Oxford Studies in Ancient Philosophy, Volume 62 2023933494, 9780192885180, 0192885189

Table of contents :
Cover
Oxford Studies in Ancient Philosophy
Copyright
Advisory Board
Contents
Opposites and Explanations in Heraclitus
1. Introduction
2. Preliminaries on explanation and opposites
3. Opposites in Heraclitus’ predecessors
4. Opposites as explananda
5. Against a metaphysical foundation
6. Metaphysical interdependence
7. Conclusion: Opposites and explanation
Bibliography
Evaluative Illusion in Plato’s Protagoras
1. Introduction
2. Akrasia in the Protagoras
3. The puzzle of temporal distortion
4. The non-rationalinterpretation
5. What are appearances?
6. What is the power of appearance?
7. Why do proximate pleasures appear greater?
8. Conclusion
Bibliography
That Difference Is Different from Being: Sophist 255 c 9–e 2
1. Introduction
2. Text, translation, and linguistic concerns
3. ‘Participated in both forms’ at 255 d 4 and the ὥσπερ τὸ ὄν clause
4. Reading the argument with ὥσπερ τὸ ὄν adjoined to the apodosis
5. Conclusion
Bibliography
Is Plato a Consequentialist?
1. Introduction
2. Consequentialist principles in the Republic and the Laws
2.1. Laws and consequentialist principles
2.2. Principle H and consequentialism
3. Objections to consequentialism in Plato
4. Evidence for and against consequentialism in Plato
4.1. Demandingness
4.2. The do/allow distinction
4.3. Doing bad things
4.4. Impartiality
5. The harm principle
6. The Laws’ theology
6.1. The petteia-player
6.2. Why does the Coincidence obtain?
6.3. Rational eudaemonism, A1, and the Coincidence
6.4. Book 10’s theology and our self-understanding
Bibliography
Aristotle’s Argument for the Necessity of What We Understand
1. We think we only understand necessities?
2. Understanding as a relative
2.1. Understanding is not simultaneous with its object
2.2. What are the objects of understanding?
2.3. Is the dependency principle specific to scientific understanding?
3. Understanding as a state
3.1. The tension between the two principles
4. A tempting solution
5. Aristotle’s argument for the necessity of what we understand
5.1. Why does Aristotle hold that contingencies are KOWT?
5.2. How does the claim that contingencies are KOWT establish that understanding is of necessities?
6. Objections
6.1. Objection 1: Durability without eternal truth
6.2. Objection 2: Eternal truth without necessity
7. Concluding remarks
Bibliography
Aristotle on Digestion, Self-Motion, and the Eternity of the Universe: A Discussion of Physics 8. 6 and De somno
1. Introduction
2. Physics 8. 6: The difficulty and its context
3. The passage (Physics 8. 6, 259b1–20)
4. Three interpretations and their shortcomings
5. A new interpretation and its evidence in Physics 8
6. De somno on sleep and digestion
7. Temporal continuity and the ou kuriōs claim in Physics 8. 6
8. The broader context
Bibliography
Giving Gifts and Making Friends: Seneca’s De beneficiis on How to Expand One’s Sphere of Ethical Concern
1. Introduction
2. The Stoic ideal
3. Benefaction, joint activity, and friendship
4. The uniqueness of benefaction and self-sufficiency
5. Conclusion: Friendship and oikeiōsis
Bibliography
Hierocles’ Concentric Circles
1. Four puzzles about the circles
2. Is Hierocles advocating complete impartiality?
3. Do the circles illustrate the theory of oikeios̄ is?
4. Finding out what is appropriate
5. Hierocles’ investigation of duties
6. Organic parts within parts
7. The circles as organic parts of the universal body
8. Solving the puzzles about the circles
Bibliography
Archaic Epistemology: A Discussion of Jessica Moss, Plato’s Epistemology: Being and Seeming
I
II
III
IV
V
VI
VII
VIII
Bibliography
Index Locorum
Notes for Contributors to Oxford Studies in Ancient Philosophy

Citation preview

OX F OR D ST U DIE S IN A NCIEN T PHIL O S OPH Y

OX FOR D ST UDIE S IN A NCIEN T PHIL OSOPH Y E DI T OR : V IC T OR CA S T ON

a s s o c i at e e di t or: r ac h a na k a m t e k a r VOLU M E L X II su m m e r 2 0 22

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © the several contributors 2023 The moral rights of the authors have been asserted All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2023933494 ISBN 978–0–19–288518–0 (hbk.) ISBN 978–0–19–288526–5 (pbk.) DOI: 10.1093/oso/9780192885180.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

A DVISORY BOA R D Professor Rachel Barney, University of Toronto Professor Gábor Betegh, University of Cambridge Professor Susanne Bobzien, All Souls College, Oxford Professor Riccardo Chiaradonna, Università degli Studi Roma Tre Professor Alan Code, Stanford University Professor Brad Inwood, Yale University Professor Gabriel Lear, University of Chicago Professor A. A. Long, University of California, Berkeley Professor Stephen Menn, McGill University and Humboldt-­ Universität zu Berlin Professor Susan Sauvé Meyer, University of Pennsylvania Professor Jessica Moss, New York University Professor Martha Nussbaum, University of Chicago Professor Marwan Rashed, Université Paris-­Sorbonne Professor David Sedley, University of Cambridge Professor Richard Sorabji, King’s College, University of London, and Wolfson College, Oxford Professor Raphael Woolf, King’s College, University of London Contributions and books for review should be sent to the Associate Editor, Professor Rachana Kamtekar, Sage School of Philosophy, Cornell University, 218 Goldwin Smith Hall, Ithaca, New York, USA (email [email protected]). Contributors are asked to observe the ‘Notes for Contributors to Oxford Studies in Ancient Philosophy’, printed at the end of this volume.

vi

Advisory Board

Up-­to-­date contact details, the latest version of Notes for Contributors, and publication schedules can be checked on the Oxford Studies in Ancient Philosophy website: www.oup.com/academic/content/series/o/oxford-­studies-­in-­ancient-­philosophy-­osap

CON T EN T S Opposites and Explanations in Heraclitus richard neels

1

Evaluative Illusion in Plato’s Protagoras41 suzanne obdrzalek That Difference Is Different from Being: Sophist 255 c 9–e 2 michael wiitala Is Plato a Consequentialist? christopher bobonich

85 105

Aristotle’s Argument for the Necessity of What We Understand167 joshua mendelsohn Aristotle on Digestion, Self-­Motion, and the Eternity of the Universe: A Discussion of Physics 8. 6 and De somno233 wei wang Giving Gifts and Making Friends: Seneca’s De beneficiis on How to Expand One’s Sphere of Ethical Concern allison piñeros glasscock Hierocles’ Concentric Circles ralph wedgwood

261 293

Archaic Epistemology: A Discussion of Jessica Moss, Plato’s Epistemology: Being and Seeming333 matthew evans Index Locorum

363

OPPOSITES AND EXPLANATIONS IN HERACLITUS richard neels 1. Introduction Heraclitus was clearly interested in opposites.1 He was also interested in explaining the cosmos and its constituents.2 In fragment B 1 he promises to explain the nature of each thing (ἐγὼ διηγεῦμαι, I would like to thank Keith Begley, Victor Caston, Patricia Curd, Tom Davies, Daniel Graham, Mahadi Hassan, David Hitchcock, Tyler Hopkins, Mark Johnstone, Howard Jones, Melle van Duijn, and an anonymous referee for Oxford Studies in Ancient Philosophy for their helpful comments on various drafts of this paper. 1  Heraclitus was famous in antiquity for claiming that opposites were identical. See Aristotle, who claims that Heraclitus—­or at least followers of Heraclitus—­ transgressed the Law of Non-­Contradiction by claiming that opposites are ­identical: Top. 8. 5, 159b30–3; Phys. 1. 2, 185b19–25; Metaph. Γ. 7, 1012a24–b22. It should be noted that Aristotle calls this the ‘Heraclitean account’ (τὸν Ἡρακλείτου λόγον) in the Physics passage. Elsewhere he shows sensitivity to the fact that Heraclitus may have meant something different from what he appears to say about the identity of opposites (Metaph. Γ. 3, 1005b23–5). Some modern scholars have also argued that Heraclitus thought that opposites were identical. See K.  Popper, ‘Back to the  Presocratics’, Proceedings of the Aristotelian Society, 59 (1958), 1–24 at 13; C.  Emlyn-­Jones, ‘Heraclitus and the Identity of Opposites’, Phronesis, 21(1976), 89–114; J. Barnes, The Presocratic Philosophers (London, 1982), 57. The issue with this view is that none of the extant fragments identifies opposites in such a way that transgresses the Law of Non-­Contradiction. The more popular view today seems to be what is known as the Unity of Opposites—­or what I call the standard view. See n. 4 below for prominent advocates of this weaker view. 2  By explanation, I mean an account of why something is thus and so. See B. van Fraassen, The Scientific Image (Oxford, 1980), ch. 5, who argues that explanations are answers to why-­questions. Heraclitus never poses why-­questions specifically, but his fragments are puzzling so as to induce why-­questions in his reader. See also R. J. Hankinson, Cause and Explanation in Ancient Greek Thought [Cause and Explanation] (Oxford, 1998), 3–4 for an account of explanation particularly relevant to early Greek philosophy. Finally, see J. Moravcsik, ‘Heraclitean Concepts and Explanations’, in K. Robb (ed.), Language and Thought in Early Greek Philosophy (La Salle, 1983), 134–52. Moravcsik argues that there are three levels of explanation in very early Greek thought: (1) explanation in terms of origin, (2) explanation in terms of constituency, and (3) explanation in terms of a thing and its attributes (134). According to Moravcsik, Heraclitus’ writings represent a shift from the second to the third level of explanation (135). I am in agreement with Moravcsik’s general thesis, though this paper attempts to give much more detail to Heraclitus’ theory and method of explanation.

Richard Neels, Opposites and Explanations in Heraclitus In: Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Richard Neels 2023. DOI: 10.1093/oso/9780192885180.003.0001

2

Richard Neels

κατὰ ϕύσιν διαιρέων ἕκαστον) and to show how things are (ϕράζων ὅκως ἔχει).3 But how exactly do opposites feature in Heraclitus’ explanation of the cosmos? Most scholars believe that Heraclitus espoused a so-­called ‘unity of opposites’ doctrine.4 According to this standard view, Heraclitus claimed that opposites were somehow united and that opposites permeate the world to some extent. So, 3  I have adopted the standard Diels-­Kranz number of the fragments found in H. Diels and W. Kranz, Die Fragmente der Vorsokratiker, 6th edn [Fragmente] (Berlin, 1951), but am using the Greek text from D. W. Graham, The Texts of Early Greek Philosophy: The Complete Fragments and Selected Testimonies of the Major Presocratics [Texts] (Cambridge, 2010), where available, and Diels–­Kranz otherwise. Translations are also generally from Graham, though I have made significant variations throughout. 4  The view in its modern conception was initiated by G. S. Kirk, Heraclitus: The Cosmic Fragments [Cosmic Fragments] (Cambridge, 1954), esp. at 69. Most scholars since Kirk have followed his view or something very similar to his view. See, for example, G.  Vlastos, ‘On Heraclitus’, American Journal of Philology, 76 (1955), 337–68 at 367; G. E. R. Lloyd, Polarity and Analogy (Cambridge, 1966), 96–102; M. Marcovich, Heraclitus: Greek Text with a Short Commentary, 2nd edn [Heraclitus] (Sankt Augustin, 2001), 158–9; M. C. Stokes, One and Many in Presocratic Philosophy (Washington DC, 1971) at 97; A. P. D. Mourelatos, ‘Heraclitus, Parmenides and the Naïve Metaphysics of Things’ [‘Naïve Metaphysics’], in E. N. Lee, A. P. D. Mourelatos, and R. M. Rorty (eds.), Exegesis and Argument: Studies in Greek Philosophy Presented to Gregory Vlastos (Assen, 1973), 16–48 at 33 and 35; C. Kahn, The Art and Thought of Heraclitus [Art and Thought] (Cambridge, 1979), 192; E. Hussey, ‘Epistemology and Meaning in Heraclitus’, in M. Schofield and M. Nussbaum (eds.), Language and Logos: Studies in Ancient Greek Philosophy Presented to G.  E.  L.  Owen [Language and Logos] (Cambridge, 1982), 33–59 at 38; D.  Wiggins, ‘Heraclitus’ Conception of Flux, Fire and Material Persistence’, in Schofield and Nussbaum (eds.), Language and Logos, 1–32 at 10; G. S. Kirk, J. E. Raven, and M. Schofield, The Presocratic Philosophers, 2nd edn (Cambridge, 1983), 188–9; T. M. Robinson, Heraclitus: Fragments (Toronto, 1987), 81 and 86; M. M. Mackenzie, ‘Heraclitus and the Art of Paradox’, Oxford Studies in Ancient Philosophy, 6 (1988), 1–37 at 7–8; D. O’Brien, ‘Heraclitus on the Unity of Opposites’, in K. I. Boudouris (ed.) Ionian Philosophy (Athens, 1989), 298–303; P. Curd, ‘Knowledge and Unity in Heraclitus’, Monist, 74 (1991), 531–49 at 539; Hankinson, Cause and Explanation, 29; E. Papamichael-­Paspalides, ‘The Concept of One in Heraclitus’, Revue de Philosophie Ancienne, 23 (2005), 41–54 at 42; A. Long, ‘Wisdom in Heraclitus’, Oxford Studies in Ancient Philosophy, 33 (2007), 1–17 at 1–2; J. Warren, Presocratics (Berkley, 2007), 67–70. There are three notable exceptions. For alternative views that break from the standard view, see R. Dilcher, Studies in Heraclitus [Studies] (Hildersheim, 1995), 103–16; C.  Osborne, ‘Heraclitus’, in C.  C.  W.  Taylor (ed.), Routledge History of Philosophy, vol. i: From the Beginning to Plato (London, 1997), 80–116 at 92–6; D. W. Graham, ‘Heraclitus’ Criticism of Ionian Philosophy’, Oxford Studies in Ancient Philosophy, 15 (1997), 1–50. Graham expanded and revised his argument in D.  W.  Graham, Explaining the Cosmos: The Ionian Tradition of Scientific Philosophy [Explaining the Cosmos] (Princeton, 2006.) I will make reference to the expanded and revised version of his argument in this paper. Osborne and Graham in particular make notable advancements in our understanding of Heraclitus’ use of opposites. I discuss these views in Section 2 below.



Opposites and Explanations in Heraclitus

3

on this view, Heraclitus was encouraging his readers to comprehend the hidden connectedness of opposites and, perhaps, the connectedness of all things.5 I will argue that the standard view is incomplete. Opposites are indeed connected, and I believe Heraclitus (and his Ionian predecessors) recognized this feature of the world. However, I do not think the demonstration of this fact was his primary goal in composing the opposites fragments. The central issue with the standard view is that any clear formulation of a unity of opposites principle is too restrictive to make complete sense of Heraclitus’ interesting and varied statements about op­pos­ites.6 In place of the standard view, I argue that Heraclitus’ treatment of opposites was a reaction to the way opposites were being used by his Ionian predecessors, Anaximander and Anaximenes.7 Opposites, for the earlier Ionians, seem to have been explanatory principles: fundamental explanantia. The physical world, which includes events, stuffs, and things, was explained by a limited set of oppositional pairs (e.g. hot and cold, condensation and rarefaction). Heraclitus’ treatment of opposites, I submit, is best understood in relation to these earlier schemes of philosophical explanation. I will argue that Heraclitus was the first Ionian to treat opposites not as explanatory principles but as philosophical problems in need of explanation. Heraclitus’ opposites are related to his philosophical explanation of the cosmos insofar as they are (a) negatively, explananda in need of explanantia and (b) positively, the clearest expression of the interdependent nature of the metaphysical structure of the cosmos and things in it. Heraclitus appears to have held a principle of non-­ 5 For example, Kirk writes: ‘The Logos is undoubtedly connected with the opposites, in fact it is the unity which underlies them and which binds together into one nexus all the components of the apparently discrete phenomenal world’ (Cosmic Fragments, 188–9). He also claims that ‘The Logos is the formula, structure, plan, of each and all things . . . As such it results in the fact that “all things are one” . . .  because they all connect up with each other because of this common structure’ (70, emphasis original). 6  I defend this claim below in Section 2. See Dilcher, Studies, 103–8 for another critique of the essential connection thesis. He points out that ‘it is equally easy to state this idea of unity-­in-­opposites and to apply it to any opposition that [we] might encounter, as it is exceedingly difficult to explicate with any reasonable precision in  what way these opposites actually “coincide.” The formula cannot be directly elicit­ed from any Heraclitean fragment. The relevant fragments present “op­pos­ ites” of the most diverse kind and are as varied in their formulation as they could conceivably be’ (104). 7  Anaximander and Anaximenes were residents of the Ionian city of Miletus, while Heraclitus lived in the nearby Ionian city of Ephesus.

4

Richard Neels

well-­foundedness, according to which there are no metaphysically foundational entities, and to a principle of reciprocal de­ter­min­ ation, whereby the cosmos and its parts are caught in a web of explanatory interdependence. In the end, the opposites too are part of this web of metaphysical interdependence and, as such, are both explananda and (non-­fundamental) explanantia in the cosmos.

2.  Preliminaries on explanation and opposites Fragment B 1 is taken by most to have been the opening of Heraclitus’ book. It will prove a helpful place to begin understanding what Heraclitus thought about explanation: τοῦ δὲ λόγου τοῦδε ἐόντος αἰεὶ ἀξύνετοι γίνονται ἄνθρωποι, καὶ πρόσθεν ἢ ἀκοῦσαι, καὶ ἀκούσαντες τὸ πρῶτον· γινομένων γὰρ πάντων κατὰ τὸν λόγον τόνδε ἀπείροισιν ἐοίκασι, πειρώμενοι ἐπέων καὶ ἔργων τοιούτων, ὁκοῖα ἐγὼ διηγεῦμαι, κατὰ ϕύσιν διαιρέων ἕκαστον καὶ ϕράζων ὅκως ἔχει. τοὺς δὲ ἄλλους ἀνθρώπους λανθάνει ὁκόσα ἐγερθέντες ποιοῦσιν, ὅκωσπερ ὁκόσα εὕδοντες ἐπιλανθάνονται. (22 B 1 DK) Of this logos which holds always, humans prove to be uncomprehending, both before they hear it and after hearing it for the first time. For although all things come to be in accordance with this logos, they are like the inexperienced experiencing words and deeds such as I explain when I distinguish each thing according to its nature and show how it is. Other humans are unaware of what they do when they are awake just as they are forgetful of what they do when they are asleep.8

Perhaps the most striking feature of this fragment is Heraclitus’ novel use of the Greek term logos.9 Prior to Heraclitus, the term logos was used to indicate a structured account of events.10 Heraclitus claims that ‘all things’ (πάντα) come to be in accordance with the 8  This fragment is preserved for us by both Sextus Empiricus (M 7. 132–3) and Hippolytus (Haer. 9. 9. 3). Both Aristotle (Rhet. 3. 5, 1407b14–15) and Sextus claim that B 1 is the opening of Heraclitus’ book. 9  Most interpreters have supposed that the term refers to some cosmic principle. For example, see Kirk, Cosmic Fragments, 39. This has been challenged by some who claim that it simply refers to Heraclitus’ account (i.e. his book). The most notable proponent of this view is perhaps M.  L.  West, Early Greek Philosophy and the Orient [Orient] (Oxford, 1971), 124–9. For my part, I agree with M. Johnstone, who argues for a middle position. See M. Johnstone, ‘On “Logos” in Heraclitus’ [‘Logos’], Oxford Studies in Ancient Philosophy, 47 (2014), 1–29. I deal with the philo­soph­ic­al significance directly in Section 5 below. 10  See Johnstone, ‘Logos’, 13–17.



Opposites and Explanations in Heraclitus

5

logos of which he speaks. Furthermore, the logos holds always (ἐόντος αἰεί ). At the very outset, Heraclitus seems to commit himself to a cosmos that has a stable structure.11 Furthermore, he seems optimistic that the stable structure of the cosmos renders the cosmos intelligible and explicable, even though he seems to think this is a very difficult task. Let’s examine some of the verbs contained in this fragment. Heraclitus claims to ‘explain’ (διηγεῦμαι) things. (I gloss ἕκαστον as ‘things’, since the ‘each’ refers to the ‘all things’ (πάντα) earlier in the fragment.) διηγεῦμαι has the sense of describing something in detail to someone who isn’t in the know.12 He also claims to show how things are (ϕράζων ὅκως ἔχει). ϕράζειν has the sense of showing or telling someone a piece of information.13 It was often used prior to Heraclitus in the sense of showing or telling someone the way to get somewhere.14 It was also used in the sense of relaying a command from a place of authority.15 Hence the term connotes the passing on of reliable, practical information from a source of authority. Finally, Heraclitus promises to differentiate (διαιρέω) things according to their nature (κατὰ ϕύσιν). διαιρέω has the sense of division or taking things apart.16 Heraclitus’ term for nature or phusis seems to mean something like a thing’s metaphysical character.17 The main 11  Of course, this has been the subject of great controversy. However, I believe it is safe to say that most scholars today doubt Heraclitus held a theory of radical flux whereby everything is always changing in every respect. See Graham, Explaining the Cosmos, 113–22 for a summary of the debate over the last century. I take it that the standard view today regarding flux is that Heraclitus’ cosmos does change but that the change is orderly and the order is stable. In Heraclitus’ own words: ‘Changing, it rests’ (μεταβάλλον ἀναπαύεται, B 84a). See Graham, Explaining the Cosmos, 129–37 for an account of flux in Heraclitus’ stable world. 12 For later uses of the term see Ar., Birds 198 and Antiph., Against the Stepmother, 13.8. See also Kirk, Cosmic Fragments, 41, who implies that the term for Heraclitus means ‘explanation’ in a strong sense of the term. 13 E.  Cunliffe, A Lexicon of the Homeric Dialect (Norman, 2012) includes ­‘display’, ‘indicate’, ‘instruct’, and ‘direct’ as translations of this term. 14  See Hom., Od. 11. 2 and 14. 3. 15  See Hom., Il. 10. 127 and Od. 10. 549. 16  Kirk claims that this term ‘means something more than merely “judging”, and implies a process of analysis’ (Cosmic Fragments, 41). This is, of course, a precursor to the great notion of diairesis in Plato and Aristotle. See Plato, Soph. 253 d 1–3 and Arist., Pr. An. 1. 31. 17  Phusis prior to Heraclitus meant something like ‘visible characteristic’. See Hom., Od. 10. 302–6. Pindar, writing around the same time as Heraclitus, also used the term ‘phusis’ to reference visible characteristics. See Pind., Isthm. 4. 50. However, Heraclitus elsewhere claims that ‘a nature tends to be hidden’ (ϕύσις κρύπτεσθαι ϕιλεῖ,

6

Richard Neels

verb of the clause is ‘explain’ (διηγεῦμαι) and Heraclitus uses the participle ‘differentiating’ (διαιρέων) most plaus­ibly to indicate his means of explanation. If this is right, Heraclitus is interested in explaining things according to their nature by differentiating them. These words apply to ‘all things’ (πάντα) referenced earlier in the fragment. So, it seems safe to say that Heraclitus was interested in explaining the way things are. But I do not think he was merely interested in cataloguing and describing the existents of the world. Elsewhere, in B 50, he claims that ‘all things are one’ (ἓν πάντα εἶναι, with Diels). Significantly, Heraclitus calls the world ‘kosmos’ in the sense of an ordered whole.18 All this strongly suggests that Heraclitus saw himself as explaining the structure and nature of all things (πάντα) which constitute the cosmos (κόσμος). It is difficult to know exactly what Heraclitus took to constitute the class of ‘things’. We might think it is all ‘things’ in a robust sense whereby things are the bearers of properties. But I think this is too strong for a Presocratic thinker writing before the incredibly nuanced discussions of Plato and Aristotle. Instead, I propose we treat ‘all things’ (πάντα) as referring to all the aspects of the cosmos: things, properties, stuffs, and events. At any rate, these are what he discusses in his fragments. All these aspects of the cosmos have a

B 123). I take this to indicate Heraclitus has metaphysical character (i.e. something closer to an essence) in mind rather and not just a loose description of visible characteristics. See A. P. D. Mourelatos, The Route of Parmenides, revised and expanded edn (Las Vegas, 2008), who states that ‘The term clearly carries for [Heraclitus] a sense of real constitution or inner nature’ (6). See also K. Begley, ‘Heraclitus against the Naïve Paratactic of Mere Things’, Ancient Philosophy Today, 3 (2021), 74–97, for an interpretation of Heraclitus according to which his interest in phusis is in terms of essences. For a more thorough analysis of the term ‘phusis’, see my ‘Phusis, Opposites and Ontological Dependence in Heraclitus’ [‘Phusis’], History of Philosophy Quarterly, 35 (2018), 199–217 at 200–3. 18  See C. Kahn, ‘Anaximander’s Fragment: The Universe Governed by Law’, in A. P. D. Mourelatos (ed.), The Pre-­Socratics: A Collection of Critical Essays (Princeton, 1993; originally published, 1974), 99–117. He states: ‘Precisely con­sidered, the kosmos is a concrete arrangement of all things, defined not only by a spatial dis­pos­ ition of parts, but also by the temporal taxis within which opposing powers have their turn in office’ (111). In B 30 Heraclitus refers to the world as κόσμον τόνδε. This is the first known philosophical use of the word κόσμος (although it is possible it was used before Heraclitus). See also G. Betegh and V. Piano, ‘Column IV of the Derveni Papyrus: A New Analysis of the Text and the Quotation of Heraclitus’, in C. Vassallo (ed.), Presocratics and Papyrological Tradition (Berlin, 2019), 179–220 at 198–200.



Opposites and Explanations in Heraclitus

7

character—­a nature or phusis—­and Heraclitus seems interested in explaining how all these aspects hang together as an orderly whole. B 1 has an analytic flavour. Heraclitus promises to break down (διαιρέω) the cosmos into smaller components in order to reveal the nature of its parts (ἕκαστον). This may give us the sense that Heraclitus intends to explain the cosmos by division. However, this is only partially true. If we turn to the extant fragments, we will see that Heraclitus spends as much time synthesizing discrete parts of the cosmos as he does analysing the cosmos into its parts. Some of the most famous fragments speak to Heraclitus’ synthetic project. B 50 claims that ‘all things are one’ (ἓν πάντα εἶναι, with Diels) and B 41 claims that ‘thought . . . steers all things through all things’ (γνώμην . . . ἐκυβέρνησε πάντα διὰ πάντων). I think B 10 provides a clear statement of Heraclitus’ balanced interest in synthesis and analysis: συλλάψιες ὅλα καὶ οὐχ ὅλα, συμϕερόμενον διαϕερόμενον, συνᾶιδον διᾶιδον, καὶ ἐκ πάντων ἓν καὶ ἐξ ἑνὸς πάντα. (22 B 10 DK) Collectives: wholes and not wholes; brought together, pulled apart; sung in unison, sung in conflict; both from all things one and from one all things.19

The point, I think, is that Heraclitus saw himself explaining the way the cosmos is and why it is that way. I will have a great deal more to say about the details of Heraclitus’ scheme of explanation below. For now, let it suffice that he seems to have thought that he needed to analyse the cosmos into its parts and synthesize its parts into a whole in order to explain the cosmos and its parts. But now let’s turn to the opposites so we can begin to inquire into their relation to Heraclitus’ scheme of explanation. To begin, it will prove useful to examine just a few of the opposites fragments in brief.20 Much more can and has been said concerning the following fragments. I will simply focus on features relevant to our question. My initial goal here is twofold: (1) to provide a glimpse 19  I discuss this fragment in much more detail in Section 5 below. 20 See the ‘Table of Opposites in Heraclitus’ Doctrine on the Logos’, in Marcovich, Heraclitus, inserted between pages 160–1, for a useful introductory guide to the opposing terms in Heraclitus. For each pair of opposites, Marcovich lists their ‘Reason for Unity’. But in the end, he finds eleven different reasons for the unity of various opposites. Marcovich does not attempt to explain how the eleven different reasons are related to one another. I take this as evidence that Marcovich has not given a full explanation of Heraclitus’ use of opposites.

8

Richard Neels

of the richness and variety of the fragments concerned with op­pos­ ites and (2) to show that the standard view—­that Heraclitus is merely displaying the internal connection of opposites—­is unable to account for that variety and richness. If this is true, then the standard view cannot account for how the opposites feature in Heraclitus’ explanation of the cosmos. I don’t mean to suggest that the standard view is incompatible with these fragments, but just that it fails to bring out what is most interesting about opposites and indeed what seems to be Heraclitus’ point in repeatedly referring to them. First, consider fragment B 57: διδάσκαλος δὲ πλείστων Ἡσίοδος· τοῦτον ἐπίστανται πλεῖστα εἰδέναι, ὅστις ἡμέρην καὶ εὐϕρόνην οὐκ ἐγίνωσκεν· ἔστι γὰρ ἕν. (22 B 57 DK) The teacher of the multitude is Hesiod; they believe he has the greatest knowledge—­who did not comprehend day and night: for they are one.21

I begin with this fragment because it is, perhaps, the strongest evidence for the standard view. As is well known, Hesiod treated the divinities Day and Night as distinct entities.22 Heraclitus appears to be correcting this by claiming that they are one, that is, day and night are opposing parts of a single meteorological process. The deeper point seems to be this: what-­it-­is-­to-­be-­day and what-­it-­is-­ to-­be-­night are connected by being conceptually interdependent—­ they belong to a single conceptual sequence. So, it does appear that Heraclitus recognized that opposites were connected. But it does not follow that this was a new discovery on Heraclitus’ part,23 nor does it follow that this was Heraclitus’ only or even primary interest in opposites. The interpretive strategy of the standard view has been to take this somewhat clear expression of a principle and read it onto all the other instances of Heraclitus’ opposites. But we shall 21 Dilcher, Studies, 109 takes an even stronger stance against the standard view. He argues that the ἕν cannot refer to day and night since day and night are feminine nouns and ἕν is neuter. Dilcher thinks this is good evidence against any notion of unity of opposites in Heraclitus. But this might be wrong, since it is possible, at least in slightly later Greek prose, for a neuter noun to refer to two feminine ones. Compare e.g. the use of the neuter ταὐτόν at Arist., Pol. 1. 7, 1255b16–17 (οὐ ταὐτόν ἐστι δεσποτεία καὶ πολιτική Ross). 22  See Mourelatos, ‘Naïve Metaphysics’, 33–4. 23  I think it is implausible that Anaximenes, for example, could use rarefaction and condensation as he does without realizing that the two processes are ­conceptually connected.



Opposites and Explanations in Heraclitus

9

see just how restrictive this move is when we examine other instances of Heraclitus’ opposites in their own right. Consider fragment B 61: θάλασσα ὕδωρ καθαρώτατον καὶ μιαρώτατον, ἰχθύσι μὲν πότιμον καὶ σωτήριον, ἀνθρώποις δὲ ἄποτον καὶ ὀλέθριον. (22 B 61 DK) Sea is the purest and most polluted water: for fish drinkable and healthy, for men undrinkable and harmful.24

On the surface, Heraclitus is making the true observation that seawater is both pure and polluted. This is a judgement of value: ­seawater is both good and bad. But Heraclitus is obviously not supposing that seawater is both A and not-­A at the same time and in the same respect. He specifies that seawater has these properties relative to different species. For fish, it is good to drink, but for humans, it is not good to drink. The deep meaning of the fragment seems to indicate something about value: opposing values are dependent on different kinds of subjects.25 This is an interesting and important philosophical insight concerning opposites and ­values. Now, according to the standard view, the meaning of the fragment must merely be that these opposites, as with all opposites, are connected; so, this fragment shows that purity and pollution are connected. But this seems to miss the deep insight concerning kind-­relative value, an insight that has to do with opposites but that cannot be cashed out in terms of the standard view. Here is another fragment concerning opposites: οὐ ξυνιᾶσιν ὅκως διαϕερόμενον ἑωυτῶι ὁμολογέει. παλίντροπος ἁρμονίη ὅκωσπερ τόξου καὶ λύρης. (22 B 51 DK) They do not understand how disagreeing with itself it agrees with itself: reciprocal structure as of a bow or a lyre.26 24  This fragment is also preserved by Hippolytus (Haer. 9. 10. 5). Hippolytus claims ‘[Heraclitus] says that the polluted and the pure are one and the same thing and that the drinkable and the undrinkable are one and the same thing’ (καὶ τὸ μιαρόν ϕησιν καὶ τὸ καθαρὸν ἓν καὶ ταὐτὸ εἶναι, καὶ τὸ πότιμον καὶ τὸ ἄποτον ἓν καὶ τὸ αὐτό). 25  This thought resurfaces in a few other fragments: B 9, B 13, B 11. For a further discussion of this insight concerning value, see Neels, ‘Heraclitus on the Nature of Goodness’, Ancient Philosophy, 41 (2021), 1–22. 26  This fragment is preserved for us by Hippolytus (Haer. 9. 9. 2). There is some disagreement over the word παλίντροπος; some scholars prefer παλίντονος (i.e. back-­ stretched), since the term was known in ancient times as a Homeric epithet for ‘bow’ (τόξον). Furthermore, Plutarch, who cites the fragment three times, once renders it παλίντονος. Marcovich, Heraclitus, and Kirk, Cosmic Fragments, support παλίντονος,

10

Richard Neels

The opposites cited in this fragment are disagreement and agreement, and the claim is that something can agree with itself while disagreeing with itself. This principle is applied to a bow and a lyre as illustrative examples of the principle. It is clear to see how a bow and a lyre disagree while agreeing with themselves. Take a bow. The stave and string are at odds with one another in that they exert opposing forces on one another. However, the bow agrees with itself in that these opposing parts constitute the very nature of the bow. So, at least for some objects, the opposites present in them are constitutive of the very nature of the objects in question. But again, the standard view must conclude that this fragment is merely saying that disagreement and agreement are connected. Again, this misses a rich philosophical insight about opposites and their role in constituting the nature of various objects.27 Finally, consider the following fragments together: ψυχρὰ θέρεται, θερμὰ ψύχεται, ὑγρὰ αὐαίνεται, καρϕαλέα νοτίζεται. (22 B 126 DK) Cold things warm up, hot things cool off, wet things become dry, dry things become moist.28 πυρὸς θάνατος ἀέρι γένεσις, καὶ ἀέρος θάνατος ὕδατι γένεσις. γῆς θάνατος ὕδωρ γενέσθαι καὶ ὕδατος θάνατος ἀέρα γενέσθαι καὶ ἀέρος πῦρ καὶ ἔμπαλιν. (22 B 76 b–­c DK) The death of fire is the birth of air, and the death of air the birth of water. It is death for earth to become water, and death for water to become air, and death for air to become fire and contrariwise.29

B 126 contains a true observation about changing opposites: cold and hot, wet and dry are connected in terms of change. B 76 contains a set of elemental transformations. I believe the two fragments resonate thematically. The opposites listed in B 126 happen to be what later Greeks associated with the four elements listed in while Kahn, Art and Thought, and Graham, Texts, support παλίντροπος. For my part, I am happy to leave the text as it has been handed down to us as παλίντροπος. 27  See Neels, ‘Phusis’, for more details on other fragments that seem to illustrate the same insight concerning opposing properties and their role in constituting objects. Importantly, rivers (B 12) seem to bear this oppositional structure. 28  This fragment is preserved by Tzetz., Scholia ad exeg. in Iliadem, 126 Hermann. 29 Diels, Fragmente, provides three variants of this fragment together. I have provided a combination of the second (from Plutarch) and third (from Marcus Aurelius). Graham, Texts, places these two together, but they may very well have been disparate quotations. I think it is helpful to see the two together, but I do not think my philosophical point rests on this precise composition.



