - and this is not the end of either of the other
parts).
Those who claim that logic is not a part but an instrument of
philosophy reply as follows. For something to be a part of an art or
5 science it is not enough that the art or science is concerned with it8
in the same way as it is concerned with each of the other parts which
it studies. For something is judged to be a part not merely because it
is an object of attention and study, but when, in addition, its end and
construction do not make reference to anything other than those
10 very things which the science in question studies, i.e. when it is not
investigated and constructed for the sake of these other items. For if
something makes reference to the needs of other items which fall
under the same science or art, then it cannot properly be taken as
co-ordinate with them, since it is for their sake that it exists and
comes into being; nor can it be a part in the way in which they are,
since it exists for their sake. For if something makes reference to
15 certain items, and if its end is given attention insofar as it
contributes to the discovery and construction of other items, then it
is the instrument of those items. With distinct arts, the product of
one is the instrument of another if its end makes reference to the
needs of what is made by the art whose instrument it is: similarly, if
the things made by a single science or art have such a ranking in

philosophy; but Top. 163b9-ll, which uses the word organon in a pertinent context,
may have been in Alexander's mind (cf. in Top. 584,9-12).
6
1,9-2,2 = SVF II 49a.
7
cf. Ammonius, in An. Pr. 9,26-7. Note that here Alexander uses axioma
('statement') in the Stoic sense: see e.g. Frede (1974a), p. 32 n. 1.
8
Omittingtomorion.
 1. Preface 43
relation to each other, then one will be an instrument and the other 20
- the higher one - will be both the product and a part of the science.
A hammer and anvil are not precluded from being an instrument of
the smith's art by the fact that they are its product.9
Moreover, those who call logic a part of philosophy must admit
that a part of philosophy is an instrument of those other sciences
and arts which use syllogisms and demonstrations to establish and
construct the objects of their own concerns. For they use them, but 25
not as parts of themselves (since it is impossible for the same thing
to be a part of different sciences, nor does any of these sciences study
the construction and discovery of syllogisms); hence they will be
using them as instruments. But if so, and if one art or science
dominates another whenever it uses in its own production the
product and end of the other art, whose own product makes 30
reference to its needs (as bridle-making in relation to the art of
riding and ship-building in relation to the pilot's art),10 then
according to these people there will be other sciences and arts more
perfect than philosophy, namely those which have a part of
philosophy as their instrument.

1.1.2 Useless parts of logic?11

They might say that not all the study of logic makes reference to the 35
discovery and construction of what philosophy investigates or of
what some other science or art theorises about or investigates, but
that there are things logic theorises about and investigates which 3,1
have no utility at all.
In that case, first, they will concede that the earlier thinkers12
who developed the study of logic to the extent of its utility were right
to call it an instrument and not a part.
Secondly, according to them, what is useful in logic is an 5
instrument and it is what is not useful which will be a part.13 But if
what is useful is better and worth more attention and study, then
for them the better subpart of logic will be an instrument of
philosophy and the other sciences and arts, and what is worth less
9
10
For the analogy see Ammonius, in An. Pr. 9,37-10,1; Philoponus, in An. Pr. 7,31.
See Aristotle, EN 1094a6-16.
11
For the utility of logic see further 8,19-22; 18,14-19,3; 20,12-13; 28,17-30;
30,29-30; 39,19-40,5; 44,4; in An. Pr. 164,23-165,6; cf. Aristotle, Top. 101a25-b4 (with
Alexander, in Top. 9,20-10,16). See Barnes (1990b).
12
The earlier thinkers, hoi arkhaioi or hoi palaioi, contrast with the more recent
thinkers, hoi neoteroi. The designations have different referents in different contexts;
but by and large it is true to say that, in in An. Pr., Alexander is thinking of Aristotle
and his immediate followers when he refers to the ancients and of the Stoics when he
refers to the more recent thinkers (cf. 17,12; 19,5; 22,18; etc). See further Kieffer, pp.
130-3; Barnes (1990a), IV.3.
13
Omitting on.
 44 Translation
10 attention will be a part. Yet in all cases, an instrument has a lower
status than any product which is a part.
Again, if logic is given attention in order to exercise the mind for
making discoveries among items which the parts of philosophy
investigate,14 then in this way too it will have the status of an
instrument. If, on the other hand, logic is studied in order to gain
knowledge of the truths it contains in itself,15 then it will become a
subpart of theoretical philosophy; for it is this which has knowledge
as its end. But if you theorise about useless items in accordance with
15 logical method, what good does that contain which might make it
worth attention as a part of philosophy? What is worth attention
should either make reference to something else which is desirable in
itself, or else contain in itself something worthy of attention. For it is
not all and any knowledge which is worthy of philosophy (some things
it is better to be ignorant of), but only knowledge of what is divine and
20 valuablethat is, of the things created by the divine art of nature.16 It
is clear from its very name that theorising is concerned with the sight
and knowledge of what is divine; for it means seeing what is divine.17
That is why we say that theoretical philosophy is knowledge of what
is divine and of what comes about or is constituted by nature;18 for
knowledge of such things is in itself worth attention. But where the
25 things theorised about neither make reference to anything else nor
contain in themselves anything excellent19 and valuable, then
knowledge of them, being utterly superfluous, is not appropriate to
philosophy - it is a waste of labour. For it is above all appropriate to
philosophers that, as they should never act at random, so too they
should never theorise in this way: their theorising, like their actions,
should be delimited.
30 Geometry is not, as they think, similar to the useless part of logic -
their comparison is mistaken. First, as they themselves say, geo-
metry is not a part of what is called philosophy in the strict sense.20
14
For this use of logic see Aristotle, Top. 101a28; Alexander, in Top. 27,7-31 (cf.
below, 8,24-7).
15
Reading autei for autois.
16
For creative nature, he demiourgesasa phusis, see esp. Aristotle, PA 654a6-23; cf.
Alexander, in Metaph. 103,5-104,18.
17
i.e. thedrein is fancifully derived from theia and horan. Note that 'theory',
'theorise' and their cognates are not ideal translations. (The standard 'contemplate',
'contemplation' and so on are no better.) To 'theorise' about an area is simply to study
the area with a view to grasping the truth about it (and not with any further practical
end in mind).
18
cf. Aristotle, Metaph. 983a5-10; Alexander, in Metaph. 18,5-13. Compare the
standard Stoic definition of philosophy as 'knowledge of things divine and human'
(e.g. Seneca, Ep. 89,5; Sextus, MIX13).
19
Retaining peritton (Wallies prints terpnon with the Aldine). But note that
perittos occurs in the following line with the sense of'superfluous'.
20
cf. e.g. Posidonius, F 90 EK (Seneca, Ep. 88,21-8).
 1. Preface 45
21
Then, astronomy, being in a way a subpart of geometry, theorises 4,1
about divine and natural substances, knowledge of which is in itself
noble and valuable.
In addition, very many aspects of geometry are useful for philo-
sophy, and because of them it is reasonable to give it attention.22 For
the fact that geometry deals not only with perceptible things but with
things which escape perception and are intelligible is very useful for 5
philosophical theory since the incorporeal and intelligible substances
about which philosophers theorise are primary and more valuable
than the perceptible substances. Further, geometry accustoms us
from our youth to deal with lines and planes and solids, none of which
are perceptible, and instructs us about each of these things by
themselves. Now if you are to engage in philosophical theorising, it is
absolutely necessary to be able to distinguish from one another in
account things which are different from one another in substance but 10
cannot be apart from one another in subsistence and existence. For
this is how philosophy grasps the principles of natural compounds,
i.e. matter and form: these are inseparable from one another in
subsistence and neither of them can be apart from the other. Again, it 15
is by separating in account what is confused23 that philosophy makes
distinctions among intelligible objects and finds the number of the
genera under which all beings fall; for the other nine genera depend
on substance, which underlies them, and cannot subsist apart from
it.24 And for this25 geometry has proved necessary and useful by
accustoming us to talk separately about lines, as if they constituted a
certain nature by themselves, and separately about planes and 20
solids, even though none of these things can subsist outside physical
bodies.
Further, by using syllogisms and demonstrations in proofs about
its objects, geometry accustoms us to take the same attitude in
philosophy too, so that we do not follow and believe silly stories26
about its objects, but demand demonstrations and regard as justified 25
only what is either known through itself or posited by way of a
demonstration.
21
That astronomy is a mathematical science is a commonplace: see esp. Plato, Rep.
528D-530C; cf. e.g. Aristotle, Phys. 194a7-9; Metaph. 989b31-3 (Alexander, in
Metaph. 72,9-12); Geminus, quoted by Proclus, in Eucl. 38,4-10; Ptolemy, Synt.
6,21-7,4. But we have found no other text affirming that astronomy is a subpart of
geometry.
22
An idea developed by the Platonists: see e.g. Albinus, Didasc. 161H;
lamblichus, Comm. Math. Sc. 55,21-2; Proclus, in Eucl. 21,15-22,16; cf. O'Meara. The
classic text is Plato, Rep. 526C-527C.
23
cf. Aristotle, Phys. 184a21-3.
24
Alexander refers to the ten Aristotelian genera or 'categories' of being, the first of
which is substance.
25
Readingpros touto for pros toutoi.
26
For muthodeisphluariai see e.g. Ammonius, in Int. 181,5-7; 249,1-25.
 46 Translation
Again, geometry uses proportions and symmetries for its own
proofs: many of the things which concern philosophy are also proved
by these means, none of which can be taken from the useless part of
dialectic and made to contribute to philosophy.
For these reasons this part of dialectic is completely useless and
superfluous and will rightly be banished from philosophical
theorising.

1.1.3 The value of logic27

30 But if analytics is an instrument and not a part of philosophy, it is
not therefore any less worthy of our attention. For the attention due
to an instrument is judged by the worth of what is made or proved
by it. But the objects of philosophy are proved by demonstration,
and they are of the greatest worth. For everything in philosophy is
5,1 worth full attention, and the theory and knowledge of the truth
above all else. For the truth is both most appropriate to men and
also their highest good.
That it is most appropriate is clear from the following fact. Just as
something is appropriate and natural to each of the animals
(running to one, flying to another, hunting to one, guarding to
5 another, bearing burdens to another, and so on), so knowledge and
theory are appropriate and natural to men. 'For all men by nature
seek after knowledge'28 - and there is evidence for this in the facts
that they prefer those modes of perception which provide them with
more knowledge,29 and that they are immediately from their youth
fond of listening to stories. For small children listen attentively to
10 stories30 although they gain nothing from them apart from the
knowledge of them; and the more gifted children are distinguished
by the attention with which they listen to the stories - which
suggests that knowledge is natural for men.
Further, men are distinguished from other animals above all and
most evidently by knowledge of the truth.31 You can find some
15 traces of the virtues and of virtuous actions in non-rational animals
too (for some of them seem to share in courage, some in temperance,
27
According to Todd (1976), p. 17 n. 83, this passage (with which cf. e.g.
Philoponus, in An. Pr. 4,26-9) is 'an apologia for the study of logic', written at a time
when logical studies were often disparaged. But note that the question of the 'utility'
of logic was also seriously discussed among logicians (see above, 2,33 note). (Later,
the value or utility of a subject was one of the standard topics in the Procline canon of
commentatorial procedure: see Introduction, p. 8.)
28
Aristotle, Metaph. 980a21.
29
cf. Aristotle, Metaph. 980a21-6; Alexander, in Metaph. 1,10-2,3.
30
A commonplace exploited by philosophers to various ends; see e.g. Olympiodorus,
in Gorg. 46,3; Elias, Proleg. 27,15-22.
31
cf. Aristotle, Metaph. 980a27-981al2; Alexander, in Metaph. 2,24-5,13.
 1. Preface 47
and some in justice, and further some of them have actually been
thought in addition to be sagacious and others to be generous, as we
can learn from the History of Animals,32 which Aristotle wrote in
several books); but for truth and theoretical understanding they 20
have no sense at all.
Secondly, theorising is the highest of human goods. 'For truth it is
which brings all goods to gods and to men.'33 If we may express an
opinion about what is above us, we must deem that apart from this
there exists no other activity for the gods.34 It cannot be allowed
that they are active in accordance with any of the other virtues,
since the virtues concern the emotions (insofar as they measure and 25
shape them), whereas the divine is free from emotion. Further,
moral virtues involve choice. For they have been proved35 to be
dispositions connected with choice. And choice is deliberative
desire,36 and deliberation is about objects of deliberation, and we
deliberate about what is up to us but of unclear outcome. (No one
deliberates about scientific matters.37) So if for the gods none of the 30
things they bring about is of unclear outcome, then for them there
will be no objects of deliberation. If so, no deliberation either; if no
deliberation, no choice; and if no choice, no moral virtue. 6,1
Now, for the gods, theorising about truth is continuous and
uninterrupted. But for men it is not possible to be continuously
active in this way38 - for many of the conditions of life which were
allotted to them lead them away from things of higher value. Yet if a
man emerges,39 as far as he can, from the emotions and conditions of 5
human life, he may see the things of highest value and be active in a
theorising which is divine and worthy of its name. Now when he is
active with this faculty of his soul and exercises activities like those
of the gods, then he will become like the gods. Thus if becoming like
god is the greatest good for men,40 and if this is attained by the
32
See e.g. Aristotle, HA 488bl2-26 (cf. EN 1144b8-9); for sagacity (phronesis) in
animals, see e.g. HA 611al5-19; 614bl8-21; 618a25-30 (cf. Metaph. 980b22-5).
33
Plato, Laws 730C. For Alexander's knowledge of Plato see Sharpies (1990), pp.
90-2.
34
The following argument is a paraphrase of Aristotle, EN 1178b7-23 (see also
Metaph. 1072bl3-30; cf. e.g. Cicero, Nat. Deorum III xv 38; Alexander, in Metaph.
2,3-21; 17,5-18,14; Quaest. 141,2-4: Dooley, pp. 13-14; 38-9).
36
36
See Aristotle, EN 1106b36.
EArill3alO-ll.
37
AT1112a34-bll.
38
39
See Aristotle, EN 1177b26-31 (but note 1177a21-2); cf. Metaph. 1072bl4-16.
Anakuptein: for the metaphor see Plato, Phaedrus 249C, to which Alexander
probably alludes.
40
For homoiosis theoi as the greatest good for man see esp. Plato, Rep. 613A;
Theaetetus 176B; and for Aristotelian parallels see e.g. EN 1177b30-1178a4;
1178b25-8. The idea is a commonplace in later Platonism (e.g. anon., in Theaet. 7
14; Albinus, Didasc. 153, 179-80H; Plutarch, Ser. Num. 550D; Apuleius, Dog. Plat.
126,3-6); but it is also found elsewhere (e.g. Seneca, Ep. 92,27; Epictetus, Diss. II xiv
 48 Translation
10 theory and knowledge of what is true, and if the knowledge of what
is true comes by way of demonstration, then demonstration will
rightly be held most valuable and worthy of most study - and so too,
therefore, will syllogistic, since a demonstration is a sort of
syllogism.41

1.2 The Prior Analytics

1.2.1 Contents and title42

That is why Aristotle first studies syllogisms in this work, the title of
15 which is Prior Analytics, before he discusses demonstration. He tells
us what a syllogism is, what syllogisms are compounded from, how
many syllogistic figures there are, and what are the differences
among them. (The figures are like a sort of common matrix: by
fitting matter into them, it is possible to mould the same form in
different sorts of matter.43 For just as things fitted into one and the
20 same matrix differ not in form and figure but in matter, so it is with
the syllogistic figures.) He44 says further how many combinations
there are in each figure, which of them are syllogistic and which are
non-syllogistic, what are the differences among the syllogisms in
each figure, which of the syllogisms are perfect and directly known
25 and in no need of demonstration, and which are imperfect and not
indemonstrable,45 and how it can be that imperfect syllogisms are
said to be syllogisms at all. In this work he also discusses the
discovery of premisses, saying how we may discover premisses and
so obtain appropriate syllogisms in each figure. He also discusses
the analysis of syllogisms and in general whatever is proper to the
study of syllogistic.
30 Having discussed these issues in the two books of the Prior
Analytics, he will next proceed to the discussion of demonstration -
11-13; and cf. Alexander, DA 90,10-91,6). See e.g. Praechter (1909a), pp. 541-3;
Merki; Jaeger.
41
Aristotle, An. Pr. 25b30; below, 6,32-7,11; 42,17-31.
42
See also Alexander, in An. Pr. 340,11-12; Ammonius, in An. Pr. 5,5-7,25;
Philoponus, in An. Pr. 5,15-6,6. For analogous discussions in other commentaries see
e.g. Simplicius, in Cat. 10,8-19 (quoting Alexander's lost commentary); cf. Praechter
(1909a), pp. 530-1; Todd (1976), p. 9 n. 39. Note that Aristotle himself refers to the
Analytics by the phrase to. analutika. But he does not use the distinguishing
qualifications 'Prior' and 'Posterior', which are first found in the catalogue of
Aristotle's writings preserved by Diogenes Laertius, V 23.
43
On matter and form in logic see e.g. below, 52,19-25; cf. Patzig, p. 171; and esp.
Barnes (1990a), III (to the references there add Alexander, Conv. 56).
44
6,21-6 =FDS 1099.
45
The later tradition regularly connects 'perfection* with indemonstrability (on
which see Introduction, p. 21): see below 24,4-5; 55,3; 69,28; 113,7; Galen, Inst. Log.
viii 1-4; Apuleius, Int. 188,4-11; Sextus, PH II 198; Boethius, Syll. Cat. 823A; cf.
Frede (1974a), p. 128; Flannery.
 1. Preface 49
which he does not propose to deal with here - in the books entitled
Posterior Analytics. (These are also two in number.)
Now the reason why he entitles the work on syllogisms Prior
Analytics and the work on demonstration Posterior Analytics is that
the syllogism is by nature prior to demonstration. We have learnt in
the Categories that one thing is prior by nature to another if it 7,1
follows when the other is posited but does not convert with it as to
implication of existence.46 Genera are things of this sort: every
genus is prior by nature to each of the species which fall under it; for
if a species is posited, the genus must necessarily follow, whereas
the species does not follow the genus.47 Similarly with species in
relation to the things of which they are species: species are prior. 5
And thus too is the syllogism related to demonstration. For if
there is a demonstration, there must be a syllogism, since a
demonstration is a sort of syllogism; but if there is a syllogism, there
need not be a demonstration,48 because there are also dialectical
and sophistical syllogisms.49 Thus since the syllogism is prior and
demonstration posterior, it is reasonable that Aristotle entitled 10
those books in which he discusses what is prior Prior Analytics and
those in which he discusses what is posterior Posterior Analytics.
They are called Analytics because the reduction of any compound
to the things from which it is compounded is called analysis.50
Analysing is the converse of compounding;51 for compounding is a
route from the principles to what depends on them, whereas
analysing is a return route from the end up to the principles.52 15
Geometers are said to analyse when they begin from the conclusion
and proceed in order through the assumptions made for the proof of
46
See Aristotle, Cat. 14a29-35: Alexander alludes to the second of five types of
priority there distinguished. (Aristotle does not explicitly call this natural priority;
but note 15a7-ll.) The tortuous phraseology, which Alexander takes from Aristotle,
hides a simple thought: x is prior by nature to y provided that (i) if y exists, then x
exists, and (ii) it is not the case that if x exists then y exists.
47
See e.g. Aristotle, Cat. 15a4-7; cf. e.g. Alexander, in Top. 367,16-20.
48
Aristotle, An. Pr. 25b29: below, 42,17-27.
49
On the different kinds of syllogism see below 8,19-29; 18,22-31; 28,24-30 (cf.
12,7-14,6); cf. e.g. Alexander, in An. Pr. 331,12-24; in Top. 2,15-3,24; 10,26-8; in
Metaph. 260,1-20. It is a commonplace to distinguish three kinds of syllogism (cf. e.g.
Albinus, Didasc. 158H); but note that at Conv. 57-9 Alexander adds a fourth kind, the
'examinatory' syllogism (see above 1,4; cf. e.g. Ammonius, in An. Pr. 2,18-29). At in
An. Pr. 2,29-3,30, Ammonius offers a schematic derivation of the division of
syllogisms into their kinds (cf. Philoponus, in An. Pr. 2,22-4,14).
50
For analysis in Alexander see also in An. Pr. 275,32-7; Quaest. 4,4-7. On the
varieties of analysis see e.g. Albinus, Didasc. 156-7H; Ammonius, in An. Pr.
5,10-7,25; Philoponus, in An. Pr. 307,6-8; scholium on Aristotle, 140a35-41; see e.g.
Lloyd (1990), pp. 8-11; and esp. lerodiakonou.
51
'Compounding" gives sunthesis, which is standardly translated as 'synthesis'.
52
See e.g. Aristotle, EN 1095a30-b4; Plato, Rep. 511BC (and for the 'return route'
see 521C, 532B); cf. e.g. Ammonius, in Porph. /sag. 37,7-13; Proclus, in Eucl.
43,18-21; Eustratius, in An. Post. 3,4-7.
 50 Translation
the conclusion until they bring the problem back to its principles.53
Again, if you reduce compound bodies to simple bodies, you use
analysis; and if you reduce each of the simple bodies to the things on
20 which their being depends - that is to say, to matter and form - you
are analysing.54 Again,55 if you divide speech into the parts of speech,
or the parts of speech into their syllables, or the syllables into letters,
you are analysing.56 If you reduce compound syllogisms to simple
ones you are said to analyse in a special sense of the word, and so too if
you reduce simple syllogisms to the premisses on which their being
25 depends.57 Again, reducing imperfect syllogisms to perfect ones is
called analysing.58 Again, the reduction of a given syllogism to its
appropriate figure is said to be analysis and it is in this sense of
analysis in particular that the books are entitled Analytics. For at the
end of the first book Aristotle outlines a method for us by means of
which we shall be able to do this59 - and he also explains how we shall
30 be able to effect the reduction of simple syllogisms to the appropriate
premisses on which their being depends.60 We shall also find him
saying61 how compound syllogisms come from simple ones and how
we may reduce the former to the latter.62
8,1 The books in which he studies the analysis of syllogisms he entitles
Prior Analytics; those in which he studies the analysis of demon-
strations he also calls Analytics -but Posterior Analytics.

1.2.2 Procedure
When you are giving instruction it is always63 very useful to state
53
The main texts on geometrical analysis are Pappus, 634-6, and a scholium to
Euclid XIII1-5 (printed in Heiberg's Euclid, IV 363-6); see also e.g. Albinus, Didasc.
157H; Philoponus, in An. Post. 162,16-28. Geometrical analysis is said to have been
encouraged by Plato: Diogenes Laertius, III 24; Proclus, in Eucl. 211,18-212,4. See
e.g. Hintikka and Remes; lerodiakonou, pp. 173-90.
64
See e.g. Aristotle, Gael. 300a7-ll; GC 329a20-4.
55
7,22-7 = FDS 1106.
56
See e.g. Apollonius Dyscolus, Synt. 265,9-10; 326,11-327,12.
57
cf. e.g. Ammonius, in An. Pr. 6,2-4; in Porph. /sag. 36,7-9; Philoponus, in An.
Post. 334,25-335,3; see lerodiakonou, pp. 166-73.
58
e.g. by Galen (Inst. Log. ix 1, 2; x 2). Aristotle does not use 'analyse' in precisely
this sense in An. Pr.; but he does use it in a closely related way, of reducing a
syllogism in one figure to a syllogism in another (e.g. 51a2, 3,18,22).
59
i.e. at An. Pr. 46b40-47bl4 (for analusis see 47a4): see Alexander, in An. Pr.
340,5-21 (with a reference to Theophrastus" work On the Analysis of Syllogisms: F 31
Graeser).
60
i.e. at An. Pr. 47bl5-50a4 (for analusis see 49al9); see Alexander, in An. Pr.
372,26-373,9. (At 7,30-31 we follow the punctuation proposed by Wallies on p. 711 of
his edition of in Top.)
61
Reading pou auton (Aldine) for tina autdn.
62
See An. Pr. 50a5-15 (but Aristotle does not use the word 'compound' of
syllogisms); Alexander, in An. Pr. 381,28-386,2; cf. lerodiakonou, pp. 21-3.
63
Reading aei (cj. Wallies) for dein.
 1. Preface 51
64
the aim and purpose of what you are going to say; for those who
know what each thing you say refers to learn more easily than those
who do not know. (The difference between such learners is like that 5
between people walking along the same road, when some know the
destination to be reached and others are ignorant of it: those who
know walk with more ease and accomplish their purpose without
exhaustion, whereas those who are ignorant tend to tire. In all cases
ignorance is like a sort of inexperience.65) Since this practice is 10
useful, Aristotle usually follows it in his other studies - and he does
so here too. At the very beginning he states what his purpose is; and
having done so quickly and briefly, he next tackles the items which
lead up to this purpose.66 For just as 'that for the sake of which' is a
starting-point for things which come about, so too is it for things
which are said; for that for the sake of which certain things are said, 15
i.e. their aim, is the explanation for the things said for its sake. And
if this is a starting-point, we should use it as a starting-point and
start from it. Moreover, if the aim is recognized, it will help those
who speak to judge what to say and those who learn to judge what
they hear.
At the same time, by saying that the purpose of his study of
syllogisms is to speak about demonstrations, he tells us that the 20
account of demonstration must be considered the primary product of
syllogistic method as a whole. For the study of the other forms of
syllogism is a matter for philosophers only to the extent that dealing
with them is useful for demonstration and for the discovery of what
is true: if you have exercised in dialectical syllogisms and can 25
recognize what is plausible and what lies close to what is true, then
you can more easily discover what is true, not being deceived by the
similarity of the plausible to the true but being aware of their
difference;67 and if you know how sophistical arguments come about
you will be able to guard against falsehood - and avoiding this helps
very greatly in the discovery of what is true.68
Aristotle does the same in the Ethics as he does here. There he 30
first says that 'our method aims at these things, being in a sense
political',69 and he states that his purpose has to do with
64
Alexander's general comment, and also the specific analogy with walking, allude
to Aristotle's celebrated remarks on Plato's lecture On the Good (see Aristoxenus, -II
30; cf. Aristotle, Metaph. 995a33-b2).
65
'Inexperience' gives apeiria. The word may also mean 'infinity'; and Alexander
may be punning. At Probl. 955b9-21, [Aristotle] asks why a road seems longer if we do
not know how long it is; and he replies that it is in a sense infinitely long (apeiros).
66
See An. Pr. 24alO-15; cf. below 10,9-23.
67
See Aristotle, Top. 101a27; Alexander, in Top. 27,24-31.
68
See Alexander, Conv. 59. Note the practice of the Stoic Zeno (Plutarch, Stoic.
Rep. 1034E); and cf. Ebbesen (1981a), I, pp. 88-9.
69
SeeAT1094blO-l.
 52 Translation
constitutions or political power. But only after ten books does he
give an account of these things,70 thereby suggesting that it is
9,1 necessary first of all to speak of human character and to say which
characters are required in those who are to make up a State, since
they are the primary parts of a State.71

70
i.e. at EN 1179a33-1181b23? Or does Alexander rather mean that the whole of
EN is, so to speak, a preface to Pol.?
71
See e.g. Aristotle, Pol. 1290b38-1291a8 (cf. Plato, Rep. 552A).
 2
Introductory Explanations

2.1 The goal of the Analytics1

First, to say about what and of what the inquiry is: it is about
demonstration and demonstrative science. [1.1,24alO-ll]
He states briefly what is the purpose and the aim of the science of 9,5
analytics as a whole.
Having proposed2 to say 'about what' ('about' demanding the
accusative case) and 'of what' (which is in the genitive case), he gives
his answer in the accusative case, saying only 'about demonstration
and science', and leaving it to us to reformulate his remarks for the
genitive case too. But in some copies3 -demonstrative science' is 10
written not with a nu but with a sigma;4 and with this reading he
will have replied to both the questions he propounded - to 'about
what' with 'about demonstration', and to 'of what' with 'of
demonstrative science'. If 'demonstrative science' is written with a 15
sigma, it is possible that his words 'about what' and 'of what' do not
both refer to the subject matter of the work. Rather, one of them
('about what') refers to the subject matter (for that which something
is about5 is the subject matter), and 'of what' refers to the disposition
which theorises about the subject matter. Thus the work is about
demonstration as its subject matter, and of demonstrative science,
which theorises about demonstration.6 In this way, he will have
1
The issue was disputed: some commentators said that the goal was
demonstration, others that it was the syllogism in general: see Ammonius, in An. Pr.
4,36-7; Philoponus, in An. Pr. 10,3-25; scholium to Aristotle, 139a36-140alO.
2
Reading protheis forprostheis.
3
For other references in Alexander to variant readings see e.g. in An. Pr. 144,4-6;
151,14-16; 210,30-2. For textual criticism in Alexander see Moraux (1986b), pp.
136-7.
4
i.e. in the genitive rather than in the accusative. All our MSS of An. Pr. have the
genitive, and this is the only reading mentioned by Ammonius (in An. Pr. 12,6-10)
and Philoponus (in An. Pr. 9,28-10,25).
5
Beading peri (Aldine) for pan.
6
Alexander means that the inquiry is 'of demonstrative science in the sense that
you exercise your demonstrative skills in pursuing it (so too Philoponus, in An. Pr.
9,28-32). But the genitive is certainly objective rather than subjective, and 'of

53
 54 Translation
20 given an appropriate reply to each of the questions he proposed.
A demonstration is a demonstrative syllogism; and demonstrative
science is the disposition by means of which it is possible to syllogize
demonstratively - for to speak about demonstrations7 is the mark of
demonstrative science and of someone who possesses it.

2.2 Propositions8

2.2.1 The general account

Then to determine what a proposition is, <and what a term, and
what a syllogism, and what sort of syllogisms are perfect and
what sort imperfect; and after that, what it is for this to be or not
be in that as in a whole, and what we mean by being predicated
of every or of none. > [1.1, 24all-15]
25 Since an account of the syllogism is indispensable for an account of
demonstration (as we have already said9), and syllogisms are
compounded of propositions and propositions of terms, Aristotle
reasonably speaks about the items on which the being of a syllogism
depends before speaking about syllogisms themselves. (In the same
way it is not possible to know about words unless you know about
30 syllables and letters.) Next he proposes to speak about syllogisms,
indicating to us that it is because of syllogisms that he has given an
account of these items - just as he has given an account of
10,1 syllogisms because of the account of demonstration. Proceeding
further, he will make known which syllogisms are perfect and which
imperfect. And he will explain that 'to be in as in a whole' and 'in no'
(for this is what is meant by 'or not be') mean the same as 'of every'
5 and 'of no'. And 'of every" and 'of no' mean universal affirmation and
universal negation, as he will go on to explain. But since, in his
presentation of syllogisms and in general of combinations in the
figures, he will make use of 'one thing is in another as in a whole'
and 'in no', for this reason he mentions these items first.10
10 Now a proposition is an utterance affirming or negating
something of something. It is either universal or particular or
indeterminate. <By universal, I mean holding of all or of none,
demonstrative science' simply means 'about demonstrative science' (cf. An. Post.
99bl5-17: see e.g. Ross, p. 288; Mignucci (1969), p. 181). But note the subtle
interpretation in Brunschwig(1981).
7
As a matter of fact, demonstrative sciences do not speak about (peri)
demonstrations. Hence we toyed with emending pert to dia (or meta): demonstrative
sciences speak through or with demonstrations.
8
On Alexander's account of propositions see Lee, pp. 55-8.
9
Above, 6,32-7,11.
10
See below, 24,23-25,11 and note to 25,11.
 2. Introductory Explanations 55
by particular holding of some or not of some or not of all, by
indeterminate holding or not holding, without being universal
or particular ...> [1.1, 24al6-20]
One account of propositions will be the account Aristotle gave of
assertions in On Interpretation, namely: an utterance 'in which
there is truth or falsity'.11 But, in fact, he gives a definition specific 15
to propositions; for even if propositions and assertions are the same
in what underlies them, they differ in account:12 insofar as they are
either true or false, they are assertions;13 insofar as they are
expressed affirmatively or negatively, they are propositions. Or: for
something to be an assertoric utterance is simply a matter of its
being true or false; for something to be a proposition is a matter of 20
how it contains truth and falsity. This is why propositions which do
not contain truth and falsity in the same way are the same
utterance but not the same proposition. The proposition which says
that justice is good is similar to the proposition which says that
injustice is bad - they are both true and both affirmations; but they
are not the same proposition, since their subjects and predicates
differ. Again, the affirmation and the negation, which are true,14 are 25
similar in this respect and are the same utterance; but they are not
the same proposition, since the quality of the assertion is different
in them: they are not the same proposition but they are the same
assertion.
He seems to define propositions in terms of what falls under
them.15 The reason is this. In On Interpretation he said that 'the
primary single assertoric utterance is affirmation; and next is 30
negation'.16 Now where things are ranked - where one of the
subordinate items is primary and one posterior, so that the primary
shows up in the posterior - there the predicate which they share is
not predicated of them as a genus.17. For in genera the proximate 11,1
species are co-ordinate with one another, and it is not the case that
one of them is primary and another posterior. Now since proposi-
11
See Int. 17a2-3.
12
For the difference between protasis and apophansis see e.g. Ammonius, in An.
Pr. 13,17-14,4; 15,30-16,9; Philoponus, in An. Pr. 11,25-36; cf. Maier II 2, pp. 359-66;
Lee, p. 32.
13
cf. e.g. Ammonius, in Int. 2,21-5; Boethius, Int. Syll. Cat. 767C.
14
The phrase is obscure, but Alexander must have in mind some particular pair of
affirmation and negation; and it is most likely that he expects us to supply the
affirmation of 10,22 ('Justice is good") with an appropriate negation (e.g. 'Justice is
not bad'): cf. Alexander, in Top. 12,10-24. For a different interpretation of the
argument see Volait, pp. 1-3.
15
i.e. in terms of affirmation and negation, which might be taken to be the species
of proposition.
16
See Aristotle, Int. 17a8-9.
17
For the thought see e.g. Alexander, DA 16,18-17,8; in Metaph. 208,28-209,34;
and note Aristotle, DA 414b20-32; Metaph. 999a6-13; EE 1218al-8.
 56 Translation
tions are not of a genus or nature distinct from the things of which
they are predicated and on which their being depends (as is the case
5 with genera),18 for this reason he makes clear what they are by way
of affirmation and negation, of which they are predicated.
What he said in On Interpretation - namely, 'Thus a simple
assertion is a significant sound concerning something's holding or
not holding'19 - is equivalent to the present account. Equivalent too
is what he says in the Posterior Analytics: 'The one part of a
contradictory pair,20 one thing of one thing.'21 (In this definition he
10 no longer added an indication of quality or of times, because these
are comprised in the notion of a contradictory pair - contradictory
pairs actually involve these features.22) Similar, too, is: 'An
utterance assertoric of something about something'23 - except that
this is more general, since in it the notions of affirmation and
negation have not yet been introduced.
15 (Theophrastus, in his On Affirmation, seems to think of
'proposition' as having several senses. At any rate, he does not
define propositions but rather affirmation and negation.24)
These definitions apply not to all propositions but to simple and
so-called predicative propositions. Saying something of something,
and being universal or particular or indeterminate, are features
peculiar to predicative propositions; for in hypothetical propositions
20 truth and falsity depend not on something's being said of something
but rather on implication or conflict.25 Moreover, the definition he
18
Alexander's view (which may be Theophrastean in origin: below, n. 24) that prota-
sis is not the genus of affirmation and negation was criticised by Ammonius (in Int.
15,16-30; 66,14-28; 67,30-68,9; 80,15-35), Philoponus (in An. Pr. 12,23-20,3), and Boe-
thius (in Int. ed. sec. 16,6-18,26; 119,10-125,27, which cites Porphyry as rejecting
Alexander's view). Philoponus and Boethius report Alexander's arguments at some
length, presumably drawing on his lost commentary on Int. (Note that in Alexander's
viewprotasis does function as a genus for problems, axioms, and so on: below, 44,16-23.)
19
See Int. 17a23-4 (Alexander has oun where our MSS of Aristotle have d'h men).
20
Reading antiphaseos for apophanseds: the emendation is demanded by 11,10-11.
21
See An. Post. 72a8-9.
22
The definition of'contradictory pair' includes an explicit reference to affirmation
and negation (An. Post. 72al3-14); hence there is no need to make an express
reference to 'quality' in the An. Post, definition of propositions. What about 'times?
Aristotle's account of'simple assertion' contains the clause (which Alexander does not
cite) 'in one of the divisions of time' (Int. 17a24): hence his definition of'contradictory
pair' implicitly refers to times; and hence - perhaps - no explicit reference to time is
needed in the Are. Post, definition.
23
This account is not found in Aristotle.
24
See Theophrastus, F 2 Graeser. See Bochenski (1947), pp. 40-1, who argues that
Theophrastus was the author of Alexander's view (above, n. 18) that protasis is not
the genus of affirmation and negation.
26
The distinction between predicative (or 'categorical') and 'hypothetical'
propositions was a commonplace by Alexander's time: see e.g. Albinus, Didasc. 158H;
Galen, Inst. Log. iii 1; Apuleius, Int. 177,3-10. Roughly speaking, a hypothetical
proposition is either a conditional ('implication') or a disjunction ('conflict'): see
Introduction, p. 23; cf. e.g. Sullivan, pp. 24-30.
 2. Introductory Explanations 57
gives does not apply to dialectical propositions (for they involve
requests) but only to syllogistic propositions.26
He used 'something of something'27 of negations too, since in
negations too there is a subject term and a predicate term. For here
'of something' means not that something is affirmed but that 25
something is predicated - and it may be predicated either
affirmatively or negatively. In On Interpretation, however, when
denning affirmation and negation, he used 'of something" for
affirmations - for negations he used not 'of something' but rather
'from something', saying that a negation asserts 'something from
something".28
There are two things by which predicative propositions are 30
determined, namely quality and quantity, and he introduces both of
them into his account of propositions: quality, in saying that a
proposition is an utterance which is affirmative or negative (this is
the primary and common quality of propositions); quantity, in
saying 'It is either universal or particular or indeterminate'. Here he 12,1
sets out the meanings of these words - in On Interpretation he
discussed them at greater length.29 Note that he does not present
universals in the same way here and in the Posterior Analytics.30

2.2.2 Demonstrative and dialectical propositions31

A demonstrative proposition differs from a dialectical
proposition in that a demonstrative proposition is the 5
assuming of one part of a contradictory pair <(for if you are
demonstrating you do not request but assume), while a
dialectical proposition is a request for one of a contradictory
pair. There will be no difference with regard to there being a
syllogism in each case. For whether you are demonstrating or
requesting, you syllogize by assuming that something holds of
something or that something does not hold. Hence a syllogistic
proposition simpliciter affirms or negates something of
something in the way already mentioned. It is demonstrative if
it is true and is assumed by way of the initial hypotheses. It is
dialectical if it is a request for one of a contradictory pair (if you
26
For dialectical propositions see the next section, and esp. 13,30-14,5: it is plain
that here Alexander uses 'dialectical proposition' in the first of the two senses which
(following Aristotle) he later distinguishes.
27
Retaining ti (del. Wallies) before kata.
28
See Int. 17a25-6 (with Ammonius, in Int. 80,31-5); cf. Ammonius, in An. Pr.
17,10-32.
29
Alexander is presumably referring to Int. 17a38-bl6.
30
See An. Post. 73b26-7: 'I call universal whatever holds of every case and in its
own right andgua itself See also below, 25,11-13.
31
See above, 7,9 note.
 58 Translation
are inquiring) or an assumption of what is apparent and
reputable (if you are syllogizing) - as has been said in the
Topics.>[l.l,24a22-b3]
Having given his general account of propositions, he adds the
differences by which demonstrative and dialectical propositions are
distinguished from each other, thereby showing that the general
10 account of propositions which he has given applies to them and that
it is not in this respect that they differ. The account he has given
applies to all syllogistic propositions. But it does not apply to
requests: requests are not propositions simpliciter but dialectical
propositions, the phrase being taken as a whole. Now when he
defined propositions in the Topics,32 he gave an account to the effect
that dialectical propositions are reputable requests.33 So lest anyone
15 should think that this is the account of propositions in general, he
marks the distinction here and makes clear the differences among
propositions, showing that the account of propositions in the general
sense - i.e. of propositions assumed for syllogizing in general - is the
account given here and not the account stated in the Topics; for the
latter is an account of dialectical propositions, which are not as yet
syllogistic.
20 The differences among propositions correspond to the differences
among syllogisms, there being a general account of propositions as
there is of syllogisms. In the Posterior Analytics34 he describes at
greater length how demonstrative propositions differ from other
propositions: they differ both with regard to their matter (for they
are true and prior and known and necessary), and also in the way in
25 which they are used and assumed (for someone who is
demonstrating does not make requests).35 Here he indicates their
difference from dialectical propositions first by the manner in which
they are assumed: demonstrative propositions do not involve
requests, i.e. are not assumed by way of a request. Someone who
requests one of a contradictory pair does so inasmuch as he is
prepared to object to anything supposed by his respondent and to
30 reject it by way of the syllogism which he will construct from what
the respondent grants. Such is the dialectician, who syllogizes from
reputable propositions; for it is possible to obtain reputable
propositions even for opposites.36 But if you are demonstrating you
will not demonstrate from anything which is granted; for it is not
32
Aristotle, Top. 104a8-9; cf. Alexander, in Top. 69,13-70,11; Ammonius, in An. Pr.
19,3-21,30; Philoponus, in An. Pr. 22,23-23,17.
3
- Por endoxa or 'reputable' items see Aristotle, Top. 100a29-30 (cf. e.g. Metaph.
995b23-4); and see Barnes (1980b).
34
Aristotle, Are. Post 71b20-72a24.
36
On these two modes of difference see Ebbesen (1981b), p. 5.
36
i.e. there are cases in which both 'P' and 'not-P' are reputable.
 2. Introductory Explanations 59
possible for opposites to be true. So if you demonstrate you do not
request but make your own assumptions: you assume what follows 13,1
from the principles and hypotheses. We see this to be the case in
geometry; for if you want to prove that the three angles of a triangle
are equal to two right angles you do not prove it by way of propositions
granted after a request. (Aristotle says here that a dialectical
proposition is 'a request for one of a contradictory pair' in the sense of 5
a proposition assumed by way of a request for one of a contradictory
pair this is clear from the next phrase 'There will be no difference
...'. For someone who makes a request syllogizes on the basis not of
the request but of what is assumed by way of the request.) The
principles of demonstrations are hypotheses; for there is no demon-
stration of such propositions, i.e. of principles, but they are posited as
obvious and directly known or else as following from propositions of 10
this type; and something which is assumed without demonstration
they call a hypothesis (and also, more generally, a posit37) and say
that it is hypothesized.38 It is on items of this sort that demon-
strations are based.
Having said how they differ in use, he next adds their difference in
respect of matter, having first shown that the account of propositions
which he gave is general and applies both to demonstrative and to
dialectical propositions when they are assumed for the generation of 15
a syllogism. (He says: 'There will be no difference with regard to there
being a syllogism in each case. For whether you are demonstrating or
requesting, you syllogize by assuming that something holds of
something or does not hold' - and this is the account he gave of
propositions.) For although the manner in which they are assumed is
different, still the assumptions share the general account of proposi- 20
tions. Now it is dialectical propositions as assumed, not as requested,
to which the general account of propositions applies; and what holds
generally of propositions, whether assumed demonstratively or
dialectically, is the fact that they affirm or negate something of
something. It is propositions in this sense which we use for syllogisms
in the general and unqualified sense; for it is the fact that they are
affirmative or negative which is useful for syllogisms. 25
When he says 'something holds of something' and 'something39
does not hold', his expression makes clear to us that 'of something'
applies to negations too. And when he adds to the next sentence 'in
the way already mentioned', he seems to mean that 'of something' is
common to both sorts of proposition.40 Or else 'in the way already 30
37
cf. Ammonius, in Int. 9,7-10.
38
On this broad sense of 'hypothesis', which became common in later Greek, see
Barnes (1990c), ch. 4.
39
Reading kai ti with the MSS: Wallies prefers e (Aldine, Aristotle).
40
See above, 11,22-6.
 60 Translation
mentioned' means 'universal or particular or indeterminate'.41
He also makes a division among dialectical propositions: he
14,1 distinguishes between the dialectical proposition of someone inquir-
ing and requesting ('a request for one of a contradictory pair'), which
consists in demanding an answer from someone else and which is not
yet an assertoric utterance nor an affirmation or a negation, and the
dialectical proposition of someone syllogizing, which is part of the
syllogism being constructed and which depends upon its matter; for a
5 request for one of a contradictory pair is not part of a syllogism.
In the Topics*"* he has said what he means by 'apparent and
reputable.'

<In what follows we shall say with precision> what a proposi-

tion is and how syllogistic and demonstrative and dialectical
propositions differ. [1.1,24bl2-14]
He has said that propositions are utterances affirming or negating
10 something of something, and this is his account of syllogistic
propositions. Demonstrative propositions are, as he puts it here,
'true and ... assumed by way of the initial hypotheses',43 whereas in
the Posterior Analytics he says that they are true and immediate
and primary and better known than and explanatory of the
15 conclusion.44 A dialectical proposition is 'a request for one of a
contradictory pair (if you are inquiring) or an assumption of what is
apparent and reputable (if you are syllogizing)'.45 There are also
sophistical propositions - which he does not mention at this point -
which are assumptions of what is apparently reputable.46 For
propositions in which the plausibility is superficial and rather easy
to detect47 are sophistical propositions.
He discusses them 'with precision' - demonstrative propositions
20 in the Posterior Analytics, dialectical propositions in the Topics, and
sophistical propositions in the Sophistical Refutations. These books
all follow the general study of syllogisms. By saying 'In what follows
we shall say' he makes it clear that the present books are prior in
order to those studies.

41
i.e. it refers back to An. Pr. 24al7.
42
Top. 100b21-3: 'Reputable opinions are those held by everyone or by most people
or by the wise (and either by all the wise or by most of them or by the best known and
most reputable)'.
43
An. Pr. 24a30-bl.
44
An. Post. 71b21-2.
45
An.Pr. 24blO-12.
46
Omitting kai before endoxou with LM: see Aristotle, Top. 100b23-5.
47
i.e. it is relatively easy to discern that the propositions are merely plausible.
 2. Introductory Explanations 61
2.3 Terms

2.3.1 Subjects and predicates

I call a term that into which a proposition resolves <- i.e. the
predicate and that of which it is predicated ...>[!.!, 24bl6-17]
By saying 'I call a term' Aristotle seems to indicate that the word 25
'term' was not in common use or well known in its application to a
part of a proposition.48
He deals first with propositions because he is going to give his
account of terms by way of propositions. The terms in a simple
proposition are noun and verb.49 The parts of propositions are called
terms because propositions are determined by50 and compounded 15,1
from them. The term of which something is predicated is the subject
term; and about it the utterance is made. The term which is
attached to the subject and said about it is the predicate.51 For every
predicative proposition is composed of a subject term and a
predicate term.

2.3.2 The copula

<... (when you add or remove 'is' or 'is not').> [1.1, 24bl7-18]52
He adds 'when you add or remove53 "is" or "is not"' lest out of 5
ignorance people should divide up propositions in which 'is' is
co-predicated as a third item,54 and should think that 'is' or 'is not' is
a third term. For in propositions of this kind 'is' is not a term.
Rather, when it is added, it signifies a compounding of the predicate
and the subject and betokens an affirmation; and when it is said in 10
the negative form, it removes and separates the terms from one
another and betokens a negation. Every proposition is divided into
48
cf. Ammonius, in An. Pr. 14,13-22; scholium to Aristotle, 146a9-18.
49
cf. Apuleius, Int. 178,1-18; Galen, Inst. Log. ii 2-3; Ammonius, in An. Pr. 14,22-5;
22,18-24 (the ultimate source is Plato: see Theaetetus 206D; Sophist 262D; Cratylus
431B; but note, in Aristotle, Int. 16al-2; 17a9-10). This grammatical analysis of
propositions is at odds with the standard logical analysis into subject term and
predicate term: the predicate term is not a verb it has the same logical syntax as the
subject term, with which it is interchangeable (see Barnes (1983)).
60
They are called terms, horoi, because they are determined, horizontal.
51
cf. e.g. Ammonius, in Int. 7,32-8,4; in An. Pr. 22,12-30.
52
The interpretation of this clause was (and is) controversial: see esp. Ammonius,
in An. Pr. 22,34-24,24; Philoponus, in An. Pr. 25,30-30,20. (Ammonius and Philoponus
discuss and reject Alexander's interpretation, preferring the more elaborate
suggestion put forward by lamblichus.) See also Ross, pp. 290-1, who cuts the knot by
excising e diairoumenou.
53
For this use ofdiairein see e.g. Ammonius, in Int. 27,1-3.
54
See Aristotle, An. Pr. 25b22; Int. 19bl9; cf. esp. Ammonius, in Int. 165,4-30 (cf.
8,5-8); in An. Pr. 23,25-24,5.
 62 Translation
two terms. That 'is' is neither a term nor part of a term55 seems to be
clear from the fact that opposite affirmations and negations are
compounded from the same terms.
15 But 'is' does seem to be a term when it stands by itself, i.e. when it
is predicated of the subject, as in 'Socrates is'.56 For here the 'is'
seems to be the predicate term. But in fact, if you consider the case
precisely, 'is' is not even here a term in its own right. For the
proposition which says 'Socrates is' is equivalent to 'Socrates is
being', in which not 'is' but 'being', which goes with 'is',57 forms the
20 predicate term. Since 'is' seems to be equivalent to 'being' (for it is an
inflected form of it), for the sake of brevity and in order not to appear
to say the same thing twice it alone is connected to the subject. And
when it is connected in this way it becomes a term and a part of the
proposition.
After 'when you add58 or remove', he adds ' "is" or "is not" '. The
addition of 'is not' seems to make the expression somewhat
25 incongruous; for 'when you add or remove "is" ' seems to be in itself
sufficient, since the addition of 'is' makes an affirmation and its
removal a negation. But in fact he adds 'is not' in order to make clear
how a negation comes about when 'is' is removed. For you remove 'is'
30 both if you delete it completely and also if you connect it with the
negative particle, to make 'is not'. If you subtract 'is' completely from
the proposition and remove it in this sense, then what remains is not
an affirmation nor indeed a proposition at all. For if you subtract 'is'
from the proposition 'Socrates is white' and leave 'Socrates white',
16,1 you have not made a negation nor even kept what remains as a
proposition. For propositions are complete utterances which lack
nothing.59 Hence in order to show the way in which you should
remove 'is' in order to get a negation, he added 'is not', so as to mean
'when you add "is" in affirmations, and remove the same thing in
5 negations by making it "is not" '. For 'is' is not completely removed
and subtracted.
Or else he adds 'when you add or remove "is" or "is not" ' in order
to indicate that these items, i.e. 'is' and 'is not', are neither parts of
the proposition nor terms, but that both 'is' and 'is not' are external
to the terms, either being added externally to the predicate terms
56
But see e.g. Alexander, in An. Pr. 406,32-5 (below, 44,28 note).
56
The observation was made by Eudemus (scholium to Aristotle, 146al9-27 = F 27
Wehrli); see e.g. Ammonius, in Int. 14,18-26; 57,29-32 (cf. Boethius, Int. Syll. Cat.
769A): Ammonius repeats Alexander's analysis, which derives ultimately from
Aristotle (see Int. 21b9-10; Metaph. 1017a27-30).
67
Reading to on <to> meta tou estin.
58
59
Readingprostithemenou for suntithemenou.
See e.g. Ammonius, in Int. 44,11-19; 62,10-12; Apuleius, Int. 176,13-14. Compare
to the Stoic definition of an axioma as a lekton autoteles: Sextus, PHII 104; M VIII
70-1; Diogenes Laertius, VII65; Varro, cited by Aulus Gellius, XVI viii 1-8.
 2. Introductory Explanations 63
when propositions are divided into terms or else being separated 10
from them. For the addition or subtraction of these items
contributes nothing to the division of propositions into terms: the
terms in 'Socrates is white' seem to be 'Socrates' and 'white'. His
expression would be more congruous if it were put60 in this way:
'when you add "is" or "is not" or remove them'.
Or is it absurd to claim that the 'is' in these propositions is not 15
predicated in any sense? (Eudemus, in the first book of his On
Expression, shows this at some length.61)

2.4 Syllogisms62

2.4.1 The premisses

A syllogism is an utterance in which, certain things being
posited, something different from the suppositions comes 20
about by necessity inasmuch as they are the case. <By
'inasmuch as they are the case' I mean that it comes about
because of them; and to come about because of them is to need
no external term for the generation of the necessity.> [1.1,
24bl8-22]
Having discussed propositions and the parts of propositions (which
he called terms), he duly presents his account of what a syllogism
is.63 Now that a syllogism is a sort of utterance - that it falls within
the genus of utterances is clear.
He first takes its genus64 - the genus is what betokens the
common nature of the definienda, by which they are distinguished 25
from what differs from them in genus. This is the point of positing
the genus for the definienda; for the definiendum must be
completely distinguished by the definition. Of the things which
differ from the definiendum, some are further away (namely, those
which do not even share in the same genus), and others are closer
and belong to the same genus. Now the genus itself distinguishes it 30
from items in a different genus, while the differentiae distinguish it
from items in the same genus. That is why the genus is taken first in
definitions, since it distinguishes items which differ most from one
60
Retaining legoito (Wallies emends to legoi to).
61
See Eudemus, F 26 Wehrli; cf. the scholium to Aristotle, 146a24-7 [= F 27 Wehrli]:
'Eudemus, in the first book of his On Expression shows at some length that the "is" in
simple propositions (such as "Socrates is", "Socrates is not") is predicated and is a
term.' (Thus the final paragraph of this section refers back to 15,14-22.)
62
With sections 2.4.1-4 compare in Top. 7,10-15,14.
63
For Aristotle's definition see also Top. 100a25-7; SE 164b27-165a2; Rhet.
1356bl6-18; cf. Kapp; Frede (1974b), pp. 110-22; Barnes (1981), pp. 21-5.
64
With this account of definition by genus and differentia cf. e.g. Alexander, in Top.
421,3-34.
 64 Translation
another. Well, then, having taken the genus of the syllogism,
17,1 Aristotle next shows the appropriate and peculiar essence of the
syllogism by taking the differentiae of utterances and compounding
them with the genus.
By 'certain things being posited', he indicates that some
propositions must be assumed or conceded by whoever is syllogizing
- something which is not the case with every utterance. For what is
65 66
5 posited in a prayer or a command or a request or an invocation?
Some people think that 'posited' does not mean simply 'assumed' but
makes clear what sort of thing must be assumed.67 For that the
propositions assumed for a syllogism must be predicative68 is - they
say - shown by the word 'posited' (it is these propositions which
Aristotle has denned); for he will prove that hypothetical
propositions, in and of themselves, do not produce syllogisms.69 This
10 - they say - is why he says 'posited' and not 'assumed'.
Furthermore,70 that there must be not just one posit or one
proposition but at least two,71 is made clear by the plural 'things
posited'.72 For the 'single-assumption' arguments, as the more
recent thinkers73 call them, are not syllogisms. A syllogism by its
very name betokens a sort of compounding of utterances. For the
prefix sun- means a conjunction or compounding or collection of the
15 things it is added to. Thus suntrekhein ['run in company'] signifies a
compounding or concourse of runners; summakhein ['fight
alongside'], a combination of fighters; sumpalaiein ['wrestle with'], a
group of wrestlers; suze'tein ['investigate with'], a group of
investigators. So too sumpsephizein ['compute'] signifies a compoun-
ding of counters. And thus, a syllogism is a compounding of
utterances.74
66
Adding & en tei erotesei after entole (see Introduction, p. 16 n. 83).
66
For the five species of logos which the later Peripatetics distinguished see e.g.
Ammonius, in Int. 2,9-3,6; 5,1-23; 64,26-65,2; in An. Pr. 26,30-3; Bosthius, Syll. Cat.
797BC. The ideas derive ultimately from Aristotle, Int. 17a4. For full details see
Schenkeveld.
87
So Alexander himself, in An. Pr. 348,29-32; in Top. 8,8-14; the interpretation is
rejected by Ammonius, in An. Pr. 27,6-14; cf. Philoponus, in An. Pr. 33,6-10.
68
See above, 11,17-20 and note.
69
See Aristotle, An. Pr. 50al6-28.
70
17,10-18,7 =FDS 1051.
71
And also at most two: e.g. Aristotle, An. Pr. 41b36-42a40; Alexander, in An. Pr.
257,6-12.
72
cf. Alexander, in An. Pr. 257,8-13; in Top. 8,14-9,19; Ammonius, in An. Pr.
27,14-33; Philoponus, in An. Pr. 33,10-23. That a syllogism requires at least two
premisses is explicit in Aristotle: An. Pr. 34al6-19; 40b30-7; 53bl6-23; An. Post.
73a7-ll.
73
See above, 3,3 note.
74
i.e. sullogismos is derived from sun- and logos (cf. Alexander, in Top. 9,17-19;
Ammonius, in An. Pr. 26,2-6; Philoponus, in An. Pr. 2,14; 64,11-15). The etymology is
false: sun- is rather compounded with logismos, 'calculation'.
 2. Introductory Explanations 65
75
The so-called single-assumption arguments are sometimes
thought to be syllogisms inasmuch as when you hear them you add 20
the second premiss because it is well known. For
You are breathing,
Therefore: you are alive,
seems to be a syllogism because the hearer himself adds the second
premiss:
Everyone who is breathing is alive,
which is well known. Were this not well known, no one would
concede the conclusion
Therefore: you are alive,
on the basis of
You are breathing,
- rather, they would demand a reason. A syllogism is supposed to be
the assembling of a conclusion from the premisses,76 and this cannot 25
be done from a single premiss.77
Even if something quite certainly follows from some single item by
necessity, it does not thereby follow syllogistically. For it is not the
case that, because a syllogism has a conclusion which follows
necessarily from the suppositions, then if one thing follows
necessarily from another it follows syllogistically as well. For
necessity is of wider extension than the syllogism.78 Thus in 30
implications and relations and in sound continuous79 or conditional
propositions, the consequent follows the antecedent by necessity,
but they are not thereby syllogisms. In the case of
If it is day, it is light,
its being light follows by necessity from its being day - but not
syllogistically. Similarly, to err follows from to steal, and to wrong
from to assault; but these are not therefore syllogisms. 18,1
It is the proper characteristic of a syllogism that it proves
necessity from a plurality of items. That is why even if there is a
75
On which see below 21,25; Alexander, in Top. 8,14-9,19; 13,25-8; 574,10-15;
Sextus, PH II 166-7; M VIII 441-3; Apuleius, Int. 184,19-23; Ammonius, in An. Pr.
27,14-33; Philoponus, in An. Pr. 33,10-23; Boethius, Syll. Cat. 821BC; scholium to
Aristotle, 147b47-148a2. See Mates, p. 66; Mueller (1969), pp. 175-7; Frede (1974a),
pp. 118-19; (1974b), p. 114; Maroth, pp. 84-5. (17,18-22 = SVF III Ant 27.)
76
cf. Ammonius, in An. Pr. 26,6.
77
What, then, is the status of'conversions'? Within Aristotelian syllogistic, 'B holds
of some A' follows from 'A holds of every B'; and this appears to be a
single-assumption argument. The ancient texts give no satisfactory explanation of
this: see Barnes (1990a), III.8.
78
Aristotle, An. Pr. 47a22-3; cf. Alexander, in An. Pr. 344,9-13.27-31; 379,20-1; in
Top. 9,5-8. See Lee, p. 98; Maconi, p. 95; Barnes (1990a), III.6.
79
'Continuous (sunekhes)' was the word used by 'the earlier thinkers' (i.e. probably
by the early Peripatetics) for conditional propositions: Alexander, in An. Pr. 262,32-5;
390,3-4; Galen, Inst. Log. iii 3; cf. Bochenski (1947), p. 108; Frede (1974a), p. 16. (The
idea is that a conditional proposition links two things together and so makes them
continuous, whereas a disjunction separates them.)
 66 Translation
plurality of expressions for what is posited, still, if the rest80 signify
the same as the first, there will be no syllogism from them in such a
case. For an argument of this sort is virtually a single-assumption
argument - e.g.
5 It is day.
But it is not the case that it is not day.
Therefore: it is light.
For 'It is not the case that it is not day' differs from 'It is day' only in
expression.81

2.4.2 The conclusion

This, then, is what 'being posited' means. As for 'something different
from the suppositions comes about by necessity', this separates
10 syllogisms from inductive arguments; for it means that the
conclusion necessarily follows from the suppositions - something
which holds neither of paradigmatic nor of inductive arguments,
even though in them too you posit some thing or things.82
Aristotle83 was quite right to add that the conclusion must be
something different from what has been posited.84 It is useless85 - it
destroys any syllogistic utility - to infer what is agreed and
15 supposed. No syllogistic utility is provided by the argument:
If it is day, it is light,
But it is day,
Therefore: it is day86
nor generally by those which the more recent thinkers call
non-differently concluding arguments.87 (These include the dupli-
cated arguments,88 such as
If it is day, it is day.
But it is day.
80
Reading talla for tauta.
81
But at in Top. 12,31-13,10, Alexander argues that T' and 'Not-not-F do not mean
the same thing primarily (proegoumenos). On double negations, or huperapophatika,
see Diogenes Laertius, VII69; cf. O. Becker, p. 29; Frede (1974a), p. 72.
82
See below, 43,17-44,2.
83
18,12-24 = FDS 1171; 18,12-18 = SVFII261.
84
cf. Ammonius, in An. Pr. 27,34-28,20.
85
See above, 2,34 note.
86
Reading hemera forphos with B and the Aldine (so Wallies, in his corrigenda: in
Top. p. 711); cf. Frede (1974a), p. 184 (Lee, p. 99 n. 8, offers an alternative).
87
On these arguments see Alexander, in An. Pr. 164,30; in Top. 10,8-13; 566,25-7;
Galen, PHP V 224; Apuleius, Int. 184,23-31; Ammonius, in An. Pr. 28,5-9; 32,12-14;
scholium to Aristotle, 294b23-9. See Frede (1974a), p. 184 n. 21.
88
On these arguments see below, 20,10-12; Alexander, in Top. 10,7-10; Apuleius,
Int. 184,23-31; Ammonius, in An. Pr. 27,35-28,5; 32,13; Philoponus, in An. Pr.
33,23-6; Boethius, Syll. Cat. 821C-822A. (On 'duplicated' propositions add: Sextus,
PHII 112; M VIII 93-5, 108-9; Diogenes Laertius, VII 68.) See Frede (1974a), p. 50
n. 5; (1974b), pp. 116-18.
 2. Introductory Explanations 67
Therefore: it is day.)
This sort of thing may indeed have a syllogistic figure and a
syllogistic combination, but it is certainly not a syllogism. For 20
syllogisms are instruments, and they are introduced for some useful
purpose - for proving something. Hence what is not useful is not a
syllogism.
We shall see that arguments of this kind are useless if we run
through the species of syllogism89 and ask in which of them it could
be appropriate to infer one of the suppositions. Is it appropriate in
demonstrative syllogisms? - No: they attempt to reveal what is
unclear by way of what is clear and known, what is posterior by way 25
of what is primary. In dialectical syllogisms? - No: what an
interlocutor is unwilling to concede they attempt to prove by way of
what is reputable and conceded by him, bringing him to a
contradiction. In eristic syllogisms? No: their purpose, too, is to
bring the respondent to a contradiction or an apparent contradiction
on the basis of what he grants. They do not conclude to what he 30
grants: they conclude, on the basis of what he grants, to what he is
unwilling to grant. Clearly, then, they will infer something different 19,1
from what is granted. Now if a genus exists in its species,90 and if
the syllogism is the genus of its species, and if in none of the species
is what is inferred the same as what is assumed, then this will not
be so of syllogisms in general.
As91 for disjoint syllogisms constructed from contradictories,92
they do not infer their conclusions qua the same as the
re-assumption or (as the more recent thinkers put it) the 5
co-assumption.93 If you say:
89
On the kinds of syllogism see above, 7,9 and note. For the form of argument
which Alexander uses here see Aristotle, Top. 121a27-39; Alexander, in Top.
302,23-304,6.
90
cf. e.g. Alexander, in Top. 321,26-8; DA 90,2-9. Alexander's view on universals is
controversial: see e.g. Lloyd (1981), pp. 49-61; Tweedale; Sharpies (1990), pp. 101-3.
91
19,3-12 = SVF II261.
92
With the following paragraphs cf. Alexander, in Top. 10,30-13,10; Lee, pp. 101-3.
By a 'disjoint' syllogism Alexander means a two-premissed argument in which one
premiss is a disjunction ('Either P or Q') and the other premiss is one of the disjuncts
(i.e. 'P', or 'Q'). Disjunction is construed exclusively: Either P or Q - but not both (but
see below, 66,10 note). Hence the syllogism yields the conclusion 'not-Q' (or: 'not-P'). A
disjoint syllogism is 'constructed from contradictories' if the two disjuncts are
contradictories of one another. See further below, notes 99 and 102.
93
'Co-assume', proslambanein, is a standard term in Stoic logic (e.g. Diogenes
Laertius, VII 76; cf. Frede (1974a), p. 118 n. 1 - the verb is also found in Aristotle: e.g.
An. Pr. 40b36); 're-assume', metalambanein, was the corresponding Peripatetic term
(Alexander, in An. Pr. 263,26-36; Ammonius, in Int. 3,19-20; [Ammonius], in An. Pr.
68,7-13; Philoponus, in An. Pr. 243,8-10 - who notes, however, that proslambanein
had come into general use). Metalambanein derives from Aristotle (see e.g. An. Pr.
41a39). It was used in a variety of other senses (see e.g. Alexander, in An. Pr. 262,5-9;
386,5-30). See also Volait, pp. 86-9; Ross, p. 373; and also - for its use in the
grammarians Sluiter, pp. 111-17.
 68 Translation
Either it is day or it is not day,
and then co-assume one of the disjoints - either the negative:
But it is not day,
or the affirmative:
It is day,
- you do indeed deduce either:
Therefore: it is not day,
or:
Therefore, it is day.
10 And this does seem to be the same as what was co-assumed - either
as:
But it is not day,
or as:
But it is day.
It is not, however, inferred qua being the same as the
co-assumption, but rather qua contradictory of the other disjoint. In
syllogisms of this sort it so happens that this contradictory is the
same as the co-assumption. But there is a world of difference
between taking as the conclusion from the outset something which
15 is the same as one of the suppositions and taking as the conclusion
something qua something different which then turns out to be the
same.
Now inferences of this sort never occur unless the contradictory of
one disjoint94 is the same as the co-assumption; and so in these
cases too,95 it is taken qua different - since being this particular
proposition is indeed different from being the contradictory of the
other disjoint. (They are the same in expression but not in force.96)
20 And so it is in these cases - as97 the Stoics too must admit, since
they say that in disjoint or disjunctive syllogisms it is by the
co-assumption of one of the disjuncts that the contradictory of the
remaining disjunct98 follows.99 But if this is so, there will no longer
be a syllogism, according to them, where one element in the
disjunction is co-assumed and you infer not the contradictory but
25 something which is the same as the co-assumption. Hence either the
94
Reading heterdi for hepomenoi.
95
i.e. 'What is true in general of arguments of the form "Either P or Q; P: therefore,
not-Q", is also true of the special case of such arguments in which "Q" is "not-P".' (But
it is difficult to find this in the text; and the repetition of kai tote at 19,17 and 19
suggests the possibility of textual corruption.)
96
cf. Alexander, in Top. 12,31-13,10: in the co-assumption, 'It is day' is used to
affirm that it is day; in the conclusion 'It is day' is used to deny that it is not day.
Hence, according to Alexander, 'It is day' has different senses (or different primary
senses) in these different contexts.
97
19,20-3 = SVF II261.
98
Omitting t&s sullogistiks sumplokes.
99
This is the fourth of Chrysippus' five indemonstrable syllogisms: see e.g. Sextus,
PHII158; Galen, Inst. Log. vi 6; Frede (1974a), p. 152.
 2. Introductory Explanations 69
100
conjunction in question is not syllogistic (if it infers the same
disjunct as the one co-assumed) or else, if it is syllogistic, then
according to them too what is inferred has not been taken qua the
same as the co-assumption.
Even if it happens to be the case that one item is both the
contradictory of one of the disjoints and also the same as the other, it 30
does not thereby follow that it is the same thing for it to be the same as
the one and to be the contradictory of the other. For if these
characteristics were the same as one another, then in all cases the
contradictory of one disjoint would thereby also be the same as the
other; but since they are different in most cases, then where they do 20,1
hold of the same thing at the same time, they do not do so in virtue of
being the same. Hence if you take one of the disjoints qua contra-
dictory of the one, you do not take it qua the same as the other.
Thus101 in disjoints which are not constructed from contradictories -
say, in those constructed from contraries - the conclusion is not the
same even in expression as either of the suppositions, because in
these cases it does not happen that the contradictory of the one is the 5
same as the other.102 In the argument:
Either it is day or it is night,
But it is not day,
Therefore: it is night,
'It is night' is not the same as either of the assumptions, neither as
what they call the 'tropic' (that is, the whole disjunction, 'Either it is
day or it is night')103 nor as the co-assumption (the co-assumption is
'But it is not day'). 'It is night' is different from each of these.104 10
The105 arguments which they call duplicated106 are not syllogistic -
100
i.e. the conjunction or combination of the two premisses.
101
20,2-10 = SVF II265.
102
The illustrative argument which Alexander produces is not an instance of the
fourth Chrysippean indemonstrable (above, n. 99); hence it is not, strictly speaking, a
'disjoint' syllogism in the sense in which Alexander has up to now been
understanding that expression (above, n. 92). Rather, the illustrative argument has
the form: either P or Q; not-P; therefore, Q. (This is the fifth of Chrysippus' five
indemonstrables: see e.g. Sextus, PHII 158; Galen, Inst. Log. vi 6; Frede (1974a), pp.
152-3.)
103
For this (Stoic) use of tropikon see e.g. Alexander, in An. Pr. 262,29-31;
263,11-12; rival explanations of the term in Galen, Inst. Log. vii 1; [Ammonius], in
An. Pr. 68,6-7; Philoponus, in An. Pr. 243,4-8. See Frede (1974a), p. 101 n. 25.
104
The gist of Alexander's argument is this: 'An argument of the form "P or not-P;
P: therefore, P" is valid not in virtue of the fact that it exhibits that form, but rather in
virtue of the fact that it exhibits (is a special case of) the form "P or Q; P: therefore,
not-Q". Hence the conclusion of the argument really has the form "not-Q", and hence
is not formally the same as the co-assumption, "P".' (See the summary version of this
material at Alexander, in Top. 11,23-7.) All Alexander's ingenuity is wasted: in the
end, he (and Aristotle) must either modify their definition of the syllogism or else
reject disjunctive arguments from contradictories.
105
20,10-16 = SVF II261.
108
See above, 18,17 note.
 70 Translation
for example:
If it is day, it is day.
But it is day.
Therefore: it is day.
First, they do not offer any syllogistic utility. For107 even if it is true
that a thing follows from itself,108 a conditional of this sort is not
thereby useful for syllogisms - no more than one which does not
allow a co-assumption, such as 'If you are a horse, you are an
animal'.109
Secondly, if it is non-syllogistic to co-assume the consequent of a
continuous proposition, and if in continuous propositions of this sort
15 the antecedent is the same as the consequent, then the
co-assumption is no more of the antecedent than of the
consequent.110
Again, if it is non-syllogistic to co-assume both of them - both the
20 antecedent and the consequent then in this way too arguments of
this sort will be refuted. For where the antecedent is the same as the
consequent, the co-assumption is either of both or of neither. (Or111
is it rather of the consequent - and in this way is more like it?)
Again, if it is non-syllogistic to assume the starting-point,112 then
107
The text of 20,12-22 is odd; and although the general train of Alexander's
thought seems reasonably plain, we are not sure what words he actually wrote. Here
we transpose ougar ... zoon el (20,16-18) to followparekhetai (20,13).
108
i.e. if every conditional of the form 'If P, then P' is true - as it is on some, but not
on all, ancient accounts of conditional propositions: see Sextus, PHII110-12.
109
But such conditionals may, of course, be 'useful' in reduction to the impossible.
(Note that there is indirect evidence, from Avicenna, of some ancient dispute over the
truth-value of conditionals with impossible antecedents: see Maroth, pp. 128-9.)
110
i.e. 'From a conditional proposition and its consequent (from "If P, Q" and "Q")
nothing follows. But in duplicated arguments the antecedent is the same proposition
as the consequent. Hence in duplicated arguments nothing follows from the
conditional and its antecedent.' (This must be what Alexander means.)
111
Placing e mallon ... ginetai (20,21-2) in parentheses. We tentatively paraphrase
this obscure paragraph as follows: 'Nothing follows from a conditional together with
both its antecedent and its consequent. Nothing follows from a conditional together
with neither its antecedent nor its consequent. But in duplicated arguments the
antecedent and the consequent are the same. Hence either you co-assume both or you
co-assume neither. Hence nothing follows from the premisses of a duplicated
argument. Or else perhaps there is a difference between antecedent and consequent;
but in that case the co-assumption is actually more like the consequent than like the
antecedent, and so (once again) nothing follows.' (Compare in Top. 10,28-30, which
appears to be a shortened and simplified version of the present text.) This
interpretation raises two questions. First, why suppose that nothing follows from a
conditional together with both its antecedent and its consequent? Because such
arguments have redundant premisses, and arguments with redundant premisses are
not syllogisms (see below, 23,1). Secondly, why might the co-assumption be more like
the consequent than like the antecedent of the conditional? - Perhaps because the
antecedent is thought of as including the word 'if: the co-assumption and the
consequent are each simply 'P', whereas the antecedent is 'if P'. (cf. [Ammonius], in
An. Pr. 68,8-13.)
112
i.e. to 'beg the question': see esp. Aristotle, An. Pr. 64b28-65a37 (cf. Alexander,
 2. Introductory Explanations 71
in this way too we may fault an argument which infers something
which is the same as one of the suppositions - for it co-assumes what
it purports to prove.
Again, Aristotle himself announced, at the beginning of the study, 25
what its end and aim are, namely demonstration and demonstrative
knowledge.113 Hence whatever makes no contribution to these
things will not be appropriate to the study before us.
Again, it would be absurd to deny the name of induction to an
inductive argument in which one of the assumptions or suppositions
is inferred and yet to claim that duplicated arguments are
syllogisms.