Opposites and Explanations in Heraclitus

11

B 76.30 The fragments, if they do go together, suggest that the elem­en­tal stuffs which constitute material reality change and in virtue of this change things become warm or cool, wet or dry.31 This, again, is a deep philosophical insight about reality. But on the standard view, B 126 merely indicates that cold and hot, wet and dry are connected and B 76 merely indicates that death and birth are connected. While it is true that these opposites are connected, the interpretive strategy of the standard view misses the rich connection between elemental transformations and the changes between opposing properties. I don’t mean to suggest that scholars who accept the standard view can’t also recognize these various insights; however, the standard view does not explain how Heraclitus’ concern about the unity of opposites relates to his other concerns about opposites. There are two alternatives to the standard view that are worth considering. Catherine Osborne has argued that for Heraclitus ‘[i]dentity, similarity, difference, opposition’ are ‘all determined by the significance acquired in context’.32 She states that, according to Heraclitus, ‘what counts as the same and what counts as opposed is decided by a significance acquired in a social or temporal context, and is not determined absolutely by a fixed or material constitution in the entities we observe’.33 This view goes a long way towards breaking from the standard view. It is particularly good at explaining fragments like B 61 concerning seawater. However, it does not seem to be able to make good sense of a fragment like B 51 concerning the bow and the lyre. The opposition inherent in and the resulting identity of the bow and the lyre do not seem to be context dependent, and they do, in fact, seem to have a fixed constitution: without certain material conditions, the bow and the lyre fail to be what they are. Furthermore, oppositions qua opposites don’t seem to be determined by context: hot and cold may be relative to 30  This is most clearly seen in Arist., GC 2. 3. 31 See my discussion of Daniel Graham’s view later in this section. See also G. Betegh, ‘On the Physical Aspect of Heraclitus’ Psychology’ [‘Physical Aspect’], Phronesis, 52 (2007), 3–32, who incorporates Heraclitus’ soul into the physical flux of elements. On Betegh’s view too, there is an association between the elements and the elemental properties, including mental properties. There may be some disagreement about the precise relations of the properties to the elements, but that there is a relation seems to be agreed on by most scholars. 32  Osborne, ‘Heraclitus’, 94. 33  Osborne, ‘Heraclitus’, 80.

12

Richard Neels

context, but qua opposites they seem to be relative merely to one another. While I think Osborne captures something Heraclitean and while her interpretation does much better than the standard view in accounting for deep philosophical insight, it cannot fully account for the variety of opposites. As such, it fails to provide a satisfying solution to the relation between Heraclitus’ opposites and his philosophical explanation of the cosmos. Daniel Graham presents another alternative to the standard view: opposites are the same just in the sense that opposite things or stuffs turn into one another . . . They are, moreover, quantitatively equivalent in the sense described, by bearing a determinate ratio to one another. To say that opposites are the same is simply to say that they are transformationally equivalent.34

Graham’s interpretation of Heraclitus’ use of opposites works well with fragments like B 126 and B 76. But Graham’s thesis cannot make sense of the opposing values of seawater in B 61 or of the opposing forces of the bow and the lyre in B 51. These fragments cite opposites that have nothing to do with transformation. Graham doesn’t force a false thesis on Heraclitus, but as with Osborne, the interpretation does not account for the variety of opposites. Heraclitus’ opposites clearly have something to do with his explanation of the cosmos. I have argued that there are multiple philosophical insights concerning opposites in the world. The insights do not seem to be reducible to the single thesis ‘opposites are essentially connected’. So, we are left with two problems: how exactly do Heraclitus’ opposites feature in Heraclitus’ explanation of the cosmos? And what unifies Heraclitus’ interest in opposites? To answer these questions, I believe it is helpful to examine how opposites were used in philosophical explanations prior to Heraclitus.

3.  Opposites in Heraclitus’ predecessors Heraclitus was preceded by two other Ionian thinkers: Anaximander and Anaximenes. Both thinkers appealed to opposites in their explanations of the cosmos. In what follows, I present a sketch of 34 Graham, Explaining the Cosmos, 129.



Opposites and Explanations in Heraclitus

13

Anaximander’s and Anaximenes’ ideas, which I take to be, for the most part, uncontroversial. According to Anaximander, there is something called the apeiron, which seems to be a sort of primordial entity from which all things are generated. Aristotle claims that the apeiron is Anaximander’s arche — ̄ his starting point of metaphysical ex­plan­ation.35 Anaximander seems to have constructed a cosmogony based on the apeiron: ϕησὶ δὲ τὸ ἐκ τοῦ ἀιδίου γόνιμον θερμοῦ τε καὶ ψυχροῦ κατὰ τὴν γένεσιν τοῦδε τοῦ κόσμου ἀποκριθῆναι καί τινα ἐκ τούτου ϕλογὸς σϕαῖραν περιϕυῆναι τῶι περὶ τὴν γῆν ἀέρι ὡς τῶι δένδρωι ϕλοιόν. (12 A 10. 7–10 DK) [Anaximander] says that the part of the everlasting [i.e. the apeiron] which is generative [γόνιμον] of hot and cold separated off [ἀποκριθῆναι] at the coming to be of the world order and from this [ἐκ τούτου] a sort of sphere of flame grew around the air about the earth like bark around a tree.36

It is hard to know with certainty, but it seems that this bit of testimonia is evidence that there are two stages of cosmic generation. First, a ‘generative part’ (γόνιμον) of the apeiron separates off from the apeiron itself. Second, this mysterious part of the apeiron produces physical contrarieties (hot and cold) as well as physical stuffs (fire, air, and earth).37 It seems that Anaximander’s entire cosmos is explained by these physical stuffs and their opposing properties. From Theophrastus (quoted by Simplicius), we learn that from the apeiron things come to be, but also that it is into the apeiron that things perish. Consider Anaximander B 1, which is embedded in a paraphrase from Simplicius: ἐξ ὧν δὲ ἡ γένεσίς ἐστι τοῖς οὖσι, καὶ τὴν ϕθορὰν εἰς ταῦτα γίνεσθαι κατὰ τὸ χρεών· διδόναι γὰρ αὐτὰ δίκην καὶ τίσιν ἀλλήλοις τῆς ἀδικίας κατὰ τὴν τοῦ χρόνου τάξιν, ποιητικωτέροις οὕτως ὀνόμασιν αὐτὰ λέγων. (12 B 1 DK) 35  By ‘metaphysical explanation’ I mean an explanation for why things are thus and so. Note that metaphysical explanations do not preclude anything physical. See Arist., Phys. 3. 4, 203b6–28. Simplicius claims that Anaximander is the first to introduce the term apeiron as an archē (In Phys. 24. 13–16 Diels). We don’t know for certain whether Anaximander himself used the term archē, but Aristotle is right to point out that the apeiron is a metaphysical principle in the sense that it is the metaphysical ground for everything. 36  A 10 DK is a testimony from pseudo-­Plutarch, Strom. 2. Though it is, on the whole, clearly a testimony and not a verbatim quotation, Graham, Texts, thinks the words italicized are verbatim quotations. 37  Simplicius tells us that ‘[Anaximander’s] contrarieties are hot, cold, dry, moist, and the others’ (ἐναντιότητες δέ εἰσι θερμὸν ψυχρὸν ξηρὸν ὑγρὸν καὶ τὰ ἄλλα, In Phys. 150. 24–5 Diels). It isn’t clear here what Simplicius means by ‘the others’.

14

Richard Neels

From what things existing objects come to be, into them too does their destruction take place, according to necessity: for they give recompense and pay restitution to each other for their injustice according to the ordering of time, expressing it in these rather poetic terms.38

According to the standard interpretation of Anaximander, the most basic entities are the elemental powers, which participate in a system of retaliation with one another, such that if one opposite becomes too powerful, it is overtaken by another opposite. According to Aëtius, Anaximander thought that meteorological events are consequences of wind (A 23). But we also learn that wind is accounted for by elemental powers in strife: ‘Anaximander [says] wind is a rush of air when the most fine and moist parts of it are moved or dissolved by the sun’ (Ἀναξίμανδρος ἄνεμον εἶναι ῥύσιν ἀέρος τῶν λεπτοτάτων ἐν αὐτῷ καὶ ὑγροτάτων ὑπὸ τοῦ ἡλίου κινουμένων ἢ τηκομένων, A 24). Presumably the sun has the power to heat and thereby dry out the moisture in the air, causing the air to move and become wind. Wind accounts for the meteorological phenomena, but the striving of opposing powers accounts for wind. If this is correct, Anaximander seems to have treated the opposites as fundamental explanantia with reference to the rest of the physical world.39 Graham writes: ‘In Anaximander we see a closed system of explanation, in which a set of items, apparently including elem­ental stuffs and their contrary properties, accounts for all the phenomena of experience.’40 Since the opposites seem to pose the ultimate physical explanation of the physical stuffs and their operations, I believe it makes most sense to suppose that for Anaximander the war of opposites is the most basic principle of explanation for this world. The apeiron

38 The italicized words are generally thought by scholars to be original to Anaximander. 39  C. Kahn supports this view (Anaximander and the Origins of Greek Cosmology [Anaximander] (New York, 1960), 178). Vlastos too seems to suggest this when he speaks of ‘the opposites which constitute this world’ (‘Equality and Justice in Early Greek Cosmologies’ [‘Equality and Justice’], Classical Philology, 42 (1947), 156–78 at 169). G.  Freudenthal claims that the statement ‘the basic constituents of Anaximander’s world are equal opposite powers’ is ‘uncontroversial’ (‘The Theory of Opposites in an Ordered Universe: Physics and Metaphysics in Anaximander’ [‘Theory of Opposites’], Phronesis, 31 (1986), 197–228 at 198). Graham, on the other hand, argues against this pure power ontology (Explaining the Cosmos, 42). But he still promotes, as do I, the interpretation that the powers are the ex­plana­tor­ ily basic factors in Anaximander’s thought. 40 Graham, Explaining the Cosmos, 42.



Opposites and Explanations in Heraclitus

15

may explain why the opposites and their stuffs exist at all,41 but it is the opposites that provide the ultimate explanation why stuffs are the way they are and behave the way they do.42 I turn now to Anaximenes, who is generally thought to be one of Anaximander’s rough contemporaries.43 According to the testimonia, Anaximenes provides an explanation of the physical world by means of two opposing processes, rarefaction and condensation: διαϕέρειν δὲ μανότητι καὶ πυκνότητι κατὰ τὰς οὐσίας. καὶ ἀραιούμενον μὲν πῦρ γίνεσθαι, πυκνούμενον δὲ ἄνεμον, εἶτα νέϕος, ἔτι δὲ μᾶλλον ὕδωρ, εἶτα γῆν, εἶτα λίθους, τὰ δὲ ἄλλα ἐκ τούτων. (13 A 5. 3–6 DK) [Air] differs in essence in accordance with its rarity (μανότητι) and density (πυκνότητι). When it is thinned (ἀραιούμενον), it becomes fire, while when it is condensed, it becomes wind, then cloud, when still more condensed, it becomes water, then earth, then stones. Everything else comes from these.44

Here, as with Anaximander, Anaximenes in a way treats opposites as the fundamental explanantia of the physical world. It is true that Aristotle claims Anaximenes’ arche ̄ is air, but the real explanatory power seems to come from the opposing processes that account for the differences in essence of the various stuffs into which air can transform.45 This is not to say that the opposites are of the same sort for Anaximenes and Anaximander or that they account for the physical world in the same way. For Anaximander, it seems opposing powers explain the physical world, while for Anaximenes opposing processes explain the physical world. But it does seem that both attempted to offer an explanation of the cosmos by appealing to fundamental opposites. To put it simply, for both Anaximander and Anaximenes, stuffs, things, and events in the physical world are explananda and certain opposites serve as their explanantia.

41 This interpretation of the apeiron is supported by Vlastos, ‘Equality and Justice’, but contested by Freudenthal, ‘Theory of Opposites’. 42  Some may think that cosmic justice is a fundamental, explanatory principle for Anaximander. However, it seems to me that Anaximander’s opposites are self-­ regulating as a matter of necessity. Cosmic justice simply is the self-­regulating war of opposites. 43 Diogenes Laertius (2.3) and Simplicius (In Phys. 24. 26 Diels) claim that Anaximenes was the student of Anaximander. But these later reporters loved to arrange all historical philosophers into long chains of teachers and students, often without any real evidence. 44  This testimony is provided by Simpl., In Phys. 24. 26–25. 1 Diels. 45 Arist., Metaph. Α. 3, 984a5–7.

16

Richard Neels

Anaximander and Anaximenes seem to be concerned with explanatory fundamentality.46 Their respective opposites appear to be fundamental explanantia for the rest of the physical world. As  such, we can classify their views in line with Metaphysical Foundationalism. According to foundationalists, the entire cosmos is explained by one or more metaphysically foundational (i.e. absolutely fundamental) entities.47 Explanations flow in one direction from the foundational entity/entities through the varying levels of

46  In the terms of modern metaphysics, we might say they are concerned with grounding. Grounding relations are generally thought to be ‘in virtue of’ relations, such that when y holds in virtue of x, x grounds y. If x grounds y, then x is more fundamental than y and x explains y. While these early thinkers are clearly not concerned with the finer points of grounding relations being debated today, their systems of explanation seem to follow the general pattern of grounding as expressed above. See F. Correia and B. Schnieder (eds.), Metaphysical Grounding: Understanding the Structure of Reality [Metaphysical Grounding] (Cambridge, 2012) for a recent anthology on ground. There is a current debate concerning the precise nature of ground; for a sample of this debate, see G. Rosen, ‘Metaphysical Dependence: Grounding and Reduction’, in B. Hale and A. Hoffman (eds.), Modality: Metaphysics, Logic, and Epistemology (Oxford, 2010), 347–83; C. S. Jenkins, ‘Is Metaphysical Grounding Irreflexive?’, Monist, 94 (2011), 267–76; J.  Schaffer, ‘Grounding, Transitivity, and Contrastivity’, in Correia and Schnieder (eds.), Metaphysical Grounding, 122–38. K. Fine, ‘Guide to Ground’ [‘Guide’], in Correia and Schnieder (eds.), Metaphysical Grounding, 37–80; K.  Koslicki, ‘Varieties of Ontological Dependence’, in Correia and Schnieder (eds.), Metaphysical Grounding, 186–213; M. J. Raven, ‘Ground’, Philosophy Compass, 10 (2015), 322–33; J. Zylstra, ‘The Essence of Grounding’[‘Essence’], Synthese, 196 (2019), 5137–52. The notion of grounding I believe to be at work in Anaximander is endorsed by Raven, ‘Ground’. According to Raven, ‘Ground is . . . supposed to serve a certain job description: it is the common factor in diverse in virtue of questions, the structuring relation in the project of explaining how some phenomena are “built” from more fundamental phenomena, and a key part of a venerable tradition concerned with metaphysical explanation’ (324, emphasis original). He also explains that ‘ground is metaphysical because it concerns the phenomena in the world itself, but also explanatory because it concerns how some phenomena hold in virtue of others’ (326, emphasis original). We might worry that it is anachronistic to interpret these thinkers as being concerned with grounding. However, the idea is not that they are consciously employing a concept of grounding, but merely that they seem to have had some basic notion of ontological and ex­plana­ tory fundamentality, and that this notion, untheorized among these early Greek philosophers, is the same one contemporary metaphysicians interested in grounding investigate. 47  Raven, ‘Ground’, summarizes foundationalism: ‘if explanations must begin, then so too any grounded fact must ultimately be grounded in facts which themselves are ungrounded . . . [this] entails that this ordering terminates in minimal elem­ents, like an explanatory chain beginning from unexplained explainers’ (327). These unexplained explainers are foundational entities.



Opposites and Explanations in Heraclitus

17

fundamental entities to the derivative entities.48 Anaximander’s metaphysical foundation or arche ̄ is the apeiron, and Anaximenes’ is air insofar as these pose the existential conditions for the cosmos. These, in turn, ground the existence of the opposites, and, in the words of Simplicius, ‘everything else comes from these’ (τὰ δὲ ἄλλα ἐκ τούτων). The opposites are existentially derivative from the foundation, but the opposites are fundamental to the rest of the physical world (observable stuffs, things, and events). In short, their re­spect­ ive opposites do all the explanatory work with respect to the nature of the physical world. In sum, Anaximander and Anaximenes accepted two principles concerning opposites and explanations: (1) the physical stuffs, things, and events in the world are explananda, and certain opposites are their explanantia, and (2) these opposites are explanatorily fundamental in the physical world, and everything else in the physical world is explained with reference to them.49

4.  Opposites as explananda When we turn to Heraclitus, we see a change in how opposites are featured, but it is difficult to see clearly what that change amounts to.50 In Anaximander and Anaximenes, the featured opposites are 48  A contemporary expression of this principle is called a Strict Partial Order: explanatory relations are irreflexive, transitive, and asymmetric. See M. J. Raven, ‘Is Ground a Strict Partial Order?’, American Philosophical Quarterly, 50 (2013), 193–201. Anaximander’s and Anaximenes’ explanations do appear to follow a Strict Partial Order. According to this view, chains of grounding obtain. In the grounding chain, X → Y → Z, X is foundational and non-­derivative; Y and Z are derivative, though Z is more derivative than Y; and Y is fundamental with respect to Z, but derivative with respect to X. 49  This isn’t to say that Anaximander and Anaximenes don’t have their differences with respect to their schemes of explanation, but I do think it is safe to say that they held these principles in common. It’s worth pointing out that this in­ter­ pret­ation fits nicely with Aristotle’s argument that the ‘physicists’ used contraries to explain change in the cosmos (Phys. 1. 5, 188a27–31). 50 Lloyd, Polarity and Analogy, tries to make sense of Heraclitus’ new approach within the tradition of appealing to opposites: ‘while Heraclitus’ theory was exceptional in that he particularly emphasised the interdependence or “unity” of op­pos­ ites, it was typical in so far as he too analysed the data of experience generally into pairs of opposites’ (17). However, Lloyd does not discuss exactly how and why Heraclitus analysed the data of experience into opposites. Furthermore, it seems Heraclitus treated the pairs of opposites as data of experience rather than analysing the data of experience by means of opposites (esp. B 126: ‘cold things warm up’ (ψυχρὰ θέρεται)).

18

Richard Neels

few and of the same order. For Heraclitus, the opposites are numerous and varied. This discrepancy suggests an implicit criticism on the part of Heraclitus, but it is difficult to see what the criticism might be.51 Many scholars have suggested that Heraclitus’ B 80 is an implicit criticism of Anaximander, and so it is worth examining this line of thought:52 εἰδέναι δὲ χρὴ τὸν πόλεμον ἐόντα ξυνόν, καὶ δίκην ἔριν, καὶ γινόμενα πάντα κατ᾽ ἔριν καὶ χρεών. (22 B 80 DK) It is necessary to know that war is common, and strife is justice, and all things come to be in accordance with strife and necessity.

While Anaximander in B 1 (quoted in Section 3 above) says that the opposites pay for their injustice, Heraclitus claims that the striving of opposites is justice. Some interpreters see this as evidence for the single unity of opposites thesis in Heraclitus: op­pos­ites are essentially connected.53 But if this is a criticism of Anaximander, surely Heraclitus’ point is merely semantic and not substantial, since both Anaximander and Heraclitus seem to be claiming that a war or strife of some sort is the proper course of the cosmos.54 Hence it does not seem that B 80 really is a criticism of Anaximander after all; it even seems to be a point of agreement. To be clear, there are differences. For Anaximander, this strife has a beginning (insofar as he thinks there is a beginning to the cosmos). For Heraclitus, strife is eternal. Furthermore, Anaximander’s retaliation of opposites 51  According to Graham, Explaining the Cosmos, Heraclitus contributes to the development of early Ionian natural philosophy by solving the problem of primacy in what Graham calls the Generating Substance Theory (see esp. chs 4 and 5). Graham’s view helps to situate Heraclitus in the development of Ionian science. However, it does not fully explain why Heraclitus used opposites in the way he did. 52  See Kirk, Cosmic Fragments, 240; Vlastos, ‘On Heraclitus’, 358; Mourelatos, ‘Naïve Metaphysics’, 35; Kahn, Art and Thought, 206. 53  Mourelatos argues that, for Anaximander, the ‘opposites are essentially incompatible’, but for Heraclitus, ‘they are one, they are internally or conceptually related by being opposed determinations within a single field’ (‘Naïve Metaphysics’, 35). But Anaximander’s opposites would also have to be ‘opposed determinations within a single field’ by the same logic. 54  Kahn writes, ‘The elements feed one another by their own destruction, since what is life to one is death for its reciprocal. The first law of nature is a lex talionis: life for life’ (Anaximander, 183). There is a debate concerning whether or not, for Anaximander, the opposites govern themselves (Kahn, Anaximander, 167–8; Vlastos, ‘Equality and Justice’, 156–8) or whether the opposites require the apeiron to govern them (Freudenthal, ‘Theory of Opposites’, 208).



Opposites and Explanations in Heraclitus

19

is just, but not all strife is just on his account. For Heraclitus, strife as such, seems to be just. So, what is Heraclitus’ implicit criticism of the way opposites are featured in the intellectual milieu of his day? Here is my suggestion: while Heraclitus’ predecessors explained the physical world by means of a simple set of fundamental op­pos­ ites, Heraclitus proliferates examples of many different pairs of opposites to make the point that we cannot so simply explain the world by reducing our explanantia to a pair (or limited set of pairs) of opposites.55 Perhaps we could even read Heraclitus’ numerous and varied instances of opposites as an argument of indifference against his predecessors.56 An indifference argument uses a lack of reason to establish a definite conclusion. Such arguments are not foreign to Presocratic thinkers before and after Heraclitus.57 Heraclitus’ implicit criticism of Anaximenes could plausibly be construed thus: ‘Why pick rarefaction–­condensation as the fundamental explanatory pair of opposites when there are a number of other opposites?’ Heraclitus cites numerous pairs of opposites that could rival Anaximenes’ opposites as contenders for fundamental explanantia: wet–dry, war–peace, life–death. It stands to reason that we could just as plausibly construe a fundamental principle out of each of these as we could out of Anaximander’s or Anaximenes’ opposites. At the same time, it is implausible to think that all op­pos­ites could equally be fundamental. For simplicity’s sake, just consider Anaximander’s and Anaximenes’s opposites: hot–­cold, rare–­dense. Both sets are equal contenders to be fundamental 55  It is possible that Parmenides (presumably writing after Heraclitus, although we don’t know for certain) made a similar point about the use of opposites in his predecessors. See Parmenides, 28 B 8. 53–61 DK. 56  I would like to thank Victor Caston for this suggestion. I appeal to the work of S. Makin, Indifference Arguments (Oxford, 1993). The main form of an indifference argument is this: (1) There is no more reason for p than there is for q; (2) either both p and q are true or neither p nor q is true. Independent argumentation is required to determine which disjunct in the initial conclusion is warranted. 57  It is thought that Anaximander appealed to an indifference argument when he reasoned that the earth is stable and at the centre of the universe, since it has no more reason to be at one side than it does to be at the other (Anaximander, A 26 DK). See Makin, Indifference Arguments, 101–5 for an analysis of Anaximander’s indifference argument. Democritus, writing after Heraclitus, employed indifference arguments to conclude that atoms are partless and that there exists an infinite var­ iety of atoms. See Makin, Indifference Arguments, 49–84 for an extensive analysis of indifference arguments in Democritus.

20

Richard Neels

explanantia. But both can’t be fundamental. Either rare–­dense explains hot–­cold (as Anaximenes thinks) or hot–­cold explains rare–­ dense (as Anaximander thinks).58 Neither individually is right (according to Heraclitus); and both can’t be right. And so, neither set of opposites can serve as the fundamental explanans of the cosmos.59 His further point, I think, is that opposites require an ex­plan­ ation as much as the physical world does. Indeed, we experience them as evident in the physical world (e.g. cold things warm up), and it stands to reason that we need an explanation why they behave the way they do. Thus, we cannot merely treat them as explanantia; we must also treat them as explananda. I reconstruct the argument as follows: 1. There are many sets of opposites of a varied sort that are equal contenders to be taken as fundamental explanantia.60 2. We have no more reason to choose one set of opposites (e.g. hot–­cold) over other equal contenders as a fundamental explanans. 3. Either these sets of opposites are all fundamental explanantia or none is. 4. It is implausible that all these sets of opposites are fundamental explanantia. 5. Hence, none of these sets of opposites is a fundamental explanans. 6. Everything is either a fundamental explanans or an explanandum. 7. Therefore, all sets of opposites are explananda.61

58  At least, this is true on a foundationalist framework for explanation. It need not be true on a coherentist framework. 59  This isn’t to say that they cannot be explanantia at all, just that they cannot be fundamental explanantia. 60  Note that my formulation of the argument does not rest on the claim that all sets of opposites for Heraclitus are equal contenders for fundamental explanantia. It is simply the case that at least some sets of opposites are equal contenders to be fundamental explanantia. And this is enough to secure the argument. I do not spell out which sets these are, nor need I for the argument to work. 61  One may wonder why I slide from ‘these sets’ in the premises to ‘all sets’ in the conclusion. The answer is simple: in the initial premises, we are considering sets of opposites that are plausible contenders as fundamental explanantia. All remaining sets (e.g. possibly up–­down) are not fundamental explanantia by default. So, the move from ‘these sets’ to ‘all sets’ is warranted by this default exclusion.



Opposites and Explanations in Heraclitus

21

Note that being an explanans is compatible with being an explanandum—­something can explain something else while requiring an explanation itself.62 But being a fundamental explanans is incompatible with being an explanandum—­fundamental explanantia are unexplained explainers. Of course, Heraclitus did not formulate this argument. However, I do think it is plausible that the conclusion is Heraclitus’ and that the argument captures his ­reasons for holding that conclusion. Two things strike me as true: (1) Anaximander and Anaximenes did choose a limited set of op­pos­ ition­al pairs to explain the entire cosmos; and (2) they are wrong for arbitrarily choosing one set of opposites or several sets from a number of equally plausible pairs of opposites. I think Heraclitus was aware of (1) and saw the truth in (2). I submit that this reconstruction makes good sense of why Heraclitus cited so many varied pairs of opposites. It also makes good sense of why many of his pairs of opposites are presented in a puzzling fashion. If this is right, we have a partial answer to the question ‘How do opposites feature in Heraclitus’ explanation of the cosmos?’ Our initial answer is that they are not fundamental explanantia; they are explananda. I want to be clear. I do not think the negative argument exhausts Heraclitus’ interest in opposites. I will have more to say about Heraclitus’ positive interest in opposites in Section 7 below. Before I do so, it is worth pointing out that if I’m right and Heraclitus saw the opposites as explananda, and if I’m right that Heraclitus is interested in explaining the cosmos, then he must have thought there was an explanans for the opposites. But what is the explanans for the opposites now conceived as explananda?

5.  Against a metaphysical foundation When dealing with Heraclitus’ predecessors, I argued that they were foundationalists of a sort. Foundationalists hold the belief that everything in the cosmos is either directly or indirectly explained by a foundational entity or a set of foundational entities. Foundational 62  For example, a water molecule both explains the properties of blood plasma and is explained by its subatomic particles. The water molecule is both an explanans (with respect to blood plasma) and an explanandum (with respect to the subatomic particles).

22

Richard Neels

here means ‘absolutely fundamental’. What is regarded as foundational is something that explains other things but is not itself explained by anything else. Does Heraclitus have a foundational entity or set of foundational entities? For Heraclitus, there seem to be four potential candidates for a non-­cosmogonical, foundational principle of explanation for the cosmos: fire, logos, God, or the cosmos itself. However, I will argue that none of these candidates can satisfy the conditions necessary to be considered a foundational principle of explanation. Fire is cited in B 30 as something that might appear to be a foundational principle. Heraclitus identifies the cosmos with an ever-­ living fire, and many have interpreted this literally.63 According to this material monist view, Heraclitus thought all things are fire as substratum and the changes in the cosmos are mere accidental changes of this basic substance. If this is true, then Heraclitus’ theory of elemental change is reduced to the alterations of a more basic stuff: fire. But this view does not square well with Heraclitus’ theory of radical elemental change. We saw from B 76 that the elem­ents transform into one another such that the birth of one elem­ent is the death of another element. Importantly, fire is also part of this birth–­death cycle. Heraclitus says that ‘the death of fire is the birth of air’ (πυρὸς θάνατος ἀέρι γένεσις). If the fire of B 30 is ever-­living (ἀείζωον), it cannot be the same fire spoken of in B 76 of which death (θάνατος) is predicated. At the very least, I think it is safe to say that elemental fire is not what everything is made of and therefore not a metaphysical foundation in that sense. However, fire might be foundational in another sense. Consider fragment B 90: πυρός τε ἀνταμοιβὴ τὰ πάντα καὶ πῦρ ἁπάντων ὅκωσπερ χρυσοῦ χρήματα καὶ χρημάτων χρυσός. (22 B 90 DK) All things are an exchange for fire and fire for all things, as goods for gold and gold for goods. 63  Aristotle is famous for his material monist interpretation of Heraclitus and other Presocratics. See Metaph. A. 3, 983b6–18 and 984a5–11. Barnes, The Presocratic Philosophers, 45–8, has argued more recently that Heraclitus is a material monist who supposed that all things are actually fire in altered guise. See Graham, Explaining the Cosmos, 122–9 (esp. at 127), for an extended argument against the material monist view. See also H.  Cherniss, Aristotle’s Criticism of Presocratic Philosophy (Baltimore, 1935) and id., ‘The Characteristics and Effects of Presocratic Philosophy’, Journal of the History of Ideas, 12 (1951), 319–45.



Opposites and Explanations in Heraclitus

23

Heraclitus does seem to give a privileged status to fire. For this reason, some have argued that fire has metaphysical primacy over the other elements. Charles Kahn, for example, has argued that fire is the original element out of which all the elements are born and into which all the elements die.64 Indeed, fragment B 31 seems to indicate that the elemental transformations begin with fire and are called the ‘turnings of fire’ (πυρὸς τροπαί ). But metaphysically speaking, the elements still seem to exist in a closed cycle of trans­form­ ation. The essences of the other elements are no more ­dependent on the essence of fire than the essence of fire is ­dependent on the other elements. Furthermore, this again does not square well with B 30, which claims that the cosmos always was, is, and will be fire. For these reasons, I think it makes most sense to suppose that fire itself is not a metaphysical foundation for the cosmos; instead, it is a symbol for the cosmos as a dynamic, regulated entity.65 Instead of fire, some might claim that the logos is a foundational principle of explanation for Heraclitus.66 We saw that in fragment B 1 Heraclitus claims that the logos, whatever it is, holds forever (τοῦ δὲ λόγου τοῦδ᾿ ἐόντος αἰεί). He also claims in B 1 that all things come to be in accordance with the logos (γινομένων γὰρ πάντων κατὰ τὸν λόγον τόνδε). In B 2, he claims that the logos is common (τοῦ λόγου δ᾿ ἐόντος ξυνοῦ), even though most humans have a private understanding of the world. In B 50, Heraclitus says: ‘It is wise for them, having listened not to me, but to the logos [τοῦ λόγου ἀκούσαντας], to agree that all things are one’ (οὐκ ἐμοῦ, ἀλλὰ τοῦ λόγου ἀκούσαντας ὁμολογεῖν σοϕόν ἐστιν ἓν πάντα εἶναι, with Diels). Many scholars have interpreted Heraclitus’ logos as a cosmic law.67 The greatest evidence seems to come from the second occurrence in B 1, in which Heraclitus claims that all things come to be in accordance with the logos. There is a problem with interpreting Heraclitus’ logos as a 64 Kahn, Art and Thought, 132–8. 65  In taking fire as a symbol for organized, cosmic change, I follow Graham, who says that fire is ‘fundamental just by being symbolic of the constant change that the elements undergo’ (Explaining the Cosmos, 127). Interestingly, Kahn, Art and Thought, 136, also claims that Heraclitus chose fire as his primary element for meta­ phor­ic reasons. 66  Most recently R. McKirahan, Philosophy before Socrates, 2nd edn (Indianapolis, 2010), 133, and P. Curd, ‘The Divine and the Thinkable: Toward an Account of the Intelligible Cosmos’ [‘Intelligible’], Rhizomata, 1 (2013), 217–24 at 236–8. 67 Graham, Explaining the Cosmos, 143–4; McKirahan, Philosophy before Socrates, 136; Curd, ‘Intelligible’, 237.

24

Richard Neels

cosmic law: rendering the term in this manner would mean that Heraclitus’ use of the term logos bears no resemblance to the standard use of the term in and around the time Heraclitus was ­writing.68 In response to this problem some scholars have argued that the cosmic law interpretation forces a Stoic conception of logos back onto Heraclitus and that it is plausible to read logos as simply referring to what Heraclitus is saying.69 That’s what a logos was in Heraclitus’ day: something someone says (legein). However, this interpretation also has its problems; it seems im­plaus­ible that Heraclitus wasn’t using the term in a special sense, especially with the second usage of logos in B 1: it is implausible that Heraclitus thought all things came to be in accordance with his utterances. This interpretation also makes nonsense of B 50: ‘having listened not to me, but to the logos’. It is implausible that Heraclitus wanted us to listen not to him but to what he says. So, on the one hand, a cosmic law interpretation of logos seems to be too strong, but the deflationary reading seems to be too weak. Mark Johnstone has recently offered a moderate solution to this issue that to my mind provides the most plausible understanding of logos in Heraclitus. Johnstone examines the common usage of logos around the time Heraclitus was active. He finds that ‘while the term “logos” most commonly referred to something appearing in language, it was not merely anything that happened to be said. Rather, a logos was an organized presentation of things as being “thus and so”.’70 When things are presented in a certain way, they can be understood. On these grounds, Johnstone argues that ‘Heraclitus denotes by the term “logos” neither his own discourse nor a cosmic law, but rather the world’s orderly and intelligible (i.e. comprehensible, understandable) presentation of its nature to us throughout our lives.’71 Thus, the logos doesn’t belong to Heraclitus but to the cosmos itself. On this reading, the logos is simply the world’s presentation of its nature as orderly and intelligible, 68  See Johnstone, ‘Logos’, 2. There he argues: ‘Yet the single biggest problem with this “cosmic-­law” interpretation, as it might be called, is that it risks completely detaching Heraclitus’ employment of the word “logos” from any other attested use of it in and around his time.’ 69 Burnet, Early Greek Philosophy, 4th edn (London, 1930), 133 n. 1; West, Orient, 124–9; Barnes, The Presocratic Philosophers, 59. 70  Johnstone, ‘Logos’, 16. 71  Johnstone, ‘Logos’, 21.



Opposites and Explanations in Heraclitus

25

not the foundational explanatory principle of the cosmos itself. It is worth noting that if this interpretation is correct—­and I think it is—­ the question of fundamentality gets passed on to the cosmos. Is the cosmos itself the foundation of all things? We will examine this view shortly. But first it will be helpful to examine Heraclitus’ theology. Some have claimed that Heraclitus’ God is the foundational principle of explanation for the cosmos.72 We’ve already examined fragment B 30 which claims that the cosmos was not created by a god. But Heraclitus does think God exists and is somehow identified with the cosmos: ὁ θεὸς ἡμέρη εὐϕρόνη, χειμὼν θέρος, πόλεμος εἰρήνη, κόρος λιμός [τἀναντία ἅπαντα . . . ], ἀλλοιοῦται δὲ ὅκωσπερ , ὁκόταν συμμιγῆι θυώμασιν, ὀνομάζεται καθ᾽ ἡδονὴν ἑκάστου. (22 B 67 DK) God is day night, winter summer, war peace, satiety hunger [all the contraries . . . ], and he alters just as when it is mixed with spices, is named according to the aroma of each of them.73

Heraclitus here identifies God with several opposing pairs—­perhaps all the opposites if we accept Hippolytus’ interpolation.74 Heraclitus elsewhere refers to this divine being as ‘one’ (ἕν) and ‘wise’ (σοϕόν, B 32 and B 41).75 Incidentally, Heraclitus claims that all things are one: οὐκ ἐμοῦ, ἀλλὰ τοῦ λόγου ἀκούσαντας ὁμολογεῖν σοϕόν ἐστιν ἓν πάντα εἶναι. (22 B 50 DK, accepting Miller’s emendation) It is wise for them, having listened not to me, but to the logos, to agree that all things are one.76

72  Curd, ‘Intelligible’, suggests at 237 that the logos is divine and that the divine logos is foundational. 73 There is a lacuna in this fragment. Diels, Fragmente; Graham, Texts; Marcovich, Heraclitus; and Kirk, Cosmic Fragments, all fill it with ‘fire’. However, there is some evidence to suggest that the missing word could well be ‘oil’, since this was a common mixture with spices. See H. Frankel, ‘Heraclitus on God and the Phenomenal World’, Transactions and Proceedings of the American Philological Association, 69 (1938), 230–44. 74  Hippolytus, the supplier of this fragment, adds that God is all the contraries. 75  ‘The wise is one, it alone wishes and does not wish to be called by the name of Zeus’ (B 32: ἓν τὸ σοϕὸν, μοῦνον λέγεσθαι οὐκ ἐθέλει καὶ ἐθέλει Ζηνὸς ὄνομα). ‘Wisdom is one: knowing the thought/plan that steers all things through all’ (B 41: εἶναι γὰρ ἓν τὸ σοϕόν, ἐπίστασθαι γνώμην, ὁτέη ἐκυβέρνησε πάντα διὰ πάντων). 76  This fragment is preserved for us by Hippolytus (Haer. 9. 9. 1).

26

Richard Neels

Since Heraclitus’ God is not the creator of the cosmos (B 30), is somehow immanent in the cosmos (B 67), and is identified with the cosmos (B 67 and B 50, by way of B 32 and B 41), I submit it makes most sense to suppose that Heraclitus’ God simply is the cosmos: all things together as a whole, integrated, orderly, intelligible, and ever-­living entity. I cannot defend this view fully here, but I believe it makes the most sense of Heraclitus’ divine fragments. If I’m right, we are again pushed to consider whether the cosmos itself is explanatorily basic (i.e. foundational). To be explanatorily basic, the cosmos as an integrated whole must explain its parts and not the other way round. This is a question about fundamental mereology: are the parts prior to the whole or is the whole prior to the parts? The priority pluralist will argue that the parts explain the whole; the priority monist will argue that the whole explains the parts.77 Importantly, a foundationalist must come down on one side and one side only, since, according to their view, something must be foundational, and a foundation cannot be something that is itself explained by anything else. Does Heraclitus pick a side? Consider again fragment B 10 (quoted above): Collectives: wholes and not wholes; brought together, pulled apart; sung in unison, sung in conflict; both from all things one and from one all things.

This fragment is concerned with mereological priority—­or so I will argue.78 The first word (συλλάψιες) refers to things that are taken together: collective entities. Heraclitus then affirms that collective entities are both whole and not-­whole. There are three ways we might interpret the meaning of the opposing pairs in the middle of the fragment: (1) we might think it is a diachronic description of the cosmos: at one time it is pulled apart (a mereological sum); at another time it is brought together (a proper whole). But this interpretation is ill-­advised, as we have no other evidence of such an Empedoclean cycle in Heraclitus. Furthermore, it fails to make sense of Heraclitus’ use of the plural ‘collectives’ at the beginning of the fragment. (2) we might think that Heraclitus’ is using this 77  I borrow the terms ‘priority pluralism’ and ‘priority monism’ and their meanings from J. Schaffer, ‘Priority Monism’, Philosophical Review, 119 (2010), 31–76. See this work for a contemporary defense of Priority Monism. 78 Graham, Explaining the Cosmos, 143, also recognizes the mereological nature of this fragment.



Opposites and Explanations in Heraclitus

27

fragment to parse out different types of collectives: some are proper wholes; others are mereological sums. But this in­ter­pret­ation suffers, since the point in the first half would have very little to do with the final statement about the cosmos and its parts. Finally, (3) we might think that the pairs of opposites parse out the different aspects of part–­whole structures (i.e. collectives). The different aspects of such a structure are their parts (not-­whole, pulled apart, sung in conflict) and the integrated whole (whole, brought together, sung in unison). I adopt this aspectual reading of part–­whole structures in B 10, since it fares better than the other two in­ter­pret­ ations in that it can provide a unified interpretation of the ­fragment: the entire fragment is about wholes and their parts. The plural collectives at the beginning of the fragment lets us know that Heraclitus thinks there are whole–­part structures within the cosmos. The final words reveal that such wholes can also be the parts of the  cosmos and that the cosmos itself is a whole–­part structure. Heraclitus ends the fragment with the cosmic claim that ‘all things come from one and one comes from all things’. The ‘one’ must indicate the cosmos. The ‘all things’ must indicate the parts of the cosmos. If this is so, Heraclitus appears to be ambivalent with respect to mereological priority: the whole and its parts stand to one another in a relation of metaphysical interdependence. And since a Metaphysical Foundationalist cannot be ambivalent with respect to mereological priority, Heraclitus cannot be a Metaphysical Foundationalist. But let us consider the last part of B 10 more carefully: ‘both from all things one and from one all things’. If ‘one’ (ἕν) designates the cosmos and ‘all things’ (πάντα) its parts, can we make good sense of the odd and reciprocal use of ‘from’ (ἐκ)? In context, the Greek term ἐκ (‘from’) plausibly denotes either generation or composition.79 If Heraclitus means the former, he would be claiming that all things are generated from the cosmos and that the cosmos is generated from all things. But reading ‘from’ (ἐκ) as signifying generation seems to contradict B 30, which claims that the world is ever-­living and not created. If he intends the latter, he would be claiming that the cosmos is made out of all things and 79  There are other uses of ἐκ: place (the place from which) or time (thereafter, or a point in time at which something occurs). But this fragment does not seem ­concerned with time or place.

28

Richard Neels

that all things are made out of the cosmos. But reading ‘from’ (ἐκ) as composition seems to contain a logical difficulty: the cosmos cannot be composed of all things as matter if all things are composed of the cosmos as matter. Pages can be said to compose a book; but that same book cannot then be said to compose the pages. Composition is a non-­ reciprocal relation. A third option is possible. The ‘from’ (ἐκ) might suggest cosmogonical language. In Anaximander A 10 DK, it is claimed that ‘the part of the everlasting which is generative of hot and cold separated off at the coming to be of the world order and from this (ἐκ τούτου) a sort of sphere of flame grew’. The preposition ‘from’ (ἐκ), when used in a cosmogonical context, might signal a sort of cosmogonical movement from explanans to explanandum. Heraclitus’ point in B 10 might then be that ‘all things’ and ‘one’ are mutually explanatory, such that neither is more fundamental than the other. In other words, ‘all things’ hold in virtue of the ‘one’, and the ‘one’ holds in virtue of ‘all things’.80 Neither is explanatorily basic. Both the cosmos and its parts explain one another in a reciprocal fashion.81

80  This usage of ἐκ seems to be a plausible construal of the ‘in accordance with’ usage cited by LSJ, s.v. ἐκ III.7. In this instance, ἐκ seems to be citing the ground of information. Compare the following usages: ‘he purified the island of Delos in accordance with the oracles’ (Hdt. 1. 64, τὴν νῆσον Δῆλον καθήρας ἐκ τῶν λογίων); ‘he cancelled the time that was in accordance with the statute’ (Dem., Against Timocrates, 28.3, ἀνελὼν τὸν ἐκ τῶν νόμων χρόνον); ‘they struck [with their oars] the briny deep in accordance with the command’ (Aesch., Pers. 397, ἔπαισαν ἅλμην βρύχιον ἐκ κελεύματος). In all these cases ἐκ seems to be citing the origin of some information. In the Herodotus passage, Pisistratus cleanses the island of Delos having been told to do so by the oracles; Pisistratus understands x (to cleanse the island) from y (the oracle). In the Demosthenes passage, Demosthenes is describing a man who drafts a decree but alters the date; he cancels the date that was given or told by the statute. In this situation we understand x (the correct date) from y (the decree). In the Aeschylus passage the rowers strike oar because of the command; they understood x (to row) from y (the command). 81  It is worth pointing out that B 90, discussed earlier in this section, cor­rob­­ orates this conclusion. Heraclitus claims that fire is an exchange (ἀνταμοιβή) for all things and all things for fire. If I am right in thinking that fire is a metaphoric symbol for the cosmos itself (as evidenced by B 30), then B 90 is saying that the cosmos is exchangeable for all things and all things for the cosmos in the same way that goods and gold are exchangeable. What is the foundation for the value of goods and the value of gold? Their values are interdependent and arise from the exchange itself. The point of B 90 seems to be that there is a metaphysical currency whereby all things (i.e. the parts of the cosmos) and the cosmos itself are worth the same. This point resonates with the idea that the cosmos and its parts are explanatorily inter­depend­ent.



Opposites and Explanations in Heraclitus

29

Let’s explore the relation between the cosmos and its parts more carefully. If Heraclitus is claiming that the whole cosmos is explained by its parts and the parts are explained by the whole cosmos, does this commit him to symmetrical explanation whereby X and Y explain one another in exactly the same manner? It need not, at least not if he assumed the following: Part–­Whole Principle: Wholes and parts stand to one another as mutually, though not exhaustively, explanatory. Heraclitus’ point is rather intuitive. If you want to understand the cosmos, you’ll need to examine its parts. And if you want to understand the parts of the cosmos, you’ll need to examine the whole cosmos. But this isn’t the whole story. In B 1 Heraclitus claims to differentiate each part (διαιρέων ἕκαστον) of the cosmos in order to explain its nature. The success of explaining the nature of things by division entails the explanatory interdependence of the things being explained. Any part of the cosmos is what it is (in part) because of its surrounding parts. As such, I believe it makes good sense to think that Heraclitus adopted the following principle: Difference Principle: Any part of a whole is different from every other part and is what it is, in part, because of the other parts of that whole.82 By differentiating the parts (διαιρέων ἕκαστον) of the cosmos, we come to understand how things are explained by their surrounding ­entities.83 For example, we recognize that land and sea (parts of the 82  More evidence for the Difference Principle might be found in B 53 and B 80, where Heraclitus claims that ‘war [πόλεμος] is the father of all’ (B 53) and ‘all things happen in accordance with strife [ἔριν]’ (B 80). Some might think that, on the basis of these fragments, war/strife is a candidate for a foundational principle of ex­plan­ ation. But it seems to me that claiming that ‘war is the father of all’ just is to say that nothing in or about the cosmos is a foundational principle of explanation. Strife itself cannot be a foundational principle of explanation for Heraclitus since it has an opposite, peace, that helps explain what strife is. Instead, by saying that war is the father of all, I take Heraclitus to be saying that things in the cosmos are always pitted against one another in an explanatorily productive manner. For more on Heraclitus’ interest in differentiating, see B 7: ‘If all things became smoke, noses would distinguish them’ (εἰ πάντα τὰ ὄντα καπνὸς γένοιτο, ῥῖνες ἂν διαγνοῖεν). 83  I don’t mean to suggest that the Difference Principle is fully expressed in B 1. Rather, the Difference Principle is a plausible and meaningful interpretation of B 1, especially in light of Heraclitus’ lack of a clear and stable foundation for ex­plan­ ation. I should also make clear that I do not think that non-­identity entails explanatory

30

Richard Neels

cosmos) are distinct from one another. And we can begin to understand what they are by explicating the differences between them. For starters, the one is dry, stable, and fit for humans (but not fish) and the other is wet, flowing, and fit for fish (but not humans). This differentiation reveals to us something about the metaphysical character of land and sea. At least some of their intrinsic and extrinsic properties help to explain one another. As such, land and sea themselves are what they are (at least in part) because of one another. The point of the Difference Principle is this: in the absence of a metaphysical foundation, the different explanatory factors of a given whole all help to explain one another.84 By combining the Difference Principle and the Part–­Whole Principle, we can get a fuller picture of Heraclitus’ scheme of expla­ nation. If we want to understand any part of the cosmos fully, we will need to see how it is differentiated by its surroundings and how it fits into the nested part–­whole structures of the cosmos. An ultimate explanation of a Heraclitean entity will make reference to  the cosmos itself. By the Part–­Whole Principle, the cosmos is explained via its parts and an explanation of the parts will run back dependence. Instead, it seems that the explanatory dependence of parts is understandable by recognizing the difference between the parts. 84  One might worry that the Difference Principle overgenerates. For example, a slice of pizza and the number 3 are distinct, but not obviously explanatorily dependent on one another. (Thanks to Victor Caston for this objection.) However, if we assume, as is reasonable, that Heraclitus thought explanatory chains were transitive, this worry may dissipate. A slice of pizza is indeed not directly ex­plana­tor­ily dependent on the number 3. But a slice of pizza is explanatorily dependent on the number 1; after all, it is a single slice of pizza as opposed to two slices of pizza. And the number 1 and the number 3 are explanatorily interdependent just in the sense that all numbers can be considered inter-­explanatory. (This is an admittedly controversial claim. But I submit that it is a Heraclitean claim in light of the fact that he has an aversion to metaphysical foundations.) A slice of pizza may, therefore, be explanatorily dependent on the number 3 by transitivity through the number 1. Someone might object further that the number 3 is not explanatorily dependent on the slice of pizza. However, we might think that numbers as conceptual entities are dependent on worldly concreta and that a given slice of pizza is a part of the set of entities that constitute the concreta of the world. The number 3, therefore, may be explanatorily dependent on a slice of pizza by transitivity through the set of worldly concreta. There seems to be a principle of proximity here. The closer parts are, the more explanatory power they hold for each other. But there is nothing preventing the Difference Principle from radiating out via transitivity. I don’t mean to suggest Heraclitus thought precisely in this manner, but just that the Difference Principle can be a coherent principle of explanation especially if we reject explanatory ­foundations and allow for the transitivity of explanatory relations.



Opposites and Explanations in Heraclitus

31

to the cosmos itself. By the Difference Principle, the cosmos is what it is, in part, by being different from its parts (its designation is singular rather than plural; it is an integration of parts rather than a mere collection of parts). The parts of the cosmos are explananda for the cosmos as explanans. And the cosmos is an explanandum for the parts as explanantia. In this manner, Heraclitus takes the cosmos as an explanandum and an explanans. This sets Heraclitus apart in a striking fashion from his predecessors—­indeed, from any Presocratic thinker.

6.  Metaphysical interdependence If, as I’ve argued, Heraclitus thought that neither the cosmos nor its parts are fundamental but that they stand to one another as mutually (though not strictly symmetrically) explanatory, then Heraclitus seems to have subscribed to a principle of metaphysical interdependence: explanatory relations are reciprocal rather than unidirectional (i.e. from fundamental to derivative). Such a view has been labelled Metaphysical Coherentism in contemporary metaphysics.85 So far, we have conducted the examination of Heraclitus’ views on explanation at a cosmic level. It will prove helpful to explore some specifics. Importantly, we must be clear about exactly what the explanatory relata are for Heraclitus and how exactly they hang together. This will require us to investigate the parts of the cosmos. But what are its parts for Heraclitus? This is incredibly difficult to answer. However, I have suggested above that we treat as parts the things that Heraclitus discusses in his fragments: material elements (stuffs), middle-­sized objects (things), opposites ­(properties), meteorological phenomena (events), etc. In short, I think it is safest to treat the aspects of the cosmos—­the ones mentioned by Heracli­ tus—­as its parts. For now, I will restrict my current discussion to two aspects of the cosmos discussed by Heraclitus: (1) material elements and (2) middle-­sized objects. The ultimate goal, as will become clear, is to situate Heraclitus’ positive interest in opposites in this discussion of explanation.

85 N.  Thompson, ‘Metaphysical Interdependence’, in M.  Jago (ed.), Reality Making (Oxford, 2016), 38–55.

32

Richard Neels

Recall from Section 2 that Heraclitus held a theory of elements and that these transform into one another. I cannot give a comprehensive account of this theory here. The following discussion is based in part on the important scholarship done by Daniel Graham.86 Consider fragments B 31 and B 76: πυρὸς τροπαὶ· πρῶτον θάλασσα, θαλάσσης δὲ τὸ μὲν ἥμισυ γῆ, τὸ δὲ ἥμισυ πρηστήρ . . .  θάλασσα διαχέεται, καὶ μετρέεται εἰς τὸν αὐτὸν λόγον, ὁκοῖος πρόσθεν ἦν ἢ γενέσθαι γῆ. (22 B 31. 2–3 and 9–10 DK) The turnings of fire: first sea, and of sea half is earth, half fire-­wind . . .  is liquefied as sea and measured into the same proportion it had before it became earth. πυρὸς θάνατος ἀέρι γένεσις, καὶ ἀέρος θάνατος ὕδατι γένεσις. γῆς θάνατος ὕδωρ γενέσθαι καὶ ὕδατος θάνατος ἀέρα γενέσθαι καὶ ἀέρος πῦρ καὶ ἔμπαλιν. (22 B 76 b–­c DK, also quoted above) The death of fire is the birth of air, and the death of air the birth of water. It is death for earth to become water, and death for water to become air, and death for air to become fire and contrariwise.

Whether we think that Heraclitus held that there were three or four elements, he clearly thought that any element was transformable into any other.87 First, we should note that if this is so, then no one element is foundational with respect to the rest. If we were to explain earth, we might begin by citing where earth comes from: air. In

86 Graham, Explaining the Cosmos, 122–9. I explore a slightly different view in line with Graham’s in Neels, ‘Elements and Opposites in Heraclitus’, Apeiron, 51 (2018), 427–52. The clearest differences between my view and Graham’s view are these: (1) I take Heraclitus to have a four-­element theory, while Graham takes Heraclitus to have a three-­element theory; and (2) I take Heraclitus’ elements to be associated closely with a single property each, while Graham seems to think that they have a vague range of properties. On this last point, see also Betegh, ‘Physical Aspect’, 23, who claims that we should not interpret Heraclitus as thinking that the elements had a rigid set of properties. 87  There is some discussion as to whether or not Heraclitus held a three-­element theory (Earth–­Fire–­Water) or a four-­element theory (Earth–­Air–­Fire–­Water). The debate turns somewhat on the authenticity of B 76. However, recent work done on Heraclitus’ notion of soul may suggest evidence beyond B 76 that Heraclitus did think that air was an elemental stuff. See A. Laks, ‘How Preplatonic Worlds Became Ensouled’, Oxford Studies in Ancient Philosophy, 55 (2018), 1–34 at 14–23. Laks argues that Heraclitus’ soul is air and that it is breathed in from around us. If Laks’s argument is successful, we probably have good reason to think air is a genuine Heraclitean element. We should note that Laks’s argument rests heavily on doxo­ graphic evidence as opposed to verbatim fragments. Still, the evidence does bear on the question of the authenticity of Heraclitean elemental air.



Opposites and Explanations in Heraclitus

33

part, this is true: air serves as the partial explanans for earth.88 It is only partial, since the generating source of earth does not fully explain what it is to be earth. Furthermore, we would do well to treat air as an explanandum in turn, and it quickly becomes apparent that we will end up in a circle, back to earth. Does earth explain earth? Not obviously. But a partial explanation of earth has been gained from this exercise: earth is ontologically dependent on all the other elements insofar as it depends for its existence on all the other elements. But all the elements are ontologically dependent on one another and so they comprise a system of elements, that is, a network composed of interdependent, dynamic parts in flux. The elements also seem to be conceptually interdependent amongst themselves. What each element is is distinct from what every other element is. For, what-­it-­is-­to-­be-­water is different from what-­it-­is-­ to-­be-­fire (Difference Principle). When one attempts to explain a given element, one can do so by referencing its powers: presumably fire has the power to heat (among others) while water has the power to quench (among others). Heraclitus certainly must have thought that the elements had powers in some way distinct from one another or they would not be distinct elem­ents. When one attempts to explain what Heraclitean earth is, one must explain how earth’s function differs from those of the other material elements. In this way, all the elements are conceptually interdependent, that is, what they are (their essence) is dependent, at least in part, on the whatness of all the other elements. Combining this result with the result from the previous paragraph, we can conclude that for Heraclitus the material elements are each ontologically and conceptually dependent on one another. And if that is true, they explain one another in a reciprocal fashion. But I don’t mean to suggest that the elements are wholly explained by one another. According to the Part–­Whole Principle, a full ex­plan­ation of an element must ul­tim­ ate­ly refer to its role in the cosmos (e.g. constituting middle-­sized objects). And as we’ve seen, an ex­plan­ation of the cosmos will push us to examine its parts once again.89 88  It is common for contemporary metaphysicians to make a distinction between x partially grounding y and x fully grounding y. See Fine, ‘Guide’; Raven, ‘Ground’; and Thompson, ‘Metaphysical Interdependence’. 89  We might worry that this introduction of explanatory loops may be unsatisfying as a metaphysical principle. A central worry may be this: it seems as though anything can only be partially (i.e. not fully) explained by this coherentist approach

34

Richard Neels

Next let’s examine what Heraclitus said about middle-­sized objects. I will restrict my discussion to one fragment already mentioned in Section 2: οὐ ξυνιᾶσιν ὅκως διαϕερόμενον ἑωυτῶι ὁμολογέει. παλίντροπος ἁρμονίη ὅκωσπερ τόξου καὶ λύρης. (22 B 51 DK) They do not understand how while differing with itself it agrees with itself: reciprocal structure as of a bow or a lyre.

For simplicity, let’s examine the bow. Heraclitus claims that the bow agrees with itself while disagreeing with itself. The stave and string are at odds with one another (i.e. they oppose one another), but the bow itself is what it is because of this tension. The stave and string are the parts of the bow. In a philosophically interesting way, the stave and string of a bow explain one another via the Difference Principle. To explain a bowstring, one must refer to the bow stave. A bowstring is the part of a bow that gets attached to the bow stave and pulls the bow stave inward. But to explain a bow stave, one must refer to the bowstring. A bow stave is the part of a bow to which the bowstring is attached and pulls the bowstring outward. Importantly, both stave and string are the partial cause of the tension in one another. As such, the parts of the bow explain one another in a non-­trivial way (Difference Principle). But things can only be partially explained by the Difference Principle, since the Difference Principle is blind to the full being of the higher-order entity of which the parts are mere parts. As such, the parts of the bow must refer to the whole bow itself for a more complete ex­plan­ ation via the Part–­Whole Principle. When we ask what a bow is, our explanation will require an analysis of its parts in tension, but it will also push us closer to the cosmic level: to fully understand the bow, we must understand the larger context in which it operates to explanation, since grounding chains never terminate. See R. Bliss, ‘Viciousness and Circles of Ground’, Metaphilosophy, 45 (2014), 245–56 for a contemporary argument defending the plausibility of grounding loops. See also Thompson, ‘Metaphysical Interdependence’. If readers are worried about the plausibility of Heraclitus’ view, they can consider a recent example cited by Zylstra: ‘The volume of a substance is the quotient of its mass (dividend) and density (divisor). The density of a substance is the quotient of its mass (dividend) and volume (divisor). Finally, the mass of a substance is the product of its density and volume’ (‘Essence’, 5147). Zylstra concludes that volume is grounded in mass and density, density is grounded in mass and volume, and mass is grounded in density and volume. This is a plausible case of a local grounding loop.



Opposites and Explanations in Heraclitus

35

(i.e. hunting and war). The Part–­Whole Principle will also require a push in the opposite direction: towards the ma­ter­ial constitution of the parts and the properties of the material elem­ents, since this explains the pre-­existing nature of the bowstring and bow stave such that they can compose a bow. This, I submit, is a very clear picture of Heraclitus’ theory of explanation, which I think can be summarized in two statements: (1) The parts of a thing partially explain one another (Difference Principle), and (2) wholes further explain their parts and parts further explain wholes (Part–­Whole Principle)—from the basic material constituents to the cosmos itself.

7.  Conclusion: Opposites and explanation We have been trying to understand the relation between Heraclitus’ interest in opposites and his scheme of explanation. We are now in a better position to answer this question—­indeed, the answer has already surfaced. We have just now seen that the opposites are caught in the middle of this interdependent web called the cosmos. Opposites are parts of the cosmos, not independently existing en­tities (indeed, nothing is independent if my interpretation is correct), but real aspects of the cosmos that need to be explained. What, then, explains the opposites? If the conclusion of Section 6 is true, then (a) the opposites partially explain one another (Difference Principle) and (b) the opposites are further explained with reference to their role in the cosmos (Part–­Whole Principle). But we also saw that the cosmos is explained by its parts, so, im­port­ant­ly, the cosmos is partially explained by the opposites. As such, the op­ pos­ ites, as with all the parts of the cosmos, are both (non-­ fundamental) explanantia and explananda for Heraclitus. B 51 told us that things differ from one another while agreeing with one another. Differing and agreeing are opposites; they are real aspects of the world. On the first analysis, differing and agreeing as opposing terms explain one another: they are metaphysically interdependent. One cannot explain what-­it-­is-­to-­differ without explaining what-­it-­is-­to-­agree and vice versa. This is an instance where a part of the cosmos partially explains another part and vice versa (Difference Principle). Differing and agreeing are further explained by their role in the cosmos—­constitution of bows and lyres, for example (Part–­Whole Principle towards the whole). Differing and

36

Richard Neels

agreeing are also explained by the material conditions in which they are realized—­here, the flexibility of the horn in the bow stave and the elasticity of the sinew of the bowstring (Part–­ Whole Principle towards the parts).90 In the case of a bow, a harmony (agreement) of opposing parts (difference) constitutes the nature of the bow. As such, the opposites, in part, explain what the bow is. So, these opposites, being caught in the web of inter­depend­ence, both partially explain the bow and are explained, in part, by the bow. Opposites explain things like the bow (whose name is life and whose work is death), but the opposed tensions and forces must themselves be understood, both how they are connected—­ the standard view has that much right—­and the ways in which they work together to do what they do (and explain what they produce). What I’ve said here about differing–­agreeing can be said about all the opposites in Heraclitus; the result is not a banal thesis, but a rich philosophical explanation of the cosmos and its constituents. We have seen how Heraclitus’ opposites factor into his ex­plan­ ation of the cosmos, but we might still wonder why Heraclitus was centrally interested in opposites. We could answer this by claiming that Heraclitus felt the need to respond to his predecessors. And I think that would be a satisfying answer. But I think the answer runs deeper. I mentioned in Section 1 that Heraclitus was aware that opposites are essentially connected (though this does not exhaust his interest in opposites). The essential connection of opposites entails that opposing pairs partially explain one another. In Heraclitus, opposing terms present the clearest case of the Difference Principle: what-­it-­is-­to-­be-­hot must refer to what-­it-­is-­to-­be-­cold and vice versa. So why was Heraclitus so interested in opposites? I submit it is because opposites are the clearest sign of metaphysical inter­ depend­ence evident in the cosmos. Metaphysical interdependence happens to be Heraclitus’ most basic principle of explanation for all things, including the cosmos. To cite B 51, the cosmos has a palintropic (i.e. reciprocal) structure (ἁρμονίη), just like a bow. For Heraclitus, the role of opposites in his explanation of the cosmos and its constituents is summarized in three statements: (1) opposites and oppositions cannot function as fundamental explanantia, contra his Ionian predecessors; (2) all explanatory 90 Hom., Il. 4. 105–26, tells us that Lycaon’s bow was made from the horn (κέρας) of an ibex and the string was made from ox tendon (νεῦρα βόεια).



Opposites and Explanations in Heraclitus

37

relata are complexly interdependent, just as opposing terms are simply interdependent; and (3) opposites are both explananda and (non-­fundamental) explanantia. The cosmos has a reciprocal structure (παλίντροπος ἁρμονίη) whereby all things are caught in a web of metaphysical interdependence. We can call this view Metaphysical Coherentism. Interestingly, the cosmos itself is implicated in this reciprocal structure, since it and its parts stand to one another in a relation of interdependence: all things from one and one from all things (B 10). Opposites for Heraclitus are not fundamental explanatory principles as thought by his Ionian predecessors, but opposition itself stands as a symbol of the ­metaphysical inter­depend­ence in and of the cosmos. Heraclitus rejects Metaphysical Foundationalism and embraces Metaphysical Coherentism. Oklahoma State University

BIBLIOGR A PH Y Barnes, J., The Presocratic Philosophers (London, 1982). Begley, K., ‘Heraclitus against the Naïve Paratactic of Mere Things’, Ancient Philosophy Today, 3 (2021), 74–97. Betegh, G., ‘On the Physical Aspect of Heraclitus’ Psychology’ [‘Physical Aspect’], Phronesis, 52 (2007), 3–32. Betegh, G. and Piano, V., ‘Column IV of the Derveni Papyrus: A New Analysis of the Text and the Quotation of Heraclitus’, in C. Vassallo (ed.), Presocratics and Papyrological Tradition (Berlin, 2019), 179–220. Bliss, R., ‘Viciousness and Circles of Ground’, Metaphilosophy, 45 (2014), 245–56. Burnet, J., Early Greek Philosophy, 4th edn (London, 1930). Cherniss, H., Aristotle’s Criticism of Presocratic Philosophy (Baltimore, 1935). Cherniss, H., ‘The Characteristics and Effects of Presocratic Philosophy’, Journal of the History of Ideas, 12 (1951), 319–45. Correia, F. and Schnieder, B. (eds.), Metaphysical Grounding: Understanding the Structure of Reality [Metaphysical Grounding] (Cambridge, 2012). Cunliffe, R. J., A Lexicon of the Homeric Dialect (Norman, 2012; 1st edn, 1924.) Curd, P., ‘The Divine and the Thinkable: Toward an Account of the Intelligible Cosmos’ [‘Intelligible’], Rhizomata, 1 (2013), 217–47. Curd, P., ‘Knowledge and Unity in Heraclitus’, Monist, 74 (1991), 531–49.

38

Richard Neels

Diels, H. and Kranz, W., Die Fragmente der Vorsokratiker, 6th edn [Fragmente] (Berlin, 1951). Dilcher, R., Studies in Heraclitus [Studies] (Hildersheim, 1995). Emlyn-Jones, C., ‘Heraclitus and the Identity of Opposites’, Phronesis, 21 (1976), 89–114. Fine, K., ‘Guide to Ground’[‘Guide’], in Correia and Schnieder (eds.), Metaphysical Grounding, 37–80. Frankel, H., ‘Heraclitus on God and the Phenomenal World’, Transactions and Proceedings of the American Philological Association, 69 (1938), 230–44. Freudenthal, G., ‘The Theory of Opposites in an Ordered Universe: Physics and Metaphysics in Anaximander’, Phronesis, 31 (1986), 197–228. Graham, D. W., Explaining the Cosmos: The Ionian Tradition of Scientific Philosophy [Explaining the Cosmos] (Princeton, 2006.) Graham, D.  W., ‘Heraclitus’ Criticism of Ionian Philosophy’, Oxford Studies in Ancient Philosophy, 15 (1997), 1–50. Graham, D.  W., The Texts of Early Greek Philosophy: The Complete Fragments and Selected Testimonies of the Major Presocratics [Texts] (Cambridge, 2010). Hankinson, R. J., Cause and Explanation in Ancient Greek Thought [Cause and Explanation] (Oxford, 1998). Hussey, E., ‘Epistemology and Meaning in Heraclitus’, in Schofield and Nussbaum (eds.), Language and Logos, 33–59. Jenkins, C. S., ‘Is Metaphysical Grounding Irreflexive?’, Monist, 94 (2011), 267–76. Johnstone, M., ‘On “Logos” in Heraclitus’ [‘Logos’], Oxford Studies in Ancient Philosophy, 47 (2014), 1–29. Kahn, C. H., Anaximander and the Origins of Greek Cosmology [Anaximander] (New York, 1960). Kahn, C.  H., ‘Anaximander’s Fragment: The Universe Governed by Law’, in A. P. D. Mourelatos (ed.), The Pre-Socratics: A Collection of Critical Essays (Princeton, 1993; originally published, 1974), 99–117. Kahn, C.  H., The Art and Thought of Heraclitus [Art and Thought] (Cambridge, 1979). Kirk, G.  S., Heraclitus: The Cosmic Fragments [Cosmic Fragments] (Cambridge, 1954). Kirk, G. S., Raven, J. E., and Schofield, M., The Presocratic Philosophers, 2nd edn (Cambridge, 1983). Koslicki, K., ‘Varieties of Ontological Dependence’, in Correia and Schnieder (eds.), Metaphysical Grounding, 186–213. Laks, A., ‘How Preplatonic Worlds Became Ensouled’, Oxford Studies in Ancient Philosophy, 55 (2018), 1–34. Liddell, H. G. and Scott, R., A Greek-English Lexicon, 9th edn, revised by H. Stuart Jones and R. McKenzie (Oxford, 1925–40).