2.4.3 Necessitas consequentiae114

As for 'comes about by necessity', this does not mean that the 30
conclusion is necessary, as some have thought (this is so only when
the premisses are necessary), but rather that the conclusion follows
by necessity from the suppositions, whether it is actual115 or 21,1
contingent or indeed necessary.116 For even if a conclusion is
contingent, still in syllogistic combinations it too follows from the
premisses by necessity. Aristotle's expression does not mean that
the proposition which is concluded must be necessary: rather, it 5
marks a certain relation between the conclusion and the premisses.
If you assume the premiss:
Every man is lying down,
and co-assume the premiss which says:
Everyone lying down is asleep,117
then:
Every man is asleep
will follow by necessity - yet it will not be necessary that every man
is asleep.

2.4.4 Non-syllogistic arguments

As to why he added 'inasmuch as they are the case'118 to the 10
definition of the syllogism, he himself explained this when he said:
'By "inasmuch as they are the case" I mean that it comes about
in An. Pr. 257,1-4; in Top. 10,14-16; 566,25-31; Philoponus, in An. Pr. 451,15-453,33).
na
An.Pr. 24alO-ll; above, 9,5-23.
114
See Patzig, pp. 16-36.
115
For our use of the word 'actual' here see Introduction, p. 30.
116
cf. Ammonius, in An. Pr. 29,20-34.
117
Note that in presenting this argument Alexander has reversed the order of the
premisses: see Patzig, pp. 59; 76-7.
118
For further comments on this clause see below, 68,27-69,4; Alexander, in An. Pr.
343,21-344,6; 350,11-16 (cf. Barnes (1980a), pp. 168-73).
 72 Translation
because of them.' This itself might still seem less than plain. For
"because of them' betokens an explanation, and yet there can be
syllogisms which do not proceed by way of explanations - for
example, syllogisms by way of signs which prove what is primary
119
15 from what is posterior. This feature - viz. being syllogized by way
of explanations - is a proper characteristic of demonstrations.120 For
although the premisses must indeed be explanatory of the
conclusion if there is to be a syllogism, what is meant by the
premisses need not always be explanatory of what is meant by the
conclusion. (For you can also syllogize what is prior by way of what
20 is posterior - proving that she has given birth from the fact that she
is lactating, or that there was a fire from the ashes - and in general,
syllogisms by way of signs are of this sort: for the posterior is not
explanatory of the prior.) This121 is why he also explained 'it comes
about because of them', by saying that they 'need no external term
for the generation of the necessity', i.e. that the terms laid down are
sufficient in themselves for the conclusion.122
25 By this addition he shows at the same time that the so-called
single-assumption arguments are not syllogisms.123 For these need
some external term and premiss for their conclusion - a term and
premiss which those to whom the argument is addressed add for
themselves, inasmuch as it is well known. In general,124 if
something is deduced, yet not from the suppositions but by the
co-assumption of some other premiss, then this will be called
30 necessary but it is not thereby a syllogism. The arguments which
the Stoics say conclude unmethodically are of this sort.125 Thus you
might say:
119
For sign-inferences see Aristotle, An. Pr. 70a2-b38; cf. Burnyeat.
120
See Aristotle, An. Post. 71b22; above, 14,13.
121
21,21-22,26 =FDS 1087.
122
The paragraph is obscure. Alexander means to distinguish two ways in which
the premisses of a syllogism may be explanatory of its conclusion. One way is
reasonably plain: for 'P, Q; therefore, R' to be a demonstrative syllogism, it is
necessary that the states of affairs referred to by P and Q should explain the state of
affairs referred to by R. The other way is harder to grasp, and Alexander does nothing
to elucidate it. In the second way, the premisses of every syllogism must explain, or be
aitia for, the conclusion. At Phys. 195al6-21 (= Metaph. 1013bl7-23) Aristotle asserts
that the premisses are aitia for the conclusion inasmuch as they are 'that from which'
the conclusion comes about. In his commentary, Alexander claims that the premisses
are not the matter or 'material cause' of the conclusion (rather, they are the matter of
the syllogism as a whole); but that they should be regarded as the efficient cause,
aition poiStikon, of the conclusion (in Metaph. 351,9-15; cf. Simplicius, in Phys.
320,1-10, quoting Alexander's lost commentary). It is a plausible guess that our
passage in in An. Pr. alludes to this claim.
123
Above, 17,19 and note.
124
125
21,28-22,3 = SVF II260.
On unmethodical arguments see also Alexander, in An. Pr. 344,9-346,6 (cf.
below 68,21-4); in Top. 14,17-15,14; Galen, Inst. Log. xix 6 (cf. xvi-xviii); [Ammonius],
in An. Pr. 70,11-13; Philoponus, in An. Pr. 36,5-13 (cf. 321,7-322,18). See Mueller
 2. Introductory Explanations 73
The first is greater than the second.
But the second is greater than the third.
Therefore: the first is greater than the third.
This follows necessarily, but not syllogistically - not unless you 22,1
co-assume a premiss from outside which says:
What is greater than something greater is also greater than
what is smaller than it.
The same sort of argument occurs126 in the first book of Euclid's
Elements,IZ1 viz:
This is equal to this.
But this is equal to this.
Therefore: this is equal to this. 5
This is indeed true - but there is need of a universal premiss if it is
to be deduced syllogistically. The premiss is this:
Things equal to the same thing are also equal to one
another.128
The argument in Plato's Republic is similar:
If he was the son of a god he was not avaricious.
But if he was avaricious he was not the son of a god.
Therefore: he was not both.
This too is not deduced by way of the suppositions but by the 10
co-assumption of a universal premiss. The premiss is this:
When from each of two things130 there follows the
contradictory of the other, they cannot both hold of the same
thing at the same time.
For if with this premiss, which is universal, you co-assume:
From each of being the son of a god and being avaricious
there follows the contradictory of the other,
you get as conclusion:
They da not both hold of the same thing. 15
For from being the son of a god there follows the contradictory of
being avaricious (i.e. its negation), and conversely, from being
avaricious there follows the negation of being the son of a god.
In general,131 the arguments which the more recent thinkers say
conclude unmethodically are of this sort. For example:
(1969), pp. 175-8; (1974), pp. 59-67; Frede (1974a), pp. 121-4; (1974b), p. 102; Lee, pp.
108-14; Maroth, pp. 178-85; Barnes (1990a), IV.
126
Deleting thedrema.
127
For other references by Alexander to Euclid see in An. Pr. 260,23-261,2; 268,7-8.
128 Thjs proposition is the first of Euclid's 'common notions'; it is explicitly invoked
in Euclid's proof at Elements 11.
129
See Rep. 408C (but Alexander offers a feeble interpretation of the text). For
syllogizing interpretations of Platonic arguments see e.g. anon, in Theaet. 66,12-43;
Albinus, Didasc. 158-60; Galen, Inst. Log. xviii 2-4; cf. Moraux (1973/84), II, pp.
487-90; O'Meara, p. 42 n. 33.
130
Omitting ton antikeimendn - but see Lee, p. 110 n. 22.
131
22,17-19 = SVF II260.
 74 Translation
It is day.
But you say that it is day.
Therefore: you are telling the truth.
20 This is not a syllogism. But it will be one if you add the universal
premiss:
Anyone who says that what is the case is the case, is telling
the truth,
and add to this:
Someone who says that it is day when it is day says that
what is the case is the case.
The conclusion from the suppositions is:
Therefore: someone who says that it is day when it is day is
telling the truth.
Hence it is because of single-assumption arguments and
25 arguments which conclude unmethodically that 'inasmuch as they
are the case' is added - and also because of arguments which do not
have canonical132 premisses and which must be transformed in
order for there to be a syllogism. For example, the argument which
runs:
If a non-substance is destroyed, a substance is not
destroyed.
But if the parts of a substance are destroyed, the substance
is destroyed.
Therefore: the parts of substances are substances.
(Aristotle refers to this argument at the end of the book.133) The
30 conclusion does not come about 'inasmuch as they are the case'
rather, the premisses must be transformed.
23,1 The134 added phrase will also rule out, as he said,135 arguments
which contain a superfluous addition,136 among which are those
sophisms which make the non-explanatory explanatory.137
*
132
kurios: i.e. the propositions presented as premisses are not premisses in the
strict sense.
183
See An. Pr. 47a22-31; cf. Alexander, in An. Pr. 346,10-14.
134
22,30-23,1 =FDS 1118.
135
Reading eipen for eipon: the reference is to Aristotle, SE 168b22-6.
136 por arguments with superfluous or redundant premisses, which were deemed
not to be syllogisms, see Aristotle, Top. 161b28-30 (cf. An. Pr. 47al6-20); Alexander,
in Top. 8,29-9,5; 13,28-14,2; 568,18-23; Sextus, PH II 156-67; M VIII 438-43;
Boethius, Syll. Cat. 822C. See Barnes (1980a), pp. 164-75.
137
See Aristotle, SE 167b21-36; 168b22-6; Top. 162a32-4; An. Pr. 65a38-b40; cf.
e.g. Philoponus, in An. Pr. 456,24-457,7; anon., Logica 62; [Alexander], in SE
50,8-52,28. A standard example. Suppose that, in a reduction to the impossible of the
claim that all white things are coloured, you argue as follows: 'All white things are
coloured; all winged things are white; all crows are winged: therefore, all crows are
white'. Your argument will be rejected with a cry of 'ou para touto': the conclusion
does not depend on the premisses (the first premiss is redundant). But in a syllogism
the premisses must explain the conclusion. Hence your argument treats what is not
explanatory as explanatory.
 2. Introductory Explanations 75
5
2.4.5 'Positing
Some138 people try to quibble at the term 'posited', holding that it is
not used as it should be. They say that position is properly ascribed
to bodies which occupy places, but that utterances are incorporeal, 5
and that - according to Aristotle himself139 - in definitional
utterances one should not use words in metaphorical senses. These
people seem not to understand the reason why in definitional
utterances one should not use words in metaphorical senses. It is
because the utterances then become less plain. Hence it is
metaphors of this sort and in these circumstances which should be
guarded against. But when through common usage a metaphor is 10
better known than the strict sense, there is no longer anything
wrong in using it. And the metaphor in question is ubiquitous in
common usage. (That is why Plato can use it as though it were
obvious when, in the Theaetetus, he says: 'Posit a wax tablet in the
soul.'140)

2.4.6 Perfect syllogisms141

I call a syllogism perfect if it needs nothing else apart from the 15
assumptions in order for the necessity to be evident; <and
imperfect if it needs one or more items which are necessary by
way of the terms supposed but have not been assumed by way
of premisses>. [1.1, 24b22-7]
If you look at it carelessly, you will think that the definition he has
already given142 is a definition of perfect syllogisms; for in it he said
that syllogisms need no term for their necessity to come about, and
now he calls a syllogism imperfect 'if it needs one or more items ...'. 20
138
23,3-6 = FDS 691.
139
See Are. Post. 97b37-9; cf. e.g. Top. 123a33-7; 139b32-140a2; Meteor. 357a24-32;
Alexander, in Top. 425,16-426,5; Philoponus, in An. Post. 416,13-31.
140
See Theaetetus 191C8-9 (Alexander's citation is not exact - but the crucial word
thes is in Plato's text); cf. e.g. Philebus 33D2. As Alexander rightly says, the
metaphorical use of posit' ('suppose') is as common as the strict or 'literal' use.
141
See Mignucci (1969), pp. 191-2; Flannery; Patzig, pp. 43-61, and esp. 69-83 (for a
history of the interpretations of perfection). Note that Ammonius (in An. Pr. 32,30-3)
and Philoponus (in An. Pr. 36,19-37) give a different account from Alexander's: they
stress that in imperfect syllogisms the necessity is not evident. Note, too, that some
Peripatetics held that syllogisms in all figures were perfect: see esp. Themistius,
Max.; Ammonius, in An. Pr. 31,14-32,7; scholium to Aristotle, 156b43-157b9. This
view was championed by Boethus (see Moraux (1973/84), I, pp. 165-8, and later
defended by Porphyry, lamblichus and Maximus. Themistius opposed Maximus: the
dispute was put to imperial arbitration, and the Emperor Julian decided for
Maximus. Ammonius (in An. Pr. 31,22-3 = F 20 Graeser) says that 'Theophrastus
seems (phainetai) to have taken the view opposed to Aristotle'; but see e.g. Bochenski
(1947), pp. 64-5. (See also below, 30,6 note; 31,9 note.)
142
i.e. at Are. Pr. 24bl8-22, the immediately preceding sentence.
 76 Translation
But in that case, imperfect syllogisms will not be syllogisms. But
the syllogisms of the second and third figures are imperfect, and he
insists that they too are syllogisms. Again, at the beginning of the
book he spoke of imperfect syllogisms as syllogisms; for he proposed
to investigate 'what sort of syllogisms are perfect and what sort
25 imperfect',143 suggesting that imperfect syllogisms are indeed
syllogisms.
But this is not what is said here. In the definition of the syllogism
imperfect syllogisms too are included; for Aristotle does not mean
that the things which he says imperfect syllogisms need should be
assumed from outside - rather, he means that they are present in
the suppositions but stand in need of proof. For he says of them that
30 they 'are necessary by way of the terms supposed but have not been
assumed by way of premisses'. And if the items from which the
conclusion derives are present in the suppositions, then both in
perfect and in imperfect syllogisms the conclusion will be deduced
24,1 'inasmuch as they are the case'; and hence imperfect syllogisms too
will be syllogisms. (Arguments144 which conclude unmethodically145
are not even imperfect syllogisms; for in them the item which
explains the deduction is added from outside.)
Imperfect syllogisms which need a single conversion in order to be
reduced to one of the perfect and indemonstrable syllogisms in the
5 first figure need one item; and those which are reduced to a perfect
syllogism by two conversions need more items, as he will later
show.146 Again, syllogisms which are proved by reduction to the
impossible are also imperfect.
Thus arguments which need the addition of some account from
outside in order for their conclusion to be deduced are not syllogisms
at all (arguments which conclude unmethodically are of this sort),
whereas arguments in which what is needed is present potentially
10 in the suppositions but which need to be helped and unveiled are
indeed syllogisms but they are imperfect syllogisms (such are the
non-indemonstrable arguments of the second and third figures).
Why do the assumptions in impossibles147 not come from outside?
No doubt the assumptions do indeed come from outside and are not
even potentially present in the suppositions (as converses are); but
15 the assumptions do not syllogize the point at issue but rather
something else. For the syllogism yields something which is rejected
inasmuch as it is impossible and which explains why we may posit

r. 24al2-13.
144
24,1-12 =FDS 1100.
145
See above, 21,31 and note.
146
i.e. in his account of second and third figure syllogisms in An. Pr. 1.5-6.
147
If the text is right, 'impossibles' here must be shorthand for 'reductions to the
impossible".
 2. Introductory Explanations 77
the point at issue. That is why we say that it is proved not by a
syllogism but from a hypothesis - the syllogism yields something
else.148

2.5 Kinds of proposition

2.5.1 Universal propositions

For one thing to be in another as in a whole and for the other to 20
be predicated of all the one are the same thing. <We say that
one thing is predicated of all another when it is not possible to
take any of it of which the other is not said. And similarly for of
none.>[l.l,24b26-32]
As his account proceeds, he will make use of the phrases
'such-and-such holds of every so-and-so', 'is in as in a whole', 'holds
of none' and 'is in none'. So first he makes these known to us. He
tells us what 'to be in as in a whole' and 'of every' mean, saying that
they are the same as one another and are used for universal 25
affirmative propositions; and he tells us what 'of none' and 'in none'
mean, saying that these too are always the same as one another and
148
Alexander's later accounts of reduction to the impossible are clearer than this
brief sketch: see below, 31,10-19; 77,20-5; 93,25-30; 112,25-31; in An. Pr. 259,8-260,6;
317,31-318,10; 386,5-30 - and for an example of his perfectly competent use of
reduction see e.g. below 79,27-80,24. (For Aristotle's treatment of reduction, on which
Alexander draws, see An. Pr. 41a23-32; 61al8-33; cf. e.g. Lukasiewicz, pp. 54-5;
Patzig, pp. 144-56). The problem Alexander is here trying to solve is this: Tfou hold
that reduction to the impossible is a syllogistic procedure; and yet a reduction surely
introduces something from outside. Therefore syllogisms may depend on external
propositions' (cf. below, 84,19-31; Philoponus, in An. Pr. 37,16-38,29). Alexander
replies that a reduction does indeed introduce something external, but that what a
reduction syllogizes is not the point at issue but something else - and the reduction
introduces nothing external to this item. An example will show more clearly what
Alexander means. Suppose we want to prove Baroco by reduction to the impossible
(see below, 83,24-84,3; and cf. Appendix 2, p. 211). We assume
(DM holds of every N,
and
(2) M does not hold of some O.
(These are the premisses of Baroco.) Now hypothesize:
(3) N holds of every O.
From (1) and (3) we can infer, by Barbara, that
(4) M holds of every O.
But (4) is impossible, since it contradicts (2). Hence we must reject the hypothesis, (3),
and thus posit its contradictory,
(5) N does not hold of some O.
Here the external item introduced by the reduction is the hypothesis, (3): it is
external to the premisses of Baroco, inasmuch as it is neither identical with nor
derivable from them. The reduction is the argument (l)-(5). It is not, as a whole, a
syllogism. It also contains an item external to Baroco. This item is indeed used in a
syllogism. It is used in the instance of Barbara contained in (l)-(5); and in this
syllogism it is not, of course, external.
 78 Translation
are used for universal negative propositions.
Thus he says that 'of every' holds 'when it is not possible to take any
of it of which the other is not said', i.e. when none of the subject can be
30 taken of which the predicate is not said. For every predicative
proposition consists of a subject term and a predicate term,149 and the
predicate is truly said of every one of the subject when none of the
subject can be taken of which the predicate is not said. For example,
25,1 animal is said of every man - for no man can be taken of which animal
is not said. This being so, man is in animal as in a whole, i.e. it is
included in it as in a whole;150 for what is in something as in a whole is
in a sense part of the universal: 'as in a whole' signifies that none of
what is said to be in something as in a whole falls outside that which it
5 is said to be in as in a whole. For it is not only what is completely
co-extensive with something else which is said to be in it as in a whole:
parts too are in things as in a whole. What is said of every so-and-so
may be of wider extension (as for example animal extends further
than man); and it may also be of equal extension (as with risibility
and man).151
10 Now in such cases we say 'of every' when we start from the
predicate (for it is this which is 'of every'); and we say 'as in a whole'
when we start from the subject (for it is this which is in something 'as
in a whole').152
Note that he does not present universals in the same way here and
in the Posterior Analytics.153
After giving an account of 'of every', he says 'And similarly for of
15 none'. By this he does not mean that 'of every' and 'of none' are the
same (they are contraries), but rather that it is possible to define 'of
none' on the basis of what has been said about 'of every'.154 For if what
has been said is presented the other way about, we shall get 'of none'.
For when none of the subject can be taken of which the predicate is
20 said, 'of none' can truly be said. For example, 'Neigher holds of no
man': there is no man of ivhom neigher is predicated. Thus neigher
holds of no man and man s in no neigher. (Both of these formulations
are used for universal nej ative propositions.)

149
See above, 14,25-15,4.
iso por 'as jn a wnoie' see Introduction, p. 28.
151
Contrast Philoponus, in An. Pr. 39,15-24, who maintains that 'A is in B as in a
whole' implies that A has a smaller extension than B.
152
See above, 10,2-9; below, 53,19-25; 54,2-29; Ammonius, in An. Pr. 33,28-34,5;
Philoponus, in An. Pr. 12,5-12. See e.g. Mignucci (1969), pp. 216-17; 224-7.
153
See An. Post. 73b26-7; above, 12,2-3. (The words at 25,11-13 are exactly the
same as those at 12,2-3: it is difficult to avoid the suspicion that their second
occurrence is not due to Alexander.)
154
Omitting eipein peri tou kata mgdenos (with the Aldine).
 2. Introductory Explanations 79
2.5.2 Modality
Now every proposition is either of holding or of holding by
necessity or of holding contingently. [1.2,25al-2] 25
With 'every proposition' one has to understand 'predicative'. For
Aristotle is here dealing only with propositions and syllogisms of
this sort - they are the ones of which he has given an account.155
Now in every predicative proposition a term is predicated of a term
either affirmatively or negatively, i.e. as holding of the subject or not 26,1
holding of it. Of those terms which hold of something, some hold
always while others sometimes hold and sometimes do not hold. If
what is said to hold holds always and is taken as holding always,
then the proposition is a true necessary affirmative. A true 5
necessary negative is a proposition in which what by nature never
holds of something is taken never to hold of it. If the predicate does
not always hold of the subject, then if it holds of it at the present
moment, the proposition which marks this is a true actual156
affirmative; and in the same way, a true actual negative is a
proposition which says that what does not now hold does not hold. If 10
the predicate does not hold of the subject at the present moment,
although it is possible for it to hold, and if it is taken in this way, i.e.
as being possible, then the proposition is a true contingent
affirmative; and a proposition which says of something which either
holds or does not hold but can both hold and not hold, that it is
contingent for it not to hold, is a true contingent negative.157
155 156
See above, 11,17-20; 17,5-8. For our use of'actual' see Introduction, p. 30.
157
(1) Aristotle normally operates with three (or four) different modalities:
necessity, possibility, contingency (or 'two-sided' possibility) and, as an optional
extra, impossibility. It is contingent that P if and only if it is possible that P and also
possible that not-P (see e.g. Int. 23a7-26; An. Pr. 32al8-b2; cf. A. Becker, pp. 11-16;
Hintikka (1973), pp. 27-30, 47-51). See also below, 41,10 note. (2) The later
Peripatetics regularly gloss these modalities in temporal terms (see e.g. Ammonius,
in Int. 153,13-15; 215,11-14), so that 'Necessarily P' is equivalent to 'Always P' and
'Possibly P' to 'Sometimes P'. Hence 'Contingently P' becomes equivalent to
'Sometimes P and sometimes not-P', and 'Impossibly P' to 'Never P'. Traces of these
equivalences are to be found in Aristotle, but it is disputed how thoroughly
Aristotelian they are (see e.g. Hintikka (1973), pp. 93-113; Barnes (1977)). (3)
Alexander's account in the present text (cf. e.g. in An. Pr. 183,19-24) is clearly based
on some idea of temporal equivalences; but it is muddled. He preserves the equations
between 'Necessarily P' and 'Always P' and between 'Necessarily not-P' (i.e.
'Impossibly P') and 'Never P'. But for 'Contingently P' he offers the curious 'Possibly P
and now not-P' (which he finds in Aristotle at An. Pr. 29b29-32; cf. in An. Pr.
119,22-8). For 'Contingently not-P' he offers the odder 'Possibly P and possibly not-P
and (either now P or now not-P)'. And, what is oddest of all, the accounts he gives of
'Contingently P' and 'Contingently not-F evidently do not run parallel. We can find
no way of explaining - let alone of justifying - this muddle. In addition, Alexander
glosses the non-modal 'P', or 'Actually P', in temporal terms: 'P' is equivalent to 'Now
P and not always P'. This has some strange consequences. Thus 'Necessarily P' does
not entail 'F; 'P or not P' is not logically true\ It also fails to consist with Aristotle's
remark at An. Pr. 34b7-18 (cf. An. Post. 73a29; 75b25-6), according to which 'you must
 80 Translation
15 Such propositions are false when they mark something as having
a mode which it does not in fact have. For since propositions make
clear that what is meant by them holds, the propositions
themselves, if they are assimilated to the mode of what they mean,
correspond to the way in which those items hold.
Further, everything which holds of something either is
inseparable from it and holds of it in its own right or else is
20 separable. If it is inseparable, the proposition which means this is
necessary, and if it is separable, contingent. If a contingent
proposition means what is already present, it is actual; if it means
either what has been separated or what is not yet present but can
hold, then it is contingent in the specific sense.158

2.5.3 Quality and quantity

Of these some are affirmative and others negative, in respect of
each adjunct. [1.2,25a2-3]
25 Having mentioned three kinds of proposition - necessary, actual
and contingent - he says that in respect of each of these kinds some
propositions are affirmative and others negative. For there are
necessary affirmatives and necessary negatives, as we said, and it is
the same with the actual and the contingent. Hence there are six
kinds of proposition.
'In respect of each adjunct' means 'in respect of each different type
159
27,1 of predication', i.e. 'in respect of each addition of a mode'. He
claims that propositions take their necessity, actuality and
contingency not from what underlies and is meant by them, but
from the addition which is annexed or co-predicated, and which says
5 that this holds of this by necessity or that it holds or that it holds
contingently.160 For in this way there will be some true and some
false necessary propositions - and similarly for the actual and
contingent. For there would be no false utterances if something
take "holds of every" in an unqualified way and without limiting it in time (e.g. to now
or to this time)' (cf. Hintikka (1973), pp. 68-9; 136-8).
158 With the analysis of the modalities offered in this paragraph compare e.g.
Alexander, in Top. 257,22-3. The analysis is problematical in itself; and it appears to
yield results which are inconsistent with those yielded by the earlier temporal
analysis. (But note Ammonius, in Int. 153,13-15: "what is necessary in the unqualified
and proper sense is that which (a) always holds of the subject in such a way as (b) not
even to be able to subsist separately from it'. Here Ammonius combines Alexander's
first account, in (a), with his second, in (b).)
159
On 'modes' or tropoi see Introduction, p. 29.
160
i.e. Alexander distinguishes between (a) propositions which say that P, where in
fact it is necessary that P; and (b) propositions which have the form 'Necessarily P';
and he holds, correctly, that when Aristotle talks about 'necessary propositions' in the
context of his syllogistic he has (b) and not (a) in mind. Cf. Ammonius, in Int.
88,12-28; 215,7-28; Philoponus, in An. Pr. 43,18-44,1; Barnes (1991).
 2. Introductory Explanations 81
were only called an utterance if it signified the nature of the objects
and were of the same kind as the things it signified. But we say that
assertoric utterances admit both truth and falsity. Thus, just as we
say that some affirmations161 are true and others false (namely,
those which do not describe their subjects as they are), so we will 10
also say that some necessary propositions are false. False necessary
propositions will be those which say that what is not necessary is
necessary. These adjuncts and modes must therefore be annexed to
the propositions.

Again, of affirmatives and negatives, some are universal, some

particular, some indeterminate. [1.2,25a3-5]162 15
With these differences added to the propositions mentioned before,
there are eighteen kinds of predicative proposition in all. Three
necessary affirmatives, universal, particular and indeterminate; the
three negatives which are opposite to these; similarly three actual
affirmatives and three actual negatives; and the same for the 20
contingent. Hence there will be in all three times six propositions
differing from one another in form. From them predicative
syllogisms are constructed.
So much for the kinds and number of propositions. Next he sets
out to discuss their conversions. For most of the syllogisms163 in the
second and third figures, which are all imperfect, require one or 25
more prepositional conversion to prove them syllogistic, as will be
shoNvn.

2.5.4 Modes164
It seemed to me worth asking why, when he is dealing in these books
with syllogisms and figures, he also introduces material differences
among propositions.165 For the fact that a predicate belongs in this
way rather than in that way is a material difference. Differences of
this sort among propositions will seem to bear not on an argument's 28,1
being a syllogism simpliciter but on its being this or that kind of
syllogism - demonstrative, say, or dialectical.166
161
Alexander must mean: 'some actual [i.e. non-modal] affirmative propositions'.
162
cf. 24al7.
163
Not all, since some can only be reduced by using reduction to the impossible.
164
cf. Barnes (1990a), III.3-4; (1991).
165
Note that 27,1-5 has effectively answered this question, by distinguishing
between propositions which say that P, where it is necessary that P (where necessity
introduces a 'material' difference) and propositions which have the form 'Necessarily
P' (which introduces a formal difference). The present section breaks the flow of the
argument: Alexander perhaps incorporated it into his commentary at a late stage.
166
See above, 7,9 and note.
 82 Translation
Or are these differences among propositions in general necessary
5 for the method and study of syllogistic? For since propositions which
differ in respect of the modes we have mentioned do not convert in
the same way, and since most of the syllogisms in the figures other
than the first are proved to be deductive by means of conversions, it
is indeed necessary for him to divide propositions according to the
modes. For it is by converting in one way in the case of the necessary
and the actual and in another way in the case of the contingent, that
10 the propositions will produce syllogisms. Moreover, when proposi-
tions which differ in respect of these modes are mixed167 with one
another, they produce different syllogisms: for this reason too, the
difference among propositions in respect of these modes is necessary
for the study of syllogistic.
Thus, ignoring the material aspect, he annexes the modes
themselves to the propositions and produces universal proofs for
them; and he thereby shows that the difference among the
15 syllogisms depends not on there being this or that matter, but
rather on the annexed mode. So the proofs, being universal, are a
proper object of syllogistic study.
Aristotle mentions these modes insofar as they are useful for
syllogistic method - as is clear from the fact that there are
20 differences among propositions in respect of other modes too. If'well'
or 'badly' or 'at length' or 'briefly' or 'fast' or 'slowly' is annexed, these
are modes and differences among propositions (e.g. in 'Socrates
talks well' or 'at length' or 'concisely').168 But Aristotle does not
mention any of these when he makes his division of propositions,
since they do not bear on the generation or differentiation of
syllogisms.
25 Now these different modes are useful for the generation and
construction of syllogisms, but not for the generation of the species of
syllogism. For the different kinds of syllogism (i.e. the demon-
strative, the dialectical and the sophistical) depend not on this
division but on the one which Aristotle discusses in the studies
30 appropriate to them. For there he says much about this theory and
makes it splendidly plain.169

167
For such 'mixing* see 123,28-124,7: mixis was the standard term for 'mixed'
modal syllogisms (cf. the title of Alexander's work on the subject: Introduction, p. 3).
168
cf. Ammonius, in Int. 214,31-215,3 (see Barnes (1991), XI). For other modes in
Alexander see in An. Pr. 270,10-28; 329,30-330,5; 411,35-7 (with reference to
Aristotle, Int. 22al3, where 'truly' is implicitly taken as a mode).
169
See above, 7,9 note (and cf. 14,18-21).
 3
Conversions

3.1 Conversions in general1

And in the case of propositions of holding, it is necessary for the 29,1
universal privative to convert in its terms ... [1.2, 25a5-6]
Next he discusses the conversions of propositions. The exposition of
this topic is of use to him in relation to the imperfect syllogisms, 5
which, as we said,2 are constructed in the second and third figures;
for most of these are shown to be syllogistic by means of one or more
conversion.
There3 are several sorts of conversion. Thus conversion is
performed on syllogisms, and syllogisms are said to convert - an
argument which converts from a syllogistic argument is itself said to
be syllogistic. These arguments convert together with an opposition. 10
For when we assume the opposite of the conclusion and co-assume
one of the premisses, we deduce syllogistically the opposite of the
other premiss. For what converts in this way from a syllogistic mode
is itself syllogistic. Aristotle will instruct us about the conversion of
syllogisms in the second book of the study before us.4 15
There is conversion together with an opposition among
propositions too. For the proposition saying "What is not an animal
is not a man' converts from 'Man is an animal'.5
There is also conversion among propositions when an affirmation

1
cf. e.g. below, 45,14-46,16; Alexander, in Top. 191,6-12; 582,23-6; and esp. Cow.;
also Galen, Inst. Log. vi 3-7; Apuleius, Int. 181,19-183,6; Ammonius, in An. Pr.
35,7-36,15; Philoponus, in An. Pr. 39,32-42,34; Boethius, Int. Syll. Cat. 785B-790C
(cf. Syll. Cat. 804A-809A). See Lee, pp. 79-94; Mardth, pp. 140-3.
2
See above, 27,22-6 (or perhaps 23,21-24,7, if the previous section is a later
addition).
3
29,7-13 =FDS 1163.
4
See An. Pr. 2.8-9. On conversion of syllogisms see also Alexander, Conv. 60;
Galen, Inst. Log. vi 5; Patzig, pp. 152-3.
5
cf. below, 46,6-8; Alexander, in Top. 190,26-193,7 (on Aristotle, Top. 113bl5-26,
whence Alexander's example in in An. Pr. is taken); Conv. 60-1; Galen, Inst. Log. vi 4;
[Ammonius], in An. Pr. 68,28; Lee, pp. 86-7.

83
 84 Translation
6
is true together with the corresponding negation. In this sense a
contingent affirmative is said to convert from a contingent negative.
20 For example, if it is true (as a contingency) that it is contingent that
every man is walking, then it is also true that it is contingent that
no man is walking.
Conversion of terms when the items are true together is also
called conversion of propositions. Conversion of terms occurs when
we interchange and make the subject term the predicate term and
25 what was the predicate7 the subject, preserving the quality of the
converted proposition. The interchange of terms in the manner
described is called conversion simpliciter^ - for example:
Every man is an animal.
Every animal is a man.
Here the terms have been converted. When in addition to the
conversion, the propositions which are converted, being assumed in
the same way,9 are also true together, then the propositions are said
to convert from themselves.

3.2 Non-modal conversions

3.2.1 General remarks
30 It is this sort of propositional conversion which he now discusses,
and he shows which of the propositions we have mentioned convert
from one another and which do not. It will be shown that
propositions convert from one another if their opposites convert
from one another and that they do not convert if their opposites do
30,1 not convert.10 Thus particular affirmatives are contradictorily
opposed to universal negatives and particular negatives to universal
affirmatives. Now universal negatives convert from themselves, as
5 Aristotle will show: given that a universal negative is true, then the
proposition assumed in the same way but with its terms reversed is
also true. For example, 'No man is a horse' is true, and 'No horse is a
man'is true.11
6
The verb sunale'theuein is used in two ways: 'x is true together with y' sometimes
means 'When y is true, x is true' (e.g. Ammonius, in Int. 161,11-16); more often,
however, it means 'x is true if and only if y is true' (e.g. Ammonius, in Int. 272,3). The
ambiguity leads to confusion and obscurity in some passages: see Lee, pp. 88-92.
7
Reading teds forpds. cf. below, 75,13; 110,28.29.
8
Or anastropM: see Alexander, in Top. 264,19; Ammonius, in Int. 194,15;
Philoponus, in An. Pr. 42,20; Sextus, M VIII 108; Galen, Inst. Log. vi 3 - and esp.
Simp. Med. Temp. XI463-6K; 498-500K. See Lee, pp. 80-1.
9
i.e. having the same quantity and quality.
10
cf. Alexander, Conv. 63.
11
Some people denied that universal negatives converted and produced putative
counterexamples: see Philoponus, in An. Pr. 48,18-49,5, and esp. Alexander, Conv.
69-74 (cf. Themistius, Max. 186). If the conversion is rejected, then the second figure
 3. Conversions 85
Particular affirmatives, which are opposite to universal negatives,
also convert. For, given that a particular affirmative is true, then the
proposition assumed in the same way but with the terms reversed is
also true. For example, 'Some man is musical' is true, and so also is 10
'Some musical is a man.' So these too convert from themselves.
Universal affirmatives do not convert from themselves. For it is not
the case that, if the proposition that every man is an animal is true,
then the proposition which says that every animal is a man is thereby
also true (were that so, they would convert from themselves). But the 15
particular affirmative which says that some animal is a man does
convert from it - hence the particular affirmative converts both from
itself and from the universal affirmative.
Just as the universal affirmative does not convert from itself,
neither does the opposite of the universal affirmative - i.e. the
particular negative - convert from itself. For it is not the case that if it 20
is true that some animal is not a man, it is also true that some man is
not an animal. (If some propositions of this type can be assumed and
found to be true together with their converses, this does not disprove
what we have just said. For in all such cases, in order to reject a
universal claim it is sufficient to prove that it is not so in some
instances. For items which are not true together in all instances do 25
not derive their convertibility from their own nature but from the
special nature of their matter.12)
Aristotle discusses these prepositional conversions: he attempts to
show that the conversions hold not only by appealing to the obvious
but also by invoking certain universal arguments. He does not
discuss indeterminate propositions because they are not useful for 30

syllogisms Cesare and Camestres cannot be reduced to Celarent and must therefore
be treated as perfect (Alexander, Corey. 69; Themistius, Max. 180): see above, 23,14
note. Themistius explicitly and Alexander implicitly ascribe the rejection to the
Megaric philosopher Eubulides; and Alexander explicitly says that he defended,
against Aristotle, the thesis that particular negatives convert (see below, 30,26 note).
Eubulides is known from Greek texts as an enemy of Aristotle's (e.g. Diogenes
Laertius, II 109: see Giannantoni II B, 8-11); but the two Arabic texts provide the
only evidence thus far discovered that Eubulides attacked Aristotle's philosophical
position rather than his life and person. (It must be added that the attacks show no
logical refinement at all.) The two texts'also provide evidence that thinkers outside
Aristotle's immediate circle knew the contents of his Analytics.
12
cf. Alexander, Corey. 61; Boethius, Syll. Cat. 805B-806A; in Cic. Top. 1133BC. In
Conv. Alexander goes on to remark that 'Eubulides argued against Aristotle's view
that particular negatives do not convert from any proposition, saying that they convert
sometimes from particular negatives and sometimes from universal affirmatives' (66:
on 66-8 Alexander rightly scouts this suggestion). See above, 30,7 note.
 86 Translation
syllogisms and because they are equivalent to particular propo-
sitions.13

3.2.2 Universal negatives

31,1 First, then, let the proposition AB be a universal privative.
Now if A holds of no B, B will hold of no A. <For if of some, e.g.
of C, it will not be true that A holds of no B. For C is something
ofB.>[1.2,25a!4-17]
Theophrastus and Eudemus gave a simpler proof that universal
5 negatives convert from themselves. (By 'universal privative'
Aristotle means universal negative.) They conduct their proof in this
way. Suppose that A is said of no B. If it is said of no B, then A is
disjoined and separated from B; but what is disjoined from
something is disjoined from something disjoined from it; therefore B
is disjoined from every A, and if this is so, it is said of none of it.14
10 This is the way in which they argue. Aristotle, however, seems to
use a reduction to the impossible to prove this conversion. We15 have
a reduction to the impossible when we hypothesize the opposite of
what we wish to prove, co-assume one of the items agreed and
supposed, and reject syllogistically something which is evident.16
For the hypothesis because of which what is evident and agreed is
15 rejected is proved to be false by way of such a reduction, and when
this has been proved false, its opposite is found to be true, since in
every case one part of a contradictory pair is true and the other
false. And this opposite is the point at issue which is to be proved.
Geometers make most17 use of this method, but logicians use it
20 too; and Aristotle seems to use it here. For he says that 'if A holds of
13
'B holds of A' is equivalent to 'B holds of some A'; 'B does not hold of A' to 'B does
not hold of some A'. (For 'B holds of A' will be true if either all As are B or some As are
B. But if all As are B, then some As are B. Hence 'B holds of A' is true if and only if
some As are B.) This view (which finds support at e.g. Int. 17b29-37; An. Pr. 26a29-30;
29a27-9) is a commonplace in the commentators: e.g. below, 62,24; 94,20;
111,30-112,2; Alexander, in Top. 288,13-289,8; Apuleius, Int. 177,15-17; Ammonius,
in Int. 100,21-3; 106,13-16; Philoponus, in An. Pr. 79,4-5; and see esp. Ammonius, in
Int. 111,10-120,12, which argues at length for the equivalences (cf. Boethius, Int.
Cat. Syll. 776C-778A; Syll. Cat. 802C-803B) against unnamed adversaries. See also
Brunschwig (1968); Back.
14
F 13 Graeser: cf. below, 34,13-15; Alexander, in An. Pr. 124,18-22; 132,24-32;
220,12-15; Conv. 65; Ammonius, in Int. 185,14-18; Philoponus, in An. Pr. 48,11-18.
See Maier Ha, pp. 20-1; 44 n. 1; Bochenski (1947), pp. 54-6. Note that the idea of
'separation' was later adopted by Boethus (and hence Maximus) in connexion with
the thesis that all syllogisms are perfect: Themistius, Max. 190-1; scholium to
Aristotle, 157al3-24. (See above, 23,14 note.)
15
31,11-20 = FDS 1184.
16
Reading enargon sullogistikos for enargos [enargon M] sullogistikon (cf.
Philoponus, ire An. Pr. 90,20).
17
Readingpleista (LM, Aldine) forpleistoi.
 3. Conversions 87
no B, B will hold of no A', which is what he wishes to prove. If
someone should not concede this, clearly he will say that its opposite
is true, i.e. that B holds of some A. For 'of no' and 'of some' are
contradictorily opposed, as Aristotle has shown in On Interpreta-
tion.18 Hypothesizing that B holds of some A, he takes some of A of 25
which B holds, namely C, which is a part of A, and by means of this
he proves that A will hold of some B.
He proved the point at issue in a very brief compass; and that is
why some think he conducted the proof by taking particular
affirmatives to convert - and then accuse him of using reciprocal
proof.19 For, on the one hand, when he wants to prove that universal 30
negatives convert, he invokes the particular affirmative (so they
say) and its conversion. On the other hand, when a little later on he 32,1
proves that particular affirmatives convert, he invokes the
conversion of universal negatives. But such a method of proof is
agreed to be unsound.
But Aristotle does not proceed as they think he does. It is agreed
that he proves that particular affirmatives convert by means of the 5
conversion of universal negatives; but he does not here prove that
universal negatives convert by assuming that particular affir-
matives convert. He proves the point at issue, but he does not use
the particular conversion as something agreed and supposed. He
proves it by means of items already proved and supposed, i.e. 'of
every' and 'of none' and 'in as in a whole' and 'in none'. These are the 10
items he invokes in his proof of the conversion of actual universal
negatives.20
If it is supposed that A holds of no B, he says that it follows from
this that B holds of no A. For if B holds of some A (this is the opposite
of what was supposed, and one of the two must be true), let it hold of
C; for let C be some of A of which B holds. Now C will be in B as in a 15
whole and will be some of B, and B will be said of every C, since 'in as
in a whole' and 'of every' mean the same. But C was some of A; C,
therefore, is also in A as in a whole. But if it is in A as in a whole, A
18
See Int. 17bl6-20.
19
The same charge is reported and rejected below, 100,24-6; cf. Alexander, Conv.
64-5; Philoponus, in An. Pr. 49,6-31. Bochenski (1947), p. 55, guesses that the charge
may have been levelled by Theophrastus, who therefore elaborated his new proof of
the conversion. (On reciprocal proof see Barnes (1990c), ch. 3.)
20
i.e. the conversion can be proved on the basis of the definitions of'of every' and 'of
no': it depends on what the mediaevals called the dictum de omni et nullo. See also
below, 54,2-29; 55,1-3; 61,3-5; 69,16-20; Alexander, in An. Pr. 126,1-8, where perfect
syllogisms are put on the same basis. (Cf. e.g. Philoponus, in An. Pr. 188,20-1;
198,19-20; 201,18-19; Boethius, Syll. Cat. 809C-810C.) Thus according to Alexander -
and almost all the later commentators - Aristotle makes his syllogistic depend on the
dictum, which is itself a 'dialectical' or 'topical' rule. Lukasiewicz, pp. 46-7, dismisses
Alexander's interpretation; but see e.g. Smith (1989), p. 111. On the later controversy
over the relation between syllogistic and topics see Stump.
 88 Translation
will be said of every C. But it was supposed that C was some of B.
Therefore A will be predicated of some of B. But it was supposed
20 that A is said of no B - and 'of none' meant that there is no B of
which A is predicated.21
That he does not conduct the proof by assuming particular
affirmatives to convert is clear from the very expression he uses. He
did not say: 'For if B holds of some A, A holds of some B' - this would
25 be to conduct the proof by means of the particular affirmative. Thus
when he later proves that particular affirmatives convert, he
invokes the conversion of universal negatives by saying: 'For if of
none, A will hold of no B' - there he uses the conversion as
something agreed.22 But here he does not assume the conversion of
particular affirmatives as something agreed. Rather, he sets out
30 some of A (namely C) and conducts the proof in terms of this - and B
is no longer said of some of this.23 That is why we should say that for
the proof Aristotle used not the particular conversion but rather 'of
every' and 'in as in a whole'.
Or is it better and most faithful to the text to say that the proof
proceeds by means of exposition and perception, and not by the
33,1 method described, i.e. not syllogistically? For the method of
exposition depends on perception and is not syllogistic.24
As C, the item set out, we take something which is perceptible and
is a part of A. For if B is said of a perceptible individual, C, which is a
part of A, then this same C will also be a part of B, since it is in it.
5 Hence C will be a part of both and in both. So A, insofar as it is
predicated of C, which is a part of itself, will also be predicated of
some B, since C is a part of B, being in it. For example, if it is
supposed that man holds of no horse, then if someone refuses to
concede that horse holds of no man,25 he will say that it holds of
10 some man. If this man - call him Theo - is taken, it will turn out
that man does hold of some horse, since Theo was assumed to be
21
If B holds of some A, then there is something - call it C - such that B holds of
every C and A holds of every C. (If B holds of some A, then consider the ABs, i.e. the
things which are both A and B: A holds of all of them, and so does B.) Hence, by a
syllogism in Darapti, we have it that A belongs to some B - which contradicts the
supposition that A belongs to no B. See Patzig, pp. 161-3, who finds this thought in
Aristotle, at An. Pr. 43b43-44a2 (and who maintains that 31,6-34,22 is 'one of the
finest parts of <Alexander's> commentary1).
22
See An. Pr. 25a20-2.
23
sc. 'butofaWofit'.
24
On 'exposition' or ekthesis see below, 67,11-12; 89,12-13; 96,17; and esp.
99,27-100,14; 103,25-104,9; 112,33-113,2; Alexander, in An. Pr. 121,15-123,24; see
e.g. Ross, p. 318; Kneale, p. 77; Mignucci (1969), pp. 261-2; Lee, p. 93 n. 24; Smith
(1982); and esp. Thorn, pp. 164-76. (See also Appendix 2, p. 211.) With Alexander's
account in the next paragraph compare Conv. 63-4; see Lukasiewicz, pp. 59-67;
Patzig, pp. 156-68.
25
Omitting toutdi (with L). M offers to, the Aldine touto, for toutoi: either reading
will give the same sense as our preferred text.
 3. Conversions 89
both a man and a horse. But this is impossible, since it was agreed
that man holds of no horse.
When in the third figure, he explicitly uses proof by exposition,26 he
uses it as a perceptual and not a syllogistic proof. And this is why he
uses it here too: we do not yet know anything about syllogistic proofs. 15
Now by using this proof you can also show that particular
affirmatives convert inasmuch as this too follows from what has
already been said. Hence some deem him to conduct the proof by
means of the conversion of particular affirmatives. But he does not do
this: he does not invoke the conversion of particular affirmatives. For 20
he assumes that A holds of some B not because it converts but because
C, which is a part of A, is also a part of B. Thus he continues: 'For C is
something of B.' The particular conversion follows from this; but he
did not invoke it. After all, it is possible to use the same proof to show
that particular affirmatives convert from universal affirmatives.27
But this does not mean that Aristotle is using it here. 25
That proof by means of exposition is something other than proof by
means of conversion he has shown most evidently in the case of the
third figure.28 For he takes the combination in which both extremes
are predicated of all the middle term, and first proves the syllogistic
conclusion by conversion of one of the premisses. Then, after this, he 30
conducts a proof by means of exposition, suggesting that it is a
different proof and distinct from proof by means of conversion. And 34,1
the proof he uses here in the case of the conversion is very similar to
the proof by exposition which he uses there.
This, then, is the proof in the case of universal negatives. Aristotle
conducts it without using either reciprocal proof or syllogistic 5
reduction to the impossible. The latter would not have been timely,
and even if a proof by way of the impossible fits what he says, it does
not follow that he actually used it to prove the point at issue. For the
expository mode of proof is different from the reductive.
A29 proof of the point at issue by way of the impossible comes about
in the third figure: hypothesize that B holds of some A, i.e. of every C
(where C is something of A), and co-assume the premiss 'A holds of 10
every C', which is obvious: then it can be deduced in the third figure
that A holds of some B. But this is impossible, since it was supposed to
hold of none. However, a proof through the third figure is untimely
here.30 So it is better to say that his remarks show that B must be
disjoined from A if A is disjoined from B - this is what Theophrastus 15
26
ForDarapti see Aristotle, An. Pr. 28a22-6; below, 99,19-100,26.
27
So Alexander, Conv. 63.
28
See Are. Pr. 28a22-6; below, 99,19-100,26.
29
This paragraph and the next are repetitive and inconsistent: they probably
represent an early version of Alexander's commentary.
30
Later commentators did find a reduction in the text (e.g. Philoponus, in An. Pr.
47,14-15).
 90 Translation
assumes without proof, as being obvious.
It is also possible to prove it through a syllogism in the first figure
(just as Aristotle invokes reduction to the impossible31). If it is
denied that universal negatives convert, then suppose that A holds
of no B, and - if it does not convert - let B hold of some A. We get, in
20 the first figure, that A does not hold of some A - which is absurd.32
But it is untimely to prove something by a syllogism when we do not
yet know about syllogisms. Hence we should use the method of proof
which Aristotle set out.33

3.2.3 Other propositions

<If A holds of every B, then B will hold of some A. For if of none,
A will hold of no B. But it was supposed to hold of every B. -
Similarly if the proposition is particular. For if A holds of some
B, then it is necessary for B to hold of some A. For if of none,
then A will hold of no B.> [1.2,25al7-22]
Having proved the negative conversion, he invokes what he has
proved for what he will prove next. For he proves that particular
25 affirmatives convert by invoking the conversion of universal
negatives. If A holds of some B, then B holds of some A; for if of none
(this is the opposite of 'of some'), then A will hold of no B, since
universal negatives have been proved to convert. But it was
supposed that A holds of some B. Therefore B will hold of some A.34
30 That particular affirmatives convert from universal affirmatives
is proved in the same way: if A holds of every B, then B holds of some
35,1 A. For if of none, we have a universal negative, which has been
proved to convert from itself. Hence A will hold of no B. But it was
supposed to hold of every B.
He does not mean that universal affirmatives never convert from
31
The Greek seems to mean something like this: 'We can also prove the conversion
by using a reduction which involves a first figure syllogism -just as Aristotle proves
the conversion by using a reduction which involves a third figure syllogism.' But
Alexander expressly denies that Aristotle uses a reduction. Perhaps, then, the text
means rather this: 'We can also prove the conversion by using a first figure syllogism,
invoking a reduction of the sort which Aristotle uses elsewhere.' But it is hard to
believe that the Greek can convey this sentiment.
32
"This is perhaps the neatest example of an argument by substitution derived
from an ancient source' (Lukasiewicz, p. 10).
33
With the paragraph cf. Alexander, Conn. 65.
34
Note that Alexander inverts Aristotle's order of exposition, taking the
particular-particular conversion before the universal-particular conversion. So too at
Cony. 63, where the universal-particular conversion is proved from the
particular-particular conversion together with the law of subalternation. (If A holds
of every B, then A holds of some B. If A holds of some B, then B holds of some A. Hence
if A holds of every B, B holds of some A.) So too e.g. Boethius, Int. Syll. Cat.
786D-787A.
 3. Conversions 91
themselves - they do convert in certain material instances (e.g. in the
case of co-extensional terms and properties).35 But since they do not
always convert, whereas particulars always do (for the particular is 5
true even when the universal converts), for this reason he says that
particulars convert from universals.36 For conversions - and in
general deductions in the figures - do not depend, as I have already
said,37 on the peculiarities of matter (which is different in different
cases) but on the nature of the figures themselves. That is why proofs
in their case are universal.
<If A does not hold of some B, it is not necessary that B should
not hold of some A. E.g. if B is animal and A man: man does not
hold of every animal, but animal holds of every man.> [1.2,
25a22-6]
That particular negatives do not convert he proved by way of a 10
material instance. For if there are any occasions or cases where it
does not convert, universality is lost. Although it is true that
Man does not hold of every animal (or: does not hold of some
animal),
the proposition saying that
Animal does not hold of every man (or: does not hold of some
man)
is not true. For if its opposite is assumed, nothing impossible
follows, as it does - so we have proved - with propositions which do 15
convert. 'Of every' is opposed to 'not of some'. If we assume that B
holds of every A, it will follow that A holds of some B. It was
supposed not to hold of some B. But it is not impossible that what
holds of some should also not hold of some. Thus man, which does
not hold of some animal, also holds of some animal.38

3.3 Modal conversions39

3.3.1 Necessary propositions

3.3.1.1 Necessary universal negatives

The case will be the same for necessary propositions. [1.3, 20
25a27-8]
35
On this argument see Barnes (1990a), III.7.
36
Reading autei (with B and the Aldine) for hautei (Wallies: LM have heautei).
'Particular affirmatives always convert from universal affirmatives; for even when "A
holds of every B" yields "B holds of every A", it also - and thereby - yields "B holds of
some A".'
37
See above, 30,21-6.
38
With this paragraph compare Alexander, Conv. 65-6; and see above, 30,26 note.
39
On the issues raised in this section see esp. Wieland.
 92 Translation
He says that necessary propositions convert in the same way as
actual propositions, and here too he first proves that universal
privatives convert from themselves, thus:

25 For if it is necessary for A to hold of no B, it is necessary for B to

hold of no A; for if it is contingent that it holds of some, then it
will be contingent that A holds of some B. [1.3,25a29-31]
Actual universal negatives have been proved to convert; and since
36,1necessary universal negatives differ from actual universal negatives
inasmuch as they hold by necessity, he conducts the proof with
reference to the point at which what is now being proved differs
from what has already been proved (i.e. with reference to the
necessity). Let it be the case that B does not hold of no A by necessity
5 but that it is contingent that it holds of some A (this is what rejects
or contradicts the necessary universal negative, as has been proved
in the discussion of affirmation40). Then it will be contingent, he
says, that A holds of some B. But it was supposed that A holds by
necessity of no B.
Here again he seems to have used the conversion of contingent
particular affirmatives in his proof for necessary universal
10 negatives, even though he has not yet discussed conversions of
contingent propositions.41 Or should we rather say this? He holds it
to be agreed that contingent particular affirmatives are opposite to
necessary universal negatives (they are contradictories), and
therefore assumes this. Having assumed it, then since what holds of
some (if it does not hold by necessity) is said to be contingent or to
15 hold of it contingently, and since he has proved that actual
particular affirmatives convert from themselves, it is them which he
invokes. Thus he rejects the necessity by saying that it is contingent
that A holds of some B; and he invokes the fact that what holds of
some - when it holds - converts:42 if it is contingent that it holds of
some, then either it already holds of it or else it can hold of it at some
time; and whenever it holds, it converts. In this way what by
40
Reading kataphaseos, with LM and the Aldine, for antiphaseos: we follow the
majority reading, for there is nothing else to go by. Alexander does not appear to be
referring to a work of Aristotle's (but Philoponus, in An. Pr. 52,4-56,5, reads the point
into Int. 22a24-33); perhaps Alexander refers to a work of his own - a section of his
lost commentary on Int. or an otherwise unknown essay?
41
For the charge see Alexander, Conv. 76-7; Philoponus, in An. Pr. 52,27-9. A.
Becker, p. 90, argues that the text of An. Pr. does contain the 'vicious circle' which
Alexander and Philoponus try to explain away; and he infers that An. Pr. 25a29-34
was interpolated by a later hand. Ross, pp. 294-5, agrees that there is a vicious circle,
but he thinks that 'the reasoning is so natural' that Aristotle himself is probably
responsible for it.
42
Alexander's syntax is uncertain here (if indeed the text is not corrupt): we take
toi... antistrephein to depend onprosekhresato, understood from 36,15.
 3. Conversions 93
necessity holds of none will at some time hold of some, which is 20
impossible. That 'contingent' is said of what holds but does not hold
by necessity he says a little later on when he distinguishes kinds of
contingency: he says that it signifies both what is necessary and also
what is not necessary but actual43 - and this is the sense in which he
now uses it. And the opposite of what by necessity holds of none is
what contingently holds or will hold of some.44
What is necessary either is necessary simpliciter or is called 25
necessary with a qualification45 - e.g.
Man holds by necessity of every grammarian so long as he is a
grammarian.
This proposition is not necessary simpliciter. (Theophrastus showed
the difference between them.46) For there are not always
grammarians, and men are not always grammarians. Since they
differ in this way, we must recognise that Aristotle is here 30
discussing what is necessary simpliciter and in the strict sense - it
is necessary propositions of this sort which convert.

3.3.1.2 Other propositions

If A holds by necessity of all or of some of the Bs, then it is 37,1
necessary for B to hold of some of the As. <For if it is not
necessary, A will not hold of some B by necessity.> [1.3,
25a32-4]
That necessary particular affirmatives convert both from necessary
43
See An. Pr. 25a38-9 (cf. e.g. Int. 23a6-ll; Metaph. 1019b30-3).
44
(1) Philoponus elaborates the following argument on Aristotle's behalf (ire An. Pr.
56,15-57,13):
Suppose that necessarily A holds of no B and that it is not necessary that B holds
of no A. Then it is possible that B holds of some A. Then at some time, t, B does
hold of some A. Then at t A holds of some B. Then it is possible that A holds of
some B. Hence it is not necessary that A holds of no B. But this contradicts the
supposition.
Philoponus' reduction is sound, and it does not invoke the conversion of 'Possibly
some A is B' in any vicious fashion. (2) Alexander appears to argue thus:
Suppose that necessarily A holds of no B and that it is not necessary that B holds
of no A. Then it is possible that B holds of some A. Then B holds of some A. (For
here 'Possibly P has the force of'P.) Then ...
This is unsound; for it says nothing of the cases in which 'Possibly P' does not have the
force of'P. (Thus A. Becker, pp. 84-6, who finds the error again in 37,26-41,29 - and
also, presumably, at 37,15-18.) But this is an uncharitable interpretation of
Alexander. If we disregard, as a trifling slip, the words hoion te ('... or else it can hold
of it ...') at 36,18, then we may find the sound Philoponan argument in Alexander's
text. (No doubt Philoponus himself found it there.) And a similar defence can be made
in the other two passages where Becker finds the same error.
45
For the distinction see also Alexander, in An. Pr. 140,14-141,6; 155,20-5;
156,27-157,2; 201,21-4; cf. e.g. Ammonius, in Int. 153,13-26; 240,7-13. The ultimate
source is Aristotle, Are. Pr. 30b35-40.
46
F14 Graeser: see Bochenski (1947), pp. 73-4.
 94 Translation
universal affirmatives and from necessary particular affirmatives, he
5 proves in the same way as he did in the case of universal privatives. If
A holds of every or of some B by necessity, and B does not hold of some
A by necessity, it will be contingent that B holds at some time of no A;
for the negation of 'It is necessary for it to hold of some' is 'It is not
necessary for it to hold of some', which is equivalent to 'It is
contingent that it holds of none'. ('It is not necessary for it to hold of
some' and 'It is contingent47 that it holds of none' are the same.) But
10 when B holds of no A, A will hold of no B (this has been proved).
Hence, A will not hold by necessity either of every or of some B.48
(Aristotle omitted 'of every' and was content to take it that A no longer
holds of some B by necessity.49 For instead of saying 'For if it is
contingent50 for it to hold of none', he simply said 'For if it is not
necessary'.)
It is clear that he has not conducted the proof with contingent
15 negatives; for he holds that they do not convert. Rather, he brings it to
an actual negative, thus51 subtracting the necessity. (He makes this
clear by no longer using 'contingent', but simply 'For if it is not
necessary'.) For he is supposing that actual negatives convert. (It is
clear from this that in the previous proof too he used 'It is contingent
20 that it holds of some' for an actual proposition; for there 'For if it is
contingent that it holds of some' should be understood to be used
instead of'For if it holds contingently of some'.52)

<The particular privative does not convert for the same reason
as we gave earlier.> [1.3, 25a34-6]
He says that no proposition converts from a necessary particular
negative, because none converted from an actual particular
negative; i.e. because it is proved in the same way - by way of terms
and material instances - that they do not convert. The refutation
uses the same terms:
By necessity man does not hold of some animal,
25 and
By necessity animal holds of every man.