Opposites and Explanations in Heraclitus

39

Lloyd, G. E. R., Polarity and Analogy (Cambridge, 1966). Long, A., ‘Wisdom in Heraclitus’, Oxford Studies in Ancient Philosophy, 33 (2007), 1–17. MacKenzie, M. M., ‘Heraclitus and the Art of Paradox’, Oxford Studies in Ancient Philosophy, 6 (1988), 1–37. Makin, S., Indifference Arguments (Oxford, 1993). Marcovich, M., Heraclitus: Greek Text with a Short Commentary, 2nd edn [Heraclitus] (Sankt Augustin, 2001). McKirahan, R., Philosophy before Socrates, 2nd edn (Indianapolis, 2010). Moravcsik, J., ‘Heraclitean Concepts and Explanations’, in K. Robb (ed.), Language and Thought in Early Greek Philosophy (La Salle, 1983), 134–52. Mourelatos, A. P. D., ‘Heraclitus, Parmenides and the Naïve Metaphysics of Things’ [‘Naïve Metaphysics’], in E. N. Lee, A. P. D. Mourelatos, and R.  M.  Rorty (eds.), Exegesis and Argument: Studies in Greek Philosophy Presented to Gregory Vlastos (Assen, 1973), 16–48. Mourelatos, A. P. D., The Route of Parmenides, revised and expanded edn (Las Vegas, 2008). Neels, R., ‘Elements and Opposites in Heraclitus’, Apeiron, 51 (2018), 427–52. Neels, R., ‘Heraclitus on the Nature of Goodness’, Ancient Philosophy, 41 (2021), 1–22. Neels, R., ‘Phusis, Opposites and Ontological Dependence in Heraclitus’ [‘Phusis’], History of Philosophy Quarterly, 35 (2018), 199–217. O’Brien, D., ‘Heraclitus on the Unity of Opposites’, in K. I. Boudouris (ed.) Ionian Philosophy (Athens, 1989), 298–303. Osborne, C., ‘Heraclitus’, in C. C. W. Taylor (ed.), Routledge History of Philosophy, vol. i: From the Beginning to Plato (London, 1997), 80–116. Papamichael-Paspalides, E., ‘The Concept of One in Heraclitus’, Revue de Philosophie Ancienne, 23 (2005), 41–54. Popper, K., ‘Back to the Presocratics’, Proceedings of the Aristotelian Society, 59 (1958), 1–24. Raven, M. J., ‘Ground’, Philosophy Compass, 10 (2015), 322–33. Raven, M. J., ‘Is Ground a Strict Partial Order?’, American Philosophical Quarterly, 50 (2013), 193–201. Robinson, T. M., Heraclitus: Fragments (Toronto, 1987). Rosen, G., ‘Metaphysical Dependence: Grounding and Reduction’, in B.  Hale and A.  Hoffman (eds.), Modality: Metaphysics, Logic, and Epistemology (Oxford, 2010), 347–83. Schaffer, J., ‘Grounding, Transitivity, and Contrastivity’, in Correia and Schnieder (eds.), Metaphysical Grounding, 122–38. Schaffer, J., ‘Priority Monism’, Philosophical Review, 119 (2010), 31–76. Schofield, M. and Nussbaum, M. (eds.), Language and Logos: Studies in Ancient Greek Philosophy Presented to G.  E.  L.  Owen [Language and Logos] (Cambridge, 1982).

40

Richard Neels

Stokes, M. C., One and Many in Presocratic Philosophy (Washington DC, 1971). Thompson, N., ‘Metaphysical Interdependence’, in M. Jago (ed.), Reality Making (Oxford, 2016), 38–55. van Fraassen, B., The Scientific Image (Oxford, 1980). Vlastos, G., ‘Equality and Justice in Early Greek Cosmologies’, Classical Philology, 42 (1947), 156–78. Vlastos, G., ‘On Heraclitus’, American Journal of Philology, 76 (1955), 337–68. Warren, J., Presocratics (Berkley, 2007). West, M. L., Early Greek Philosophy and the Orient [Orient] (Oxford, 1971). Wiggins, D. ‘Heraclitus’ Conception of Flux, Fire and Material Persistence’, in Schofield and Nussbaum (eds.), Language and Logos, 1–32. Zylstra, J., ‘The Essence of Grounding’ [‘Essence’], Synthese, 196 (2019), 5137–52.

EVALUATIVE ILLUSION IN PLATO’S PROTAGORAS suzanne obdrzalek 1. Introduction In the late 1960s and early 1970s, a group of scientists conducted a series of experiments which have since become known as the ‘Stanford Marshmallow Experiments’. A scientist would take a child between the ages of 3 and 5 into a room, sit them at a table, and place one delicious marshmallow chick in front of them. The scientist would then tell the child that they were going to leave the room and that the child had a choice. They could eat one marshmallow now. On the other hand, if they waited fifteen minutes until the scientist returned, they could get two marshmallows instead. In the end, only about a third of the children were able to resist. A few of the children ate the sweet immediately, but many held out for some time before capitulating to the lure of the marshmallow. Those who tried to resist would hide their heads in their arms, pound the floor with their feet, stare up at the ceiling, tug on their pigtails, or stroke the marshmallow as if it were a tiny stuffed animal.1 What we see here appears to be a classic case of akrasia. Ordinarily, we would explain it as follows: the children judged that it was best to resist the marshmallow that lay before them, but were unable to I am extremely grateful to Victor Caston for his invaluable, extensive comments on this paper, as well as to my anonymous referees. I am also indebted to Rachel Singpurwalla and to Nick Smith for their generous written comments; this paper would not have been possible without the springboard provided by their provocative engagement with these issues. Thanks also to my colleague, Piercarlo Valdesolo, for introducing me to relevant areas of psychological research; and to audiences at the West Coast Plato Workshop at the San Diego State University, the Ancient Philosophy and Science Network at Humboldt University, and the Munich School of Ancient Philosophy at Ludwig Maximilian University. 1 See  W.  Mischel and E.  B.  Ebbeson, ‘Attention in Delay of Gratification’ [‘Attention’], Journal of Personality and Social Psychology, 16 (1970), 329–37; W.  Mischel, E.  B.  Ebbesen, and A.  Raskoff Zeiss, ‘Cognitive and Attentional Mechanisms in Delay of Gratification’, Journal of Personality and Social Psychology, 21 (1972), 204–18; and W. Mischel, Y. Shoda, and M. Rodriguez, ‘Delay of Gratification in Children’, Science, 244 (1989), 933–8.

Suzanne Obdrzalek, Evaluative Illusion in Plato’s Protagoras In:Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Suzanne Obdrzalek 2023. DOI: 10.1093/oso/9780192885180.003.0002

42

Suzanne Obdrzalek

control their desire for it, and therefore gave into their desire and ate the marshmallow. One might even describe them as ‘overcome by pleasure’ and ‘unwilling to do what is best, even though they know what it is’.2 Readers familiar with the Protagoras will have noticed that, in describing the plight of these children, I quoted the account of akrasia that Socrates attributes to hoi polloi in that dialogue, an account that Socrates insists is false. He maintains that if the children knew or even believed that it was best not to eat the sweet in front of them, they would not do so. Their decision is, therefore, the fault of ignorance, not weakness. What does this ignorance consist in? Socrates gives us a very specific account: just as objects of a given size appear larger when close at hand than at a distance, so proximate pleasures appear greater than those that are far off in time. The children’s error is one of poor hedonic calculus: the imminence of the pleasure from eating the marshmallow in front of them causes them to falsely overestimate its magnitude and to conclude that they can maximize their pleasure by eating the one marshmallow now rather than the two marshmallows later. Interpreters of Plato’s Protagoras have devoted a great deal of attention to determining Socrates’ reasons for rejecting the common account of akrasia—­Socrates declares that it is ridiculous, but it is controversial what its ridiculousness consists in.3 Many also raise the question of whether his treatment of akrasia does justice to its phenomenology: can it account for the feeling of inner conflict that 2  Prot. 352 d 8–e 1: ὑπὸ ἡδονῆς ϕασιν ἡττωμένους. 352 d 6–7: γιγνώσκοντας τὰ βέλτιστα οὐκ ἐθέλειν πράττειν. I use the translations in J.  M.  Cooper (ed.), Plato: Complete Works (Indianapolis, 1997) throughout, with occasional modifications; Greek texts are from the Burnet OCTs. 3 For discussion, see e.g. A.  G.  Callard, ‘Ignorance and akrasia-Denial in the Protagoras’, Oxford Studies in Ancient Philosophy, 47 (2014), 31–80; J. Clark ‘The Strength of Knowledge in Plato’s Protagoras’, Ancient Philosophy, 32 (2012), 237–55; M.  Dyson, ‘Knowledge and Hedonism in Plato’s Protagoras’ [‘Knowledge and Hedonism’], Journal of Hellenic Studies, 96 (1976), 32–45; D. Gallop, ‘The Socratic Paradox in the Protagoras’, Phronesis, 9 (1964), 117–29; T. Irwin, Plato’s Ethics [Ethics] (Oxford, 1995), 83–4; R. Kamtekar, Plato’s Moral Psychology [Moral] (Oxford, 2017), 40–7; T. Penner, ‘Socrates on the Strength of Knowledge: Protagoras 351 b–357 e’ [‘Socrates on the Strength of Knowledge’], Archiv für Geschichte der Philosophie, 79 (1997), 117–49 at 126–7; G. Santas, ‘Plato’s Protagoras and Explanations of Weakness’, Philosophical Review, 75 (1966), 3–33; C. C. W. Taylor, Plato: Protagoras [Protagoras] (Oxford, 1991), 181–6; G. Vlastos, ‘Socrates on Acrasia’ [‘Acrasia’], Phoenix, 23 (1969), 71–88; and D.  Wolfsdorf, ‘The Ridiculousness of Being Overcome by Pleasure: Protagoras 352 b 1–358 d 4’ [‘Ridiculousness’], Oxford Studies in Ancient Philosophy, 31 (2006), 113–36.



Evaluative Illusion in Plato’s Protagoras

43

so often characterizes our experience of akrasia?4 In this paper, I intend to explore a different problem. I want to take a closer look at Socrates’ account of what actually occurs in supposed cases of akrasia, of the evaluative error he attributes to the would-­be akratic. His account is, in fact, puzzling. Consider again our children and their marshmallows. On the face of it, it seems quite extraordinary to propose that what is going on is that the children judge that eating one marshmallow now offers more pleasure than eating two later. Why, then, does Socrates make this claim?5 Getting to the bottom of this is crucial, since Socrates’ alternative explanation of the supposed phenomenon of akrasia stands at the heart of his denial of akrasia in the Protagoras. On the standard interpretation of the Protagoras, the error is simply calculative. Henceforth, I shall refer to this as the trad­ition­al in­ter­pret­ ation.6 Recently, several interpreters—­notably, Devereux, Brickhouse and Smith, and Singpurwalla—­have objected that this leaves too much unexplained.7 Why, exactly, should we perform this faulty 4  See e.g. D.  Devereux, ‘Socrates’ Kantian Conception of Virtue’ [‘Kantian Conception’], Journal of the History of Philosophy, 33 (1995), 381–408 at 389–96; T. Penner, ‘Plato and Davidson: Parts of the Soul and Weakness of Will’ [‘Plato and Davidson’], Canadian Journal of Philosophy, 16 (1990), 35–74 at 67–71; N. Reshotko, Socratic Virtue: Making the Best of the Neither-­ Good-­ Nor-­ Bad [Socratic Virtue] (Cambridge, 2006), 74–91; and R.  Singpurwalla, ‘Reasoning with the Irrational: Moral Psychology in the Protagoras’ [‘Reasoning with the Irrational’], Ancient Philosophy, 26 (2006), 243–58 at 247 and 253. 5  Moss raises a closely related problem: why the desire for pleasure, more than other desires, should be susceptible to distance illusions (‘Pleasure and Illusion in Plato’, Philosophy and Phenomenological Research, 72 (2006), 503–35 at 509–10). According to Moss, the solution, which Plato recognizes in the Gorgias and Republic, is that the desire for pleasure is non-­ calculative, and hence unquestioningly accepts the appearance that proximate pleasures are better. Moss claims that this solution is unavailable to Plato in the Protagoras, since there he takes the desire for pleasure to be responsive to the measuring art and hence calculative. By contrast, the problem I am examining concerns the appearances themselves, why proximate pleasures should appear greater than distant ones. 6  Among its numerous proponents are M. Frede, ‘Introduction’, in S. Lombardo and K. Bell (trans.), Plato: Protagoras (Indianapolis, 1992), vii–­xxxiv at xxix–­xxx; T. Irwin, Plato’s Moral Theory (Oxford, 1977), 78–9 and 117; Irwin, Ethics, 209–10; Penner, ‘Plato and Davidson’; Penner, ‘Socrates on the Strength of Knowledge’; Reshotko, Socratic Virtue; and H. Segvic, ‘No One Errs Willingly: The Meaning of Socratic Intellectualism’, Oxford Studies in Ancient Philosophy, 19 (2000), 1–45. 7  Devereux, ‘Kantian Conception’; T. C. Brickhouse and N. D. Smith, ‘Socrates on akrasia, Knowledge, and the Power of Appearance’ [‘Socrates on akrasia’], in C. Bobonich and P. Destrée (eds.), Akrasia in Greek Philosophy (Leiden, 2007), 1–17; T. C. Brickhouse and N. D. Smith, Socratic Moral Psychology [Moral Psychology] (Cambridge, 2010); Singpurwalla, ‘Reasoning with the Irrational’.

44

Suzanne Obdrzalek

hedonic calculus in comparing pleasures near and far? This has led these interpreters to reject the trad­ition­al interpretation as hopelessly arbitrary.8 Instead, they propose that Socrates must hold that there are non-­rational desires; it is our non-­rational desire for proximate pleasures which causes us to falsely inflate their value. Henceforth, I shall refer to this as the non-­rational interpretation. The non-­rational interpretation takes an interpretive risk in attributing to Socrates the belief that there are non-­rational desires, since on the most obvious and popular reading of the Socratic dialogues, he holds that all desires are rational, in the sense that all desires are aimed at the perceived good.9 However, if their reading could solve this puzzle, if it could explain why proximate pleasures appear larger than they actually are, then it would possess considerable appeal. In this paper, I defend the traditional interpretation against the non-­rational interpretation.10 I begin by outlining Socrates’ argument against akrasia in the Protagoras, paying particular attention to his discussion of the evaluative error of the would-­be akratic. Next, I provide further motivation for my puzzle concerning Socrates’ treatment of evaluative illusion. After setting out the non-­rational interpretation, I then argue that it offers no advantage over the traditional interpretation in explaining the evaluative illusion and, 8  Brickhouse and Smith, ‘Socrates on akrasia’, 10. 9  Note, however, that Singpurwalla differs from Devereux and Brickhouse and Smith in defining non-­rational desires as desires that arise independently of our reasoned conception of the good. 10 Like me, Storey seeks to provide an account of the evaluative error in the Protagoras that does not appeal to non-­rational passions (‘Sex, Wealth, and Courage: Kinds of Goods and the Power of Appearance in Plato’s Protagoras’ [‘Sex, Wealth, and Courage’], Ancient Philosophy, 38 (2018), 241–63 at 250). Our accounts differ sharply, insofar as Storey argues that the error is not that of miscalculating the relative magnitudes of proximate and distant pleasures, but, rather, of failing to recognize that goods such as health and wealth produce pleasure and are therefore commensurable with proximate pleasures such as those relating to food, drink, and sex. Storey raises an important issue in observing that the goods being compared are different in kind. However, the fact that Socrates refers to the form of wisdom needed to combat the power of appearance as the measuring art (ἡ μετρητικὴ τέχνη, 356 d 4), and specifies that it is ‘the study of relative excess and deficiency and equality’ (ὑπερβολῆς τε καὶ ἐνδείας οὖσα καὶ ἰσότητος πρὸς ἀλλήλας σκέψις, 357 b 2–3) weighs de­cisive­ly against his interpretation. Furthermore, Socrates offers his alternative ex­plan­ation of akrasia after securing the agreement of the many to the principle that pleasure and pain are the only goods and evils; if their error were a failure to recognize that the later goods are, in fact, to be valued for the sake of the pleasure they produce, then his subsequent discussion of the evaluative illusion would be redundant.



Evaluative Illusion in Plato’s Protagoras

45

furthermore, that it is saddled with difficulties. In the final part of the paper, I offer a close examination of Socrates’ treatment of evalu­ ative illusion. In particular, I attempt to answer the following three questions: first, what, exactly, are the appearances that cause the agent to err? Second, what does their power consist in and how is the measuring art supposed to remove this power? And finally, what resources could Socrates appeal to in order to explain why proximate pleasures appear greater than distant ones? I propose that appearances in the Protagoras are vivid and imagistic mental representations that the agent forms of imminent pleasures; their power lies in the fact that they appear true and hence invite assent. I conclude by speculating that proximate pleasures appear greater than distant ones because we represent them more vividly, which yields greater anticipatory pleasure, causing us to overestimate their magnitude.

2.  Akrasia in the Protagoras To begin, I will provide a brief summary of Socrates’ treatment of akrasia in the Protagoras. In the Protagoras, Socrates famously argues that weakness of will is impossible; what appears to be weakness of will is, in fact, ignorance due to poor hedonic calculus. In developing this argument, Socrates is opposing the position he attributes to hoi polloi; henceforth, I will refer to this as the common account of akrasia. According to the common account, knowledge: οὐκ ἰσχυρὸν οὐδ’ ἡγεμονικὸν οὐδ’ ἀρχικὸν εἶναι . . . ἐνούσης πολλάκις ἀνθρώπῳ ἐπιστήμης οὐ τὴν ἐπιστήμην αὐτοῦ ἄρχειν ἀλλ’ ἄλλο τι, τοτὲ μὲν θυμόν, τοτὲ δὲ ἡδονήν, τοτὲ δὲ λύπην, ἐνίοτε δὲ ἔρωτα, πολλάκις δὲ ϕόβον, ἀτεχνῶς διανοούμενοι περὶ τῆς ἐπιστήμης ὥσπερ περὶ ἀνδραπόδου, περιελκομένης ὑπὸ τῶν ἄλλων ἁπάντων. (352 b 4–c 2) is not a powerful thing, neither a leader nor a ruler . . . while knowledge is often present in a man, what rules him is not knowledge but rather anything else—­ sometimes anger, sometimes pleasure, sometimes pain, at other times love, often fear; they think of his knowledge as being utterly dragged around by all these other things as if it were a slave.

As a result, we often fail to act in accordance with what we know to be best. Even when we know that it is best to, say, resist another drink, the pleasure of the drink somehow compels us to take another sip. Socrates, by contrast, insists that knowledge is in­vin­cible:

46

Suzanne Obdrzalek

καλόν τε εἶναι ἡ ἐπιστήμη καὶ οἷον ἄρχειν τοῦ ἀνθρώπου, καὶ ἐάνπερ γιγνώσκῃ τις τἀγαθὰ καὶ τὰ κακά, μὴ ἂν κρατηθῆναι ὑπὸ μηδενὸς ὥστε ἄλλ’ ἄττα πράττειν ἢ ἃν ἐπιστήμη κελεύῃ (352 c 3–6) knowledge is a fine thing capable of ruling a person, and if someone were to know what is good and bad, then he would not be forced by anything to act otherwise than knowledge dictates.

According to Socrates, if one knows what is best, one will always act in accordance with one’s knowledge; akrasia is impossible. Socrates’ attack on the common account of akrasia takes the following form. First, he takes great pains to establish that his hypothetical adversaries, the proponents of the common account, agree to the following hedonic assumption: pleasure is the only good, pain the only evil (355 a 1–5). Next, Socrates encapsulates the common account in the following two claims: 1. ‘frequently a man, knowing the bad to be bad, nevertheless does that very thing, when he is able not to do it, having been driven and overwhelmed by pleasure’.11 2. ‘a man knowing the good is not willing to do it, on account of immediate pleasure, having been overcome by it’.12 Socrates then uses the agreed-­upon hedonic assumption to perform a series of substitutions. First, he substitutes ‘good’ for ‘pleasure’ in (1), which yields the following: frequently a man, knowing the bad to be bad, nevertheless does that very thing, when he is able not to do it, having been driven and overwhelmed by the good. Then, he substitutes ‘painful’ for ‘bad’ in (1), which yields the following: frequently a man, knowing the painful to be painful, never­ the­less does that very thing, when he is able not to do it, having been driven and overwhelmed by pleasure. Socrates specifies that in both cases, the good or the pleasure that supposedly overpowers the agent is less than the bad or the pain that he acquires through his action. According to Socrates, the two claims that result from these substitutions are somehow ridiculous; interpreters are divided over where the ridiculousness lies. 11 355 a 7–b 1: πολλάκις γιγνώσκων τὰ κακὰ ἄνθρωπος ὅτι κακά ἐστιν, ὅμως πράττει αὐτά, ἐξὸν μὴ πράττειν, ὑπὸ τῶν ἡδονῶν ἀγόμενος καὶ ἐκπληττόμενος. 12 355 b 2–3: γιγνώσκων ὁ ἄνθρωπος τἀγαθὰ πράττειν οὐκ ἐθέλει διὰ τὰς παραχρῆμα ἡδονάς, ὑπὸ τούτων ἡττώμενος.



Evaluative Illusion in Plato’s Protagoras

47

We now come to the section of the argument that interests me most. Socrates imagines his adversaries objecting that present pleasures differ very much from distant ones, to which he replies that they do not differ in any other way than in quantity. Pleasures and pains must be weighed against one another as if on a scale; the path which secures the greatest amount of pleasure versus pain is the one that must be pursued. What, then, accounts for all the cases where we choose the less pleasant course of action? Here, I will quote Socrates’ exchange with his imagined interlocutors at length: ϕαίνεται ὑμῖν τῇ ὄψει τὰ αὐτὰ μεγέθη ἐγγύθεν μὲν μείζω, πόρρωθεν δὲ ἐλάττω· ἢ οὔ; Φήσουσιν. Καὶ τὰ παχέα καὶ τὰ πολλὰ ὡσαύτως; καὶ αἱ ϕωναὶ ἴσαι ἐγγύ-θεν μὲν μείζους, πόρρωθεν δὲ σμικρότεραι; Φαῖεν ἄν. Εἰ οὖν ἐν τούτῳ ἡμῖν ἦν τὸ εὖ πράττειν, ἐν τῷ τὰ μὲν μεγάλα μήκη καὶ πράττειν καὶ λαμβάνειν, τὰ δὲ σμικρὰ καὶ ϕεύγειν καὶ μὴ πράττειν, τίς ἂν ἡμῖν σωτηρία ἐϕάνη τοῦ βίου; ἆρα ἡ μετρητικὴ τέχνη ἢ ἡ τοῦ ϕαινομένου δύναμις; ἢ αὕτη μὲν ἡμᾶς ἐπλάνα καὶ ἐποίει ἄνω τε καὶ κάτω πολλάκις μεταλαμβάνειν ταὐτὰ καὶ μεταμέλειν καὶ ἐν ταῖς πράξεσιν καὶ ἐν ταῖς αἱρέσεσιν τῶν μεγάλων τε καὶ σμικρῶν, ἡ δὲ μετρητικὴ ἄκυρον μὲν ἂν ἐποίησε τοῦτο τὸ ϕάντασμα, δηλώσασα δὲ τὸ ἀληθὲς ἡσυχίαν ἂν ἐποίησεν ἔχειν τὴν ψυχὴν μένουσαν ἐπὶ τῷ ἀληθεῖ καὶ ἔσωσεν ἂν τὸν βίον; . . . Τί δ’ εἰ ἐν τῇ τοῦ περιττοῦ καὶ ἀρτίου αἱρέσει ἡμῖν ἦν ἡ σωτηρία τοῦ βίου, ὁπότε τὸ πλέον ὀρθῶς ἔδει ἑλέσθαι καὶ ὁπότε τὸ ἔλαττον, ἢ αὐτὸ πρὸς ἑαυτὸ ἢ τὸ ἕτερον πρὸς τὸ ἕτερον, εἴτ’ ἐγγὺς εἴτε πόρρω εἴη; τί ἂν ἔσῳζεν ἡμῖν τὸν βίον; ἆρ’ ἂν οὐκ ἐπιστήμη; καὶ ἆρ’ ἂν οὐ μετρητική τις, ἐπειδήπερ ὑπερβολῆς τε καὶ ἐνδείας ἐστὶν ἡ τέχνη; ἐπειδὴ δὲ περιττοῦ τε καὶ ἀρτίου, ἆρα ἄλλη τις ἢ ἀριθμητική; . . . Εἶεν, ὦ ἄνθρωποι· ἐπεὶ δὲ δὴ ἡδονῆς τε καὶ λύπης ἐν ὀρθῇ τῇ αἱρέσει ἐϕάνη ἡμῖν ἡ σωτηρία τοῦ βίου οὖσα, τοῦ τε πλέονος καὶ ἐλάττονος καὶ μείζονος καὶ σμικροτέρου καὶ πορρωτέρω καὶ ἐγγυτέρω, ἆρα πρῶτον μὲν οὐ μετρητικὴ ϕαίνεται, ὑπερβολῆς τε καὶ ἐνδείας οὖσα καὶ ἰσότητος πρὸς ἀλλήλας σκέψις; (356 c 5–357 b 3) ‘Do things of the same size appear to you larger when seen near at hand and smaller when seen from a distance, or not?’

48

Suzanne Obdrzalek

They would say they do. ‘And similarly for thicknesses and pluralities? And equal sounds seem louder when near at hand, softer when farther away?’ They would agree. ‘If then our well-­being depended upon this, doing and choosing large things, avoiding and not doing the small ones, what would we see as our salvation in life? Would it be the art of measurement or the power of appearance? While the power of appearance often makes us wander all over the place in confusion, often changing our minds about the same things and regretting our actions and choices with respect to things large and small, the art of measurement in contrast, would make the appearance lose its power by showing us the truth, would give us peace of mind firmly rooted in the truth and would save our life.’ . . . ‘What if our salvation in life depended on our choices of odd and even, when the greater and lesser had to be counted correctly, either the same kind against itself or one kind against the other, whether it be near or remote? What, then, would save our life? Surely nothing other than knowledge, specifically some kind of measurement, since that is the art of the greater and the lesser? In fact, nothing other than arithmetic, since it’s a question of odd and even?’ . . . ‘Well, then, my good people, since it has turned out that our salvation in life depends on the right choice of pleasures and pains, be they more or fewer, greater or lesser, farther or nearer, doesn’t our salvation seem, first of all, to be measurement, which is the study of relative excess and deficiency and equality?’

Socrates begins by making an observation about visible objects: objects appear greater when near at hand and smaller when far away, even when their size remains fixed. He then extends this to the cases of thicknesses, pluralities, and sounds: they are all subject to the same form of distortion. Socrates then draws a contrast between the power of appearance and the measuring art. The power of appearance causes us to wander in confusion, to change our minds, and to feel regret in our actions and choices concerning the large and small. The measuring art, by contrast, through showing us the truth, makes appearances lose their power; it gives us peace of mind rooted in truth and saves our lives. Socrates then applies this analysis to the case of pleasure. He implies that, in the case of pleasure, poor choices are due to the power of appearance. Since pleasure is assumed to be the good, our well-­being depends on the correct measurement of pleasures and pains, and the measuring art is our salvation.



Evaluative Illusion in Plato’s Protagoras

49

Earlier, I referred to what I called the traditional interpretation of Socrates’ argument in the Protagoras. What sort of an account can this interpretation provide of the passage we are focusing on? First, Socrates makes clear that he is seeking to redescribe the phe­nom­enon of akrasia. This means that, on the one hand, Socrates must do justice to at least some aspects of the experience that the common account views as akrasia; on the other hand, his ex­plan­ation must differ from the common account in some significant way. According to the common account, knowledge is dragged around (περιελκομένης, 352 c 2) by pleasure and pain and hence does not determine how we act. Interestingly, in Socrates’ redescription of akrasia, his language comes very close to that of the common account: he writes of how the power of appearance makes us wander (ἐπλάνα), often changing our minds (356 d 5). Nonetheless, there are two significant differences between these accounts. First, whereas, on the common account, pleasure drags us and enslaves us, on Socrates’ new account, it causes us to wan­ der. This implies a greater degree of agency. Second, the cause of  our downfall is s­ubtly different. Whereas, on the common account, it is pleasure that drags us, on Socrates’ alternative account, it is the power of appearance. We can put these pieces together as follows: on the common account, pleasure somehow directly causes us to act against our judgement of what is best. By contrast, on Socrates’ new account, what is op­era­tive is an appearance, an appearance of a pleasure as greater than it actually is; it is this appearance that causes us to commit an error. Thus, on the traditional in­ter­pret­ation, the error of the would-­be akratic is purely intellectual: it is a matter of her assenting to appearances that are false. It is in this sense that her error is due to ignorance and not due to irrational desire. What does Socrates mean when he speaks of the appearances as making us wander and change our minds? What occurs is that the agent assents to shifting appearances depending on the proximity of the pleasure in question. When a pleasure is far off—­either pro­spect­ ive­ly or retrospectively—­it appears smaller, but when it becomes imminent, it appears greater, causing a shift in judgement.13 After 13  It is worth noting that nothing in how Socrates presents the case at 356 c 5–357 b 3 indicates whether the error is due to an inflation of the proximate pleasure, a deflation of the distant one, or an error at both ends of the spectrum. He merely

50

Suzanne Obdrzalek

the fact, the agent is able to look back and see that she misjudged the pleasure’s value; it is this recognition that causes regret. These shifts in the agent’s judgement are what make her case differ from mere intemperance: whereas the intemperate person, say, consistently overvalues eating sweets in comparison with being healthy, the akratic shifts her assessment of their relative values depending on their temporal proximity. This instability in the akratic’s evaluations enables Socrates to account for at least some aspects of the ex­peri­ence that the common account treats as akrasia. While Socrates cannot and does not allow for synchronic akrasia, he can allow for diachronic akrasia, the phenomenon of acting against an earlier considered judgement of what is best.14 If the agent had the measuring art and hence knowledge of what is best, she would not, of course, be susceptible to this confusion. She would see through the power of appearance, and it would fail to motivate her. The difficulty with mere belief, in the Protagoras, as in the Meno, is that it is shifty, given to wandering; the power of knowledge lies in the fact that it remains unwavering in guiding correct action.

3.  The puzzle of temporal distortion Why, then, do I find Socrates’ explanation of the evaluative error of the would-­be akratic puzzling? Think back to our case of the child who chooses one marshmallow now over two later. It seems reasonable to assume that in many cases, consuming two states that the same objects appear greater when near at hand and smaller when far away (ϕαίνεται ὑμῖν τῇ ὄψει τὰ αὐτὰ μεγέθη ἐγγύθεν μὲν μείζω, πόρρωθεν δὲ ἐλάττω, 356 c 5–6). As I read these lines, the comparison is between how large objects appear when they are near at hand and how large they appear when far away (and not, say, between how large objects appear when near or far away and how large they actually are in some objective sense); Socrates’ claim concerns the relative magnitudes of the objects as they appear to us. As I shall go on to suggest in Section 7, in the hedonic case, the distortion is a result of a key feature in how we represent proximate pleasures: the vivacity with which we represent proximate pleasures causes us to overinflate their value. This is why, as the common account captures, we blame the proximate pleasure for causing us to err (355 b 3). But, of course, this overinflation of the proximate pleasure causes us to overestimate its magnitude relative to the distant one, consistently with Socrates’ characterization of the case at 356 c 5–6. 14  See, among numerous others, Penner, ‘Plato and Davidson’ and, following Penner, Reshotko, Socratic Virtue, 80–2.



Evaluative Illusion in Plato’s Protagoras

51

marshmallows would actually give the child more pleasure than consuming one. And the fact that the children visibly struggled to resist the marshmallow that lay in front of them suggests that they recognized this fact. But according to Socrates, they were guilty of a very specific error. Their error was not desiring the present marshmallow more. It was not even supposing that eating the present marshmallow is better. It was believing that eating one marshmallow now would afford them more pleasure than eating two fifteen minutes later. And that seems both implausible and odd. It is important to emphasize that the error Socrates attributes to the would-­be akratic is that of overestimating the magnitude of imminent pleasures. The reason this is worth emphasizing is that virtually every interpreter who discusses our passage slips into describing the error as though it concerned the goodness of the pleasure in question. Thus, for example, Devereux describes Socrates’ position as follows: ‘Just as an object’s being very close may make it appear larger than it is, so the fact that some pleasant (or painful) experience is immediately at hand can make it seem more important or valuable than it really is’ (‘Kantian Conception’, 391, emphasis added). Singpurwalla, similarly, writes, ‘One likely explanation for the appearances is that how good or bad a certain pleasure or pain appears to us is a reflection of the strength of our current attraction or aversion for it’ (‘Reasoning with the Irrational’, 249, emphasis added).15 But of course, Socrates does not argue that proximate pleasures appear more important, valuable, or good than distant ones; rather, he claims that they appear bigger. This is why it is the measuring art that Socrates proposes as our salvation in life, describing it as ‘the study of relative excess and deficiency and equality’ (ὑπερβολῆς τε καὶ ἐνδείας οὖσα καὶ ἰσότητος πρὸς ἀλλήλας σκέψις, 357 b 2–3). Granted, on the hedonist assumption, in judging that an imminent pleasure is greater, one also judges that it is better. But it is telling that interpreters slip into treating the power of appearance as though it simply made 15  Though this tendency is less pronounced in Brickhouse and Smith, they describe non-­rational desire as desire that ‘inclines us to believe that what it is attracted to is a good, and if sufficiently strong it causes us to believe that what it is attracted to is good by preventing us from seeing or being persuaded by reasons for thinking it is not good’; they add that those with strong appetites are compelled ‘to see illicit pleasures or enjoyments as good’ (‘Socrates on akrasia’, 17, emphasis added). The emphasis is on the pleasure’s appearing good, rather than on its ­seeming large.

52

Suzanne Obdrzalek

imminent pleasures appear better. It is, arguably, difficult to determine which of two pleasures is better, and one can see how this might be a subject of confusion that could give rise to error. But if we identify the goodness of a pleasure with its overall magnitude, then the comparisons seem more straightforward and the error more odd.16 One might attempt to explain this miscalculation along a variety of paths. For example, one might propose that one who chooses proximate pleasures does so because he holds that their being proximate in itself makes them more valuable. However, Socrates cannot avail himself of this manoeuvre. If, on the one hand, we take more valuable to mean more pleasant, then we are back to our original question: why do proximate pleasures appear greater than distant ones? If, on the other hand, the agent values proximate pleasures for their proximity, in addition to or independently of their magnitude, then, as Taylor (Protagoras, 193) observes, Socrates would not propose the measuring art as the solution to his woes. Furthermore, before offering his alternative explanation of akrasia, Socrates secures his opponents’ agreement to a principle of tem­ poral neutrality: pleasures only differ in magnitude; when they occur makes no difference to their value (356 a 5–c 3). Alternatively, one might argue that the one who chooses prox­im­ate pleasures does so because he wishes to minimize the pain of an­tici­ pa­tion. In a similar vein, one might propose that people tend to choose proximate pleasures because the insecurity of future pleasures diminishes their value. The difficulty with both these explanations is that the agent would be acting rationally and not commit an evaluative error. But it is agreed upon by all that the agent who chooses the proximate pleasure commits an error.17 Finally, one might maintain that the evaluative error is due to the fact that the pleasures and pains are difficult to compare with one another. Thus, one might argue that my opening example of the marshmallows was overly simple, since it seems clear that two marshmallows later will yield more pleasure than one now. But the 16  But see A. W. Price, Mental Conflict (London, 1995), 22–4. 17  Would it help to modify this proposal such that the agent tends to overestimate the pain of anticipation or the insecurity of future pleasures? Not really—­we would then simply replace our original problem, of why we overestimate the magnitude of proximate pleasures with a new one, of why we overestimate the pain of a­ n­tici­pa­tion or the insecurity of future pleasures.



Evaluative Illusion in Plato’s Protagoras

53

examples Socrates considers are eating, drinking, and enjoying sex now, versus being poor and sick later (353 c 6–d 4).18 Perhaps these pleasures and pains are more difficult to compare, making the evalu­ative error less jarring. Against this, I would argue that we do tend to compare all sorts of disparate pleasures and pains: indeed, such comparisons are a ubiquitous feature of everyday practical deliberation. Furthermore, it is not clear how different in kind these pleasures are: sex and sickness are both corporeal conditions and, in that regard, similar in kind. Lastly, observing that certain pleasures and pains are difficult to compare gives us no explanation of why we tend to inflate the value of the proximate rather than the distant ones. Suppose, though, that one allows that there is something odd or even implausible about Socrates’ proposal that proximate pleasures appear greater than distant ones. One might still maintain, against me, that this does not call out for further explanation. The problem, such an interpreter might argue, is not Socrates’ problem. It is commonly assumed that Socrates does not, himself, endorse the hedonist assumption; it is, rather, a view he attributes to hoi polloi.19 Thus, my opponent might argue that in proposing that what appears to be akrasia results from the tendency of proximate pleasures to appear magnified, Socrates is just pointing out to the hedonist that this is the only explanation he can provide of cases where people appear to behave akratically. If this is a poor ex­plan­ation, then fault lies with the hedonist, not Socrates.20 In response, I would note that this interpretation is prima facie implausible, in light of the fact that Socrates makes virtually the same claim in the Philebus, as part of his analysis of false pleasures (41 a 7–42 c 3). In that later dialogue, there is no reason to suppose that Socrates does not endorse the claim; thus, one cannot straightforwardly argue that it is so implausible that when Socrates 18  One might also take 354 a 4–b 5 to imply the possibility of cases of akrasia where one forgoes athletic training or medical treatment or avoids military campaigns and as a result suffers ill health, poverty, or political defeat later; the same argument would apply. 19  See e.g. Dyson, ‘Knowledge and Hedonism’, 40–3; G. R. F. Ferrari, ‘Akrasia as Neurosis in Plato’s Protagoras’ [‘Akrasia as Neurosis’], Proceedings of the Boston Area Colloquium in Ancient Philosophy, 6 (1990), 115–39 at 131–2; Frede, ‘Introduction’, xxvii–­viii; Vlastos, ‘Acrasia’, 78; and Wolfsdorf, ‘Ridiculousness’, 133–4. 20  Dyson, ‘Knowledge and Hedonism’, and Ferrari, ‘Akrasia as Neurosis’, offer dialectical interpretations of Socrates’ argument in a similar vein.

54

Suzanne Obdrzalek

advances it in the Protagoras, he must be doing so dialectically, and not in propria persona. Furthermore, if, indeed, Socrates were offering what he takes to be a patently implausible explanation of apparent cases of akrasia, then he would do little to convince his interlocutors that akrasia is impossible; rather than abandoning hedonism, they would more likely reject Socrates’ analysis and conclude that akrasia is, indeed, possible. But this would be catastrophic for Socrates’ philosophical aims in the dialogue: his denial of akrasia undergirds his subsequent argument that courage is identical to wisdom (358 b 6–360 e 5), which, in turn, serves as the capstone of his argument for the unity of the virtues. Relatedly, if, Socrates does not, in fact, hold that we are subject to hedonic errors, then there will be no role for the measuring art, but his praise of the measuring art is part of his overarching aim in the dialogue of promoting the value of wisdom.21 Finally, it is worth noting that the sorts of cases that Socrates seeks to explain in this passage are cases where the goods being weighed are naturally understood in hedonistic terms: when one chooses to pursue food or sex and then later regrets the consequent ill health, one’s calculations largely involve bodily pleasure and pain. This does not imply that Socrates thinks that bodily pleasure and pain are the only or the chief goods. But if he wishes to show that akrasia is impossible, he must have resources to deal with such cases, cases in which, if there is an evalu­ative error, the error involves miscalculating the relative values of proximate and distant pleasures and pains.

4.  The non-­rational interpretation The worry that I have attempted to articulate, that Socrates’ account of evaluative error as due to temporal distortion is somehow wanting, has led several recent interpreters to propose that the error Socrates is describing cannot be purely calculative. They argue that the only way to explain the evaluative error is to posit that the agent is subject to non-­rational desires for proximate pleasures, desires which cause him to falsely judge the proximate pleasures to be larger than 21  M. C. Nussbaum emphasizes the importance that Socrates assigns to making the case for an ethical measuring art (The Fragility of Goodness [Fragility] (Cambridge, 1999), 108–10).



Evaluative Illusion in Plato’s Protagoras

55

they actually are. I have been referring to this as the non-­rational interpretation. Its primary proponents are Devereux, Brickhouse and Smith, and Singpurwalla; I will now briefly set out their views. In his article, ‘Socrates’ Kantian Conception of Virtue’, Devereux argues that the traditional interpretation lacks the resources to explain the evaluative error Socrates describes in the Protagoras. He writes: If the distortion is understood as a ‘purely intellectual’ error, it is hard to see what might account for it. In the case of perceptual distortion, we can appeal to the fact that when an object is very near it takes up a large proportion of the visual field, and this can sometimes lead to overestimation of its size. But there does not seem to be an analogous factor, separate from desire, that might account for our ‘overestimating’ pleasures and pains that are near in time. (395 n. 27)

Devereux, therefore, argues that, contrary to the traditional in­ter­ pret­ation of the early dialogues, we must attribute to Socrates the view that there exist non-­rational desires. It is the agent’s non-­ rational desire for the proximate pleasure that accounts for his belief that it is greater. Like Devereux, Brickhouse and Smith maintain that the trad­ ition­al interpretation cannot offer an adequate explanation of ­temporal distortion. They offer the following case: suppose that someone who has just eaten an enormous meal is presented with a chocolate tart. He considers the health repercussions and decides not to consume it. However, a little later, when he is not quite so full, he changes his mind and decides to eat the tart after all. What could possibly explain his shift in judgement? On the traditional interpretation, the error is purely calculative, but the agent has been given no new information. Brickhouse and Smith write, ‘If we are to avoid what appears to be the hopeless arbitrariness of the position the traditionalist ascribes to Socrates, we must think that Socrates recognizes that nonrational desires have an explanatory role to play in [the person’s] decision to devour [the chocolate tart] at [the later time].’22 Brickhouse and Smith differ from 22  ‘Socrates on akrasia’, 10. Brickhouse and Smith’s example is, perhaps, infelicitous for their purposes. For, in fact, it is not as though the agent gets no new information. Per their example, when the agent reconsiders, he ‘has managed to digest enough of his previous meal to lose his feelings of complete satiety’ (‘Socrates on akrasia’, 9)—thus, he does have new information about his bodily condition, namely that he is no longer so full as to be unable to enjoy the tart; this information supports

56

Suzanne Obdrzalek

Devereux in maintaining that the measuring art is compatible with the existence of weak, but not of strong non-­ rational desires. According to Devereux, knowledge makes appearance lose its power because it always generates stronger desires for what it knows to be best. Brickhouse and Smith worry that this will not secure the agent the peace which Socrates promised would ensue from the measuring art (‘Socrates on akrasia’, 14–15). They address this by distinguishing between strong and weak non-­ rational desires. The measuring art brings peace because it is incompatible with the existence of strong non-­rational desires, desires that compel the agent to believe that their objects are good. Most recently, Singpurwalla, like Devereux and Brickhouse and Smith, argues that we must take Socrates to assume the existence of non-­rational desires in order to make sense of the evaluative error. She writes: It is a fact that our judgments of value can be erroneously affected by the relative proximity of the goods or bads in question. But now this fact demands an explanation. . . . One likely explanation for the appearances is that how good or bad a certain pleasure or pain appears to us is a reflection of the strength of our current attraction or aversion for it. Moreover, the strength of our attraction or aversion to certain pleasures or pains is, no doubt, often affected by the relative proximity of the pleasure or pain in question. But the strength of our attraction or aversion to certain pleasures or pains can arise independently of reason and can conflict with our reasoned judgment of the worth of the pleasure or pain. Thus, just as facts about our visual apparatus can cause us to see things differently from how they are, facts about the way we desire can cause us to ‘see’ things as worth more or less than they are. (‘Reasoning with the Irrational’, 249)

Singpurwalla differs from Devereux and Brickhouse and Smith in terms of her account of non-­rational desires. Where they treat non-­rational desires as pure feelings of attraction, she proposes, instead, that they are themselves beliefs about what is good, beliefs that are based on appearances and that are resistant, but not immune to rational revision. a new assessment of whether the pleasure of the tart will outweigh the eventual health repercussions. Thus, this example would, in fact, seem to support the trad­ ition­al interpretation on which the appetites can provide information about our bodily condition that can then factor into our rational deliberation and desires. (See also R. Jones, ‘Review of Thomas C. Brickhouse and Nicholas D. Smith, Socratic Moral Psychology’, Polis, 28 (2011), 147–52 at 150.)



Evaluative Illusion in Plato’s Protagoras

57

While I am entirely sympathetic with these interpreters’ desire for a further explanation of the evaluative error, I will argue, first, that their proposed solution does not provide any advantage over the traditional account in explaining the evaluative error and, second, that it is saddled with difficulties. 1. That the non-­rational interpretation offers no advance over the traditional interpretation in accounting for the evaluative illusion is troubling, since, as we have seen, they present their su­per­ ior­ity in this regard as one reason to favour their interpretation. But in fact, their explanation multiplies, rather than dissolves the puzzle concerning the evaluative illusion. On the traditional in­ter­ pret­ation, we are faced with the problem of why proximate pleasures should appear greater than distant ones. The non-­rational interpretation addresses this by proposing that the proximity of the pleasures causes us to form a non-­rational desire for them; this non-­rational desire, in turn, causes us to overestimate their magnitude. But now we have two puzzles where before we had one: we must now ask, on the one hand, why we should desire proximate pleasures more than distant ones and, on the other, why our desiring them more causes us to overestimate their magnitude. In response to the first question, proponents of the non-­rational interpretation perhaps take it to be a brute fact that we just do desire proximate pleasures more than distant ones. But if, indeed, the pleasures are of equal magnitude, then it seems that there is no good reason to prefer the proximate to the distant, which renders it odd, or even inexplicable that we should desire them more. Thus, it is not clear why proponents of the non-­rational account should be in a position to, at once, criticize the traditional account for not explaining why proximate pleasures appear greater, while helping themselves to the claim that they are more desired. Suppose, though, that we grant that we just do desire proximate pleasures more. The non-­rational interpretation must now confront the difficulty of explaining how this desire causes us to over­ esti­mate their magnitude. Singpurwalla alludes to how ‘facts about the way we desire can cause us to “see” things as worth more or less than they are’ (‘Reasoning with the Irrational’, 249). But on the face of it, it is not obvious why wanting to enjoy a pleasure now, rather than later, should cause one to think that it will be more pleasant to do so. How, exactly, does the desire cause the inflated hedonic prediction? Brickhouse and Smith attempt to address this

58

Suzanne Obdrzalek

by claiming that non-­rational desires are present-­focused: ‘Because a nonrational desire demands immediate satisfaction, it can explain why the pleasure of C appears to be larger at t2 than it did at t1, when P did not possess a nonrational desire for C’ (‘Socrates on akrasia’, 10). But why should the fact that the non-­rational desire demands immediate satisfaction cause us to overestimate the ensuing pleasure, whereas desires that are future-­focused do not introduce such a distortion? They later add, ‘Something acquires the power of appearance when it becomes the object of a nonrational desire and so becomes recognized by an agent as a way to satisfy some appetite or passion—­for example, as a pleasure or as a relief from some pain’ (‘Socrates on akrasia’, 11). If their thought is that the non-­rational desire introduces a new reason to pursue the prox­ im­ate pleasure, in virtue of the pleasure that satisfying this new desire will produce, then what we have is a version of the trad­ ition­al interpretation: one would be revising their assessment of the worth of the proximate pleasure in response to new information.23 Alternatively, proponents of the non-­rational interpretation might argue that our heightened desire for the proximate pleasure causes us to focus on it more, ignoring other alternatives.24 But per the parameters of the problem of akrasia, it is not as though the agent is ignorant of other alternatives; at most, she is less attentive to them. And it is not obvious why increased attention should result in overestimation and not, say, underestimation or some other form of cognitive distortion. This second problem, of why, exactly, non-­rational desire should cause overvaluation does not arise on Singpurwalla’s account; at most, she faces the first difficulty, of explaining why we should form greater non-­rational desires for proximate, rather than distant pleasures. The reason the second problem does not arise is that, on her interpretation, the non-­rational desire just is the belief that the proximate pleasure is better and hence bigger. Thus, there is no question of how the desire should give rise to the belief, since it simply is the belief. But this reveals a different challenge for her account. Namely, if Singpurwalla treats the non-­rational desire as an evaluative belief, then it is not clear why her account should offer any improvement on the traditional interpretation. The supposed 23  See Reshotko, Socratic Virtue, 86. 24  I owe this suggestion to Rachel Singpurwalla.



Evaluative Illusion in Plato’s Protagoras

59

advantage of the non-­rational interpretation, that it appeals to some­ thing besides belief to explain belief formation, is lost if the non-­ rational desire simply is an evaluative belief.25 2a. In addition to possessing no explanatory advantages, the non-­rational interpretation faces significant difficulties. First, as Singpurwalla concedes, it lacks direct textual support.26 2b. A further set of difficulties arises for Devereux and for Brickhouse and Smith, depending on whether or not non-­rational desires can coexist with the measuring art.27 On Devereux’s account, they can; as he concedes, this raises the problem of what guarantee there is that they will not overpower reason. Furthermore, as Brickhouse and Smith argue, even if we grant Devereux that reason will always generate stronger, rational desires, the agent does not appear to have secured the tranquillity which the measuring art 25  I am indebted to Singpurwalla for generous correspondence that has enabled me to develop a better understanding of her position. For the purpose of simplicity, I focus on the role of non-­rational desire in Singpurwalla’s account, but a more precise rendering of her position would be that we have a heightened attraction to proximate pleasures, where the attraction just is their appearing to be better and hence greater. When we assent to this attraction, we form a non-­rational desire. My worry can be reposed as follows: if it is puzzling why proximate pleasures should appear greater than distant ones, it will not help to attribute this to our being more attracted to them, if our being more attracted to them is identical to their appearing greater. 26  Singpurwalla, ‘Reasoning with the Irrational’, 247–8. Devereux (‘Kantian Conception’, 388–9) and Brickhouse and Smith (Moral Psychology, 50–62) offer an extensive discussion of passages from the early dialogues that they take to imply that Socrates does acknowledge the existence of non-­rational desires. In this paper, I do not take a stand on whether Socrates allows for non-­rational desires in other dialogues; I am simply claiming that there is no evidence that he deploys such desires to explain evaluative illusion in the Protagoras. Devereux cites Protagoras 357 c 3–4, where Socrates claims to agree with Protagoras that knowledge ‘always prevails, whenever it is present, over pleasure and everything else’ (ἀεὶ κρατεῖν, ὅπου ἂν ἐνῇ, καὶ ἡδονῆς καὶ τῶν ἄλλων ἁπάντων) as evidence for the existence of non-­rational desires. He parses this line as claiming that ‘in other words, there are resistant mo­tiv­at­ing factors, but knowledge always has the upper hand’ (‘Kantian Conception’, 389). But it is far from clear that by ‘pleasure and everything else’ Socrates means resistant motivating factors. Indeed, he has just argued against the many that if pleasure compels us to act against our perceived best interest, it does not do so by directly causing us to act, but rather by generating a mistaken assessment of its magnitude. 27  This constellation of difficulties does not arise for Singpurwalla, since, on her account, the measuring art prevents the agent from assenting to the appearance that the proximate pleasure is better and hence from forming the non-­rational desire for it. Insofar as the agent lacks a non-­rational desire for the proximate pleasure, she has achieved tranquillity.

60

Suzanne Obdrzalek

promised. Brickhouse and Smith address this difficulty by maintaining that the measuring art cannot coexist with strong non-­ rational desires; it is only active in relation to desires that are ‘weak, and thus, disposed to fall in line with knowledge of what is best’ (‘Socrates on akrasia’, 15). But then, as they acknowledge, we are left to wonder what role there is left for the measuring art (‘Socrates on akrasia’, 15). Assuming it is not totally impotent, these weak non-­rational desires must still possess some force, but then we are back to Devereux’s problem—­what non-­arbitrary reason is there to suppose that knowledge will prevail?28 Furthermore, this appears to mischaracterize the role of the measuring art. Socrates describes it as saving our lives by ‘making the appearances lose their power by showing us the truth’ (ἄκυρον μὲν ἂν ἐποίησε τοῦτο τὸ ϕάντασμα, δηλώσασα δὲ τὸ ἀληθές, 356 d 8–e 1). But if the measuring art cannot even emerge until some other faculty extinguishes strong non-­ rational desires, then it would appear to be that other faculty that deserves the title of a life-­saving capacity to defeat the power of appearances. 2c. Finally, Devereux and Singpurwalla each face difficulties when it comes to accounting for the synchronic conflict that appears to characterize akrasia. This is striking, since each maintains that it is an advantage of the non-­rational interpretation that it can explain the phenomenology of conflict, whereas the trad­ition­al in­ter­pret­ ation cannot. Devereux claims that only on the non-­rational in­ter­ pret­ation can Socrates’ alternative account plausibly explain the phenomenon that the common account wishes to see explained; in positing the existence of non-­rational desires, it has the resources to explain the ‘inner conflict between [the agent’s] belief as to how he ought to act and his desire for the immediate pleasure’ (‘Kantian Conception’, 396). Similarly, Singpurwalla complains that the trad­ition­al account cannot explain how ‘if we do take the action that is in fact bad, we feel conflicted about it at the time of action’ (‘Reasoning with the Irrational’, 247), and promises to explain this by positing non-­rational desires. Devereux accounts for the phenomenology of inner conflict by proposing that the agent’s non-­rational desire causes his belief to 28  For a similar line of criticism, see G.  R.  Carone, ‘Calculating Machines or Leaky Jars? The Moral Psychology of Plato’s Gorgias’, Oxford Studies in Ancient Philosophy, 26 (2004), 55–96 at 89.



Evaluative Illusion in Plato’s Protagoras

61

shift at the time of action: ‘The agent starts out with a true belief, but this belief is unable to withstand the opposition of a strong desire for an immediate pleasure, and it is temporarily displaced by a belief supporting the desire’ (‘Kantian Conception’, 395). Thus, ‘Socrates has argued that an agent will never act contrary to his belief—­at the time of action—­as to what is best’ (393). It is hard to see how Devereux has preserved the phenomenon he wishes to account for: at most there will be a diachronic tension between the agent’s desire for the immediate pleasure at the time of action and his belief, prior and posterior to acting, that it is not worth pursuing. But in that case, his account is no improvement on the trad­ ition­al interpretation, since it too can account for diachronic conflict in terms of shifts in the agent’s evaluative judgements.29 Singpurwalla does better when it comes to synchronic conflict, but this comes at a cost. She writes, ‘So we have a prima facie reason at least for thinking that in both the Protagoras and the Republic the irrational part of the soul is home to beliefs based on appearances, and that motivational conflict is a conflict between two beliefs—­one based on reasoning and one based on the way things appear’ (‘Reasoning with the Irrational’, 255–6). Singpurwalla is able to give a robust account of synchronic conflict, since she proposes that the agent has compresent opposed beliefs: her non-­ rational desire for the present pleasure and her rational belief that it would be best not to pursue it. But this interpretation appears to contradict Socrates’ explicit position in the dialogue. Socrates claims at 358 b 7–c 1 that ‘no one who knows or believes that there is something else better than what he is doing, something possible, will go on doing what he had been doing when he could be doing what is better’.30 This decisively precludes the possibility of an agent pursuing a present pleasure in the face of a belief that it would be best not to do so.31 29  Thus, I am in agreement with Devereux’s account of the form of conflict that Socrates’ account allows for, namely belief instability; my point is that neither his account nor the traditional interpretation allows for synchronic conflict. 30 358 b 7–c 1: οὐδεὶς οὔτε εἰδὼς οὔτε οἰόμενος ἄλλα βελτίω εἶναι ἢ ἃ ποιεῖ, καὶ δυνατά, ἔπειτα ποιεῖ ταῦτα, ἐξὸν τὰ βελτίω. 31  Singpurwalla perhaps seeks to address this when she continues, ‘When these beliefs are about the value of a certain object or course of action, then no matter which belief we act upon we will be acting in accordance with our beliefs about value’ (‘Reasoning with the Irrational’, 256). But even if in acting on our non-­ rational desire, we act on a belief about value, we would still be acting against our

62

Suzanne Obdrzalek

In fact, though Devereux and Singpurwalla attack the traditional account as unable to account for the phenomenon of synchronic conflict, this does not pose an interpretive problem for the trad­ ition­al interpretation. For it is far from clear that Socrates wishes to preserve this aspect of the common account. He appears to reject the phenomenon of synchronic conflict precisely because he  holds that it is impossible to act against what we believe to be best.32 This does not mean, however, that Socrates is unable to capture any of the experience of conflict that characterizes akrasia. Though this may fail to do full justice to the experience of akrasia, Socrates is able to appeal to the diachronic instability of the akratic’s evalu­ative beliefs in order to preserve at least some of her ex­peri­ence of conflict.33 The power of appearance is said to ‘make us wander all over the place in confusion, often changing our minds about the same things and regretting our actions and choices with respect to things large and small’.34 What Socrates is describing here is the phenomenon of changing one’s mind repeatedly in response to shifting appearances. And since these conflicting judgements all belong to the same agent, she might look back upon herself as having been in tension with herself, insofar as her beliefs and ensuing desires shifted back and forth and conflicted with one another.35 Singpurwalla complains that the traditional interpretation cannot account for why, ‘if we do take the wrong action, we hold ourselves culpable; we blame ourselves’ (‘Reasoning with the Irrational’, 247). But in fact, the instability of belief can explain post factum regret: what we regret is not that we acted against what we knew to rational belief that it would be best to resist, and it is this possibility that Socrates explicitly rejects at 358 b 7–c 1. It is worth noting that, to the extent that Singpurwalla is able to account for the phenomenology of conflict, it is not, in fact, because she posits non-­rational desires but, rather, because she allows for the compresence of opposing beliefs about value. But most would maintain that there is no strong evidence that Plato allows for such a possibility until we come to the Republic. 32  See also Reshotko, Socratic Virtue, 89–91. 33 Here I am in full agreement with Devereux, ‘Kantian Conception’, 391–2. Devereux goes on to argue that ‘the desire thus explains the fact that the pleasure appears greater than it is, and not the other way around’ (395), claiming that this is the most plausible way to explain Socrates’ claim that knowledge is more stable than belief. But as I shall argue, the instability of belief lies not in its being easily overcome by desire but, rather, in its lacking full justification and hence being sus­cep­ tible to appearances that merely appear true. 34 356 d 5–7: ἡμᾶς ἐπλάνα καὶ ἐποίει ἄνω τε καὶ κάτω πολλάκις μεταλαμβάνειν ταὐτὰ καὶ μεταμέλειν καὶ ἐν ταῖς πράξεσιν καὶ ἐν ταῖς αἱρέσεσιν τῶν μεγάλων τε καὶ σμικρῶν. 35 Cf. Phaedo 79 c 2–8.



Evaluative Illusion in Plato’s Protagoras

63

be best at the time of action, but, rather, that we let ourselves be taken in by appearances. Perhaps, then, we should not be so quick to dismiss the trad­ ition­al interpretation. If it can explain at least as much as the non-­ rational interpretation, while falling prey to fewer difficulties, then it has much to recommend it. In Sections 5–7, I take a closer look at what sort of account the traditional interpretation can provide of the power of appearance. In particular, I address the following three questions. First, what are the appearances that cause the would-­be akratic to err? Second, what is their power and how does the measuring art counter it? And finally, how might Socrates account for the fact that these appearances tend to exaggerate the magnitude of proximate over distant pleasures?

5.  What are appearances? 1. To turn to our first question, it will be helpful to first consider Socrates’ motives in introducing appearances into his discussion of akrasia. He does so at a very specific juncture: they enter as part of his alternative explanation of why the agent chooses what is in fact worse. Socrates begins by making a series of claims about how objects near and far appear—­the same magnitude appears (ϕαίνεται) larger when nearby, smaller when far off, and so forth (356 c 5–8). Next, he contrasts the art of measurement with the power of appearance. The power of appearance (ἡ τοῦ ϕαινομένου δύναμις) causes us to wander and change our minds concerning the same things and to regret our choices and actions regarding the large and small (356 d 4–7). The measuring art, by contrast, robs the appearance (τὸ ϕάντασμα) of its power. This passage presents us with an account of perceptual illusions concerning distance: appearances confuse us by causing proximate magnitudes to appear greater than distant ones. This is intended to suggest a parallel analysis of evaluative illusion: the power of appearance causes prox­im­ate pleasures to falsely appear greater than distant ones. But here we should pause to observe that there is an im­port­ant disanalogy between the perceptual and the evaluative cases.36 In the perceptual 36  See also Dyson, ‘Knowledge and Hedonism’, 40–3, and Taylor, Protagoras, 195–6. I do not agree with Dyson’s further argument that since the pleasures and

64

Suzanne Obdrzalek

case, the two objects being compared are potentially both present to the viewer at the time when he is manipulated by the power of appearance. But this cannot occur in the evaluative case. Whereas in the perceptual case, we might, say, compare a nearby tree with one that is at a distance, in the evaluative case neither pleasure is  present at the time when the agent compares them.37 At the moment of decision, the agent is not undergoing either pleasure; he is deciding which of two prospective pleasures to pursue—­one that is in the very near future or one that is in the more distant future.38 But if the agent is comparing two pleasures, neither of which currently exists, then what is the appearance that deceives him? It must be an internal mental representation that the agent forms of the imminent pleasure which falsely represents it as larger than it in fact is.39 The role played by the agent’s mental representations of prospective pleasures is obscured on the common account of akrasia, which describes the agent as dragged by pleasure to act against his knowledge of what is best, as though the pleasure itself were a force external to and separate from the agent that compels him to act. This error is facilitated by the fact that ‘pleasure’ is ambiguous between the source of the pleasure, the pleasant activity, and the ensuing psychological state.40 Thus, at 353 c 1–8, being mastered by pleasure (ὃ ἡμεῖς ἥττω εἶναι τῶν ἡδονῶν) is equated with being mastered by food, drink, and sex, they being pleasant (ὑπὸ σίτων καὶ ποτῶν καὶ ἀϕροδισίων κρατούμενοι ἡδέων ὄντων): pleasure (ἡδονή) is treated as roughly equivalent to the objects that produce it. But whereas the psychological state is something internal, its sources—­ food, drink, beautiful people, etc.—are typically external to the pains being measured do not yet exist, there can be no knowledge of them. According to this line of argument, Socrates must hold that we can have no know­ ledge whatsoever of the future, since future events do not yet exist, but such a conclusion goes far beyond anything explicitly stated in the text. 37  Thus, when Socrates describes the many as claiming that the agent is defeated by αἱ παραχρῆμα ἡδοναί (e.g. 355 b 3), he presumably means pleasures that are imminent but not currently occurring. 38  This is occasionally missed, since akrasia often occurs when we directly perceive that which will occasion the proximate pleasure. But even when we perceive the cake before us, we still represent to ourselves the pleasure of eating it in the not too distant future. 39  Cf. Plato’s discussion of desire at Phileb. 35 b 1–d 3. 40  See V. Harte, ‘The Philebus on Pleasure: The Good, the Bad and the False’ [‘Philebus’], Proceedings of the Aristotelian Society, 104 (2004), 113–30 at 114.



Evaluative Illusion in Plato’s Protagoras

65

agent. The conflation of these senses of pleasure, in turn, gives rise to a tendency to view akrasia as involving the agent’s being mastered by something external to himself—­the food and drink or the beautiful person—­and hence beyond his control. Thus, the common account describes the agent as ruled (ἄρχειν, 352 b 7), dragged around (περιελκομένης, 352 c 2), overcome (ἡττωμένους, 352 e 1, 353 a 1, 353 a 5), and conquered (κρατουμένους, 352 e 2, 353 c 6–7) by pleasure, suggesting that he is the victim of outside forces. The common account of the Protagoras, in fact, appears to reflect a general tendency among Plato’s contemporaries to externalize the emotions and to treat them as having compulsive force.41 Thus, Gorgias famously proposes that if Helen left due to sexual desire, she acted out of compulsion, not choice.42 Again, Aeschines describes Timarchus as ‘a slave to the most shameful pleasures, to fancy food and dining extravagance, to flute girls and prostitutes, to dice, and to all other things, none of which should master a man who is well born and free’.43 While Timarchus can be faulted for failing to develop sufficient strength of character to resist these pleasures, the pleasures themselves are presented as if they were external forces that compel him to give in.44 Thus, we can see that Socrates is introducing a radical in­nov­ ation in proposing that it is not pleasant things or pleasure itself that lead the agent to choose poorly, but, rather, the agent’s own mental representation of these.45 The introduction of these mental representations, on his argument, is required even to make sense of supposed cases of akrasia. On the common account, there are only two players at work: the present pleasure and the future one. This would suggest that the present pleasure compels the agent to act by 41  Thus, Dover: ‘we may be struck by their treatment of sexual desire as an irresistible god, their externalization and personification of emotions, and their readiness to believe that a god may pervert the course of a man’s thinking without his necessarily displaying to others any outward sign of insanity’ (Greek Popular Morality in the Time of Plato and Aristotle (Indianapolis, 1994), 144). 42 Gorgias, Hel. 19. 43 Aeschin., In Tim. 42 Martin and Budé: δουλεύων ταῖς αἰσχίσταις ἡδοναῖς, ὀψοϕαγίᾳ καὶ πολυτελείᾳ δείπνων καὶ αὐλητρίσι καὶ ἑταίραις καὶ κύβοις καὶ τοῖς ἄλλοις ὑϕ’ ὧν οὐδενὸς χρὴ κρατεῖσθαι τὸν γενναῖον καὶ ἐλεύθερον. 44  Here, one might compare G. Watson’s contemporary treatment of akrasia, on which the agent can be held to blame for not developing strength of will, but is powerless to resist his desires at the moment of action (‘Skepticism about Weakness of Will’, Philosophical Review, 86 (1977), 316–39 at 334). 45  For a similar analysis, see Kamtekar, Moral, 44–9.

66

Suzanne Obdrzalek

being greater than the future one; the problem, as Socrates points out, is that this is precluded by the shared assumption that the agent is committing an error, forsaking a greater pleasure for the sake of a smaller one (355 d 3–6). Socrates is able to resolve this by introducing a third player: the agent’s mental representation of the present pleasure. It is the gap between the apparent and actual magnitudes of the present pleasure that can explain, at once, why the agent pursues the present pleasure and, at the same time, why, in so doing, he errs. But once Socrates proposes that the agent is led to act by his mental representation of the present pleasure, then he is also able to internalize the source of the agent’s error and assign him full responsibility. Far from being driven to act by food, drink, and sex, the agent is being driven by his own faulty representations of the pleasures of enjoying these as being greater than they actually are. Thus, at 356 c 5, when Socrates introduces appearances, part of what he is doing is pointing out that the responsibility for his poor choices lies with the agent himself: it is not pleasure that compels him to act, as though he were at the mercy of some external force, but rather, it is the appearance that he himself generates that is the cause of his downfall.46 2. What sort of a mental representation is this appearance?47 Here, it will be helpful to consider Plato’s use of the word phan­ tasma (pl. phantasmata) throughout his corpus. Plato’s usage is by no means univocal; indeed, phantasma has considerable semantic range. Plato uses it to refer to ghosts (e.g. Phaedo 81 d 2), shadows, and reflections (e.g. Soph. 266 b 9–10); to mental items such as dreams (e.g. Tim. 46 a 2), illusory mental representations (Rep. 584 a 9), and, more broadly, to imagistic mental representations (e.g. Phileb. 40 a 9); as well as to artistic products, such as poems (e.g. Rep. 599 a 2), to name a few. Nonetheless, there are certain threads in his usage that shed light on his deployment of the term in the Protagoras. First, the term phantasma has obvious connections to phainesthai, to how things appear. Plato, therefore, frequently uses 46 Cf. Phaedo 82 d 9–83 a 1. 47 Storey offers a helpful discussion of appearances in the Protagoras (‘Sex, Wealth, and Courage’, 247–8). Like me, he takes appearances to be quasi-­ perceptual mental presentations that incline us to believe them. Storey emphasizes the passivity of appearances (248) more than I do—­on my account, it is important that appearances are representations that we generate and for which we are responsible; I also place greater emphasis on the tendency of phantasmata to be deceptive.



Evaluative Illusion in Plato’s Protagoras

67

phantasma to denote representations that have an imagistic aspect. These can be external representations, such as artistic products, or internal ones, such as mental imagery or dreams; the latter are of  primary interest for our investigation into the Protagoras. For ex­ample, in the Philebus, Plato refers to the painted images in our souls that accompany our judgements alternately as images (εἰκόνες) and as phantasmata (e.g. 39 b 7, 40 a 9). And in the Timaeus, he describes reason as generating visible images, phantasmata and eidōla, on the surface of the liver so as to persuade appetite, which is resistant to reasoning (71 a 3–d 4). Second, in both these passages, we see, coupled with an emphasis on the imagistic aspect of phan­ tasmata, an implicit contrast with logoi (assertions, propositions).48 In the Republic (599 a 2) and Sophist (234 e 1), though Plato presents the poet and sophist as creating phantasmata with their words, the implication is that they are not presenting arguments, but are, instead, using words so as to create vivid and hence persuasive images; in the Republic, Plato at­tri­butes this to their use of rhythm, metre, and harmony (601 a 4–b 4). Finally, related to this imagistic and vivid aspect of phantasmata is a tendency to be deceptive.49 Plato thus frequently uses phantasma to denote appearances that are manipulative or even false.50 To name just a few examples, in the Republic, Plato writes that when a state of calm occurs after pain, it falsely appears pleasant, and calls this an unsound phan­ tasma and witchcraft (584 a 7–10). He describes the poet as devoid of knowledge, because he produces phantasmata, not onta—­images, not realities (599 a 1–3). Earlier, Plato specifies that the gods would not wish to lie by word or deed by producing phantasmata, illusions (382 a 1–2). In the Sophist, the sophist is defined as one who deceives us via phantasmata (240 d 1–2); phantasmata are contrasted with likenesses (εἰκόνες), since they are distorted representations that 48  See e.g. Phileb. 40 a 6–9 and Tim. 71 a 3–5. For further discussion of the relation of these passages, see H.  Lorenz, The Brute Within [Brute] (Oxford, 2006), 99–110, and J. Moss, ‘Appearances and Calculation: Plato’s Division of the Soul’, Oxford Studies in Ancient Philosophy, 34 (2008), 35–68 at 55–7. 49  This may draw on the primary sense of phantasma, namely an apparition or ghost. 50  See also M. Schofield’s excellent discussion (‘Aristotle on the Imagination’, in M. C. Nussbaum and A. O. Rorty (eds.), Essays on Aristotle’s De Anima (Oxford, 1992), 249–76 at 265–6). For a broader discussion of the deceptiveness of phantasia, see A.  Silverman, ‘Plato on Phantasia’, Classical Antiquity, 10 (1991), 123–47 at 139–47.