47
Reading endekhetai (with LB2: endekhesthai M) for ananke.
48
See above, 36,11 note, for the argument.
49
And rightly, since from 'It is not necessary for A to belong to some B' we may
infer 'It is not necessary for A to belong to every B'.
50
Reading endekhetai for ananke. Note that Philoponus, in An. Pr. 57,24-58,27,
knows - and rejects - a variant reading in Aristotle's text: endkhetai medeni for me
anank.
51
Reading tautei for tautSs (aphelon here has the same sense as anelon at 36,16).
52
Above, 36,20-3.
 3. Conversions 95
53
3.3.2 Contingent propositions

3.3.2.1 Contingent affirmatives

As for contingent propositions, since being contingent is meant
in several ways <(we say that the necessary and the
non-necessary and the possible are contingent), in affirmatives
the case will be similar in respect of conversion in all instances.
If it is contingent for A to hold of every or of some B, then it will
be contingent for B to hold of some A. For if of none, then A of
no B (this has been proved earlier).> [1.3, 25a37-b3]
He showed us the homonymy of 'contingent' in On Interpretation
too.54 For we use 'contingent' both of what is necessary, when we say 38,1
that it is contingent that animal is said of every man, and of what is
actual, if we say of what holds of something that it is contingent that
it holds. Here in saying 'the non-necessary' Aristotle means what
holds; for what holds differs in this way from what is necessary
while sharing with it the fact of holding at the present time. (Note 5
the expression: what holds contingently is the same as what is
signified by an actual proposition.) 'Contingent' is also said of what
is possible. He will explain what this means a little later on when he
refers to 'those which are said to be contingent inasmuch as they
hold for the most part and by nature - this is the way in which we 10
determine the contingent.. ,'.55
Now given that 'contingent' is homonymous and is predicated of
these several cases, he says that however the contingency is taken
(whether as necessary or as actual or as possible) contingent
affirmative propositions convert in the same way: particular
affirmatives convert from universal affirmatives, and similarly 15
particular affirmatives convert from particular affirmatives, i.e.
from themselves. He proves this by invoking what he has previously
said and proved. If it is contingent that A holds of every or of some B,
then it will be contingent that B holds of some A; for if it is
contingent that B holds of no A, then we have a universal negative,
and this has been proved to convert from itself. If contingency
signifies what is possible, then the negative which is assumed and 20
hypothesized is a necessary universal (for 'It is not contingent that it
holds of some' has this character), and this has been proved to
convert. If it signifies what is actual, then the proposition
hypothesized is an actual universal negative, and this too has been
53
The lines of An. Pr. to which Alexander now turns are extraordinarily difficult:
see e.g. Maier Ha, p. 25 n. 1; A. Becker, pp. 59-65; 83-91; Ross, pp. 295-8; Hintikka,
pp. 35-8.
54
See Int. 23a7-20. On the present passage see Sharpies (1982).
65
Aristotle, ATI. Pr. 25bl4-15.
 96 Translation
proved to convert. And if 'It is contingent that A holds of every or of
some B' signifies what is necessary, it has been proved in this case
25 too that its opposite, a contingent universal negative, will be such
that at some time it holds of no B (for this is the way in which what
is contingent is true); and it has been laid down that actual
negatives convert.56 (This was how he conducted the proof in the
case of necessary universal and particular affirmatives.57)

3.3.2.2 Contingent negatives

<In negatives it is not the same. With those which are said to
be contingent inasmuch as it is by necessity that they do not
hold or inasmuch as it is not by necessity that they do hold, the
case is indeed similar. E.g. if someone were to say that it is
contingent for man not to be a horse or that it is contingent for
white to hold of no cloak. ... For if it is necessary for it to hold of
some, then white will hold of some cloak by necessity - this was
proved earlier. (Similarly too for particular negatives.)> [1.3,
25b3-14]58
He says that the conversions will hold in the same way in the case of
all the affirmatives, but that this is no longer so in the case of
30 negatives. Rather, where contingency signifies what necessarily
does not hold or what holds not by necessity (he means actual
negatives in saying 'or inasmuch as it is not by necessity that they
do hold') - when contingency is taken in these senses, the
56
(1) Alexander argues as follows. If it is not contingent that B holds of some A,
then it is contingent that B holds of no A. Then, since 'contingent' is ambiguous, there
are three cases to consider. 'It is contingent for A to hold of some B' means either
'Possibly A holds of some B' or 'A holds of some B' or 'Necessarily A holds of some B'.
The two latter propositions have already been proved to convert, and the first is
readily proved to convert too. Hence in whatever way 'contingent' is construed,
contingent particular affirmatives convert. Thus Alexander's argument rests on the
thought that 'contingent' is ambiguous. But Aristotle does not hold, and it is not true,
that 'contingent' is ambiguous; and Alexander's argument is thus based on a
misconception. (2) Given that 'It is contingent that P' means 'Possibly P and possibly
not-F, then the conversion can be shown as follows. Suppose that it is not contingent
that B holds of some A. Then either it is necessary that B holds of some A or it is
necessary that B does not hold of some A. Hence either it is necessary that B holds of
some A or it is necessary that B holds of no A. Hence (by the conversions already
established) either it is necessary that A holds of some B or it is necessary that A
holds of no B. Hence it is not contingent that A holds of some B. But it is laid down
that it is contingent that A holds of some B. (3) Aristotle's text remains obscure. When
he says 'For if of none ...' he ought to mean (as Alexander sees) 'For if it is contingent
that B holds of no A'; but it is hard to see how this hypothesis, which is not the
negation of'It is contingent that B holds of some A', can ground the proof.
67
58
See above, 36,16-21; 37,15-18.
Note that Alexander's text of Aristotle differs from the modern text at a crucial
point in this passage (see Appendix 1). In consequence, his interpretation of Aristotle
will not match our text.
 3. Conversions 97
conversions are the same. And this is reasonable enough. For in
them only the name is changed and what is signified is the same as
in the cases which have already been proved. In the case of
contingencies of this kind, universal negatives convert from 39,1
universal negatives, and no proposition converts from particular
negatives.
Of his examples,
It is contingent for man not to be a horse
signifies a necessary negative, whereas
It is contingent for white to hold of no cloak
signifies an actual negative. The conversions which he sets down are 5
plain. He said 'For if it is necessary for it to hold of some, then white
will hold of some cloak by necessity', since necessary particular
affirmatives must be opposed to contingent universal negatives, and
the actual proposition supposed was expressed in terms of
contingency. Its opposite, then, will contain necessity in its
expression, but what is signified by it will be an actual particular 10
affirmative; for this is opposite to an actual universal negative. And
if we take a particular affirmative which is necessary in the strict
sense, this too has been proved to convert.
He says 'Similarly too (sc. it will hold) for particular negatives'; for
no proposition converts from contingent particular negatives, when
they signify either what is necessary or what is actual. 15

But as for those which are said to be contingent inasmuch as

they are for the most part and by nature <- this is the way in
which we determine the contingent - here the case will not be
similar for negative conversions. Rather, the universal
privative proposition does not convert whereas the particular
does convert. This will be evident when we discuss the
contingent.> [1.3,25bl4-19]
The third sense of the contingent was the possible. Here in saying
what it is, he speaks of 'those which are said to be contingent
inasmuch as they are for the most part and by nature'. He set down
only this sort of contingency59 - what holds for the most part and is 20
by nature (for what is by nature holds for the most part), since only
this sort is useful for syllogistic purposes. The possible also covers
what holds in equal part and what holds for the lesser part, but
syllogisms with matter of this kind are of no use. For no art and no
science deals with this sort of contingency and possibility. For in the
case of such matter the opposite will hold no more and no less than 25
what is proved syllogistically to hold - and in the case of what holds
69
Perhaps we should read dunatou for endekhomenoul
 98 Translation
for the lesser part, it is actually the opposite which holds more than
what is deduced as Aristotle himself will say when he discusses
syllogisms of the contingent.60 Thus he deals only with the kind of
contingency and possibility which holds for the most part and which
is by nature or according to nature (which itself holds for the most
30 part). For some natural events are indeed proved on such a basis,
and there are some arts concerned with what is in this way
contingent - for instance, the conjectural arts.61 A doctor assumes
40,1 that someone who is ill in such-and-such a way is for the most part
unwell from surfeit, and that someone who is unwell from surfeit is
for the most part cured by venesection; and he deduces that it is
contingent that someone who is ill in this way will be cured by
venesection. Taking this to be so, he performs a venesection.62 For
5 this reason, there is syllogistic utility in these things.
In the case of what is said to be contingent in this sense, he says
that negative conversions are no longer the same as those he proved
in the case of necessary and actual propositions. For in the latter
cases, universal negatives convert from themselves and nothing
converts from particular negatives. But in the former case he says
10 that contingent universal negatives do not convert with themselves
(but do convert with particulars: this is what he means), and that
particular negatives convert from themselves. Thus it is the other
way about here in comparison with the former modes. He says that
he will explain the reason for such conversions when he discusses
the syllogisms which come about from contingent propositions.63

15 For the moment, in addition to what we have said, let it be

clear that 'It is contingent that it holds of none or not of some'
has an affirmative figure. <For 'It is contingent' is ranked in
the same way as 'is', and 'is' (where it is co-predicated) always
and in every case makes an affirmation - e.g. 'It is not good' or
60
See An. Pr. 32bl8-22 (cf. e.g. An. Post. 87bl9-27; Metaph. 1027al9-26;
Alexander, in An. Pr. 164,17-165,15 (cf. 168,31-169,10); Philoponus, in An. Pr.
61,16-62,4). For syllogisms involving what holds 'for the most part' see also An. Pr.
43b32-6; An. Post. 96a8-19; cf. Maier, Ha, p. 138 n. 1; A. Becker, pp. 76-83; Ross,
p. 328; Mignucci (1969), pp. 303-4; Barnes (1969), pp. 133-5; (1975), pp. 184, 229; and
esp.
61
Mignucci (1981).
For which see, elsewhere in Alexander, in An. Pr. 165,8-15; 300,3; in Top.
32,12-34,5; Quaest. 61,1-28. The distinction is a commonplace (e.g. Philodemus, Rhet.
v 34-vi 19; Sextus, MII13) and goes back by way of Aristotle (e.g. EN 1112a34-b8) to
Plato (Philebus 55E-56C). The rules or general principles of an art guarantee success
in the sense that if you apply them correctly you will achieve your end. An art is
conjectural if it can only guarantee success for the most part.
62
Medicine was a stock example of a conjectural art (e.g. Galen, Meth. Med. X 206K;
Opt. Sect. 1114-15K); cf. the reference to the art of medicine at Alexander, in An. Pr.
165,10 (and see Aristotle, Metaph. 1027a23-4).
63
See ATI. Pr. 36b35-37a31; Alexander, in An. Pr. 219,35-221,13.
 3. Conversions 99
'It is not white' or generally 'It is not such-and-such' (this too
will be shown in what follows).> [1.3, 25bl9-25]
Aristotle says here that those propositions in which 'to be
contingent' is co-predicated are affirmations (just like those
propositions in which 'is' is predicated without 'to be contingent'). 20
This is true, and he stated and proved it in On Interpretation.64 Yet
the fact that contingent negatives are actually affirmations is not a
sufficient explanation of why they convert in the same way as the
other affirmations, i.e. of why particulars convert both with
universals and with particulars. If this were sufficient the same
would have to hold of necessary negative propositions; for he showed 25
that propositions in which necessity is co-predicated are also
affirmations.65 But even though they have this character, he proved
that necessary universal negatives convert from themselves -
although they are affirmations (for necessary negatives, although
they are affirmations simpliciter, are called necessary negatives as a
whole); and no proposition converted from necessary particular
negatives. Thus he was not giving an explanation of the fact that 41,1
contingent universal negatives do not convert when he said what he
posits here; rather he hypothesized it as something to be made
evident later on, and he intimates to us that we must regard them as
affirmations and investigate their conversions in this light.
He will give the true explanation of why they must convert like 5
affirmations when he proceeds further. For since contingent
universal negatives and contingent universal affirmatives convert
from each other and are true together if their terms are preserved in
the same order (in the case of what is called contingent in this way),
because of this the conversions of contingent universal negatives are
like the conversions of affirmatives. For what contingently holds of
none also contingently holds of all, and what contingently holds of 10
all also contingently holds of none.66 This is not so for other
propositions but only for those which are contingent in this sense.
Thus it is because universal negatives are in this respect equivalent
64
See Aristotle, Int. 21a34-22al3 (cf. An. Pr. 32a31-b3; 51b3-52b34) and the long
discussion in Ammonius, in Int. 221,11-229,11 (esp. 227,3-25, on Are. Pr. 25bl9-25).
See also Philoponus, in An. Pr. 53,15-56,5 (cf. 62,28-63,3) for a dispute over the
matter (cf. Maier Ila, p. 27 n. 1).
65
i.e. at/^. 21a34-22al3.
68
See Aristotle, An. Pr. 32a29-bl; cf. e.g. Alexander, in An. Pr. 158,24-161,2;
218,20-4. The conversion was rejected by Theophrastus and Eudemus (whom the
later Platonists followed): Alexander, in An. Pr. 199,7-10; [Ammonius], in An. Pr.
45,42-46,1 [= Eudemus, F 13 Wehrli]; 49,7-12 [= F 14 Wehrli] (all placed under
Theophrastus, F 16 Graeser). The conversion depends on Aristotle's 'two-sided'
conception of contingency (above, 26,14 note): in effect, Theophrastus and Eudemus
jettisoned the two-sided conception, and thus rejected the conversions associated
with it (cf. below, 41,23 note). See e.g. A. Becker, pp. 14-15; Bochenski (1947), pp.
74-6; Barnes (1969), p. 135.
 100 Translation
to universal affirmatives - and are in any case affirmations - that
they convert in the same way, and not because they have
'contingent' co-predicated.
For given that the negation is equivalent to the affirmation, if the
15 universal negative converted, then contingent universal affir-
matives would also convert with themselves. But this is false. For
suppose someone were to say that, if it is contingent that A holds of
no B and also of every B, then it is also contingent that B holds of no
A: then inasmuch as it would also be contingent that it holds of
every A, contingent universal affirmatives would convert from
themselves, which is not true. For if it is contingent that every man
20 is walking, it is not thereby contingent that everything which walks
is a man.
This, then, is why Aristotle denies that contingent universal
negatives convert from themselves.67 But Theophrastus says that in
fact they convert in the same way as other negatives.68 I shall
discuss the difference between them in more detail when I talk
about contingent propositions.69

25 And in respect of conversions the case will be similar to that of

the others. [1.3, 25b25]
Having said that contingent negatives are affirmatives, he says that
in respect of conversion they will behave in the same way as the
others, sc. as the affirmatives.70 Contingent particular negatives71
convert both from contingent universals and from contingent
particulars. For this is how the affirmatives convert.

67
Alexander has in effect summarized the argument of .An. Pr. 36b35-37a31: cf. in
An. Pr. 221,16-227,9.
68
F 15 Graeser; cf. Alexander, in An. Pr. 219,35-221,5 [= Eudemus, F 16 Wehrli];
anon, in Int. An. Pr. 100 Mynas [= Eudemus, F 15 Wehrli]; cf. A. Becker, pp. 65-7;
Bochenski (1947), p. 74 (see above, 41,10 note).
69
i.e. at in An. Pr. 219,35-221,5.
70
So too Philoponus, in An. Pr. 63,19-64,7; Ross, pp. 299-300. For a contrary view
see Maier, Ila, p. 27 n. 1; see also Mignucci (1969), pp. 209-10.
71
Reading apophatikei (Aldine) for kataphatikSi.
 4
The First Figure

4.1 Introductory

4.1.1 General remarks

Having determined these issues, let us now say by what means 41,30
and when and how every syllogism comes about. We must
discuss demonstration later. [1.4, 25b26-8]
After these preliminary remarks he proposes to discuss the issues to 42,1
which they were preliminary.
As to by what means syllogisms come about, they come about by
means of premisses. Since syllogisms come about by means of a
particular compounding of premisses, he added 'when' and 'how' to
'by what means'. 'When' signifies the combinations and figures: in
the figures and combinations in which syllogisms come about the 5
premisses must share a term. It also signifies the quality of the
premisses: syllogisms do not come about from all compounded
premisses (they do not come about from two negatives or from two
particulars, as he will show).1
'How' means the combinations in each figure2 and the modes of
the syllogisms. For if a syllogism is to come about from the
premisses, they must be compounded not in any way whatsoever but 10
in the appropriate order. For syllogisms come about by means of
certain kinds of premisses and in figures (which he will discuss), and
in these figures they come about according to a particular
conjunction of the premisses with each other. For in each figure
there are both non-syllogistic and syllogistic combinations,
depending on the particular compounding of the premisses.
Thus: 'By what means?' - By means of premisses. 'When and 15
how?' - From premisses of such-and-such a sort, compounded in
such-and-such a way - where 'when' makes clear that one must take
such-and-such a sort of premisses and 'how' a compounding of
such-and-such a type.
1
See Aristotle, An. Pr. 41b6-31; cf. Alexander, in An. Pr. 266,8-270,8.
2
Reading skhema for skhematon.

101
 102 Translation
<We must discuss syllogisms before demonstration because
syllogisms are more universal: a demonstration is a syllogism,
but not every syllogism is a demonstrations [1.4,25b28-31]
Here he plainly presents the reason why, after having proposed to
discuss demonstration, he first deals with syllogisms: it is because
20 syllogisms are 'more universal'. And he tells us in what sense they
are universal by showing that there is no conversion: 'a
demonstration is a syllogism, but not every syllogism is a
demonstration.' Since syllogisms stand in this relation to
demonstrations they are primary with regard to them. This indeed
is the explanation of the titles Prior and Posterior Analytics which
25 have been given to these studies, as we have said.3 For in order to
define demonstrations and to place them in the genus of syllogisms,
one must learn first what a syllogism is, and by what means it comes
about, and when, and how these means must relate to each other.
He has added 'every syllogism',4 although he discusses only
predicative syllogisms, because he thinks that only these are
30 syllogisms in the strict sense, as he will show later on. For he thinks
that no syllogism from a hypothesis syllogizes the point at issue.5

4.1.2 Syllogisms and justification

43,1 Now when three terms so stand to one another that the last is
in the middle as in a whole ... [1.4, 25b32-3]
Now comes his account of the syllogistic figures - the figures in
which all syllogisms, both perfect and imperfect, have their
5 construction; and he first discusses what is called the first figure, in
which the perfect syllogisms are generated.
A syllogism is a type of justification;6 for when you syllogize
something, you justify and prove it by means of certain things - that
is to say, when you syllogize, you justify a disputed point by means
of things which have been justified. But not all justifications are
10 syllogisms. So we could do worse, perhaps, than say a few things
about justification; for in this way syllogistic justification will be

3
See above, 6,32-7,11.
4
i.e. in An. Pr. 25b27, in the previous lemma.
5
See An. Pr. 50al6-b4; Alexander, in An. Pr. 386,5-30 (cf. above, 24,18 note); on
syllogisms 'from a hypothesis' in Aristotle see Striker; Lear, ch. 3. Alexander is no
doubt alluding, generally, to 'hypothetical' syllogisms: a hypothetical syllogism is a
syllogism in which at least one premiss is a hypothetical proposition (on hypothetical
propositions see above, 11,20 note). For the stock contrast between categorical and
hypothetical syllogistic see e.g. the little essay in [Ammonius], in An. Pr. IX 22-XI36;
cf. e.g. Mueller (1969); Frede (1974b); Barnes (1983), (1985); Maroth.
6
cf. Aristotle, Are. Pr. 68b8-14.
 4. The First Figure 103
7
better known.
Now anyone who justifies something or wishes to make it known,
justifies it on the basis of some other item or items. Nothing is
proved and justified on its own basis, otherwise everything would be
justified. Since, then, justification is based on some other item,
either you prove the point at issue on the basis of what is whole and
universal and inclusive (the point at issue is either some part of that 15
by means of which the justification is effected or, more generally, is
something of it; and it is proved by means of it), or you prove the
whole on the basis of its parts or a part on the basis of some part
falling under the whole.
When a part is proved or justified on the basis of a part, such a
proof (and the justification so conducted) is called a paradigm.8 For
instance, someone opposes Dionysius' demand that the Syracusans
give him a bodyguard on the grounds that if he gets it he will aim at 20
a tyranny; and he justifies this by observing that when Pisistratus
got the bodyguard he demanded from the Athenians, he became a
tyrant; and that Theagenes did the same to the Megarians. Each of
the items - both what is being proved and that by means of which
the proof is effected (and the demand opposed) - are parts of a
whole: the whole being the fact that all those who demand a
bodyguard aim at tyranny.9 So he produces a justification10 by way 25
of the one part, which he uses to establish the other; and the
argument proves an unknown and disputed part by means of a
known and justified part.
When the whole is justified and proved on the basis of the parts,
such a proof is called an induction.11 For instance: men move their
7
Aristotle says that 'a justification is a sort of demonstration; for we take
something to be justified par excellence when we deem it to have been demonstrated'
(Rhet. 1355a4-6: cf. Grimaldi, pp. 349-56; Sprute, pp. 58-67); and he analyses
justifications by way of the part/whole relation (Rhet. 1357b26-30; cf. An. Pr.
49b37-50al; 64al5-17, bll-13; 69al3-16; see Introduction, p. 28, on 'in as in a whole',
and in general Barnes (1988)). Usually he distinguishes two varieties of justification:
syllogisms (the rhetorical version of which is the enthymeme) and inductions (the
rhetorical version of which is the paradigm) see Rhet. 1356a35-blO; An. Pr. 68b8-14;
Top. 103b3-7; SE 165b27-8 (followed by Alexander, in Top. 62,6-13; Apuleius, Int.
185,10-20). But at An. Pr. 69al6-19 he implicitly marks three varieties of
justification, inasmuch as there he distinguishes between inductions and paradigms;
and the later tradition generally adopts this three-fold division (e.g. Alexander in the
present passage; Galen, Simp. Med. Temp. XI 470K; Sem. IV 58 IK; Ammonius, in
An. Pr. 28,21-29,19; Philoponus, in An. Pr. 34,12-30; David, in Porph. hag.
88,12-89,10). Note that here Alexander, following Aristotle, confounds the notion of
an argument with the notion of a proof: see esp. 44,14, where syllogisms are simply
saidtobeproo/s.
8
The definition and the example of Dionysius (for whom see Diodorus Siculus, XIII
91-5) are taken from Aristotle, Rhet. 1357b25-1358a2 (cf. 1393a25-1394al8).
9
See Plato, .Rep. 566B.
10
Reading ho de [with the MSS: ho de Wallies]pism [piston MSS].
11
cf. e.g. Aristotle, Top. 105al3-16; 108blO-ll; see Ross, pp. 481-5; Hintikka
 104 Translation
lower jaw, as do horses, dogs, cows and sheep; therefore every
animal does.12 Induction is a path by way of known and justified
44,1 particulars toward the universal which is unknown. Both these
things (induction and paradigm) contain plausibility but not
necessity.13
Aristotle discusses these types of justification at greater length in
the second book,14 showing how they differ from syllogistic
justification, that they too are useful, and how15 they are subsumed
5 under syllogistic justification. Thus for present purposes what we
have said about them is enough.
When a part is justified from the whole, such a justification is
called a syllogism; and this is the most compelling16 type of
justification. For anything which applies to or holds of a universal
and a totality, by necessity also holds of what is within it and is
10 included in it. If you assume that all justice is noble and that
everything noble is good, then it is by means of the whole and the
universal - that is, by means of the fact that everything noble is
good - that you prove justice to be good, this being a part of the fact
that everything noble is good, since justice is indeed noble. (Aristotle
has already given an account of what a syllogism is.17)

4.1.3. Problems
15 Now since a syllogism is a proof, by means of agreed items, of
something disputed, and since what is disputed and proposed for
proof is called a problem, let us say a little about problems.18
Problems belong to the genus of propositions;19 for problems and
assumptions and agreements and conclusions and axioms are all the
same in genus - they are all propositions, differing from one another
in aspect.20 When a proposition is put forward for proof as not being
20 known, it is called a problem; when it is assumed in order to prove
(1980); Burnyeat, pp. 200-1.
12
The example is a stock one: e.g. Sextus, PH II 185; Apuleius, Int. 185,10-20;
Ammonius, in An. Pr. 28,32-29,2; Philoponus, in An. Pr. 34,21-6. The usual point of
the example is that the induction is misleading - for the crocodile does not move its
lower jaw (Herodotus, II68; Aristotle, HA 492b23-6; PA 691b5-16).
13
See above, 18,8-12; cf. Apuleius, Int. 185,10-20; Ammonius, in An. Pr. 28,23-5;
Philoponus, in An. Pr. 34,10-14.
14
Paradigms at An. Pr. 68b38-69al9, inductions at 68bl3-37.
15
Reading kaipos forpos.
16
anankastikdtatos: cf. Aristotle, Top. 105al6-19 (cf. An. Pr. 68b35-7); Alexander,
in Top. 86,21-87,6; cf. Long, pp. 137-9.
17
i.e. at An. Pr. 24bl8-22.
18
cf. Aristotle, Top. 101bl6-36 (cf. Alexander, in Top. 40,13-41,16): a 'problem' has
the form of a question: 'Is it the case that P, or not?' (see e.g. below, 46,25).
19
On protasis as a genus see above, 11,4 note.
20
cf. Apuleius, Int. 183,22-6; Ammonius, in Int. 10,2-4; in An. Pr. 26,34-27,4;
Philoponus, in An. Pr. 11,25-36; 22,4-15; in Cat. 11,20.
 4. The First Figure 105
21
something else, it is an assumption or an agreement (as Plato calls
it22) or a proposition in the special sense;23 it is an axiom if it is true
and known through itself; and if it has been proved, it is a conclusion
(for anything which has been proved is a conclusion from the items by
means of which it was proved).
Thus problems belong to the genus of propositions, and every
proposition consists of a subject term and a predicate term - the
subject being that about which something is said, the predicate that 25
which is said about it;24 and again, the subject is that to which the
quantitative determination of the proposition is annexed ('every' or
'no' or 'some' or 'not every')25 and the predicate is that to which is
annexed 'is' or something equivalent to 'is' which includes 'is'
potentially in itself.26 Well then, since there are two terms in a
problem - the subject, and the predicate, which must be proved to 30
hold or not to hold of the subject - there must be some third term to
prove it, a term which, when it is co-assumed, will either bring the 45,1
terms in the problem together or else separate and part them. This
term which is co-assumed from outside must stand in a certain
relation to the two terms in the problem. When it is taken, it becomes
a middle term for them both, dividing the problem and making the
one proposition two. It is itself introduced into each of the two 5
premisses, being compounded with each of the terms of the problem
in turn, now with the subject and now with the predicate. You can
recognize the middle term, which has been introduced from outside,
inasmuch as it is taken twice and is present in both the premisses.
A conjunction of propositions of this sort is called a combination:
there is a combination when two propositions share one term, which 10
is a middle term. For when propositions differ from one another, then
either they share nothing at all with each other (e.g. 'Everything just
is noble' and 'Every pleasure is good'), or else they share something;
and if they share something, then either they share one term or else
they share the two terms.

21
cf. e.g. in Top. 23,21.
22
See Theaetetus 155AB; cf. Mardth, pp. 25-32.
23
i.e. a premiss: see Introduction, p. 22.
24
See above, 14,27-15,4.
25
For the four 'quantitative determinations' or quantifiers see e.g. below, 65,26-7;
100,11-14; Galen, Inst. Log. ii 5; Ammonius, in Int. 89,2-36. The quantifiers are taken
to modify the subject term (cf. Alexander, in An. Pr. 297,22-3; Philoponus, in An. Pr.
277,29-31; in An. Post. 155,24-5): the predicate term is not quantified (see below,
58,29 note). See also Introduction, pp. 28-9.
26
On the copula see above, 15,23-16,17 and notes. For the thought that it is
annexed to the predicate see above, 16,8-10 (cf. e.g. Alexander, in An. Pr. 406,32-5).
Ammonius says, more persuasively, that 'is' is added to 'the terms' and serves to 'bind
them together' (in Int. 165,10-16).
 106 Translation
4.1.4 Propositions which share two terms27
Now it is among propositions which share the two terms that
15 oppositions and conversions of propositions are found.28 For when
they share their terms in such a way that the terms preserve the
same order in both propositions but differ either in quality or in
quantity or in both, then we find the oppositions, i.e. the different
types of opposition.29
When their terms are supposed in the same way and the quantity
20 is the same in both, but they differ in quality, then if they are
universal they are contraries. For universal affirmatives and
universal negatives, when they have the same terms in the same
order and differ only in respect of being affirmative and negative,
are contraries. When both are particular, they are called
subcontraries (these seem to stand in opposition to one another but
30
25 are not in fact opposed ). When their quality and also the order of
their terms are the same but they differ in quantity, then they are
subalterns (and these are not in any way opposed). Universal
affirmatives stand in this relation to particular affirmatives, and
universal negatives to particular negatives. When they differ both
30 in quantity and in quality and have the same terms ordered in the
same way, then they are contradictories. Universal affirmatives and
particular negatives stand in this relation to one another, as do
46,1 universal negatives and particular affirmatives - these are
contradictorily opposed to one another.
When propositions share their two terms with one another but the
terms in them are not in the same order but are taken inversely - it
is among propositions which share in this way that prepositional
31
5 conversions are found. For the conversion of propositions is a
matter of their sharing their two terms, inversely posited, and in
addition being true together. When they differ in quality, such
prepositional conversions require an opposition - and they are
called 'conversions with opposition'. When they are the same in
quality, the conversions which are taken in this way and are true
10 together come about without an opposition. Of those which convert
in this way, some preserve their quantity as well as their quality -
27
On this and the following section see Lee, pp. 65-74.
28
On the koinonia of terms see e.g. Ammonius, in An. Pr. 36,8; Philoponus, in An.
Pr. 42,22; Boethius, Syll. Cat. 798C-799A. Galen calls such pairs of propositions
sunoroi or 'co-terminal': Inst. Log. vi 3.
29
On the different types of opposition, which produce the so-called 'square of
opposition', see Introduction, pp. 26-7; cf. e.g. Apuleius, Int. 179,16-181,17;
Ammonius, in Int. 91,4-93,18. The terminology is Aristotelian, apart from the words
'subcontrary' and 'subaltern', which we find as logical terms first in Alexander.
30
Note that Aristotle counts them as opposites at An. Pr. 59b8-ll; but at 63b23-30
he remarks that they are opposites only kata ten lexin, in expression.
31
See above, 29,1-30,6.
 4. The First Figure 107
as was proved in the case of universal negatives, both actual and
necessary (and similarly for particular affirmatives). Others have
the same quality but disagree in quantity, namely those which
convert with universal affirmatives (they are particular affir-
matives), and similarly those which convert with contingent 15
universal negatives (it is contingent particular negatives which
convert with them).

4.1.5 Combinations and figures32

These are the different ways in which propositions may share their
two terms. When they share one term, they make combinations, as I
have already said;33 and the different ways in which the middle
term is shared in relation to the extremes - the terms in the problem 20
- determine the different syllogistic figures. For this middle term,
which is taken twice and connected with each of the extremes
(which were the parts of the problem) may be so taken that it
actually holds a middle position between them, being predicated of
the subject in the problem and being subject for the predicate. Thus
consider the problem: Is it the case that every just thing is good, or 25
not? If we introduce a term, noble, from outside and predicate it of
just (which was subject) and make it subject for good (which was
predicate), then we get the following combination:
Everything just is noble.
Everything noble is good.
When the middle term has this position, and when the premisses
share the middle term in this way, we call this the first figure: when
there is a combination in which the middle term, which is taken 30
twice, is predicated of one of the terms in the problem and made
subject for the other, then this is called the first figure.34
Or again, the middle term may be predicated of both the terms in
the problem, as in the following combination. Suppose we are
investigating whether men are neighers or not. We take a third
term, rational, and predicate it of both the others - both of man and
of neigher (affirmatively of man and negatively of neigher). This 47,1
makes the following combination:
Every man is rational.
No neigher is rational.
This way of sharing the middle term in relation to the extremes
32
On the definition of the figures see esp. Patzig, pp. 88-108.
33
Above, 45,8-10.
34
cf. e.g. below, 53,2-8; Alexander, in An. Pr. 349,5-7. Note that Alexander's
definition is subtly but significantly different from Aristotle's at An. Pr. 25b32-5: in
effect Alexander excludes, while Aristotle implicitly leaves room for, a fourth figure
(below, 47,12 note). See esp. Lukasiewicz, pp. 27-8.
 108 Translation
makes what is called the second figure.
There is also a third way in which the middle term may be shared
5 in relation to the extremes, different from the ways so far
mentioned. This occurs when the term taken as middle is subject for
both the terms in the problem. Imagine that we are investigating
whether or not any substance is animate. We take animal as middle
term, and make it subject for both terms (both for substance and for
animate). Then we get:
Every animal is a substance.
Every animal is animate.35
10 In this combination the middle term is subject for both. Such a
conjunction of middle and extreme terms makes what is called the
third figure.
There are only these three figures;36 for (1) it is in general the way
of taking the middle term which generates the combinations and the
figures, and (2) it is impossible for the middle term to be connected
to the two extremes by any relation among three terms other than
15 those already described, and (3) every simple syllogism depends on
three terms and two premisses, as Aristotle will prove.37

4.2 Why is the first figure first?38

Since in each figure there are some syllogistic and some
non-syllogistic combinations, he will discuss these and show how
many syllogisms there are in each figure and what they are. And
20 first, as I said, he discusses the first figure.
It is reasonable for that figure to be first in which the middle term
is middle not only in its relation to the extremes but also in order
and in position;39 for since the middle term explains the generation
of the figures, it is reasonable that it should also be authoritative in
the matter of their order. And so he places before the others the
figure in which the same term is actually middle in all respects.
25 Again, the syllogisms which come about in this figure are perfect,
35
Omitting tis am ousia empsukhos estin.
36
See Aristotle, An. Pr. 1.23 (esp. 41a2-20); Alexander, in An. Pr. 256,32-258,25;
Cony. 56-7; Albinus, Didasc. 158H; Galen, Inst. Log. xii 1; Ammonius, in An. Pr.
7,6-14. On the history of the fourth figure see Thouverez; on the logical and exegetical
problems connected with it see e.g. Patzig, pp. 109-27; Rescher; Ebert (1980); Thorn,
pp. 24-7.
37
See Are. Pr. 41b36-42b26; Alexander, in An. Pr. 271,16-272,10.
38
With this section compare Philoponus, in An. Pr. 65,4-66,26 (cf. 86,4-13);
Themistius, Max. 181-2; scholium to Aristotle, 151a46-b4. And see below, 94,5-17, on
the third figure.
39
'Order', taxis, sometimes refers to the role of a term as subject or predicate of a
proposition, and sometimes to the relative position of a word in a (standard) formula
or diagram; and exactly the same is true of'position' or thesis. We cannot tell which
term here refers to which feature.
 4. The First Figure 109
while those in the other figures are imperfect and are helped by this
figure.40 And the perfect is prior to the imperfect.
Moreover, the two figures owe their generation to the first.41 For
there are two premisses in the first figure, one on the major term and
the other on the minor. (The major term is the term predicated in the
problem - and also in the conclusion42 - being called major because it 30
has a wider extension than the term of which it is predicated.
Actually, this is not always so: sometimes it has an equal extension.
But that they are sometimes equal is a feature common to subject and
predicate alike, whereas when they are not equal, it is a proper
characteristic of the predicate that it is said to extend more widely 48,1
and of the subject that it is said to extend less widely -1 mean, in true
propositions; for it is never the other way about.43 Now each takes its
name from its proper characteristic, even when they are equal: the
predicate is called major even when it is equal, because it is its proper
characteristic sometimes to be major and never to be minor. The 5
subject is called minor, since this is its proper characteristic: it is
never major in extension.) Well then, there are two premisses in the
first figure. One of them contains the middle term connected to the
predicate and major term, and for this reason is itself called major.
The other contains the middle term connected to the subject and
minor term, and for this reason it too is called minor. Now it is the 10
conversion of each of these two premisses which generates each of the
other two figures.44 Hence for this reason too it is first in relation to
them: it actually generates the others.
In syllogisms the universal is more authoritative (for proving
something from a universal is the proper characteristic of syllogistic
justification45), and the major premiss in the first figure is universal if
40
cf. Aristotle, An. Post. 79a29-32; below, 53,8-10.
41
On the 'generation' of the figures see below, 71,12-21; 94,10-17; 95,14-24;
97,14-30; cf. e.g. Alexander, in An. Pr. 136,1-2; Boethius, Syll. Cat. 812D-813B;
Themistius, Max. 180-3. (Note that, according to Themistius, Max. 184, Aristotle,
Theophrastus and Eudemus did not discuss the generation of the figures: the topic
was introduced by 'the more recent Peripatetics'.) See Lee, pp. 120-3; Maconi, pp.
96-8; Lloyd (1990), pp. 21-3.
42
See below, 60,16-18; 75,10-19; cf. e.g. Philoponus, in An. Pr. 67,18-30.
43
What Alexander says holds only of true universal affirmatives: if it is true that A
holds of every B, then the extension of'A' is at least as great as that of'B'. This is not
so for any other type of proposition, and the standard explanation of 'major' and
'minor' in terms of relative extension will not in general work. See e.g. Patzig, pp.
97-100.
44
By 'conversion' here Alexander means conversion of terms (see above, 29,23-9):
'Man holds of every animal; animal holds of no stone' is a first figure combination.
Convert the first proposition and you get the second figure combination 'Animal holds
of every man; animal holds of no stone'. Convert the second proposition and you get
the third figure combination 'Man holds of every animal; stone holds of no animal'.
The conversions are not logical derivations: they are mere interchanges of terms.
45
See above, 44,6-7.
 110 Translation
15 the combination is syllogistic. Hence the conversion of the premiss
which is major and more authoritative will generate the more
authoritative and first of the remaining figures. But when this
premiss is converted, the middle comes to be predicated of both
extremes. It is therefore reasonable that the middle and second
figure should be the one in which the middle is predicated of both
extremes.46 The third and last figure comes about by the conversion
20 of the minor premiss in the first figure: when it is converted, the
middle comes to be subject for both extremes.
Again, the first figure is also first for the following reason.47 In
this figure all problems can be deduced universal affirmative and
universal negative and particular affirmative and negative. Hence
for this reason too it is rightfully first, being complete and perfect in
25 every way. In the remaining two figures not everything can be
deduced. Every proposition has two features, quality and quantity.
In quantity, the universal is more valuable, in quality the
affirmative.48 Each figure excels the other in one respect: the second
excels the third, inasmuch as universals can be deduced in it, and
30 they are superior in quantity (nothing affirmative is proved in this
figure); the third excels the second, inasmuch as affirmatives can
again be deduced in it, and they are more valuable in quality
49,1 (nothing universal is deduced in the third figure). Now it is
universality rather than affirmativeness which is the proper
characteristic of syllogisms; for the generation of syllogisms depends
on universals, and it is by the universal that they differ from other
sorts of justification. Reasonably enough, then, the figure which
excels in the proper characteristic of syllogisms, i.e. which possesses
this characteristic, is ranked before the figure which proves no
5 universal conclusion. (In the same way, the second indemonstrable
syllogism excels the third and is ranked before it.49)
Again,50 some syllogisms are demonstrative, some dialectical,
some sophistical.51 The demonstrative are the most valuable, the
dialectical hold the second rank, and the sophistical the third. Now
the first figure is more appropriate for demonstrative syllogisms.
10 For those who demonstrate in the strict sense conclude to universal
affirmatives (demonstrations come about by means of them), and
these are proved through the first figure alone.52 For this reason too,
then, it is rightfully first. The second figure is more appropriate for
dialectical syllogisms; for dialecticians, I suppose, always attempt to
46
See below, 71,11.
47
Later this is called the chief reason: below, 69,23-5 (cf. Apuleius, Int. 183,15-21).
48
cf. Aristotle, Are. Post. 85al3-86b39; Barnes (1975), pp. 175-80.
49
i.e. Celarent is ranked before Darii: below, 51,12-16.
50
On this paragraph see Patzig, pp. 154-5, who calls it 'logically grotesque'.
51
On the species of syllogisms see above, 7,8 note.
52
cf. Aristotle, Are. Post. 79al7-32.
 4. The First Figure 111
refute what has been posited by their partners and hence deduce
negatives - and to these the second figure is dedicated. Sophists
request and deduce indeterminate propositions. Inasmuch as the 15
indeterminate is equivalent to the particular,53 and all propositions
deduced in the third figure are particular, this will be the
appropriate figure for them.

4.3 First figure syllogisms in general

4.3.1 General rules54

So much for the quality of the figures and their order and the reason
for their order. In the first figure, the major term must be predicated 20
of the middle and the middle of the last and minor term: the minor
and last term - either the whole or a part of it - must by necessity be
in the middle term as in a whole,55 if there is to be a syllogism. If the
terms stand in this relation to one another, then the minor premiss
(i.e. the premiss on the minor extreme) will be either a universal
affirmative or a particular affirmative - but in any case an 25
affirmative. For if this premiss is not affirmative in the first figure, it
is impossible for there to be a syllogism. The major premiss, for its
part, is necessarily universal, if there is to be a syllogism in the first
figure. For since the minor is in the middle term as in a whole56 and
may be a part of it (even if it is sometimes taken as equal to it), and
the middle term may be major in extension compared to the minor, 30
then if the major extreme is predicated not of all the middle but of a
part of it, the minor term is not necessarily encompassed by it. For
the major may be predicated of some other part of the middle, if it is 50,1
particular, and not of that part which is either the whole minor term
or a part of it. Someone who has part of his face wounded is not
necessarily wounded in the eye, but someone who is wounded over
his whole face has his eye wounded, since the eye is in the face as in
a whole: similarly in the case of the combination before us - an 5
53
See above, 30,31 note.
54
Compare e.g. Apuleius, Int. 185,23-186,10; and esp. Philoponus, in An. Pr.
69,30-71,17, who makes a more elaborate and systematic attempt to formulate rules
for all syllogisms (cf. Lee, pp. 119-20).
65
'A is in B as in a whole' means 'B holds of every A' (above, 24,21-25,11). Hence
'The whole of A [or: a part of A] is in B as in a whole' makes no immediate sense.
Nonetheless, it is reasonably clear what Alexander means. "The whole of A is in B as
in a whole' simply means 'A is in B as in a whole', and hence is equivalent to 'B holds
of every A'. 'A part of A is in B as in a whole' means 'B holds of all of a part of A', i.e. 'B
holds of every D, where D is a part of A', i.e. 'B holds of every D, where A holds of
every D'. Hence (by the equivalence explained above, 32,20 note), 'A part of A is in B
as in a whole' is equivalent to 'B holds of some A'.
56
i.e., since either the whole of the minor or a part of the minor is in the middle as
in a whole ...: see above, 49,19-22.
 112 Translation
extreme which encompasses the whole of the middle also co-
encompasses all its parts, while an extreme which encompasses a
part of it does not necessarily thereby encompass the part which is
either the whole minor term or a part of it. So the major premiss will
by necessity be universal, either affirmative or negative; and in the
first figure the major premiss will have its quantity determined (it is
10 universal) and its quality indeterminate, while with the minor it will
be the other way about - its quality will be determined (it will always
be affirmative) and its quantity indeterminate (there will be a
syllogism both when it is a universal affirmative and when it is
particular).
Hence the major premiss in the first figure will be either a
universal affirmative or a universal negative, and the minor either a
15 universal affirmative or a particular. Since we thus have four
propositions - two of them universal (either affirmative or negative)
and two affirmative (either universal or particular) - there are four
conjunctions when they are conjoined. Hence there are four syllo-
gisms in the first figure: either the affirmative minor premiss will be
universal and be connected to a major which is itself also universal
20 and affirmative; or, remaining a universal affirmative, it will be
compounded with a major which is a universal negative; or it will be a
particular affirmative and be connected to a major which is in one
case a universal affirmative and in the other a universal negative.
In the first figure the conclusion will always have its quality from
the major premiss; for whatever quality the major premiss has -
25 whether it is affirmative or negative - the conclusion will be the
same.57 On the other hand, it will have its quantity from the minor
premiss; for if the minor is universal the conclusion too will be
universal, and if it is particular the conclusion too will be particular.
Thus the conclusion is determined by what is indeterminate in
each premiss - and this is itself evidence of the fact that syllogisms
30 depend on what has been posited, i.e. on what is conceded. For it is
what can also not be conceded which is conceded; whereas what is
51,1 necessary - i.e. what is such that if it is rejected no syllogism will
come about - is not posited and conceded. Thus that syllogisms
depend on what has been posited is clear from the fact that the
conclusion is similar to what has been conceded and not to what holds
by necessity. Moreover, if the conclusion were similar to what is
5 determined in the premisses, one and the same thing would always
be deduced (namely a universal affirmative, an affirmative because of
the minor and a universal because of the major); and it would not be
possible to prove anything else syllogistically in the first figure.

57
See Apuleius./rct 186,5-10.
 4. The First Figure 113
Further, in this way it turns out that the conclusion is similar to
the inferior assumption.58

4.3.2 Ranking59
When both premisses are universal and affirmative, the conclusion
too is such. And it is reasonable for this syllogism to hold the first 10
rank; for in its conclusion it possesses both what is superior in
quantity and what is superior in quality.60 When the major premiss
changes to a universal negative, we get the syllogism which holds
the second rank. For this syllogism again possesses what is superior
in the proper characteristic of syllogisms (but not both types of
superiority); for it is universal. If the minor premiss changes and 15
becomes particular, then since the major is a universal affirmative
and the minor a particular, the conclusion will be affirmative. This
syllogism is third in rank; for it excels the next syllogism by being
superior in quality: it is affirmative, whereas the other is negative.
When the minor premiss is a particular affirmative and the major is
a universal negative, the conclusion is a particular negative. This 20
syllogism is the last in the first figure, being excelled by its
predecessors - by one with respect to quantity, by one with respect
to quality, and by one with respect to both.

4.3.3 The number of combinations61

The syllogistic combinations in the first figure are this many, and
they are found in this way, if we take combinations of determinate
premisses. But if we also count in the combinations of indeterminate 25
premisses,62 two other syllogisms will be found in the first figure,
58
This is the earliest formulation of the peiorem-rule (peiorem semper conclusio
sequitur partem); cf. e.g. Philoponus, in An. Pr. 71,12-17; Thorn, pp. 183-5; Lee,
p. 119. (Note that Theophrastus and Eudemus introduced a correspondingpeiorem rule
in modal logic: Alexander, in An. Pr. 124,8-30; 126,29-127,2; 173,32-174,19;
Philoponus, in An. Pr. 123,12-17; 129,16-19; [Ammonius], in An. Pr. 38,38-39,2;
40,2-40 (texts collected in F 24-5 Graeser; F 11 Wehrli). See Bochenski (1947), pp.
78-87.)
59
cf. below, 76,29-77,31; 95,25-31; 96,22-97,11, on the ranking of syllogisms in the
second and third figures.
60
See above, 48,25-8.
61
See Philoponus, in An. Pr. 68,8-69,29.
62
Since indeterminates are equivalent to particulars (above, 30,31 note), it might
be thought that indeterminate combinations should not be added (see Apuleius, Int.
193,9-16). There is some support for this view in Aristotle (see esp. An. Pr. 26a28-30;
29a27-9), and it is the view which Alexander takes below, 111,30-112,2 (cf. 94,20).
But at 61,1-3 he says that particular and indeterminate premisses produce similar
conclusions (not the same conclusions); and in Philoponus it is clear that the
indeterminates provide new and distinct combinations of premisses (e.g. in An. Pr.
79,4-9).
 114 Translation
when the minor is assumed as an indeterminate affirmative and
connected either to a major which is a universal affirmative (the
conclusion of this combination will be an indeterminate affirmative)
or to a major which is assumed as a universal negative (the
30 conclusion will be an indeterminate negative). For it is impossible
for there to be a syllogism if the major is assumed as indeterminate.
And it seems that the conclusion is always similar to the inferior of
the premisses assumed, both in respect of quantity and in respect of
quality.63
The syllogistic combinations in the first figure are this many. The
combinations in thefirstfigure- and also in the other twofigures-
35 will be thirty-six in all, if you count in the indeterminates. In each
figure the one premiss can be compounded in six ways with the
52,1 other, which itself can similarly be taken in six ways. For if the
major is a universal affirmative, there will be six combinations, the
minor being taken either as a universal affirmative or as a universal
negative or as a particular affirmative or as a particular negative or
5 as an indeterminate, affirmative or negative. And if, again, the
major is a universal negative, there will again be six further
combinations, since the minor can be compounded with it in the six
ways we have just mentioned. If the major is taken as an
indeterminate affirmative, the minor will again be compounded
with it in six ways. Similarly and in the same number of ways, if the
10 major is taken as an indeterminate negative. In the same way, if the
major is taken as a particular, whether affirmative or negative.
The major changes in six ways and itself becomes sometimes a
universal affirmative, sometimes a universal negative, sometimes
an indeterminate affirmative or negative, and sometimes a
15 particular affirmative or negative. Hence inasmuch as in each
different case the minor can be compounded with it in six ways, the
conjunctions of premisses in this figure will be six times six in all.
(They are the same in the other two figures as well.) However, if the
indeterminates are set aside, the conjunctions will be four times
four in all, which make sixteen. If the four syllogistic combinations
are subtracted, the remaining non-syllogistic combinations are
20 twelve. Combinations are called syllogistic and reliable if they do
not alter together with differences in the matter - i.e. if they do not
deduce and prove different things at different times, but always and
in every material instance preserve one and the same form in the
conclusion.64 Combinations which change and alter configuration
together with the matter and acquire different and conflicting
conclusions at different times, are non-syllogistic and unreliable -
63
See above, 51,7-8.
64
For the error illustrated here see Introduction, pp. 12-13.
 4. The First Figure 115
just as a man is unreliable if he is not stable and firm in his
judgment.65

4.3.4 Aristotle's general comments

He discusses the first figure first and indicates what it is by way of 25
examples, saying:

Now when three terms so stand to one another that the last is
in the middle as in a whole and the middle is or is not in the
first as in a whole, <it is necessary for there to be a perfect
syllogism of the extremes.' I call a term middle when both it is
in another and another is in it - it is middle by position too.>
[1.4,25b32-6]
These are examples, one the first and the other the second
syllogism. He next gives the account of the first figure, which he is 53,1
discussing, in his remarks about the middle term: the first figure is
that in which the middle term 'both ... is in another and another is
in it'. For in this figure the middle is predicated of one of the terms
and is subject for the other. He does not give a general account of the 5
middle term but one which holds for the first figure. In general, the
middle term in a combination is the one which is taken twice, which
occurs in both premisses, and which the premisses in the
combination share with each other.66
The expression he uses also makes clear the reason why this is the
first figure. (He says: 'it is necessary for there to be a perfect 10
syllogism of the extremes.') For it is reasonable that the figure with
the perfect syllogisms should be the first figure.67
He does not mean that combinations in which the middle term
'both ... is in another and another is in it' and in which the middle
holds the middle position, are always syllogistic or that a perfect
syllogism always comes about from them. For in that case all the
combinations in this figure would be syllogistic. Rather, he means 15
that this is so when the terms stand in the way he has just
described. He sets down this way by saying: 'so ... that the last is in
the middle as in a whole and the middle is or is not in the first.' For
when the premisses stand in this way, there is then a perfect

65
The terms 'reliable' and 'unreliable' are not used in this logical sense by Aristotle,
and Alexander's explanation perhaps implies that they are new (or at least
unfamiliar). Galen knows them: Inst. Log. xi 6.
66
On the middle term see above, 44,29-45,10; 46,21-32.
67
See above, 47,24-7.
 116 Translation
syllogism of the extremes in relation to one another. (Here he calls the
conclusion a syllogism.68)
20 He begins his exposition of the syllogisms now with 'as in a whole'
and now with 'of every', since each of these items is a sort of principle
and is primary.69 'As in a whole' is primary and a principle in relation
to us. For things which are 'as in a whole' and subjects are better
known to us than things said with a wider extension; for they are
closer to perceptible things. On the other hand, 'of every' is primary
25 by nature; for it is more common and more general, and common
items are primary by nature and thus are also principles.70

4.4 Combinations of universal premisses

4.4.1 Barbara and Celarent

If A is said of every B and B of every C, <then it is necessary for
A to be predicated of every C. For we have earlier said what we
mean by 'of every'.> [1.4,25b37-40]
He uses letters in his exposition in order to indicate to us that the
30 conclusions do not depend on the matter but on the figure, on the
conjunction of the premisses, and on the modes.71 For so-and-so is
deduced syllogistically not because the matter is of such-and-such a
54,1 kind but because the combination is so-and-so. The letters, then,
show that the conclusion will be such-and-such universally, always,
and for every assumption.
He has told us that for one thing to be in another as in a whole and
for the other to be said of all the one are the same.72 So, having used
5 'as in a whole' in setting out the figure, he now uses 'of every'
instead. Both of these formulae occur in universal affirmative
propositions. He draws our attention to the account he gave of 'of
every' ('when it is not possible to take any of the subject of which the
predicate is not said') in order to indicate that in these deductions
there is no need for anything external to make the necessity evident,
10 but that the suppositions are sufficient. For 'of every', which is

68
For the word sullogismos used in the sense of 'conclusion' see e.g. Galen, Inst.
Log. i 5; cf. Patzig, pp. 95-6.
69
See above, 25,9-11 note.
70
On the primacy of universals in Alexander see Tweedale.
71
For other comments on Aristotle's use of letters see below 77,32-78,5; 98,20-3;
Alexander, in An. Pr. 125,26-8; 379,14-380,27; 414,9-10; 415,10-12; in Top. 2,16-29;
Conv. 61; Philoponus, in An. Pr. 46,25-47,11; Boethius, Syll. Cat. 810CD. Cf.
Bochenski (1956), p. 157; Lukasiewicz, pp. 7-10 (but the letters should not be
construed as variables: Frede (1974b), p. 113).
72
See An. Pr. 24b26-8; above, 24,26-25,11.
 4. The First Figure 111
supposed and taken in the premisses, is sufficient for the proof of the
deduction.73
This74 is why syllogisms of this kind are perfect and indemon-
strable in the strict sense.75 Let A be the major extreme, B the
middle term, and C the minor extreme. If C is in B as in a whole,
B is said of every C. For these formulae convert from each other.
Therefore it is not possible to take any of C of which B is not said. 15
Again, if B is in A as in a whole, A is said of every B. Hence it is not
possible to take76 any of B of which A is not said. Now, if nothing of B
can be taken of which A is not said, and C is something of B, then by
necessity A will be said of C too. The deduction in syllogisms of this
sort is thus directly obvious, being justified by means of the 20
suppositions - by means of'of every' and 'of none' - and not needing
anything else from outside.
He77 uses 'of every1 and 'of none' in his exposition because by
means of these formulae the deduction becomes known, and because
when they are stated in this way the subject and the predicate are
better known, and because 'of every' is primary by nature in relation
to 'in it as in a whole', as I have already said.78 (But syllogistic usage 25
is normally the other way about. Virtue is not said of every justice,
but the other way about - 'All justice is virtue'. This is why we have
to train ourselves in both types of formulation, so that we can follow
both normal usage and Aristotle's exposition.79)

Similarly, if A is said of no B <and B of every C, A will hold of 30

no C.> [1.4, 25b40-26a2]
Having set out the first syllogism in the first figure, which deduces a
universal affirmative from two universal affirmatives, he next
mentions the second, which deduces a universal negative from a
universal negative major and a universal affirmative minor. He says 55,1
that in this case too the deduction is known by way of 'of none' and
'of every'.80 This81 is why this syllogism, too, is both indemonstrable
73
See above, 32,11 note.
74
54,11-12 = FDS 1101.
75 Why 'in the strict sense [kuriosY? Alexander does find a sort of proof for these
syllogisms, based on the dictum de omni et nullo; presumably he holds that this proof
is not a demonstration in the strict sense - perhaps because it cannot be expressed
within categorical syllogistic.
76
Reading labein, with M and the Aldine. Wallies omits the verb.
77
On this paragraph see Lukasiewicz, p. 17; Mignucci (1969), pp. 185-6; Thorn,
p. 20; and esp. Patzig, pp. 8-12 (but the construal which he canvasses at p. 15 n. 24 will
hardly fit the text) and Ebert (1977).
78
See above, 53,24-6.
79
See Introduction, pp. 30-1.
80
See above, 32,11 note.
81
55,3 =FDS 1101.
 118 Translation
and perfect. For if B is said of every C, there will be nothing of C of
5 which B is not said. Therefore C is something of B. Now if A is said of
no B, there will be nothing of B of which A is predicated. But C was
something of B. Therefore A is said of no C.

4.4.2 Non-syllogistic combinations82

But if the first follows all the middle and the middle none of the
last, there will not be a syllogism <of the extremes. For
nothing necessary follows inasmuch as they are the case. For it
is possible for the first to hold of all and also of none of the last.
Hence neither the particular nor the universal is necessary.
And if nothing is necessary by way of these items, there will
not be a syllogism. Terms for holding of every: animal, man,
horse. For holding of none: animal, man, stone.> [1.4,26a2-9]
10 He calls being predicated 'following'; for the predicate follows what
falls under it.83
Having discussed the syllogistic84 combinations of the first figure
in which both premisses are universal, he sets down the
non-syllogistic combinations in which the premisses are universal.
There are four combinations in the first figure, given that both
15 premisses are universal. Two of them are syllogistic and reliable
(these he has set out); two are non-syllogistic - the one with a
negative minor and an affirmative major (which he mentions here)
and the one in which both premisses are negative. The reason why
the combination set out here is non-syllogistic is that the minor has
been assumed as negative; for we have already said that it is
20 impossible for there to be a syllogism in the first figure where the
minor proposition is negative.85
He sets down material instances to prove that, when the
premisses stand thus, nothing necessary can be deduced (which is
the proper characteristic of a syllogism86). For he will prove that in
some material instances it is possible for a universal affirmative to
be deduced, and in other material instances a universal negative87 -
25 and this is the most obvious sign that this combination has no
syllogistic force, since contraries and opposites are proved in it and
they cancel one another. For nothing can be deduced syllogistically
82
On Aristotle's ways of showing that a combination is non-syllogistic see Patzig,
pp. 168-92; Lear, pp. 54-75.
83
cf. Alexander, in An. Pr. 326,31-2; Philoponus, in An. Pr. 302,17-19; and note
Aristotle's use ofhepesthai at e.g. An. Pr. 43b3; 44al3. See Barnes (1983), pp. 309-13.
84
Reading sullogistikon for sullogismon ton.
85
See above, 49,22-50,7.
86
See above, 18,8-12.
87
For the confusion exhibited here see Introduction, pp. 12-13.
 4. The First Figure 119
in a combination in which universal affirmatives and universal
negatives can be proved. What is deduced syllogistically must be the
same in every material instance and must be either a universal
affirmative, or a universal negative, or a particular affirmative or 30
negative; but both affirmatives are cancelled by the universal
negative and both negatives by the universal affirmative. This is
what he means when he says 'Hence neither the particular nor the
universal is necessary'. 56,1
He does not mean that if you prove a contradictory pair to be
deducible in some combination then you cancel the possibility that
something may be syllogistically deduced in it. For if a universal
affirmative and a particular negative are deduced, there is nothing
to prevent a particular affirmative from being deduced syllogisti-
cally since this is cancelled by neither of the suppositions. And if the 5
other contradictory pair is proved, it is still possible for a particular
negative to be deduced syllogistically. We will say more about this
shortly;88 for the moment let us add the reason why the combination
set out here is non-syllogistic.
If C is in no B, B is said of no C. Therefore, it will not be possible to
take anything of C of which B is said. Thus C has been separated 10
from B and is in no way linked to it. A is supposed to be said of every
B. But what is said of all of something may actually have a wider
extension than it (it is for this reason that the predicate term is
called 'major').89 You may take A in such a way as to encompass all
of B and yet have some parts which fall outside the compass of B. 15
Thus A may either encompass or not encompass C with those parts
of itself which fall outside B. So, depending on the difference of
matter, opposites are true when premisses are conjoined in this
way, and such conjunctions are non-syllogistic. Let A be animal, B
man, and C horse: animal holds of every man, man of no horse, and 20
animal of every horse. It is by the part which falls outside man that
animal includes horse - the sharing of the middle term is not the
explanation. If stone is posited instead of horse, the premisses are
true in the same way: animal is said of every man, man of no
stone.90 Therefore the combination is unreliable; for it does not itself 25
impose a form on the matter but changes together with the
differences in the matter and deduces nothing necessary from the
suppositions.91
Having said 'for it is possible for the first to hold of all and also of
88
See below, 89,29-91,33.
89
See above, 47,29-48,6.
90
Omitting kai to zoion kat' oudenos lithou, which Wallies adds to the text on the
authority of the Aldine.
91
With this argument compare below 61,24-9; 62,16-21; 64,15-22; 67,17-22;
81,28-82,1; 85,27-9.
 120 Translation
30 none of the last', he continues: 'Hence neither the particular nor the
universal is necessary.' For, as we have said, it is not only the
universal affirmative which is cancelled by the universal negative
57,1 but also the particular affirmative (since they are contradictories);
and again, it is not only the universal negative which is cancelled by
the universal affirmative but also the particular negative. Hence in
proving that contraries are deduced, he has proved that no particular
is deduced with syllogistic necessity.

5 Nor when the first holds of none of the middle and the middle of
none of the last will there be a syllogism in this case either.
<Terms for holding: science, line, medicine. For not holding:
science, line, unit.> [1.4, 26a9-13]
He has moved on to the combination of two negative universals in
the first figure, and he proves that this too is non-syllogistic. Again,
he sets down terms such that, in the different cases, A evidently
10 holds of every C and also of no C. The reason why nothing is deduced
syllogistically in this combination is that the middle bears no
relation to either of the extremes (it is as if a middle had not been
taken at all - and syllogisms depend on the middle term); and when
it stands thus, C may fall under A and also not fall under A. If A is
15 science, B line, and C medicine, both universal negative premisses
are true and science will hold of all medicine. But if unit or stone is
posited instead of medicine, the premisses are true in the same way,
but science will hold of no unit and of no stone.

4.4.3 Concluding remarks

20 Thus when the terms are universal it is clear ... in this figure
...[1.4,26al3]
We have said that there are four combinations in the first figure
when the premisses are taken universally, and that two of them are
syllogistic and two non-syllogistic.

... and that if there is a syllogism, it is necessary for the terms

25 to stand as we have said, <and that if they stand in this way
there will be a syllogism.> [1.4,26al4-16]
He converts his statement: if there is a syllogism in the first figure
with universal premisses, then the terms must have the position
and the mutual order which he has described; and if the terms stand
thus and in this position, then there must be a syllogism.
58,1 Now if there is to be a universal affirmative conclusion, then it is
clear that it must come about in the first figure from premisses of
 4. The First Figure 121
this sort supposed in the way we have shown - for no universal
affirmative is deduced in any other figure. But how is this still true in
the case of universal negative conclusions? For it will seem not to be
necessary that if a universal negative conclusion has come about 5
syllogistically, then the terms stand as we have described them in the
second combination of the first figure. For the same conclusion can be
proved syllogistically when the terms stand otherwise: universal
negatives are proved in the second figure, and in two ways. Hence it
will seem untrue that 'if there is a syllogism, it is necessary for the
terms to stand as we have said'. 10
Thus if what Aristotle says is to be true, we must understand 'in
this figure'. He said earlier 'when the terms are universal, it is clear
... in this figure .. ,'.92 Hence what he goes on to say also refers to the
syllogisms in this figure.
One may also say that what he has said does actually fit the
syllogisms which deduce a universal negative in the second figure.
For there too, even if the order in the premisses is not the same - the 15
terms stand to one another in a different relation and not as they do in
the first figure -, nonetheless there too they stand in this relation
potentially. At any rate, it is in virtue of an analysis into the first
figure by way of conversions of their premisses that these combin-
ations are proved to be syllogistic. Hence here too the terms do stand 20
to one another in the way he described in the case of the second
combination of the first figure. For the syllogisms which deduce a
universal negative in the second figure are reduced to this combin-
ation, as he will later prove.93

4.5 Combinations of universal and particular premisses

4.5.1 Darii andFerio
If one of the terms is universal and one particular in relation to
the other... [1.4, 26al7]
'In relation to the other' is not, in my view, idly added to 'and one 25
particular'. For it is possible that in one and the same premiss one of
the terms is universal and the other particular - the subject, if
taken as a whole, being universal, and the predicate, part of which is
said of the subject, being particular. Not every animal is said of
man: a part of animal is said of every man in the proposition which
says 'Every man is an animal' - some animal is said of man. Again, 59,1
'Animal is said of some animate' has one of its terms universal and

92
An. Pr. 26al3-14 (see also below, 60,9).
93
See An. Pr. 27a5-14; cf. below, 78,21-2; 79,7-8.
 122 Translation
94
the other particular. (That is why the converses of such propositions
are particular.) Thus it was in order to make it clear that the
universality and particularity of the terms occur in the conjunction of
5 the premisses and not in a single premiss nor yet in a converse of a
premiss, that he added 'in relation to the other': i.e. 'when one term is
predicated universally, and this is predicated particularly of some-
thing else'; or, more simply, 'when the same term is universal in
relation to one term and particular in relation to the other in the
conjunction described'. For it is when it is taken particularly in this
10 way that we get a combination or conjunction of premisses.
Having discussed combinations in the first figure with two uni-
versals, he turns next to combinations in which one premiss is
particular and the other universal, showing which and how many
they are and which of them are syllogistic. First, he sets them out. Of
15 these combinations, two again are reliable, viz. the combination of a
universal affirmative major premiss and a particular affirmative
minor, and the combination of a particular affirmative minor and a
universal negative major. He mentions both of them together when
he says:

... when the universal is posited on the major extreme (whether

it is predicative or privative) and the particular on the minor is
20 predicative, it is necessary for there to be a perfect syllogism.
[1.4, 26al8-20]
When the major is universal and affirmative, what is deduced will
be a particular affirmative; and if the major is assumed as a
universal negative, a particular negative will be deduced. Inasmuch
as one premiss must be assumed as particular, and the major must
certainly be universal, the minor will be particular; and he has said
25 that it will also be predicative.

94
A puzzling passage; for in what sense can we talk about the terms in a
proposition being universal or particular? It might seem that Alexander is supposing
that the proposition 'Every man is an animal' predicates some animal of every man, so
that man is universal and animal particular. And it might then further seem that
Alexander is here admitting - or even insisting - that the predicate as well as the
subject bears a quantifier. But quantification of the predicate was rejected by
Aristotle (Int. 17bl2-16; An. Pr. 43bl7-22; An. Post. 77b30), and the commentators
followed him: e.g. Alexander, in An. Pr. 297,4-23 (cf. above, 44,28 note); Philoponus,
in An. Pr. 277,26-32; and esp. the long essay in Ammonius, in Int. 101,14-108,36, on
which see Mignucci (1983). Rather, Alexander has the following simple fact in mind:
in some cases where A holds of every B, only some As are Bs, and in other cases all As
are Bs. In the former cases A is 'particular', in the latter 'universal'. (And this has
nothing to do with quantifying the predicate.)
 4. The First Figure 123
But when it is on the minor or the terms stand in any other
way, it is impossible. [1.4, 26a20-l]
Having referred to the major premiss and said that there will be a
syllogism when it is privative, he continues 'But when it (sc. the
privative) is (sc. posited instead) on the minor' so that it is a 30
particular negative and the major is universal, either affirmative or
negative. 'Stand in any other way' means 'If the major is instead 60,1
assumed as a particular and the minor as a universal'. For when
they stand so, there is no syllogism.
'But when it is on the minor' may also have the force of 'when the
universal is instead assumed on the minor' - and to judge by what
comes next, this is indeed rather what he is saying.95 Then 'or the
terms stand in any other way' will be said in reference to the minor 5
premiss 'if this is not assumed instead as a universal but remains
particular and is assumed as a negative'.

I call the major extreme the one in which the middle is, and the
minor the one under the middle. [1.4, 26a21-3]
Everything he says must be understood to be said of the first figure.
For what he says here is this: 'In the first figure, the major term is 10
the one which the middle is in and the minor the one under the
middle.' For in this figure, as we saw, the middle term is subject for
one term and predicate of the other. In the second figure both the
major and the minor are in the middle term - the middle is under no
term. Conversely, in the third figure the middle is not only under the 15
major but also under the minor - no term is in the middle term.96
He has shown plainly that the term predicated in the conclusion is
the major term in the premisses and that the major premiss is major
because of it.97

For let A hold of every B and B of some C. <Then if being 20

predicated of every is what we said at the beginning, it is
necessary for A to hold of some C.> [1.4, 26a23-5]
This is the combination of the third indemonstrable, which has a
particular affirmative conclusion. If B holds of some C, then
something of C is in B as in a whole.98 But A is said of every B.
95
i.e. An. Pr. 26a30-l; below, 63,11-12 (cf. Philoponus, in An. Pr. 77,22-9).
96
See above, 46,21-47,12. On the unannounced use of 'under [hupo]' here see
Patzig, pp. 98-9.
97
See above, 47,30-48,10.
98
For the locution 'some C is in B as in a whole' see above, 49,22 note. Note that
Alexander again tacitly invokes the thesis that if A holds of some B, then there is
some D such that A holds of every D and B holds of every D (above, 32,20 note).
 124 Translation
Therefore there is nothing of B of which A is not said. But C is
something of B. Therefore A will be said of it.

<And if A holds of no B and B of some C, it is necessary for A

not to hold of some C. For we have also denned what we mean
by 'of none'. Hence there will be a perfect syllogism.> [1.4,
26a25-8]
25 Similarly with the fourth indemonstrable, which has a universal
negative major premiss and a particular affirmative minor. For if
something of C is in B as in a whole, and B is in no A, then A will not
hold of some C. For something of C is under B; but nothing of B can
be taken of which A is said. Hence A will not be said of that item of C
61,1 which is something of B.

<Similarly too if BC is indeterminate ...> [1.4,26a28-9]

Deductions similar to these" are found if we make the minor
premiss an indeterminate affirmative and keep the major either as a
universal affirmative or as a universal negative.
All the syllogisms thus far described are perfect; for in all of them
5 the deduction is evident, invoking only 'of every' or 'of none', which
are supposed.100

4.5.2 Non-syllogistic combinations

4.5.2.1 Major particular, minor universal
If the universal, either predicative or privative, is posited on
the minor extreme, there will not be a syllogism, <whether the
indeterminate or particular is affirmative or negative>. [1.4,
26a30-3]
He is discussing combinations which have one premiss universal
and the other particular. He first discussed those in which the
10 universal is on the major premiss and the particular, which is
affirmative, on the minor, proving that such combinations are
syllogistic. He then added the combinations in which the minor
premiss is indeterminate instead of particular, everything else
remaining the same, and found two syllogisms when the minor is
taken in this way. Now he turns to the complementary cases - I
15 mean the combinations which have a universal minor premiss.
When this is universal, the major premiss is either particular or
99
See above, 51,25 note.
100
See above, 32,11 note.
 4. The First Figure 125
indeterminate, and if it is particular or indeterminate, then - as we
have already said101 - all the conjunctions are non-syllogistic. Thus
he sets out these conjunctions or combinations and proves that they
are non-syllogistic. Once again, he sets down material instances in
which contraries are found to be deduced. 20
He said 'the particular' instead of 'the major'102 - for the major
premiss is particular.

I.e. if A holds or does not hold of some B <and B holds of every

C. Terms for holding: good, disposition, sagacity. For not
holding: good, disposition, ignorance.> [1.4, 26a33-6]
If the major premiss is particular, whether affirmative or negative,
the conjunction is non-syllogistic. For if A is said of some B, and B of 25
every C, there will be nothing of C of which B is not said; hence C is
something of B. But since A is said of some B, it may be said of that
part of B which is C, and it may also be said of some other part; for B
is predicated of C with a wider extension.103 Therefore A will neither
hold nor not hold of C by necessity. For example, suppose that A is
good, B disposition, and C sagacity: good is said of some disposition, 62,1
disposition of all sagacity, good of all sagacity. But if C is ignorance
(here he takes ignorance as a disposition,104 i.e.105 as folly or vice),
good, the major extreme, will hold of no ignorance and of no folly,
although disposition, which was the middle term, is universally 5
predicated of them. Hence the combination is unreliable.
If A does not hold of some B, the proof is similar and uses the same
terms. For as A holds of some B, so it may also not hold of some -
good will hold of some disposition in this way, because it does not
hold of some.

Again, if B holds of no C and A holds of some B or does not hold 10

(or does not hold of every), <in this case too there will not be a
syllogism. Terms: white, horse, swan; white, horse, raven. The
same terms serve if AB is indeterminate. > [1.4, 26a36-9]
'Does not hold of some' and 'does not hold of every' both mean
particular negatives, differing only in expression.106 (To 'does not
hold' you must add 'of some', which goes with both verbs.)
101
See above, 49,26-50,8.
102
Retaining Wallies' tou for the MS reading to, pace Ross, p. 303. The MS reading
is ungrammatical.
103
See above, 56,8-27.
104
For this sense of'ignorance' see Aristotle, An. Post. 79b23-5; cf. Barnes (1975),
p. 145.
105
Reading kai for he.
106
This was affirmed by Aristotle (An. Pr. 62a9-10) and denied by Theophrastus,
 126 Translation
He proves that conjunctions which have a universal negative minor
15 and a particular major are not deductive.107 By following the same
path as before,108 we will find that A sometimes holds of every C and
sometimes holds of none. Let A be white, B horse, and C swan: white
holds of some horse and also does not hold of some horse; horse of no
swan; white of every swan. For white is of wider extension than horse
20 and is predicated of it; and in that part by which it exceeds horse it
now encompasses C, which is swan, of none of which horse holds. And
if we make C raven instead of swan, horse will hold of no raven, and
nor will white.
If the major premiss is taken as indeterminate, whether affir-
mative or negative, the proof will be given by the same terms - for
indeterminates are taken as equivalent to particulars.109
25 Note that when the minor is a universal affirmative and the major
is particular, the combination is non-syllogistic because of the major's
being particular; whereas when the major is particular or indeter-
minate and the minor is a universal negative, the combination is
non-syllogistic on account of both premisses - on account of the major,
because it is particular, and on account of the minor, because it is
30 negative.

4.5.2.2 Minor particular and negative

63,1 When the term on the major extreme is universal, either
predicative or privative, and the term on the minor is privative
and particular, there will not be a syllogism (whether it is
assumed as indeterminate or as particular). I.e. if A holds of
5 every B, and B does not hold of some C (or if it does not hold of
every C). <For the first will follow all and also none of that of
some of which the middle does not hold. Suppose the terms to be
animal, man, white. Then of the whites of which man is not
predicated, take swan and snow. Then animal is predicated of
all the one and of none of the other. Hence there will not be a
syllogism.> [1.4, 26a39-blO]
'Does not hold of some' and 'does not hold of every' are again the
same. no

who suggested that 'not of every' implies that the predicate holds of several of the
subject and that 'not of some' implies that it holds of one (scholium to Aristotle,
145a30-7 = F 5 Graeser - but the sense of the passage is uncertain); cf. Bochenski
(1947), p. 43.
107
'Deductive' translates sunaktikos (from sunagein, 'to deduce'): it is here used as
a synonym for 'syllogistic'.
108
See above, 61,24-9.
109
See above, 30,31 note.
110
See above, 62,11-12.
 4. The First Figure 127
Having first proved non-syllogistic the combinations in which the
major is particular, he now proves non-syllogistic those in which the
major remains universal, whether affirmative or negative, and the 10
minor is a particular negative. (He thereby shows what he meant
when he said above: 'But when it is on the minor or the terms stand
in any other way.'111) These combinations are non-syllogistic
because of the minor, which is negative.
His proof112 that these combinations are non-syllogistic uses the
part of C of which it is supposed that B does not hold. For taking 15
this, he proves, again by setting down terms, that A holds of all and
also of none of it. He intended in this way to give an appropriate
proof of the fact that the combination set out is non-syllogistic; for if
the proof used every C, it would no longer be suitable in the same
way. When B held of no C, the combination was proved to be
non-syllogistic inasmuch as A could hold both of all and also of none 20
of that of none of which B held:113 in the same way, when B is
supposed not to hold of some C, the combination would be proved
non-syllogistic, if it were proved that A could hold of all and also of
none of that part of C of which B does not hold.114
Again, since the minor is a particular negative, the same premiss
can also be a particular affirmative in the same respect;115 for 25
nothing prevents that which does not hold of some so-and-so from
holding of some of it as well. But when the minor is a particular
affirmative, there is a syllogism if the major is universal, whether
affirmative or negative - a syllogism which is either a particular
affirmative or a particular negative, as has been proved116 (for each
of the premisses possesses its proper characteristic, the minor being 30
affirmative and the major universal). And since this combination is
syllogistic, it was not possible for him to take terms both for which A 64,1
holds of every C and also for which it holds of no C. Hence, since it is
supposed that B does not hold of some C, then if it also holds of some
C, it will not be possible to take terms for which A holds of every C
and also for which it holds of no C. For in that case we should also
disprove and reject the syllogistic combinations in which B holds of
111
An. Pr. 26a20-l; see above, 59,28-60,6.
112
On this see Maier, Ha, p. 87; Ross, p. 304; Patzig, pp. 177-80.
113
See Are. Pr. 26a2-9; above, 55,20-56,1.
114
Alexander's thought seems to be this: since 'B does not hold of some C' only
concerns some C (some white things), a disproof of the combination which invoked all
C would be too general and hence not appropriate. This is a confused thought; and it
is intended to justify a procedure which is unjustifiable. (Aristotle's argument fails
because, in Alexandrian terms, he does not take that part of C of which B does not
hold: he takes two different parts of C.)
115
Alexander must mean: 'The particular affirmative, "B holds of some C", may be
true when the particular negative premiss is true.'
116
i.e. Darii and Ferio: Aristotle, An. Pr. 26al7-29. (Note that here 'syllogism' is
used in the sense of'conclusion': see above, 53,19 note.)
 128 Translation
5 some C and A of every B, and in which B holds of some C and A of no
B. For in the case of the combination which deduces an affirmative,
it is impossible to take terms for which A holds of no C, since it holds
syllogistically of some C; and in the case of the combination which
syllogizes a particular negative, it is impossible to take terms for
which A holds of every C, since it is syllogistically proved that A does
10 not hold of some C. Each of these combinations would be rejected if,
in the case of the combination before us, terms were found for A's
holding of every C and also for A's holding of no C.117
For this reason, since it is not possible, he does not give the proof
in terms of C simpliciter, but rather takes from C those parts of
which B does not hold (for it is supposed that B does not hold of some
C); and by setting down material instances he proves that there are
15 parts of C (of which B does not hold) of all of which A holds, and
parts of none of which A holds.
It is also possible to prove that the conjunction we have set out118
is non-syllogistic by predicating B universally and negatively of the
parts of C, assuming the minor as a universal negative. For suppose
that A is said of every B and B of none of the parts of C which have
20 been taken: since A may have a wider extension than B (for it is
possible that A is such as to exceed B), it may include the parts of C
of which B does not hold, and also not include them.119 Such a
combination is unreliable.120
The terms which Aristotle sets down are animal for A, man for B,
and white for the whole C, of some of which man holds and of some
25 of which man does not hold. Next, since BC is a particular negative
and it is impossible to prove and deduce a universal negative of C
when the terms are so taken (or a universal affirmative if AB is
taken as a universal negative), let us take some parts of white, of
none of which man is said - let them be swan and snow. (Man is said
117
This is false: the valid combinations would be rejected only if we could find
terms to verify 'A belongs to every B and B does not belong to some C and B belongs to
some C" as well as 'A belongs to every C' or 'A belongs to no C'.
118
Reading ekkeimenen for eirSmenen (Wallies, from the Aldine: the MSS have
antikeimenen).
119
See above, 56,8-27.
120
Alexander's procedure in this paragraph is correct, but his exposition is
compressed. The argument depends on the fact that 'B does not hold of some C' is true
if and only if there is some D such that C holds of every D and B holds of no D. (See
above, 32,20 note, for the corresponding fact about 'B holds of some C'.) Hence if a
syllogism of the form 'P, B does not hold of some C: therefore, Q' is valid, then so too is
the corresponding syllogism of the form 'P, B holds of no D: therefore, Q'. Hence a
counterexample to the latter will refute the former. According to Alexander, this
train of reasoning, although probative, was not Aristotle's (see below, 65,16-32:
Alexander implies that some critics had both ascribed it to Aristotle and found it
wanting). Later Philoponus expresses the line of thought more clearly - and reads it
into Aristotle (in An. Pr. 82,21-7; cf. 82,34-83,4; and note metalambanein at
109,20-1).
 4. The First Figure 129
neither of swan, which is white, nor of snow.) Now animal is said of
every swan but of no snow. In this way such a combination is proved 65,1
unreliable and non-syllogistic in terms of a part of C to which B does
not belong.
(As for conjunctions with two particulars,121 when he sets down
the terms by which he proves that such conjunctions are
non-syllogistic, he no longer argues in terms of a part of C of some of
which it is supposed that B does not hold, but rather in terms of the 5
whole of C; for a conjunction with two particulars is non-syllogistic,
both if B does not hold of some C while also holding of some C and
also if B holds of no C. Now inasmuch as there is no syllogistic
combination to set alongside such a conjunction of premisses, it is
possible to take the whole of C and obtain terms both for A's holding
of every C and also for its holding of none. And if something holds of
the whole of C, clearly it also holds of each of its parts - and hence of 10
the part of which B does not hold.)

<Again, let A hold of no B and let B not hold of some C; and let
the terms be inanimate, man, white. Then of the whites of
which man is not predicated, take swan and snow. Inanimate
is predicated of all the one and of none of the other.> [1.4,
26blO-14]
The refutation is similar if AB is taken as a universal negative. Let
the terms be inanimate for A, man for B, white for C; and let us
again take swan and snow as whites of which man is not predicated.
For inanimate holds of all snow and of no swan. (The same terms,
swan and snow, were used to prove that the combination is 15
non-syllogistic when the major is a universal affirmative.)
He does not, as some think, invoke a universal negative in setting
out the terms, transforming the particular negative of the minor
proposition into a universal negative. (Nor, if someone does invoke
it, should one therefore deem him not to prove the given 20
combination to be non-syllogistic.) For you invoke what is universal
if you prove 'of every' and 'of none' for all C rather than for some part
ofC.122
Again, if it were not possible to take any parts of the last term of
which the middle did not hold universally, then the objectors would
be right in objecting to the transformation. This would be the case if
C were indivisible and did not have parts. But if it is indivisible, it 25
cannot be true that the middle does not hold of some of it. For if the
determinations - namely 'of every' and 'of none' and 'of some' and
121
Aristotle, An. Pr. 26b21-5; below, 68,9-21.
122
See above, 64,15-24 and note.
 130 Translation
123
'not of some' - are annexed to the universal, as has been shown in
On Interpretation,12* it is clear that the last term is universal and
not indivisible;125 and sometimes there will be not only some thing
but some things of which it is predicated, if the middle has been
30 taken not to hold of some of it - as in the examples Aristotle sets
down. If it is of this sort, then both the exposition126 and the
transformation into the universal are correct, and the proof by way
of the universal that the given combination is non-syllogistic is
sound.

4.5.2.3 The indeterminacy of particulars127

<Again, since 'B does not hold of some C' is indeterminate and
it is true both if it holds of none and if it does not hold of all
(because it does not hold of some), and since there is no
syllogism if terms are taken such that it holds of none (this has
been said earlier), it is evident that there will not be a
syllogism inasmuch as the terms stand in the way we are
considering. For then there would be one in this case too.>
[1.4, 26bl4-20]
66,1He also proves in the following way that the two combinations set
out here are non-syllogistic. 'Again, since "does not hold of some" is
said indeterminately ....' For 'does not hold of some' is true both if it
holds of none and also if it holds of some. Here he calls propositions
indeterminate if they do not have their truth determined. Thus it is
5 true to say that hot does not hold of some snow - and hot holds of no
snow. In the same way it is true that man does not hold of some
horse, because man holds of no horse. Theophrastus, too, in his On
Affirmation, mentions this sort of indeterminacy. He calls
indeterminate 'Some of these are ...' and 'One of them is ...': 'Some of
these are ...' because it is true both if all are and also if some are and
123
See above, 44,27 note.
124
See Aristotle, Int. 17a37-bl2; cf. e.g. below, 100,11-14; Boethius, Int. Syll. Cat.
778B.
125
Alexander argues that since the quantifiers are annexed to the term C, then C
must be a general and not a singular term. The argument thus seems to turn on the
vexed question of the place of singular terms in Aristotelian syllogisms (on which see
Lukasiewicz, pp. 4-7; Patzig, pp. 4-8; Ackrill; Thorn, pp. 174-6). But in fact
Alexander's point requires less than he supposes: it is enough for the argument to go
through that 'B does not hold of some C' does not exclude the possibility that there is
more than one C. (Moreover, Alexander could safely allow that C may sometimes be a
singular term - as in fact he does, e.g. at in An. Pr. 350,30-352,26.)
126 By 'exposition' (ekthesis) Alexander probably refers to Aristotle's device of
taking parts of C: cf. 67,11; 89,13.15. But note that at 65,17 Alexander has used
ekthesis of the setting down of terms (i.e. as equivalent toparathesis).
127
On this see Lukasiewicz, pp. 67-72; Patzig, pp. 180-3; Brunschwig (1969);
Thorn, pp. 59-62. See Introduction, p. 29.
 4. The First Figure 131
some are not, 'One of them is ...' because this too is true if both are 10
and also if only one of them is.128
Now particular negatives are not determinate but are true both
for negatives which are particular in the specific sense and for
universal negatives. For inasmuch as what is called subaltern129 to
something is true together with it, and particular negatives are
subaltern to universal negatives, the particular negative is true
when the universal negative is true even if it is assumed in its own 15
right. Now suppose we assume that B does not hold of some C: even
if it is not possible to obtain suitable terms for every such
combination, yet it is enough to set down a single case, given that its
truth is indeterminate, to prove the conjunction non-syllogistic.
When it is true because it holds of none, and we set down suitable
terms, then the conjunction will be proved non-syllogistic for
negatives which are particular in this way, inasmuch as A holds of 20
every C and also of no C.130 (This has been proved, when the minor
is a universal negative, by means of the terms animal, man, horse,
and stone.131) And if the combination with a particular negative
minor is proved to be non-syllogistic in the case of one material
instance, it is universally non-syllogistic. For if it were syllogistic,
then you should not have been able to disprove and refute it by 25
setting down material instances when the particular negative is
taken to be true in this way (i.e. inasmuch as it falls under the
universal negative) - for it is no less true then than when it is taken
in its own right.132
Aristotle made this clear by saying 'For then there would be one in
this case too.'
In the following expression you must take the connective 'for'
instead of 'and'. In this way the phrase becomes plainer and more 67,1
congruous: 'For it will be proved in the same way as it would be if a
128
Theophrastus, F 4 Graeser - see Brunschwig (1982). (The other three texts
which Graeser assigns to F 4 make a perfectly different point, namely that
Theophrastus called particular propositions indeterminate because they do not refer
to any one individual. Like Aristotle, Theophrastus evidently used the word
'indeterminate' in different senses.) Note that in saying that 'One of them is ...' is
indeterminate, Theophrastus in effect construes disjunctions as inclusive. Cx ory is F'
will be true if both x and y are F.) The later Peripatetic tradition construed
disjunctions as exclusive: e.g. Galen, Inst. Log. iv 1-4; Alexander, in Top.
174,5-176,26; cf. Maroth, pp. 55-69 (see above, 19,23 note).
129
See above, 45,26.
130
Omitting toi before medeni.
131
See Aristotle, Are. Pr. 26a2-9; above, 55,21-56,27.
132
Alexander's exposition is cumbersome. The point is this. Given that 'B holds of
no C' entails 'B does not hold of some C', then if we can infer Q from P together with 'B
does not hold of some C' we can infer Q from P together with 'B holds of no C'. But no
syllogistic conclusion follows from 'A holds of every B' together with 'B holds of no C'.
Hence no syllogistic conclusion follows from 'A holds of every B' together with 'B does
not hold of some C'.
 132 Translation
universal privative had been assumed.'133
He calls 'not of some' indeterminate not in the sense that the
proposition is indeterminate134 (for it is determinate, since it is
5 particular and the particular is as determinate as the universal) but
rather in the sense that the truth which is signified by it is not
determinate. For it may be true both if it holds of none of that of some
of which it is assumed not to hold, and also if it holds135 of some, as I
have already said.
In the case of this figure, since the particular negative is true in an
indeterminate way (being true both in its own right and because of
10 the universal), he has proved that it is non-syllogistic in both cases -
both if it is assumed in its own right and if it holds because of the
universal. If in its own right, by exposition (i.e. by taking a term of
some of which B does not hold); if because of the universal, by taking
the terms he gives here. For the middle, which was taken not to hold
of those items which are in the last term as in a whole, is taken not to
hold of them when they are separated from the whole and taken by
136
15 themselves. In the second and third figures he is content to prove
similar conjunctions non-syllogistic solely by the indeterminacy of
the particular negative.137
The explanation of why the combination with a particular negative
minor is non-syllogistic is this. Part of C is separated from B; and A
20 includes all B. Because of B, A will not include the part of C which lies
outside B; yet there are parts of A which exceed B (since A is major in
relation to B), and with these parts it may encompass the part of C -
in the case of some material instances - and also, conversely, not
encompass the same part.138
When the major is a universal negative, the terms with which he
refuted this conjunction were inanimate, man, swan, and snow.139

25 And it will be proved in the same way if the universal is posited

as privative. [1.4, 26b20-l]
The phrase means either 'For it will be proved in the same way ...' or
'Thus ... in the same way ...', so that he uses 'and' either instead of
133
See below, 67,27-68,6, which is an expansion of this little paragraph. The second
of the two accounts which Alexander gives in the later paragraph is plainly the
correct interpretation of Aristotle's meaning.
134
i.e. unquantified.
135
Omitting me: see 66,2-3.
136 fjjjg sentence applies to the first of the two cases distinguished in the previous
sentence ('If in its own right,...').
137
See Aristotle, An. Pr. 27b20-3, 27-9; 28b24-31; below, 87,5-88,14; 104,25-105,8;
cf. Alexander, in An. Pr. 203,15-35.
138
See above, 56,8-17.
139
See An. Pr. 26bll-13.
 4. The First Figure 133
140
'for' or instead of thus'.
For it will be proved that the combination which has a particular
negative minor is non-syllogistic in the same way as the
combination with a universal negative minor. But he has already 30
proved the latter case.
Or else he uses the proof by way of the indeterminate separately 68,1
for each conjunction - for the one in which the major is a universal
affirmative and the minor a particular negative, and for the one
which has a universal negative major and a particular negative
minor. Thus, after he has first given an account of the combination 5
which has a universal affirmative major, he says that it will be
proved in the same way if the major, being universal, is transformed
into a negative.

4.6 Combinations of particular premisses

Again, if both the intervals141 are particulars, either
predicatively or privatively <(or one is said predicatively and
the other privatively), or one indeterminate and the other
determinate, or both indeterminate, in none of these cases will
there be a syllogism. Terms common to all cases: animal,
white, horse; animal, white, stone.> [1.4, 26b21-5]
He proves that the combinations are also non-syllogistic if both
premisses are particular, whichever quality they have, whether 10
both are affirmative or both negative or the major affirmative and
the minor negative or conversely. Nor are they syllogistic when the
premisses are indeterminate, whether both are affirmative or both
negative or the major negative and the minor affirmative or
conversely. The terms are common to all eight combinations. For
holding: animal, white, horse. For not holding: animal, white, stone. 15
The reason why all such conjunctions are non-syllogistic is clear. For
inasmuch as nothing is assumed universally, the middle term does
not share with the other two terms in virtue of the same part but
may be posited in virtue of different parts, so that it is hardly even a
middle term for them. Reasonably enough,142 nothing is deduced
syllogistically from two particular propositions, since it is supposed 20
that a syllogistic justification justifies by means of a universal and
proves one of the items under it.143
140
See above, 66,29-67,2. For comparable comments on aberrant particles in
Aristotle see Alexander, in An. Pr. 129,9; 203,3-5; 221,6; cf. e.g. Ammonius, in Int.
257,10-13.
141
Diastemata: i.e. the premisses.
142
68,19-69,9 = FDS 1090 (68,21-69,4 = SVFll 260).
143
See above, 44,6-13. For the fact that every syllogism must have at least one
universal premiss see Aristotle, An. Pr. 41b6-22; Alexander, in An. Pr. 266,8-267,27.
 134 Translation
As for those who think that something can be deduced syllogisti-
cally from two particulars144 - for example, those who in proof of this
adduce the arguments which the Stoics say conclude unmethodic-
ally145 and who also collect certain other examples - either they must
25 disprove the examples cited by Aristotle to show that such conjunc-
tions are non-syllogistic, proving that they are146 false (and then they
would be saying something), or else they must recognize that setting
down even one example is sufficient to prove the combination
non-syllogistic.
Besides, in the examples they adduce, the conclusion does not
follow by necessity from what is assumed and posited, i.e. it does not
follow inasmuch as they are the case,147 but rather inasmuch as in
30 these cases the universal premiss is true: although they get the
conclusion from this universal premiss, they leave it out when they
assume the premisses and they divide the minor into two.148
69,1 Of the arguments which they call unmethodically concluding, all
those which have two particular premisses deduce in this way; and it
is easy to prove this for the examples they adduce. And the other
examples which they mistakenly take hold of in their effort to prove
that such combinations are syllogistic are also not difficult to refute.

4.7 Additional syllogisms

5 Thus it is evident from what has been said that if there is a
particular syllogism in this figure, it is necessary for the terms
to stand as we have said; <for if they stand in any other way,
there is not.> [1.4, 26b26-8]
He has proved that in the particular syllogisms (there were two of
these) the major must be universal and either affirmative, in which
10 case the conclusion is affirmative, or negative, in which case the
conclusion is negative; and the minor must be a particular
affirmative in both syllogisms.
144
On these people see Barnes (1990a), IV.5. Note that [Ammonius], in An. Pr.
XII. 10-16, also allows that there may be syllogisms with two particular premisses.
145
See above, 21,32 note. For the view that unmethodically concluding arguments,
all or some of them, have particular premisses see Alexander, in An. Pr. 345,18-20;
Philoponus, in An. Pr. 321,12-14; [Ammonius], in An. Pr. 70,20-2; 71,3-4. See Barnes
(1990a),V.2-3.
146
Reading auta for auten.
147
See Aristotle, Are. Pr. 24b20; cf. above, 21,10 note.
148
Alexander's general view is this: corresponding to every unmethodical
argument, 'Pi, Pz, ..., Pn: therefore Q', there is a categorical syllogism, U, P*:
therefore Q', where U is a universal proposition and P* is some sort of amalgamation
of the premisses of the unmethodical argument. Thus Alexander represents his
opponents as starting from this syllogism, omitting U, and dividing P* into PI and ?2.
See the examples above, 21,25-22,24; cf. the other references in 21,10 note, and see
esp. Barnes (1990a), V.8-11.
 4. The First Figure 135
It is clear too that all the syllogisms in it are perfect <(for they
are all perfected by means of the initial assumptions) .. .> [1.4,
26b28-30]
He defined a perfect syllogism as one which 'needs nothing else
apart from the assumptions in order for the necessity to be 15
evident'.149 He says that all the syllogisms which have been proved
in this figure are perfect, since they all are perfected by means of
what was initially assumed and posited and need nothing else. The
items assumed initially are those by means of which the necessity is
evident - namely, 'of every', which is equivalent to 'in as in a whole',
and 'of none' or 'in no'.150 20

... and that all problems are proved by means of this figure ...
[1.4,26b30-l]
This is the main reason why this figure is the first.151 This is why
Aristotle, having said this, continues: 'I call this the first figure',152
suggesting that he has given the reasons why this figure is rightly 25
called the first.
Aristotle153 has proved that the four syllogisms which have been
set out are the primary syllogisms in the first figure. To these four
Theophrastus adds five more, which are no longer perfect or
indemonstrable. Aristotle, too, will mention these, some at a later
point in this book and some at the beginning of the next, i.e. the 70,1
second, book. Three of them, which come about from the first three
indemonstrables by conversion of the conclusion, he will mention at
the beginning of the second book, where he is investigating whether
it is possible for there to be several conclusions from the same
suppositions.154 The other two he will mention in this book, where 5
he says that of the non-syllogistic combinations those similar in
form are perfectly155 non-syllogistic, while in non-syllogistic
combinations which have a universal negative and are dissimilar in
form something can be deduced with the minor term said of the
major.156 These two conjunctions are in the first figure, one with a
universal affirmative major and a universal negative minor and the 10
other with a particular affirmative major and a universal negative
minor. (The remaining combinations are either syllogistic or similar
149
Aristotle, Are. Pr. 24b23-4.
150
See above, 32,11 note.
151
See above, 48,21-49,6.
152
Are. Pr. 26b33.
153
69,26-9 =FDS 1101.
154
Are. Pr. Bl,53a3-14.
155
Reading teleon for teleioi with Wallies (see his corrigenda on p. 426).
156
Are. Pr. 29al9-27.
 136 Translation
in form or do not have a universal negative minor.) Theophrastus
calls one of them the eighth and the other the ninth. When both
15 their premisses are converted, a particular negative conclusion is
deduced with the minor term said of the major. If particular
negatives converted, then each of the two combinations would be
syllogistic necessarily and primarily, proving the point at issue by
conversion of the conclusion. But since they do not convert, the
combinations are non-syllogistic with regard to the point at issue -
20 but something else can be deduced syllogistically by means of them.
We will say something about these syllogisms when we come to the
passages in question.157
157
See below, 109,4-111,27, on An. Pr. 29al9-27. (Alexander's commentary on An.
Pr.'B- and hence his discussion of 53a3-14 - is lost). On the 'additional' syllogisms
see e.g. Volait, pp. 30-6; Bochenski (1947), pp. 56-61; Rose, pp. 57-79; 109-32; Patzig,
pp. 112-14; Thorn, pp. 52-5 (also Appendix 2, pp. 213-15). (1) Aristotle says at An. Pr.
53a3-14 that whenever the conclusion of a syllogism converts, then there is a further
syllogism to hand. He thus recognizes - but only implicitly and in entirely general
terms - eight further syllogisms; for we may convert the conclusions of Barbara,
Celarent and Darii in the first figure, of Cesare and Camestres in the second figure,
and of Darapti, Disamis and Datisi in the third figure. Let us call these new
syllogisms Group A. (2) At An. Pr. 29al9-27 Aristotle remarks that, of the
combinations which he has rejected as non-syllogistic, some will in fact yield a
conclusion in which the minor term is predicated of the major: the combinations are
those which consist of a universal negative and a universal or particular affirmative.
Aristotle thus recognizes - but only implicitly and in entirely general terms - five
further syllogisms, namely Fapesmo and Frisesomorum in the first figure, Firesmo in
the second figure, and Fapemo and Frisemo in the third figure. (There are, in all,
twelve combinations - four in eachfigure- which contain a universal negative and a
universal or particular affirmative. Seven of these combinations are recognised as
syllogistic in An. Pr. 1.4-6.) Let us call these new syllogisms Group B. (3)
Theophrastus added five syllogisms to the first figure (F 17 Graeser: in addition to the
present text see below, 109,29-110,21; Alexander, Corey. 60; Apuleius, Int. 193,7-13
[but the text is corrupt: see the discussion in Sullivan, pp. 155-7]; Martianus Capella,
IV 411; Boethius, Syll. Cat. 813BC, 814C-816C [Boethius mentions Eudemus - and
Porphyry - alongside Theophrastus]; scholium to Aristotle, 188a4-12). The five are
Baralipton, Celantes and Dabitis, from Group A, and Fapesmo and Frisesomorum,
from Group B. Theophrastus seems to have described these syllogisms in detail and
to have shown how they can be reduced to the four direct syllogisms in the first figure.
No text hints that Theophrastus made any additions to the other figures. Later
logicians refer to these syllogisms as 'reflected' (kat anaklasin or antanaklomenoi:
Boethius, Syll. Cat. 815A; Philoponus, in An. Pr. 79,10); the terminology was in use
before the time of Apuleius, who translates it by reflexim, and it may go back to
Theophrastus. (4) According to Apuleius, 'Ariston and some of the more recent
Peripatetics' added a further five syllogisms (Int. 193,16-20; see Sullivan, pp. 165-6).
(Note that the best MSS give Aristoteles instead of Aristo. At Int. 193,24 the MSS
again give Aristoteles; but here the text should not be changed to Aristo: see Moraux
(1973/84), I, pp. 190-1.) They are the so-called subaltern moods, i.e. Barbari,
Celaront, Celantos, Cesaro, and Camestrop. (Any combination which yields a
universal conclusion will also yield the particular proposition which falls under the
universal.) Call them Group C. Apuleius regards the addition of Group C as 'very
silly'. (5) At Inst. Log. xi 3-7 Galen enumerates certain additional syllogisms. (The
text is desperate: we follow here the version suggested by Jonathan Barnes and
Michael Frede in their forthcoming edition.) To the first figure Galen adds the five
 4. The First Figure 137

Theophrastean syllogisms (three from Group A, two from Group B) and the two
Aristonian syllogisms (from Group C). To the second figure he adds Cesaro and
Camestrop from Group C (he omits Celantop). To the third figure he adds Daraptis
from Group A. (6) At in An. Pr. 109,4-111,27 Alexander adds Firesmo to the second
figure and Fapemo and Frisemo to the third, all from Group B. (7) Philoponus accepts
the three first figure syllogisms from Group A (in An. Pr. 79,10-20); he accepts the five
syllogisms from Group B (in An. Pr. 112,21-113,20); and he implicitly endorses all
eight syllogisms from Group A (in An. Pr. 388,18-399,9). (8) The general disregard for
Group C is not, perhaps, surprising; for (as Apuleius in effect remarks) the subaltern
moods do not add anything to the probative power of categorical syllogistic. But why
are Group A syllogisms not added to the second and third figures? The explanation is
to be found in Galen; for he observes that the reflected form of Cesare is simply
Camestres, and vice versa, and that the reflected form ofDisamis is simply Datisi and
vice versa (Inst. Log. xi 7). (Convert the conclusion of Cesare and you get: B holds of no
A; B holds of every C: therefore, C holds of no A. And this is Camestres, with the
premisses in reversed order.) It is, then, strange that Galen allows the reflected form
ofDarapti; for Daraptis is in fact identical with Darapti, as Apuleius observes (Int.
189,23-5). Here Galen seems to be following Theophrastus (see below, 95,30 note).
 5
The Second Figure

5.1 General remarks

When the same item holds of all the one and of none of the
other, or of all or none of each, <I call this the second figure. >
[1.5, 26b34-6]
Having spoken of the first figure and the combinations in it, and
70,25 having shown which of them are syllogistic and which non-
syllogistic, he moves to the second figure; and he will discuss its
combinations in the same way. The second figure, as we have said,1
is the one in which the common or middle term is predicated of both
terms in the problem at issue. He shows this by saying 'When the
30 same item holds of all the one and of none of the other, or of all or
none of each'; for this is equivalent to 'When the same term (that is,
71,1 the middle term, which is taken twice) holds of all of one of the
extremes - i.e. is predicated of all of it - and of none of the other - i.e.
is predicated of none of it - or is said of all or none of each of the
extremes.' Again, with this figure as with the first,2 he set down
5 examples to make evident how the middle term stands, viz. that it is
predicated of both extremes. The examples which he uses betoken
the combinations of two universal premisses. He discusses them
first, as he did in the first figure too. And he mentioned all
combinations of universal premisses in saying 'When the same item
10 holds of all the one and of none of the other, or of all or none of each.'
Now, we have said that this is the second figure, and for what
reasons.3 It is clear that by the conversion of the major premiss in
the first figure, the middle term comes to be predicated of both
extremes.4 For when it is converted, the middle comes to be
15 predicated of the term for which it was subject (it was the subject for
the major extreme); and it is also predicated of the minor. Therefore
1
Above, 46,32-47,4.
2
See above, 52,30.
3
Above, 48,12-18.
4
On the 'generation' of the second figure see above, 47,27 note.

138
 5. The Second Figure 139
it comes to be predicated of both terms once the major has been
converted. And as predicated, it has a superior place in the second
figure since being predicated is more important than being subject.5
Reasonably, therefore, this is the second figure, since the middle
(because of which syllogisms come about), having lost the position
appropriate to it (which it had in the first figure), has the nobler of the 20
two remaining positions.
A combination is syllogistic in the second figure when the major
premiss is universal. When it is particular there will be no syllogism
in the second figure - and with good reason. For the syllogistic 25
combinations in the second figure preserve the proper characteristic
of the proposition by the conversion of which the second figure was
generated from the first.6 The second premiss, the minor, must be
either particular or universal (it is syllogistic both when it is
universal and when it is particular). And it must by necessity be
opposite in quality to the major, i.e. dissimilar in form: if the latter is 30
affirmative, it must be negative; if negative, affirmative. For if they
are similar in form a syllogism does not come about in the second
figure: neither from two negatives (because in no figure is such a
combination syllogistic), nor from two affirmatives, as will be proved.
The syllogistic premisses standing thus, in this figure too there will 72,1
be four syllogistic combinations: if the major premiss is affirmative, it
is connected either to a universal or to a particular negative minor; if
the major is assumed as negative, the minor is either a universal or a
particular affirmative. 5

I say that in it the middle term is the one predicated of both ...
[1.5, 26b36-7]
This is actually the account of the second figure: it is the figure in
which the middle is predicated of both the extremes for which the
deduction must be made.

... and the major extreme is the one supposed on the middle.
The middle is posited outside the extremes and first in 10
position. [1.5,26b37-9]
By the diagram of the terms7 and the expression8 he uses he has
made clear to us that it is when the major premiss in the first figure
5
See above, 48,25-7.
6
cf. below, 95,14-19; Themistius, Max. 182-3.
7
For other references to diagrams see e.g. below, 78,4; Alexander, in An. Pr.
301,9-19; 381,8-12; Philoponus, in An. Pr. 65,20-3; 87,8; Boethius, Syll. Cat. 814B.
For illustrative examples see [Ammonius], in An. Pr. VTII.20-1, 24-5; X.10-XI.1; 39,9
app. crit. See Rose, pp. 133-6.
8
Reading lexeos (with LM and the Aldine) for taxeos.
 140 Translation
is converted that the second figure comes about. For the position and
order of the terms which he describes - the fact that the middle is
put first in order and the major supposed after it - make clear that it
9
15 is the major premiss which was converted.

5.2 Major and minor terms

... and the major extreme is the one supposed on the middle.
[1.5,26b37-8]10
People investigate whether by nature there are major and minor
extremes in the second figure, and how we can judge which they
are.11 For if it is indifferent, then it will be possible to call major
whichever sort of term we like to connect in position to the middle
20 term. This will be so because the conclusions in this figure are
negative, and universal negatives convert from one another. Hence
for this reason one term will be no more major than the other in
universal negations - since the predicate is major and these terms
are counterpredicated each of the other with an equal extension. In
the case of affirmatives, the major is the term predicated
25 universally, since it has a wider extension. (That is why they do not
convert.) Hence being major holds of it by nature. But in the case of
universal negatives this is no longer the case.
Now Herminus takes the following view.12 If in the second figure
the two terms of which the middle is predicated are co-generic, then
the major extreme is the one which is nearer to their common
73,1 genus.13 Suppose the extremes are bird and man: bird is closer than
9
In the 'diagram' for first figure syllogisms the terms are arranged thus:
A-B-C.
For the second figure the corresponding diagram is:
B-A-C
Hence it is clear that the major premiss (represented by 'A - B') has had its terms
converted. (See below, 98,2-12.)
10
It is curious that Alexander makes no comment at all on this phrase; nor does he
discuss the next phrase in Are. Pr. ('and the minor is the one further from the middle'),
which
11
his own text of Aristotle apparently lacked.
See Philoponus, in An. Pr. 67,18-68,8; 87,10-19, who concludes that it is
convention and not nature which determines the major and minor terms in the
second and third figures (cf. Themistius, Max. 181-2). See Lukasiewicz, pp. 30-2 (who
remarks that 'what John Philoponus writes on this subject deserves to be regarded as
classic'); Mignucci (1969), pp. 219-21; and esp. Patzig, pp. 118-27.
12
On Herminus' view see Moraux (1973/84), II, pp. 383-5. Bochenski (1947), p. 64
n. 218 observes that 'this is one of the most instructive passages we possess about the
Peripatetic cast of mind'.
13
Herminus' view presupposes that every term which may appear in a syllogistic
proposition has a unique and fixed position in some 'division' of terms: it represents a
node in a Porphyrean tree. The 'distance' between one term and another, given that
they appear in the same tree (i.e. are co-generic), is then determined by the number of
nodes between them. Then take a middle term, B, and two extreme terms, A and C.
 5. The Second Figure 141
man to their common genus, animal, and is produced by the primary
division. Hence bird is the major extreme. And in general, when the
terms are co-generic, the one related in this way to the common
genus is major.
If both terms are equally distant from the common genus, as man
and horse are, then one must look at the middle term which is 5
predicated of them and see of which it is predicated because of itself
and of which because of another term. If it is predicated of one
because of itself and of the other because of another term, then you
should compare the term because of which it is predicated of
something else with the term of which14 it is predicated in its own
right.
If the former term - the term because of which it is predicated of
something else - is closer to their common genus, then you should
call major the term of which the middle is predicated because of the 10
term nearer to the common genus. For example, suppose the
extremes are horse and man, and rational is predicated of them
(negatively of horse, affirmatively of man). Now rational is not
denied of horse in its own right but because horses are non-rational,
whereas rational is affirmed of man in his own right; and
non-rational is closer than man to their common genus, animal.15 15
Thus horse will be the major extreme in relation to man, even
though man is equally distant from their appropriate genus,
because the term because of which the predicate is predicated of it is
There are two possibilities. (1) A and C appear in the same tree. Then there must be
some term, D, higher in the tree under which both A and C fall. (A cannot fall under
C, nor C under A; or at least, this is apparently excluded by the fact that the first two
syllogisms in the second figure have universal negative conclusions.) Again, there are
two possibilities, (i) A and C are not equidistant from D. Then the term closer to D is
the major term, (ii) A and C are equidistant from D. Then we must consider the terms
because of which B is predicated of A and of C. Suppose that B is predicated of A
because of E and of C because of F. Then if E is closer to D than F is, A is the major
term; and if F is closer, C is major. (What if E and F are equidistant from D?
Herminus tacitly excludes this case by supposing that either E will be identical with
A or F will be identical with C; i.e. either B is predicated of A because of itself or it is
predicated of C because of itself.) (2) A and C appear in different trees, T and T*. Then
consider the highest genus in each tree, G and G*. There are, again, two possibilities.
(i) The distance between A and G is not the same as the distance between C and G*:
then the term nearer its own highest genus is the major term, (ii) A and C are
equidistant from G and G*. Then - as in case (l)(ii) - consider E and F. If E is closer to
G than F is to G*, then A is major; if F is closer, C is major. Numerous comments
might be made on this elaborate suggestion; but the matter does not merit detailed
discussion. For, as Alexander rightly says (74,5-6), the whole thing is 'a waste of
time'.
14
Reading toi hou for toi di'hon.
15
The transmitted text of this sentence has Herminus contradict himself in an
obvious (and wholly irrelevant) fashion. The text is surely corrupt, but no convincing
emendation suggests itself. We add ekeino de before engutero in line 14 and delete ho
hippos eper anthrdpos in line 15. We do not suppose that Alexander wrote these
precise words; but we do suppose that they give the sense of what he wrote.
 142 Translation
major. For rational is denied of horse qua non-rational, not qua
horse, whereas rational is affirmed of man in his own right.
20 If the extreme terms are not co-generic but belong to different
genera, then one should posit as major the one which is closer in its
own genus. For example, if something is predicated of colour and of
man, colour is the major extreme. For colour is closer to quality than
man is to substance - man is an indivisible species whereas colour is
not. If, again, they are both equally distant from their appropriate
25 genera, then you should ascend to the predicate and seek of which of
them it is predicated because of itself and of which because of
another term. And if the term because of which it is predicated of
something else is closer to its own genus, then you should deem the
major extreme to be the one of which the middle is predicated
because of that term. For example, suppose the terms are white and
man, both being indivisible species (the one of quality and the other
of substance), and suppose that rational is predicated affirmatively
of man and negatively of white. Now it is affirmed of man qua man
74,1 but denied of white not qua white but qua inanimate. Then since
inanimate, because of which rational is denied of white, is more
common and more universal and closer to inanimate substance than
man is to animate substance, then for this reason white is the major
term in relation to man.
5 Now to say all this, and to seek to prove that the major extreme in
the second figure is major by nature, is not only a waste of time - it is
not even true. For, first, if we examine not the terms taken, in and
by themselves, but rather those terms in virtue of which the
predicated term does not hold of them, then the major term will
always be in the negative proposition. For it will be either equal or
10 major in relation to the middle term, either having been taken in
this way from the beginning or becoming such because of the term
because of which the middle term is denied of it, so that16 it is major
in relation to the middle term which is denied of it. For the middle
term will not hold of that of which it is supposed not to hold, because
its opposite and co-ordinate holds of the subject; and what is
opposite to the middle term and co-ordinate with it is equal to it. It
will fail to hold either because of this term itself or because of
15 something else which is of wider extension than the middle term -
as when rational is denied of something because of inanimate. It is
equal when it holds because of non-rational - for rational is equal to
non-rational, because of which rational is negatively predicated of
horse. Thus the middle term is either equal to that of which it is
negated or minor in relation to it - when rational is denied of
something because of inanimate. For inanimate is equal to animate,
16
Reading hosts for houtos.
 5. The Second Figure 143
under which rational falls - and rational is major in relation to the 20
other term, of which it is affirmed. Now since the term predicated
affirmatively is major in relation to its subject, the term of which the
middle is denied is also major in relation to the term of which the
middle is affirmed - since the reason why it is denied is equal or
major in relation to the middle itself, which is major in relation to its
subject in the affirmative proposition. Hence the negative
proposition will always be major in relation to the affirmative.17 But
Aristotle says that the negative can be posited on the minor term as 25
well - thus the second syllogism in this figure has a negative
minor.18
Again, why will it only be in the case of negatives that we
transform the issue and seek the term because of which the middle
is negatively predicated? The same thing will be sought in the case
of affirmation. For even if rational is said of man in his own right, it 30
is not said of man primarily or qua man but qua rational. Hence if it
is said of horse because of non-rational and of man because of
rational, and if rational is equal to non-rational (since they arise 75,1
from the same division), then the method before us will not yet have
yielded the major term.
For these reasons one should not judge which is the major premiss
in the second figure in this way. For, in general, inasmuch as both
the affirmative and the negative may be on the major term in this 5
figure, then whatever term is found to be major by the method we
have described will make the combination syllogistic whether it is
taken as major or as minor. But in that case it will no longer be the
major term in this figure. (We are not seeking the major term
simpliciter but the major term in this figure.)
Yet we should not simply say that the term predicated in the 10
conclusion of the syllogism is the major term, as some people
think.19 For it is not clear which this term is: sometimes it will be
the one and sometimes the other - it is not determined (inasmuch as
universal negatives convert), and what is now major will later be
minor, and it will be up to us to make the same term both major and
minor.20 Thus in negatives there is by nature no major term, nor 15
should one take the major simply from the conclusion. For it will not
17
Alexander's argument is convoluted. It may be summarized as follows. Consider
the term, D, because of which B holds of no C. Either D is identical with not-B, in
which case it is on a level with B, or it holds of every B, in which case it is above B. But
B holds of every A. Hence B is above A. Hence C is always major in relation to A.
18
i.e. Camestres.
19
So later e.g. Philoponus, in An. Pr. 67,27-30; cf. above, 47,30; Alexander, in An.
Pr. 341,15-19.
20
Thus the combination 'A holds of every B, B holds of every C' will yield 'A holds of
no C' (by Celarent) and also 'C holds of no A' (by Celantes). Hence we cannot say which
of A and C is 'the predicate in the conclusion'; for there are two conclusions.
 144 Translation
be determined which this term is. Moreover, the conclusion has as
predicate the term taken as major in the premisses, so that the
conclusion does not show the major term - rather, the reason why a
term is predicated in the conclusion is that it is taken as the major
term.21
20 Yet we cannot say that there is no major term in the figure. For it
is determined that in it the major premiss must be universal if there
is to be a syllogistic conjunction,22 and the major premiss is the one
containing the major term. Now we must deem to be major and posit
as primary the term which we want to prove and deduce in the
problem, i.e. the term which we have predicated in it. For everyone
25 who syllogizes first defines for himself what he wants to prove and
so obtains premisses appropriate to it.23 He does not hit on the
conclusion by chance. Now it is the term predicated in the problem
set up for proof which should be posited as the major term. For even
if it converts, and for this reason the same term becomes subject,
nonetheless it was and remains predicated in the problem which is
30 set up for us to prove. (That is why if we get a different conclusion
we convert it.) Hence for us, who prove and syllogize and order the
terms, this is the major term; for it is not in their own natures that
terms are major and minor in negatives but rather with reference to
our purpose in the conclusions - and it is clear that the term
predicated in the problem is also predicated in the conclusion.24

5.3 Second figure syllogisms

76,1 Now there will not be a perfect syllogism in this figure, but
there will be a potential one both when the terms are universal
and when they are not universal. [1.5,27al-3]
He called perfect syllogisms those in which the necessity is evident
5 from the suppositions (the syllogisms in the first figure have been
shown to be such) and imperfect syllogisms those which need some
externalitem or items - items which are necessary because of the
terms laid down but have not been taken in a premiss.25 The
syllogisms in this figure and in the third are of this sort. For three of
the syllogisms in this figure need conversion in order for their
10 necessity to become evident; and while the fourth is not proved by
conversion, the necessity of its deduction is proved by reduction to
the impossible.
21
This is not Alexander's normal view: see above, 47,29-48,6.
22
See above, 71,22-6.
23
See esp. Kapp for this way of looking at syllogisms. But it is not the only way in
which Alexander (or Aristotle) regards them.
24
See above, 47,30.
25
See Aristotle, Are. Pr. 24b22-6; above, 23,17-24,12.
 5. The Second Figure 145
The sentence 'But there will be a potential one both when the
terms are universal and when they are not universal' means that in
this figure too it is possible to get a syllogism both when both
premisses are universal and also when they are not both
universal.26 (In the latter case, clearly only one of them is particular, 15
namely the minor. For there is no syllogism if both premisses are
particular, nor - as we have said27 - if the major is particular.)

When they are universal, there will be a syllogism when the

middle holds of all the one and of none of the other, on
whichever the privative may be; and in no other case. [1.5, 20
27a3-5]
We have said28 that if we are to produce a syllogistic combination,
the premisses in this figure must be dissimilar in form with regard
to quality. When both premisses are universal and are also
dissimilar in form, there will be a syllogism - whether the major is
negative and the minor affirmative or conversely. For both 25
combinations of this sort are syllogistic, deducing a universal
negative.
As we have already said,29 there are four syllogistic combinations
in this figure, as there are in the first: two with universal premisses
and two in which one, the minor, is particular.
The order of these syllogisms is as follows.30 First are the two 30
which deduce a universal negative from universal premisses. (No
affirmative is deduced in this figure, because nothing is deduced
from premisses similar in kind.) These are first, because the
universal is more valuable than and prior to the particular.31 Of
these two, the first will be the one which has the major premiss 77,1
universal and negative and the minor universal and affirmative. For
this syllogism needs a single conversion to make its necessity
evident: by converting the major premiss it is reduced to the second 5
indemonstrable, as will be proved.32 The syllogism with the major
premiss universal and affirmative will be ranked after this; for it
28
By 'a potential one' Aristotle here means 'a potentially perfect syllogism': see Are.
Pr. 41b33 (cf. 28al6), with Philoponus, in An. Pr. 255,27-9 (cf. Alexander, in An. Pr.
271,2-6, who takes the word to mean 'potentially evident'); Patzig, p. 46. Here,
however, Alexander appears to take 'there will be a potential one" to mean 'it is
possible to get a syllogism' (so too Philoponus, in An. Pr. 87,30-88,2).
27
Above, 71,22-6.
28
Above, 71,28-33.
29
Above, 72,1-5.
30
cf. Apuleius, Int. 190,15-191,5; Philoponus, in An. Pr. 88,26-33; 94,3-7; see above,
51,8-24, on the order of syllogisms in the first figure, and below, 95,25-31 and
96,22-97,11, on the order in the third figure.
31
See above, 48,27.
32
See Aristotle, Are. Pr. 27a5-9; below, 78,12-22.
 146 Translation
needs two conversions to prove that the major extreme holds of the
minor extreme. Thus since it needs more help than the syllogism
before it, it is reasonable to rank it after it, being more imperfect
than it.33 It will be shown that in this syllogism two propositions are
converted.34
10 As we have said,35 the two syllogisms which deduce particular
negatives must have their major premisses universal - the minor in
both cases is particular. The syllogism in which the major is
universal and negative and the minor a particular affirmative will
be third. Its necessity becomes evident by the conversion of the
15 major premiss; for if this is converted we get the fourth
indemonstrable in the first figure. (If the converse has not been
explicitly assumed in the premisses, it is there potentially.36) For
this reason, this syllogism is less imperfect than the one after it,
which has the major premiss universal and affirmative and the
minor particular and negative. This cannot be proved by conversion,
20 but needs more help. Its necessity is proved by reduction to the
impossible, and in this proof we introduce from outside a premiss
which is not disclosed by what has been laid down. Nevertheless, the
point at issue is not deduced by using other terms or by proving
something else. For, as will be shown, the syllogism deduces the
opposite, and, since this is evidently impossible, by rejecting it we
25 justify the point at issue which is its opposite.37 On the other hand,
syllogisms which deduce by way of conversions syllogize the point at
issue itself. As the syllogism in question needs so much help, it
stands to reason that it is more imperfect and has the last place
among the syllogisms of this figure. For the three syllogisms before
it which are proved by conversion can also be proved by reduction to
30 the impossible, whereas the fourth cannot be proved by conversion
and is justified as a syllogism only by reduction to the impossible.
While examining the combinations in this figure - which and how
many they are and which are syllogistic and which non-syllogistic -
78,1 Aristotle sets out the combinations by way of letters, as he did in the
first figure.3 The letters he uses are not A, B, C, as in the first
figure, but M, N, O: as middle term he takes M, which is predicated
of both the other terms and has first position in the diagram;39 as
33
On degrees of imperfection see Patzig, pp. 48-9 (on degrees of perfection see
below, 113,8 note).
34
See Aristotle, An. Pr. 27a9-14; below, 78,25-79,25.
35
See above, 71,29.
36
See above, 58,13-22.
37
This account of reduction to the impossible is compact to the point of
unintelligibility; for a fuller and clearer account see below, 83,24-84,3, to which
Alexander refers us (cf. above, 24,18 note).
38
See above, 53,30 note.
39
On logical diagrams see above, 72,11 note.
 5. The Second Figure 147
major extreme N, which is laid down next after the middle; and as 5
last and minor O.
By saying 'on whichever the privative may be; and in no other
case' he makes it clear that if there is no privative premiss, there
will not be a syllogistic combination in this figure. And, as he proves,
there are twelve non-syllogistic combinations in this figure apart
from the indeterminate ones, since there are sixteen combinations
in all.

5.4 Combinations of universal premisses

5.4.1 Cesare and Camestres

Let M be predicated of no N and of every O. Since the privative 10
converts, <N will hold of no M. But M was supposed to hold of
every O. Hence N of no O (this has been proved earlier).> [1.5,
27a5-9]
It has been proved above that actual universal privatives convert
from themselves.40 Thus since the major premiss has been assumed
as a universal negative (for it is supposed that M holds of no N), then
if it is converted, N will hold of no M. But it is supposed that M holds 15
of every O. With such premisses we get the second syllogism of the
first figure, which deduces a universal negative, 'N holds of no O'.
Thus this will also be deduced from the combination set out here.
Let M be animal, N inanimate, O man. If animal is said of no
inanimate, by conversion inanimate is said of no animal. But animal 20
is said of every man. Therefore inanimate is said of no man. This is
the first syllogism of the second figure and it is perfected with such
help.41

Again, if M holds of every N and of no O, O will hold of no N.

<For if M holds of no O, O holds of no M. But M held of every N.
Therefore O will hold of no N. For we have got the first figure
again. And since the privative converts, N will hold of no O.
Hence there will be the same syllogism.> [1.5,27a9-14]
We ought first to infer the conclusion which should be proved in this 25
conjunction (namely, that N holds of no O - for it is supposed that N
is the major term, and so we must predicate it in the conclusion),
and then to turn to the proof of how this conclusion is deduced. But
Aristotle42 omitted the final conclusion, and instead gives the 79,1
40
See Aristotle, An. Pr. 25al4-17; above, 31,4-34,23.
41
On proofs by conversion see esp. Patzig, pp. 137-44.
42
Reading ho de (conjectured by Wallies, index verborum s.v. hode) for hode.
 148 Translation
demonstration by which the following conclusion will be proved: he
says that O will hold of no N. And he proves that this is the first
conclusion to result from this combination. For since it is supposed
that the minor premiss, MO, is a universal negative, and since
5 universal negatives convert, O will hold of no M. But it was
supposed that M holds of every N. From this it is deduced that O
holds of no N. (The major was a universal affirmative.) With these
suppositions we again get the second indemonstrable in the first
figure; for O holds of no M, M holds of every N, and from this it can
be deduced that O holds of no N. Thus by converting the minor
10 premiss he proves what he said would be the conclusion. But since
this was not the point at issue to be proved (for it was supposed in
the combination that the major term is N and the major term must
be predicated in the conclusion), he converts the conclusion, which is
a universal negative, and thus proves that N holds of no O - and this
was what he had to prove to be deducible. Thus we needed two
15 conversions for the proof of the point at issue: we converted both the
minor premiss and the conclusion which then resulted. Let M be
animal, N man, O inanimate. Animal holds of every man, animal of
no inanimate. Proof: inanimate holds of no animal, animal of every
man, inanimate of no man, and therefore man of no inanimate.43
20 He added 'Hence there will be the same syllogism' to show that
the conclusion that N holds of no O will be proved by the same
syllogism by which it was proved that O holds of no N. This was
proved to be deducible through the second syllogism in the first
figure, by converting the universal negative. Hence the former
conclusion is also proved by this syllogism. For it is supposed to be
agreed that the universal negative, which was the conclusion,
25 converts from itself, and no other syllogism is needed for this.44

It is also possible to prove these by reduction to the impossible.

[1.5,27al4-15]
He says that by using reduction to the impossible45 you can also
prove that the combinations just mentioned are syllogistic and
deduce a universal negative conclusion.
30 The first syllogism we will prove in the following way. Suppose the
premisses that M holds of no N and that M holds of every O. I say
that N holds of no O. For if this is impossible, let N hold of some O,
which is the opposite of'of none'. But either the latter or the former
must be deduced. For in every case one part of the contradictory pair
43
On this paragraph see Patzig, pp. 140-1.
44
Other interpretations of Aristotle's phrase are possible: see Ross, p. 308; Patzig,
p. 186 n. 18.
45
On proofs by reduction to the impossible see above, 24,18 note.
 5. The Second Figure 149
holds. Since, then, it is supposed that N holds of some O, it is clear 80,1
that O will hold of some N, because particular affirmatives convert
from themselves. But it is also supposed that M holds of every O. Thus
we get the combination: M holds of every O, O holds of some N.
Therefore M holds of some N, which is impossible. For it is supposed 5
that M holds of no N. Now if this impossibility follows because it was
hypothesized that N holds of some O, then the hypothesis is
impossible and false. But if it is false, its opposite is true. But the
opposite of'N holds of some O' is 'N holds of no O'. Therefore N holds of
no O, which we aimed to prove.
Or is it better to prove it in the following way?461 say that N holds of 10
no O in the combination set out. For if someone says that this is false
let its opposite, that N holds of some O, be hypothesized. But it is also
supposed that M holds of no N. Then it can be deduced in the first
figure that M does not hold of some O. But it was supposed that M
holds of every O. Therefore something impossible has been deduced,
namely that M does not hold of some O, on the hypothesis that N 15
holds of some O. Therefore47 the hypothesis is false. Therefore its
opposite, that N holds of no O, is true.
Again, suppose the second combination: M holds of every N; M
holds of no O. I say that N holds of no O. For if of some and if it is also
supposed that M holds of every N, then M will hold of some O, which is
impossible. For M held of no O. Therefore the hypothesis on which 20
this has been deduced, i.e. the hypothesis which posits that N holds of
some O, is false. Therefore its opposite, that N holds of no O, is true.
If we convert the hypothesis and produce 'O holds of some N' and
co-assume 'M holds of no O', again something impossible is deduced,
namely that M does not hold of some N. For it was supposed to hold of
every N.

Thus it is evident that there is a syllogism when the terms stand 25

thus <- but not a perfect one. For the necessity is perfected not
only from the initial assumptions but also from other items. >
[1.5, 27al5-18]
After he has proved by conversion and reduction to the impossible
that these two combinations in the second figure are syllogistic, he
remarks that the syllogisms are not perfect inasmuch as they
needed a demonstration from outside in order to make their 30
necessity evident.

46
The first reduction uses Darii, the second Ferio; Alexander presumably thinks
that the second is better because it does not require us to convert the hypothesis.
Both proofs are, of course, equally 'good'.
47
Reading ara for gar.
 150 Translation
5.4.2 Non-syllogistic combinations
81,1 If M is predicated of every N and O, there will not be a
syllogism. <Terms for holding: substance, animal, man. For
not holding: substance, animal, number. (Substance is the
middle.)> [1.5, 27al8-20]
Having proved the syllogistic combinations with two universal
premisses in the second figure, he sets down the non-syllogistic
5 combinations and shows that they are non-syllogistic, setting down
material instances to prove that in them the major extreme may
hold of all and also of none of the minor. The remaining
combinations with two universal premisses are the ones which are
similar in form, i.e. the combination with two universal affirmatives
and the combination with two universal negatives.
First, then, he tackles the combination with two universal
10 affirmatives, in which M holds of every N and O. For the case that N
holds of every O, he sets down the terms substance for M, animal for
N, man for O. For substance holds both of every animal and of every
man. And animal, which was N, holds of every man, which was
supposed for O.
For the case in which N holds of no O he posits number for O.
15 What he says would be more evident if stone had been posited. For
the premisses are again true in the same way (substance holds both
of every animal and of every stone), and animal holds of no stone.
Either he took number to stand for unit - for they say48 that units
are substances. (They say that units are not quantities inasmuch as
they are neither continuous nor determinate.49 They are substances
inasmuch as they are capable of receiving contraries50 - for a
20 beginning and an end are contraries, and numerical units possess
both.) But if a unit is what is undivided in quantity, a point what is
undivided in magnitude, and an instant what is undivided in time,51
then they will have one genus52 - they will be quantities. Further,
the unit is a part of number. For numbers are compounded from
units.53 But what is a part of a quantity is a quantity.
25 It is better to say that he uses the example rather loosely - in
other cases too we shall find him using examples without deliberate
48
Who? According to Philoponus, the view is Pythagorean (in An. Post. 301,26-8);
but Alexander's 'they' are no doubt earlier scholars who have in mind Aristotle's own
remark that 'a unit is a substance without position' (An. Post. 87a36).
49
Quantities are either continuous or determinate: Aristotle, Cat. 4b20.
50
This is the proper characteristic of substances: Cat. 4alO-ll.
51
See e.g. Aristotle, An. Post. 77a22-3 (units); Metaph. 1016b24-6 (points); Phys.
233b33 (instants).
52
Reading hen genos for en genei (Wallies' emendation of en genesi, which the MSS
present).
53
See e.g. Aristotle, Metaph. 1053a30; cf. Cat. 4b23 for number as a quantity.
 5. The Second Figure 151
54
precision - since the doctrine is Pythagorean. For according to the
Pythagoreans numbers are substances since they are principles of
substances.55
The reason why there is no syllogistic combination from two
universal affirmatives in the second figure is that the middle is
predicated of both extremes, and the predicate is the major term. 30
Thus, being major in relation to both, it may be predicated of one
extreme term in virtue of one of its parts, and of the other in virtue of
another. And in this way there is nothing which the extremes share
with each other, if they each share with the middle term in virtue of
different parts of it. For the extremes must share in one and the same
thing if there is to be a syllogism.56 82,1

<Nor when M is predicated of no N and of no O. Terms for

holding: line, animal, man. For not holding: line, animal,
stone.> [1.5, 27a20-3]
He proves in the same way that the combination with two universal
negatives is non-syllogistic, setting down terms for holding of all and
of none: line for M, animal for N, man and stone for O. For line holds
of no animal, of no man, and of no stone, and animal holds of every 5
man and of no stone. We have already said for what reason nothing
is deduced from two negative premisses.57
After this he continues as follows: if there is a syllogism with a
universal conclusion in the second figure,58 it must have premisses
dissimilar in form of the sort we have described59 - one a universal
negative, the other a universal affirmative. Otherwise there will be 10
no syllogism. For it has been proved that, apart from these, the
conjunctions of universal premisses in this figure are non-
syllogistic.