68

Suzanne Obdrzalek

nonetheless appear true (235 d 6–236 c 7).51 Later, Plato suggests that if there are no falsehoods, then there can be no phantasmata (264 c 12–d 1).52 Thus, consideration of Plato’s use of phantasma in other dialogues suggests that phantasmata are vivid and imagistic representations that have a tendency to be false. In the case of the Protagoras, the implication is that the appearances that cause the agent to err are misleading but vivid imagistic mental representations that she forms of imminent pleasures. Are they fully fledged beliefs? No. As, for example, Singpurwalla and Storey argue, the measuring art removes the power from appearance, but does not remove the appearance itself.53 The agent who possesses the measuring art clearly no longer forms the belief that the proximate pleasure is greater and hence better; the role of the measuring art is to forestall such belief formation by revealing that the appearance is non-­ veridical. The appearance, therefore, is a mental representation we form of the pleasure, which gives rise to belief when we accept it as veridical.54

6.  What is the power of appearance? To turn to our second question, what does the power of appearance consist in? Socrates specifies that the power of appearance makes us wander, change our minds concerning the same things many 51  See J. Beere, ‘Faking Wisdom: The Expertise of Sophistic in Plato’s Sophist’, Oxford Studies in Ancient Philosophy, 57 (2019), 153–90 at 161. 52  Note that this feature appears to be absent from Plato’s usage in the Philebus, where phantasma appears to be used more broadly to refer to images that accompany logoi in our souls, but not exclusively to false appearances. 53  Singpurwalla, ‘Reasoning with the Irrational’, 252; Storey, ‘Sex, Wealth, and Courage’, 247–8. See also e.g. M. D. Boeri, ‘Socrates, Aristotle, and the Stoics on the Apparent and Real Good’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, 20 (2004), 109–41 at 120; N. Denyer, Plato: Protagoras (Cambridge, 2008), 192; and J. Warren, The Pleasures of Reason in Plato, Aristotle, and the Hellenistic Hedonists (Cambridge, 2014), 114. For a contrasting perspective, see Penner, ‘Socrates on the Strength of Knowledge’, 137. 54  I should emphasize that I am not making the more general claim that Plato subscribes to a proto-­Stoic appearance/assent theory of belief formation. My discussion solely concerns Plato’s treatment of hedonic error in the Protagoras; my argument is that, in his discussion of hedonic error, Plato is proposing that prox­im­ ate pleasures cause us to err because they look a certain way, namely they misleadingly appear greater than they actually are.



Evaluative Illusion in Plato’s Protagoras

69

times, and feel regret. I take this to mean that appearances cause us to believe them; as the appearances change, so they cause our beliefs to shift. In virtue of what do appearances have the power to cause us to believe them? Here, it will be helpful to take a closer look at Plato’s treatment of appearances in the Sophist. In the Sophist, Plato contrasts the art of making likenesses (εἰκαστική) with the art of making appearances (ϕανταστική). The art of making likenesses preserves the proportions of the subject it is imitating. However, when sculptors make very large statues, if they were to reproduce the proportions of their subject, then its upper parts would appear smaller than they ought, and its lower parts larger. Plato writes that such craftsmen ‘say goodbye to the truth, and produce in their images the proportions that seem to be beautiful instead of the real ones’.55 He then describes this appearance as that which ‘appears to be like a beautiful thing, but only because it’s seen from a viewpoint that’s not beautiful, and would seem unlike the thing it claims to be like if you came to be able to see such large things adequately’.56 The sophist, in turn, creates appearances with his words in order to deceive and manipulate his audience. Appearances, in the Sophist, are not merely representations that are false; they are also deceptive, manipulating us into believing that they are true. They do so because our undeveloped perceptual and cognitive faculties are such that, in certain circumstances, appearances seem more true to us than faithful representations.57 They possess what Stephen Colbert famously refers to as truthiness.58 The power of appearance, then, resides in its tendency to appear true and hence to induce assent. How, then, does the measuring art counter this power? Socrates states that the measuring art is the ‘study of relative excess and deficiency and equality’ regarding pleasures and pains.59 It ‘would make the appearance lose its power and, by showing us the truth, 55  Soph. 236 a 4–6: Ἆρ’ οὖν οὐ χαίρειν τὸ ἀληθὲς ἐάσαντες οἱ δημιουργοὶ νῦν οὐ τὰς οὔσας συμμετρίας ἀλλὰ τὰς δοξούσας εἶναι καλὰς τοῖς εἰδώλοις ἐναπεργάζονται; 56  Soph. 236 b 4–6: τὸ ϕαινόμενον μὲν διὰ τὴν οὐκ ἐκ καλοῦ θέαν ἐοικέναι τῷ καλῷ, δύναμιν δὲ εἴ τις λάβοι τὰ τηλικαῦτα ἱκανῶς ὁρᾶν, μηδ’ εἰκὸς ᾧ ϕησιν ἐοικέναι. 57  Thus, the cave dwellers in the Republic initially find images more persuasive than reality (515 c 4–516 a 3), and in Book 10, hoi polloi are presented as finding the images produced by poets more persuasive than the truth (602 b 1–4). 58 S.  Colbert, The Colbert Report (17 October 2005): ‘We’re not talking about truth; we’re talking about something that seems like truth.’ 59 357 b 2–3: ὑπερβολῆς τε καὶ ἐνδείας οὖσα καὶ ἰσότητος πρὸς ἀλλήλας σκέψις.

70

Suzanne Obdrzalek

give the soul peace of mind firmly rooted in the truth and save our life’.60 Thus, the measuring art is the art of correctly determining the relative magnitudes of pleasures and pains. Through showing us the truth—­that is, showing us the correct relative magnitudes—­it disables the power of appearance. Since the power of appearance resides in its tendency to appear true and compel assent, the measuring art must reveal that this appearance is not, in fact, true, so that it ceases to appear true and compel assent. Thus, the measuring art undermines the power of appearance not by changing how things appear, but by robbing appearances of their truthiness. Compare this with how an oar in water might continue to look as if it were bent even to those familiar with the illusion, but they will see this as a mere seeming. The measuring art accomplishes this via a two-­step process. First, it generates a correct judgement of the relative magnitudes of pleasures and pains. Second, the correct judgement conflicts with the appearance, revealing it to be a mere appearance and causing it to no longer produce assent. The resulting state is one of knowledge, which is stable and unaffected by appearance. We might compare this with the perceptual case. Suppose that, unaware of distance-­illusions, I form the false judgement that a nearby horse is larger than a distant one. The way to disabuse me would be to take a measuring stick and place it next to each horse. While the distant horse might continue to look as though it were smaller than the nearby one, I would cease to believe that it is.61 This account raises two difficulties. First, as we have seen, there are certain disanalogies between the perceptual and the hedonic cases. In the perceptual case, we can frequently counter the illusion by putting ourselves in the same perceptual relation to both objects being compared. We might walk up to the distant horse, place it 60 356 d 8–e 2: ἄκυρον μὲν ἂν ἐποίησε τοῦτο τὸ ϕάντασμα, δηλώσασα δὲ τὸ ἀληθὲς ἡσυχίαν ἂν ἐποίησεν ἔχειν τὴν ψυχὴν μένουσαν ἐπὶ τῷ ἀληθεῖ καὶ ἔσωσεν ἂν τὸν βίον. Plato’s use of the μέν . . . δέ construction in this passage might be taken to imply that the measuring art first disempowers the appearance, then shows us the truth. However, given that Socrates describes the measuring art as the study of ‘relative excess and deficiency and equality’ (357 b 2–3) the mechanism by which it disempowers the appearance must be by measuring the pleasures correctly, i.e. by showing us the truth. Thus, I take the contrast implied by the μέν . . . δέ construction to not be between two distinct mechanisms deployed by the measuring art, but rather to be between two distinct, albeit related results: first the appearance is dis­em­ powered, then we gain stable knowledge and peace of mind. 61 Cf. Euthph. 7 b 2–c 8; it is not clear how optimistic Socrates is in the Euthyphro concerning the possibility of objectively measuring evaluative properties.



Evaluative Illusion in Plato’s Protagoras

71

next to the nearby horse, or compare the size of both horses with the same object, the measuring stick. But despite Socrates’ appeal to the image of hedonic scales, it is not as though we can literally place the two pleasures next to one another; as noted earlier, the pleasures do not even exist yet. So how does the measuring art generate a correct comparison of their magnitudes? Socrates defers: ‘What exactly this art, this knowledge is, we will inquire into later.’62 But what the image of the scales and the comparison with the perceptual case suggests is that just as in the perceptual case, we often correct the illusion by placing ourselves in the same perceptual relation to the two objects being compared, so in the evaluative case, we must place ourselves in the same cognitive relation to the two pleasures being compared. This will forestall a tendency to represent the proximate pleasure differently from how we represent the distant one, such that it appears greater than it actually is. A second concern is whether this will, in fact, work. Earlier, I raised a problem for the non-­rational interpretation: what guarantee is there that knowledge will escape the distorting effects of non-­rational desires? But one might raise a similar difficulty for the traditional interpretation: what guarantee is there that once we correctly assess the relative magnitudes of the proximate and distant pleasures, we will cease to be taken in by the appearance of the proximate pleasure as greater? In response, both the non-­rational and the traditional interpretations can appeal to the nature of knowledge: to have knowledge is to have a stable grasp of the truth. Thus, as a matter of definition, the one with knowledge will not be taken in by appearances. But the traditional interpretation has an advantage in terms of explaining why this is so: whereas it is difficult to see how, exactly, knowledge disempowers non-­rational desires, it is relatively uncontroversial that when presented with compelling evidence, we will abandon a belief that conflicts with it. It might take a while to get there: the cave dwellers are initially strongly persuaded by the shadows among which they dwell; eventually, however, they are compelled by the sight of truth to revise their beliefs (Rep. 515 c 4–516 c 6). In the case of hedonic errors, this process might be assisted by coming to realize the ways in which appearances tend to manipulate us. Once we become aware, through repeated experience, of the tendency of proximate pleasures 62 357 b 5–6: Ἥτις μὲν τοίνυν τέχνη καὶ ἐπιστήμη ἐστὶν αὕτη, εἰς αὖθις σκεψόμεθα.

72

Suzanne Obdrzalek

to appear inflated, then we will become less likely to be taken in by them.63 7.  Why do proximate pleasures appear greater? This brings us to our final question: what resources does Socrates have to explain why proximate pleasures appear greater than distant ones? Socrates does not offer an explanation of this phenomenon; he simply treats it as a brute fact. Proponents of the non-­rational interpretation find this unsatisfying; they, therefore, propose non-­ rational desires as the means to fill in this lacuna in his account. But before we consider what sort of explanation we might offer on Socrates’ behalf, we should first ask whether Socrates might not, after all, be entitled to treat the tendency of proximate pleasures to appear greater as a basic psychological fact. In fact, there is considerable empirical evidence that people often commit precisely this sort of evaluative error. Psychologists have repeatedly demonstrated that our preferences for rewards over time tend to shift according to hyperbolic and not exponential curves.64 The rate at which we discount future rewards does not remain consistent over time; instead, as a reward becomes more proximate, we tend to inflate its value more rapidly. Thus, for example, people will sometimes choose a short period of relief from a noxious noise over a longer later period of relief from that noise, but only if the shorter period is imminent.65 However, this does not settle the question of whether our preference for prox­im­ate pleasures is due to a calculative error; perhaps, as on the common account, we simply desire them more. But in a recent study, Kassam, Gilbert, Boston, and Wilson set out to answer exactly that question. Instead of presenting their subjects with an intertem­ poral choice, they asked them to predict how happy they would be now to receive $20 now versus how happy they would be at a series of future times to receive the money at those future times.66 63 Cf. Rep. 10, 595 b 5–7. 64  See e.g. G. Ainslie, Breakdown of Will (Cambridge, 2001). 65 Ainslie, Breakdown of Will, 30. This result has been replicated in animals: pigeons will choose earlier access to less grain over later access to more grain, but only when the earlier access is imminent and not when it is delayed. 66  In drawing upon this study, I am, admittedly, assuming that the results would remain consistent if the question were posed in terms of future pleasure rather than future happiness.



Evaluative Illusion in Plato’s Protagoras

73

The subjects consistently predicted that receiving $20 now would make them happier now than how happy receiving $20 at some future time would make them at that future time. The authors conclude: The standard account of temporal discounting suggests that there is something wrong with people’s decisions about the future but nothing wrong with their perceptions of it. Our studies show that shortsighted perception is, in fact, one of the causes of time discounting. One reason why people appear to be so unconcerned about the future is that they mispredict how they will feel when it arrives.67

What this suggests is that the sort of evaluative illusion that Socrates describes in the Protagoras is, in fact, fairly widespread, and hence, even if he cannot offer a further explanation for what causes it, perhaps we ought not to demand this of his account. I have to confess that this leaves me unsatisfied. In the perceptual case, we can appeal to various facts about optics to explain the illusion, and this seems to call out for a parallel explanation of the evaluative illusion. Granting that Socrates does not offer any such an explanation, what we need to ask is whether, as the proponents of the non-­rational interpretation claim, the only possible ex­plan­ ation is non-­rational desires. I have argued that non-­rational desires, even if they exist, would not do much to explain the evalu­ative illusion. I shall now propose an alternative explanation of the evaluative illusion. I should be very clear that I am not claiming that Socrates offers this further explanation. I am simply arguing that we do not need to appeal to non-­rational desires to provide such an ex­plan­ ation, that there are resources within the traditional interpretation of Socrates’ position to account for the evaluative illusion. To see this, we should return, briefly, to the psychological research I alluded to earlier. According to construal level theory, we tend to represent proximate events differently from how we represent ­distant ones.68 While we represent proximate events with a high 67  K. A. Kassam, D. T. Gilbert, A. Boston, and T. D. Wilson, ‘Future Anhedonia and Time Discounting’ [‘Future Anhedonia’], Journal of Experimental Social Psychology, 44 (2008), 1533–7 at 1536. 68  E.g. N. Liberman, M. Sagristano, and Y. Trope, ‘The Effect of Temporal Distance on Level of Mental Construal’, Journal of Experimental and Social Psychology, 38 (2001), 523–34. This difference in representation is reflected in the neural systems at work: when we select proximate rewards, the limbic system is activated, while when we select distant rewards, there is increased activity of the

74

Suzanne Obdrzalek

degree of specificity and detail, we represent distant events more schematically and abstractly. Kassam et al. offer the following account of how this might affect our hedonic predictions: since we represent proximate pleasures with more specificity and concrete detail than distant ones, our representations of them cause more intense affect. And since people often use the affect they experience upon representing a future pleasure to predict its magnitude, this leads to an overestimation of proximate pleasures.69 If we apply this to our problem in the Protagoras, the proposed ex­plan­ation is that we tend to represent proximate pleasures with more specificity and concrete detail than distant ones.70 This causes greater anticipatory pleasure, and since we use the anticipatory pleasure we experience upon representing a pleasure to forecast the amount of pleasure we will experience, this causes us to overestimate the magnitude of the proximate pleasure.71 lateral prefrontal cortex and the posterior parietal cortex (S. M. McClure, D. I. Laibson, G. Loewenstein, and J. D. Cohen, ‘Separate Neural Systems Value Immediate and Delayed Monetary Rewards’, Science, 306 (2004), 503–7). 69  Kassam, Gilbert, Boston, and Wilson, ‘Future Anhedonia’, 1533. In a later iteration of the marshmallow experiment, W. Mischel and N. Barker demonstrate a related and highly suggestive result (‘Cognitive Appraisals and Transformations in Delay Behavior’, Journal of Personality and Social Psychology, 31 (1975), 254–61). If the children were instructed to think of the rewards in consummatory, arousing (‘hot’) imagery, they capitulated significantly faster than if they were instructed to think of the rewards in non-­consummatory, abstract (‘cold’) imagery. Thus, for example, if the rewards in question were pretzel sticks, the children capitulated faster if they were instructed to think about how crunchy pretzels are when you bite into them and how salty they are when you lick them than if they were instructed to think about how pretzels are brown, long, and cylindrical (257). This demonstrates that one factor influencing how likely we are to pursue a reward is whether we represent it via imagery related to enjoying the consumption of it. 70  D. Parfit also attributes this view to Plato (Reasons and Persons (Oxford, 1987), 162); he does not offer a defence of this as an interpretation of Plato but, rather, sets this up as a theory that cannot fully account for the phenomenon of future discounting. 71  One might wonder whether this account of evaluative error depends on ethical hedonism or whether it could be applied to other conceptions of the good. Whether it can be applied to other kinds of goods will depend on whether we tend to forecast the magnitude of such goods on the basis of vivid, imagistic representations of ourselves experiencing such goods. Thus, this account will work best for goods whose value lies in our experience of them, less well for goods whose value is independent of our experience of them. At Phileb. 40 e 22–3, Plato extends his analysis of anticipatory pleasure to ‘fear, anger, and everything of that sort’ (περὶ ϕόβων τε καὶ θυμῶν καὶ πάντων τῶν τοιούτων). Thus, at the very least, we could extend this analysis to all the pathēmata that the many list as causes of akrasia at Prot. 352 b 7–8: anger, pleasure, pain, erōs, and fear. That this account could not be applied to goods whose value is independent of our experience of them does not seem



Evaluative Illusion in Plato’s Protagoras

75

Given the role of anticipatory pleasure in causing us to pursue the proximate pleasure, is it still fair to say that knowledge is not dragged around by pleasure, as on the common account? Yes, for two reasons. First, on the common account, what it means for knowledge to be dragged around is for the agent to pursue the proximate pleasure, despite knowing that it would be better not to do so. By contrast, on my proposed interpretation, at the moment of action, the agent acts in accordance with her erroneous belief that the proximate pleasure is better. Second, on the common account, the agent is directly caused to act by pleasure or, perhaps, a non-­rational desire for pleasure. By contrast, on my in­ter­pret­ation, what causes the agent to act is her belief that the proximate pleasure is greater and hence better. This belief is affected by her mental representation of the proximate pleasure and the resultant anticipatory pleasure, but it is ultimately her belief that it would be best to pursue the proximate pleasure which determines how she acts.72 As I have noted, Socrates simply does not offer any explanation of the evaluative error in the Protagoras. Nonetheless, the ex­plan­ ation I am proposing is clearly consistent with his presentation of the evaluative error: it explains why the appearance we form of the proximate pleasure misleads us into thinking it is greater than it actually is. Furthermore, if we look forward to the Philebus, we see some indications that Socrates might not be averse to this ex­plan­ ation. In the Philebus, Plato again observes that proximate pleasures appear greater than distant ones; his discussion is complicated by the fact that he blends into his treatment of temporal discounting another phenomenon—­namely that proximity to pain causes us to overestimate the magnitude of pleasure.73 For our purposes, problematic, as we are much less likely to show present bias in the case of such goods, if indeed they exist. For example, it seems probable that people are less likely to show present bias in pursuing posthumous goods; if that is the case, then this may be due to the fact that they do not imagine themselves enjoying those goods when deciding what to do. 72  Note that this interpretation differs from the non-­rational interpretation insofar as it claims that it is our representation of the proximate pleasures that causes us to erroneously form an inflated judgement of their magnitudes and not, instead, a non-­rational desire for these pleasures. 73  S. Delcomminette takes the minority view that temporal distortion is not at issue, only distortion due to the comparison of pleasure to pain (‘False Pleasures, Appearance and Imagination in the Philebus’ [‘False Pleasures’], Phronesis, 48 (2003), 215–37 at 231–2).

76

Suzanne Obdrzalek

what is significant is the following: first, Plato describes the vehicle by which we experience anticipatory pleasure as an appearance (ϕάντασμα). Accompanying the assertion (λόγος) in our soul that describes a future pleasure—­for example, that I will be rich—­is an appearance, which Plato depicts as an image painted in our soul (40 a 6–12). Plato emphasizes that it is through this appearance that we experience anticipatory pleasure: he invokes the painted images at 40 a 9–12 as part of his analysis of anticipatory pleasure and describes the anticipatory pleasure as involving envisioning (ὁρᾷ) oneself taking great pleasure in the possession of heaps of gold.74 Why does Plato need to introduce an appearance here in addition to the assertion? Because we don’t experience anticipatory pleasure when we dispassionately consider propositions about the future; we have to actively imagine ourselves undergoing the pleasure.75 Thus, the appearance brings with it the level of vivacity and immersive detail required for us to experience some version of the future pleasure now.76 Second, in the Philebus, Plato considers two cases of false an­tici­pa­tory pleasure. In the first, the judgement (δόξα) infects the 74  There is some controversy over how, exactly, to understand Plato’s treatment of anticipatory pleasure: does he take it to be the anticipation of future pleasure, pleasure taken in the anticipation of future pleasure, or, as Harte puts it, ‘an advance installment of the pleasure anticipated’ (‘Philebus’, 125)? I favour the last option, in part because, as Harte (n. 12) and S. Delcomminette (Le Philèbe de Platon [Philèbe] (Leiden, 2006), 384) observe, this is suggested by Plato’s use of προχαίρειν at 39 d 4. For a contrary perspective, see A. W. Price, ‘Varieties of Pleasure in Plato and Aristotle’, Oxford Studies in Ancient Philosophy, 52 (2017), 177–208 at 180–2. Delcomminette’s suggestion that in experiencing anticipatory pleasure, one identifies with the representation contained within the phantasma of oneself being pleased (Philèbe, 386) goes some way to addressing Price’s concern that Plato is depicting acentral imagining. 75  See Delcomminette, Philèbe, 385 and J.  Moss, ‘Pictures and Passions in the Timaeus and Philebus’, in R. Barney, T. Brennan, and C. Brittain (eds.), Plato and the Divided Self (Cambridge, 2012), 259–80 at 266. Delcomminette, ‘False Pleasures’, 228, and Lorenz, Brute, 105–6, emphasize the role of phantasmata in representing future pleasures when perception is not available. 76  I understand vivacity to refer not only to the content of the representation—­its concrete detail—­but also to its mode of presentation. Part of what is distinctive about phantasmata is that they present their content in an immersive, quasi-­ perceptual manner; this is the key to their verisimilitude. Plato alludes to this in his account of the manipulative power of poetry in the Republic: poets’ words only seem true to us when clothed in metre, rhythm, and harmony; stripped of these, they lose their power of bewitchment, like a boy who has lost the bloom of youth (601 a 4–b 7). Thus, it is not so much the propositional content of poetry as its mode of presentation that accounts for its ability to deceive.



Evaluative Illusion in Plato’s Protagoras

77

an­tici­pa­tory pleasure: a mistaken judgement that one will be rich renders the anticipatory pleasure one feels false.77 But in the second case, the causal direction switches (42 a 5–9). Plato explicitly claims that when we overestimate the amount of pleasure we will feel—­either due to temporal proximity or the proximity of pain—­ the falsehood lies in the pleasure itself. In claiming that the falsehood lies in the pleasure itself and not the judgement, Plato suggests that it is the appearance that is at fault. How does it deceive us? As in the Protagoras, Plato explains the hedonic error by a comparison with visual distance illusions: ‘Does it happen only to eyesight that seeing objects from afar or close by distorts the truth and causes false judgements? Or does not the same thing happen also in the case of pleasure and pain?’78 Just as the way that objects appear to vision can cause us to misestimate their size, so the way that pleasures appear in advance—­that is, the appearances we form of them—­can cause us to misestimate their magnitude. In the visual case, the objects look larger or smaller than they actually are; similarly, in the hedonic case, the pleasures must appear greater or lesser than they will be. Plato goes on to refer to the amount of excessive an­tici­pa­tory pleasure we feel as an appearance (ϕαινόμενον) that should be cut off as false (42 b 8–c 3). This suggests that our representation of ourselves enjoying future pleasures contains a certain quantity of pleasure and, furthermore, that this quantity can be distorted. Thus, the appearance renders judgement false because we use the quantity of anticipatory pleasure we experience to predict how great a future pleasure will be. In sum, it is the image we form of ourselves as pleased through which we experience an­tici­ pa­tory pleasure; the amount of anticipatory pleasure we feel, in turn, is the basis of our false hedonic prediction.79 77  There is considerable debate over the content of this false judgement: possibilities include the judgement that a certain event will occur, that it will occur under a certain description, and that it will be pleasant should it occur; for helpful discussion, see Harte ‘Philebus’. For my purposes, all that matters is that there is some judgement made about what the future will be like and that it is this judgement which renders the anticipatory pleasure false. 78  Phileb. 41 e 9–42 a 3: ἐν μὲν ὄψει τὸ πόρρωθεν καὶ ἐγγύθεν ὁρᾶν τὰ μεγέθη τὴν ἀλήθειαν ἀϕανίζει καὶ ψευδῆ ποιεῖ δοξάζειν, ἐν λύπαις δ’ ἄρα καὶ ἡδοναῖς οὐκ ἔστι ταὐτὸν τοῦτο γιγνόμενον; 79 For a full defence of this interpretation of Phileb. 41 a 5–42 c 5, see N. Mooradian, ‘What to Do about False Pleasures of Overestimation? Philebus 41 a 5–42 c 5’, Apeiron, 28 (1995), 91–112. Our interpretations of the passage differ insofar as I take the false hedonic prediction to derive from the amount of pleasure

78

Suzanne Obdrzalek

Of course, this does not get us all the way to the account that Kassam et al. offer of the evaluative illusion. In particular, Plato nowhere claims that we tend to represent proximate pleasures more vividly than distant ones. But what we do get from our con­sid­er­ ation of the Philebus is not insignificant. It makes clear that Plato conceives of our representations of future pleasures as imagistic appearances and that it is in virtue of the anticipatory pleasure we feel through these appearances that we forecast the amount of pleasure we will experience.80

8. Conclusion In this paper, I have sought to defend the traditional interpretation of the Protagoras against the non-­rational interpretation. The non-­ rational interpretation, importantly, draws attention to the need to explain why Socrates should hold that proximate pleasures appear greater than distant ones. However, introducing non-­rational desires into Socrates’ account does not, I argue, solve this difficulty, and, furthermore, it generates additional problems. I then offer a close examination of Socrates’ account of the evaluative illusion. I claim that in the Protagoras, appearances are vivid and imagistic mental representations that appear true and hence invite assent, but that contained in one’s representation of the future pleasure, whereas Mooradian treats the anticipatory pleasure as an additional cognitive state that is directed at one’s representation of the future pleasure but is not identical to it. J. Moss, Aristotle on the Apparent Good (Oxford, 2012), offers a similar treatment of the role of phantas­ mata in Aristotle’s account of practical deliberation. 80  Storey offers a different account of the source of the error in the Protagoras. According to Storey, the error of the many is not to falsely measure the relative magnitudes of the pleasures, but to fail to recognize that the later goods against which they measure the imminent pleasures are also pleasure-­producing (‘Sex, Wealth, and Courage’, 252–4; see also Nussbaum, Fragility, 115–16, who makes the case that the many fail to recognize the commensurability of the goods being weighed). While I do not agree with this aspect of Storey’s analysis (see Section 1, n. 10 above), his observation concerning the kinds of goods being compared is significant. The immediate pleasures are things like food, drink, and sex, the later goods things like wealth and health. One confronted with an opportunity to eat a cake might immediately and vividly represent to himself the pleasure of eating the cake. However, as Storey notes, it requires instrumental reasoning to conclude that eating the cake will lead to eventual pain via ill health. The fact that a more complex reasoning process is involved in representing the later evil as unpleasant may explain why we are less likely to represent it imagistically and so undergo anticipatory pain.



Evaluative Illusion in Plato’s Protagoras

79

have a tendency to be false. I conclude by suggesting that prox­im­ ate pleasures have a tendency to produce inflated hedonic predictions because we represent them more vividly than distant ones, which yields greater anticipatory pleasure which, in turn, causes us to overestimate their magnitude. This final proposal is, of course, speculative. My aim is not to claim that Socrates had this sort of account in mind but, rather, to demonstrate that he had the resources to develop his account along these lines. Some will find this explanation unsatisfactory.81 It might seem dubious that it could fully account for the marshmallow case with which I opened this paper. Could the children really forecast less pleasure when the temporal gap is a mere fifteen minutes? Perhaps, though, that case is complicated by the fact that children tend to experience time differently from how adults do; thus, while fifteen minutes might seem a short stretch of time to an adult, it might seem an eternity to a child.82 This account also forces us to accept certain phenomena as brute facts, namely that we represent imminent pleasures more vividly and that this causes us to over­ esti­mate their magnitude.83 But in the end, every ex­plan­ation requires us to accept certain phenomena as given; the question is whether these phenomena seem implausible and mysterious or whether we are comfortable accepting them. Part of what is distinctive about Socrates’ intellectualism is that he does not treat desire as given; he 81  Parfit argues powerfully against Plato that even if the vividness of our representations of proximate pleasures accounts for some of our present bias, it cannot fully account for the phenomenon (Reasons and Persons, 161–2). 82  It is generally held that children do not have a solid grasp of time until around the age of 7 (see e.g. S. Droit-­Volet, ‘Children and Time’, Psychologist, 25 (2012), 586–9). As regards hedonic forecasting (rather than intertemporal choice), I am not aware of any studies that focus on such short temporal distances. However, it is noteworthy that in Kassam et al.’s study, 43% of the discounting occurred at a gap of a mere twenty-­four hours (‘Future Anhedonia’, 1535). 83  One might attempt to explain this fact in a variety of ways. Generally, we have more detailed and reliable information about proximate, rather than distant pleasures, which would account for a tendency to represent proximate goods in more vivid detail (see e.g. K. Fujita, Y. Trope, and N. Liberman, ‘On the Psychology of Near and Far: A Construal Level Theoretic Approach’, in G.  Keren and G.  Wu (eds.), The Wiley Blackwell Handbook of Judgment and Decision Making (Chichester, 2016), 404–31 at 405–6). Furthermore, more detailed representations might be of more use to us regarding proximate goods than distant ones, given that we are more likely to have to act soon in relation to proximate than distant goods, and low-­level construals are more germane to action. It is possible that both of these general facts produce a tendency to represent proximate goods with greater specificity and detail than distant ones, even in cases where these facts do not, in fact, obtain.

80

Suzanne Obdrzalek

views desire as reasons-­responsive, and when someone claims to desire something, he maintains that it is reasonable to ask why. At the same time, Socrates does recognize that our cognitive mechanisms can become distorted. Thus, at the very opening of the d ­ ialogue, Socrates warns Hippocrates that the danger of going to the sophist is that one’s very reasoning capacity will be perverted through his teachings, so much so, that one will become incapable of recognizing the same teachings’ falsity (313 c 7–314 b 4). What recourse is there? How can we prevent this distortion from occurring? Here, we might call to mind Socrates’ comment in Republic 10, that ­imitative poetry has the power to ruin one’s intellect unless one possesses the antidote of knowing what it is (595 b 5–7). Thus, in the case of the sophist, the solution—­apart from avoiding him altogether—­is perhaps to be taught to recognize the distortions he threatens to introduce into our reasoning. Similarly in the case of pleasure: in highlighting the tendency of proximate pleasures to appear inflated, Socrates is, perhaps, en­ab­ling us to safeguard ourselves against being deceived. Claremont McKenna College

BIBLIOGR A PH Y Ainslie, G., Breakdown of Will (Cambridge, 2001). Beere, J., ‘Faking Wisdom: The Expertise of Sophistic in Plato’s Sophist’, Oxford Studies in Ancient Philosophy, 57 (2019), 153–90. Boeri, M. D., ‘Socrates, Aristotle, and the Stoics on the Apparent and Real Good’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, 20 (2004), 109–41. Brickhouse, T. C. and Smith, N. D., ‘Socrates on akrasia, Knowledge, and the Power of Appearance’ [‘Socrates on akrasia’], in C. Bobonich and P. Destrée (eds.), Akrasia in Greek Philosophy (Leiden, 2007), 1–17. Brickhouse, T.  C. and Smith, N.  D., Socratic Moral Psychology [Moral Psychology] (Cambridge, 2010). Callard, A. G., ‘Ignorance and akrasia-Denial in the Protagoras’, Oxford Studies in Ancient Philosophy, 47 (2014), 31–80. Carone, G. R., ‘Calculating Machines or Leaky Jars? The Moral Psychology of Plato’s Gorgias’, Oxford Studies in Ancient Philosophy, 26 (2004), 55–96. Clark, J., ‘The Strength of Knowledge in Plato’s Protagoras’, Ancient Philosophy, 32 (2012), 237–55.