5.5 Combinations of universal and particular premisses

5.5.1 Festino
If the middle is universal in relation to one term ... [1.5, 27a26]
54
We follow LM and the Aldine: after paradeigmasin B has all'hosper with te writ-
ten above -per and Wallies prints allos te.
65
So Philoponus, in An. Pr. 92,9-12 (for the evidence about Pythagorean number
metaphysics see Burkert (1972), pp. 28-52). Cf. Aristotle, Are. Pr. 27b6-8; below, 86,5-6.
56 57
For the style of argument see above, 56,27 note. See above, 57,10-14.
58
Alexander is paraphrasing Are. Pr. 27a23-5. In modern texts Aristotle continues
as follows: '... if there is a syllogism when the terms are universal', i.e. 'if there is a
syllogism with universal premisses'. It is unlikely that Alexander should have
misunderstood this, so we suppose that he had a different text before him.
59
See above, 71,29-31.
 152 Translation
Having discussed the combinations with two universal premisses,
15 and having proved some of them to be syllogistic and the others
non-syllogistic, he now discusses those which are compounded from
a universal and a particular. He shows which they are, and which of
them are syllogistic and which non-syllogistic. He shows that here
too the combinations dissimilar in form are syllogistic - provided
that the universal is preserved for the major premiss. There are two
20 syllogistic combinations: one in which the major is a universal
negative and is compounded with the minor which is a particular
affirmative; the other in which the major is a universal affirmative
and the minor a particular negative.

For if M holds of no N and of some O, it is necessary for N not to

hold of some O. <For since the privative converts, N will hold
of no M. But M was supposed to hold of some O. Hence N will
not hold of some O. For we get a syllogism by way of the first
figure.> [1.5,27a32-6]

25 He proves this combination to be syllogistic by invoking the

conversion of the universal negative MN. For if M holds of no N, N
holds of no M. But it is supposed that M holds of some O. We get the
combination of the fourth syllogism in the first figure (N holds of no
M, M of some O), from which it was deduced that N does not hold of
30 some O. This then will also be the conclusion in the second figure
combination before us.
83,1 It is also possible to prove it by reduction to the impossible. For
suppose that M holds of no N and of some O: I say that N does not
hold of some O. For if this is not so, then its opposite, that N holds of
every O, is the case. But it is also supposed that M holds of no N. It
5 will be deduced, according to the second syllogism in the first figure,
that M holds of no O, which is impossible - for it is supposed that M
holds of some O. The hypothesis by means of which this was deduced
is therefore false. Its opposite, therefore, is true. And opposed to 'N
holds of every O' is the proposition stating that N does not hold of
some O.

5.5.2 Baroco
10 Again if M holds of every N and not of some O, it is necessary
for N not to hold of some O. For if it will hold of every O, <and
M is predicated of every N, it is necessary for M to hold of every
O. But it was supposed not to hold of some.> [1.5,27a36-bl]

He does not reduce this syllogism to one of the indemonstrables by

 5. The Second Figure 153
conversion and thus prove it as he proved the three before it in the
second figure. The reason is that it is not possible to conduct such a
proof for the present combination. For if we convert the major, 15
which is a universal affirmative, it no longer remains a universal,
since a universal affirmative converts with a particular affirmative;
and neither in the first figure nor in the second is there a syllogistic
combination in which the major is particular. Nor in general is there
a syllogistic combination in which the two premisses are particular.
(The major becomes particular through the conversion, and the 20
minor was taken as a particular at the beginning.) For this reason,
then, it is not possible to use conversion of the major premiss. And it
is impossible to convert the minor, which is a particular negative:
for particular actual negatives do not convert with particular
negatives.60 So he uses reduction to the impossible to prove that the 25
combination under consideration is syllogistic. For if it is not true in
the case of the supposed premisses that N does not hold of some O,
then its opposite will be true. This is that N holds of every O. But it
is supposed that M holds of every N. Thus we get the combination in
the first figure with two universal affirmatives, from which it will be 84,1
deduced that M holds of every O. But this is impossible; for it is
supposed that it does not hold of some O. What was hypothesized,
that N holds of every O, is therefore false. Its opposite, that N does
not hold of some O, is therefore true.

And if M holds of every N and not of every O, <there will be a 5

syllogism that N does not hold of every O. The demonstration is
the same.> [1.5, 27bl-3]
He61 conducted the proof in terms of 'does not hold of some'. And
since 'does not hold of some' is equivalent to 'not of every', differing
only in expression,62 he now makes the transformation and instead
of supposing that M does not hold of some O, he posits that M does
not hold of every O. He says that there will be the same syllogism,
and by way of the same proof, if the particular negative is 10
transformed into this expression. For the conclusion too will differ in
expression in the same way - it will be deduced that N does not hold
ofeveryO.
This is an argument of the sort which the more recent thinkers
call subsyllogistic:63 it takes something equivalent to the syllogistic
60
See below, 103,15-19, on Bocardo.
61
84,6-19 = FDS 1084 (84,11-19 = SVFII264).
62
See above, 62,12 note.
63
The 'more recent thinkers' are Stoics; and subsyllogistic arguments had been
discussed by Chrysippus. See Galen, Inst. Log. xix 6 ('the arguments called
subsyllogistic <are expressed> in formulations equivalent to the arguments said to
be syllogistic'); and for examples of such arguments see Alexander, in An. Pr.
 154 Translation
premiss and deduces the same thing from it. ('Does not hold of some'
15 has been transformed into 'does not hold of every', which is equivalent
to it.) The more recent thinkers deny64 that such arguments are
syllogisms, since they look to the words and the expression.65
Aristotle, however, looks to the meanings (when the same things are
meant) rather than to the words and says that the same syllogism is
deduced when the expression of the conclusion is transformed in this
way - granted that the conjunction is in general syllogistic.
20 It is worth observing how a reduction to the impossible, which
takes an external hypothesis in order to prove the point at issue, can
be syllogistic;66 for a syllogism is an argument which needs nothing
external for the generation of the necessity (as Aristotle said when he
denned the syllogism67). Perhaps in a reduction to the impossible the
point at issue is in fact proved and justified by means of the
25 suppositions, while the hypothesis proves something else and is
assumed with a different conclusion in view. This conclusion is
proved to be impossible by means of the suppositions and nothing
else; when it is rejected, the hypothesis is rejected together with it;
and when the hypothesis is rejected, the point at issue is justified. It is
the items by means of which the hypothesis is proved to be impossible
which prove that what is called the conclusion68 follows by necessity
from the premisses which have been assumed. Here the conclusion
30 was that N does not hold of every O; and this was proved by the
suppositions that M is predicated of every N and not of every O.
Note the following point. In the second figure, in the syllogistic
85,1 combination which contains the particular negative, the major must
be a universal affirmative and the minor a particular negative. It is
only when the terms stand thus69 that the proof by means of the
impossible can proceed - and it is only in this way that this
combination is proved to be syllogistic. If the terms are taken the
other way about, proof by means of the impossible no longer has any
5 force; and that is why the combination is no longer syllogistic when it
is taken in this way.70 Now in the third figure, the sixth syllogistic
combination also has a particular negative for one of its premisses
373,16-35; Diogenes Laertius, VII 78. See Kneale and Kneale, p. 158; Frede (1974b),
p. 102; Barnes (1990b).
64
The negative particle ou is omitted by LM and the Aldine, and von Arnim prints
oun in its stead (SVF II 264). But the Stoics did not regard subsyllogistic arguments
as genuine syllogisms, and the negative particle is required.
66
For this common criticism of Stoic logic see e.g. Alexander, in An. Pr. 372,29-30;
373,16-30; Galen, Inst. Log. iii 5.
66
See above, 24,18 note.
67
An. Pr. 24bl8-22; above, 21,10-23,2.
68
i.e. the conclusion of the syllogism whose validity the reduction is establishing:
not the conclusion of the hypothetical argument.
69
Retaining ekhontdn (Wallies emends toekhousdn).
70
See Aristotle, An. Pr. 27b4-6; below, 85,17-29.
 5. The Second Figure 155
and a universal affirmative for the other; and it too is proved by
reduction to the impossible. But here things are the other way about.
For in this figure the major premiss must contain the particular 10
negative and the minor the universal affirmative, inasmuch as in this
figure the major premiss must always be negative and the minor
affirmative if there is to be a syllogism. When the combination is of
this sort, it is proved to be syllogistic by means of the impossible.71 If
the premisses are the other way about, the conjunction is non-
syllogistic, as we shall affirm and prove shortly.72

5.5.3 Non-syllogistic combinations

5.5.3.1 Major universal, minor particular

5.5.3.1.1 Aristotle's proofs

If it is predicated of every O and not of every N, there will not be 15
a syllogism. [1.5, 27b4-5]
He has shown which conjunctions are syllogistic when the premisses
are dissimilar both in quality and in quantity. Now he sets down the
non-syllogistic combinations and shows which they are when the
premisses are dissimilar in both the ways mentioned.
We said that if there is to be a syllogism in the second figure, then 20
the major premiss must contain the universal.73 So every
combination not of this sort is unreliable and non-syllogistic.
He proves the combination before us to be non-syllogistic by way
of the following terms: for 'N holds of every O', animal, substance,
raven (animal does not hold of every substance and holds of every
raven; and substance holds of every raven); for 'N holds of no O', 25
animal, white, raven (animal does not hold of every white and holds
of every raven; and white holds of no raven).
The combination is non-syllogistic because there are parts of the
major which74 fall outside and do not share in the middle term, and
with these it may non-syllogistically include the minor term and
also not include it.75

Nor when of no O and of some N. <Terms for holding: animal, 86,1

substance, unit. For not holding: animal, substance, science. >
[1.5, 27b6-8]
71
Bocardo: Aristotle, An. Pr. 28bl5-20; below, 103,9-25.
72
Below, 104,13-106,2, on ATI. Pr. 28b22-4.
73
See above, 71,22-6.
74
Retaining kata ta moria (kata ha moria [sic] Wallies), and adding ha after autou.
76
See above, 56,27 note.
 156 Translation
In this combination too the major premiss is again particular. Terms
for 'N holds of every O': animal, substance, unit (animal is middle).
Animal holds of some substance and of no unit; and substance of
5 every unit. (He uses a doctrine of the Pythagoreans, who think that
units are substances - a little earlier he used number in the same
way.76 But the proof is the same if you take stone instead of unit.)
Terms for not holding: animal, substance, science. Animal holds of
some substance and of no science; and substance of no science.
10 The reason why this combination is non-syllogistic is the same.

Thus when the universal is opposite to the particular, <we

have said when there will be a syllogism and when not. When
the premisses are similar in form - i.e. both privative or
affirmative - there will in no case be a syllogism.> [1.5,
27b9-12]
For the cases where the premisses are dissimilar in form77 both in
quality and in quantity (he means difference in quality when he
says Thus when the universal is opposite to the particular'), he has
15 proved which combinations are syllogistic and which non-syllogistic.
The combinations in which the major premiss is a universal
negative and the minor a particular affirmative, and in which the
major is a universal affirmative and the minor a particular negative
- these are both syllogistic. The others are non-syllogistic.
Next he proves that, just as the combinations of universals similar
20 in form are non-syllogistic, so too it is with the combinations which
have one premiss universal and the other particular (where the
universal is no longer opposite to the particular). For the
combinations in the second figure which are similar in form are all
of them non-syllogistic, however they stand with regard to quantity.
He makes clear which combinations are similar in form - namely
25 those which are similar in quality - by saying 'i.e. both privative or
affirmative'. He takes as the first combination of premisses similar
in form the one which has a universal negative major and a
particular negative minor.

<Let them first be privative, and let the universal be supposed

on the major extreme.> I.e. let M hold of no N and not of some
O. Now it is possible for N to hold of every O and also of no 0.
<Terms for not holding: black, snow, animal.> [1.5, 27bl2-16]
76
Aristotle, An. Pr.
27a20; above, 81,17-28.
77
'Dissimilar in form [anomoioskhemon]' means 'different in quality' (see
Introduction, p. 27): either Alexander uses the word loosely here, or else he wrote
anomoion (as he does in the parallel sentence at 85,17).
 5. The Second Figure 157
He undertakes to refute the conjunction and prove it non-syllogistic 87,1
by setting down terms; for it is possible to find terms for this
combination to prove that N holds of every O and also of no O.78 For
'N holds of no O' he takes the terms black, snow, animal; for black
holds of no snow (which is the major extreme, answering to N), and
does not hold of some animal (which is the minor extreme, 5
answering to O); and snow holds of no animal.

<It is not possible to take terms for holding if M holds of some

O and not of some O. For if N holds of every O and M of no N, M
will hold of no O; but it was supposed to hold of some. Thus in
this case it is not possible to take terms - it must be proved
from the indeterminacy. For since 'M does not hold of some O'
is true even if it holds of none, and since there was no syllogism
when it held of none, it is evident that there will not be one in
the present case either.> [1.5, 27bl6-23]
Then he turns to take terms for 'N holds of every O' (for this is how
the combination was to be proved non-syllogistic). He first draws our
attention to the fact that if the particular negative, MO, is taken to
be true in one way, then it is possible to obtain terms for 'N holds of
every O', but that if it is taken in another sense it is impossible. For
particular negatives, just like affirmatives, are true both inasmuch 10
as the universal propositions over them are true and also inasmuch
as their subcontraries are true.79 The proposition saying that M
does not hold of some O is true even if M holds of no O; for it is
ranked under this proposition and follows it and is subaltern to it. It
is also true if the proposition saying that M holds of no O is false and
the proposition positing that M holds of some O is true. (This is 15
subcontrary to the proposition saying that M does not hold of some
O; for they both fall under the universals which are contraries of
each other.80)
Since, then, particular negatives are true in two ways, if the
premiss saying that M does not hold of some O is taken as true
because M also holds of some O, we will not be able to obtain terms
proving that N holds of every O. The reason for this is that we have 20
proved that the combination saying that M holds of no N and of
some O is syllogistic and deduces a particular negative; for it is
necessary for N not to hold of some O.81 But if a particular negative
is deduced syllogistically and by necessity when M does not hold of
78
Deleting to asullogiston at 86,30.
79
On the indeterminacy of particular negatives see above, 66,1-67,24 (where the
argument is closely parallel to that of the present section).
80
On subcontraries see above, 45,24-5.
81
i.e. Festino: Aristotle, An. Pr. 27a32-6; see above, 82,25-83,8.
 158 Translation
some O in such a way that it also holds of some O, then it is clear
25 that we will not obtain terms which predicate N of every O. For in
this way, we should reject the fact that a particular negative is
deduced by necessity in the combination in which M holds of no N
and of some O, since we are now assuming that M does not hold of
some O in such a way that it also holds of some O.
This is the reason why it is not possible to take terms for 'N holds
30 of every O', if we take MO to be true in this way. Aristotle proves by
means of the impossible that if it is supposed that the particular
negative MO is true in this way, it will be impossible for N to hold of
88,1 every O. For if it is supposed that N holds of every O, then as it is
supposed that M holds of no N, we get in the first figure that M holds
of no O. But this is impossible; for it was hypothesized that M does
not hold of some O in such a way that it also holds of some O. So if we
take the particular negative in this way, it is not possible to obtain
terms for 'N holds of every O'.
5 But 'M does not hold of some O' is true not only when M also holds
of some O, but also when it holds of none; for the truth of particular
negatives is indeterminate. Thus if 'M does not hold of some O' is
taken as true in such a way that it holds of none, i.e. if we take a
material instance in which M holds of no O (for the particular
10 negative is no less true in this case), then we will obtain terms to
prove that N holds of every 0. For we have proved that nothing is
deduced from two universal negative premisses.82 Terms for 'N
holds of every O': line for M, animal for N, man for O. Line holds of
no animal and does not hold of some man (since it holds of no man);
and animal holds of every man.

<Again, let them be predicative and let the universal be

supposed in the same way. I.e. let M hold of every N and of
some O. ...> [1.5,27b23-5]
15 He uses the same proof when both premisses are affirmative, the
major MN being universal and the minor MO particular. He obtains
terms for holding of none, positing white for M, swan for N, stone for
O. White holds of every swan and of some stone, and swan of no
20 stone. But not for holding of all - if the particular affirmative has
been assumed in such a way that it is true together with the
particular negative. For, as has been proved, if the minor is again a
particular negative and the major a universal affirmative, a
particular negative conclusion can be deduced syllogistically in this
figure.83 Thus since 'not of some' is true in this combination, it is
82
See Aristotle, ATI. Pr. 27a20-3; above, 82,1-7.
83
i.e. Baroco: Aristotle, An. Pr. 27a36-bl; above, 83,13-84,3.
 5. The Second Figure 159
impossible to obtain terms for 'N holds of every 0'. For in that case 25
the syllogistic combination will be cancelled and disproved: if it
holds of all, it will also hold of that of some of which it was proved
syllogistically not to hold. So if the particular affirmative is true in
this way, it will again not be possible to obtain terms for 'N holds of
every O'. For if N holds of every O, and M of every N, M will hold of
every O; but it was supposed that it does not hold of some. For the 30
particular affirmative MO was true in this way.
But the truth of particular affirmatives is indeterminate; for they
are true not only when their subcontraries are true as well, but also
when the universal affirmatives under which they fall are true. So
when it is assumed to be true in this way, i.e. because of the
universal under which it falls, the combination will be disproved in 89,1
the case of such terms - as it has already been disproved. For it was
proved that nothing is deduced in this figure from two universal
affirmatives.84 Hence for this reason nothing will be deduced, if the
one premiss is taken as a universal affirmative and the other as a
particular affirmative. For the particular affirmative is true in both 5
cases, both when the universal affirmative is true and also when its
subcontrary is true. And to prove that a combination is non-
syllogistic, it is sufficient to find some cases and some material
instances which prove that it holds of all and also of none. Terms for
'N holds of every 0': substance, animal, man. Substance holds of
every animal and of some man (since it holds of every man).
He invoked the indeterminacy of particular propositions earlier, 10
in the first figure, when he proved that the combination with a
universal negative major and a particular negative minor is
non-syllogistic.85 But there he also refuted the combination by
exposition, while here he was content with the indeterminacy of the
particular. However, in the case of the conjunctions we have just 15
mentioned, you can also prove through terms and by exposition that
they are non-syllogistic.86
If both premisses are negative, take as terms science, vice,
disposition. For science belongs to no vice and does not belong to
some disposition. And for those parts of disposition of which science
does not hold, take courage and cowardice. (For the virtues are not
sciences.) Thus vice holds of no courage and of all cowardice. If it is 20
assumed that virtue holds of no vice and does not hold of some
disposition, and if for those parts of disposition of which virtue does
not hold medicine and cowardice are taken, then vice holds of all
cowardice and of no medicine.
84
See Aristotle, Are. Pr. 27al8-20; above, 81,3-82,1.
85
Aristotle, An. Pr. 26bl4-20; above, 66,1-29.
86
i.e. Aristotle's argument at An. Pr. 26blO-16 (see above, 64,1-65,22) can be
applied to the present case.
 160 Translation
If both premisses are affirmative, let the terms be disposition,
science, quality. For disposition holds of every science and of some
quality. And for those parts of quality of which disposition holds,
25 take self-indulgence and grammar. For science holds of all grammar
and of no self-indulgence. The same argument holds also if
substance is taken to hold of every animal and of some white. Take
swan and snow for those whites of which substance holds. For
animal holds of every swan and of no snow.

5.5.3.1.2 The view of Herminus87

Aristotle, then, disproved the combinations we have just described
30 in this way: they are non-syllogistic inasmuch as one can obtain
material instances for 'N holds of every O' and also for 'N holds of no
O'. He thinks that this is sufficient proof that the combinations are
non-syllogistic.
Some think that you can disprove the combinations in question if
you prove that N holds of no O and also of some O, since these are
90,1 contradictories. Some of the earlier thinkers say this, and so does
Herminus. (Herminus says that if you can prove that a pair of
contradictories can be deduced, it is reasonable to call the
combination no less non-syllogistic than one in which contraries can
be deduced. For contradictories, like contraries, cannot hold at the
same time.) As terms for 'holds of some', they set down inanimate,
animate and fleshly body.88 For inanimate holds of no animate, and
5 does not hold of some fleshly body - or simply body -; and animate
holds of some fleshly body - or simply body. But this is not at all
sound, nor is it sufficient for the disproof of a combination.
It will be conceded that neither of the contradictories can be
deduced syllogistically, because they cancel each other. However,
one may say that - as far as these suppositions go - a particular
negative can still be deduced. For the particular negative is
10 cancelled by neither of the contradictories in question. The items set
down to refute a non-syllogistic combination must not only not be
able to hold at the same time as each other: they must also cancel
everything which can be proved syllogistically. But how is the
particular negative disproved, if you do not obtain a material
instance for 'holds of every"?89
87
On this section see Patzig, p. 177; Moraux (1973/84), II, pp. 385-94.
88
i.e. they accept Aristotle's terms for 'N holds of no O' and supply their own terms
for its contradictory, 'N holds of some O'.
89
The essence of Alexander's point is this. You may show by material instances
that a combination is compatible both with 'A holds of no C' and with 'A holds of some
C'. But this leaves open the possibility that the combination entails 'A does not hold of
some C'. This is correct - but Alexander mars it by making his characteristic error
(see Introduction, pp. 12-13).
 5. The Second Figure 161
Further, in combinations which are agreed to be syllogistic you
can find contradictory pairs being deduced. Either, then, we will call 15
these combinations non-syllogistic as well, or else the present
combination is not non-syllogistic on that count. For example, in the
first figure, where the combination of a universal negative major
and a particular affirmative minor with a particular negative
conclusion is syllogistic, you can obtain terms for holding of none
and also for holding of some. For holding of none:90 man holds of no 20
horse, horse holds of some quadruped, man holds of no quadruped.
For holding of some: snow holds of no swan, swan of some white,
snow of some white. (In the combination with a universal
affirmative major one can take terms for holding of all and also for
not holding of some. For holding of all: substance, animal, man. For
not holding of some: substance, animal, white.) Moreover, if we take
the combination before us, i.e. the one with a universal negative 25
major, and hypothesize that A holds of some C,91 nothing impossible
meets us;92 hence 'holds of some' is not cancelled. But it has been
proved that A holds of no C.93
But surely, they say, proof by means of the impossible will be
rejected if opposites can be proved in a syllogistic combination? For
we shall no longer deduce only the proposition whose opposite is 30
impossible we shall also deduce something else, the opposite of
which can also be proved. But if this is so, then proof by means of the
impossible will no longer have any standing. For if in the
combination which deduces a particular affirmative in the first
figure it is also possible to deduce both a universal affirmative and a 91,1
particular negative, but neither of these can be proved by reduction
to the impossible by assuming and hypothesizing the opposite of the
particular affirmative,94 then the deduction95 will be disproved as
being useless.
Again, it seems that the conclusion of every syllogistic combi-
90
Adding tou men medeni before anthropos.
91
Reading to A tini toi G (with M and the Aldine) for to prdton tini toi tritoi.
92
Take as premisses 'A holds of no B' and 'B holds of some C', and hypothesize the
contradictory of the putative conclusion, viz. 'A holds of some C'. Then you will be able
to infer 'B does not hold of some C'. But this is not 'impossible', i.e. it is consistent with
'B holds of some C'.
93
The train of thought is obscure. We guess that 'the combination before us' is 'A
holds of no B, B holds of some C' (from which, as Ferio, we may deduce 'A does not
hold of some C'). Alexander's point is this: the combination is compatible with 'A holds
of some C'. Yet we have shown that you can deduce 'A holds of no C' from the
combination. Hence you can get contradictories from it. (Of course, we have not
shown that 'A holds of no C' can be deduced from the combination. But Alexander
thinks that we have.)
94
Removing Wallies' parentheses at 91,1 and 3.
95
Alexander must mean the reduction to the impossible; and perhaps we should
emend sunagoge to apagdg.
 162 Translation
5 nation can also be proved by means of reduction to the impossible.
But this is not possible in a combination in which opposites are
deduced. For in the case of such a combination neither of the
opposites is impossible. Either, then, no syllogistic conjunction
deduces syllogistically, or else the opposite cannot also be true.
But this is not what we have said. Hence proof by means of the
10 impossible is not rejected. For the only thing deduced syllogistically is
that of which the opposite is impossible. Nonetheless it is not
impossible that there should be other propositions, distinct from
what can be proved syllogistically and opposite to one another, and
that in certain material instances there should be a deduction - but
no longer a syllogistic deduction - from such propositions: if they are
hypothesized, nothing impossible turns out to be deduced. For in the
deduction of the particular affirmative it is impossible to obtain terms
15 for its opposite, i.e. for holding of none. However, it is not impossible
to obtain terms for holding of all and for not holding of some: neither
of these is opposite to 'of some', nor is it proved to be syllogistically
deduced. Neither96 when the universal affirmative nor when the
particular negative is hypothesized does it follow that anything
impossible is deduced. It is only if the opposite of the particular
affirmative is hypothesized that something impossible follows. For
20 this was what followed syllogistically.
The proof, then, which Herminus uses is not sufficient to disprove a
combination and to assert that it is non-syllogistic. Hence Aristotle is
right to refute the non-syllogistic combinations by setting down only
what holds of all and of none. For here nothing can be proved
syllogistically inasmuch as, whatever you assume to be syllogistically
25 deduced, you will find that its opposite, which cancels it, is true in the
case of some material instances. In general, consider syllogistic
combinations with a particular conclusion. Particulars are true in an
indeterminate way; for a particular is true both with the universal
under which it falls and with its own subcontrary. Thus we can
assume it to be true in either way: sometimes we will obtain terms
30 which prove the universal under which the particular we have
assumed falls; sometimes we will obtain the subcontrary of the
particular which we have assumed and deduced, which does not in
any way cancel its particular subcontrary.

5.5.3.2 Minor universal, major particular

92,1 If the universal is on the minor extreme, and M holds of no O and
not of some N ... [1.5, 27b28-9]
Having proved non-syllogistic the combinations in the second figure
96
Omitting the gar which Wallies adds.
 5. The Second Figure 163
which are similar in form and have a particular on the minor and a
universal on the major, he now takes up those which are similar in 5
form and have, conversely, a particular on the major and a universal
on the minor; and he proves them to be non-syllogistic by setting
down terms. These combinations in this figure are non-syllogistic on
two counts: both because of the similar form of the premisses and
because the major is assumed as particular. And that these are not97
syllogistic he proves simply by setting down terms, this time not 10
invoking the indeterminacy of the particular. For in whatever way
the major, being particular, is true - whether because of the universal
or in its own right (inasmuch as the particular affirmative which is
subcontrary to it is true) -, the combination will be non-syllogistic
inasmuch as there is no syllogism in this figure when the premisses 15
are similar in form.
As terms for 'N holds of every O' he sets down white, animal, raven:
animal holds of every raven. For 'N holds of no O': white, stone, raven
for stone holds of no raven. (White holds both of some animal and of
some stone, but of no raven.)

<If the premisses are predicative ...> [1.5, 27b32-3]

He conducts the proof similarly when both premisses are taken as
affirmative and the major as particular. For not holding, the terms 20
are white, animal, snow. Animal holds of no snow, while white holds
of some animal and of all snow. For holding: white, animal, swan.
Animal holds of every swan, while white holds of some animal and of
every swan.

5.6 Combinations of particular premisses

Nor if it holds of some of each, or does not hold, <or of some of
one and not of some of the other, or of all of neither, or
indeterminately.> [1.5,27b35-8]
He has discussed all the combinations which have one universal and 25
one particular premiss, and has proved that only two among them
are syllogistic. (Those which have a universal major and a minor
opposite in quality have been proved to be syllogistic,98 and all the
others six in number to be non-syllogistic.99) He now turns to
combinations of two particulars. He discusses them all at the same 30
time and proves all of them to be non-syllogistic by means of the

97
Retaining mede (medenos Wallies).
98
Aristotle, Are. Pr. 27a26-b2.
99
Aristotle, An. Pr. 27b3-36.
 164 Translation
same examples, just as he did in the first figure.100 And he counts
the indeterminates along with the particulars as being, like them,
non-syllogistic - they are equivalent to them101 and are refuted by
the same terms.
93,1 Having said, 'Nor if it holds of some of each or does not hold (sc. of
some of each)', he adds 'or of all of neither'. This differs only in
expression from 'does not hold of some of each'. For both expressions
betoken particular negatives.102
5 Terms for holding of every: white, animal, man. For holding of no:
white, animal, inanimate. For white holds of some animal and of
some man, and also does not hold of some; and it holds of some of one
of them and does not hold of some of the other; and it holds of each
indeterminately, and also does not hold (and complementarily); and
indeterminately of the one and as a determinate particular of the
10 other. And animal holds of every man. Again, white similarly holds
of animal and inanimate - through all the particular and
indeterminate permutations; and animal holds of no inanimate.

5.7 Concluding remarks

Thus it is evident from what we have said that if the terms
stand to one another in the way we have said, there is a
syllogism by necessity, and that if there is a syllogism it is
15 necessary for the terms to stand in this way. [1.5,28al-3]
We must again103 understand 'in the second figure'. For it is not the
case simpliciter that if there is a universal negative or a particular
negative syllogism, the premisses must stand in this way. For there
are syllogisms of these sorts in the first figure104 and not only in the
second figure.

<... they are all perfected when certain items are co-assumed
20 which> either inhere by necessity in the terms or are posited
as hypotheses <- i.e. when we prove by means of the
impossible ....> [1.5,28a5-7]
What 'inhere by necessity in the terms' (or in the premisses
supposed) are the converses, which we have invoked for the
perfecting of those syllogisms where the proof is by way of
100
Aristotle, Are. Pr. 26b21-5; above, 68,9-21.
101
See above, 30,31 note.
102
See above, 62,12 note.
103
i.e.'as we did at Are. Pr. 27a23': see above, 82,8, where the point is not explicit
but is implicit in Alexander's paraphrase. (See above, 58,10 and 60,9, on the first
figure, and below, 108,22 on the third.)
104
i.e. Celarent and Ferio (and also, in the third figure, Bocardo and Ferison).
 5. The Second Figure 165
conversion. For the converses of the premisses inhere in the 25
premisses which are assumed.105
What are 'posited as hypotheses' are the hypotheses we use for the
proof of imperfect syllogisms in reductions to the impossible, as
Aristotle has made clear. For we take as an hypothesis the opposite
of what we are proving and thus produce a syllogism and a
deduction of the impossible. Then by means of this we reject the
hypothesis and establish its opposite, which was the point at issue. 30

<... there is no affirmative syllogism in this figure, but... all of

them ... are privative.> [1.5,28a7-9]
It is clear too that all the conclusions in this figure are negative.

106
cf. above, 58,13-22.
 6
The Third Figure

6.1 General remarks

94,1 When one term holds of all and the other of none of the same
term, <or both of all or none, I call this the third figure.> [1.6,
28alO-12]
He turns to the third figure, and makes clear what it is and what
position the middle term has in it. He does this, again,1 by example,
taking the combinations of universal premisses in this figure. In it
2
5 the middle is subject for both the terms. This, as we said, is the
third figure.
It is third, and comes last in order, being ranked thus for several
reasons. (1) The middle term, which is the reason why the extremes
share something in common, is ranked last in order in this figure.3
(2) No universal is deduced in it.4 (3) Sophistical syllogisms, which
conclude to indeterminates and particulars, occur in this figure
10 especially, and sophistical syllogisms are the last sort of syllogism.5
(4) The second and third figures being each generated from the first,
the third is generated from the inferior premiss. For each of the two
premisses in the first figure, when converted, produces one of these
figures, as we have already said - when the major is converted, the
6
15 second figure is generated and when the minor the third. Hence if it
is generated from the inferior premiss in the first figure (for the
minor is inferior qua minor7), then the figure itself will reasonably
be last.
In this figure, as in the one before it, there are sixteen
combinations in all, not counting the combinations of indeterminate
20 propositions. (We have omitted them as being equivalent to the
1
See above, 52,30, for the same procedure in the second figure.
2
Above, 47,4-11.
3
See above, 47,21-4.
4
See above, 49,1.
5
See above, 49,14-18.
6
On the generation of the figures see above, 47,27 note.
7
The same point, implicitly, at 48,15.

166
 6. The Third Figure 167
8
particulars. ) But there are more syllogistic combinations in this
figure - there are six, whereas each of the earlier figures contained
four syllogistic combinations.
The reason is this. In the third figure the minor premiss must
again9 by necessity be affirmative if there is to be a syllogism (if it is
negative, there is no syllogism in the third figure); but so long as it is 25
affirmative, there is a syllogism no matter how the major is
assumed, provided only that we ensure that the premisses are
neither both negative nor both particular. For in no figure is there a
syllogism from two particulars or from two negatives.10 (The major
and the minor premisses will be grasped in the same way in this
figure as they are in the second.11) Now if these conditions are 30
observed, there are six syllogistic combinations. Hence this is the
number of syllogistic combinations in this figure. If the minor is a
universal affirmative, there will be four syllogistic combinations, the
major being taken either as universal, whether affirmative or 95,1
negative, or as particular, whether affirmative or negative; and if
the minor is a particular affirmative, there are two further
syllogistic combinations, the major being universal and either
affirmative or negative.
The reason why there are more syllogistic combinations in the 5
third figure is this. In the other figures more features were
determined, and these had to be preserved in order for there to be a
syllogistic conjunction. Both in the first and in the second figure two
features were determined. In the first figure, the major premiss was
determined (it had to be universal) and so was the minor (it had to
be affirmative).12 In the second figure the major was determined (it
had to be universal), and it was also determined that the two 10
premisses must not be similar in form (it is the proper characteristic
of the second figure that nothing is deduced from two affir-
matives).13 But in the third figure only the minor premiss is
determined - it must be affirmative. (The fact that the premisses
may not be both particular or both negative is common to the three
figures and not a proper characteristic of the third.) And what is less
determined is more general and has a wider extension.
Given that the first figure has both its premisses determined, the
major in quantity (it is universal) and the minor in quality (it is
affirmative), it stands to reason that the other figures (I mean the 15
second and the third) should preserve the proper characteristic of
8
See above, 30,31 note; 51,25 note.
9
i.e. as well as in the first figure: above, 49,25.
10
See above, 71,32-3.
11
i.e. the major premiss is the premiss containing the major term, and the major
term is the term which is predicate in the problem: see above, 75,19-34.
12
See above, 49,19-50,14.
13
See above, 71,22-33.
 168 Translation
the premiss by the conversion of which each of them was
14
20 generated. The second figure was generated by conversion of the
major, and it preserves the proper characteristic of that premiss in
syllogistic combinations15 - in the second figure, too, the major
premiss is universal in the syllogistic conjunctions. The third figure
was generated by conversion of the minor and retains its proper
characteristic16 - for in this figure the minor is again determined, as
we said, inasmuch as it is affirmative.

6.2 Third figure syllogisms

17
25 Of the six syllogisms in this figure, the first in order deduces a
particular affirmative from two universal affirmatives by conversion
of the minor premiss. It can also be generated by converting the
major, but in that case the conclusion too will have to be converted.
30 For this reason, some add this syllogism, as being different from its
predecessor, and say that there are seven syllogisms in this figure.18
This conjunction is first because each of its premisses is both
universal and affirmative.
One might raise the following question. In the second figure there
14
Reading to alia skhemata, to te deuteron ... kai to triton for ton allon skhematon,
tou te deuteron ... kai tou tritou.
15
See above, 71,24-6.
16
See above, 94,15.
17
On the order of the syllogisms in the third figure see below, 97,11 note. Compare
above, 51,8-24 and 76,29-77,31, on the order in the first and second figures.
18
What is 'this syllogism' which some people added to the third figure? (1)
Alexander's text here, and also his ensuing discussion, indicates that it is Darapti, i.e.
that some people thought that the form 'A holds of every B, C holds of every B:
therefore, A holds of some C' actually marked out two syllogisms. And Alexander
further implies that this view was based on the fact that the form can be reduced to
Darii in two ways, either by converting the major premiss or by converting the minor
premiss. (2) From other sources we learn that Theophrastus added a 'second Darapti'
to the third figure (Apuleius, Int. 189,19-27 = F 19 Graeser); he was followed by Galen
(Inst. Log. xi 7) and later by Porphyry (Boethius, Syll. Cat. 813C, 819B). See
Bochenski (1947), pp. 61-4; Sullivan, pp. 102-3. It seems clear from these texts that
the syllogism in question was Daraptis ('A holds of every B, C holds of every B:
therefore, C holds of some A'), which may be derived by converting the conclusion of
Darapti. (See above, 70,21 note.) It is a puzzle why Daraptis, which is a mere
notational variant on Darapti should have been added to the third figure, when
Cesares and Camestre (which are, in the same way, notational variants on Cesare and
Camestres) were - for that very reason - not added to the second. (As Apuleius puts it:
'It does not matter from which premiss you make the particular affirmative, since it
does not matter which you propound first.' Bochenski and Sullivan attempt, vainly, to
defend Theophrastus here.) (3) It is tempting to suppose that Alexander's remarks
reflect his understanding of the Theophrastean position, and that he thought that
Daraptis simply was a second Darapti inasmuch as it is a mere notational variant on
Darapti. We might further imagine that Theophrastus argued that Darapti is
genuinely different from Daraptis because the proof of the former is different from
the proof of the latter. But this - as Alexander plainly shows - is such a bad argument
that we might well be reluctant to ascribe it to Theophrastus.
 6. The Third Figure 169
were two syllogistic conjunctions having a universal negative as one
premiss and a universal affirmative as the other, inasmuch as now
the major and now the minor, being universal negatives, convert.19 96,1
Why, then, in the third figure are there not said to be two syllogistic
conjunctions with two universal affirmatives, even though here too
now the minor and now the major may be converted? Well, in the
second figure, inasmuch as the premisses are dissimilar in form, the 5
universal negatives are different, so that by necessity a different
premiss - now the major and now the minor - is converted: it is not
up to us to convert whichever we wish. But in the third figure, when
the two premisses are universal affirmatives, it is not the way in
which the premisses are posited which explains why now one and
now the other of them converts; for this is similar and the same. 10
Rather, the choice is up to us, and it does not depend on the
combination. This is why, although each of the premisses can be
converted in the same way inasmuch as they are posited in the same
way, we take the premiss whose conversion yields the least
laborious proof, and we determine the syllogistic status of the
combination by way of this premiss.
Again, the differences among the syllogisms in each figure depend
on the differences in the combinations and not on differences in the 15
proofs. Thus the same conjunction can be proved syllogistic both by
conversion and by reduction to the impossible - and also by
exposition. But this does not mean that there are several syllogisms,
because there is one combination for which the different proofs are
made. Hence if there is one combination of two universal
affirmatives in the third figure, there is also one syllogism, even if 20
there are different proofs inasmuch as now one and now the other of
the premisses can be converted.
After this comes the syllogism which deduces a particular
negative from a universal negative major and a universal
affirmative minor. This has both premisses universal, but not both
affirmative. It too is proved to be deductive by conversion of the 25
minor.
The third deduces a particular affirmative from a universal
affirmative major and a particular affirmative minor. It too is
proved by conversion of the minor. It is prior in rank to the syllogism
after it, which deduces a particular affirmative from similar
premisses, because it is proved by one conversion to deduce the point 30
at issue, whereas the syllogism after it, with a particular affirmative
major and a universal affirmative minor, is proved by two
conversions to deduce the point at issue. (You must convert both the
major, which is a particular affirmative, and the conclusion.)
19
i.e. Cesare and Camestres: An. Pr. 27a5-14; above, 78,10-79,25.
 170 Translation
97,1 The fifth deduces a particular negative from a universal negative
major and a particular affirmative minor. It is proved by conversion of
the minor. It comes after the others because it deduces a negative and
it too has one particular premiss.
Sixth rank is held by the syllogism which deduces a particular
negative from a particular negative major and a universal affir-
5 mative minor. This syllogism is no longer proved by conversion, but
rather by reduction to the impossible - like the fourth syllogism in the
second figure, which has a universal affirmative major and a
particular negative minor.20 For this reason, just as that syllogism
was last in the second figure, so this is last in the third.
10 This is the ranking of the syllogisms in this figure, but Aristotle did
not follow it in his presentation - he changed it.21
The following point stands to reason. In the second figure the
syllogisms proved by means of conversion are proved by the conver-
15 sion of the major premiss. For the second syllogism, which has a
universal affirmative major and a universal negative minor, and
which comes about22 by conversion of the negative, in fact makes
what is supposed as the minor into the major, as far as the proof goes.
That is why we need to convert the conclusion too in this case, in order
that the term which was supposed in the premisses as major may be
20 predicated in the conclusion.23 In the third figure, on the other hand,
it is the minor which is converted. For in this figure, in the
combination of a particular affirmative major and a universal
affirmative minor, the particular affirmative, which is supposed as
major, is converted, and as far as the syllogism goes it in fact becomes
the minor. That is why, once again, in this combination too we had to
25 convert the conclusion too in order that the term supposed as major
might be kept as predicate in the conclusion. For as far as the
syllogism by means of conversion goes, it became the minor.
This, as I said, stands to reason. For each of these figures is
generated from the first figure by the conversion of a premiss; and
when this premiss is again converted, the figures are analysed into

20
i.e. Baroco: Aristotle, An. Pr. 27a36-b3; above, 83,12-84,3.
21
Alexander follows Theophrastus in placing Datisi before Disamis and Ferison
before Bocardo (see Philoponus, in An. Pr. 105,28-106,2; Boethius, Syll. Cat.
819A-821A; scholium to Aristotle, 155b8-19: - F 21 Graeser; cf. Bochenski (1947),
p. 65; Sullivan, pp. 123-6). The later tradition generally took the Theophrastean line
(e.g. Apuleius, Int. 190,25-191,5; Martianus Capella, IV 413). Galen (Inst. Log. x 1-6)
follows Aristotle in placing Disamis before Datisi, but like Theophrastus he has
Ferison before Bocardo.
22
Or 'is generated': see below, 97,30 note.
23
See above, 78,25-79,19. Here Alexander seems to treat the proof of Camestres as
consisting simply in the application of Celarent, which yields the conclusion 'O holds
of no N'. The conversion of this conclusion is then superadded to the proof proper.
 6. The Third Figure 171
and reduced to the first figure - and by the analysis they are proved 30
to be syllogistic.24

... I say that in it the middle is that of which both are

predicated,.... [1.6, 28al2]
This is an account of the third figure, in which both extremes are
predicated of the middle.

... the major is the extreme further from the middle, <and the 98,1
minor is the one nearer.> [1.6,28al3-14]
Since the major premiss in the first figure was converted, the middle
term came to be posited before the major extreme in the second
figure.25 In the same way, since the third figure has been generated
by converting the minor premiss in the first figure, the middle term 5
in the third figure comes to have position after the minor extreme.
For by the conversion of each premiss the middle term falls in the
position near the extreme with which it is combined in the converted
premiss. In the third figure the middle term comes to be near the
minor extreme and becomes subject for both terms: for the minor
because of the conversion of the minor premiss, for the major 10
because it was supposed from the beginning in the first figure to be
its subject. Being subject for both, it reasonably holds the last
position.26

6.3 Combinations of universal premisses

6.3.1 Darapti
... but there will be a potential one both when the terms are
universal and when they are not universal in relation to the
middle. [1.6, 28al6-17]
'Not universal' does not mean 'neither of them being universal' (for it 15
24
On the generation of the figures see above, 47,27 note. Here Alexander, in trying
to correlate generation with analysis, conflates figures with syllogisms: the figures,
properly speaking, are generated but not analysed or proved (it makes no sense to
speak of proving a figure); the syllogisms, properly speaking, are analysed or proved
but not generated (cf. Themistius, Max. 184).
26
See above, 72,15 note.
26
Alexander tacitly alludes to the standard diagrams (see above, 72,11 note). In
the first figure the terms are set out thus:
A-B-C.
By converting the minor premiss, i.e. by inverting the order of B and C, we get:
A-C-B.
And A, the major term, is further from B than C is.
 172 Translation
is impossible, as we have already said,27 to get a syllogism from two
particular propositions). It means 'not both being universal'.

If they are universal, then when both P and R hold of every S,

<P will hold of some R by necessity. For since the predicative
converts, S will hold of some R. Hence since P holds of every S
and S of some R, it is necessary for P to hold of some R. For we
get a syllogism by way of the first figure. It is also possible to
effect the demonstration by means of the impossible and by
exposition. For if both hold of every S, then if something of S -
say N - is taken, both P and R will hold of this. Hence P will
hold of some R.> [1.6,28al7-25]
20 In this figure he again uses letters28 - this time P, R, S, where P
signifies the major extreme, R the minor, which should be subject in
the conclusion, S the middle term. First he sets out the combination
of two universal affirmative premisses. To prove that this is
25 syllogistic, he converts the minor premiss RS, which is a universal
affirmative, and assumes RS as a particular affirmative; and in this
way he reduces it to the third syllogism of the first figure, which
deduces a particular affirmative from a universal affirmative major
and a particular affirmative minor.
As I said,29 it is also possible to prove that the same thing is
deduced by converting the major premiss, if we also convert the
99,1 conclusion. For if P holds of every S, S holds of some P. But it is also
supposed that R holds of every S. It follows that R holds of some P.
Now if we convert this too, P will be said of some R. And this was to
be proved. Aristotle does not set this proof down, because it is
inferior to the one which we have already discussed, since it needs
two conversions.
5 He also uses reduction to the impossible to prove that in this
combination a particular affirmative can be deduced. For
hypothesize the opposite of'P holds of some R' - i.e. 'P holds of no R'.
It is also supposed in the combination that R holds of every S. It is
deduced that P holds of no S according to the second syllogism in the
first figure, which deduces a universal negative from a universal
10 negative major and a universal affirmative minor. But it is
impossible for P to hold of no S; for it was supposed in the
combination before us that it holds of every S. Therefore the
hypothesis on which the impossible conclusion depended is false.
But if the hypothesis positing that P holds of no R is false, then its
15 contradictory opposite, which posits that P holds of some R, is true.
27
Most recently at 94,27-8.
28
See above, 53,30 note.
29
See above, 95,25-31.
 6. The Third Figure 173
He says that it is possible to prove that the conclusion is a
particular affirmative by reduction to the impossible; but he refrains
from setting down the proof itself since it is already familiar - he
used it in the second figure.30 But he does add a third kind of proof,
which can also be invoked to prove that in the combination before us 20
it can be deduced that P holds of some R. He calls this a proof 'by
exposition', and he shows in outline what the method of exposition
is.31 Since it is supposed that both P and R hold of every S, if instead
of S we take one of the items under S, it is clear that both P and R
hold of this, since they hold of everything under S. In this way, it 25
will be proved that P holds of some R. He takes N: given this, he
says, P will hold of some R.
But it seems that in this way nothing is gained with regard to
proving the point at issue. For what is the difference between
assuming that both P and R hold of every S, and taking it that they
hold of some part of S, viz. N? The case remains the same if N is 30
taken; for the combination is the same, whether they are each
predicated of every N or of S. Or is this not the proof he uses? For the
method of exposition proceeds by perception. He does not tell us to
take something of S, of all of which both P and R are said (if we did
that we would gain nothing), but to take something which falls
within our perception and which is evidently both in P and in R. For 100,1
example, suppose that animal is taken for P and holds of every man,
which is S, and rational is taken for R and also holds of every man.
Then if we take some perceptible item of S - i.e. some man, say
Socrates -, then inasmuch as it is evident and perceptible that he is 5
both animal and rational, it becomes obvious that P, i.e. animal,
shares in and holds of some R, i.e. rational.32
That proof by exposition is perceptual is indicated first by the fact
that if it is not taken in this way, there will be no proof. Next, there
is the fact that in the case of N, which is something of S, Aristotle no
longer says that both P and R hold of every N,33 but simply posits 10
that P and R hold of it. Note also that he does not convert either
premiss. For in the case of things which are perceptible and one in
number, neither 'of every' nor the determinations in general are
appropriate; for the determinations of propositions have their place
with universals, and individuals are not universal.34
Proof by exposition is peculiar to the third figure because in this 15
figure there is a single subject for both extremes.35 For exposition
30
See Aristotle, An. Pr. 27a36-b3.
31
See above, 32,32-34,2.
32
On this paragraph see Lukasiewicz, p. 63; Patzig, pp. 159-60.
33
Deleting to ekkeimenon, which Wallies emends to toi ekkeimenoi.
34
See above, 65,27 note.
35
Excising ditton to meson esti <kai>.
 174 Translation
needs to be done from what is both a middle term and a subject.36
(Some people think that if they take some one item, which is
perceptible, then with respect to this item they will find the
predicates either linked with or separated from each other.37) But if
an item is taken from several subjects, it will not be one thing,
20 inasmuch as something different will be taken from each -
especially if the predicate is predicated affirmatively of one term
and negatively of the other. And in this way there is no evident and
perceptible sharing of the terms.
It can also be shown by the fact that everything deduced in this
figure is particular; for if you take some one item, i.e. something
perceptible, that is sufficient for conclusions of this sort.
25 Aristotle's proof that actual universal negatives convert is
patently of this type and not, as some think, by way of the particular
affirmative.38

6.3.2 Felapton
And if R holds of every S and P of none, <there will be a
syllogism that P does not hold of some R by necessity. There is
the same mode of demonstration, the premiss RS being
converted. It can also be proved by way of the impossible, as in
the earlier cases.> [1.6,28a26-30]
He has turned to the second combination, which has a universal
negative major and a universal affirmative minor. This too he
30 proves syllogistic by conversion of the minor, RS, which is a
universal affirmative. He takes RS as a particular affirmative and
reduces the syllogism to the fourth syllogism in the first figure. (This
was the syllogism which deduces a particular negative conclusion
from a universal negative major and a particular affirmative
minor.39)
It is clear that one can also prove this syllogism by reduction to
101,1 the impossible, as he says; and it is plain what impossibility is
V deduced. For if you hypothesize that P holds of every R, then since R
was supposed to hold of every S, it will be deduced that P holds of
every S - but it was supposed to hold of no S.
36
Aristotle also uses exposition for a second figure syllogism, viz. Baroco with two
necessary premisses (An. Pr. 30a5-14); but Alexander argues that this is a different
type of exposition (in An. Pr. 121,15-123,9).
37
Who are these people and what is their view? The style of reference leads us to
think that Alexander is mentioning a rival interpretation of exposition; but we are
unable to find anything here incompatible with what Alexander himself says. (It is
possible that Alexander has Boethus in mind: see Themistius, Max. 191-2 - but the
text is obscure.)
38
See above, 31,27-32,32.
39
Ferio: Aristotle, An. Pr. 26a25-8.
 6. The Third Figure 175
And, once again, you can also prove it by invoking exposition,
although Aristotle does not mention this. If we take some perceptible 5
item of S, then inasmuch as it is obviously in R and participates in it
but does not share with P, it will be evident that P does not hold of
some R. For it will not hold of what has been taken and set out, and
this is some item of R.

6.3.3 Non-syllogistic combinations

If R holds of no S and P of every S, <there will not be a syllogism. 10
Terms for holding: animal, horse, man. For not holding: animal,
inanimate, man.> [1.6, 28a30-3]
Having set out the syllogistic combinations with two universal
premisses, he now mentions those which are non-syllogistic. First
he sets down the combination with a universal negative minor and a
universal affirmative major. He shows that it is non-syllogistic by
setting down terms, proving that if this combination is supposed, P 15
may hold of every R and also of no R. Animal, which is P, holds of
every man, which is S; horse, which is R, holds of no man; and
animal holds of every horse, i.e. P of every R. Again, animal holds of
every man; inanimate, which is R, of no man; and animal of no
inanimate, i.e. P of no R. The reason why the combination is 20
non-syllogistic is again the same as in the first figure where the
major is a universal affirmative and the minor a universal
negative.40

Nor when both are said of no S <will there be a syllogism.

Terms for holding: animal, horse, inanimate. (Inanimate is the
middle.)> [1.6,28a33-6]
He also proves non-syllogistic the combination with two universal
negative premisses, which is the remaining combination of 25
universal premisses. Terms for 'P holds of every R': animal, horse,
inanimate. Neither animal nor horse holds of inanimate, which is S;
and animal holds of every horse, i.e. P of R. For holding of none:
man, horse, inanimate. Again, neither man nor horse holds of 30
inanimate, which is S; and man does not hold of horse.

Thus it is evident in this figure too ... [1.6,28a36] 102,1

He draws our attention to the point (which has already been made)
that in this figure two combinations of universal premisses will be

40
See Aristotle, An. Pr. 26a2-9; above, 56,7-27.
 176 Translation
syllogistic (he says which they are41) and two will be non-syllogistic.

6.4 Combinations of universal and particular premisses

6.4.1 Disamis andDatisi
5 If one term is universal in relation to the middle and the other
particular, <then if both are predicative, it is necessary for
there to be a syllogism, whichever of the terms is universal. For
if Rholds of every S and P of some S, it is necessary for P to hold
of some R. For since the affirmative converts, S will hold of some
P. Hence since R holds of every S and S of some P, R will hold of
some P. Hence also P of some R.> [1.6,28b5-ll]
Having discussed the combinations of universal premisses in the
third figure, he turns to those with only one universal. For this sort
of combination too, he proves which of them are syllogistic and
which non-syllogistic. He says first that there will be a syllogism if
10 both premisses are affirmative, whichever is universal, the major or
the minor. Hence there are again two syllogistic combinations here,
when both premisses are affirmative and one is universal.
First he sets out the one having the minor, RS, as a universal
affirmative and the major, PS, as a particular affirmative.42 (We
15 ranked this combination fourth, inasmuch as it needs two
conversions.43) If the particular affirmative, PS, which was the
major, is converted, we get'S holds of some P'. It is also supposed
that R holds of every S. By the third syllogism in the first figure it is
deduced that R holds of some P. But since P is supposed as major
and must be predicated in the conclusion, the conclusion too will be
20 converted. It was deduced that R holds of some P. Therefore P will
hold of some R. Thus it needed the conversion of the conclusion as
well, and not only of the premiss PS. Aristotle makes this clear by
adding 'Hence also P of some R' after he has proved that by the
syllogism we get the conclusion that R holds of some P.

25 Again, if R holds of some S and P of every S, <it is necessary for

P to hold of some R. There is the same mode of demonstration.
It is also possible to demonstrate by means of the impossible
and by exposition, as in the earlier cases.> [1.6, 28bll-15]
Now he takes the premisses the other way round. He has kept them
both as affirmative but has made the major universal. He proves
41
At An. Pr. 28a37-b4.
42
On Aristotle's proof of Disamis see Patzig, pp. 141-2.
43
See above, 96,28-34.
 6. The Third Figure 177
that this combination too is syllogistic by converting the minor
premiss, which is a particular affirmative, and by reducing the 30
combination to the third indemonstrable in the first figure, which
has a universal affirmative major and a particular affirmative
minor. (We ranked this combination third, since the point at issue 103,1
becomes evident and is proved by one conversion and not by two, as
was the case with the previous combination.44) He remarks of this
combination too that one can prove it by reduction to the impossible
and also by exposition. By adding 'as in the earlier cases' he made 5
clear that not only proof by means of the impossible but also proof by
exposition is common to all the syllogisms in this figure.45

6.4.2 Bocardo
If one is predicative and the other privative <and the
predicative is universal, then when the minor is predicative
there will be a syllogism. For if R holds of every S and P does
not hold of some S, it is necessary for P not to hold of some R.
For if of every, and R of every S, then P will hold of every S. But
it did not hold. It is also proved without the reduction if
something of S is taken of which P does not hold.> [1.6,
28bl5-21]
From premisses similar in form he moves on to premisses dissimilar
in form and the combinations constructed from them, first 10
preserving one of the premisses as a universal affirmative. He shows
which of these conjunctions are syllogistic combinations. Thus, if the
minor is universal and affirmative, it is clear that the major will be
particular and negative and the combination syllogistic. For if R
holds of every S and P does not hold of some S, it can be deduced that 15
P does not hold of some R. The proof can no longer proceed by
conversion; for the particular negative does not convert, and if we
convert the universal affirmative we will produce two particulars -
and from two particulars nothing necessary can be deduced in any
figure.46 Rather, he proves by reduction to the impossible that we 20
get a particular negative conclusion with this combination. For if,
when the premisses are supposed in the way we have said, someone
should not concede that P does not hold of some R, then let the
opposite be hypothesized, i.e. let P hold of every R. But it is supposed
that R holds of every S. Thus it will be deduced that P holds of every
S. But this is impossible; for it was supposed that P does not hold of 25
some S.
44
See above, 96,26-32.
45
It is hardly true that Aristotle 'made this clear'; but see below, 106,21.
46
Compare the similar remarks about Baroco, above 83,12-24.
 178 Translation
He says that you can prove this same conclusion to be deducible
even without reduction to the impossible, 'if (he says) something of S
is taken of which P does not hold'. He means again the method of
exposition. Quite rightly he has not taken one of the Ss of which R
holds but rather one of them of which P does not hold. For, since R
30 holds of every S and P does not hold of some, R will certainly hold of
any item of S of which P does not hold; but it is not necessary for P
not to hold of any item of S of which R holds - for it is possible to take
something of S of which P will hold. (It was supposed not that it
holds of no S but that it does not hold of some S; and what does not
104,1 hold of some may also hold of some.) But if you take one of the Ss47
(an individual thing) of which P does not hold, then since this is
certainly in R (for R was said of every S), P will not hold of some R.
It is possible to conduct the proof for this combination if you take
5 not some perceptible and individual item of S but rather some of S of
none of which P will be predicated. For then P will be predicated of
none of it and R of all, and it has been proved that such a
combination deduces syllogistically that P does not hold of some R.48
Note that he uses the method of exposition for negatives too.
10 We made this the final syllogistic combination since it is not
possible to prove its conclusion by conversion.49

6.4.3 Non-syllogistic combinations

When the major is predicative, there will not be a syllogism
<- i.e. if P holds of every S and R does not hold of some S.
Terms for holding of every: animate, man, animal. It is not
possible to take terms for holding of none if R holds of some S
and does not hold of some. For if P holds of every S and R of
some S, then P will hold of some R. But it was supposed to hold
of none. We must take terms as in the earlier cases. For 'does
not hold of some' is indeterminate, and it is true to say of what
holds of none that it does not hold of some. But if it holds of
none, there was no syllogism. So it is evident that there will not
be a syllogism>. [1.6,28b22-31]
He is still discussing combinations of premisses dissimilar in form
where the affirmative is universal. He has proved that if the
15 affirmative, being universal, is on the minor premiss, there is a
syllogism, and he has justified this by means of a reduction to the
impossible. Now he takes the converse case, in which the universal
affirmative is on the major premiss and the particular negative on
47
Reading ti for touto (the Aldine prints ti toiouto).
48
i.e. Felapton: Aristotle, An. Pr. 28a26-30; above, 100,28-101,8.
49
See above, 97,3-9.
 6. The Third Figure 179
the minor, and he proves that this combination is non-syllogistic,
showing its unreliability by setting down terms. (This is how he
usually shows combinations to be unreliable.) He proves that the 20
major extreme may hold of all and also of none of the minor.
As terms to prove that P holds of every R in this combination he
set down animate for P, man for R, animal for S. For animate holds
of every animal, man does not hold of some animal, and animate 25
holds of every man.
But he says that it is not possible to take terms for 'P holds of no
R', if the particular negative which says that R does not hold of some
S is taken as true in such a way that R also holds of some S. (This is
the case with the terms he set out: man, which does not hold of some
animal, also holds of some animal.) Now when the particular 30
negative is true in such a way that the particular affirmative is true
together with it, it will not be possible to take terms for 'of no R'. The
reason, again,50 is that, if the major is supposed as a universal
affirmative and the minor as a particular affirmative, it can be 105,1
deduced syllogistically that P holds of some R; and this conjunction
is present potentially when the particular negative is taken as true
in its own right, i.e. in the way I have just described. And since,
when the premisses are posited in this way, it can be deduced
syllogistically that P holds of some R, then it is impossible to obtain 5
terms by means of which we may prove that P holds of no R. For in
that case the syllogistic combination would be rejected. This is what
Aristotle shows when he says 'For if P holds of every S and R of some
S, then P will hold of some R'.
Having shown this, he continues: 'But it was supposed to hold of
none.' He does not mean that it was somehow supposed in the 10
premisses that P holds of no R; for what he wanted to prove was that
terms cannot be taken for 'P holds of no R'. What he means is rather
this: if someone hypothesizes that P holds of no R, when it is
supposed that P holds of every S and that R does not hold of some S
in such a way as also to hold of some S, then he will make an
impossible hypothesis. For P will hold syllogistically of some R. And
if this is so, it is impossible for it to hold of none. It is possible to 15
assume the hypothesis that P holds of no R and to prove it
impossible by a proof which he himself ordinarily uses. For if P holds
of no R and if R holds of some S (because it is supposed that R does
not hold of some S in such a way as also to hold of some S), then it
comes about syllogistically that P does not hold of some S - and this
is impossible. For it was supposed in the combination that P holds of 20
every S.
Thus when the particular negative is taken to be true in this way,
50
See above, 66,1-67,24; 87,5-88,14.
 180 Translation
you can neither obtain terms for T holds of no R', nor prove the
combination to be non-syllogistic. But since particular negatives are
true not only when their subcontraries are true but also when the
25 universal negative is true (the truth of particular negatives is
indeterminate, as we said earlier51), then when the particular
negative is taken as true in such a way that the universal under
which it falls is also true, the combination is non-syllogistic; for it
consists of a universal affirmative major and a potentially universal
negative minor, and he has proved that this combination is
52
30 non-syllogistic by setting down terms. It is clear, then, that the
present combination will also be non-syllogistic, inasmuch as in
some cases the particular negative is true together with the
universal negative.
If the terms are animal, inanimate, man, then animal holds of
every man, inanimate will not belong to some man (since in fact it
106,1 holds of no man), and animal holds of no inanimate. Hence in this
way we can obtain terms for 'of no R', just as we did for 'of every R',
and we can prove the combination non-syllogistic.

6.4.4 Ferison
If the privative is the universal term, <then when the major is
privative and the minor predicative, there will be a syllogism.
For if P holds of no S and R holds of some S, P will not hold of
some R. For again there will be the first figure, the premiss RS
being converted.> [1.6,28b31-5]
He is still discussing premisses dissimilar in form and combinations
5 which contain only one universal. Having discussed those in which
the universal is affirmative, he now discusses those which have the
negative universal (and the particular, of course, affirmative). In
combinations of such premisses he says that if the major is a
universal negative, there will be a syllogism, but not in the converse
case.
10 He proves that there is a syllogism when the major is a universal
negative and the minor a particular affirmative, setting out the
combination with letters. If P holds of no S and R of some S, P will
not hold of some R. For when the minor premiss, RS, is converted,
15 we get the first figure and its fourth syllogism, which deduces a
particular negative from a universal negative major and a
particular affirmative minor.
We said that this syllogism was the fifth, inasmuch as it is proved
by way of conversion, whereas the syllogism which Aristotle placed
51
See above, 66,1-10.
52
See Aristotle, An. Pr. 28a30-3; above, 101,11-22.
 6. The Third Figure 181
53
fifth cannot be proved by conversion. For proof by conversion is
more authoritative and more appropriate than proof by means of the
impossible, as Aristotle himself will say.54 It is possible to use 20
reduction to the impossible and the method of exposition for this
syllogism too; but he does not say so since it is well known and has
already been said.55

6.4.5 More non-syllogistic combinations

When the minor is privative, <there will not be a syllogism.
Terms for holding: animal, man, wild. For not holding: animal,
science, wild. (Wild is middle in both cases.)> [1.6,28b36-8]
He has turned to the combination in which the minor premiss is a
universal negative and the major a particular affirmative, and he
proves that it is non-syllogistic by setting down terms. For 'of every', 25
he sets down animal, man, wild: animal (which is P) holds of some
wild (which is S); man (which is R) holds of no wild; and animal
holds of every man. For 'of none': animal, science, wild. Again,
animal holds of some wild, science (supposed for R) of no wild, and 30
animal of no science. The reason is that the minor is negative.

Nor when both are posited as privative <and one is universal 107,1
and the other particular^ [1.6, 28b38-9]
He assumes both premisses as negative with one of them universal,
and he proves that the combinations of these premisses are
non-syllogistic, whichever is the universal. Now, if the minor is a
universal negative, the terms he sets down for 'of every' are animal, 5
man, wild. For animal does not hold of some wild, man holds of no
wild, animal of every man. For 'of none': animal, science, wild. For
animal does not hold of some wild, science holds of no wild, and
animal of no science.

When the major, <for not holding: raven, snow, white. It is not
possible to take terms for holding if R holds of some S and does
not hold of some S. For if P holds of every R and R of some S, P

63
See above, 96,34-97,3.
64
Aristotle nowhere says exactly this. Alexander is probably thinking of ATI. Post.
87al-30, where Aristotle argues that a direct proof of a proposition is 'superior' to a
proof by reduction to the impossible (see Barnes (1975), pp. 180-1). (But note that
Themistius, Max. 192, says that Aristotle makes the point in the second book of
An.Pr.)
56
Aristotle has not in fact said this; but Alexander wrongly takes him to have done
so at ATI. Pr. 28bl5 (see above, 103,4-7).
 182 Translation
holds of some S. But it was supposed to hold of none. It must be
proved from the indeterminacy. > [1.6,29a2-6]
10 He is trying to prove that the combination in which the major is a
universal negative and the minor a particular negative is
non-syllogistic. Now for 'P holds of no R' the terms he sets down are
raven, snow, white; for raven holds of no white, snow does not hold
of some white, raven holds of no snow.
15 He says again that it is not possible to take terms for T holds of
every R' when R does not hold of some S in such a way that it also
holds of some S. (This was the case with the terms he set down, snow
and white.56) The reason is this. We again get a syllogistic
combination which has a universal negative major and a particular
affirmative minor; and as we have proved,57 this deduces a
particular negative, the opposite of which is impossible. Thus if we
20 hypothesize that P holds of every R, then given that R does not hold
of some S in such a way that it also holds of some S, our hypothesis
will be impossible. For it will be cancelled, since it was proved
syllogistically that P does not hold of some R. But since the truth of
particulars is indeterminate, if we take the particular negative to be
true inasmuch as the universal negative is true, we will then obtain
25 terms for 'of every', and we will prove that the combination is
non-syllogistic. For it is potentially the same as the combination of
two universal negatives.58 Aristotle proved by reduction to the
impossible that it is not possible to obtain terms for 'P holds of every
R', given that the particular negative RS is true in its own right. For
if it is hypothesized that P holds of every R and it is also the case
30 that R holds of some S, then P will also hold syllogistically of some of
the Ss. But this is impossible - for it is supposed that P holds of no S.

6.5 Combinations of particular premisses

108,1 And if each holds of some of the middle,... [1.6, 29a6-7]
He proves that all combinations of two particulars are non-
syllogistic, if they are assumed both as affirmative, or both as
negative, or as dissimilar in form, or both as indeterminate, or the
5 one as indeterminate and the other as particular. He proves that all
these conjunctions are non-syllogistic by way of the same terms,
setting them down and showing that P will hold of no R and also of
every R when the combinations are assumed in these ways. For
holding of every: animal, man, white. For you can assume animal
10 and man each to hold of some white, and each not to hold of some, or
56
See also above, 104,25-9.
57
i.e. Ferison: Aristotle, An. Pr. 28b31-5; above, 106,4-21.
58
See Aristotle, Are. Pr. 28a33-6; above, 101,24-30.
 6. The Third Figure 183
one to hold of some and the other not to hold of some, and also
indeterminately; and animal holds of every man, however the
premisses are taken. For holding of none: animal, inanimate, white.
For again you can assume animal and inanimate each to hold of
some white and each not to hold of some - and also in the
complementary cases, and indeterminately: when any of these
combinations is true, animal holds of no inanimate. 15

6.6 Concluding remarks

Thus it is evident in this figure too when there will be a
syllogism and when there will not... [1.6, 29all-12]
He has shown in this figure, as in the one before it, how many and
which are the syllogistic combinations and why. And he has also
shown which are the non-syllogistic conjunctions. 20

... and if there is a syllogism, it is necessary for the terms to

stand thus. <It is evident too that all the syllogisms in this
figure are imperfect ... and that it is not possible to syllogize a
universal, whether privative or affirmative, by means of this
figure.> [1.6, 29al3-18]
Here too59 we must understand 'in this figure'. For what he first
said, namely 'that when the terms stand as has been said there is a
syllogism', is true without qualification; but 'if there is a syllogism, it
is necessary for the terms to stand thus' is not true without 25
qualification - rather, if there is a syllogism in this figure. For it is
possible to syllogize particular affirmatives and particular negatives
even if the terms are not taken in this way: it is possible to prove a
particular affirmative by the first figure (in which the terms stand in
a different way) and a particular negative both in the first and in the 30
second. It is evident too, as he adds, that the syllogisms in this figure
are all imperfect, like those of the second figure, and that no
universal is deduced in the third figure.

59
See above, 93,16 note.
 7
Further Reflections

7.1 Syllogisms with non-standard conclusions1

109,1 It is clear too that in all the figures, when there is no syllogism,
if both the terms are predicative or privative nothing necessary
comes about at all; <but if they are predicative and privative,
then if the privative is taken as universal there is always a
syllogism of the minor extreme in relation to the major. E.g. if
A holds of every or of some B and B of no C. For if the premisses
are converted, it is necessary for C not to hold of some A.
Similarly in the other figures too: there is always a syllogism
by means of conversions [1.7, 29al9-27]2
After he has set out all the combinations in the three figures and has
5 shown in each figure which conjunctions are syllogistic and which
non-syllogistic, he discusses the non-syllogistic conjunctions and
records the following facts about them. The non-syllogistic
combinations in each figure which are similar in form not only do
not prove the point at issue - they do not prove or deduce anything
else either. But as for the non-syllogistic combinations in each figure
10 which are dissimilar in form and have a universal negative -
although they are non-syllogistic in relation to proving the point at
issue, it is possible to syllogize and prove something else from them.
In the first figure the combinations similar in form are the two
particular affirmatives, the negatives (universal and particular),
and the complementary cases;3 in the second figure all affirmatives
1
On this section see 70,21 note.
2
On this paradoxical text see esp. Patzig, pp. 93-6 (cf. pp. 56,110,128 n. 14).
3
Alexander should describe six cases here: (1) the combination of two particular
affirmatives, which we suppose him to intend by the inaccurate phrase 'the two
particular affirmatives' (there are not two affirmative combinations); those (2) of two
universal negatives, and (3) of two particular negatives, which we take to be intended
by the phrase 'the negatives (universal and particular)'; and those of (4) a particular
affirmative major and a universal affirmative minor, (5) a universal negative major
and a particular negative minor, and (6) a particular negative major and a universal
negative minor - which must somehow be encompassed by the phrase 'the
complementary cases' (for which see above, 108,14). It is extremely difficult to believe

184
 7. Further Reflections 185
and all negatives; in the third figure the particular affirmatives and 15
all negatives. These conjunctions of premisses, then, prove nothing
syllogistically at all.
Of conjunctions dissimilar in form (i.e. differing according to their
quality) and having a universal negative, the following were
non-syllogistic: in the first figure, the combination of a universal
affirmative major and a universal negative minor and that of a
particular affirmative major and a universal negative minor; in the 20
second figure, the combination having a particular affirmative
major and a universal negative minor; and in the third figure those
having an affirmative major, either universal or particular, and a
universal negative minor.
It has been proved that these combinations do not syllogize the 25
point at issue.4 Yet something else can be syllogized from them,
either by converting both premisses and interchanging the terms5
(in the first figure), or by converting only one premiss (in the other
two figures), so that the minor extreme becomes the major and is
predicated in the conclusion.
Take first the non-syllogistic combination in the first figure with 30
'A holds of every B' and 'B holds of no C'. In this combination there is
no deduction from A to C. This is why the combination is
non-syllogistic, being of no use with regard to the point at issue. Yet
something can be deduced syllogistically from the minor to the 110,1
major term. For if both premisses are converted, we get 'C holds of
no B' and 'B holds of some A' (since the universal affirmative
converts with the particular). From these premisses it can be
deduced that C does not hold of some A, where C takes the place of
the major term in the conclusion and A that of the minor, conversely 5
to how they were supposed. (The particular negative does not
convert, so that we cannot convert and have the conclusion we
proposed. In that case the conjunction would deduce the point at
issue, though not indemonstrably.) From this combination, then,
this is what can be proved and this is the way in which it is proved.
The other combination - with a particular affirmative major and a 10
universal negative minor - also proves by conversion of both
premisses that C does not hold of some A, taking the terms
conversely to how they were supposed.
These are the last two syllogisms of the five which Theophrastus -
who says that there are nine syllogisms in the first figure - adds to
the four laid down in the first figure.6 They are last because they do 15

that the text yields these six cases: either Alexander has made a trivial error or else
his scribes have miscopied.
4
See Aristotle, An. Pr. 26a2-9,36-9; 27b6-8; 28a30-3,36-8.
5
i.e. making the major term minor and the minor major.
6
Theophrastus, F17 Graeser (see above, 70,21 note).
 186 Translation
not prove the point at issue at all, unlike the three before them in
which the conclusion is converted. (The remaining three, which come
about by conversion of the conclusions of the first, second, and third
syllogisms in the first figure, are also mentioned by Aristotle, at the
20 beginning of the second book of the Prior Analytics .7 We will remark
on them again in our notes there. In Theophrastus they are ranked
before the other two.)
The non-syllogistic conjunction in the second figure with a parti-
cular affirmative major and a universal negative minor is indeed
non-syllogistic with regard to proving that something is deduced from
25 the major to the minor term; but it deduces something from the minor
to the major if the universal negative minor is converted and the
minor term takes instead the major place. For if the order is changed8
and the minor converted, then the premiss which was the minor is the
major and a universal negative and the premiss which was the major
is the minor and a particular affirmative, the terms remaining in the
30 same order. From these premisses it can be deduced that the term
originally hypothesized as minor does not belong to some of the one
originally posited as major.
Again, the two combinations in the third figure - one with a
111,1 universal affirmative major and a universal negative minor and the
other with a particular affirmative major and a universal negative
minor - are non-syllogistic with regard to the point at issue. Again,
both of them syllogize a particular negative from the minor to the
5 major term, again by conversion of the affirmative major premiss.
These and thus many are the conjunctions in combinations which
are non-syllogistic with regard to the point at issue but syllogistic
with regard to something else. They all deduce their conclusions in
the first figure. Aristotle shows how the deduction comes about and
what is deduced in one of the combinations in the first figure: he omits
the other combinations which syllogize something in a way similar to
10 the one he mentions. He makes only this one clear because these
syllogisms are generated and proved in the same way in all the
figures. This is so because they all yield the fourth indemonstrable.

<It is clear too that if an indeterminate is posited instead of a

particular predicative, it will effect the same syllogism in all the
figures.> [1.7, 29a27-9]
One might ask why he says 'an indeterminate ... instead of a
particular predicative' and not 'and negative' as well. Either, then,
15 he means to refer to the latter in referring to the former, or else it is
7
See Aristotle, An. Pr. 53a3-14: Alexander's commentary on the second book of
An. Pr. has not survived.
8
i.e. the order of the terms in the premiss (see above, on taxis), not the order of the
 7. Further Reflections 187
because in syllogistic combinations the particular affirmative has a
certain advantage over the particular negative. For in those
syllogistic combinations in which the particular premiss is
affirmative, the point at issue is proved probatively - either
indemonstrably, as in the first figure, or by conversion of a premiss,
as in the second and third figures. But in those syllogistic 20
combinations in which the particular is negative, the point at issue
is proved only by reduction to the impossible, as in the fourth
syllogism of the second figure and the sixth of the third. (The
particular negative occurs only in these syllogisms.) Perhaps, then,
this is why he mentions the affirmative - because it has a certain
advantage. For even in reductions to the impossible we do not use 25
the particular negative for the syllogism of the falsehood rather, it
is used to prove the impossibility of the conclusion.

It is clear too that if an indeterminate is posited instead of a

particular predicative, it will effect the same syllogism in all
the figures. [1.7,29a27-9]
We have said that indeterminate propositions are equivalent to 30
particulars.9 For even if they sometimes cover the universal case,
nonetheless, they are not equivalent to universals. For it is not the
case that if an indeterminate is true, the universal is always true.
But the particular is always true - just as, conversely, when a
particular is true, so also is the indeterminate. Hence, if an 112,1
indeterminate is assumed, then since it is equivalent to a particular,
it will produce the same syllogism as the particular produced when
it was posited.

7.2 Reduction

7.2.1 All syllogisms reduced to the first figure

It is evident too that all the imperfect syllogisms are perfected
by means of the first figure. <For they are all brought to a
conclusion either probatively or by means of the impossible,
and in both ways we get the first figure when they are
perfected probatively, because they are all brought to a
conclusion by means of conversion and conversion produces the
first figure; when they are proved by means of the impossible,
because when the falsity is posited the syllogism comes about
by means of the first figure. (E.g. in the last figure, if A and B
premisses (pace Volait, pp. 30-3), to which Alexander (rightly) shows himself
indifferent.
9
See 30,31 note.
 188 Translation
hold of every C, A holds of some B. For if of none and B of every
C, then A of no C. But it held of every C. Similarly in the other
cases too.)> [1.7,29a30-9]
5 He has proved above that the syllogisms of both the second and the
third figures, which are imperfect, are all perfected by means of the
first figure, and now he brings the fact to our attention. For they
were perfected either by conversion - those in the second, by
conversion of the major premiss, those in the third, by conversion of
10 the minor - or by reduction to the impossible. And those which are
perfected by conversion were reduced to the first figure, while those
which are perfected by reduction to the impossible were proved to
deduce the point at issue insofar as it was proved in the first figure
that an impossibility follows from the hypothesized premiss.
He says that those proved by conversion are proved 'probatively'
because in them, once the conversion is made, it is the point at
15 issue10 which is primarily proved. With 'because they are all
completed by means of conversion', we must understand 'those
which are proved probatively'; for they are not all proved by
conversion.
Those which are proved by way of the impossible assume a
hypothesis and produce a syllogism and a proof of something else
and not primarily of the point at issue; rather, with them, the point
20 at issue is proved accidentally. Reduction to the impossible comes
about by way of the first figure both in the case of syllogisms in the
second figure and in the case of those in the third. He shows this by
taking one of the combinations of the third figure which is also
proved by conversion.11 This is the combination which deduces a
particular affirmative from two universal affirmatives - which is
proved to be syllogistic, as we have said,12 both by conversion and
25 also by reduction to the impossible. For if it is supposed that A and B
hold of every C, then A holds of some B. For if not, hypothesize the
opposite - which is that A holds of no B. But it was supposed that B
holds of every C. We get the second combination in the first figure,
which deduces a universal negative, 'A holds of no C'. Since this is
30 impossible (for it is supposed that A holds of every C), the hypothesis
from which this followed and which posited 'A holds of no B' is
rejected, and it is established that A therefore holds of some B. And
just as in the case before us the reduction to the impossible has come
about by means of the first figure, so it does too in all the other
syllogisms of both the second and the third figures.
10
Reading prokeimenou for keimenou.
11
Removing the commas after gignetai (112,20) and edeixen (112,21) and adding a
comma after tritoi (112,21).
12
Darapti: see above, 98,20-99,15 (on An. Pr. 28al8-23).
 7. Further Reflections 189
That proof by exposition is perceptual and not syllogistic is clear 113,1
from the fact that he does not mention it here as coming about by
means of a syllogism.13

7.2.2 All syllogisms reduced to Barbara and Celarent14

It is also possible to reduce all the syllogisms to the universal
syllogisms in the first figure. [1.7,29bl-2]
He15 first proved that all the syllogisms in the second and the third 5
figures reduce to the first figure and are perfected by it; for in this
sense the latter are perfect and indemonstrable. Here, since the first
two syllogisms in the first figure are more perfect,16 containing two
universal premisses, he proves that all the syllogisms can be
reduced to them. He thus sets us an exercise,17 and at the same time 10
shows that it is possible, on the basis of combinations which are
syllogistic par excellence and by general agreement,18 to justify the
view that the others are indeed syllogisms. There are two remaining
syllogisms in the first figure (those with particular affirmative
minors), four in the second figure, and six in the third. Thus he
reduces all these twelve to the first two in the first figure. 15

<For the second figure syllogisms are evidently perfected by

them (though not all in the same way: the universals by
conversion of the privative, each of the particulars by means of
reduction to the impossible).> [1.7,29b2-6]
The first two syllogisms in the second figure have an evident
reduction.19 They were proved when their universal negative was
converted, and by this conversion we got the second syllogism in the
first figure with a universal negative major and a universal
affirmative minor. The two syllogisms which deduce a particular 20
negative are reduced to first figure syllogisms by reduction to the
impossible. In the case of the syllogism which deduces a particular
negative from a universal negative major and a particular
affirmative minor,20 we hypothesize the opposite of what is deduced
(which is a particular negative), i.e. a universal affirmative, and we
13
See above, 33,1 note.
14
On this section see Volait, pp. 48-53.
15
113,5-7 = FDS 1101.
18
'Degrees of perfection' are not found in Aristotle: see Patzig, pp. 48-9 (on degrees
of imperfection see above, 77,9).
17
gumnasion: for a comparable comment see e.g. Ammonius, in Int. 251,31.
18
Reading homologoumends sullogistikon for homologoumenon.
19
Cesare, Camestres: Aristotle, An. Pr. 27a5-14; above, 78,11-79,25.
20
Festino: Aristotle, An. Pr. 27a32-6 (where it reduces to Ferio by conversion);
above, 82,25-83,8.
 190 Translation
25 co-assume the major, which is a universal negative. In this way we
get, in the first figure, a universal negative major and a universal
affirmative minor, which we hypothesized. The conclusion is a
universal negative, which is impossible - for the middle term, which
was deduced to hold of none of the last term, holds of some of it. And
in this way the third syllogism in the second figure is reduced to the
30 second in the first.
The fourth, which has a universal affirmative major and a
particular negative minor,21 is reduced to the first syllogism in the
first figure, which has two universal affirmatives, as has been
proved. For if it is not true that the major does not hold of some of
the minor, then suppose it to hold of all. But the middle is supposed
to hold of all the major. Thus there are two universal affirmatives in
35 the first figure, and from them it will be deduced that the middle
holds of all the minor. But it was supposed not to hold of some of it.

<The particular syllogisms of the first figure are indeed

perfected by means of themselves; but it is also possible to
prove them by means of the second figure by reducing them to
the impossible. E.g. if A holds of every B and B of some C, then
A of some C. For if of none and of every B, then B will hold of no
C - we know this by means of the second figure. The
demonstration will be similar in the case of the privative
syllogism. For if A holds of no B and B of some C, A will not
hold of some C. For if it holds of every C and of no B, B will hold
of no C. (This was the middle figure.) Hence since all the
syllogisms in the middle figure are reduced to the universal
syllogisms in the first figure, and the particular syllogisms in
the first figure to the syllogisms in the middle figure, it is
evident that the particular syllogisms will be reduced to the
universal syllogisms in the first figure.> [1.7, 29b6-19]
114,1 Having proved that all the syllogisms in the second figure reduce to
the two in the first figure with two universal premisses, he next
discusses the two in the first figure which have a particular
affirmative minor and a universal major (in the one case affirmative
5 and in the other negative). He says that for the proof of their
necessity they are self-sufficient and need nothing else since they
are perfect, but that they can in fact be reduced to the two
syllogisms in the first figure with two universals, if they are first
reduced by reduction to the impossible to the first two in the second
figure.
The third, which deduces a particular affirmative, he reduces to
21
Baroco: Aristotle, An. Pr. 27a36-bl; above, 83,12-84,3.
 7. Further Reflections 191
the second in the second figure, which has a universal affirmative 10
major and a universal negative minor. For it is supposed that A
holds of every B and that B holds of some C. The conclusion is that A
holds of some C. If this is not so, then its opposite, that A holds of no
C, is the case. But it was supposed that A holds of every B. This is
the combination of the second syllogism in the second figure which,
by conversion of the negative premiss, was reduced to the second 15
syllogism of the first figure. (Its conclusion was also converted.) So if
the third syllogism in the first figure has been reduced to the second
in the second and the second in the second to the second in the first,
then the third in the first reduces to the second in the first.
The fourth syllogism, which has a particular negative conclusion, 20
he reduces by reduction to the impossible to the first in the second
figure, which has a universal negative major. For if it is not true
that A does not hold of some C, it will hold of every C. But it is
supposed that it also holds of no B. This is the combination of the
first syllogism in the second figure, which is itself reduced by
conversion of the universal negative to the second in the first. The 25
fourth in the first will therefore also reduce to the second in the first.
Having reduced the two syllogisms in the first figure which
deduce a particular to the two22 which deduce a universal by
reducing them to the first two in the second which had been reduced
to the latter, he next attempts to reduce the six syllogisms in the 30
third figure to the same two, and says:

The syllogisms in the third figure, when the terms are

universal, are perfected immediately by means of those
syllogisms;... [1.7, 29bl9-21]23
What he means is this: in the third figure the first two syllogisms
contain two universal premisses - one of them has two affirmatives, 115,1
the other a universal negative major.24 He says that these can be
immediately reduced to and perfected by the syllogisms in question.
This is so because each is immediately reduced to one of them by
reduction to the impossible, the one with two universal affirmatives 5
to the second and the one with a universal negative major and a
universal affirmative minor to the first. For if we hypothesize the
opposite of what is deduced (which is either a particular affirmative
or a particular negative), we have either a universal negative or a
22
Reading tous (Aldine edition) for toutou tous (Wallies: toutous MSS).
23
With the following paragraphs compare [Ammonius], in An. Pr. XIII, a scholium
which paraphrases and comments on Alexander.
24
i.e. Darapti (An. Pr. 28al7-23 - for the reduction to the impossible by way of
Celarent see above, 99,4-15); and Felapton (An. Pr. 28a26-30 - for the reduction to the
impossible by way of Barbara see above, 100,34-101,3).
 192 Translation
universal affirmative; we also have the premiss co-assumed, which
10 is both universal and affirmative (both premisses were universal);
and we get a syllogism in the first figure with two universal
premisses. The deduction is either an affirmative or a negative
universal.
He cannot mean25 that the reduction proceeds by conversion. For
conversion makes one of the premisses particular. Nor is the fact
that both their premisses are universal in itself sufficient for them
to be reduced to the syllogisms in question. For the reduction must
15 be done by way of conversion, and when the conversion is made, the
premisses no longer remain universal. (In the case of the syllogisms
in the second figure with two universals, he used conversion and by
means of this proved that they reduce to these syllogisms.)

<... and when the terms are taken as particulars, they are
perfected by means of the particular syllogisms in the first
figure - but these were reduced to the former: hence so too are
the particulars in the third figure.> [1.7,29b21-4]
As for the next three - two of them with a universal major, either
affirmative or negative, and a particular affirmative minor, and the
20 other with a particular affirmative major and a universal
affirmative minor - these no longer seem to derive from two
universals by way of reduction to the impossible. For the premiss
hypothesized in their case must certainly be universal, since it is
taken as opposite to the conclusion, which is particular; but the
co-assumption is particular.
Or does this hold only of the two cases in which the minor is a
25 particular affirmative?26 For it is this which is assumed together
with the hypothesis, which was universal; and when it is assumed it
produces the first figure - not one of the syllogisms with two
universals but one with one particular premiss. Hence, since they
reduce by reduction to the impossible to the same syllogisms to
which they reduce by conversion, he reasonably used proof by
conversion in their case.27
30 It is possible, however, to reduce them by reduction to the
impossible to the syllogisms in the second figure with two universal
premisses; and these are reduced by conversion of the negative
25
sc. despite the fact that he says 'immediately', which here means not 'by using
one or more conversions' but 'without first being reduced to some other syllogism'.
26
For the third case, Disamis, can be treated by reduction to the impossible using
Celarent since the co-assumption is universal: below, 116,20-9.
27
Datisi is reduced by conversion to Darii (An. Pr. 28bll-14; above, 102,27-103,3)
and by reduction to the impossible using Ferio. Ferison is reduced by conversion to
Ferio (An. Pr. 28b31-5; above, 106,9-16) and by reduction to the impossible using
Darii.
 7. Further Reflections 193
premiss to the second syllogism in the first figure. For if it is
supposed that A holds of every C and B of some C, then I say that A
holds of some B. For if not, then A holds of no B. But it is supposed
that it also holds of every C. The proof is similar if AC is a universal 35
negative.28
The reduction by way of the second figure to the first syllogisms in
the first figure will seem easier than the proof which reduces them 116,1
by conversion to the two in the first figure which have a particular
minor premiss, inasmuch as these in turn must first be reduced by
reduction to the impossible to the syllogisms in the second figure in
which both premisses are universal.29
Again, it is possible to say that these syllogisms too are reduced to
the first syllogism in the first figure by conversion. In the case of the 5
syllogism with a universal affirmative major, we must hypothesize
not the opposite of what is deduced but the converse of the opposite.
For if it is supposed that A holds of every C and B of some C, it can
be deduced that A holds of some B. If not, then its opposite, that A
holds of no B, is the case. But since this converts, B will hold of no A. 10
But it was supposed that A holds of every C. And from this it can be
deduced that B holds of no C, which is impossible.
If AC is a universal negative, what is deduced is that A does not
hold of some B. For if not, then let its opposite, that A holds of every
B, be the case. If this is assumed, and if the universal negative AC is
converted so that C is taken to hold of no A, then again we get the 15
second syllogism in the first figure.30
But if in a reduction to the impossible you must hypothesize the
opposite of what is being proved and co-assume the other
supposition, the first syllogism I stated will seem not to satisfy the
suppositions, because it hypothesizes not the opposite of what is
being proved but the converse of the opposite; nor will the second,
because it co-assumed not one of the suppositions but the converse 20
of a supposition.31
The syllogism with a particular affirmative major and a universal
affirmative minor can be reduced by conversion to the third
syllogism in the first figure, which does not contain two universals.32
However, it can also be reduced by reduction to the impossible to the
28
i.e. Datisi may be treated by a reduction to the impossible using Cesare, and
Ferison by a reduction using Camestres.
29
Thus Aristotle reduces Datisi to Celarent by way of first Darii and then
Camestres. Alexander reduces Datisi to Celarent by way of Cesare alone.
30
The combination is: A holds of no C, B holds of some C. Hypothesize that A holds
of every B. Convert the first premiss to: C holds of no A. From this and the hypothesis,
infer that C holds of no B (by Celarent). Convert the conclusion to: B holds of no C -
and this is inconsistent with the second premiss of the original combination.
31
A trifling objection to a sound procedure.
32
cf. Aristotle, An. Pr. 28b7-ll; above, 102,6-24.
 194 Translation
second syllogism in the first figure which has a universal negative
25 major and a universal affirmative minor, to which the first
syllogism, with two universal affirmatives, was also reduced.33
Hence this too, like the first two with two universal premisses, is
also immediately perfected by means of the universal syllogisms in
the first figure. (Those syllogisms reduce by conversion to the third
and fourth34 just as this reduces to the third.)
30 Now this case was overlooked in what Aristotle said and still
more the case of the syllogism with a particular negative major and
a universal affirmative minor. For this is not reduced in any way to
the particular syllogisms in the first figure - it is not proved by
conversion at all, as he said when he discussed it,35 but only by
35 reduction to the impossible, by which it is reduced to the first
syllogism of the first figure.
In this way,36 then, the syllogisms with two universal premisses
are directly reduced to the syllogisms in question, according to
117,1 Aristotle. The other four, he says, which have a particular premiss,
are reduced by conversions to the two particular syllogisms in the
first figure. These were proved to reduce to the first two syllogisms
by way of the first two in the second figure. Hence the syllogisms in
the third figure will be reduced to them by the same means. For the
5 syllogisms in the third figure reduce by conversion to the particular
syllogisms in the first figure, and the particular syllogisms in the
first figure reduce to the first two in the second figure by reduction to
the impossible, and the first two in the second figure reduce to the
second in the first figure by conversion. Hence the syllogisms in the
third figure with one particular premiss will be reduced to this same
syllogism.
But perhaps, as I said,37 this is so not for the four syllogisms but
10 only for the three which are proved by conversion. The syllogism
with a particular negative major and a universal affirmative minor
was neither proved by conversion nor reduced to one of the
particular syllogisms in the first figure. Rather, it too was directly
reduced by reduction to the impossible to the first syllogism in the
first figure, as has been proved.38 But the same is true of one of the
33
See above, 114,33-115,12.
34
Darapti to Darii (An. Pr. 28al8-22); Felapton to Ferio (An. Pr. 28a26-9).
35
See An. Pr. 28bl5-20 - but Aristotle does not say this. Alexander himself does, at
103,16-19. Should we change eipe Che said') to eipon ('I said')?
36
This paragraph does not fit. We suspect that in the original version of
Alexander's discussion, 115,17 was immediately followed by 116,36-117,9. When he
considered the matter again, Alexander added some new thoughts in 115,17-116,36.
Having done so, he re-read 116,36-117,9 which he corrected and supplemented by
appending 117,9-25. (See, in general, Introduction, p. 9.)
37
Above, 116,30-5.
38
Aristotle, Are. Pr. 28bl5-20; above, 103,9-25.
 7. Further Reflections 195
other three: the syllogism with a universal affirmative minor and a 15
particular affirmative major can indeed be reduced by conversion to
the third syllogism in the first figure - but by reduction to the
impossible it too can be reduced directly to the second syllogism.39
And if in this case we must use proof by conversion because it is
more authoritative,40 then we must do the same in the case of the
syllogisms with two universal premisses. And in that case they too
will no longer be perfected immediately by means of the first two 20
syllogisms in the first figure but will also be reduced to the two
particular syllogisms, the one to the third and the other to the
fourth.
Perhaps, then, with 'when the terms are taken as particulars' we
should understand 'if they cannot be directly perfected by means of
the syllogisms in question' inasmuch as they are reduced to the
particular syllogisms in the first figure. For this is surely missing, 25
and when it is added there is no longer anything to investigate.

7.3 Final remarks

Thus <we have said how> syllogisms which prove that
something holds or does not hold < stand - both in their own
right, when they are from the same figure, and in relation to
one another, when they are from different figures ...> [1.7,
29b26-9]41
There is a threefold distinction among propositions according to
their modes: some are actual, some necessary, some contingent.42
He has discussed the combinations of actual premisses and the 30
syllogisms involving such premisses in each figure; and now that he
is about to move on to the combinations compounded from necessary
premisses and to discuss them, he informs us that his account of the
combinations of actual premisses has been completed. 118,1
When he says 'in their own right, when they are from the same
figure, and in relation to one another, when they are from different
figures', he is referring to the differences among them. The
difference among the syllogisms in each figure in relation to one
another is evident; for one syllogism in the first figure has two
universal affirmatives, another has a universal negative major - 5
39
See above, 116,20-9.
40
See above, 106,19 note.
41
At this point, Alexander's text of An. Pr. is significantly different from ours. (He
signals no variant readings: the later commentators do not mention Alexander's
reading.) Moreover, Alexander's double set of comments is inconsistent and
perplexing. Perhaps 118,9-25 represents Alexander's second thoughts, later inserted
into the original text.
42
On modes see above, 26,2-22; 27,26-28,30.
 196 Translation
and so on, in each figure. Again, the difference among the syllogisms
in the different figures in relation to one another is familiar; for the
position of the middle term is not the same in all of them.

- both in their own right, when they are from the same figure
...[1.7,29b27-8]
10 This depends on Thus it is evident'.43 For it is evident from what he
has said that all syllogisms will be reduced to the first two
syllogisms in the first figure, and how each syllogism is syllogistic in
its own figure, and why each of them is syllogistic. All this has been
proved.

... and in relation to one another, when they are from different
figures... [1.7, 29b28]
44
15 This too still depends on 'Thus it is evident'. For it is also evident
from what he has said how the figures, and the syllogisms in the
different figures, differ in relation to one another - the difference
depends on the position of the middle term and on the different ways
in which they are reduced to the first two syllogisms in the first
20 figure. For some are reduced directly and some by reduction to one
another (i.e. the two syllogisms in the first figure and the three in
the third with particular propositions); and of the latter again, some
inasmuch as they are reduced to the first syllogisms in the second
figure (i.e. the two in the first figure), and others inasmuch as they
are reduced to those in the first figure which prove particular
premisses (i.e. the three syllogisms in the third figure which are
25 reduced to these by conversion). The difference in their reductions
shows the differences among these syllogisms.

<... (either by the middle term> or inasmuch as they are

reduced to the first figure).
Having said 'either by the middle term', he adds 'or inasmuch as
they are reduced to the first figure'. For they differ in this respect too
119,1 - their reductions are different. For the particular syllogisms in the
first figure are reduced in one way (they are reduced by means of
those in the second) and the syllogisms in the second figure in
another way - and these are themselves reduced in different ways,

43
At An. Pr. 29b24; but unless Alexander's text of An. Pr. was even further from
ours than we suppose, his construal is impossible.
44
Again, at An. Pr. 29b24; and again, Alexander's construal is impossible.
 7. Further Reflections 197
some by conversion and some by reduction to the impossible. So too
with the syllogisms in the third figure: some are reduced by
reduction to the impossible and some by conversion first to the 5
particular syllogisms in the first figure.
 Appendix 1
Translation of Prior Analytics 1.1-7

This Appendix gives a continuous translation of Aristotle, Prior Analytics

1.1-7, from which the translated lemmata in the commentary are drawn.
The translation is deliberately spare, and at times inelegantly close to the
Greek: its purpose is not to ease the understanding of Aristotle's text but to
provide an appropriate companion to Alexander's commentary.
We have translated not the standard modern text of Aristotle but rather
the text which Alexander had before him - insofar as this text can be
reconstructed. Alexander's readings are given explicitly in the citations in
the body of his commentary; they can also sometimes be inferred from the
tenor of his comments. In addition, there are the lemmata: although their
status is uncertain,1 we here take them to represent Alexander's readings
provided that they are consistent with his commentary. In this way,
something like two-fifths of the text which Alexander read can be recovered.
For the rest, our translation simply follows the Oxford Classical Text of
An. Pr., edited by Sir David Ross and Lorenzo Minio-Paluello (Oxford, 1964).
Alexander's text differs from the OCT at a number of places. The notes to
our translation signal all these differences, most of which are trifling. In
addition, we have drawn attention to a few other interesting textual points.
(But it would be rash to use the notes as the basis for any general
conjectures about the relationship between Alexander's text and the
manuscript tradition of An. Pr.2)
'Alex1' refers to lemmata in Alexander, 'Alexc' to citations in the body of
his commentary. Other abbreviations are self-explanatory. The sigla for the
MSS are those of the OCT for Aristotle3 and of Wallies' edition for
Alexander.

Aristotle Prior Analytics 1.1-7

24alO (1) First, to say about what and of what the inquiry is: it is about
demonstration and demonstrative science.4 Then to determine what a
proposition is, and what a term, and what a syllogism, and what sort of
syllogisms are perfect and what sort imperfect; and after that, what it is for

1
See Introduction, p. 17.
2
On this question see Wallies; Ross, pp. 90-2.
3
We have consulted Williams for corrections and additions to the OCT's critical
apparatus.
4
24all: epistemen apodeiktiken Alexlc; epistemes apodeiktikes OCT, Alex: Aldina.

198
 Translation of Prior Analytics 1.1-7 199
this to be or not be in that as in a whole, and what we mean by being
predicated of every or of none. 15
Now a proposition is an utterance affirming or negating something of
something. It is either universal or particular or indeterminate. By
universal, I mean holding of all or of none, by particular holding of some or
not of some or not of all, by indeterminate holding or not holding, without 20
being universal or particular (e.g. that of contraries there is the same
science, or that pleasure is not a good).
A demonstrative proposition differs from a dialectical proposition in that
a demonstrative proposition is the assuming of one part of a contradictory
pair (for if you are demonstrating you do not request but assume), while a
dialectical proposition is a request for one of a contradictory pair. There will 25
be no difference with regard to there being a syllogism in each case. For
whether you are demonstrating or requesting, you syllogize by assuming
that something holds of something or that something5 does not hold. Hence
a syllogistic proposition simpliciter affirms or negates something of
something in the way already mentioned.6 It is demonstrative if it is true
and is assumed by way of the initial hypotheses. It is dialectical if it is a 24blO
request for one of a contradictory pair (if you are inquiring) or an
assumption of what is apparent and reputable (if you are syllogizing) - as
has been said in the Topics. In what follows we shall say with precision
what a proposition is and how7 syllogistic and demonstrative and dialectical
propositions differ; but for present purposes let what we have determined 15
be enough.
I call a term that into which a proposition resolves - i.e. the predicate and
that of which it is predicated (when you add or remove8 'is' or9 'is not').
A syllogism is an utterance in which, certain things being posited,
something different from the suppositions comes about by necessity
inasmuch as they are the case. By 'inasmuch as they are the case' I mean 20
that it comes about because of them; and to come about because of them is to
need no external term for the generation of the necessity.
I call a syllogism perfect if it needs nothing else apart from the
assumptions in order for the necessity to be evident; and imperfect if it
needs one10 or more items which are necessary by way of the terms 25
supposed but have not been assumed by way of premisses.
For one thing to be in another as in a whole and for the other to be
predicated of all the one are the same thing. We say that one thing is
predicated of all another when it is not possible to take any of it11 of which 30
the other is not said. And similarly for of none.
(2) Now every proposition is either of holding or of holding by necessity or of 25al
holding contingently. Of these some are affirmative and others negative, in
respect of each adjunct. Again, of affirmatives and negatives, some are

5
24a28: kai ti Alexc (13,27); e codd Arist, Alexc (13,17).
6
24a29: proeiremenon Alexc; eiremenon OCT.
7
24bl3: tini Alex1; ti OCT.
8
24bl7: e diairoumenou Alexc, codd Arist; del OCT.
9
24bl8: e Alexc, OCT; kai Arist: ABdA.
10
24b25: henos Alexc; e henos OCT, Alex: Aldina.
II
24b29: labein + ton hupokeimenou codd Arist; del OCT, non habet Alexc (sed cf.
24,29.32; 54,7).
 200 Appendix 1
5 universal, some particular, some indeterminate. And in the case of proposi-
tions of holding, it is necessary for the12 universal privative to convert in its
terms - e.g. if no pleasure is a good, no good will be a pleasure. Predicative
propositions necessarily convert, but to a particular and not to a universal -
10 e.g. if every pleasure is a good, then some good is a pleasure. Of particulars, it
is necessary for the affirmative to convert to a particular (if some pleasure is a
good, then some good will be a pleasure), but the privative does not
necessarily convert (it is not the case that if man does not hold of some
animal, then animal does not hold of some man).
15 First, then, let the proposition AB be a universal privative. Now if A holds
of no B,13 B will hold of no A.14 For if of some, e.g. of C, it will not be true that A
holds of no B. For C is something of B.
If A holds of every B, then B will hold of some A. For if of none, A will hold of
no B. But it was supposed to hold of every B.
20 Similarly if the proposition is particular. For if A holds of some B, then it is
necessary for B to hold of some A. For if of none, then A will hold of no B.15
If A does not hold of some B, it is not necessary that B should not hold of
25 some A. E.g. if B is animal and A man: man does not hold of every animal, but
animal holds of every man.
(3) The case will be the same for necessary propositions. For the universal
privative converts to a universal, and each of the affirmatives to a particular.
30 For if it is necessary for A to hold of no B, it is necessary for B to hold of no A.
For if it is contingent that it holds of some, then it will be contingent that A
holds of some B.16
If A holds by necessity of all or of some of the Bs,17 then it is necessary for B
to hold of some of the As.18 For if it is not necessary, A will not hold of some B
by necessity.
35 The particular privative does not convert for the same reason as we gave
earlier.
As for contingent propositions, since19 being contingent is meant in several
ways (we say that the necessary and the non-necessary and the possible are
40 contingent), in affirmatives the case will be similar in respect of conversion in
all instances.
25bl If it is contingent for A to hold of every or of some B, then it will be
contingent for B to hold of some A. For if of none, then A of no B (this has been
proved earlier).
In negatives it is not the same. With those which are said to be contingent
5 inasmuch as it is by necessity that they do not hold20 or inasmuch as it is not
by necessity that they do hold,21 the case is indeed similar. E.g. if someone
12
25a5: ton men en toi huparkhein ten men Alex1; ten men en toi huparkhein OCT;
ten men oun en toi huparkhein Alex: Aldina.
13
25al5: toi B Alexcl, OCT; ton B Arist: ABC2d, Alexc: LM.
14
25al5: tdi A Alexcl, OCT; tbn A Arist: ABCdrO; to B Arist: n2.
15
25a22: toi B Alexc, OCT; tbn B Arist: ABCdm.
16
25a31 tini toi B Alex1; t6i B tini OCT, Alex: Aldina.
17
25a32: ton B Alex1, Arist: A2C; toi B OCT. Alex: Aldina.
18
25a33: ton A Alex1, Arist: A2C; toi A OCT, Alex: Aldina.
19
25a37: epei Alex1; epeide OCT, Alex: Aldina.
20
25b4: ex anankes me huparkhein Alex ut videtur, Arist: A2B2CdFA; ex
anankes huparkhein OCT.
21
25b5: me ex anankes huparkhein Alex0, Arist: AB1CdrAri; me ex anankes
 Translation of Prior Analytics 1.1-7 201
were to say that it is contingent for man not to be a horse or that it is
contingent22 for white to hold of no cloak. In the one case it is by necessity
that it does not hold, and in the other it is not necessary that it does hold.
And the proposition converts in the same way. For if it is contingent for
horse to hold of no man, then it may be that man holds of no horse; and if it 10
may be that white holds of no cloak, then it may be that cloak holds of no
white. For if it is necessary for it to hold of some, then white will hold of
some cloak by necessity - this was proved earlier. (Similarly too for
particular negatives.)
But as for those which are said to be contingent inasmuch as they are for
the most part23 and by nature - this is the way in which we determine the 15
contingent - here the case will not be similar for negative conversions.
Rather, the universal privative proposition does not convert whereas the
particular does convert.
This will be evident when we discuss the contingent. For the moment, in
addition to what we have said, let it be24 clear that 'It is contingent that it 20
holds of none or not of some' has an affirmative figure. For 'It is contingent'
is ranked in the same way as 'is', and 'is' (where it is co-predicated) always
and in every case makes an affirmation - e.g. 'It is not good' or 'It is not
white' or generally 'It is not such-and-such' (this too will be shown in what
follows). And in respect of conversions the case will be similar to that of the 25
others.

(4) Having determined these issues, let us now say25 by what means and
when and how every syllogism comes about. We must discuss
demonstration later. (We must discuss syllogisms before demonstration
because syllogisms are more universal: a demonstration is a syllogism, 30
but26 not every syllogism is a demonstration.)
Now when three terms so stand to one another that the last is in the
middle as in a whole and the middle is27 or is not in the first as in a whole, it
is necessary for there to be a perfect syllogism of the extremes. I call a term 35
middle when both it is in another and another is in it - it is middle by
position too. Extremes are both that which is in something else and that in
which something else is.
If A is said of every B and B of every C, then it is necessary for A to be
predicated of every C. For we have earlier said what we mean by 'of every'.
Similarly, if28 A is said of no B and B of every C, A will hold of no C. 26al
But if the first follows29 all the middle and the middle none of the last,30
there will not be a syllogism of the extremes. For nothing necessary follows
inasmuch as they are the case. For it is possible for the first to hold of all 5
m22huparkhein OCT.
23
25b6: leukon + endekhesthai
cl
Alexc; non habet1 OCT.
24
25bl4: to polu Alex , OCT; to om Arist: ABd , Alex: Aldina.
25bl9: pros tois eirSmenois estd Alex1; estd pros tois eirSmenois OCT, Alex:
Aldina.
25
26
25b26: legdmen Alex1, OCT; legomen Arist: ABCd.
27
25b30: ho de sullogismos Alexc;1ho sullogismos de OCT.
28
25b34: protdi + e OCT;
1
om Alex bis [52,29 e praebet L, 53,17 & praebet Aldina].
29
25b40: d'ei kai Alex ; de kai ei OCT.
30
26a2: akolouthei Alexcl, OCT; huparkhei codd1Arist.
26a3: toi de eskhatoi to meson medeni Alex ; to de meson medeni toi eskhatoi
huparkhei OCT [huparkhei om Arist: A].
 202 Appendix 1
and also of none of the last. Hence neither the particular nor the universal is
necessary. And if nothing is necessary by way of these items, there will not
be a syllogism. Terms for holding of every: animal, man, horse. For holding
of none: animal, man, stone.
10 Nor when the first holds of none of the middle and the middle of none of
the last will there be a syllogism in this case either. Terms for holding:
science, line, medicine. For not holding: science, line, unit.
Thus when the terms are universal it is clear when there will be a
31
15 syllogism in this figure and when there will not be one; and that if there is
a syllogism it is necessary for the terms to stand as we have said, and that if
they stand in this way there will be a syllogism.
If one of the terms is universal and one particular in relation to the other,
then when the universal is posited on the major extreme (whether it is
predicative or privative) and the particular on the minor is predicative, it is
20 necessary for there to be a perfect syllogism. But when it is on the minor or
the terms stand in any other way, it is impossible. (I call the major extreme
the one in which the middle is, and the minor the one under the middle.32)
For let A hold of every B and B of some C. Then if being predicated of
25 every is what we said at the beginning, it is necessary for A to hold of some
C.
And if A holds of no B and B of some C, it is necessary for A not to hold of
some C. For we have also denned what we mean by 'of none'. Hence there
will be a perfect syllogism. (Similarly too if BC is indeterminate, being
30 predicative. For there will be the same syllogism, whether it is assumed as
indeterminate or as particular.)
If the universal, either predicative or privative, is posited on the minor33
extreme, there will not be a syllogism, whether the indeterminate or
particular is affirmative or negative. I.e.34 if A holds or does not hold of some
35 B and B holds of every C. Terms for holding: good, disposition, sagacity. For
not holding: good, disposition, ignorance. Again, if B holds of no C and A
holds35 of some B or does not hold (or does not hold of every), in this case too
there will not be a syllogism. Terms: white, horse, swan; white, horse,
raven. The same terms serve if AB is indeterminate.
26bl When the term on the major extreme is universal, either predicative or
privative, and the term on the minor is privative and36 particular, there will
not be a syllogism (whether it is assumed as indeterminate or as
particular).37 I.e. if A holds of every B and B does not hold of some C (or if it
5 does not hold of every C). For the first will follow all and also none of that of
some of which the middle does not hold. Suppose the terms to be animal,
man, white. Then of the whites of which man is not predicated, take swan
and snow. Then animal is predicated of all the one and of none of the other.
10 Hence there will not be a syllogism. Again, let A hold of no B and let B not
hold of some C; and let the terms be inanimate, man, white. Then of the
whites of which man is not predicated, take swan and snow. Inanimate is
31
26al5: ontos ge Alexc \ge om Aldina]; ontos te OCT, Alex1.
32
26a22: to meson + on OCT, Alexc: L, Aldina, Alex1: Aldina; om Alex01.
33
26a30: ton elassona Alex1 [to elasson Aldina]; to elation OCT.
34
26a33: to men A OCT, Alex: Aldina; men om Alex1.
35
26a36: toz B + e OCT; om Alex1.
36
26b2: steretikon + kai Alex1; om OCT, Alex: Aldina.
37
26b3: adioristou te kai en merei lephthentos Alex1, codd Arist; del OCT.
 Translation of Prior Analytics 1.1-7 203
predicated of all the one and of none of the other. Again, since 'B does not hold 15
of some C' is indeterminate and it is true both if it holds of none and if it does
not hold of all (because it does not hold of some), and since there is no
syllogism if terms are taken such that it holds of none (this has been said
earlier), it is evident that there will not be a syllogism inasmuch as the terms
stand in the way we are considering. For then there would be one in this case 20
too. (And38 it will be proved in the same way if39 the universal is posited as
privative.)
Again, if both the intervals are40 particular, either predicatively or
privatively (or one is said predicatively and the other privatively), or one
indeterminate and the other determinate, or both indeterminate, in none of
these cases will there be a syllogism. Terms common to all cases: animal, 25
white, horse; animal, white, stone.
Thus it is evident from what has been said that if there is a particular
syllogism in this figure, it is necessary for the terms to stand as we have said;
for if they stand in any other way, there is not. It is clear, too, that all the
syllogisms in it are perfect (for they are all perfected by means of the initial 30
assumptions), and that all problems are proved by means of this figure
(holding of every and of none and of some and not of some). I call this the first
figure.

(5) When the same item holds of all the one and of none of the other, or of all or 35
none of each, I call this the second figure. I say that in it the middle term is the
one predicated of both, the extremes are those of which this is said, and the
major extreme is the one supposed on the middle.41 The middle is posited
outside the extremes and first in position.
Now there will not be42 a perfect syllogism in this figure, but there will be a 27al
potential one both when the terms are universal and when they are not
universal. When they are universal, there will be a syllogism when the
middle holds of all the one and of none of the other, on whichever the privative 5
may be; and in no other case.
Let M be predicated of no N and of every O. Since the privative converts, N
will hold of no M. But M was supposed to hold of every O. Hence N of no O (this
has been proved earlier).
Again, if M holds of every N and of no O, O will hold of no N. For if M holds of 10
no O, O holds of no M. But M held of every N. Therefore O will hold of no N. For
we have got the first figure again. And since the privative converts, N will
hold of no O. Hence there will be the same syllogism. (It is also possible to
prove these by reduction to the impossible.) 15
Thus it is evident that there is a syllogism when the terms stand thus - but
not a perfect one. For the necessity is perfected not only from the initial
assumptions but also from other items.
If M is predicated of every N and O, there will not be a syllogism. Terms for
holding: substance, animal, man. For not holding: substance, animal, 20
number. (Substance is the middle.) Nor when M is predicated of no N and of

38
26b20: homoids de Alexcl, OCT; homoiosgar cj. Alex.
39
26b20: kai ei Alex1, OCT; kan Arist: ABd.
40
26b21: meros + ii Alex1, Arist: C; om OCT.
41
26b38: keimenon + elation de to porroterd tou mesou OCT; om Alex1, Arist: B1.
42
27al: sullogismos + oudamos OCT; om Alex1.
 204 Appendix 1
no O. Terms for holding: line, animal, man. For not holding: line, animal,
stone.
Thus it is evident that if there is a universal syllogism,43 it is necessary
25 for the terms to stand as we said at the beginning. For if they stand in any
other way, the necessity does not come about.
If the middle is universal in relation to one term, then if it is universal in
relation to the major, either predicatively or privatively, and particular in
relation to the minor and in an opposite way to the universal (by 'in an
30 opposite way' I mean that if the universal is privative the particular is
affirmative, and if the universal is predicative the particular is privative), it
is necessary for there to be a particular privative syllogism. For if M holds of
no N and of some O, it is necessary for N not to hold of some O. For since the
35 privative converts, N will hold of no M. But M was supposed to hold of some
O. Hence N will not hold of some O. For we get a syllogism by way of the first
figure. Again, if M holds of every N and not of some O, it is necessary for N
not to hold of some O. For if it will hold of44 every O, and M is predicated of
27bl every N, it is necessary for M to hold of every O. But it was supposed not to
hold of some. (And if M holds of every N and not of every O, there will be a
syllogism that N does not hold of every O. The demonstration is the same.)
5 If it is predicated of every 0 and not of every N, there will not be a
syllogism. Terms: animal, substance, raven; animal, white, raven. Nor
when of no O and of some N. Terms for holding: animal, substance, unit. For
not holding: animal, substance, science.
10 Thus when the universal is opposite to the particular, we have said when
there will be a syllogism arid when not. When the premisses are similar in
form - i.e. both privative or affirmative - there will in no case be a syllogism.
Let them first be privative, and let the universal be supposed on the major
15 extreme. I.e. let M hold of no N and not of some O. Now it is possible for N to
hold of every O45 and also of no O. Terms for not holding: black, snow,
animal. It is not possible to take terms for holding if M holds of some O and
not of some 0. For if N holds of every O and M of no N, M will hold of no O;
20 but it was supposed to hold of some. Thus in this case it is not possible to
take terms - it must be proved from the indeterminacy. For since 'M does
not hold of some O' is true even if it holds of none, and since there was no
syllogism when it held of none, it is evident that there will not be one in the
present case either.
Again, let them be predicative and let the universal be supposed in the
25 same way. I.e. let M hold of every N and of some O. Now it is possible for N
to hold of every O and also of no O. Terms for holding of none: white, swan,
stone. It is not possible to take terms for holding of all (the explanation is
the same as before) - it must be proved from the indeterminacy.
30 If the universal is on the minor extreme, and M holds of46 no O and not of
some N, it is possible for N to hold of every O and also of no O. Terms for
holding: white, animal, raven. For not holding: white, stone, raven.
If the premisses are predicative, terms for not holding are white, animal,
snow; for holding, white, animal, swan.
35 Thus it is evident that when the premisses are similar in form and one is
43
27a23: katholou + tdn hordn onion OCT; om Alex ut videtur.
44
27a38: huparxei Alex1, Arist: C; huparkhei OCT.
46
27bl5: toNtdiXMex!; tdiXtoNOCT, Alex: Aldina.
46
27b28-30: ean ...ei... huparxei Alex1; el... esti... huparkhei OCT, Alex: Aldina.
 Translation of Prior Analytics 1.1-7 205
universal and the other particular, in no case is there a syllogism.
Nor if it holds of some of each, or does not hold, or of some of one and not of
some of the other, or of all of neither, or indeterminately. Terms common to
all cases: white, animal, man; white, animal, inanimate.
Thus it is evident from what we have said that if the terms stand to one 28al
another in the way we have said, there is a syllogism by necessity, and that if
there is a syllogism it is necessary for the terms to stand in this way. It is
clear too that all the syllogisms in this figure are imperfect (for they are all 5
perfected when certain items are co-assumed which either inhere by
necessity in the terms or are posited as hypotheses - i.e. when we prove by
means of the impossible), and that there is no affirmative syllogism by
means of this figure but that all of them, both the universal and the
particular, are privative.
(6) When47 one term holds of all and the other of none of the same term, or 10
both of all or none, I call this the third figure; and I say that in it the middle
is that of which both are predicated, the extremes are the predicates, the
major is the extreme further from the middle, and the minor is the one
nearer. The middle is posited outside the extremes and is last in position. 15
There is no perfect syllogism in this figure either, but there will be a
potential one both when the terms are universal and when they are not
universal in relation to the middle.
If they are universal, then when both P and R hold of every S, P will hold
of some R by necessity. For since the predicative converts, S will hold of 20
some R. Hence since P holds of every S and S of some R, it is necessary for P
to hold of some R. For we get a syllogism by way of the first figure. It is also
possible to effect the demonstration by means of the impossible and by
exposition. For if both hold of every S, then if something of S - say N - is 25
taken, both P and R will hold of this. Hence P will hold of some R.
And if R holds of every S and P of none, there will be a syllogism that P
does not hold of some R by necessity. There is the same mode of
demonstration, the premiss RS being converted. It can also be proved by
way of the impossible, as in the earlier cases. 30
If R holds of no S48 and P of every S,49 there will not be a syllogism. Terms
for holding: animal, horse, man. For not holding: animal, inanimate, man.
Nor when both are said of no S will there be a syllogism. Terms for holding:
animal, horse, inanimate. For not holding: man, horse, inanimate. 35
(Inanimate is the middle.)
Thus it is evident in this figure too when there will be a syllogism and
when not, the terms being universal. When both the terms are predicative
there will be a syllogism that the extreme holds of some of the extreme; and
when they are privative there will not be one. When one is privative and the 28bl
other affirmative, then if the major is predicative and the other affirmative,
there will be a syllogism that the extreme does not hold of some of the
extreme; if they are the other way about, there will not be.
If one term is universal in relation to the middle and the other particular, 5
then if both are predicative, it is necessary for there to be a syllogism,
whichever of the terms is universal. For if R holds of every S and P of some
47
28alO: hotan Alex1; ean OCT.
48
28A30: medeni + tdi S Alex1, Arist: Cn2; om OCT
49
28a31: tdi S huparkMi Alex1 [tdi S om Aldina]; huparkhei tdi S OCT.
 206 Appendix 1
S, it is necessary for P to hold of some R. For since the affirmative converts, S
10 will hold of some P. Hence since R holds of every S and S of some P, R will hold
of some P. Hence also50 P of some R. Again, if R holds of some S and P of every
S,51 it is necessary for P to hold of some R. There is the same mode of
demonstration. It is also possible to demonstrate by means of the impossible
15 and by exposition, as in the earlier cases.52
If one is predicative and the other privative and the predicative is
universal, then when the minor is predicative there will be a syllogism. For if
R holds of every S and P does not hold of some S, it is necessary for P not to
20 hold of some R. For if of every, and R of every S, then P will hold of every S. But
it did not hold. It is also proved without the reduction if something of S is
taken of which P does not hold.
When the major is predicative, there will not be a syllogism - i.e. if P holds
of every S and R does not hold of some S. Terms for holding of every: animate,
25 man, animal. It is not possible to take terms for holding of none if R holds of
some S and does not hold of some. For if P holds of every S and R of some S,
then P will hold of some R. But it was supposed to hold of none. We must take
terms as in the earlier cases. For 'does not hold of some' is indeterminate, and
30 it is true to say of what holds of none that it does not hold of some. But if it
holds of none, there was no syllogism. So it is evident that there will not be a
syllogism.
If the privative is the universal term, then when the major is privative and
the minor predicative, there will be a syllogism. For if P holds of no S and R
35 holds of some S, P will not hold of some R. For again there will be the first
figure, the premiss RS being converted.
When the minor is privative, there will not be a syllogism. Terms for
holding: animal, man, wild. For not holding: animal, science, wild. (Wild is
middle in both cases.) Nor when both are posited as privative and one is
29al universal and the other particular. Terms when the minor is universal in
relation to the middle: animal, science, wild; animal, man, wild. When the
major, for not holding: raven, snow, white. It is not possible to take terms for
5 holding if R holds of some S and does not hold of some S. For if P holds of every
R and R of some S, P holds of some S. But it was supposed to hold of none. It
must be proved from the indeterminacy.
And if each holds of some of the middle, or does not hold of some, or if one
holds and the other does not hold, or one of some and the other not of all, or
indeterminately - there will not be a syllogism in any of these cases. Terms
10 common to all cases: animal, man, white; animal, inanimate, white.
Thus it is evident in this figure too when there will be a syllogism and when
there will not - that53 when the terms stand as has been said there is a
syllogism by necessity, and that if there is a syllogism it is necessary for the
15 terms to stand thus. It is evident too that all the syllogisms in this figure are
imperfect (for they are all perfected when certain things are co-assumed),
and that it is not possible to syllogize a universal, whether privative or
affirmative, by means of this figure.
20 (7) It is clear too that in all the figures, when there is no syllogism, if both the
50
28bll: hoste + kai Alexc, Arist: BC2A; om OCT.
51
28bl2: huparkhei + toiSMex1; om OCT.
52
28bl5: proteron Alexc, OCT; proteron Alex: M, Aldina, Arist: ACd.
53
29al2: ekhontdn + te OCT; om Alexc.
 Translation of Prior Analytics 1.1-7 207
54
terms are predicative or privative nothing necessary comes about at all; but
if they are predicative and privative, then if the privative is taken as
universal there is always a syllogism of the minor extreme in relation to the
major. E.g. if A holds of every or of some B and B of no C. For if the premisses 25
are converted, it is necessary for C not to hold of some A. Similarly in the
other figures too: there is always a syllogism by means of conversion.
It is clear too that if an indeterminate is posited instead of a particular
predicative, it will effect the same syllogism in all the figures.
It is evident too that all the imperfect syllogisms are perfected by means of 30
the first figure. For they are all brought to a conclusion either probatively or
by means of the impossible, and in both ways we get the first figure - when
they are perfected probatively, because they are all brought to a conclusion by
means of conversion and conversion produces the first figure; when they are 35
proved by means of the impossible, because when the falsity is posited the
syllogism comes about by means of the first figure. (E.g. in the last figure, if A
and B hold of every C, A holds of some B. For if of none and B of every C, then A
of no C. But it held of every C. Similarly in the other cases too.)
It is also possible to reduce55 all the syllogisms to the universal syllogisms 29bl
in the56 first figure. For the second figure syllogisms are evidently perfected
by them (though not all in the same way: the universals by conversion of the
privative, each of the particulars by means of reduction to the impossible). 5
The particular syllogisms in the first figure are indeed perfected by means of
themselves; but it is also possible to prove them by means of the second figure
by reducing them to the impossible. E.g. if A holds of every B and B of some C,
then A of some C. For if of none and of every B, then B will hold of no C - we 10
know this by means of the second figure. The demonstration will be similar in
the case of the privative syllogism. For if A holds of no B and B of some C, A
will not hold of some C. For if it holds of every C and of no B, B will hold of no C.
(This was the middle figure.) Hence since all the syllogisms in the middle 15
figure are reduced to the universal syllogisms in the first figure, and the
particular syllogisms in the first figure to the syllogisms in the middle figure,
it is evident that the particular syllogisms will be reduced to the universal
syllogisms in the first figure. The syllogisms in the third figure, when the 20
terms are universal, are perfected immediately by means of those syllogisms;
and when the terms are taken as particulars,57 they are perfected by means
of the particular syllogisms in the first figure - but these were reduced to the
former: hence so too are the particulars in the third figure. Thus it is evident
that all will be reduced to the universal syllogisms in the first figure. 25
Thus we have said how syllogisms which prove that something holds or
does not hold stand - both in their own right, when they are from the same
figure, and in relation to one another,58 when they are from different figures
(either by the middle term or inasmuch as they are reduced to the first
figure).59

54
29a21: ton horon ontdn Alex1; onton ton horon OCT.
55
29bl: anagein Alex1; anagagein OCT.
56
29b2: en + toi OCT; om Alex1.
57
29b21: epi merous Alex0; en merei OCT.
58
29b28: allelous Alex"1, OCT; allous Arist: ABd.
59
29b28: heteron + skhematon e toi mesdi e tdi anagesthai eis to prdton Alexcl; om
OCT [+ skhematon Arist: ABCn2].
 Appendix 2
Aristotelian Syllogistic

This Appendix contains, first, a brief outline of Aristotelian syllogistic, and

secondly, a list of the syllogistic forms which Aristotle and his successors
recognised.

I
The interpretation of Aristotle's syllogistic is controversial, and not only in
its details. Here we do not discuss the various rivals: we simply set down a
sketch of the system, without argument or defence - and also without any
attempt at technical sophistication.1
A syllogism, then, is a sort of argument; and an argument may be
represented as a sequence of propositions, one the conclusion and the rest
the premisses. Any argument, we may say, has the form:
Pi,P2 Pn: therefore, Q
- where the PIS are the premisses and Q is the conclusion. And an argument
is valid if its conclusion follows from its premisses.
An Aristotelian syllogism is a valid argument of a particular sort. First, a
syllogism is an argument which has exactly two premisses. In other words,
it has the form:
PI, P2: therefore, Q.
Moreover, the conclusion, Q, must be a different proposition from each of
the premisses; and the premisses must be different from each other.
Each of the three component propositions of a syllogism must exhibit a
particular structure; for each must be a 'subject-predicate' proposition. A
proposition is a 'subject-predicate' proposition if it contains two terms, A
and B, one of which (the predicate term) is said of the other (the subject
term). Thus 'Horses are animals' predicates animal of horses;2 and 'Dogs
are not cats' predicates cats (negatively) of dogs. If we use'... holds of ' to
1
For fuller versions of the account we outline, see e.g. Smiley; Corcoran. In
adopting this account we implicitly reject the interpretation advanced by
Lukasiewicz and elaborated by Patzig: it is nevertheless true, as our notes to the
translation suggest, that Lukasiewicz and Patzig have contributed more than any
other scholars to our understanding of Aristotle's syllogistic.
2
Or should we rather say that it predicates 'animal' of horses, or 'animal' of
'horses'? The ancients rarely saw any need for precision on this and allied points; and
we shall imitate them. Modern readers will think this a frightful muddle - rightly.
But it is a muddle which, or so we think, does not affect the presentation of the
syllogistic.

208
 Aristotelian Syllogistic 209
express the predicative tie, we may say that the component propositions of a
syllogism all have at their heart the form:
A holds of B
- where A and B are terms, A the predicate and B the subject.
Subject-predicate propositions also have a 'quality' and a 'quantity' and a
'mode'. Their quality consists in the fact that they are either affirmative or
negative: either A holds affirmatively of B or A holds negatively of B. Their
quantity consists in the fact that they are either universal or particular:
either A holds of every B or A holds of some B.3 Their mode consists in the
fact that the predicate is said to be tied to the subject either necessarily or
actually or possibly: either A holds necessarily of B or A holds actually of B
or A holds possibly of B.
Putting these three features together, we may say that an Aristotelian
subject-predicate proposition has a form which could be clumsily
represented thus:
8[A holds+/-of e/sB]
- where 8 marks the mode,'+/' indicates the quality, 'e/s' stands for the
quantity ('every', 'some').
In An. Pr. 1.4-7 Aristotle investigates syllogisms in which all the
component propositions have the neutral mode marked by 'actually1. Hence
in the rest of this Appendix we shall ignore modal differences and drop the
modal sign from our representation of the form of subject-predicate
propositions.
That being so, it is clear that there are exactly four sorts of
subject-predicate proposition: universal affirmatives, of the form:
A holds of every B;
universal negatives, of4 the form:
A holds of no B;
particular affirmatives, of the form:
A holds of some B;
and particular negatives, of the form:
A does not hold of some B.
Three distinct propositions are involved in any syllogism; but the three
propositions must be interrelated. In particular, the two premisses must
have a term in common (the so-called 'middle' term); and the two terms in
the conclusion (the 'extremes') must each appear in one of the premisses.
Thus the three propositions of any syllogism exhibit altogether precisely
three distinct terms.
It is plain that the middle term may collaborate with the other two terms
in the premisses in any one of three different ways: it may be subject for one
extreme and predicate for the other; it may be predicate for both extremes;
or it may be subject for both. In other words, the premisses may, as a pair,
show any one of the following core structures:
I A holds of B,B holds of C.
II B holds of A, B holds of C.
III A holds of B, C holds of B.
3
Aristotle sometimes also refers to 'indeterminate' propositions, i.e. to propositions
which mark no quantity. Since such propositions are treated as equivalent to
particular propositions, we may ignore them here. See 30,31 note.
4
Note that the word 'no' here combines two functions: it indicates at once the
quality - negative - and the quantity - universal - of the proposition.
210 Appendix 2
These three structures determine the three 'figures' of Aristotelian
syllogistic: a syllogism belongs to the first figure if its premisses exhibit
structure I, to the second figure if its premisses exhibit structure II, and to
the third figure if its premisses exhibit structure III.5
Since each premiss may have any of four subject-predicate forms, there
are sixteen premiss pairings with each structure. Of the sixteen with
structure I, four are shown in An. Pr. 1.4 to yield conclusions, so that the
first figure contains four syllogistic forms. Of the sixteen pairings with
structure II, An. Pr. 1.5 establishes that four yield syllogistic conclusions, so
that the second figure also contains four syllogistic forms. The third figure,
analysed in An. Pr. 1.6, has six syllogisms.
The first figure syllogisms are 'perfect'; that is to say, their validity is
evident. Thus the first syllogistic form in this figure, which is known as
Barbara6, looks like this:
(1) A holds of every B
(2) B holds of every C
(3) A holds of every C
And it is, or so we are invited to think, evident that (3) follows from (1) and
(2).
The syllogisms in the other figures are 'imperfect'; but they can be
'perfected' inasmuch as they can be reduced to first figure syllogisms.
Aristotle employs three methods of perfection or reduction - in effect, he has
three ways of proving that a given syllogistic form is valid. One method
relies on 'conversion', a second on reduction to the impossible, and the third
on 'exposition'.
Conversion is an operation on propositions. To convert a proposition of
the form 'A holds of B', you interchange the terms A and B to get 'B holds of
A'. Aristotle recognises three conversion rules, which he establishes in
An.Pr. 1.2. They may be set out thus:
(CD From 'A holds of no B' infer 'B holds of no A'.7
(C2) From 'A holds of some B' infer 'B holds of some A'.
(C3) From 'A holds of every B' infer 'B holds of some A'.
A perfection by way of conversion works in the following way. Suppose as
premisses that:
(1)A holds of no B,
and:
(2) A holds of every C.
Then from (1), by (CD, infer:
(3) B holds of no A.
And now, by the second syllogism in the first figure (Celarent), we may infer
from (3) and (2) to:
(4) B holds of no C.
B
Later versions of syllogistic produce four figures; and there is a large - and largely
futile - literature on the origin and value of the fourth figure. See 47,12 note.
6
For the 'names' of the syllogistic forms see below, pp. 212-15.
7
Strictly speaking we should write:
(Cl*) From a proposition of the form 'A holds of no B' infer the corresponding
proposition of the form 'B holds of no A'.
But here and hereafter we shall follow the harmlessly inaccurate custom of citing
prepositional forms as though they were propositions and syllogistic forms as though
they were syllogisms.
 Aristotelian Syllogistic 211
Thus we have shown that (4) follows from (1) and (2). A syllogism with
premisses of the forms (1) and (2) and with a conclusion of the form (4) is in
the second figure - it is a syllogism of the form Cesare.8
The method of reduction to the impossible works as follows. Suppose as
premisses:
(1) A does not hold of some B.
(2) C holds of every B.
Hypothesize in addition that:
(3) A holds of every C.
Then, by Barbara, we may infer from (3) and (2) to:
(4) A holds of every B.
But (4) is 'impossible'; or rather, (4) is inconsistent with premiss (1). Hence
if both (1) and (2) are true, (3) cannot be true. Hence if both (1) and (2) are
true, it follows that the contradictory of (3) is true, i.e. that:
(5) A does not hold of some C.
Thus we have shown that (5) follows from (1) and (2). A syllogism with
premisses of the forms (1) and (2) and with a conclusion of the form (5) is in
the third figure it is a syllogism of the form Bocardo.9
The third method, the method of exposition, is more controversial.10
Alexander's interpretation, which has been generally rejected on
inadequate grounds, is roughly this.11 Suppose as premisses that:
(1)A holds of every B.
(2) C holds of every B.
Now 'expose' or set out some particular B, say b. Then from (1) we may infer:
(3) A holds of b,
and by (2) we may infer:
(4) C holds of b.
But from (3) and (4) it follows that A holds of something of which C also
holds. Hence that:
(5) A holds of some C.
Thus we have shown that (5) follows from (1) and (2). A syllogism with
premisses of the forms (1) and (2) and with a conclusion of the form (5) is in
the third figure - it is a syllogism of the form Darapti.12
8
Some mav prefer a more formal presentation of the argument, thus:
1(1) A holds of no B prem
2(2) A holds of every C prem
1(3) B holds of no A 1,(C1)
1.2(4) B holds of no C 3,2Celarent
9
More lormally:
1(1) A does not hold of some B prem
2(2) C holds of every B prem
3(3) A holds of every C hyp
2,3 (4) A holds of every B 3,2 Barbara
1,2(5) A does not hold of some C 1,2,3,4 reductio
10
11
See 32,33 note.
In fact Alexander thinks that Aristotle uses two distinct types of exposition: see
100,17
12
note.
The formal version requires the interpolation of of few steps:
1(1) A holds of every B prem
2(2) C holds of every B prem
1(2*) B holds something 1
2+(2+) B holds of b exp
1,2+(3) A holds of b 1,2+
212 Appendix 2

Of the forty-eight possible premiss pairings, An. Pr. 1.4-6 proves fourteen
syllogistic. Moreover, Aristotle also proves that none of the remaining
thirty-four pairings yields a conclusion of the form 'A holds of C'. (The
method by which he proves pairings to be non-syllogistic has been described
in the Introduction.13) In 1.7, Aristotle signals two further developments.
He indicates a way of reducing all syllogisms to the first two syllogisms of
the first figure (i.e. to Barbara and Celarent); and he observes that there are
certain other syllogistic forms in addition to the fourteen which have been
set out in 1.4-6. For the former development it should be enough to con
Alexander's close commentary; for the latter we may refer to the long note
to 70,21 and to the descriptions in the second part of this Appendix.