Evaluative Illusion in Plato’s Protagoras

81

Colbert, S., The Colbert Report (17 October 2005). Cooper, J. M. (ed.), Plato: Complete Works (Indianapolis, 1997). Delcomminette, S., ‘False Pleasures, Appearance and Imagination in the Philebus’ [‘False Pleasures’], Phronesis, 48 (2003), 215–37. Delcomminette, S., Le Philèbe de Platon [Philèbe] (Leiden, 2006). Denyer, N., Plato: Protagoras (Cambridge, 2008). Devereux, D., ‘Socrates’ Kantian Conception of Virtue’ [‘Kantian Conception’], Journal of the History of Philosophy, 33 (1995), 381–408. Dover, K.  J., Greek Popular Morality in the Time of Plato and Aristotle (Indianapolis, 1994). Droit-Volet, S., ‘Children and Time’, Psychologist, 25 (2012), 586–9. Dyson, M., ‘Knowledge and Hedonism in Plato’s Protagoras’ [‘Knowledge and Hedonism’], Journal of Hellenic Studies, 96 (1976), 32–45. Ferrari, G. R. F., ‘Akrasia as Neurosis in Plato’s Protagoras’ [‘Akrasia as Neurosis’], Proceedings of the Boston Area Colloquium in Ancient Philosophy, 6 (1990), 115–39. Frede, M. ‘Introduction’, in S.  Lombardo and K.  Bell (trans.), Plato: Protagoras (Indianapolis, 1992), vii–­xxxiv. Fujita, K., Trope, Y., and Liberman, N., ‘On the Psychology of Near and Far: A Construal Level Theoretic Approach’, in G. Keren and G. Wu (eds.), The Wiley Blackwell Handbook of Judgment and Decision Making (Chichester, 2016), 404–31. Gallop, D., ‘The Socratic Paradox in the Protagoras’, Phronesis, 9 (1964), 117–29. Harte, V., ‘The Philebus on Pleasure: The Good, the Bad and the False’ [‘Philebus’], Proceedings of the Aristotelian Society, 104 (2004), 113–30. Irwin, T., Plato’s Ethics [Ethics] (Oxford, 1995). Irwin, T., Plato’s Moral Theory (Oxford, 1977). Jones, R., ‘Review of Thomas  C.  Brickhouse and Nicholas  D.  Smith, Socratic Moral Psychology’, Polis, 28 (2011), 147–52. Kamtekar, R., Plato’s Moral Psychology [Moral] (Oxford, 2017). Kassam, K.  A., Gilbert, D.  T., Boston, A., and Wilson, T.  D., ‘Future Anhedonia and Time Discounting’ [‘Future Anhedonia’], Journal of Experimental Social Psychology, 44 (2008), 1533–7. Liberman, N., Sagristano, M., and Trope, Y., ‘The Effect of Temporal Distance on Level of Mental Construal’, Journal of Experimental and Social Psychology, 38 (2001), 523–34. Lorenz, H., The Brute Within [Brute] (Oxford, 2006). Martin, V. and Budé, G. de (eds. and trans.), Aeschines: Discours (Paris, 2002). McClure, S.  M., Laibson, D.  I., Loewenstein, G., and Cohen, J.  D., ‘Separate Neural Systems Value Immediate and Delayed Monetary Rewards’, Science, 306 (2004), 503–7.

82

Suzanne Obdrzalek

Mischel, W. and Barker, N., ‘Cognitive Appraisals and Transformations in Delay Behavior’, Journal of Personality and Social Psychology, 31 (1975), 254–61. Mischel, W. and Ebbesen, E.  B., ‘Attention in Delay of Gratification’ [‘Attention’], Journal of Personality and Social Psychology, 16 (1970), 329–37. Mischel, W., Ebbesen, E. B., and Raskoff Zeiss, A., ‘Cognitive and Attentional Mechanisms in Delay of Gratification’, Journal of Personality and Social Psychology, 21 (1972), 204–18. Mischel, W., Shoda, Y., and Rodriguez, M., ‘Delay of Gratification in Children’, Science, 244 (1989), 933–8. Mooradian, N., ‘What to Do about False Pleasures of Overestimation? Philebus 41 a 5–42 c 5’, Apeiron, 28 (1995), 91–112. Moss, J., ‘Appearances and Calculation: Plato’s Division of the Soul’, Oxford Studies in Ancient Philosophy, 34 (2008), 35–68. Moss, J., Aristotle on the Apparent Good (Oxford, 2012). Moss, J., ‘Pictures and Passions in the Timaeus and Philebus’, in R. Barney, T. Brennan, and C. Brittain (eds.), Plato and the Divided Self (Cambridge, 2012), 259–80. Moss, J., ‘Pleasure and Illusion in Plato’, Philosophy and Phenomenological Research, 72 (2006), 503–35. Nussbaum, M. C., The Fragility of Goodness [Fragility] (Cambridge, 1999). Parfit, D., Reasons and Persons (Oxford, 1987). Penner, T., ‘Plato and Davidson: Parts of the Soul and Weakness of Will’ [‘Plato and Davidson’], Canadian Journal of Philosophy, 16 (1990), 35–74. Penner, T., ‘Socrates on the Strength of Knowledge: Protagoras 351 b–357 e’ [‘Socrates on the Strength of Knowledge’], Archiv für Geschichte der Philosophie, 79 (1997), 117–49. Price, A. W., Mental Conflict (London, 1995). Price, A. W., ‘Varieties of Pleasure in Plato and Aristotle’, Oxford Studies in Ancient Philosophy, 52 (2017), 177–208. Reshotko, N., Socratic Virtue: Making the Best of the Neither-Good-NorBad [Socratic Virtue] (Cambridge, 2006). Santas, G., ‘Plato’s Protagoras and Explanations of Weakness’, Philosophical Review, 75 (1966), 3–33. Schofield, M., ‘Aristotle on the Imagination’, in M.  C.  Nussbaum and A.  O.  Rorty (eds.), Essays on Aristotle’s De Anima (Oxford, 1992), 249–76. Segvic, H., ‘No One Errs Willingly: The Meaning of Socratic Intellectual­ ism’, Oxford Studies in Ancient Philosophy, 19 (2000), 1–45. Silverman, A., ‘Plato on Phantasia’, Classical Antiquity, 10 (1991), 123–47.



Evaluative Illusion in Plato’s Protagoras

83

Singpurwalla, R., ‘Reasoning with the Irrational: Moral Psychology in the Protagoras’ [‘Reasoning with the Irrational’], Ancient Philosophy, 26 (2006), 243–58. Storey, D., ‘Sex, Wealth, and Courage: Kinds of Goods and the Power of Appearance in Plato’s Protagoras’ [‘Sex, Wealth, and Courage’], Ancient Philosophy, 38 (2018), 241–63. Taylor, C. C. W., Plato: Protagoras [Protagoras] (Oxford, 1991). Vlastos, G., ‘Socrates on Acrasia’ [‘Acrasia’], Phoenix, 23 (1969), 71–88. Warren, J., The Pleasures of Reason in Plato, Aristotle, and the Hellenistic Hedonists (Cambridge, 2014). Watson, G., ‘Skepticism about Weakness of Will’, Philosophical Review, 86 (1977), 316–39. Wolfsdorf, D., ‘The Ridiculousness of Being Overcome by Pleasure: Protagoras 352 b 1–358 d 4’ [‘Ridiculousness’], Oxford Studies in Ancient Philosophy, 31 (2006), 113–36.

THAT DIFFERENCE IS DIFFERENT FROM BEING Sophist 255 c 9–e 2 michael wiitala 1. Introduction As a survey of the secondary literature on Plato’s Sophist reveals, the argument by which the Eleatic Stranger distinguishes the kind being from the kind different (255 c 9–e 2) is one of the most ­controversial in the dialogue.1 Moreover, it is often considered one of the most important.2 The argument closes the portion of the dialogue in which the Stranger demonstrates that each of the five  ‘greatest kinds’ (μέγιστα γενή) is distinct from the others (254 d 4–255 e 2). Having shown that rest, motion, same, and different are distinct from one another and that rest, motion, and same are distinct from being, the Stranger argues that different is also distinct from being. The distinction between different and being is  a crucial step in the Stranger’s overall account of non-­being, according to which non-­being is ‘the contraposing of the nature of a part of different and the nature of being’ (ἡ τῆς θατέρου μορίου ϕύσεως I am indebted to Colin Smith, Paul DiRado, and Eric Sanday for their insightful and detailed written comments on earlier versions of this paper, to David Goldstein for his help with parsing the syntax of 255 d 4–6, and to Victor Caston for his considerable assistance in distilling the argument and bringing it to print. 1  For a survey of the various ways the argument is interpreted, see P. Crivelli, Plato’s Account of Falsehood: A Study of the Sophist [Plato’s Account of Falsehood ] (Cambridge, 2012), 142–9. 2  See M. Frede, Prädikation und Existenzaussage: Platons Gebrauch von ‘ . . . ist . . . ’ und ‘ . . . ist nicht . . . ’ im Sophistes [Prädikation und Existenzaussage] (Göttingen, 1967), 12–29; L.  Brown, ‘Being in the Sophist: A Syntactical Enquiry’ [‘Being in the Sophist’], in G. Fine (ed.), Plato 1: Metaphysics and Epistemology (Oxford, 1999), 455–78 at 474–7; M.  Frede, ‘Plato’s Sophist on False Statements’, in R.  Kraut (ed.), The Cambridge Companion to Plato (Cambridge, 1992), 397–424 at 400 ff.; A. Silverman, The Dialectic of Essence: A Study of Plato’s Metaphysics [Dialectic of Essence] (Princeton, 2002), 164; M.  L.  Gill, ‘Method and Metaphysics in Plato’s Sophist and Statesman’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2016 edn), §§ 5.3–5.4, https://plato.stanford.edu/archives/win2016/entries/ plato-sophstate/, accessed 14 January 2023; cf. J. Malcolm, ‘A Way Back for Sophist 255 c 12–13’ [‘A Way Back’], Ancient Philosophy, 26 (2006), 275–89. Michael Wiitala, That Difference Is Different from Being: Sophist 255 c 9–e 2 In: Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Michael Wiitala 2023. DOI: 10.1093/oso/9780192885180.003.0003

86

Michael Wiitala

καὶ τῆς τοῦ ὄντος . . . ἀντίθεσις, 258 a 11–b 1).3 Furthermore, it is in the argument distinguishing different and being that the Stranger identifies the features of these forms that make the contraposing of a part of different and being possible. Beings, we are told, are themselves according to themselves (αὐτὰ καθ’ αὑτά) or relative to others (πρὸς ἄλλα, 255 c 13–14),4 whereas anything different is always relative to something different from it (255 d 1, 255 d 6–7). Although what these features of being and different amount to is a matter of considerable controversy, they are fundamental to the Stranger’s account of being and non-­being. While commentators have developed many careful in­ter­pret­ ations of the argument distinguishing different from being, deciding between those interpretations is difficult. Due to key ambiguities in the text, no interpretation can claim to be the only plausible one, and each comes at a cost. The main disagreement concerns the distinction between beings that are auta kath’ hauta, ‘themselves according to themselves’, and beings that are pros alla, ‘relative to others’. Following John Malcolm, I will refer to this distinction with the acronym KH/PA, for the two Greek phrases used to express it.5 There are four main reasons this distinction has become central to scholarly disputes. First, as already noted, it plays a crucial role in the Stranger’s account of being and non-­being. Second, the Stranger has neither discussed nor put forward anything like the KH/PA distinction earlier in the dialogue.6 Third, the distinction has no precise parallel elsewhere in Plato.7 Finally, it seems that the immediate context in which the Stranger introduces the distinction offers no clear indication of whether or not it is intended to be

3  All translations are my own unless otherwise indicated. There is some debate on how to construe 258 a 11–b 1; see Crivelli, Plato’s Account of Falsehood, 216; J. van Eck, ‘Not-Being and Difference: On Plato’s Sophist, 256 d 5–258 e 3’, Oxford Studies in Ancient Philosophy, 23 (2002), 68–84 at 77–8. 4  I use the text and line references of E. A. Duke, W. F. Hicken, W. S. M. Nicoll, D. B. Robinson, and J. C. G. Strachan (eds.), Platonis Opera, vol. i: Euthyphro, Apologia, Crito, Phaedo, Cratylus, Theaetetus, Sophista, Politicus (Oxford, 1995). But I correct for the typographical error at 255 c 13–14, where 255 c 14 is mislabelled as 255 c 15. 5  Malcolm, ‘A Way Back’, 275. 6  Cf. Crivelli, Plato’s Account of Falsehood, 147–8. 7  Commentators have identified a number of passages that seem relevant to the KH/PA distinction, but there are no precise parallels (see esp. Malcolm, ‘A Way Back’, 283–4; F. Leigh, ‘Modes of Being at Sophist 255 c–­e’ [‘Modes of Being’], Phronesis, 57 (2012), 1–28 at 12–15). For some relevant passages, see Plato, Theaet. 152 d, 156 e–157 b, 160 b; Phileb. 51 c; Chrm. 168 b–169 a; Rep. 4, 438 a–­d; Sym. 211 a–­b; Phaedo 78 d; Tim. 51 c.



That Difference Is Different from Being

87

exhaustive,8 or to what sort of things it is intended to apply. Readers are left wondering whether it is a metaphysical distinction, a distinction between kinds of terms, a syntactic distinction, a semantic distinction, a distinction in sentence structure, or some combin­ ation of these.9 Since the immediate context in which the KH/PA distinction is introduced seems to offer little or no guidance, commentators are forced to resort to controversial in­ter­pret­ive principles or to what they take to be Plato’s broader concerns in the Sophist in order to decide what sort of distinction it is. Although commentators have propounded many ways to interpret the distinction, one question that has not been explored is the placing of the comma in the counterfactual conditional at 255 d 4–6. Yet given that the Stranger only explicitly employs the KH/PA distinction in this conditional, how this conditional is construed is key to interpreting the distinction correctly. Since the construal of the conditional at 255 d 4–6 is what is in question here, I will first present the Greek on its own without the comma, before proceeding to the alternative translations: εἴπερ θάτερον ἀμϕοῖν μετεῖχε τοῖν εἰδοῖν ὥσπερ τὸ ὂν ἦν ἄν ποτέ τι καὶ τῶν ἑτέρων ἕτερον οὐ πρὸς ἕτερον (255 d 4–6).

If the comma is placed after ὥσπερ τὸ ὄν so that the phrase is read with the protasis, the following translation results: But if different participated in both forms, as being does, then there would sometimes also among the differents be something different not relative to something different. 8  See Malcolm, ‘A Way Back’, 282; cf. Frede, Prädikation und Existenzaussage, 19, 27, 36; Silverman, Dialectic of Essence, 165. 9  For accounts according to which the distinction is syntactic and/or semantic, although often with metaphysical implications, see A.  L.  Peck, ‘Plato and the ΜΕΓΙΣΤΑ ΓΕΝΗ of the Sophist: A Reinterpretation’, The Classical Quarterly, n.s., 2 (1952), 32–56 at 48; A.  L.  Lacey, ‘Plato’s Sophist and the Forms’, Classical Quarterly, n.s., 9 (1959), 43–52 at 49 n. 1; J. M. E. Moravcsik, ‘Being and Meaning in the Sophist’, Acta Philosophica Fennica, 14 (1962), 23–78; Frede, Prädikation und Existenzaussage; G. E. L. Owen, ‘Plato on Not-Being’, in G. Vlastos (ed.), Plato: A Collection of Essays, vol. i: Metaphysics and Epistemology (Garden City, 1971), 223–67 at 255–8; D. Bostock, ‘Plato on “Is-Not” (Sophist 254–9)’, Oxford Studies in Ancient Philosophy, 2 (1984), 89–119 at 92–4; C. D. C. Reeve, ‘Motion, Rest, and Dialectic in the Sophist’ [‘Motion, Rest, and Dialectic’], Archiv für Geschichte der Philosophie, 67 (1985), 47–64 at 54–5; Brown, ‘Being in the Sophist’; Malcolm, ‘A Way Back’. For accounts according to which the distinction is metaphysical, see R. S. Bluck, Plato’s Sophist (Manchester, 1975), 145–50; M. M. McCabe, Plato’s Individuals (Princeton, 1994), 233; D. Ambuel, Image and Paradigm in Plato’s Sophist (Las Vegas, 2007), 150–1; Silverman, Dialectic of Essence, 164–81; M.  L.  Gill, Philosophos: Plato’s Missing Dialogue (Oxford, 2012), 164–6; Leigh, ‘Modes of

88

Michael Wiitala

If, in contrast, the comma is placed before ὥσπερ τὸ ὄν so that the phrase is read with the apodosis, the translation is: But if different participated in both forms, then, just as with being, there would sometimes also among the differents be something different not relative to something different.

Editors from Heindorf to Robinson place the comma in the first way, so that ὥσπερ τὸ ὄν is read with the protasis.10 I think this placement is mistaken, obscures the character of the KH/PA distinction, and in large part explains why the distinction has become so controversial. I will argue that placing the comma in the second way, so that ὥσπερ τὸ ὄν is read with the apodosis, allows for a more natural and straightforward interpretation of the argument as a whole and indicates that the KH/PA distinction is metaphysical, rather than syntactic or semantic. 2.  Text, translation, and linguistic concerns Since determining how to interpret the KH/PA distinction and to which clause ὥσπερ τὸ ὄν should be adjoined requires a careful reading of 255 c 9–e 2 as a whole, I reproduce Robinson’s text below, omitting the comma at 255 d 5:

Being’; D. Ambuel, ‘Difference in Kind: Observations on the Distinction of the Megista Gene’, in B.  Bossi and T.  M.  Robinson (eds.), Plato’s Sophist Revisited (Berlin, 2013), 247–67 at 260–1. For a sentence structure account, see W. de Vries, ‘On Sophist 255 b–­e’, History of Philosophy Quarterly, 5 (1988), 385–94 at 390–2. For some combination of these, see S.  Rosen, Plato’s Sophist: The Drama of Original and Image [Plato’s Sophist] (New Haven, 1983), 270; R. M. Dancy, ‘The Categories of Being in Plato’s Sophist 255 c–­e’ [‘Categories of Being’], Ancient Philosophy, 19 (1999), 45–72; M.  Duncombe, ‘Plato’s Absolute and Relative Categories at Sophist 255 c 14’ [‘Absolute and Relative Categories’], Ancient Philosophy, 32 (2012), 77–86 at 77 n. 2; B. E. Hestir, Plato on the Metaphysical Foundation of Meaning and Truth [Plato on Meaning and Truth] (Cambridge, 2016), 163–4; C. C. Smith, ‘Against the Existential Reading of Euthydemus 283 e–284 c, with Help From the Sophist’, Ancient Philosophy, 42 (2022), 67–81 at 75 n. 18. 10  L. F. Heindorf (ed.), Platonis dialogi selecti, vol. iv (Berlin, 1809), 411; J. G. Baiter, J.  K.  von Orelli, and A.  W.  Winckelmann (eds.), Platonis opera quae feruntur omnia (Zurich, 1839), 127; G.  Stallbaum (ed.), Platonis Sophista (Gotha, 1840), 193; K. F. Hermann, Platonis dialogi secundum Thrasylli tetralogias dispositi, vol. i (Leipzig, 1851), 400; L.  Campbell, The Sophistes and Politicus of Plato (Oxford, 1867), 152; O. Apelt (ed.), Platonis Sophista (Leipzig, 1897), 172; J. Burnet (ed.), Platonis opera, vol. i: Euthyphro, Apologia Socratis, Crito, Phaedo, Cratylus, Theaetetus, Sophista, Politicus (Oxford, 1900); H. N. Fowler (ed. and trans.), Plato: Theaetetus, Sophist, Loeb Classical Library 123 (Cambridge, MA, 1921), 408; A. Diès (ed. and trans.), Platon: Œuvres complètes, vol. viii. 3: Sophiste, 3rd edn (Paris, 1955); Duke et al., Platonis opera, vol. i, 449.



That Difference Is Different from Being

ΞΕ. Τί δέ; τὸ θάτερον ἆρα ἡμῖν λεκτέον πέμπτον; ἢ τοῦτο καὶ τὸ ὂν ὡς δύ’ ἄττα ὀνόματα ἐϕ’ ἑνὶ γένει διανοεῖσθαι δεῖ; ΘΕΑΙ. Τάχ’ ἄν. ΞΕ. Ἀλλ’ οἶμαί σε συγχωρεῖν τῶν ὄντων τὰ μὲν αὐτὰ καθ’ αὑτά, τὰ δὲ πρὸς ἄλλα11 ἀεὶ λέγεσθαι. ΘΕΑΙ. Τί δ’ οὔ; ΞΕ. Τὸ δέ γ’ ἕτερον ἀεὶ πρὸς ἕτερον∙ ἦ γάρ; ΘΕΑΙ. Οὕτως. ΞΕ. Οὐκ ἄν, εἴ γε τὸ ὂν καὶ τὸ θάτερον μὴ πάμπολυ διεϕερέτην∙ ἀλλ’ εἴπερ θάτερον ἀμϕοῖν μετεῖχε τοῖν εἰδοῖν ὥσπερ τὸ ὂν ἦν ἄν ποτέ τι καὶ τῶν ἑτέρων ἕτερον οὐ πρὸς ἕτερον∙ νῦν δὲ ἀτεχνῶς ἡμῖν ὅτιπερ ἂν ἕτερον ᾖ, συμβέβηκεν ἐξ ἀνάγκης ἑτέρου τοῦτο ὅπερ ἐστὶν εἶναι. ΘΕΑΙ. Λέγεις καθάπερ ἔχει. ΞΕ. Πέμπτον δὴ τὴν θατέρου ϕύσιν λεκτέον ἐν τοῖς εἴδεσιν οὖσαν οἷς προαιρούμεθα. ΘΕΑΙ. Ναί.

89 255 c 10

255 c 15 255 d 1

255 d 5

255 e 1

The following is my translation, with ὥσπερ τὸ ὄν adjoined to the apodosis: stranger: What then? Must the different be said by us to be a fifth [kind]? Or is it necessary to think of the different and being as two names for one kind? theaet.: Maybe. stranger: Yet I suppose you grant that, among beings, some are [always] said [to be] themselves according to themselves, while some are always said [to be] relative to others.12 theaet.: Certainly. stranger: But a different [is] always [said to be] relative to something different, right? theaet.: Just so. stranger: This would not be the case if being and the different were not entirely distinct. But if a different participated in both forms, then, just as with being, there would sometimes also among the differents be something different not relative to something different. Yet it is now obvious to us that whatever is different has, by necessity, turned out to be this, just what it is, from something different. theaet.: It is just as you say. stranger: Then the nature of a different must be said to be a fifth among the forms we are selecting. theaet.: Yes.

255 c

255 d

255 e

11  The T and W manuscripts read πρὸς ἄλλα, while the B and D family reads πρὸς ἄλληλα. Editors prefer πρὸς ἄλλα, in part because T and W represent independent manuscript families. Duncombe, however, has recently argued in defence of the πρὸς ἄλληλα reading (Duncombe, ‘Absolute and Relative Categories’). The exegesis I develop in this essay is compatible with either. 12  The sentence in 255 c 13–14 is ambiguous in various ways. One ambiguity concerns the demonstratives τὰ μέν and τὰ δέ. The λέγεσθαι indicates they are ‘things

90

Michael Wiitala

We can begin with a brief consideration of the linguistic concerns relevant to reading ὥσπερ τὸ ὄν in the conditional at 255 d 4–6. The conditional is comprised of three interrelated clauses: the protasis (ἀλλ’ εἴπερ θάτερον ἀμϕοῖν μετεῖχε τοῖν εἰδοῖν), the apodosis (ἦν ἄν ποτέ τι καὶ τῶν ἑτέρων ἕτερον οὐ πρὸς ἕτερον), and an adjunct comparative clause, ὥσπερ τὸ ὄν.13 The ὥσπερ τὸ ὄν clause is an adjunct in that it is not a linguistic argument in either the protasis or apodosis, that is to say, it is not syntactically required by either of them.14 In the case of ὥσπερ τὸ ὄν, as often happens, the verb and other features shared by the comparative clause and the leading clause to which it is adjoined have been omitted.15 Thus, we are left with the question of whether ὥσπερ τὸ ὄν is adjoined to the protasis or apodosis. If it is adjoined to the protasis, we have: ‘But if different participated in both forms, just as being [participates in both forms] . . . ’; if to the apodosis: ‘ . . . just as being [is sometimes not relative to something different], there would sometimes also among the differents be something different not relative to something different’. Either option is linguistically possible, and which Plato intends is not immediately obvious. One might be tempted to read ὥσπερ τὸ ὄν with the protasis simply because comparative clauses introduced by ὥσπερ occur after or within their leading clauses more frequently than they precede them. Yet given that comparative clauses with ὥσπερ do sometimes precede their leading clauses—­even in cases where the comparative clause has no verb and there is no cor­res­ pond­ing demonstrative adverb (e.g. οὕτως or ὧδε) in the leading clause16—to read ὥσπερ τὸ ὄν with the protasis without carefully

said’. But ‘things said’ could be words, phrases, or statements, or alternatively could be the things about which those words, phrases, or statements are said. I adopt the latter construal. Another ambiguity concerns the scope of ἀεί, which either governs both τὰ μὲν αὐτὰ καθ’ αὑτά and τὰ δὲ πρὸς ἄλλα, or only τὰ δὲ πρὸς ἄλλα. I take it to govern both. 13  For comparative clauses introduced by ὥσπερ, see H. W. Smyth, Greek Grammar, revised by G.  M.  Messing (Cambridge, MA, 1956), 2462–7; cf. R.  Martinez, ‘Comparative Clauses’, Encyclopedia of Ancient Greek Language and Linguistics, 24 September 2013. doi: 10.1163/2214-448X_eagll_SIM_00000431. 14  E. Crespo, ‘Adjuncts’, Encyclopedia of Ancient Greek Language and Linguistics, 24 September 2013. doi: 10.1163/2214-448X_eagll_SIM_00000413. 15 Smyth, Greek Grammar, 2464; Martinez, ‘Comparative Clauses’. 16  For some cases in Plato where a ὥσπερ comparative clause with no verb, participle, or corresponding demonstrative adverb precedes its leading clause, see Plato, Parmenides, 164 d 2; Phaedo, 66 b 3–4, 82 e 3.



That Difference Is Different from Being

91

considering which reading best advances the philosophical argument the Stranger is making would be remiss. Since whether ὥσπερ τὸ ὄν adjoins to the protasis or apodosis cannot be decided on the basis of syntax or word order,17 I now turn to the philosophical content of the Stranger’s argument.

3.  ‘Participated in both forms’ at 255 d 4 and the ὥσπερ τὸ ὄν clause The philosophical consideration most relevant to determining the clause to which ὥσπερ τὸ ὄν adjoins is how to read the phrase ­‘participated in both forms’ (ἀμϕοῖν μετεῖχε τοῖν εἰδοῖν) at 255 d 4. Should the phrase be taken in a non-­technical sense, in which case ‘both forms’ would refer to the two traits or characteristics identified by the KH/PA distinction and to ‘participate’ in them would simply mean to possess them? Or should it be taken in the technical sense, such that the two forms are metaphysical entities that cause their participants to be the way they are? If taken in the non-­ technical sense, the Stranger is asking Theaetetus to consider what 17  Apart from the relative infrequency of ὥσπερ comparative clauses preceding their leading clauses without a verb or corresponding demonstrative adverb, the only other word order issue that might seem relevant is that ὥσπερ τὸ ὄν, if read with the apodosis, would precede ἄν. This might at first seem to violate a word order pattern often referred to as Wackernagel’s Law, according to which certain enclitics and postpositives such as ἄν occur in second position within their sentence or clause (D. M. Goldstein, ‘Wackernagel’s Law I’, Encyclopedia of Ancient Greek Language and Linguistics, 23 December 2013. doi: 10.1163/2214-448X_eagll_COM_00000375; cf. J. Wackernagel, ‘Über ein Gesetz der indogermanischen Wortstellung’, Indogermanische Forschungen, 1 (1892), 333–436; K. J. Dover, Greek Word Order (Cambridge, 1960), 12–19; M. H. B. Marshall, Verbs, Nouns and Postpositives in Attic Prose (Edinburgh, 1987), 2; S.  R.  Anderson, ‘Wackernagel’s Revenge: Clitics, Morphology, and the Syntax of Second Position’, Language, 69 (1993), 68–98; S.  Luraghi, ‘Clitics’, Encyclopedia of Ancient Greek Language and Linguistics, 2 October 2013, §3.d. doi: 10.1163/2214-448X_eagll_COM_00000060; D.  M.  Goldstein, ‘Variation versus Change: Clausal Clitics between Homer and Herodotus’, Indo-European Linguistics, 4 (2016), 53–97; D.  M.  Goldstein, Classical Greek Syntax: Wackernagel’s Law in Herodotus (Leiden, 2016), 4–6). Since in canonical cases, second position is the position after the first word or prosodic unit in a clause, given ὥσπερ τὸ ὂν ἦν ἄν at 255 d 5, the apodosis clause would, in a canonical case at least, begin with ἦν. However, none of this poses a problem to reading ὥσπερ τὸ ὄν with the apodosis, since doing so does not make it part of the apodosis clause. The ὥσπερ τὸ ὄν at 255 d 5 is an adjunct clause that, if adjoined to the apodosis, is not governed by the ἄν in the apodosis clause.

92

Michael Wiitala

would follow if a different possessed both the characteristic of ‘being itself according to itself’ and the characteristic of ‘being relative to others’. If taken in the technical sense, the Stranger is presumably asking Theaetetus to consider what would result metaphysically if a different participated in both the form different and the form being. I will argue that taking ‘participated in both forms’ in the technical sense is more natural and should be adopted if there is a plausible reading of the argument that permits it. The following considerations show why the technical sense is more natural. The word ‘metechein’ (to participate, share in) is used nineteen times in the Sophist.18 In fifteen of those cases it is clearly used to indicate participation in a form in the technical sense.19 Moreover, of those fifteen instances in which ‘metechein’ is clearly technical, seven occur within a Stephanus page of 255 d 4: 255 b 1, 255 b 3, 255 e 5, 256 a 1, 256 a 7, 256 d 9, and 256 e 3. Likewise, throughout the Sophist, the word ‘eidos’ (form) is almost exclusively used as a technical term. The word ‘eidos’ appears forty-­ eight times in the dialogue.20 If we except from consideration its use at 255 d 4, the only time it is used in a non-­technical sense is at 266 c 4, during the Stranger’s final attempt to define sophistry. There ‘eidos’ is used to describe the image or reflection produced by mirrors and other smooth surfaces. In all other instances, again if we except 255 d 4 from consideration, ‘eidos’ unambiguously means form in the technical sense. Thus, ‘participated in both forms’ at 255 d 4 should be understood technically, if there is a plausible construal that permits it. Furthermore, the conclusion the Stranger immediately draws from the argument of 255 c 9–d 7 is that ‘the nature of a different must be said to be a fifth among the forms [τοῖς 18 Plato, Sophist, 216 b 1, 228 c 1, 235 a 6, 238 e 2, 251 e 10, 255 b 1, 255 b 3, 255 d 4, 255 e 5, 256 a 1, 256 a 7, 256 d 9, 256 e 3, 259 a 6, 259 a 8, 260 d 3, 260 d 5, 260 d 7, 260 d 8. 19 Plato, Sophist, 238 e 2, 251 e 10, 255 b 1, 255 b 3, 255 e 5, 256 a 1, 256 a 7, 256 d 9, 256 e 3, 259 a 6, 259 a 8, 260 d 3, 260 d 5, 260 d 7, 260 d 8. The other instances of ‘metechein’ can perhaps be read in the technical sense as well, although the context does not overtly require it (see 216 b 1, 228 c 1, 235 a 6). 20 Plato, Sophist, 219 a 8, 219 c 2, 219 d 5, 220 a 6, 220 a 7, 220 e 6, 222 d 6, 222 e 3, 223 c 6, 223 c 9, 225 c 2, 226 c 11, 226 e 1, 226 e 5, 227 c 7, 227 c 8, 227 d 13, 229 c 2, 230 a 9, 234 b 2, 234 b 3, 235 d 1, 236 c 6, 236 d 2, 239 a 10, 246 b 8, 246 c 9, 248 a 4, 249 d 1, 252 a 7, 253 d 1, 254 c 3, 255 c 6, 255 d 4, 255 e 1, 256 e 6, 258 c 4, 258 d 6, 259 e 6, 260 d 7, 261 d 1, 264 c 1, 264 c 4, 265 a 8, 266 c 4, 266 d 6, 266 e 5, 267 d 6.



That Difference Is Different from Being

93

εἴδεσιν] we are selecting’ (255 d 9–e 1). This not only confirms that ‘participated in both forms’ should, if possible, be understood in the technical sense, but also that it should be read as referring to participation in two of the only five forms named and under discussion in this portion of the dialogue: rest, motion, being, same, and different. Therefore, the most natural reading of ‘participated in both forms’ would, if possible, both (1) render it in the technical sense and (2) identify the two forms as two of the five the Stranger is selecting. Yet a plausible reading that renders ‘participated in both forms’ as participation in two of the five the Stranger is selecting is impossible if ὥσπερ τὸ ὄν is adjoined to the protasis. The following translates the conditional with ὥσπερ τὸ ὄν so adjoined: But if different participated in both forms, as being does, then there would sometimes also among the differents be something different not relative to something different. (255 d 4–6)

The Stranger wants Theaetetus to conclude that the claim in the protasis is false: different does not participate in both forms. If it did, there would sometimes be a different not relative to something different, which is impossible, because a different is always relative to something different (255 d 1, 255 d 6–7). Since ὥσπερ τὸ ὄν is here read with the protasis, in order to take ‘participated in both forms’ in the technical sense, the two forms would have to be such that a being participated in both, while a different only participated in one. Yet what two forms could play this role? The forms being and different could not, since there is no plausible way to have a being participate in both different and being, while a different only participates in one of them. Whichever way we construe the cases, they ought to be treated symmetrically. Thus, different and being are not the two forms mentioned at 255 d 4 if ὥσπερ τὸ ὄν is read with the protasis. Likewise, neither are rest, motion, or same—­the other forms among the five the Stranger is selecting. Cases with those forms would also need to be treated symmetrically. Moreover, there is no indication whatsoever that rest, motion, or same are relevant to the 255 d 4–6 conditional. With ὥσπερ τὸ ὄν adjoined to the protasis, therefore, the forms mentioned at 255 d 4 cannot be any of the five greatest kinds. Since the Stranger has indicated that the five kinds are the only forms under discussion at this point in the dialogue (254 b 8–d 5,

94

Michael Wiitala

254 d 14–255 a 2), ‘participated in both forms’ does not appear to refer to forms or participation in the technical sense at all, if ὥσπερ τὸ ὄν is read with the protasis. Hence, although contemporary interpretations of the argument of 255 c 9–e 2 differ significantly from one another, most share the feature of reading ‘participated in both forms’ at 255 d 4 in a non-­technical sense, as possessing the two traits or characteristics identified by the KH/PA distinction.21 Notwithstanding, there are a handful of interpretations that take ‘participated in both forms’ in the technical sense.22 Russell Dancy, for example, argues that it refers to participation in the forms ‘standalone’ (αὐτὰ καθ’ αὑτά) and ‘relative’ (πρὸς ἄλλα)—forms he claims are identified by the KH/PA distinction. In order to maintain that

21  E.g. J. L. Ackrill, ‘Plato and the Copula: Sophist 251–259’, Journal of Hellenic Studies, 77 (1957), 1–6; A. E. Taylor (trans.), Plato: Sophist and Statesman (London, 1961), 161; J. Warrington (trans.), Plato: Parmenides, Theaetetus, Sophist, Statesman (New York, 1961), 207; Frede, Prädikation und Existenzaussage, 24; P.  Seligman, Being and Not-Being: An Introduction to Plato’s Sophist (The Hague, 1974), 61–2; Reeve, ‘Motion, Rest, and Dialectic’, 54; L. M. de Rijk, Plato’s Sophist: A Philosophical Commentary (Amsterdam, 1986), 151; Vries, ‘On Sophist 255 b–­e’, 391–2; J. Kostman, ‘The Ambiguity of “Partaking” in Plato’s Sophist’, Journal of the History of Philosophy, 27 (1989), 343–63 at 358; N. P. White (trans.), Sophist (Indianapolis, 1993), 49 n. 63; N. Notomi, The Unity of Plato’s Sophist: Between the Sophist and the Philosopher (Cambridge, 1999), 242 n. 70; Silverman, Dialectic of Essence, 170; J. Duerlinger, A Translation of Plato’s Sophist with an Introductory Commentary (New York, 2005), 59, 120; Malcolm, ‘A Way Back’, 275; Crivelli, Plato’s Account of Falsehood, 144–9; Leigh, ‘Modes of Being’, 21. Additionally, a few commentators have noted that ‘metechein’ has an unusual sense at 255 d 4, assuming, as they do, that ὥσπερ τὸ ὄν is adjoined to the protasis, e.g. R. E. Heinaman, ‘Communion of Forms’, Proceedings of the Aristotelian Society, 83 (1982), 175–90 at 186; Kostman, ‘The Ambiguity of “Partaking” in Plato’s Sophist’, 358. 22  E.g. F. M. Cornford, Plato’s Theory of Knowledge: The Theaetetus and Sophist [Plato’s Theory of Knowledge] (London, 1935), 256–7, 281 n. 2; W. G. Runciman, Plato’s Later Epistemology (Cambridge, 1962), 90–2; Bluck, Plato’s Sophist, 146 n. 1; Dancy, ‘Categories of Being’, 60 ff. (I discuss Dancy’s position in the main text). Hestir argues that the two forms are ‘itself by virtue of itself’ (αὐτὰ καθ’ αὑτά) and ‘relative to others’ (πρὸς ἄλλα), although on his view the forms in the Sophist and Statesman differ significantly from the sort of forms in dialogues like the Republic and Phaedo (Hestir, Plato on Meaning and Truth, 159–61; cf. 20–34). Heinaman thinks the two forms are ‘non-relative’ (αὐτὰ καθ’ αὑτά) and ‘relative’ (πρὸς ἄλλα), but that Plato was clearly wrong to consider them forms (‘Communion of Forms’, 186). Rosen argues that the two forms are ‘the pure forms being and otherness’ (Plato’s Sophist, 269), which makes him the only commentator I found who takes the forms mentioned in 255 d 4 as two of the five the Stranger is selecting. Yet his brief exegesis of the passage does not provide an explanation as to how construing the two forms as being and otherness coheres with his reading of ὥσπερ τὸ ὄν adjoined to the protasis (Rosen, Plato’s Sophist, 269–70).