II
Against each syllogistic form in the following list we have set its standard
mediaeval name. (Like most writers on syllogistic, we have found it
convenient to refer to the syllogistic forms by their traditional names.) The
names, which were collected into halting hexameters, are significant.14 The
first three vowels in each name indicate the quantity and quality of the
premisses and conclusion of the syllogism. (A marks a universal affirmative,
E a universal negative, I a particular affirmative, O a particular negative.)
The initial consonant indicates to which first figure syllogism a derived
syllogism may be reduced: a syllogism beginning with 'B' reduces to
Barbara, with 'C' to Celarent, with 'D' to Darii, with 'F' to Ferio. 'S' indicates
that the syllogism can be reduced if the premiss marked by the preceding
vowel is converted by an application of rule (Cl) or (C2); 'P' indicates a
conversion by rule (C3). A 'C' - except an initial 'C' - indicates the need for
reduction to the impossible. The other letters in the names have no
significance.15
We give first the standard fourteen syllogisms, in their three figures and
in the order in which Aristotle gives them.16
First figure
Barbara Darii
A holds of every B A holds of every B
B holds of every C B holds of some C
A holds of every C A holds of some C

1,2+(4) C holds of b 2,2+

1,2+(5) A holds of some C 3,4
1,2 (5*) AholdsofsomeC 2*,2+,5EE
(Why does (2*) follow from (1)? Because (within Aristotle's logic) if A holds of every B,
then A holds of some B; hence something is B.)
13
Above, pp. 12-13.
14
See e.g. Bochenski (1956), pp. 244-50; Thorn, p. 54.
15
Except 'M', which indicates that the premisses must have their order changed -
but this is not a logical operation.
16
Note that the later Peripatetics changed the Aristotelian order of the syllogisms
in the third figure: see 97,12 note.
 Aristotelian Syllogistic 213
Celarent Ferio
A holds of no B A holds of no B
B holds of every C B holds of some C
A holds of no C A does not hold of some C
Second figure
Cesare Festino
M holds of no N M holds of no N
M holds of every O M holds of some O
N holds of no O N does not hold of some O
Camestres Baroco
M holds of every N M holds of every N
M holds of no O M does not hold of some O
N holds of no O N does not hold of some O
Third figure
Darapti Datisi
P holds of every S P holds of every S
R holds of every S R holds of some S
P holds of some R P holds of some R
Felapton Bocardo
P holds of no S P does not hold of some S
R holds of every S R holds of every S
P does not hold of some R P does not hold of some R
Disamis Ferison
P holds of some S P holds of no S
R holds of every S R holds of some S
P holds of some R P does not hold of some R
Next, we list the additional syllogisms divided into the three groups which
we distinguish in the note to 70,21. The roman numeral in brackets after
the name of each syllogism indicates which figure it belongs to.
GroupA
Baralipton [I] Cesares [II]
A holds of every B M holds of no N
B holds of every C M holds of every O
C holds of some A O holds of no N
214 Annenri.ir. 2
Dabitis [I] Camestre [II]
A holds of every B M holds of every N
B holds of some C M holds of no O
C holds of some A O holds of no N
Cesares [II] Daraptis [III]
M holds of no N P holds of every S
M holds of every O R holds of every S
O holds of no N R holds of some P
Disami [III] Datisis [III]
P holds of some S P holds of every S
R holds of every S R holds of some S
R holds of some P R holds of some P
Group B
Fapesmo [I] Frisesomorum [I]
A holds of every B A holds of some B
B holds of no C B holds of no C
C does not hold of some A C does not hold of some A
Firesmo [II] Fapemo [III]
M holds of some N P holds of every S
M holds of no O R holds of no S
O does not hold of some N R does not hold of some P
Frisemo [III]
P holds of some S
R holds of no S
R does not hold of some P
GroupC
Barbari [I] Celaront [I]
A holds of every B A holds of no B
B holds of every C B holds of every C
A holds of some C A does not hold of some C
Celantos [I] Cesaro [II]
A holds of no B M holds of no N
B holds of every C M holds of every O
C does not hold of some A N does not hold of some O
 Aristotelian Syllogistic 215
Camestrop [II] Faresmo [II]
M holds of every N M holds of every N
M holds of no O M holds of no O

N does not hold of some O O does not hold of some N

Cesaros [II]
M holds of no N
M holds of every O
O does not hold of some N
 Appendix 3
Textual Notes

This appendix collects together the various textual suggestions

which are scattered in the footnotes to the translation.

2,5 Delete TO uopiov.

3,6 Delete 6v.
3,12 Read ev adxi] for EV aOtou;.
3,25 Read nepvnov, with the MSS. (Wallies prints
Tepnvov, after the Aldine edition.
4,18 Read npo<; TOUTO for npoc; TOV>T<I}.
7,30 Remove the comma after npoTdoeu; and add a comma
after TO efvm (with Wallies, in Top., p. 711).
7,31 Read nou CHJTOV (Aldine) for TIVQ QUTCDV.
8,3 Read dei (conjectured by Wallies) for 6eiv.
9,6 Read npo0ei<; for npooSeiq.
9,17 Read nepi (Aldine) for nav.
9,22 Perhaps read 8id [or UETQ] dnoSei^ecoc; for nepi
dnoSei^ecoc;?
11,8 Read dvruj>doeoo<; for dnocjxivoecoc; (see 11,10-11).
11,22 Retain TI before KQTQ, with the MSS (Wallies deletes).
13,27 Read Kai TI uq, with the MSS. (Wallies prints q uq,
after the Aldine edition and our MSS of Aristotle.)
14,15 Omit Km before v86E;ou, with LM (see Aristotle, Top.
100b23-5.)
15,18-19 Add TO before UETQ TOU ^oriv.
15,23 Read npoonOepevou for ouvTi0ejjevou.
16,14 Retain Aeyovro, with the MSS. (Wallies emends to
XeyoiTO.)
17,4 Add fj v if[ ^pwTqoei after evToAf] (see p. 16 n.83).
18,3 Read TfiAAa for TauTa.
18,16 Read quepa, with B and the Aldine edition (and
Wallies, in Top., p.711) for $&}<;.
19,16 Read TXO ^Tlpw for T0 enojievcp.
19,22 Delete Tfj<; ouAAoyionKqc; ouunAoKqt;.
20,13 Transpose ou ydp . .. SQov el (20,16-18) to follow
napexETCti.
216
 Textual Notes 217
20,21-2 Place f| uotAAov.. . yivetai between parentheses.
22,4 Delete Oewpruia.
22,11-12 Delete TCDV dvTtKEiuEvcov.
23,2 Read dnev for elnov (see Aristotle, Soph.El. 168b22-6).
25,17 Omit Eineiv nepi TOU Kara uq8v6<;, after the Aldine
edition.
29,24 Read TECDC; for ncoc;.
31,13-14 Read evapycov ouAAoyicmKCDt; for Evapyax; [evapyov,
M] ouAAoyioTiKwv.
31,18 Read nAetcna, with LM and the Aldine, for nAeiorw.
33,8 Omit TOUTCO (with L). (M offers TO, the Aldine TOUTO.)
35,6 Read auTrj, with B and the Aldine (if Wallies reports
correctly). Wallies prints ctuTr[; L and M have eamft.
36,6 Read KaTct4>doax;, with LM and the Aldine, for
dvTuJ>do(Dc; (B, followed by Wallies).
37,9 Read ouSevi v8exeTcu, with L and B2 (ev8exeo6ai, M),
for ou8evi avayicrj.
37,13 Read pq8evi ev8exeTai for jiqSEvi dvayKi].
37,16 Read TQUTJI for TOUTI]?.
39,19 Perhaps read 8uvaToO for evSexojievou?
41,28 Read dno^anKfi (Aldine) for Kaxa^aTiKi].
42,8-9 Read oxqjJta for oxrijidTcov.
43,24 Read 6 8e . . . nicrav: 6 8e. . . morov, MSS; 6 Srj . . .
nicrtov, Wallies.
44,4 Read Kai ncoc; for ncoc;.
47,9-10 Delete tic, expo ouoia ejuj/uxoc; EOTIV.
54,16 Read Aapetv after ecru, with M and the Aldine edition
(Wallies omits).
55,11 Read ouAAoyioriKcov for ouAAoyiojaov TCOV.
56,24 Omit mi TO ^wov KQT' ou8evo<; AiOou: the phrase,
printed by Wallies, is found in the Aldine edition but
not in the MSS.
62,2 Read KOI for f\.
64,17 Read EKKeijievqv: dviiKEiuevqv, MSS; EipqjiEvqv,
Wallies after the Aldine edition.
66,20 Delete Tcp before jjiqSEvi.
67,7 Delete uq before undpxoi (see 66,2-3).
68,26 Read aikd for auTfjv.
70,6 Read TEAEOV for TEAeioi, with Wallies (in his
corrigenda, p.426).
72,11 Read AE^ECOC;, with LM and the Aldine, for Td^ecoc;.
73,8 Read t(p o for TU 81' 6v.
73,14-15 The text is corrupt, but no convincing cure suggests
itself. Exempli gratia add EKEIVO SE before
EyyuiEpw and delete 6 i'nnoc; fjnEp avGpconoc;.
218 Appendix 3
74,11 Read CGOTE for OUTCOC;.
78,28 Read 6 8e (conjectured by Wallies, in the Index
Verborum to in Top.) for 68e.
80,15 Read dpct for yap.
81,22 Read ev yevoc;: ev yeveoi, MSS; EV yeveoei, Aldine; EV
yevei, Wallies.
81,26 Follow LM and the Aldine edition: after napaSeiy-
jiaoiv B has dAA' coonEp, with TE written above nep;
Wallies prints dAAcoq TE.
85,2 Retain EXOVTCOV which Wallies emends to EXOUOCOV.
85,27 Add a after KCITO id jaopia OUTOU (Wallies prints KCCTCI
d jiopia auxoO).
86,12 Perhaps read dvojioicov for dvo^oiooxqii6\'0)\' (see
85,17).
86,30-87,1 Delete TO douAAoyiotov.
90,19 Add TOU JJEV jiq8Ev( before dvGpcono^.
90,25-6 Read TO A TIVI T^ T (with M and the Aldine) for TO
npWTOV TlVl TO) TplTCp.
91,1-3 Remove Wallies' parentheses.
91,3 Perhaps read dnaywyq for ouvaycayfj?
91,6 Omit yap, added by Wallies.
92,10 Retain prjSE (jiq8Ev6<;, Wallies).
95,17-18 Read TCX dAAa oxrjjiaTa, TO TE SEUTEpov AEyco KQI TO
Tpuov for TWV dAAcov oxqpaTcov, TOU TE SsuTEpou AEyco
KQl TOU TplTOU.
100,10 Delete TO EKKEIJJEVOV, which Wallies emends to TCO
EKKElJlEVCp.
100,15 Delete 8rtTov TO }JLEOOV EOTI <KO{>.
104,1 Read TI for TOUTO (TV TOIOUTO, Aldine).
112,14 Read npOKEijiEvou for KEIJIEVOU.
112,20-1 Remove the commas after yivETai and E'SEI^EV; add a
comma after TpiTco.
113,12 Read ojaoAoyou^evwc; ouAAoyicrciKcov for opoAo-
youjiEvcov.
114,27 Read Suo TOUC;, with the Aldine edition: 8uo TOUTOVX;,
MSS; 8uo TOUTOu TOVX?, Wallies.
116,33 Perhaps read Einov for EIHEV (see 103,16-19)?
 Texts and Abbreviations

The list has a double purpose: it expands the various abbreviations which
appear in the notes to the Introduction and the Translation; and it indicates
which editions of the ancient texts we have taken as standard.
Albinus
Didasc. Didascalicus [ed. Hermann, in vol. VI of the
Teubner Plato]
Alexander
in An. Pr. Commentary on Aristotle's Prior Analytics [ed.
Wallies, CAG III]
Conv.1 On Conversion [ed. Badawi (1971)]
DA On the Soul [ed. Bruns, CAG suppt II1]
Fat. On Fate [ed. Bruns, CAG suppt II2]
in Metaph. Commentary on Aristotle's Metaphysics [ed. Hay-
duck, CAG I]
Quaest. Questions [ed. Bruns, CAG suppt II2]
in Top. Commentary on Aristotle's Topics [ed. Wallies, CAG
112]
[Alexander]2
in Metaph. Commentary on Aristotle's Metaphysics [ed. Hay-
duck, CAG I]
inSE Commentary on Aristotle's Sophistical Refutations
[ed. Wallies, CAG II3]
Ammonius
in An. Pr. Commentary on Aristotle's Prior Analytics [ed.
Wallies, CAG IV 6]
in Int. Commentary on Aristotle's On Interpretation [ed.
Busse,CAGIV5]
in Porph. Isag. Commentary on Porphyry's Introduction to Philo-
sophy [ed. Busse, CAG IV 3]
[Ammonius]
in An. Pr.3 Commentary on Aristotle's Prior Analytics [ed.
Wallies, CAG IV 6]

1
We have used an English translation generously made for us by Mr. Robert
Hoyland of Pembroke College, Oxford.
2
Square brackets around an author's name mark inauthenticity: the pseudo-
Alexandrian commentary on SE is rightly ascribed to Michael of Ephesus.
3
References using upper case roman numerals pick out texts printed in the
Introduction to Wallies' edition.

219
220 Texts and Abbreviations
anonymous
in Int. An. Pr. Commentary on Aristotle's On Interpretation and
Prior Analytics [ed. Mynas (Paris, 1844)]
in Theaet. Commentary on Plato's Theaetetus [edd. Diels-
Schubart]
Logica Logic and the Quadrivium [ed. Heidberg (Copen-
hagen, 1929)]
Apollonius Dyscolus
Conj. On Conjunctions [ed. Schneider, Grammatici Graeci
III]
Synt. Syntax [ed. Schneider, Grammatici Graeci II2]
Apuleius
Dog. Plat. On Plato and his Doctrines [ed. Beaujeu, Bud6]
Int. On Interpretation [ed. Thomas, Teubner]
Aristotle
An. Pr. Prior Analytics [ed. Ross, OCT]
An. Post. Posterior Analytics [ed. Ross, OCT]
Gael. On the Heavens [ed. Allan, OCT]
Cat. Categories [ed. Minio-Paluello, OCT]
DA On the Soul [ed. Ross, OCT]
EE Eudemian Ethics [edd. Walzer-Mingay, OCT]
EN Nicomachean Ethics [ed. Bywater, OCT]
GC On Generation and Corruption [ed. Joachim (Oxford,
1922)]
HA History of Animals [ed. Dittmeyer, Teubner]
Int. On Interpretation [ed. Minio-Paluello, OCT]
Metaph. Metaphysics [ed. Ross (Oxford 1924)]
Meteor. Meteorology [ed. Fobes (Cambridge Mass., 1918)]
PA Parts of Animals [ed. Peck, Loeb]
Phys. Physics [ed. Ross, OCT]
Pol. Politics [ed. Dreizehnter (Munich, 1970)]
Rhet. Rhetoric [ed. Kassel (Berlin, 1976)]
SE Sophistical Refutations [ed. Ross, OCT]
Top. Topics [ed. Brunschwig, Bud6; Ross, OCT]
[Aristotle]
Probl. Problems [ed. Bekker (Berlin, 1831)]
Aristoxenus Harmonics [ed. da Rios (Rome, 1954)]
Boethius
in Cic. Top. Commentary on Cicero's Topics [edd. Orelli-Baiter, in
vol. V ofCiceronis Opera (Zurich, 1833)]
in Int. Commentary on Aristotle's On Interpretation [ed.
Meiser, Teubner]
Int. Syll. Cat. Introduction to Categorical Syllogisms [ed. Migne,
Patrologia Latina 64]
Syll. Cat. Categorical Syllogisms [ed. Migne, Patrologia Latina
64]
CAG Commentaria in Aristotelem Graeca
Cassius Dio Histories [ed. Boissevain (Berlin, 1895-1901)]
Cicero
Nat. Deorum On the Nature of the Gods [ed. Ax, Teubner]
CMG Corpus Medicorum Graecorum
 Texts and Abbreviations 221
David
in Porph. Isag. Commentary on Porphyry's Introduction to Philo-
sophy [ed. Busse, CAG XVIII2].
Dexippus
in Cat. Commentary on Aristotle's Categories [ed. Busse,
CAG IV 2]
Diodorus Siculus History [ed. Dindorf, Teubner]
Diogenes Laertius Lives of the Philosophers [ed. Long, OCT]
Elias
in. An. Pr. Commentary on Aristotle's Prior Analytics [ed. West-
erink(1961)]
in Cat. Commentary on Aristotle's Categories [ed. Busse,
CAG XVIII1]
Proleg. Prolegomena [ed. Busse, CAG XVIII1]
Epictetus
Diss. Discourses [ed. Schenkl, Teubner]
Euclid Elements [ed. Heidberg/Stamatis, Teubner]
Eudemus
FnWehrli fragment n in Wehrli
Eusebius
PE Preparation for the Gospel [ed. Mras/des Places, Die
griechischen Christlichen Schriftsteller VIII1-2]
Eustathius
inOd. Commentary on Homer's Odyssey [ed. Stallbaum
(Leipzig, 1825/6)]
Eustratius
in An. Post. Commentary on Aristotle's Posterior Analytics [ed.
Hayduck, CAG XXII]
FDS Fragmente zurstoischen Dialektik [see Bibliography:
Hiilser]
Galen
Anat. Admin, Anatomical Procedures [ed. Kiihn (Leipzig, 1821-33)]
in Hipp. Aph. Commentary on Hippocrates' Aphorisms [ed. Kiihn
(Leipzig, 1821-33)]
in Hipp. Fract. Commentary on Hippocrates' On Fractures [ed. Kuhn
(Leipzig, 1821-33)]
in Hipp. Off. Commentary on Hippocrates' On the Surgery [ed.
Kuhn (Leipzig, 1821-33)]
Inst. Log. Introduction to Logic [ed. Kalbfleisch, Teubner]
Lib. Prop. On My Own Books [ed. von Miiller, Teubner]4
Meth. Med. On Therapeutic Method [ed. Kuhn (Leipzig, 1821-33)]
Opt. Sect. On the Best Sect [ed. Kuhn (Leipzig, 1821-33)]
PHP On the Doctrines of Hippocrates and Plato [ed. de
Lacey, CMG]5
Sem. On Semen [ed. Kuhn (Leipzig, 1821-33)]
Simp. Med. Temp. On the Mixture and Powers of Simple Drugs [ed.
Kuhn (Leipzig, 1821-33)]
Gellius Attic Nights [ed. Marshall, OCT]
Herodotus Histories [ed. Hude, OCT]
4
References are given by Kiihn numbering.
5
References are given by Kuhn numbering.
222 Texts and Abbreviations
lamblichus
Comm. Math. Sc. On the Common Science of Mathematics [ed. Festa,
Teubner]
Martianus Capella The Marriage of Philosophy and Mercury [ed. Dick,
Teubner]
Ocellus Lucanus On the Nature of the Universe [ed. Harder (Berlin,
1926)]
OCT Oxford Classical Texts
Olympiodorus
in Gorg. Commentary on Plato's Gorgias [ed. Westerink,
Teubner]
in Meteor. Commentary on Aristotle's Meteorology [ed. Stuve,
CAGXII2]
Pappus Collection [ed. Hultsche (Berlin, 1976-8)]
Papyri
PFayum Fayum Towns and their Papyri [edd. Grenfell-Hunt-
Hogarth (London, 1900)]
PGen Les Papyrus de Geneve [ed. Nicole (Geneva,
1896-1900)]
PHeid [see Bibliography: Carlini, 1978]
PMon [see Bibliography: Carlini, 1986]
POxy The Oxyrhynchus Papyri [edd. Grenfell-Hunt, et al.
(London, 1898-)]
Philodemus
Rhet. Rhetoric [ed. Longo Auricchio (Naples, 1977)]
Philoponus
in An. Pr. Commentary on Aristotle's Prior Analytics [ed.
Wallies,CAGXIII2]
in An. Post, Commentary on Aristotle's Posterior Analytics [ed.
Wallies,CAGXIII3]
in Cat. Commentary on Aristotle's Categories [ed. Busse,
CAGXIII1]
in DA Commentary on Aristotle's On the Soul [ed.
Hayduck,CAGXV]
Plato6
Rep. Republic [ed. Bury, OCT]
Plutarch
Ser. Num. On God's Slowness to Punish [edd. Pohlenz-
Sieveking, Teubner]
Stoic. Rep. On Stoic Self-contradictions [ed. Cherniss, Loeb]
Porphyry
Vit. Plot. Life ofPlotinus [edd. Henry-Schwyzer, in vol. I of the
OCT Plotinus]
Posidonius
FrcEK fragment n in Edelstein-Kidd
Proclus
inEucl. Commentary on the First Book of Euclid's Elements
[ed. Friedlein, Teubner]
in Rep. Commentary on Plato's Republic [ed. Kroll,
Teubner]
8
For all Plato's works we have used Bury's OCT.
 Texts and Abbreviations 223
in Tim. Commentary on Plato's Timaeus [ed. Diehl,
Teubner]
Ptolemy
Synt. Almagest [ed. Heiberg, Teubner]
scholia to Aristotle [ed. Brandis, in vol IV of Aristotelis Opera, ed.
Bekker (Berlin, 1836)]
Seneca
Ep. Letters [ed. Reynolds, OCT]
Sextus
M Against the Mathematicians [ed. Mutschmann/Mau,
Teubner]
PH Outlines of Pyrrhonism [ed. Mutschmann/Mau,
Teubner]
Simplicius
in Cat. Commentary on Aristotle's Categories [ed. Kalbflei-
sch, CAG VIII]
in DA Commentary on Aristotle's On the Soul [ed.
Hayduck.CAGXI]
in Phys. Commentary on Aristotle's Physics [ed. Diels, CAG
IX-X]
SVF Stoicorum Veterum Fragmenta [see Bibliography:
von Arnim]
Syrianus
in Metaph. Commentary on Aristotle's Metaphysics [ed. Kroll,
CAG VII]
Themistius
Max. Reply to Maximus [ed. Badawi (1971); French trans.
inBadawi(1987)]
Theophrastus
F n Graeser fragment n in Graeser
 Bibliography

The bibliography lists all the books and articles referred to in the
Introduction and in the notes to the translation. A full bibliography on
Alexander, up to 1986, can be found in Sharpies (1987); new items may be
tracked down in the regular bibliographical sections of the journal Elenchos.
Paolo Accattino: 'Alessandro di Afrodisia e Aristotele di Mitilene', Elenchos
1,1985,67-74
J. L. Ackrill: review of Patzig, Mind 71,1962,107-117
Allan Back: 'Philoponus on the Fallacy of Accident', Ancient Philosophy 7,
1987,131-146
Abdurrahman Badawi: Commentaires sur Aristote perdus en grec (Beirut,
1971)
La transmission de la philosophic grecque au monde arabe (Paris,
19872)
Jonathan Barnes: 'Aristotle's Theory of Demonstration', Phronesis 14,1969,
123-152 [reprinted in J. Barnes, M. Schofield and R. Sorabji (edd.),
Articles on Aristotle I (London, 1975)]
Aristotle's Posterior Analytics (Oxford, 1975)
review of Hintikka (1973), Journal of Hellenic Studies 1977, 97,
183-186
[1980a]: 'Proof Destroyed', in M. Schofield, M. F. Burnyeat and J.
Barnes (edd.), Doubt and Dogmatism (Oxford, 1980)
[1980b]: 'Aristotle and the Methods of Ethics', Revue Internationale de
Philosophic 133,1980,490-511
'Proof and the Syllogism', in E. Berti (ed.), Aristotle on Science: the
Posterior Analytics (Padua, 1981)
'Terms and Sentences', Proceedings of the British Academy 69, 1983,
279-326
[1985a]: Theophrastus and Hypothetical Syllogistic', in J. Wiesner
(ed.), Aristoteles: Werk und Wirkung I (Berlin, 1985)
[1985b] 'Uma terceira espcie de silogismo: Galeno e a logica das
relacdes', Andlise 2,1985,35-61
'Peripatetic Negations', Oxford Studies in Ancient Philosophy 4,1986,
201-214
'Bits and Pieces', in J. Barnes and M. Mignucci (edd.), Matter and
Metaphysics (Naples, 1988)
[1990a]: 'Logical Form and Logical Matter', in A. Alberti (ed.), Logica,
Mente e Persona (Florence, 1990)
[1990b]: 'Galen and the Utility of Logic', in J. Kollesch (ed.),
Proceedings of the Fourth International Galen Congress (Berlin, 1990)
[1990c]: The Toils of Scepticism (Cambridge, 1990)

224
 Bibliography 225
[1991]: 'Ammonius and Adverbs', in H. Blumenthal and H. M.
Robinson (edd.). Aristotle and his Successors, Oxford Studies in Ancient
Philosophy, supplementary volume 1991
Albrecht Becker: Die aristotelische Theorie der Moglichkeitsschlufle (Berlin,
1933)
Oskar Becker: 'Uber die vier Themata der stoischen Logik', in his Zwei
Untersuchungen zur Antiken Logik (Wiesbaden, 1957)
Linos G. Benakis: 'Commentaries and Commentators on the Logical Works of
Aristotle in Byzantium', in R. Claussen and R. Daube-Schackat (edd.),
Gedankenzeichen: Festschrift fur Klaus Oehler zum 60. Geburtstag
(Tubingen, 1988)
I. M. Bochenski: La Logique de Theophraste (Fribourg, 1947)
Ancient Formal Logic (Amsterdam, 1951)
Formale Logik, Orbis Academicus III 2 (Freiburg/Munich, 1956)
F. Bomer: 'Der Commentarius', Hermes 81,1953, 210-250
Jacques Brunschwig: 'Observations sur les manuscrits parisiens des Topi-
ques', in G. E. L. Owen (ed.), Aristotle on Dialectic (Oxford, 1968)
'La proposition particuliere et les preuves de non-concluance chez
Aristote', Cahierspour I'Analyse 10,1969,3-26 [reprinted in A. Menne and
N. Offenberger (edd.), Zur modernen Deutung der aristotelischen Logik I
(Hildesheim, 1982)]
'L'objet et la structure des Seconds Analytiques d'apres Aristote', in
E. Berti (ed.), Aristotle on Science: the Posterior Analytics (Padua, 1981)
' "Indetermine" et "Indefini" dans la logique de Theophraste', Revue
philosophique de la France et de I'Etranger 172,1982,359-370
Walter Burkert: Lore and Science in Ancient Pythagoreanism (Cambridge
Mass., 1972)
- "Xenarchos statt Poseidonios: zu Pap. Gen. inv. 203', Zeitschrift fur
Papyrologie und Epigraphik 67,1987, 51-55
Myles Burnyeat: The Origins of Non-deductive Inference', in J. Barnes, M. F.
Burnyeat and M. Schofield (edd.), Science and Speculation (Cambridge,
1982)
Antonio Carlini: 'Su alcuni papiri "platonici"', Proceedings of the XlVth
International Congress ofPapyrologists (London, 1975)
(ed.): Papiri letterari greci (Pisa, 1978)
(ed.): Papiri letterari greci della Bayerische Staatsbibliothek di Monaco
di Baviera (Stuttgart, 1986)
John Corcoran: 'Aristotle's Natural Deduction System', in J. Corcoran (ed.),
Ancient Logic and its Modern Interpretations (Dordrecht, 1974)
F. Edward Cranz: 'Alexander Aphrodisiensis', in P. O. Kristeller (ed.),
Catalogus Translationum et Commentariorum I (Washington D.C., I960)
Marina del Fabbro: 'II commentario nella tradizione papiracea', Stadia
Papyrologica 18,1979,69-132
H. Diels and W. Schubart (edd.): Anonymer Kommentar zu Platans Theaetet,
Berliner Klassikertexte 2 (Berlin, 1905)
John Dillon: lamblichi Chalcidiensis in Platonis dialogos commentariorum
fragmenta, Philosophia Antiqua 23 (Leiden, 1973)
Pierluigi Donini: Le scuole, I'anima, I'impero (Turin, 1982)
William E. Dooley: Alexander of Aphrodisias: on Aristotle, Metaphysics 1
(London, 1989)
Heinrich Dorrie: Der Platonismus in derAntike I (Stuttgart, 1987)
226 Bibliography
Sten Ebbesen [1981a]: Commentators and Commentaries on Aristotle's
Sophistici Elenchi, Corpus Latinum Commentariorum in Aristotelem
Graecorum VII (Leiden, 1981)
[1981b]: 'Analyzing Syllogisms, or Anonymous Aurelianensis HI - the
(presumably) earliest extant Latin commentary on the Prior Analytics,
and its Greek model', Cahiers de I'Institut du Moyen-Age Grec et Latin 37,
1981,1-20
'Ancient scholastic logic as the source of mediaeval scholastic logic', in
N. Kretzmann, A. J. P. Kenny and J. Pinborg (edd.), The Cambridge
History of Later Mediaeval Philosophy (Cambridge, 1982)
- 'Philoponus, "Alexander", and the origins of mediaeval logic', in Sorabji
(1990b)
Theo Ebert: 'Zur Formulierung pradikativer Aussagen in den logischen
Schriften des Aristoteles', Phronesis 22,1977,123-145
'Warum fehlt bei Aristoteles die 4. Figur?', Archiv fur Geschichte der
Philosophic 62,1980,13-31
L. Edelstein and I. G. Kidd: Posidonius: I The Fragments, Cambridge
Classical Texts and Commentaries 13 (Cambridge, 1972)
Kevin Flannery: 'A Rationale for Aristotle's Notion of Perfect Syllogisms',
Notre Dame Journal of Formal Logic 28,1987,455-471
P. M. Fraser: Ptolemaic Alexandria (Oxford, 1972)
Michael Frede [1974a]: Die stoische Logik (Gottingen, 1974)
[1974b]: 'Stoic vs. Aristotelian Syllogistic', Archiv fiir Geschichte der
Philosophie 56, 1974, 1-32 [reprinted in his Essays in Ancient Philosophy
(Oxford, 1987)]
J. Geffcken: 'Zur Entstehung und zum Wesen des griechischen wissenschaft-
lichen Kommentars', Hermes 67,1932, 397-412
Gabriele Giannantoni: Socraticorum Reliquiae (Naples, 1983)
John Glucker: Antiochus. and the Late Academy, Hypomnemata 56 (Got-
tingen, 1978)
review of Tarrant (1985), Journal of Hellenic Studies 109,1989,272-273
H. B. Gottschalk: 'Aristotelian Philosophy in the Roman World from the time
of Cicero to the end of the second century AD', in W. Haase (ed.), Aufstieg
und Niedergang der Romischen Welt II 36.2 (Berlin, 1987)
'The earliest Aristotelian Commentators', in Sorabji (1990b)
Richard Goulet: Dictionnaire des Philosophes Antiques I (Paris, 1989)
Andreas Graeser: Die logischen Fragmente des Theophrast (Berlin, 1973)
William M. A. Grimaldi: Aristotle, Rhetoric I: a commentary (New York,
1980)
A. Gudeman: 'Scholien', Pauly-Wissowas Realencyclopaedie IIA (Stuttgart,
1923), 625-704
Ilsetraut Hadot: 'Les introductions aux commentaires ex6getiques chez les
auteurs n^oplatoniciens et les auteurs Chretiens', in M. Tardieu (ed.), Les
regies de ^interpretation (Paris, 1987)
Pierre Hadot: 'Les divisions des parties de la philosophic dans 1'antiquite',
Museum Helveticum 36,1979,201-223
B. J. Hijmans: 'Apuleius, Philosophus Platonicus', in W. Haase (ed.), Aufstieg
und Niedergang der Romischen Welt II 36.1 (Berlin, 1987)
Jaakko Hintikka: Time and Necessity (Oxford, 1973)
'Aristotelian Induction', Revue Internationale de la Philosophie
133/134,1980,422-439
and U. Remes: The Method of Analysis (Dordrecht, 1974)
 Bibliography 227
Karlheinz Hiilser: Die Fragmente zur Dialektik der Stoiker (Stuttgart,
1987-88)
Katerina lerodiakonou: Analysis in Stoic Logic (diss. London, 1990)
Werner Jaeger: review of Merki, in his Scripta Minora II (Rome, 1960)
E. Kapp: 'Syllogistic', in J. Barnes, M. Schofield and R. Sorabji (edd.), Articles
on Aristotle I (London, 1975)
I. G. Kidd: Posidonius: II The Commentary, Cambridge Classical Texts and
Commentaries 14 (Cambridge, 1988)
John S. Kieffer: Galen's Institutio Logica (Baltimore, 1964)
William and Martha Kneale: The Development of Logic (Oxford, 19848)
Erich Lamberz: 'Proklos und die Form des philosophischen Kommentars', in
J. Pepin and H. D. Saffrey (edd.), Proclus, lecteur et interprete des anciens
(Paris, 1987)
Jonathan Lear: Aristotle and Logical Theory (Cambridge, 1980)
Tae-Soo Lee: Die griechische Tradition der aristotelischen Syllogistik in der
Spatantike, Hypomnemata 79 (Gb'ttingen, 1984)
A. C. Lloyd: Form and Universal in Aristotle (Liverpool, 1981)
The Anatomy ofNeoplatonism (Oxford, 1990)
A. A. Long: 'Reply to Jonathan Barnes: "Epicurean Signs"', Oxford Studies in
Ancient Philosophy, supplementary volume 1988
Jan Lukasiewicz: Aristotle's Syllogistic (Oxford, 19572)
John Lynch: Aristotle's School (Berkeley, 1972)
Henry Maconi: 'Late Greek Syllogistic', Phronesis 30,1985,92-98
Heinrich Maier: Die Syllogistik des Aristoteles (Tubingen, 1899/1900)
Italo Mariotti: Aristone d'Alessandria (Bologna, 1966)
Miklos Maroth: Ibn Sina und die Peripatetische 'Aussagenlogik' (Budapest,
1989)
Benson Mates: Stoic Logic (Berkeley, 19612)
H. P. F. Mercken: 'The Greek Commentators on Aristotle's Ethics', in Sorabji
(1990b)
H. Merki: Homoiosis Theoi (Fribourg, 1952)
Mario Mignucci: Aristotele: gli Analitici Primi (Naples, 1969)
'Hos epi to polu et necessaire dans la conception aristotelicienne de la
science', in E. Berti (ed.), Aristotle on Science: the Posterior Analytics
(Padua, 1981)
'La teoria della quantificazione del predicate nell'antichita classica',
Anuario Filosofico 16,1983,11-42
Paul Moraux: Der Aristotelismus bei den Griechen, Peripatoi 5 and 6 (Berlin,
1973/1984)
Le Commentaire d'Alexandre d'Aphrodisie aux 'Seconds Analytiques'
d'Aristote, Peripatoi 13 (Berlin, 1979)
[1986a]: 'Diogene Laerce et le Peripatos', Elenchos 7,1986,245-294
[1986b]: 'Les debuts de la philologie aristotelicienne', in G. Cambiano
(ed.), Storiografia e dossografia nella filosofia antica (Turin, 1986)
Ian Mueller: 'Stoic and Peripatetic Logic', Archiv fur Geschichte der Philos-
ophie 51,1969,173-187
'Greek Mathematics and Greek Logic', in J. Corcoran (ed.), Ancient
Logic and its Modern Interpretations (Dordrecht, 1974)
Iwan von Mueller: Uber Galens Werk vom wissenschaftlichen Beweis,
Abhandlungen der koniglichen Bayerischen Akademie der Wissenschaf-
ten 20 (Munich, 1897)
228 Bibliography
Vivian Nutton: 'Galen in the Eyes of his Contemporaries', Bulletin of the
History of Medicine 58,1984, 315-324
'Galen's Philosophical Testament: "On My Own Opinions"', in J.
Wiesner (ed.) Aristoteles: Werk and Wirkung II (Berlin, 1987)
J. H. Oliver: Marcus Aurelius - Aspects of Civic and Cultural Policy in the
East, Hesperia supplement 13 (Princeton, 1970)
Dominic O'Meara: Pythagoras Revived (Oxford, 1989)
Giinther Patzig: Aristotle's Theory of the Syllogism (Dordrecht, 1968)
R. Pfeiffer: A History of Classical Scholarship (Oxford, 1968)
Shlomo Pines: 'Omne quod movetur necesse est ab aliquo moveri: A
refutation of Galen by Alexander of Aphrodisias and the Theory of
Motion', Isis 52,1960,21-54 [reprinted in his Studies in Arabic Versions of
Greek Texts and in Mediaeval Science (Jerusalem, 1986)]
Karl Praechter [1909a]: review of CAG, Byzantinische Zeitschrift 18, 1909,
516-538 [reprinted in his Kleine Schriften (Hildesheim, 1973); English
translation in Sorabji (1990b)]
[1909b]: review of Diels-Schubart, Gdttinger Gelehrter Anzeiger 171,
1909, 530-547 [reprinted in his Kleine Schriften (Hildesheim, 1973)]
Carl Prantl: Geschichte derLogik imAbendlande (Leipzig, 1855)
E. Puglia: 'La filologia degli Epicurei', Cronache Ercolanesi 12,1982,19-34
Nicholas Rescher: Galen and the Syllogism (Pittsburgh, 1966)
Joyce Reynolds: Aphrodisias and Rome, Journal of Roman Studies mono-
graph 1 (London, 1982)
Marcel Richard: 'Apo Phones', Byzantion 20,1950,191-222
Lynn E. Rose: Aristotle's Syllogistic (Springfield, 1968)
W. D. Ross: Aristotle's Prior and Posterior Analytics (Oxford, 1949)
D. M. Schenkeveld: 'Stoic and Peripatetic Kinds of Speech Act and the
Distinction of Grammatical Moods', Mnemosyne 37,1984,291-353
D. N. Sedley: 'Philosophical Allegiance in the Greco-Roman World', in
M. Griffin and J. Barnes (edd.), Philosophia Togata (Oxford, 1989)
R. W. Sharpies: 'Alexander of Aphrodisias: Problems about Possibility I',
Bulletin of the Institute of Classical Studies 29,1982,91-108
'Alexander of Aphrodisias: Problems about Possibility II', Bulletin of the
Institute of Classical Studies 30,1983, 99-110
'Alexander of Aphrodisias: Scholasticism and Innovation', in W. Haase
(ed.),AufstiegundNiedergangderRomischen Welt II 36.2 (Berlin, 1987)
-The School of Alexander?', in Sorabji (1990b)
Ineke Sluiter: Ancient Grammar in Context (Amsterdam, 1990)
Timothy Smiley: 'What is a Syllogism?', Journal of Philosophical Logic 2,
1973,136-154
Robin Smith: 'What is Aristotelian Ecthesis?', History and Philosophy of
Logic 3,1982,113-127
Aristotle: Prior Analytics (Indianapolis, 1989)
Richard Sorabji [1990a]: The ancient commentators on Aristotle', in Sorabji
(1990b)
:(ed.) [1990b]: Aristotle Transformed (London, 1990)
Jiirgen Sprute: Die Enthymemtheorie der aristotelischen Rhetorik (Got-
tingen, 1982)
Gisela Striker: 'Aristoteles iiber Syllogismen "aufgrund einer Hypothese" ',
Hermes 107,1979,33-50
G. Strohmaier: 'Al-Iskander al-Afrudisi', in G. van Donzel (ed.), Encyclo-
paedia of Islam W (Leiden, 19782)
 Bibliography 229
Eleonore Stump: Dialectic and its Place in the Development of Mediaeval
Logic (Ithaca, 1989)
M. W. Sullivan: Apuleian Logic (Amsterdam, 1967)
Leonardo Taran: 'Amicus Plato sed magis arnica veritas', Antike und
Abendland 30,1984,93-124
Harold Tarrant: 'The Date of anon. In Theaetetum', Classical Quarterly 33,
1983,161-187
Scepticism or Platonism ? (Cambridge, 1985)
Pierre Thillet: Alexandre d'Aphrodise: Traite du Destin (Paris, 1984)
Paul Thorn: The Syllogism (Munich, 1981)
E. Thouverez: 'La IVme figure du syllogisme', Archiv fur Geschichte der
Philosophic 15,1902, 49-110
Robert B. Todd: 'Lexicographical Notes on Alexander of Aphrodisias'
Philosophical Terminology', Glotta 52,1974, 207-215
Alexander of Aphrodisias on Stoic Physics, Philosophia Antiqua 28
(Leiden, 1976)
E. G. Turner: Greek Papyri (Oxford, 1968)
Martin M. Tweedale: 'Alexander of Aphrodisias' Views on Universals',
Phronesis 29,1984,279-303
Mario Untersteiner: Problemi della filologia filosofica (Milan, 1980)
Hermann Usener: review of CAG, in his Kleine Schriften III (Stuttgart,
1913)
G. Volait: Die Stellung des Alexander von Aphrodisias zur aristotelischen
Schlufilehre (Bonn, 1907)
Hans von Arnim: Stoicorum Veterum Fragmenta (Leipzig, 1903-1905)
Heinrich von Staden: Herophilus: the art of medicine in early Alexandria
(Cambridge, 1989)
Maximilian Wallies: 'Zur Textgeschichte der Ersten Analytik', Rheinisches
Museum 72,1917/1918,626-632
Fritz Wehrli: Eudemos von Rhodos, Die Schule des Aristoteles 8 (Basel,
1955)
L. G. Westerink: 'Elias on the Prior Analytics', Mnemosyne 14, 1961,
126-139 [reprinted in his Texts and Studies in Neoplatonism and
Byzantine Literature (Amsterdam, 1980)]
The Greek Commentaries on Plato's Phaedo (Amsterdam, 1976)
'The Alexandrian commentators and the introductions to their
commentaries', in Sorabji (1990b)
Wolfgang Wieland: 'Die aristotelische Theorie der Konversion von
Modalaussagen', Phronesis 25,1980,109-116
Mark F. Williams: Studies in the Manuscript Tradition of Aristotle's
Analytics, Beitrage zur klassischen Philologie 161 (Konigstein, 1984)
Eduard Zeller: Die Philosophie der Griechen III i (Leipzig, 19235)
F. W. Zimmermann: Al-Farabi's Commentary and Short Treatise on
Aristotle's de Interpretatione, Classical and Mediaeval Logic Texts 3
(London, 1981)
 Indexes

English-Greek Glossary
The Glossary lists all and only those English words which occur in the Greek- English
index.
absurd: atopos clear: delos
accidentally: kata sumbebekos clear, make (v.): deloun [deldtikos]
account: logos co-assume (\.)\proslambanein
accusative: aitiatikos [aprosleptos]
actual: huparkhousa co-assumption: proslpsis
add (v.): prostithenai coincide with (v.): sumpiptein
addition: prosthesis, prosthSke combination: suzugia
adjunct: prosrSsis combine (v.): suzeugnunai
affirm (v.): kataphaskein common: koinos
affirmation: kataphasis common usage: sunetheia
affirmative: kataphatikos common use, in: suneth&s
agree (v.): homologein complementarily: enallax, parallax
agreement: hotnologema compound (v.): suntithenai
aim: skopos compounded, be (v.): sunkeisthai
alter together with (v.): summetaballein compounding: sunthesis
analyse (v.): analuein concede (v.): sunkhdrein
analysis: analusis conclude (v.):perainein, sumperainein
analytic: analutikos conclusion: sumperasma
annex (v.): prostithenai conditional: sunGmmenon
apply to (v.): harmozein, epharmozein conflict: makM
appropriate: oikeios conflict (v.): makhesthai
argument: logos congruous: katalUlos
art: tekhne conjectural: stokhastikos
aspect: skhesis conjunction: sumploke
assert (v.): apophainein connect (v.): suntattein
assertion: apophansis connection: suntaxis
assertoric: apophantikos construct (v.): sunistanai
assume (v.): lambanein [lepsis] construction: sustasis
assumption: Umma contingent, be (v.): endekhesthai
authoritative: kurios continue (v.): epipherein
axiom: axioma continuous: sunekMs
contradiction: antiphasis
beginning: arkhS contradictory: antikeimenon, anti-
betoken (v.): d&loun phatikos
contradictory pair: antiphasis
cancel (v.): anairein [anairetikos] contrary: enantios
case: ptosis contribute (v.): suntelein
change (v.): metapiptein [summeta- converse: antistrophe
piptein, summetaskhematizein] conversion: antistropM

230
 English-Greek Glossary 231
convert from (v.): antistrephein + pros external: exothen
convert with (v.): antistrephein + dative extreme: akron
co-ordinate (v.): antidiairein
co-predicate (v.Y.proskategorein fal\(v.):piptein
counterpredicate (v.): antikategorein fall in (v.): peripiptein
false -.pseudes
deduce (v.): sunagein figure: skhema
deduction: sunagdge find (v.): heuriskein
deductive: sunaktikos follow from (v.): akolouthein, hepesthai
define (v.): horizein form: eidos [eidopoiein]
definition: horismos
definitional: horistikos general: koinos
demand: attests genus :genos
demand (v.): apaitein give (an account oO (v.): apodidonai
demonstrate (v.): apodeiknunai [.apodosis]
demonstration: apodeixis greater length, at: epi pleon
demonstrative: apodeiktikos
deny (v.): apophanai, apophaskein help: boetheia
destroy (v.): anairein hold (of) (v.): huparkhein [sunuparkhein,
determination: diorismos asunuparktos]
determine (v.): diorizein homonymous: homonumos
diagram: katagraphe homonymy: homonumia
dialectical: dialektikos hypothesis: hupothesis
difference: diaphora hypothesize (v.): hupotithenai
differentia: diaphora hypothesized, be (v.): hupokeisthai
directly: autothen hypothetical: hupothetikos
discovery: heuresis
disjoin (v.): apozeugnunai, diazeugnunai immediately: eutheos, euthus
disjoint: diairetikos imperfect: ateles
disjunctive: diazeuktikos implication: akolouthia
disposition: hexis impossible: adunatos
disprove (v.): diaballein include (v.)-.periekhein
dispute (v.): amphisbetein incongruous: akatallelos
dissimilar in form: anomoioskhemon indemonstrable: anapodeiktos
distinguish (v.): diairein, khorizein indeterminate: adioristos
divide (v.): diairein indicate (v.): endeiknunai
doctrine: doxa indifferent: adiaphoros
duplicated: diphoroumenos individual: atomos
indivisible: adiairetos, atomos
encompass (v.): perilambanein [perilep- induction: epagoge
sis, sumperilambanein] infer (v.): epipherein
end: telos inferior: kheiron
equal: isos inseparable: akhoristos
equivalent: isos instrument: organon
equivalent, be (v.): ison dunasthai, intelligible: noetos
isodunamein introduce (v.): paralambanein [sum-
establish (v.): kataskeuazein paralambanein]
example: paradeigma investigate (v.): zetein
exceed (v.): huperballein, huperek-
piptein, huperekhein, huperpiptein justification: pistis
excel (v.): pleonektein justified: pistos
existence: huparxis justify (v.):pisteuein,pistousthai
explanation: aitia
explanatory: aitios keep (v.): terein
exposition: ekthesis [ekthetikos] knowledge: gnosis, episteme
expression: lexis known: gndrimos
232 Indexes
laid down, be (v.): keisthai position: thesis
last: eskhatos possible: dunatos
link (v.): sunaptein possible, be (v.): dunasthai, endekhesthai
logic: logike predicate: kategoroumenon
logical: dialektikos predicate (v.): kategorein
predication: kategoria
major: meizon predicative: kategorikos
mark (v.): menuein [menutikos] prefix (v.): protattein
material: hulikos premiss: protasis
material instance: hule present (v.): apodidonai [apodosis]
matter: hule preserve (v.): phulattein
mean (v.): deloun, menuein, semainein primary: proegoumenos
[deldtikos, semantikos} principle: arkhe
metaphor: metaphora privative: steretikos
method: methodos problem: problema
middle: mesos proof: deixis, elenkhos
minor: elation proper characteristic: to idion
mode: tropos propose (v.): protithenai
proposition: protasis
name: onoma propound (v.): proteinein
necessary: anankaios prove (v.): deiknunai [deiktikos],
necessity: ananke elenkhein
need(v.):endem purpose: prothesis
negation: apophasis purpose (v.): protithenai
negative: apophatikos put(v.): tithenai
non-different: adiaphoros
non-syllogistic: asuilogistos quality: poion,poiotes
quantity: poson
object (v.): enistastkai question, be in (v.): prokeisthai
obtain (v.): euporein
obvious: enarges rank: taxis
opposed to, be (v.): antikeisthai rank(v.): tattein
opposite: antikeimenon reason: aitia
opposition: antithesis re-assume (v.): metalambanein
order: taxis reduce (v.): anagein, apagein
order (v.): tattein reduction: anagoge, apagoge
outside, from: exothen redundant, be (v.): parelkein
own: oikeios reference: anaphora
refutation: elenkhos
paradigm: paradeigma [para- refute (v.): elenkhein
deigmatikos] reject (v.): anairein [anairesis, sun-
part: meros, morion anairein]
participate (v.): metekhein relation: skhesis
particular: enmerei, epi merous, kata reliable: dokimos
meros remain (v.): menein
perception: aisthgsis [aisthetos, reputable: endoxos
aisthetikos] request: erotesis
perfect: teleios request (v.): erdtan
perfect (v.): epitelein, teleioun reveal (v.): ekkaluptein
perfection: teleiosis right, in its own: kath' hauto
permutation: metathesis
place (v.): tithenai same genus, of the: homogenes
plausible: pithanos say (v.): legein
point at issue: prokeimenon science: episteme
posit: thesis seek (v.): zetein [zgtesis]
posit (v.): tithenai separate (v.): khdrizein
 English-Greek Glossary 233
separated: khoristos superior: beltion
separative: khoristikos suppose (v.): hupotithenai
set down (v.Y.paratithenai [parathesis] supposed, be (v.): keisthai, hupokeisthai
set out (v.): ektithenai supposition: keimenon
set out, be (v.): ekkeisthai syllogism: sullogismos
share (v.): koinonein [koinonia] syllogistic: sullogistikos
shared: koinos syllogize (v.): sullogizein
show (v.): deiknunai [deiktikos]
sign: sGmeion take (v.): lambanein
signify (v.): semainein [semantikos] term: horos
similar in form: homoioskhemon theorem: theorema
single-assumption: monolemmatos theoretical: theor&ikos
sophist: sophistes theorise (v.): theorem
sophistical: sophistikos theory: thedria
sound: hugies transform (v.): metalambanein
species: eidos transformation: metalepsis
starting-point: arkM 'tropic': tropikon
statement: axioma true: alethes [sunaUtheuein]
strict: kurios
study: pragmateia unclear: ad&los
study (v.):pragmateuesthai universal: katholikos, katholou
subaltern: hupallelos unmethodical: amethodos
subcontrary: hupenantios unreliable: adokimos
subject for, be (v.): hupokeisthai utterance: logos
subject for, make (v.): hupotithenai
subordinate (v.): hupotattein verb: rhema
subpart: morion
subsistence: hupostasis wider extension, of: epipleon
substance: ousia whole: holokleros
subsume (v.): hupagein whole, in as in a: en holoi
s\Aayl\ogistical:huposullogistikos word: onoma,phdnS
 Greek-English Index
Wallies' edition of in An. Pr. contains an adequate index verborum, and we have not
thought it necessary to duplicate his work: this Index limits itself, with a few
exceptions, to Alexander's logical terminology (but it understands the word 'logical' in
a generous sense). We have not listed every occurrence of every logical term in our
portion of in An. Pr. Where a word occurs more than about twenty-five times, we have
given the first six or seven occurrences and then indicated (by a figure printed in bold
type and between square brackets) approximately how many examples of the word a
complete list would contain. (The listings exclude occurrences in lemmata or citations
from Aristotle's text.) The Ibycus system, which we used in making this Index, is
widely available; and any reader who wishes to scan all the occurrences in our text of,
say, the word 'sullogismos' will find it far easier to consult Ibycus than to contemplate
a sequence of 290 numerical references.
For each Greek word we have given the English translation or translations which
we have normally used. (Note that the translations are normal and not invariable.) In
a few cases we have indicated that a word is normally paraphrased rather than
translated: we set between angle brackets the word or words normally used in the
paraphrase.
adelos, unclear, 5,29.31; 18,24 akolouthein, to follow, 17,31; 20,32;
adiairetos, indivisible, 81,21.22 21,7; 35,16; 55,10.10; 87,13;
adiaphoros, non-different, indifferent, 112,12.30
18,17; 72,18 akolouthia, implication, 7,1; 11,20;
adioristos, indeterminate, 11,19; 13,30; 17,30
27,18; 30,29; 49,14.15; 51,25 [70] akron, extreme, 33,27; 46,19.21;
adokimos, unreliable, 52,24.24; 56,25; 47,3.4.11.15 [50]
62,5; 64,21; 65,1; 85,22; 104,19.20 alettes, true, 5,13; 6,9.10;
adunatos, impossible, 22,11; 24,6.12.16; 8,24.25.26.27.29 [130]
31,10.11; 33,11 [125] amethodos, unmethodical, 21,30;
aisthesis, perception, 4,4; 5,7; 33,1; 22,18.24; 24,1.9; 68,23.32
99,32; 100,1 amphisbetein, to dispute, 43,8.26;
aisthetikos, perceptual, 32,33; 33,14; 44,14.15
113,1 anagein, to reduce, 7,18.25; 24,5; 58,21;
aisthetos, perceptible, 4,3.6.9; 32,33; 77,4; 98,26; 100,31 [65]
33,2.3; 53,23; 100,4.5.7.12.17.22.23; anagoge, reduction, 7,12.26.31; 97,29;
101,5; 104,4 113,16; 115,14.17.36; 118,20.25.28
aitesis, demand, 14,2; 43,21 anairein, to destroy, to reject, to cancel,
aitia, reason, explanation, 5,8; 17,24; 22,27.27.28.28; 24,16; 31,13.14;
21,13.13.15.16 [30] 50,31; 55,31; 56,2.5.30; 57,2; 64,3.9;
aitiatikos, accusative, 9,7.8 87,25; 88,25; 90,10.26.28; 91,9; 105,6;
aitios, explanatory, 8,16; 10,29; 14,13; 107,21; 112,29
21,17.21; 23,2; 24,2.16; 42,23; 47,22; anairesis, rejection, 30,24; 36,5; 77,24;
55,18; 67,17; 68,16; 75,18; 83,14; 84,26
87,20; 94,6.22; 95,4; 96,9; 101,19; anairetikos, <cancel>, 55,26; 90,8.12;
104,32; 106,30; 107,16 91,25.32
akatallelos, incongruous, 15,24 analuein, to analyse,
akhdristos, inseparable, 4,13; 26,19.20 7,15.20.22.22.23.25

234
 Greek-English Index 235
analusis, analysis, 6,28; 53,2.5; 54,6; 69,24
7,12.13.14.19.26.27; 58,19; 97,29.30 apodosis, <apodidonai>, 9,8; 10,7;
analutikos, analytic, 4,30; 14,28
6,15.30.32.33; 7,11.27; 8,1.2; 9,6; apophainein, to assert, 5,22
11,9; 12,3.23; 14,13.19; 25,12; 42,24; apophanai, to deny, 91,22
110,20 apophansis, assertion,
anankaios, necessary, 4,11.18; 8,33; 10,13.16.17.27.27; 11,7.8.28
9,25; 12,24; 17,27.28 [115] apophantikos, assertoric, 10,19.30;
ananke, necessity, 2,1; 7,3; 17,26.31.32; 11,12; 14,2; 27,8
20,32; 21,2 [50] apophasis, negation, 10,6.25.30;
anaphora, reference, 2,10.12.14.17.35; 11,5.13.15.23.23.27.27; 13.22.26;
3,16.24; 8,5 14,3; 15,11.13.27.28.32.34; 16,2.4;
anapodeiktos, indemonstrable, 6,25; 22,16.16; 37,7; 41,14; 72,22
24,4.11; 49,5; 54,12; 55,3; 60,21.25; apophaskein, to deny, 73,13.18;
69,28; 70,3; 77,5.16; 79,7; 83,13; 74,1.2.10.11.15.18.19.21.22
102.30; 110,7; 111,12.18; 113,7 apophatikos, negative, 10,18; 11,26.31;
anomoioskhemon, dissimilar in form, 13,24; 14,9; 15,10.30 [405]
70,8; 71,29; 76,21.23; 82,9.18; 86,12; apozeugnunai, to disjoin, 31,7.8.8.9;
96,5; 104,13; 106,3; 108,4; 109,9.17 34,14.14
antidiairein, to co-ordinate, 1,16.18; aprosleptos, <co-assume>, 20,18
2,12; 11,2 arkhe, principle, beginning, starting-
antikategorein, to counter-predicate, point, 4,12; 7,14.14.15.17; 8,12.14
72,23 [30]
antikeisthai, to be opposed to [anti- asullogistos, non-syllogistic, 6,22;
keimenon, opposite, contradictory], 20,14.19.23; 42,13; 47,18; 52,19 [120]
12,32.33; 15,13; 19,11.12.16.18 [95] asunuparktos, not holding at the same
antiphasis, contradictory pair, contra- time, 90,2
diction, 11,10.11; 12,28; 13,5; 14,5; ateles, imperfect, 6,25.25; 7,24; 10,1;
18,28.29.29; 19,3; 31,17; 36,5.6.12; 23,19.21.23.24 [30]
45,31; 56,1.5; 57,1; 80,1; 89,33.34; atomos, indivisible, individual,
90,15 65,24.25.28; 73,23.28; 100,14
antiphatikos, contradictory, 30,1; 31,23; atopos, absurd, 16,15; 20,27; 34,20
46,1; 99,14 autothen, directly, 6,24; 13,9; 54,18;
antistrephein, to convert [+pros, con- 116,36; 117,12.17.23; 118,19
vert from; + dative, convert with], axidma, axiom, statement, 1,19;
6,35; 28,9; 29,2.8.9.10.13 [160] 44,18.21
antistrophe, conversion, converse,
24,3.5.14; 27,23.26; 28,5.6 [140] beltion, superior, 3,6.7; 51,11.13.18;
antithesis, opposition, 29,10.16; 71,17
45,15.18.24; 46,8.9 boetheia, help, 24,10; 77,8.20.28; 78,22
apagein, to reduce, 6,4
apagoge, reduction, 24,6; 31,10.11.15; deiknunai, to prove, to show, 2,2;
34,5.17; 76,11 [55] 4,27.32.33; 5,27; 12,9.16 [345]
apaitein, to demand, 4,24; 9,7 deiktikos, <prove, show>, 34,13; 43,25;
apodeikiiunai, to demonstrate, 54,2; 75,18; 79,20; 104,23; 109,17;
12,25.32.34; 13,16; 49,9 111,18; 112,16
apodeiktikos, demonstrative, 1,4; deixis, proof, 4,22.26; 7,17; 18,21; 23,28;
9,19.21.21.22.23; 12,8.22.27; 28,14.16; 31,6 [80]
13,14.22; 14,11.19; 18,24; 20,26; delos, clear, 3,20; 5,2; 13,5; 15,12; 16,23;
28,2.27; 49,6.7.9 18,31; 28,19 [35]
apodeixis, demonstration, 2,25; deldtikos, <deloun>, 11,24; 13,30;
4,21.24.25.33; 6,10.12 [35] 15,9.11; 16,24; 18,8; 20,31; 21,4.13;
apodidonai, to give (an account of), to 26,16; 42,9.17; 59,32; 62,12; 71,6;
present, 10,14; 11,21; 76,13; 93,3
12,3.9.11.14.18; 13,13.18; 16,22; deloun, to make clear, to mean, to
23,18; 25,13.17.28; 41,5; 42,19; 44,13; betoken, 11,5; 12,16; 13,26.29; 14,23;
236 Indexes
15,28; 17,6 [40] 65,10; 68,25; 90,10; 111,26
diaballein, to disprove, 30,23; 64,3; enallax, <complementary>, 61,14;
66,27; 68,24; 88,25; 89,1.29.32; 109,14
91,3.21 enantios, contrary, 20,3; 25,15; 45,20.23;
diairein, to divide, to distinguish, 4,9; 55,25; 61,20; 81,19.19; 87,16; 90,2
7,21; 15,6.10.11.28.31; 16,2.4.5.6; enarges, obvious, 13,9; 23,12; 30,28;
36,21; 45,4; 68,31 31,13.14; 34,11.15; 54,19; 55,24;
diairetikos, disjoint, 19,4.7.12.20.29.32; 100,6; 101,5
20,3 endeiknunai, to indicate, 9,30; 12,25;
dialektikos, logical, dialectical, 1,4; 3,30; 14,25; 16,7; 17,2; 52,27; 53,28; 54,8
4,28; 7,8; 8,24; 11,21; 12,4 [30] endein, to need, 22,5; 24,9; 117,25
diaphora, difference, differentia, 4,15; endekhesthai, to be possible, to be con-
6,16.19.22; 8,6.27; 12,7 [70] tingent, 21,1.2; 25,7.8; 26,12.13.14
diazeugnunai, to disjoin, 19,22 [110]
diazeuktikos, disjunctive, 19,21.25.26 endows, reputable, 12,14.31.32; 14,5.17;
diorismos, determination, 36,26; 44,26; 18,27
65,26; 67,4; 100,13.13 en holdi, in as in a whole, 10,7; 24,22.23;
diorizein, to determine, 38,10; 51,24; 25,2.4.5.5 [30]
66,11; 67,4.6; 81,19; 93,9 enistasthai, to object, 12,29; 65,23.24
diphoroumenos, duplicated, 18,17; en merei, particular, 11,19; 13,30; 45,6;
20,11 58,27.28; 59,2.4; 93,11; 116,2;
dokimos, reliable, 52,19; 55,15; 59,14 117,1.24
doxa, doctrine, 81,27; 86,5 epagoge, induction, 20,28; 43,27.29; 44,1
dunasthai, to be possible, 4,9.11.14.17; epharmozein, to apply, 13,14.21; 34,6;
7,29.31; 8,26 [105] 111,31
dunatos, possible, 25,15; 38,7.13.19.29; epi merous, particular, 27,18;
39,17.21.24.28 30,2.3.7.8.15.16 [410]
epipherein, to infer, to continue, 13,5;
eidopoiein, to impose a form, 56,25 15,2; 18,14.24; 19,1.3.5.11.25.27.28;
eidos, form, species, 4,13; 6,17.19; 20,23.28; 33,22; 56,29; 58,12; 59,29;
7,2.3.4.4.5.20; 8,22; 11,1; 18,22.23; 60,3; 69,24; 78,25; 82,7; 102,22; 105,8
19,1.2; 22,17; 27,21; 28,26; 51,21; epipleon, at greater length, of wider
73,23.28 extension, 12,1.21; 17,29; 25,7; 41,24;
ekkaluptein, to reveal, 18,25 44,2; 47,31; 48,1; 53,23; 56,12; 61,28;
ekkeisthai, to be set out, 55,18; 56,8; 62,19; 64,19; 72,25; 74,14; 95,14
63,17; 66,1; 69,26; 75,2; 78,18; 80,10; episteme, science, knowledge, 1,6;
83,14; 85,23; 90,24; 99,21; 100,10; 2,4.5.9.11.19.23.26 [40]
101,7 epitelein, to perfect, 69,17; 115,3;
ekthesis, exposition, 32,33; 116,27; 117,20.23
33,1.13.26.30; 34,1; 54,4; 65,17.30; erotan, to request, 12,25.28.30.34;
67,11; 78,1; 89,13.15; 96,17; 99,22.32; 13,6.21; 14,1; 49,14
100,7.14.16; 103,4.6.27; 104,8; erdtesis, request, 11,22; 12,12.14.27.27;
106,20; 113,1 13,3.5.6.7; 14,4
ekthetikos, <exposition>, 34,7 eskhatos, last, 48,19; 49,20.21; 51,21;
ektithenai, to set out, 12,1; 32,30; 33,2; 65,22.28; 67,13; 77,27; 78,5;
34,21; 54,31; 55,15; 59,14; 61,18; 94,6.7.10.17; 97,8; 98,12; 113,28
98,23; 101,4.11; 102,13; 104,28; euporein, to obtain, 6,27; 12,31; 65,8;
106,11; 109,4 66,17; 75,25; 87,8.19.25;
elation, minor, 3,9; 4,31; 22,3; 39,22.26; 88,4.10.17.24.28; 89,7.30; 90,14.19;
47,29; 48,1 [230] 91,14.30; 105,5.21; 106,1.1; 107,24,27
elenkhein, to refute, to prove, 20,20; euthus (andeutheds), immediately, 5,8;
31,15; 65,3; 66,1.27; 67,23; 68,27; 8,12; 19,32; 115,2.4; 116,27; 117,19
81,6; 82,1.11; 85,23; 87,2; 89,13; exdthen, from outside, external, 16,8.9;
91,22; 92,10.30.33; 101,14; 104,19.20; 21,26; 22,2; 23,27; 24,2.7.13.13;
105,22; 107,10; 108,6; 109,26 45,2.7; 46,25; 54,8.21; 76,6; 77,21;
elenkhos, refutation, proof, 14,20; 37,24; 80,30; 84,29.22
 Greek-English Index 237
genos, genus, 4,15.17; 7,2.2.3.4; supposed, to be hypothesized,
11,1.3 [35] 1,15.18.19; 4,17; 9,16.17.17 [80]
gnorimos, (well) known, 4,25; 6,24; hupostasis, subsistence, 4,10.13
10,2.6; 12,24; 13,9; 14,13 [30] huposullogistikos,subsy\logistica\,
gnosis, knowledge, 3,13.18.21.23.25; 4,1; 84,12
5,1.8.10.13.13; 6,10.10 hupotattein, to subordinate, 4,16; 10,31
hupothesis, hypothesis, 13,1.7.10; 24,17;
harmozein, to apply to, 13,27; 58,15; 42,30; 80,7.15.20.22; 83,6;
100,12 84,20.25.27.28; 93,28.29; 99,12.13;
hepesthai, to follow, 7,1.3; 13,10; 14,21; 105,13.15; 107,21; 112,17.30
17,28.29.31 [30] hupothetikos, hypothetical, 11,19; 17,8
heuresis, discovery, 1,7.12; 2,15.27.34; hupotithenai, to make subject for, to
3,11; 6,26; 8,24.29 suppose, to hypothesize, 13,11;
heuriskein, to find, 5,15; 6,27; 7,31; 8,26; 31,12.14.24; 34,9; 38,20.22 [30]
31,16; 45,18; 51,24.26; 61,29; 62,20;
64,10; 75,2.5; 81,25; 90,15; 91,25 idion, to, proper characteristic, 18,1;
hexis, disposition, 5,27; 9,18.21; 61,30; 21,15; 47,33; 48,4.5.13; 49,1; 55,22;
62,1.2.4.8; 89,17.17.18.20.21.23.23.24 63,29; 71,25; 95,10.13.19.20.23
holokleros, whole, 48,24 isodunamein, to be equivalent to, 41,12;
homogenes, of the same genus, 16,29.30; 84,13.15
72,28; 73,3.19 ison dunasthai, to be equivalent to,
homoioskhemon, similar in form, 15,18.20; 30,30; 37,8; 41,14; 44,27;
70,6.12; 71,31; 81,8; 86,19.22.23.26; 49,15; 84,7; 92,33; 94,20; 111,30;
92,3.5.8.14; 95,10; 103,9; 109,7.12 112,1
homologein, to agree, 1,10; 5,24; 18,14; isos, equal, equivalent, 11,6.9; 13,3;
31,13.14; 32,3.5.8.28.29; 33,12; 36,12; 22,4.5.6.7 [30]
44,15; 79,25; 90,14; 113,12
homologema, agreement, 44,17.20 katagraphe, diagram, 72,11; 78,4
homonumia, homonymy, 37,28 katallelos, congruous, 16,13; 67,1
homonumos, homonymous, 38,10 kata meros, particular, 49,31; 61,21;
horismos, definition, 11,21; 16,31 86,21; 109,23
horistikos, definitional, 23,6.7 kataphasis, affirmation, 10,5.23.25;
horizein, to define, 10,15.28; 11,5.13.15.26 [35]
11,15.26.30; 12,13; 14,30 [30] kataphaskein, to affirm, 73,14.19;
fcoros.term, 3,28; 6,6; 9,27; 11,9.17.23: 74,1.20.22
14,25 [205] kataphatikos, affirmative, 10,18;
hugies, sound, 17,31; 65,32; 90,6 11,24.25.31; 13,24; 14,9; 19,8 [365]
hule, material instance, matter, 4,13; kataskeuazein, to establish, 93,29;
6,17.18.20; 7,20; 12,23; 13,13 [40] 112,30
hulikos, material, 27,29 kata sumbebSkos, accidentally, 112,19
hupagein, to subsume, 44,5 kat&gorein, to predicate [kategorou-
hupallelos, subaltern, 45,26; 87.131 menon, predicate], 10,24.32;
huparkhein, to hold (of) [huparkhousa, 11,1.4.5.24.25 [120]
actual], 5,23; 20,1.1; 21,1; 22,12; katdgoria, predication, 6,35; 26,30
26,1.2 [440] kategorikos, predicative, 11,18.29; 15,3;
huparxis, existence, 4,11; 26,16.17; 38,5 17,6; 24,30; 25,26.28; 27,17.21; 42,28;
hupenantios, subcontrary, 45,24; 59,25
87,11.15; 88,32; 89,6; 91,29.31.32; kath'hauto, in its own right, 4,20; 15,17;
92,13; 105,24 17,9; 26,19; 66,15; 67,9.10.11.14;
huperballein, to exceed, 62,19 73,13.19; 74,30; 105,3; 107,28
huperekhein, to exceed, 64,20; 67,20 katholikos, universal, 22,10; 28,14.17;
huperekpiptein, to exceed, 56,21 30,28; 35,9; 74,3
huperpiptein, to exceed, 56,15.15 katholou, universal, 6,28; 10,5.5; 11,18;
hupokeisthai, to be subject for, to be 12,3.15; 13,30 [510]
keisthai, to be supposed, to be laid down
1
See also 63,13. [keimenon, supposition], 12,17; 16,20;
238 Indexes
17,28; 18,9.10.14.23 [135] metapiptein, to change, 51,12.15; 52,11
kheiron, inferior, 43,9; 51,7.31; metaphora, metaphor, 23,6.7.9.10.12
94,12.15.16; 99,3 metathesis, permutation, 93,11
khoristikos, separative, 16,31 metekhein, to participate, 5,16; 101,6
khoristos, separated, 26,19.20 methodos, method, 1,5.17; 3,14; 7,28;
khdrizein, to separate, to distinguish, 8,21; 28,3.18; 75,2.6
4,16; 15,10; 16,10.25.27.30; 18,9; monolemmatos, single-assumption,
26,21; 31,7; 45,1; 56,10; 67,14.19; 17,12.18; 18,4; 21,25; 22,24
100,18 morion, subpart, part, 1,14; 2,5;
koinonein, to share, 13,20; 16,28; 42,5; 3,7.13.32; 11,8; 15,22.30; 31,17.25;
45,11.12.12.13.14.16; 46,2.4.17.18; 33,2.3.4.5.6.7.21.22; 80,1
53,7; 68,18; 81,33; 82,1; 100,7; 101,6
koinonia, <to share>,45,9; 46,5.20.29; noetos, intelligible, 4,5.6.15
47,3.4; 56,22; 85,28; 94,7; 100,22
koinos, common, general, shared, 6,17;
10,32; 11,12.32; 12,7.9.16 [35] oikeios, appropriate, own, 2,24.25.29;
kurios, strict, authoritative, 1,7; 3,31; 3,26.27; 4,26; 5,1 [30]
22,25; 23,4.10; 36,30; 39,11; 42,29; onoma, name, word, 3,20; 9,28; 14,25.29;
47,23; 48,12.15.16; 49,9; 54,12; 17,13; 38,34; 48,3
106,18; 117,17 organon, instrument, 1,9;
2,3.15.17.18.20.21.23.28.33;
lambanein, to take, to assume,1,6; 3,4.5.8.9.12; 4,30.31; 18,20
4,13.28; 5,11; 7,17; 12,17.27 [240] ousia, substance, 3,32; 4,6.10.17; 17,1;
legein, to say, 1,9; 2,3.21.23.33; 3,4.31 22,27.27 [40]
[435]
lemma, assumption, 44,17.20 paradeigma, example, paradigm, 39,2;
lepsis, <assume>, 12,25.26; 13,19; 43,18; 44,1; 52,26.30; 68,23.24.27;
14,16; 47,13; 68,31 69,1; 71,4.5; 81,25.26; 92,30; 94,3
lexis, expression, 13,26; 16,13.16; 18,2.6; paradeigmatikos, paradigmatic, 18,11
19,19; 20,4; 21,4; 32,23; 38,6; 39,8.9; paralambanein, to introduce, 11,30;
53,8; 62,12; 66,29; 84,7.10.11.16.18; 18,21; 27,28; 45,5.7; 46,25; 77,21
93,2 parallax, complementarity, 93,8; 108,14
logike, logic, 1,3; 2,33; 3,3.7.14 parathesis, <to set down>, 55,22; 57,9;
logos, account, utterance, argument, 61,19; 63,16; 65,2; 66,28; 81,5; 86,30;
4,4.8.9.16; 6,31; 7,10.20 [130] 91,23; 92,7; 101,15; 104,19; 105,29;
106,25
makhe, conflict, 11,20 paratithenai, to set down, 39,5.19;
makhesthai, to conflict, 17,16; 52,23 53,15; 55,12; 64,14.22; 65,30 [30]
meizdn, major, 47,28.29.30; 48,4.4.6.7 parelkein, to be redundant, 23,1
[395]2 perainein, to conclude, 18,17; 21,31;
menein, to remain, 50,19; 60,6; 61,13; 22,18.24; 24,1.9; 68,23.32
63,9; 75,29; 83,16; 99,30; 115,15 periekhein, to include, 23,26; 25,3;
menuein, to mean, to mark, 26,8.15; 43,14; 44,9.28; 56,21; 64,20.21; 67,19;
55,33 85,28.29
menutikos, marking, 21,5 perilambanein, to encompass, 49,31;
meros, part, 1,8.8.9.11.14; 2,3.4 [75]3 50,5; 56,14.16.16; 62,20; 67,21.22
mesos, middle, 33,28; 45,3.7.10; perilGpsis, <encompass>, 56,15
46,19.21.23 [105] peripiptein, to fall in, 75,26
metalambanein, to transform, to re- phone, word, 84,16.17
assume, <instead>, 19,4; 22,26.30; phulattein, to preserve, 5,4; 8,29; 23,9;
54,4; 60,1.6; 65,17; 67,1; 68,6; 41,7; 45,16; 46,11; 52,22; 71,26;
84,7.10; 110,26 82,18; 94,26; 95,6.19.21; 103,10
metalepsis, transformation, <instead>, piptein, to fall, 25,4; 57,13.14; 98,7;
65,23.31; 74,28; 84,19 100,1
2
Excluding the few occurrences in which meizdn means 'greater'.
3
Excluding occurrences which form part of the phrases en merei, epi merous and kata meros.
 Greek-English Index 239
pisteuein, to justify, 4,23; 43,8 9,12.20.30; 23,24; 41,33; 42,18; 44,19;
pistis, justification, 43,7.9.10.11.15.18; 99,11; 110,6
44,3.4.4.6.7.8; 48,13; 49,3; 68,20; pseudes, false, 8,28; 10,17.19; 11,20;
77,31; 84,24; 113,10 26,15; 27,5.6.9.10.11.11; 31,15.16.17;
pistos, justified, 4,25; 43,8.13.24.26.30; 41,16; 68,26; 80,7.7.11.15.19; 83,6;
84,27 84,2; 87,14; 99,12.13; 111,25
pistousthai, to justify, ptosis, case, 9,7.7.8.10
43,7.11.12.13.17.20.27; 54,19; 68,20;
77,25; 104,16 rhema, verb, 14,29
pithanos, plausible, 8,25.27; 14,17; 44,1
pleonektein, to excel, 48,28; 49,4.5; semainein, to mean, to signify, 3,21;
51,17 7,27; 10,3.5; 12,1; 15,8; 17,5.15.17;
poion, quality, 10,26; 11,10.30.31; 18,3; 36,22; 38,7.8.19.23.31.34;
45,17.20.25 [30] 39,10.15.17; 67,5; 84,17.17; 86,13
poiotes, quality, 11,32; 29,25; 42,6; semantikos, <semainein>, 11,7; 25,4;
49,18; 68,10; 73,22; 89,23.24.24 38,33; 39,3; 42,3; 98,15.21
poson, quantity, 11,30.32; 44,26; semeion, sign, 5,7; 21,14.20; 55,24;
45,17.19.26.29; 46,10.13; 48,26.27.29; 81,21; 100,8
50,9.11.25; 51,11.21.32; 85,18; semeioun, to infer, 12,2; 25,11; 38,5;
86,13.23; 95,16 62,25; 104,8
pragmateia, study, 1,3.13; 2,7.34; 3,4.6; skMma, figure, 6,16.17.19.20.21.23.27
6,29; 8,11.20.22; 14,21.23; 20,25.27; [335]
28,4.11.17.29; 29,15; 42,24 skhesis, aspect, relation, 17,30; 21,5;
pragmateuesthai, to study, 2,6.9.27; 44,18; 45,2; 47,14.21; 57,11
6,14; 8,1 skopos, aim, 8,4.15.18; 9,5; 20,25
problema, problem, 7,18; sophistes, sophist, 49,14
44,16.16.17.19.23.29; 45,1.3.4.5; sophistikos, sophistical, 1,4; 7,8; 8,28;
46,29.22.23.24.31.32; 47,6.29; 48,22; 14,16.18.20; 28,28; 49,7.8; 94,8.10
70,28; 75,23.27.33 ster&ikos, privative, 31,6; 35,23; 37,5;
proegoumenos, primary, 2,20; 8,20; 59,28.30; 67,2; 78,5.12
19,14; 69,27; 70,16; 112,14.18 stokhastikos, conjectural, 39,31
prokeisthai, to be in question [pro- sullogismos, syllogism, 2,24; 4,21;
keimenon, the point at issue], 1,3; 6,12.13.15.23.24 [290]
4,22.24; 8,9.12.13; 18,29 [55] sullogistikos, syllogistic, 1,3;
proskategorein, to co-predicate, 15,6; 6,11.16.20.22.29; 8,21 [200]
27,3; 40,19.25; 41,13 sullogizein, to syllogize, 9,22; 12,17.30;
proslambanein, to co-assume, 13,7.17; 14,3; 17,3; 21,15.19; 42,30;
19,5.6.9.24.24.26.28; 20,19.24; 21,6; 43,7.8; 64,8; 75,24.31; 77,26; 108,27;
22,2.12; 23,27; 29,11; 31,12; 34,10; 109,12.25.26; 111,3
44,2.31; 80,22; 113,24; summetaballein, to alter together with,
115,9.23.25.25; 116,13.16.20 52,20
proslepsis, co-assumption, 19,13.17.21; summetapiptein, to change together
20,9.9.14.15.21; 21,29; 22,10 with, 52,22; 56,25
prosresis, adjunct, 27,13 summetaskhematizein, to change confi-
prostheke, addition, 23,1; 24,7; 26,30; guration with, 52,22
27,3 sumparalambanein, to introduce
prosthesis, addition, 15,25.26; 21,24 together with, 11,13
prostithenai, to add, to annexe, 9,6; sumperainein, to conclude, 18,30; 21,3;
11,9; 12,7; 13,12.28; 15,5.8 [40] 49,10; 94,9
protasis, proposition, premiss, 1,19.20; sumperasma, conclusion, 7,16.17; 14,14;
6,26.27; 7,24.30; 9,26 [420] 17,23.24.27; 18,10 [120]
protattein, to prefix, 47,23; 49,4; 72,14 sumperilambanein, to encompass
proteinein, to propound, 44,15 together with, 50,6
prothesis, purpose, 1,16.20.20; sumpiptein, to coincide with, 19,15.28;
8,4.19.32; 9,5; 75,33 20,4
protithenai, to propose, to purpose, 6,31; sumploke, conjunction, 17,14; 19,22.26;
240 Indexes
42,13; 45,9; 47,11; 50,17 [65] suzeugnunai, to combine, 98,7
sunagein, to deduce, 2,1; 19,8; 21,28; suzugia, combination, 6,21; 10,7; 18,19;
22,6.9; 23,32; 24,8 [155] 21,3; 33,27; 42,4.5 [240]
sunagdge, deduction, 24,2; 35,7;
54,8.11.19.22; 55,3; 61,1.5; 72,8; tattein, to order, to rank, 14,26; 45,31;
76,11; 91,3.12; 93,28; 109,32; 111,8; 75,31; 77,6; 102,14.32
115,12 taxis, order, rank, 1,12; 2,19; 7,16; 10,30;
sunaktikos, deductive, 62,14 14,22; 41,7; 42,11 [35]
sunaletheuein, to be true together with, tehhne, art, 1,6;
29,18.22.28; 30,21.25; 41,7; 46,6.9; 2,4.5.11.16.16.17.18.19.23.29.30.33.
66,13; 88,20; 104,30; 105,30 35; 3,8.19; 39,23.31
sunanairein, to reject together with, teleios, perfect, 2,32; 6,23; 7,25; 10,1;
84,26 15,30; 16,1.5 [30]
sunaptein, to link, 56,11 teleiosis, perfection, 93,23
sunekhes, continuous, 6,1.2; 17,30; teleioun, to perfect, 78,22; 112,6.7.10;
20,14.14; 81,18 113,6
sunemmenon, conditional, 17,30; 20,17; telos, end, 1,19; 2,2.8.14.17.29; 3,13;
100,18 7,15.28; 20,25; 22,29
sunetheia, common usage, 23,10.11; terein, to keep, 16,1; 29,25; 61,2; 94,30;
54,25 97,25; 102,28
sunethes, in common use, 14,25 theorem, to theorise, 3,1.1.20.24.28;
sunistanai, to construct, 2,10; 3,23; 9,17.18
4,12; 29,5 theorema, theorem, 29,4
sunkeisthai, to be compounded, 6,15; thedretikos, theoretical, 1,14; 3,13.22;
9,26; 15,1.4.14; 81,23; 82,16; 117,32 4,7; 5,19
sunkhorein, to concede, 2,1; 3,2; 17,3.23; thedria, theory, 3,14.28; 4,1.5.12.29.34;
18,26.27; 31,22; 33,8; 50,30.30.31; 5,6.21; 6,2.6.9; 28,30
51,1.3; 90,6; 103,22 thesis, posit, position, 13,11; 16,26; 23,4;
suntattein, to connect, 15,21; 46,22; 24,16; 46,23.28; 47,22; 53,12; 57,27;
48,7.9; 51,27; 72,3.19 71,19.21; 72,13.19; 94,3; 96,9.12;
suntaxis, connection, 15,30 98,6.7.12; 105,4; 118,8.18
suntelein, to contribute, 2,15; 4,28; 8,17; tithenai, to posit, to place, to put, 2,1;
20,27; 28,1.24 4,25; 7,1.3.26; 13,9; 17,4 [40]
sunthesis, compounding, 7,12.13.14; tropikon, 'tropic', 20,7
15,8; 17,13.14.15.17.18; 42,2.14.17 tropos, mode4,26,15.18.30; 27,13;
suntithenai, to compound, 17,1; 27,22; 28,5.7.10.12.13.18.20.21.25; 40,12;
42,6.10.16; 45,5; 50,20; 52,1.7.7.14; 47,14; 53,30; 117,28
82,20
sunuparkhein, to hold at the same time, zetein, to investigate, to seek, 2,10.34;
90,11 3,1.2.11; 17,17; 23,24; 41,3; 46,33;
sustasis, construction, 1,7.12; 47,6; 70,4; 72,17; 73,25; 74,5.29
2,8.15.24.27.34; 28,25; 43,4 zetesis, <seek>, 74,28; 75,8

* Excluding the numerous non-technical occurrences.

 Index of Passages

Numbers in bold type refer to the works cited; numbers in ordinary type refer to the
pages of this book.
ALBINUS
Didasc. 153,47n.40; 158H, 108n.36; 158-64H, 7n.51; 156-7H, 49n.50; 157H, 50
n.53; 158-9H, 31 n.137; 158-60H, 73 n.129; 158H, 49 n.49,56 n.25; 161H, 45
n.22; 179-80H, 47 n.40
ALEXANDER
inAn.Pr. 121,15-123,9,174n.36; 121,15-123,24,88n.24; 123,28-124,7, 82n.l67;
124,8-30,113 n.58; 125,3-6,7 n.52; 125,30,3 n.23; 126,1-8,87 n.20; 125,26-8,
116 n.71; 126,29-127,2,113 n.58; 127,16,3 n.23; 129,9,133 n.140; 136,1-2,109
n.41; 140,14-141,6,93 n.45; 144,4-6,53 n.3; 151,14-16,53 n.3; 155,20-25,93
n.45; 156,27-157,2,93 n.45; 158,24-161,2,99 n.66; 164,17-165,15,98 n.60;
164,23-165,6,43 n.ll; 164,30,66 n.87; 165,8-15,98 n.61; 165,10,98 n.62;
168,31-169,10,98 n.60; 173,32-174,19,113 n.58; 188,16,3 n.23; 191,17,3 n.23;
193,21,3 n.23; 199,7-10,99 n.66; 201,21-24,93 n.45; 203,3-5,133 n.140;
203,15-35,132 n.137; 207,35,3 n.23; 210,30-2,53 n.3; 213,11-27, 9 n.60;
213,26,3 n.23; 214,12-18,9 n.60; 218,20-4,99 n.66; 219,35-221,5,100 nn.68,
69; 219,35-221,13,98 n.63; 221,6,133 n.140; 221,16-227,9,100 n.67; 236,8-11,
9 n.60; 238,22-38,9 n.60,10 n.65; 238,37,3 n.23; 240,32-241,9,9 n.60;
249,15-250,2,10 n.65; 249,25-32,9 n.60; 249,38,3 n.23; 250,2,3 n.23;
256,12-14,23 n.107; 256,32-258,25,108 n.36; 257,1-4, 70 n.112; 257,6-12,64
n.71; 257,8-13,64 n.72; 258,24,23 n.107; 259,8-260,6, 77 n.148; 260,23-261,2,
73 n.127; 261,25-6,23 n.107; 262,5-9,67 n.93; 262,29-31,69 n.103; 262,32-5,65
n.79; 263,11-12,69 n.103; 263,26-36,67 n.93; 266,8-267,27,133 n.143;
266,8-270,8,101 n.l; 268,7-8,73 n.127; 270,6-8,9 n.60; 270,10-28,82 n.168;
271,16-272,10,108, n.37; 271,2-6,145 n.26; 275,32-7,49 n.50; 282,14-15,17
n.85; 283,3,17 n.85; 284,17,3 n.23; 284,20,17 n.85; 284,29,17 n.85; 297,4-23,
122 n.94; 297,22-3,105 n.25; 300,3,98 n.61; 301,9-19,139 n.7; 317,31-318,10,
77 n.148; 326,31-2,118 n.83; 328,6,3 n.23; 328,10-30,13 n.77; 329,30-330,5,82
n.168; 331,12-24,49 n.49; 340,11-12,48 n.42; 343,21-344,6, 71 n.118; 344,9-13,
65 n.78; 344,9-346,6,72 n.125; 344,27-31,65 n.78; 345,18-20,134 n.135;
346,10-14, 74 n.133; 348,29-32,64 n.67; 349,5-7,107 n.34; 350,11-16, 71 n.118;
350,30-352,26,130 n.125; 372,26-373,9,50 n.60; 372,29-30,154 n.65;
373,16-30,154 n.65; 379,20-1,65 n.78; 379,14-380,27,116 n.71; 381,8-12,139
n.7; 381,28-386,2,50 n.62; 386,5-30,67 n.93,77 n.148,102 n.5; 388,18-399,9,
136 n.157; 390,3-4,65 n.79; 390,9,3 n.23; 392,19-26,31 n.139; 406,32-5,62
n.55,105 n.26; 411,35-7,82 n.168; 414,9-10,116 n.71; 415,10-12,116 n.71
Coreu. 83 n.l; 56,48 n.43; 56-7,108 n.36; 57-9,49n.49; 59,51 n.68; 60,4 n.25,83
n.4,136 n.157; 60-1,83 n.5; 61,85 n.12,116 n.71; 63,84 n.10,89 n.27,90 n.34;
64-5,87 n.19; 65,86 n.14,90 n.33; 65-6,91 n.38; 66-8,85 n.12; 69,84 n.ll;
69-74,84 n.ll; 76-7,92 n.41
DA 2,4-9,9 n.62; 90,2-9,67 n.90; 90,10-91,6,47 n.40
Fat. 164,1-3,1 n.l; 164,14-15,1
in Metaph. 1,10-2,3,46 n.29; 2,3-21,47 n.34; 2,24-5,13,46 n.31; 17,5-18,14,47 n.34;

241
242 Indexes
18,5-13,44 n.18; 260,1-20,49 n.49; 72,9-12,45 n.21; 103,5-104,18,44 n.16;
208,28-209,34,55 n.17; 351,9-15,72 n.122; 415,29-31,1 n.3; [531,25-6], 2 n.6
Quaest. 4,4-7,49 n.50; 61,1-28, 98 n.61; 141,2-4,47 n.34.
in Top. 2,6,23 n.107; 2,15-3,24,49 n.49; 2,16-29,116 n.71; 2,20-23,41 n.3; 2,23-25,
41 n.3; 2,25-26,41 n.3; 7,10-15,14,63 n.62; 7,11,4 n.25; 8,8-14,65 n.67;
8,14-9,19, 64 n.72,65 n.75; 8,29-9,5,74 n.136; 9,5-8,65 n.78; 9,17-19,64 n.74;
9,20-10,16,43 n.ll; 10,7-10, 66 n.88; 10,8-13,66 n.87; 10,14-16, 70 n.112;
10,26-8,49 n.49; 10,28-30, 70 n.lll; 10,30-13,10, 67 n.92; 11,23-7,69 n.104;
12,10-24,55 n.14; 12,31-13,10,66 n.81,68 n.96; 13,25-8,65 n.75; 13,28-14,2, 74
n.136; 14,17-15,14, 72 n.125; 22,10-14, 41 n.3; 23,21,105 n.21; 27,7-31, 44 n.14;
27,24-31,51 n.67; 32,12-34,5,98 n.61; 40,13-41,16,104 n.18; 62,6-13,103 n.7;
69,13-70,11, 58 n.32; 74,29-75,3,41 n.5; 86,21-87,6,104 n.16; 174,5-176,26,131
n.128; 190,26-193,7,83 n.5; 191,6-12,83 n.l; 257,22-3,80 n.158; 264,19,84 n.8;
288,13-289,8,86 n.13; 302,23-304,6,67 n.89; 321,26-8,67 n.90; 367,16-20,49
n.47; 421,3-34,63 n.64; 425,16-426,5,75 n.139; 566,25-7,66 n.87; 566,25-31,
70 n.112; 568,18-23, 74 n.136; 574,10-15,65 n.75; 582,23-6,83 n.l; 584,9-12,41
n.5
[ALEXANDER]
inSE 50,8-52,28, 74 n. 137
AMMONIUS
in An. Pr. 1 n.53; 2,18-3,30,49 n.49; 4,36-7,53 n.l; 5,5-7,25,48 n.42; 5,10-6,2-4,50
n.57; 7,6-14,108 n.36; 7,25,49 n.50; 8,15-11,21,41 n.l; 9,26-7,42 n.7;
9,37-10,1,43 n.9; 12,6-10, 53 n.4; 13,17-14,4,55 n.12; 14,13-22, 61 n.48;
15,30-16,9,55 n.12; 17,10-32,57 n.28; 19,3-21,30, 58 n.32; 22,12-30,61 n.51;
22,34-24,24,61 n.52; 23,25-24,5,61 n.54; 26,2-6,64 n.74; 26,6,65 n.76; 26,30-3,
64 n.66; 26,34-27,4,104 n.20; 27,6-14,64 n.67; 27,14-33, 64 n.72,65 n.75;
27,34-28,20,66 n.83; 27,35-28,5,66 n.88; 28,5-9,66 n.87; 28,21-29,19,103 n.7;
28,23-5,104 n.13; 28,32-29,2,104 n.12; 29,20-34, 71 n.116; 31,11-32,7,6 n.45;
31,14-32,7,75 n.141; 31,15,7 n.54; 31,22-3, 75 n.141; 32,12-14,66 n.87; 32,13,
66 n.88; 32,30-3, 75 n.141; 33,28-34,5, 78 n.152; 35,7-36,15,83 n.l; 36,8,106
n.28
in Int. 2,9-3,6,64 n.66; 2,21-5,55 n.13; 3,7-15,23 n.107; 3,19-20, 67 n.93; 5,1-23,64
n.66; 7,32-8,4,61 n.51; 9,7-10,59 n.37; 10,2-4,104 n.20; 14,18-26,62 n.56;
14,22-5, 61 n.49; 15,16-30, 56 n.18; 22,18-24,61 n.49; 27,1-3, 61 n.53; 44,11-19,
62 n.59; 57,29-32,62 n.56; 62,10-12,62 n.59; 64,26-65,2,64 n.66; 66,14-28,56
n.18; 67,30-68,9,56 n.18; 70,7, 31 n.135; 77,13-15,26 n.118; 80,15-35, 56 n.18;
80,31-5, 57 n.28; 81,13-84,25,26 n.118; 87,13,31 n.135; 88,12-28,80 n.160;
89,2-36,105 n.25; 90,19-20,29 n.127; 91,4-93,18,27 n.121,106 n.29; 94,7-24,
29 n.127; 100,21-3,86 n.13; 101,14-108,36,122 n.94; 106,13-16,86 n.13;
111,10-120,12,86 n.13; 114,22-3,19 n.93; 118,7-8,29 n.127; 122,20-26,19
n.93; 153,13-15,79 n.157; 153,13-26,93 n.45; 161,11-16,84 n.6; 165,4-30,61
n.54; 165,10-16,105 n.26; 181,5-7,45 n.26; 185,14-18,86 n.14; 194,15,84 n.8;
214,31-215,3,82 n.168; 215,7-28,80 n.160; 221,11-229,11,99 n.64; 240,7-13,93
n.45; 246,9,19 n.93; 249,1-25,45 n.26; 251,31,189 n.17 257,10-13,133 n.140;
257,30-258,4,19 n.93; 272,3,83 n.6
in Porph. Isag. 36,7-9,50 n.57; 37,7-13,49 n.52
[AMMONIUS]
in An. Pr. VIiI.20-21,139 n.7; VIII.24-5,139 n.7; IX.22-XI.36,102 n.5; X.10-XI.1,
139 n.7; XII.10-16,134 n.144; XIII, 191 n.23; 38,38-39,2,113 n.58; 39,2,7 n.54;
39,9 (app.crit), 139 n.7; 39,31,7 n.52; 40,2-40,113 n.58; 43,30-1, 7 n.54;
45,42-46,1, 99 n.66; 68,8-13, 70 n.lll; 68,6-7,69 n.103; 68,7-13,67 n.93; 68,28,
83 n.5; 70,11-13, 72 n.125; 70,20-22,134 n.145; 71,3-4,134 n.145
ANONYMOUS
in Int. An. Pr. 100,100 n.68
in Theaet. 7,14, 47 n.40; 66,12-43, 73 n.129
Logica62, 74n.l37
 Index of Passages 243
APOLLONIUS DYSCOLUS
Synt. 265,9-10,50 n.56; 326,11-327,12,50 n.56
APULEIUS
Dog. Plat. 126,3-6,47 n.40
Int. 1 n.50; 176,1-3,41 n.4; 176,13-14,62 n.59; 177,3-10,56 n.25; 177,15-17,86
n.13; 178,1-18,61 n.49; 179,16-181,17,106 n.29; 181,6,19 n.93; 181,19-183,6,
83 n.l; 183,15-21,110 n.47; 183,22-6,104 n.20; 184,19-23,65 n.75; 184,23-31,
66nn.87,88; 185,10-20,103 n.7,104nn.l2,13; 185,23-186,10, 111 n.54; 186,5,
12 n.75; 186,5-10,112 n.57; 188,4-11,48 n.45; 189,19-27,168 n.18; 189,23-5,
136 n.157; 190,15-191,5,145 n.30; 190,25-191,5,170 n.21; 192,30-193,5,31
n.137; 193,7-13,136 n.157; 193,9-16,113 n.62; 193,16-20,136 n.157; 194,23,12
n.75
ARISTOTLE*
An. Pr. 24alO-ll, 71 n.113; 24alO-15,51 n.66; 24al2-13, 76 n.143; 24al7,60 n.41,
81 n.162; 24a30-bl, 60 n.43; 24blO-12,60 n.45; 24bl8-22,75 n.142,104 n.17,
154 n.67; 24b20,134 n.147; 24b22-6,144 n.25; 24b23-4,135 n.149; 24b26-8,
116 n.72; 25al4-17,147 n.40; 25a20-22, 88 n.22; 25a29-34,92 n.41; 25a38-9,
93 n.43; 25bl4-15, 95 n.55; 25bl9-25,99 n.64; 25b22, 61 n.54; 25b27,102 n.4;
25b29,49 n.48; 25b30,48 n.41; 25b32-5,107 n.34; 26a2-9,124 n.113,131
n.131,175 n.40,185 n.4; 26a9-12,12 n.76; 26al3-14,121 n.92; 26al5,17 n.85;
26al7-29,127 n.116; 26a20-21,127 n.lll; 26a25-8,174 n.39; 26a28-30,113
n.62; 26a29-30,86 n.13; 26a30-31,123 n.95; 26a36-9,185 n.4; 26blO-16,159
n.86; 26bll-13,132 n.139; 26bl4-20,159 n.85; 26b21-5,129 n.121,164 n.100;
26b33,135 n.152; 27a5-9,145 n.32; 27a5-14,121 n.93,169 n.19,189 n.19;
27a9-14,146 n.34; 27al8-20,159 n.84; 27a20,156 n.76; 27a20-23,158 n.82;
27a23,164 n.103; 27a23-5,151 n.58; 27a26-b2,163 n.98; 27a32-6,157 n.81,
189 n.20; 27a36-bl, 190 n.21; 27a36-b3,170 n.20,173 n.30; 27b3-36,163 n.99;
27b4-6,154 n. 70; 27b6-8,151 n.55,185 n.4; 27b20-3,132 n.137; 27b27-9,132
n.137; 28al6,145 n.26; 28al7-23,191 n.24; 28al8-22,194 n.34; 28al8-23,188
n.12; 28a22-26,89 n.26; 28a22-26,89 n.28; 28a26-9,194 n.34; 28a26-30,178
n.48,191 n.24; 28a30-33,180 n.52,185 n.4; 28a33-6,182 n.58; 28a36-8,185
n.4; 28a37-b4,176 n.41; 28b7-ll, 193 n.32; 28bll-14,192 n.27; 28bl5-20,194
nn.35,38; 28b31-5,182 n.57,192 n.27; 28bl5,181 n.55; 28bl5-20,155 n.71;
28b22-4,155 n.72; 28b24-31,132 n.137; 29al9-27,135 n.156,136 n.157;
29a27-9,86 n.13,113 n.62; 29b24,196 nn.43,44; 29b29-32,79 n.157; 30a5-14,
174 n.36; 30b35-40,93 n.45; 32al8-b2, 79 n.157; 32a29-bl, 99 n.66; 32a31-b3,
99 n.64; 32bl8-22,98 n.60; 34al6-19,64 n.72; 34b7-18, 79 n.157; 36b35-37a31,
98 n.63,100 n.67; 38a29-31,12 n.75; 38bl8-20,12 n.75; 39b3-6,12 n.75;
40b30-7,64 n.72; 40b36,67 n.93; 41a2-20,108 n.36; 41a23-32, 77 n.148;
41a39,67 n.93; 41b6-22,133 n.143; 41b6-31,101 n.l; 41b33,145 n.26;
41b36-42a40,64 n.71; 41b36-42b26,108 n.37; 44al3,118 n.83; 43b3,118 n.83;
43bl7-22,122 n.94; 43b32-6,98 n.60; 43b43-44a2,88 n.21; 46b40-47bl4, 50
n.59; 47al6-20,74 n.136; 47a22-3,65 n.75; 47a22-31,74 n.133; 47bl5-50a4,50
n.60; 49b37-50al, 103 n.7; 50a5-15,50 n.62; 50al6-28,64 n.69; 50al6-b4,102
n.5; 51a2,3,18,22,50 n.58; 51b3-52b34, 99 n.64; 53a3-14,135 n.154,136
n.157,186 n.7; 53bl6-23,64 n.72; 59b8-10,27 n.119; 59b8-ll, 106 n.30;
61al8-33, 77 n.148; 62a9-10,125 n.106; 63b23-30,106 n.30; 64al5-17,103 n.7;
64bll-13,103 n.7; 64b28-65a37, 70 n.112; 65a38-b40, 74 n.137; 68b8-14,102
n.6,103 n.7; 68bl3-37,104 n.14; 68b35-7,104 n.16; 68b38-69al9,104 n.14;
69al3-16,103 n.7; 69al6-19,103 n.7; 70a2-b38, 72 n.119
An. Post. 71b20-72a24,58 n.34; 71b21-2,60 n.44; 71b22, 72 n.120; 72a8-9,56 n.21;
72al3-14,56 n.22; 73a7-ll, 64 n.72; 73a29, 79 n.157; 73b26-7, 57 n.30;
73b26-7, 78 n.153; 75b25-6, 79 n.157; 77a22-3,150 n.51; 77b30,122 n.94;
79al7-32,110 n.52; 79a29-32,109 n.40; 79b23-5,125 n.104; 85al3-86b39,110
n.48; 87al-30,181 n.54; 87a36,150 n.48; 87bl9-27,98 n.60; 96a8-19,98 n.60;
97b37-9, 75 n.139; 99bl5-17,53 n.6

* See also: SCHOLIA.)

244 Indexes
Cae/.300a7-ll,50n.54
Cat. 4alO-ll, 150 n.50; 4b20,150 n.49; 4b23,150 n.53; 14a29-35,49 n.46;
15a4-7,49n.47
ZM414b20-32,55n.l7
1218al-8,55n.l7
EN 1094a6-16,43 n.10; 1094blO-l, 51 n.69; 1095a30-b4,49 n.52; 1106b36,47 n.35;
1112a34-b8,98n.61; 1112a34-bll, 47,n.37; 1113alO-ll,47n.36; 1144b8-9,47
n.32; 1177b26-31,47 n.38; 1177b30-1178a4,47 n.40; 1178b7-23,47 n.34;
1178b25-8,47 n.40; 1179a33-1181b23,52 n.70
GC329a20-4, 50n.54
HA 488bl2-26,47 n.32; 492b23-6,104 n.12; 6Hal5-19,47 n.32; 614bl8-21,47
n.32; 618a25-30,47 n.32
Int. 16al-2,61 n.49; 17a2-3, 55 n.ll; 17a4,64 n.66; 17a8-9,55 n.16; 17a9-10,61
n.49; 17a23-4,56n.l9; 17a24,56 n.22; 17a25-6,57 n.28; 17a33,26n.ll8;
17a37-bl2,130 n.124; 17a38-bl6,57 n.29; 17bl2-16,122 n.94; 17bl6-20, 87
n.18; 17b29-37,86 n.13; 19bl9,61 n.54; 21a34-22al3,99 nn.64,65; 21b9-10,
62 n.56; 22al3, 82 n.168; 22a24-33,92 n.40; 23a7-26, 79 n.157; 23a6-ll, 93
n.43; 23a7-20,95 n.54
Metaph. 980a21-6,46 nn.28,29; 980a27-981al2,46 n.31; 980b22-5,47 n.32;
983a5-10,44 n.18; 989b31-3,45 n.21; 995a33-b2,51 n.64; 995b23-4, 58 n.35;
999a6-13, 55 n.17; 1013bl7-23, 72 n.122; 1016b24-6,150 n.51; 1017a27-30,62
n.56; 1019b30-33,93 n.43; 1027al9-26,98 n.60; 1027a23-4,98 n.62; 1053a30,
150 n.53; 1072bl3-30,47 n.34; 1072bl4-16,47 n.38
Meteor. 357a24-32, 75n.l39
PA 654a6-23,44 n.16; 691b5-16,104 n.12
Phys. 184a21-3,45 n.23; 194a7-9,45 n.21; 195al6-21, 72 n.122; 233b33,150 n.51
Pol. 1290b38-1291a8,52 n.71
Rhet. 1355a4-6,103 n.7; 1356a35-blO, 103 n.7; 1356bl6-18,63 n.63; 1357b25-
1358a2,103 n.8; 1357b26-30,103 n.7; 1393a25-1394al8,103 n.8
SE 164b27-165a2,63 n.63; 165b27-8,103 n.7; 167b21-36,74n.l37; 168b22-6, 74
nn.135,137
Top. 100a25-7,63 n.63; 100a29-30,58 n.33; 100b21-3,60 n.42; 100b23-5,60 n.46;
101a25-b4,43 n.ll; 101a28,44 n.14; 101a27,51 n.67; 101bl6-36,104 n.18;
103b3-7,103 n.7; 104a8-9,58 n.32; 105al3-16,103 n.ll; 105al6-19,104 n.16;
108blO-ll, 103 n.ll; 113bl5-26,83 n.5; 121a27-39,67 n.89; 123a33-7,75
n.139; 139b32-140a2,75 n.139; 161b28-30, 74 n.136; 162a32-4,74 n.137;
163b9-ll, 41 n.5
[ARISTOTLE]
Probl. 955b9-21,51 n.65
ARISTOXENUS
Harmonics II30,51 n.64
BOETHIUS
in Cic. Top. 1133BC, 85 n.12
in Int. (ed.sec.) 10,14-11,13,4 n.28; 12,3-16,4 n.29; 16,6-18,26,56 n.18; 119,10-
125,27,56 n.18
Int. Syll. Cat. 767C, 55 n.13; 769A, 62 n.56; 776C-778A, 86 n.13; 778B, 130 n.134;
785B-790C, 83 n.l; 786D-787A, 90 n.34;
Syll. Cat. 797BC, 64 n.66; 798C-799A, 106 n.28; 800AC, 27 n.121,802C-803B, 86
n.13; 804A-809A, 83 n.l; 805B-806A, 85 n.12; 809C-810C, 87 n.20; 810B, 28
n.124; 810CD, 116 n.71; 812D-813B, 109 n.41; 813BC, 136 n.157; 813C, 168
n.18; 814B, 139 n.7; 814C-816C, 136 n.157; 815A, 136 n.157; 819A-821A, 170
n.21; 819B, 168 n.18; 821BC, 65 n.75; 821C-822A, 66 n.88; 822C, 74 n.136;
823A, 48 n.45
CICERO
Nat. Deorum III xv 38,47 n.34
DAVID
in An. Pr. 7 n.54
in Cat. [see ELIAS]
 Index of Passages 245
in Porph. hag. 88,12-89,10,103 n.7
DEXIPPUS
in Cat. 12,29,31 n.135
DIODORUS SICULUS
History XIII91-5,103 n.8
DIOGENES LAERTIUS
Lives of the Philosophers II109,84 n.ll; III 24,50 n.53; III 65-6,6 n.41; V 23,48
n.42; V 88,5 n.37; VI19,5 n.37; VII65,62 n.59; VII68,66 n.88; VII69,66
n.81; VII76,67 n.93; VII78,153 n.63; DC 15,5 n.37
ELIAS
in An. Pr. 134, 7 n.54; 134,4-138,13,41 n.l
Proleg. 27,15-22,46 n.30
'ELIAS' [i.e., David]
in Cat. 107,24-5,8 n.57; 122,25-123,11,8 n.57,9 n.62; 123,4-8,4 n.28; 128,13,4
n.28
EPICTETUS
Diss. IIxivll-13,47n.40
EUCLID
Elements 11,73 n.128; XIII1-5,50 n.53
EUDEMUS
[F re Wehrli] 11,113 n.58; 13,99 n.66; 14,99 n.66; 15,100 n.68; 16,100 n.68; 26,63
n.61; 27,62 n.56,63 n.61
EUSEBIUS
PVIix,2n.l3
EUSTATHIUS
in Od. 1107, 6 N.41
EUSTRATIUS
in An. Post. 3,4-7,49 n.52
GALEN ,
Anat. Admin. II218K, 1 n.3
in Hipp. Aph. XVIIIA186-7K, 5 n.35
in Hipp. Fract. XVIIIB 318-19K, 8
in Hipp. Off. XVIIIB 630-2K, 9 n.59
/rest Log. i 5,116 n.68; ii 2-3,61 n.49; ii 5,105 n.25; iii 1,56 n.25; iii 3,65 n.79; iii 5,
154 n.65; iv 1-4,131 n.128; vi 3,84 n.8; vi 3-7,83 n.l; vi 5,83 n.4; vi 6,68 n.99,
69 n.102; vi 4,83 n.5; vii 1,69 n.103; vii 2,6 n.45; viii 1,21 n.98; viii 1-4,48
n.45; xi 3-7,136 n.157; xix 6, 72 n.125; ix 1-2,50 n.58; x 1-6,170 n.21; x 2, 50
n.58; xi 6,115 n.65; xi 7,136 n.157,168 n.18; xii 1,108 n.36; xviii 2-4,73
n.!29;xix6,153 n.63
Lib. Prop. XIX 42K, 6 n.48; XIX 47K, 7 n.49
Meth. Med. X 206,98 n.62
OptSectI114-5K,98n.62
PHPV224K,66n.87
Sem. IV581K, 103 n.7
Simp. Med. Temp. XI463-6K, 84 n.8; XI470K, 103 n.7; XI498-500K, 84 n.8
GELLIUS
Attic Nights XVI viii 1-8,62 n.59
HERODOTUS
Histories II68,104 n.12
IAMBLICHUS
Comm. Math. Sc. 55,21-2,45 n.22
MARTIANUS CAPELLA
DeNuptiisMercuriietPhilologiaeJV411,136 n.157; IV 413,170n.21
OLYMPIODORUS
in Gorg. 46,3,46 n.30
in Meteor. 263,21,4 n.28
246 Indexes
PHILODEMUS
Rhet.v34-vi 19,98 n.61
PHILOPONUS
in An. Pr. 7 n.54; 2,14,64 n.74; 2,22-4,14,49 n.49; 4,26-9,46 n.27; 5,15-6,6,48 n.42;
6,19-9,20,41 n.l; 7,31,43 n.9; 9,28-32,53 n.6; 9,28-10,25, 53 n.4; 10,3-25,53
n.l; 11,25-36, 55 n.12,104 n.20; 12,5-12,78 n.152; 12,23-20,3,56 n.18; 21,20-3,
4 n.28; 22,4-15,104 n.20; 22,23-23,17,58 n.32; 25,30-30,20,61 n.52; 26,5, 7
n.54; 33,6-10,64 n.67; 33,10-23,64 n.72,65 n.75; 33,23-6,66 n.88; 34,7-10,12
n.75; 34,10-14,104 n.13; 34,12-30,103 n.7; 34,21-6,104 n.12; 36,5-13, 72
n.125; 36,19-37, 75 n.141; 37,16-38,29, 77 n.148; 39,15-24, 78 n.151; 39,32-
42,34,83 n.l; 42,20,84 n.8; 42,22,106 n.28; 43,18-44,1,80 n.160; 46,25-47,11,
116 n.71; 47,14-15,89 n.30; 48,11-18,86 n.14; 48,18-49,5,84 n.ll; 49,6-31, 87
n.19; 52,4-56,5,92 n.40; 52,27-9,92 n.41; 53,15-56,5,99 n.64; 56,15-57,13,93
n.44; 57,24-58,27,94 n.50; 61,16-62,4,98 n.60; 63,12,19 n.93; 63,19-64,7,100
n.70; 64,11-15,64 n.74; 65,4-66,26,108 n.38; 65,20-23,139 n.7; 67,18-30,109
n.42; 67,18-68,8,140 n.ll; 67,27-30,143 n.19; 68,8-69,29,113 n.61; 69,30-
71,17, 111 n.54; 71,12-17,113 n.58; 74,30-75,15,12 n.75; 77,22-9,123 n.95;
79,4-5, 86 n.13; 79,4-9,113 n.62; 79,10,136 n.157; 79,10-20,136 n.157;
82,1-84,11,29 n.127; 82,21-7,128 n.120; 82,34-83,4,128 n.120; 87,10-19,140
n.ll; 87,30-88,2,145 n.26; 88,26-33,145 n.30; 90,20,86 n.16; 92,9-12,151 n.55;
94,3-7,145 n.30; 98,4-12,29 n.127; 98,6,19 n.93; 105,28-106,2,170 n.21;
109,20-1,128 n.120; 112,21-113,20,136 n.157; 123,12-17,113 n.58; 126,21,4
n.28; 129,16-19,113 n.58; 162,16-28,50 n.53; 188,20-1,87 n.20; 198,19-20, 87
n.20; 201,18-19,87 n.20; 243,4-8,69 n.103; 243,8-10,67 n.93; 255,27-9,145
n.26; 277,26-32,122 n.94; 277,29-31,105 n.25; 302,17-19,118 n.83; 307,6-8,49
n.50; 321,7-322,18, 72 n.125; 321,12-14,134 n.145; 388,18-399,9,136 n.157;
451,15-453,33, 70 n.112; 456,24-457,7,74 n.137
in An. Post. 111,31-112,6,4 n.28; 155,24-5,105 n.25; 301,26-8,150n.48; 334,25-
335,3,50 n.57; 416,13-31, 75 n.139
in Cat. 11,20,104 n.20
in DA 21,20-3,4 n.28
PLATO
Cratylus 431B, 61 n.49
Laws 730C, 47 n.33
Phaedrus 249C, 47 n.39
Philebus 33D2, 75 n.140; 55E-56C, 98 n.61
Rep. 408C, 73 n.129; 511BC, 49 n.52; 526C-527C, 45 n.22; 528D-530C, 45 n.21;
566B, 103 n.9; 613A, 47 n.40
Sophist 262D, 61 n.49
Theaetetus 155AB, 105 n.22; 176B, 47 n.40; 191C8-9, 75 n.140; 206D, 61 n.49
PLUTARCH
Ser. Num. 550D, 47 n.40
Stoic. Rep. 1034E, 51 n.68
PORPHYRY
Vit.Plot. 14,2n.l3,6n.41
PROCLUS
in Eucl. 21,15-22,16,45 n.22; 38,4-10,45 n.21; 43,18-21,49 n.52; 211,18-212,4, 50
n.53
in Rep. II96,6n.41
in Tim. 175,30-76,10,5 n.36
PTOLEMY
Synt. 6,21-7,4,45 n.21
SCHOLIA TO ARISTOTLE
139a36-140alO, 53 n.l; 140a35-41,49 n.50; 145a30-7,125 n.106; 146a9-18,61
n.48; 146al9-27,62 n.56; 146a24-7, 63 n.61; 147b47-148a2,65 n.75; 151a46-
b4,108 n.38; 155b8-19,170 n.21; 156b43-157b9,6 n.45, 75 n.141; 157al3-24,
86 n.14; 188a4-12,136 n.157; 294b23-9,66 n.87
 Index of Passages 247
SENECA
Ep. 88,21-8,41 n.l, 44 n.20; 89,5,44 n.18; 89,9,41 n.4; 92,27,47 n.40
SEXTUS EMPIRICUS
Mil 13,98 n.61; VIII70-1, 62 n.59; VIII93-5,66 n.88; VIII108,84 n.8; VIII108-9,
66 n.88; VIII223,21 n.98; VIII438-43, 74 n.136; VIII441-3,65 n.75; DC 13,44
n.18
PHll 104,62 n.59; II110-112, 70 n.108; II112,66 n.88; II156,21 n.98, 74 n.136; II
158,68 n.99,69 n.102; II166-7, 65 n.75; II185,104n.l2; II198,48 n.45
SIMPLICIUS
in Cat. 1,3-2,29,6 n.48; 7,23-32,8 n.57; 10,8-19,48 n.42; 159,23-33,6 n.48
in DA 52,27-8,4 n.28
in Phys. 80,15-16,4 n.28; 320,1-10, 72 n.122; 795,33-5,4 n.28; 707,33,4 n.28;
1170,13,4 n.28; 1176,32,4 n.28
SOCRATICORUM RELIQUIAE (Giannantoni)
IIB,8-ll,84n.ll
STOICI
CWMrT

Fragmente zur Stoischen Dialektik (FDS) 691, 75 n.138; 1051, 64 n.70; 1084, 153
n.61; 1087, 72 n.121; 1090, 133 n.142; 1099, 48 n.44; 1100, 76 n.144; 1101,
117n.74, 117 n.81, 135 n.153, 189 n.15; 1106, 50 n.55; 1118, 74 n.134; 1163, 83
n.3; 1171, 66 n.83; 1184, 86n.l5
Stoicorum Veterum Fragmenta (SVF) II 49a, 42 n.6; II 260, 72 n.124, 73 n.131, 133
n.142; II 261, 66 n.83, 67 n.91, 68 n.97, 69 n.105; II 264, 153 n.61, 154 n.64; II
265, 69 n.101; III Ant 27, 65 n.75
SYRIANUS
in Metaph. 100,6, 4 n.28
THEMISTIUS
Max. 75 n.141; 180, 84 n.ll; 180-3, 109 n.41; 181-2, 108 n.38, 140 n.ll; 182-3, 139
n.6; 184, 109 n.41, 171 n.24; 186, 84 n.ll; 190-1, 6 n.45, 86 n.14; 191-2, 174
n.37; 192, 181 n.54
THEOPHRASTUS
[F n Graeser] 4, 131 n.128; 5, 125 n.106; 13, 86 n.14; 14, 93 n.46; 15, 100 n.68; 16, 99
n.66; 17, 136 n.157, 185 n.6; 19, 168 n.18; 21, 170 n.21; 24-5, 113 n.58
 Index of Persons

Names in quotation marks occur only in examples.

Achaicus, 6 Bacchius, 5 n.35

Adrastus, 6 Boethus (Peripatetic), 6, 75 n.141,86
Alexander of Aegae, 6 n.14,174 n.37
Alexander of Aphrodisias,
'The Commentator', 4 n.28 Caracalla, 1,1 n.l
conception of logic, 11 Chrysippus, 68 n.99,69 n.102,153 n.63
and Galen, 2,10,12 Grantor, 5
influence of, 7
life of, 1-2 David, 7 n.54
logical failings of, 12-13,90-1,93 n.44, 'Dionysius', 103,103 n.ll
96 n.56,118-19,127 n.114,128
n.117,160 n.89,161 n.93,171 n.24, Elias, 7 n.54
184 n.3 Epicureans, 5 n.30
originality of, 11-12,12 n.75 Eubulides,85n.ll,n.l2
philosophy of, 3 Euclid, 73, 73 n. 127, n. 128
and Plato, 16,47,73,75 Eudemus, 4,62 n.56,63,63 n.61,86,99
predecessors, 7,9 n.66,109 n.41,113 n.58
and Stoics, 2,43 n.12 Eudorus, 6 n.43
style, 9-10,10 n.66,14,194 n.36 Eutocius, 7 n.54
teachers, 2
works, 2-3,92 n.40 Galen, 2,6,8,10,12,115 n.65,136
Alexander of Damascus, 1 n.3 n.157,137 n.157,168 n.17,170 n.21
Ammonius, 7,16,27 n.120,49 n.49,61 Geta, 1 n.l
n.52,62 n.56,75 n.141,80 n.158,
105 n.26,
Andronicus, 6 Heraclides of Pontus, 5, n.37
Antisthenes, 5 n.37 Herminus, 2,6,7 n.52,11 n.68,140,140
Apollonius (Peripatetic), 6 n.l2,n.l3,141n.l5,160,162
Apollonius Dyscolus, 18 n.91 Herophilus, 5 n.35
Apuleius, 7,136 n.157,137 n.157,168
n.17 lamblichus, 7 n.54, 75 n.141
Ariston,6,136 n.157
Aristotle Julian (Emperor), 75 n.141
commentaries on, 3-7
criticized, 7,9,9 n.60,11,84 n.ll, 87, Marcus Aurelius, 1
87 n.19,92-3,92 n.41,99 n.66,128 Marinus, 7 n.54
n.120,129,134,170,170 n.21,223 Maximus, 75 n.141,86 n.14
syllogistic, 12-14,136 n.157,208-15
Aristotle of Mytilene, 2 n.7 Peripatetics, 6,9,21,31 n.137,41 n.4,64
Aspasius, 6 n.66,71 n.141, 79 n.157,109 n.41,
Athenodorus, 6 n.43 131n.l28,212n.l6
Avicenna, 3 n.22,70 n.109 Philoponus, 7,16,56 n.18,61 n.52,67

248
 Index of Persons 249
n.93,75 n.141,78 n.151,92 n.41,93 Sosigenes, 2
n.44,94 n.50, 111 n.54,113 n.62, Sotion, 6
128 n.120,129-30,136 n.157,140 Stoics, 2,6 n.43,21,26,41 n.4, n.12,43
n.ll n.12,44 n.18,62 n.59,68,69 n.103,
Tisistratus', 103 72,134,153 n.63,154 n.64, n.65
Plato, 5,16,47 n.33,50 n.53,73,73 Syrianus, 7 n.54
n.129, 75,75 n.140,105
Platoniats, 6, n.43,7,99 Theagenes', 103
Plotinus, 2 Themistius, 7 n.54,75 n.141,85 n.ll,
Porphyry, 7 n.54,56 n.18, 75 n.141,168 181 n.54
n.17 Theo', 88
Posidonius, 41 n.l Theophrastus, 4,4 n.29,56,56 n.18,
Proclus, 5,7 n.54,8 n.57,9 n.61, n.62,46 n.24,74 n.141,86,87 n.19,89,93,
n.27 99 n.66,100,109 n.41,113 n.58,125
Pythagoreans, 150-1,150n.48,156 n.106,130,131 n.128,135-6,136
n.157,137 n.157,168 n.17,170
Septimius Severus, 1 n.21,185-6
Sextus Empiricus, 10
'Socrates', 62-3,82,173 Zeno (Stoic), 51 n.68
 Index of Subjects

actual, 30,71,79,79 n.157,195,209 conditional propositions, 65 n.79, 70

affirmative, 27,55-7,187 n.lll
agreement, 104-5 conjectural arts, 98,98 n.61
analysis, 22,49-50,121,170-1,171 n.24 contingency, 29-30,71, 79-80,79 n.157,
Aristotle criticized, 7,9,9 n.60,11,84 84,95-100
n.ll, 87,87 n.19,92-3,92 n.41,99 kinds of, 93,95,96-8
n.66,128n.l20,129,134,170,170 contradictories, 26-7,56,56 n.22,106,
n.21 119,160-1
assertion, 55 contraries, 27,106,125,157,160
assumption, 104-5 conversion, 3, 31,65 n.77,106-7,164-5,
astronomy, 45,45 n.21 181,181 n.54,195,210-11
axioms, 23 n. 109,105 and generation of figures, 109,109
n.44,138-9,168,170
Baralipton, 136n.l57,213 modal, 82,84,91-100
Barbara, 77 n.148,117,136 n.157,189, of particular affirmatives, 87,89, 90-1,
210,212 121-2
Barbari, 136 n.157,214 and particular negatives, 85 n.12,91
Baroco, 77 n.148,152-3,158,190,213 and reduction, 76,81,83,135,144,
begging the question, 70 n.112 145-6,169,170,185,186,188,189,
Bocardo, 155,177-8,211,213 192
types of, 83-4
Camestre, 168 n.17,214 of universal negatives, 84 n.ll, 86-90,
Camestres 136 n.157,147-8,168 n.18, 140
169,189,213 used in reduction, 147-8,152,172,174,
Camestrop, 136 n.157,215 176,177,180
canonical premisses, 73 co-predication, 61,80,99
categories, 45,45 n.24,142 copula, 61-3,105 n.26
Celantes, 136 n.157,213
Celantos, 136 n.157,214
Celarent, 117-8,136 n.157,189,212 Dabitis, 136 n.157, 214
Celaront, 136 n.157,214 Darapti, 88 n.21,89,136 n.157,168 n.18,
Cesare, 136 n.157,147,168 n.18,169, 171-4,188,191,211,213
189,210-1,213 Daraptis, 136 n.157,168 n.18,214
Cesares, 168 n.18,214 Darii, 123-4,127,136 n.157,149,168
Cesaro, 136 n.157,214 n.18,190-1,213
Cesaros, 215 Datisi, 136 n.157,176-7,192-3,213
co-assumption, 24,67 n.93, 70,72-3 Datisis, 214
combination, 23,105,107-8,114-15 definition, 63-4,63 n.64,74
non-syllogistic, 12-13,21,114-15, demonstration, 20,24,45,46, 67,71,81,
118-20,124-30,133-4,150-1, 82,110,148
155-64,175-6,178-80,181-3 and the Analytics, 48-9,51,53-4,53
commentaries, 4-7 n.1,71,102
on Aristotle, 6-7 and syllogism, 48,102
canons of, 8,8 n.57,9 n.62 demonstrative propositions, 57,57-60
on Plato, 5-6 diagrams, 139,139 n.7,140 n.9,146,171
conclusion, 66-71,105,112 n.26

250
 Index of Subjects 251
dialectical arguments, 24,51,67,81,82, hypothetical syllogisms, 3,6 n.45,102,
110 102 n.5
dialectical propositions, 57,57-60
dictum de omni et nullo, 87,87 n.20, imperfect syllogisms, 21,50, 75-6,135,
116-17,117 n.35,124,135 146,149,183,210
differentia, 63-4 indemonstrables, 21,76,117,189
Disami, 214 indeterminacy, 28-9,29 n.127
Disamis, 136n.l57,176,192-3,213 = absence of quantifier, 85-6,86 n.13,
disjoint arguments, 67-9,67 n.92 111, 113-14,113n.62,124,125,126,
disjunction, 131 n.128 164,167-8,187,209 n.3
disproof, 12-13,85,91,159,160 of particulars, 130-1,131 n.128,132-3,
dissimilar in form, 27,156 n.77,184-5 157,158,159,162,163,164,180,
double negation, 66 n.81 182
duplicated arguments, 66-7,66 n.88, induction, 66,71,103-4,103 n.7,104
69-70 n.12
is, 62,105
enthymeme, 103 n.7 justification, 19,102-4,103 n.7,109,110
eristic arguments, 67
etymologies, 44 n.17,64 n.74 knowledge, 46
every, of, 54,77-8
existence, 23 n. 106,62 lemmata, 8,8 n.58,15,17,17 n.85
explanation, 72,72 n.122 letters, 16,116 n.71,146,172
exposition, 26,88,88 n.24,130 n.126, logic, 41-3,67
159,169,173,174n.36,175,177, Stoic [see Index of Persons, s.v. Stoics]
178,189,211 utility of, 41,43,43 n.ll, 46 n.27,66,
70,82,97-8,104
Fapemo, 136 n.157,214 value of, 46-8
Fapesmo, 136 n.157,214 logoi, kinds of, 64,64 n.66
Faresmo, 215
Felapton, 174-5,178,191,213 major term, 32,109,109 n.43,140-4,167
Ferio, 124,127,149,161 n.93,191,213 n.ll, 176
Ferison, 180-1,192-3,213 major premiss, 109,167,167 n.ll
Festino, 152,157,189,213 matter, 30,45,48,50,58,59,81-2,81
figures, 32,48,91,101,107-8,209-10 n.165,85,114,116,119
first, 108-16,212-13 meaning, 154
fourth, 108,108 n.36,210 n.5 medicine, 98,98 n.62
ranking of, 108-10 metaphor, 74
second, 110,138-9,154,213 middle term, 32,105,107-8,115,138,
third, 110,154,166-8,213 146,166,171,209
Firesmo, 136 n.157,214 minor term, 32,109,109 n.43,140-4,
form, 30,45,48,50,114,119 166,167 n.ll
for the most part, 97-8,98 n.60 minor premiss, 109,167,167 n.ll
Frisemo, 136 n.157,214 mixed modes, 9 n.60,82 n.167
Frisesomorum, 136 n.157, 214 modal propositions, 79-80,79 n.157,80
n.158
modal syllogisms, 9 n.60,29 n.128,82
generation of figures, 108-10,109 n.41, modes, 29-30,80-1,81-2,116,195,209
138-9,166,168,170-1,171 n.24,186
genus, 45,49,55-6,56 n.18,63-4,67, nature, 97-8
140-2,140 n. 13 necessity, 29,79-80, 79 n.157,80 n.160,
geometry, 44-6,49-50,50 n.53,59, 73, 81 n.165
86-7 of the consequence, 71
gods, 44,47,47 n.40 kinds of, 93
Xsyllogism, 21,65,66,134,162
hypotheses, 25,59,179 negative, 27,55,57,59,143-4,187
hypothetical propositions, 23,56,56 non-differently concluding arguments,
n.25,64 66,66 n.87
252 Indexes
reduction of syllogisms, 11-12,22,50,76,
not of some/every, 125,125 n.106, 177,187-95
126,132,153,164 reliable, 114-15,114 n.65
nouns, 61 reputable propositions, 58,60,60 n.42
numbers, 150 requests, 58
opposition, 26,106,106 n.29 separability, 80,80 n.158,86,86 n.14
signs, 72
paradigms, 66,103,103 n.7 single-assumption arguments, 64,65-6,
parts, 28,88,103,103 n.7,111-12,125, 71,73
126,128,132,15i; 155 singular terms, 130 n.125
particles, 131-2,132-3,133 n.140 sophistical arguments, 51,74,82,
peiorern rule, 113 n.58 110-11,166
perception, 45,46,88-9,173,175,189 sophistical propositions, 60
perfect syllogisms, 6 n.45,21-2,50, 75-6, speech, parts of, 50
75 n.141,84 n.ll, 86 n.13,102, subalterns, 106,131,157
108-9,115-16,117-18,124,135, subaltern moods, 136 n.157
144,188,189,190,210 subcontraries, 27,27 n.120,106,157,
philosophy, parts of, 41-6,41 n.4 159,162,163,180
practical, 42 subject, 25,61,61 n.49,105,208
theoretical, 42,44 subpart, 28,42
Porphyrean tree, 140 n.13 subsyllogistic arguments, 153-4,153
possibility, 3,29,97-8 n.63
potential, 121,145,145n.26,146,164-5, superfluous arguments, 73,73 n.136, 73
179 n.137
predicate, 30-1,57,61,61 n.49,105,105 syllogisms, 21,63-4,104,208
n.15,118,122 n.94,208,208 n.2 additional, 134-7,184-7,212,213-15
predicative propositions, 27,30,56,56 kinds of, 49 n.49,67,82,110-11
n.25,64,78 ranking of, 113,145,145n.30,168-70,
kinds of, 81 176,177,178,180 2D1
premisses, 22 n.104,104-5 reflected, 136 n.157
principles, 59
Prior Analytics, title of, 48-50,48 n.42, terminology, 10,11-12,17-33,136 n.157
102 terms, 28,61-2,208
priority, 49,49 n.46,55-6, 71,145 universalXparticular, 121-2,122 n.94
problems, 104-5,110,144 text of Alexander's commentary, 16-17
property, 91 text of Prior Analytics, 53,53 n.3,195
propositions, 22,54-60,62,104-5,208-9 n.41,198
time, 56,56 n.22
quality, 27,56,56 n.22,57,110,112, and modality, 79-80, 79 n.157
167-8,209 truth, 22-3,26,46,55,79,80-1,84 n.6
quantification of the predicate, 105 n.25,
122 n.94 unmethodically concluding arguments,
quantifiers, 28-9,105,105 n.25,129-30, 11 n.68,20,72-4,72n.l25, 76,134,
130 n.125,173,209 134n.l44,134n.l48
quantity, 27-8,57,105,110,112,167-8, units, 150
209 universals, 28,57
unreliable, 114-15,114 n.65,119
reciprocal proof, 87,92 variables, 15 n.80,116 n.71
reduction to the impossible, 25, 70 n.109, verbs, 61
74 n.137,76-7,77n.l48,86,89,146, virtues, 47
154,158,161-2,179,187,193
used in reduction of syllogisms, 148-9, whole, in as in a, 28,28 n.124,54,77-8,
152,153,172,174,177,188,191, llln.55,112,116,123n.98
211