That Difference Is Different from Being

95

being, but not different, participates in both standalone and relative, however, Dancy admits it is necessary to attribute a ‘use-­mention ambiguity’ to Plato regarding the terms ‘being’ and ‘different’.23 Although Dancy himself claims not to find this sort of ambiguity problematic, he acknowledges others will.24 Yet leaving that aside, even if standalone and relative were forms in the tech­nical sense, as Dancy and some others have argued, they would not be two of the five forms the Stranger claims to be selecting. If standalone and relative were two forms, one would expect the Stranger either explicitly to count them as such along with the other five or else explain why they are not included. After all, if, as Dancy argues, everything participates in either standalone, relative, or both,25 then it would seem that standalone and relative are forms whose extension is as great or nearly as great as rest, motion, being, same, and different. Since the Stranger neither counts standalone and relative among the five forms nor explains why he does not, if it is possible to give a plausible reading that can render the two forms mentioned at 255 d 4 as two of the five the Stranger is selecting, that reading would be preferable to Dancy’s and to any other that claims the two forms are standalone and relative. I will argue that such a reading is possible if ὥσπερ τὸ ὄν is adjoined to the apodosis instead of the protasis.

4.  Reading the argument with ὥσπερ τὸ ὄν adjoined to the apodosis As I have shown and as the readings of contemporary commentators evince, if ὥσπερ τὸ ὄν is adjoined to the protasis of the 255 d 4–6 conditional, it is impossible to offer a plausible interpretation of the argument in which ‘participated in both forms’ is taken in the most natural way, as referring to participation in two of the five kinds. If, however, ὥσπερ τὸ ὄν is adjoined to the apodosis, as I propose, then the two forms mentioned at 255 d 4 can be the forms different and being, the two the Stranger is trying to distinguish in the immediate context. I will first present my construal of the argument and then defend it. 23  Dancy, ‘Categories of Being’, 49, 59–60. 24  Dancy, ‘Categories of Being’, 49, 60; cf. Leigh, ‘Modes of Being’, 16 n. 28. 25  Dancy, ‘Categories of Being’, 60.

96

Michael Wiitala

With ὥσπερ τὸ ὄν adjoined to the apodosis, the argument that being and different are distinct forms can be construed as a simple reductio with two premises. The first premise is the KH/PA distinction: some beings are what they are themselves according to themselves, whereas other beings are what they are relative to ­others (255 c 13–14). The second premise is the claim that something different always is what it is relative to something different (255 d 1, 255 d 6–7). After stating these two premises, the Stranger uses the counterfactual conditional in 255 d 3–4 to introduce the assumption for reductio. He asks Theaetetus to consider what would result ‘if being and the different were not entirely distinct [forms]’ (255 d 3–4), that is to say, if the forms being and different were identical. Next, in a second counterfactual conditional—­the conditional at 255 d 4–6 we have been considering throughout this paper—­he explains what would result, drawing out the contradiction. I propose the following reconstruction of the argument, which I will explain and defend afterwards: (1) Some beings are what they are themselves according to themselves, whereas some beings are what they are relative to others (premise, 255 c 13–14). (2) Every different is what it is relative to something different (premise, 255 d 1, 255 d 6–7). (3) Assume it is not the case that different and being are distinct forms (for reductio, 255 d 3–4). (4) What a thing is in virtue of participating in different (viz. what a different is as such) is identical to what a thing is in virtue of participating in being (viz. what a being is as such) (from (3), 255 d 4). (5) It is not the case that every being is what it is relative to something different (from (1), 255 d 5–6). (6) It is not the case that every different is what it is relative to something different (from (4) and (5), 255 d 5–6). (7) But (2) and (6) contradict one another (255 d 4–7). (8) Therefore, different and being are distinct forms (from (3) and (7), 255 d 3–4, 255 d 9–e 1). The main advantage of this reconstruction is that it (a) enables us to take ‘participated in both forms’ (ἀμϕοῖν μετεῖχε τοῖν εἰδοῖν) at 255 d 4 in the technical sense and (b) read the two forms as different and being. As demonstrated in Section 3, no reading that adjoins



That Difference Is Different from Being

97

ὥσπερ τὸ ὄν to the protasis of the 255 d 4–6 conditional can achieve both these desiderata. My reading, in contrast, since it adjoins ὥσπερ τὸ ὄν to the apodosis, can render the claim made by the prot­ asis as line (4) above. I will now explain how. In the 255 d 4–6 conditional and throughout the passage, the Stranger speaks not only of the forms different and being, but also about the differents and beings that participate in them. Commentators disagree about what such participant differents and beings are in this context. Some argue, for example, they are the properties different and being.26 Others claim they are instances of the forms different and being, that is to say, the qualities of different and being immanent in things.27 Still others regard them as terms ambiguous between use and mention, which refer at once to individuals, properties, linguistic subjects, and linguistic predicates.28 On my reading, a participant in different, in this context, is what a different is insofar as it is different, while a participant in being is what a being is insofar as it is. The textual basis for my reading comes near the end of our passage, where the Stranger says, ‘it is now obvious to us that whatever is different has, by necessity, turned out to be this, just what it is (ὅπερ ἐστίν), from something different’ (255 d 6–7). The Stranger there articulates more precisely what he said at 255 d 1: ‘a different [is] always [said to be] relative to something different’. In light of 255 d 6–7, the claim in 255 d 1 is that anything different is what it is relative to something different. On the grammatical level, however, the Stranger’s question to Theaetetus in 255 d 1 has an el­lip­ tic­al construction dependent on what the Stranger says in 255 c 13–14: ‘among beings, some are [always] said [to be] themselves according to themselves, while some are always said [to be] relative to others’. Thus, if ‘different’ in 255 d 1 should be glossed as ‘what a different is as such’, then ‘beings’ in 255 c 13–14 should be glossed as ‘what beings are as such’. Hence, I construe the claim at 255 c 13–14 as ‘some beings are what they are themselves according to 26  E.g. Leigh, ‘Modes of Being’; Hestir, Plato on Meaning and Truth, 162 n. 51. 27  E.g. Bluck, Plato’s Sophist, 148; Rosen, Plato’s Sophist, 269. 28  For the most concise formulation of this view, see Duncombe, ‘Absolute and Relative Categories’, esp. 77 n. 2. Others who hold this view, or a version of it, include Cornford, Plato’s Theory of Knowledge, 279–85; R. E. Heinaman, ‘Being in the Sophist’, Archiv für Geschichte der Philosophie, 65 (1983), 1–17 at 14; Vries, ‘On Sophist 255 b–­e’, 388; Dancy, ‘Categories of Being’.

98

Michael Wiitala

themselves, whereas some beings are what they are relative to o ­ thers’. A small being, for example, only is such—­is a small being—­relative to something else, namely to a large being. To put it otherwise, what something small is only participates in being relative to something else. A human being, by contrast, is a being itself according to itself. A sign of this difference between a human being and a small being is that an account of what it is to be human will identify features of human nature itself, such as rationality or animality, whereas an account of what it is to be small will require reference to things other than smallness, namely to large things. The sort of account needed differs because, while beings with natures like small, large, quicker, different, and so on are beings relative to ­others, beings with natures such as human, oneness, courage, or angling are beings themselves according to themselves. The nature of a given thing determines whether it participates itself according to itself or relative to others in the form being. Using the Stranger’s focus on the ‘what it is’ of different in 255 d 6–7 to guide construing his claims in 255 c 13–14 and 255 d 1 not only makes good sense of those claims, but more importantly renders the argument as a whole such that the two forms mentioned in the 255 d 4–6 conditional will be the forms different and being. If we take ὥσπερ τὸ ὄν in the apodosis, the 255 d 4–6 conditional reads: But if a different participated in both forms, then, just as with being, there would sometimes also among the differents be something different not relative to something different. (255 d 4–6)

I take the Stranger to be using ‘different’ here to mean ‘what a different is, insofar as it is different’. Moreover, he holds that a different is what it is in virtue of participating in the form different (see esp. 255 e 3–6). The conditional, therefore, is describing a consequence that would follow if something different were what it is in virtue of participating both in different and in being, instead of just in virtue of participating in different. Of course, everything participates in both different and being. Nevertheless, what something is in virtue of participating in different—­namely something insofar as it is different—­is not identical to what something is in virtue of participating in being—­namely something insofar as it is a being. For example, something different is something different in virtue of participating in different, whereas something different is something that is in virtue of participating in being. The 255 d 4–6



That Difference Is Different from Being

99

conditional explains what would follow if what a thing is in virtue of participating in different were identical to what a thing is in virtue of participating in being. For a different to participate in both different and being would be for anything different, just insofar as it is different, to be identical to anything that is, just insofar as it is. Yet obviously this is not the case. Instead, what it is to be different is distinct from what it is to be. This sort of construal of the protasis of the 255 d 4–6 conditional, however, is only possible if ὥσπερ τὸ ὄν is adjoined to the apodosis. Otherwise, the protasis would read, ‘if a different participated in both forms, just as being does . . .’, with the result that the two ‘forms’ would have to be the characteristics ‘being itself according to itself’ and ‘being relative to others’, despite the fact that those characteristics are not identified as forms, much less as two of the five under discussion in this portion of the dialogue. 5. Conclusion Sophist 255 c 9–e 2 is a difficult passage. Key terms and phrases in the Greek are ambiguous. Moreover, the text gives commentators little to work with in resolving these ambiguities. Some am­bi­gu­ ities are easier to decide than others, however. I have argued that the ambiguity of ‘participated in both forms’ at 255 d 4, regarding whether it should be read as referring to participation in two of the five forms the Stranger claims to be selecting, has a clear reso­ lution. Since ‘metechein’ (to participate) and ‘eidos’ (form) are both used as technical terms throughout the Sophist digression, any interpretation should, if at all possible, read ‘participated in both forms’ at 255 d 4 in the technical sense. Additionally, if possible, the two forms should be two of the five greatest kinds. I have also argued, however, that to interpret the two forms mentioned in 255 d 4 as two of the five kinds requires reading ὥσπερ τὸ ὄν with the apodosis of the conditional at 255 d 4–6. With ὥσπερ τὸ ὄν adjoined to the apodosis, it is possible to interpret the Stranger’s argument such that the forms mentioned in 255 d 4 are the forms different and being, as I demonstrated in Section 4. Moreover, reading ὥσπερ τὸ ὄν with the apodosis allows the 255 d 4–6 conditional to provide a more determinate frame of reference for interpreting both the KH/PA distinction and the argument as a whole. With

100

Michael Wiitala

ὥσπερ τὸ ὄν adjoined to the apodosis, the KH/PA distinction is clearly metaphysical, rather than syntactic or semantic, and identifies two ways in which beings can be what they are. The argument as a whole, in turn, shows that different and being are distinct forms by demonstrating that what something is in virtue of participating in different is not identical to what something is in virtue of participating in being. A small change in punctuation, therefore, leads in this case to a valid and, I think, persuasive reconstruction of the Stranger’s argument.

BIBLIOGR A PH Y Ackrill, J. L., ‘Plato and the Copula: Sophist 251–259’, Journal of Hellenic Studies, 77 (1957), 1–6. Ambuel, D., ‘Difference in Kind: Observations on the Distinction of the Megista Gene’, in B. Bossi and T. M. Robinson (eds.), Plato’s Sophist Revisited (Berlin, 2013), 247–67. Ambuel, D., Image and Paradigm in Plato’s Sophist (Las Vegas, 2007). Anderson, S.  R., ‘Wackernagel’s Revenge: Clitics, Morphology, and the Syntax of Second Position’, Language, 69 (1993), 68–98. Apelt, O. (ed.), Platonis Sophista (Leipzig, 1897). Baiter, J. G., Orelli, J. C., and Winckelmann, A. G. (eds.), Platonis opera quae feruntur omnia (Zurich, 1839). Bluck, R. S., Plato’s Sophist (Manchester, 1975). Bostock, D., ‘Plato on “Is-Not” (Sophist 254–9)’, Oxford Studies in Ancient Philosophy, 2 (1984), 89–119. Brown, L., ‘Being in the Sophist: A Syntactical Enquiry’ [‘Being in the Sophist’], in G. Fine (ed.), Plato 1: Metaphysics and Epistemology (Oxford, 1999), 455–78. Burnet, J. (ed.), Platonis opera, vol. i: Euthyphro, Apologia Socratis, Crito, Phaedo, Cratylus, Theaetetus, Sophista, Politicus (Oxford, 1900). Campbell, L., The Sophistes and Politicus of Plato (Oxford, 1867). Cornford, F. M., Plato’s Theory of Knowledge: The Theaetetus and Sophist [Plato’s Theory of Knowledge] (London, 1935). Crespo, E., ‘Adjuncts’, Encyclopedia of Ancient Greek Language and Linguistics, 24 September 2013. doi: 10.1163/2214-448X_eagll_SIM_00000413. Crivelli, P., Plato’s Account of Falsehood: A Study of the Sophist [Plato’s Account of Falsehood] (Cambridge, 2012). Dancy, R.  M., ‘The Categories of Being in Plato’s Sophist 255 c–­e’ [‘Categories of Being’], Ancient Philosophy, 19 (1999), 45–72.



That Difference Is Different from Being

101

Diès, A. (ed.), Platon: Œuvres complètes, vol. viii. 3: Sophiste, 3rd edn (Paris, 1955). Dover, K. J., Greek Word Order (Cambridge, 1960). Duerlinger, J., A Translation of Plato’s Sophist with an Introductory Commentary (New York, 2005). Duke, E.  A., Hicken, W.  F., Nicoll, W.  S.  M., Robinson, D.  B., and Strachan, J.  C.  G. (eds.), Platonis Opera, vol. i: Euthyphro, Apologia, Crito, Phaedo, Cratylus, Theaetetus, Sophista, Politicus (Oxford, 1995). Duncombe, M., ‘Plato’s Absolute and Relative Categories at Sophist 255 c 14’ [‘Absolute and Relative Categories’], Ancient Philosophy, 32 (2012), 77–86. Eck, J. van, ‘Not-Being and Difference: On Plato’s Sophist, 256 d 5–258 e 3’, Oxford Studies in Ancient Philosophy, 23 (2002), 68–84. Fowler, H. N. (ed. and trans.), Plato: Theaetetus, Sophist, Loeb Classical Library 123 (Cambridge, MA, 1921). Frede, M., ‘Plato’s Sophist on False Statements’, in R. Kraut (ed.), The Cambridge Companion to Plato (Cambridge, 1992), 397–424. Frede, M., Prädikation und Existenzaussage: Platons Gebrauch von ‘ . . . ist . . . ’ und ‘ . . . ist nicht . . . ’ im Sophistes [Prädikation und Existenzaussage] (Göttingen, 1967). Gill, M. L., ‘Method and Metaphysics in Plato’s Sophist and Statesman’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2016 edn), https://plato.stanford.edu/archives/win2016/entries/plato-sophstate/, accessed 14 January 2023. Gill, M. L., Philosophos: Plato’s Missing Dialogue (Oxford, 2012). Goldstein, D. M., Classical Greek Syntax: Wackernagel’s Law in Herodotus (Leiden, 2016). Goldstein, D.  M., ‘Variation versus Change: Clausal Clitics between Homer and Herodotus’, Indo-European Linguistics, 4 (2016), 53–97. Goldstein, D.  M., ‘Wackernagel’s Law I’, Encyclopedia of Ancient Greek Language and Linguistics, 23 December 2013. doi: 10.1163/ 2214-448X_eagll_COM_00000375. Heinaman, R. E., ‘Being in the Sophist’, Archiv für Geschichte der Philosophie, 65 (1983), 1–17. Heinaman, R. E., ‘Communion of Forms’, Proceedings of the Aristotelian Society, 83 (1982), 175–90. Heindorf, L. F. (eds.), Platonis dialogi selecti, vol. iv (Berlin, 1809). Hermann, K.  F., Platonis dialogi secundum Thrasylli tetralogias dispositi, vol. i (Leipzig, 1851). Hestir, B. E., Plato on the Metaphysical Foundation of Meaning and Truth [Plato on Meaning and Truth] (Cambridge, 2016). Kostman, J., ‘The Ambiguity of “Partaking” in Plato’s Sophist’, Journal of the History of Philosophy, 27 (1989), 343–63.

102

Michael Wiitala

Lacey, A.  L., ‘Plato’s Sophist and the Forms’, Classical Quarterly, n.s., 9 (1959), 43–52. Leigh, F., ‘Modes of Being at Sophist 255 c– e’ [‘Modes of Being’], Phronesis, 57 (2012), 1–28. Luraghi, S., ‘Clitics’, Encyclopedia of Ancient Greek Language and Linguistics, 2 October 2013. doi: 10.1163/2214-448X_eagll_COM_00000060. Malcolm, J., ‘A Way Back for Sophist 255 c 12–13’ [‘A Way Back’], Ancient Philosophy, 26 (2006), 275–89. Marshall, M. H. B., Verbs, Nouns and Postpositives in Attic Prose (Edinburgh, 1987). Martinez, R., ‘Comparative Clauses’, Encyclopedia of Ancient Greek Language and Linguistics, 24 September 2013. doi: 10.1163/2214-448X_eagll_SIM _00000431. McCabe, M. M., Plato’s Individuals (Princeton, 1994). Moravcsik, J. M. E., ‘Being and Meaning in the Sophist’, Acta Philosophica Fennica, 14 (1962), 23–78. Notomi, N., The Unity of Plato’s Sophist: Between the Sophist and the Philosopher (Cambridge, 1999). Owen, G. E. L., ‘Plato on Not-Being’, in G. Vlastos (ed.), Plato: A Collection of Essays, vol. i: Metaphysics and Epistemology (Garden City, 1971), 223–67. Peck, A.  L., ‘Plato and the ΜΕΓΙΣΤΑ ΓΕΝΗ of the Sophist: A Reinterpretation’, Classical Quarterly, n.s., 2 (1952), 32–56. Reeve, C.  D.  C., ‘Motion, Rest, and Dialectic in the Sophist’ [‘Motion, Rest, and Dialectic’], Archiv für Geschichte der Philosophie, 67 (1985), 47–64. Rijk, L. M. de, Plato’s Sophist: A Philosophical Commentary (Amsterdam, 1986). Rosen, S., Plato’s Sophist: The Drama of Original and Image [Plato’s Sophist] (New Haven, 1983). Runciman, W. G., Plato’s Later Epistemology (Cambridge, 1962). Seligman, P., Being and Not-Being: An Introduction to Plato’s Sophist (The Hague, 1974). Silverman, A., The Dialectic of Essence: A Study of Plato’s Metaphysics [Dialectic of Essence] (Princeton, 2002). Smith, C. C., ‘Against the Existential Reading of Euthydemus 283 e–284 c, with Help from the Sophist’, Ancient Philosophy, 42 (2022), 67–81. Smyth, H. W., Greek Grammar, revised by G. M. Messing (Cambridge, MA, 1956). Stallbaum, G. (ed.), Platonis Sophista (Gotha, 1840). Taylor, A. E. (trans.), Plato: Sophist and Statesman (London, 1961).



That Difference Is Different from Being

103

Vries, W. de, ‘On Sophist 255 b–­e’, History of Philosophy Quarterly, 5 (1988), 385–94. Wackernagel, J., ‘Über ein Gesetz der indogermanischen Wortstellung’, Indogermanische Forschungen, 1 (1892), 333–436. Warrington, J. (trans.), Plato: Parmenides, Theaetetus, Sophist, Statesman (New York, 1961). White, N. P. (trans.), Sophist (Indianapolis, 1993).

IS PLATO A CONSEQUENTIALIST? christopher bobonich 1. Introduction We can learn something about both ancient and contemporary value theory by considering their similarities and dissimilarities. This paper examines some similarities and dissimilarities between Plato’s ethical and political philosophy and modern consequentialism. Section 2 begins where the similarities seem greatest, that is, with the Republic’s and the Laws’ criterion for laws’ correctness. I argue that this criterion is consequentialist in some fundamental respects. Section 3 considers various objections to seeing Platonic affinities with consequentialism and argues that although the most common objections are not convincing, there is reason for caution about this attribution. Section 4 thus looks for consequentialism’s distinctive features in the Republic and the Laws. I conclude that although Plato’s criterion for legal correctness has basic consequentialist features, the case for Plato as a consequentialist is not conclusive. Section 5 points to a much neglected passage from Republic Book 1 that I argue contains an important ethical requirement that is not justified consequentially and is, in principle, inconsistent with consequentialism. Section 6 considers how Laws 10’s theology leads Plato in a fully impartialist direction and to endorse a strong form of consequentialism. This theology urges us to adopt a god’s-­eye perspective that has affinities with certain Stoic ideas and also with what Bernard Williams described as Henry Sidgwick’s ‘memorably absurd’ notion of the ‘point of view of the universe’.1 Plato here claims that there is one and the same end for all humans which is the universe’s best condition.

I would like to thank audiences at Stanford, the University of Tartu, and Yale and especially Victor Caston, David Charles, Kinch Hoekstra, Terry Irwin, Rupert Spar­ ling, and Raphael Woolf for their comments. 1 B. Williams, Moral Luck (Cambridge, 1981), xi.

Christopher Bobonich, Is Plato a Consequentialist ? In: Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Christopher Bobonich 2023. DOI: 10.1093/oso/9780192885180.003.0004

106

Christopher Bobonich

2.  Consequentialist principles in the Republic and the Laws 2.1.  Laws and consequentialist principles To see how far the Republic and the Laws go in the direction of consequentialism, let us begin where its case is strongest. This is in these dialogues’ criterion for correct laws. Although the Laws’ discussion is more complete, I think that both dialogues share the same criterion. As we shall see, its satisfaction by a set of possible laws is a decisive reason for adopting them. Consider the following passages. In both Kallipolis and Magnesia (the city sketched in the Laws), the laws’ ultimate end is the city’s greatest happiness. For example, in the Republic, replying to the charge that the rulers are not happy, Socrates remarks: [T1]  ὅτι θαυμαστὸν μὲν ἂν οὐδὲν εἴη εἰ καὶ οὗτοι οὕτως εὐδαιμονέστατοί εἰσιν, οὐ μὴν πρὸς τοῦτο βλέποντες τὴν πόλιν οἰκίζομεν, ὅπως ἕν τι ἡμῖν ἔθνος ἔσται διαϕερόντως εὔδαιμον, ἀλλ’ ὅπως ὅτι μάλιστα ὅλη ἡ πόλις. (Rep. 4, 420 b 4–8)2

it would not be surprising if these people were happiest this way, but in founding the city, we are not aiming at one group being outstandingly happy, but at the whole city being as happy as possible.

I give a more precise account of consequentialism in Section 2.2, but we should note here that, first, Socrates provides a principle for ranking sets of laws that is the same for everyone to whom it applies, in this case, for every lawgiver for Kallipolis. The second feature is that satisfying this condition decisively determines what laws to adopt. That is enough, as we shall see, to make this view consequentialist. Whether the criteria for a lawcode are consequentialist has attracted little scholarly attention. What has been debated is the distinct question of whether the city’s happiness is reducible to the citizens’ happiness.3 This concerns not whether the criterion is 2  For the Republic’s text and lineation, I use S.  Slings, Platonis Rempublicam (Oxford, 2003); for the Laws’ text and lineation, É. des Places and A. Diès, Platon, Œuvres Complètes vols. xi–­xii (Paris, 1951–6). For translations from the Laws, I use, with modification, C. D. C. Reeve, Plato Laws (Indianapolis, 2022). For other Plato translations, I use, with modification, J.  Cooper (ed.), Plato: Complete Works (Indianapolis, 1997). 3  Reductionists include J. Annas, An Introduction to Plato’s Republic (Oxford, 1981), 179–80; G. Vlastos, ‘The Theory of Social Justice in the Polis in the Republic’, in id., Studies in Greek Philosophy), 2 vols. (Princeton, 1995), ii. 69–103 at 78–90. Anti-­reductionists include G.  Cantu, ‘Individual and Polis in Plato’s Republic’



Is Plato a Consequentialist?

107

consequentialist, but the content of the lawcode’s ranking prin­ ciple. Although resolving the reducibility issue is not necessary for answering our question about consequentialism, the ranking prin­ ciple is important and deserves some discussion. Gregory Vlastos provides the main textual arguments for reducibility. First, Republic 4, 420 e 1–421 a 3, assumes that making all the citizens happy makes the whole city happy. Second, Republic 7, 519 e 1–520 a 2, shows that the benefit involved in the whole city being happy is one that the citizens give to one another, that is, this benefit consists in making each other happy. For the whole city to be happy is thus for its citizens to be so, and there seems to be nothing else for the city’s well-­being to be other than some function of individual citizens’ well-­being.4 Non-­reductionists rely on other passages, especially Socrates’ statue analogy. If someone objected to painting a statue’s eyes black because purple is the most beautiful colour, Socrates would reply: [T2]  μὴ οἴου δεῖν ἡμᾶς οὕτω καλοὺς ὀϕθαλμοὺς γράϕειν ὥστε μηδὲ ὀϕθαλμοὺς ϕαίνεσθαι, μηδ’ αὖ τἆλλα μέρη, ἀλλ’ ἄθρει εἰ τὰ προσήκοντα ἑκάστοις ἀποδιδόντες τὸ ὅλον καλὸν ποιοῦμεν. (Rep. 4, 420 d 2–5)

Do not think that we should paint eyes so beautifully that they no longer appear to be eyes, nor so paint the other parts, but consider whether by giving the appropriate colours to each of the parts, we make the whole beautiful.

This is non-­reductionists’ main textual evidence. They take Socrates to deny that a statue’s beauty is a function of its parts’ beauty and to adopt instead an organic account of beauty according to which, although each part would be most beautiful if it were painted purple, [‘Individual and Polis’], Polis, 28 (2011), 90–107; D. Morrison, ‘The Happiness of the City and the Happiness of the Individual in Plato’s Republic’ [‘Happiness’], Ancient Philosophy, 21 (2001), 1–24; M.  Schofield, Plato: Political Philosophy [Political Philosophy] (Oxford, 2006), 212–27. Plato explicitly accepts the counterfactual that if all the citizens were happy but performed their political erga poorly, then the city would still be happy (Rep. 4, 420 d 6–e 7). This is good evidence that Plato does not identify the individual’s happiness with performing her ergon or the city’s happiness with all the citizens performing their erga. 4  Cf. 4, 421 b 6 and 7, 541 a 4–7. A person’s well-­being is determined by her goods and bads (and the benefit or harm of their use). Being happy can be understood either as the highest degree of well-­being or as a range concept (someone’s happy if she is either at or near the highest point). Maximizing happiness is more precisely maximizing well-­being so that the city’s well-­being is increased if someone is made better off (or less badly off) even if she does not become happy.

108

Christopher Bobonich

a statue of a human with all purple parts would not be beautiful. So, reductionism is false. But Socrates never says that one can make each of the statue’s parts beautiful while not making the whole statue beautiful. If he did, either he would be inconsistent or his analogy would be inept, since he immediately agrees that making all the citizens happy would make the whole city happy (Rep. 4, 420 e 1–7). Non-­reductionism thus loses its primary textual evidence. In this ex­ample, Socrates does not suggest that the most beautiful eye is one painted so that it no longer looks like an eye, nor does he concede it counterfactually. He never says what a statue’s beauty consists in and so is not adopting an organic account of beauty. Rather, this example summarizes Socrates’ reply to the above objection (cf. 421 b 5–c 5): (1) its conception of happiness is false (420 b 3–7); (2) each class doing its own work is necessary for the city’s greatest happiness; and (3) their goal is making the whole city as happy as possible, even if this makes some subgroups less happy than they might be (420 b 7–8). Further non-­ reductionist interpretations, such as Donald Morrison’s, that characterize a city’s happiness in terms of some structural feature of the city either have counterintuitive results or collapse into reductionism. Morrison, for example, defines the city’s structure as ‘the arrangement of [its] fundamental parts . . . and [their] aptitude . . . to do their job’, the city’s happiness as ‘the goodness of this structure’, and the structure’s goal as ‘promot[ing] the maximal well-­being of the citizens over time’.5 But Morrison identifies no goodmaking feature of the structure that might compete with achieving its goal. Consider a case in which structural change would better promote the citizens’ well-­being. Retaining the structure would then seem to be a case of irrational structure worship, since it fails to promote the structure’s own goal. But if the structure should be changed in such circumstances, we have no reason to identify the city’s happiness with the structure’s goodness rather than directly with maximal well-­being. The structure’s goodness seems to be nothing but the extent to which it actually produces maximal well-­being, and it is the latter that has ultimate justificatory force. Moreover, the city’s structure consists of fairly coarse-­grained features that do not determine the content of every law, decree, and 5  Morrison, ‘Happiness’, 7. Morrison only considers reductionist views that maximize current well-­being. But all reductionist accounts aim at long-­term happiness.



Is Plato a Consequentialist?

109

action of the rulers, so the rulers could face a choice between options that each maintain the city’s structure, but one of which produces greater well-­being. It seems clear that the rulers should not be indifferent between the two options and thus that Morrison’s view is unsatisfactory.6 Others such as Gerald Cantu, Malcolm Schofield, and Dominic Scott identify the city’s happiness with its unity, that is, the community of pleasure and pain (Rep. 5, 462 a 2–b 6).7 I say more about this in Section 4.1 and argue that it involves the citizens including other citizens’ well-­being in their own. But this in­ter­pret­ation of the city’s happiness is problematic. First, citizens with false conceptions of well-­being might display great unity, but the laws should not aim at that. Unity should thus be such an outcome in virtuous people. But this is still unsatisfactory. Such unity might obtain among citizens enjoying more or less well-­being. But clearly the higher level is better. Thus, unity alone is not sufficient for the city’s happiness. Moreover, the happiness criterion can justify pursuing unity. Forming such a unified community is an upshot of the ­citizens being virtuous, and virtue is happiness’s most important constituent. Maximizing the citizens’ well-­being thus requires fostering unity.8 I conclude that the reductionist account is more plaus­ible than its current competitors. The Laws is more detailed and precise about the laws’ ultimate end. [T3]  ἀρχῶν περιμαχήτων γενομένων, οἱ νικήσαντες τά τε πράγματα κατὰ τὴν πόλιν οὕτως ἐσϕετέρισαν σϕόδρα, ὥστε ἀρχῆς μηδ’ ὁτιοῦν μεταδιδόναι τοῖς ἡττηθεῖσιν . . . ταύτας δήπου ϕαμὲν ἡμεῖς νῦν οὔτ’ εἶναι πολιτείας, οὔτ’ ὀρθοὺς νόμους ὅσοι μὴ συμπάσης τῆς πόλεως ἕνεκα τοῦ κοινοῦ ἐτέθησαν· οἳ δ’ ἕνεκά τινων, στασιώτας ἀλλ’ οὐ πολίτας τούτους ϕαμέν. (Laws 4, 715 a 7–b 6)

[In some cities] offices are fought over, and the victorious take over the city’s affairs so completely that they give no share of rule at all to the defeated . . . These we certainly would now say are not constitutions, nor are any laws correct that have not been established for the sake of

6  Also, if the structure remains unchanged while the citizens become happier, the city’s happiness increases, but Morrison must deny this. 7  This is different from the mathematical property of unity constituting non-­ relational goodness. Cantu, ‘Individual and Polis’; Schofield, Political Philosophy, 214–27; D. Scott, ‘The Republic’, in G. Fine (ed.), The Oxford Handbook of Plato (Oxford, 2019), 208–27 at 217–18. 8 T. Irwin, Plato’s Ethics [Ethics] (Oxford, 1995), 298–317, persuasively argues that virtue involves such unity.

110

Christopher Bobonich what is common to the entire city; where they are for the sake of some, we say that these consist of ‘partisans’ not citizens.

[T3] states at least a necessary condition for laws’ correctness, which is that they aim at what is common to the entire city. Is this simply the ‘common good’, that is, the citizens’ greatest happiness or is it a weaker condition, for example, that the laws do not entirely ignore some citizens’ well-­being? Laws 9, 875 a 2–c 2 claims that the true ‘political art’ (ἡ πολιτικὴ τέχνη) must care for what is common, which is glossed as ‘the best’ (τὸ βέλτιστον; see also 1, 628 c 6–7; 2, 664 a 3; 9, 875 a 4; 11, 923 b 2–7, 925 e 7–926 a 3). The lawgiver should thus optimize. This is made more precise in two ways. [T4]  οἱ Κρητῶν νόμοι οὐκ εἰσὶν μάτην διαϕερόντως ἐν πᾶσιν εὐδόκιμοι τοῖς Ἕλλησιν· ἔχουσιν γὰρ ὀρθῶς, τοὺς αὐτοῖς χρωμένους εὐδαίμονας ἀποτελοῦντες. πάντα γὰρ τἀγαθὰ πορίζουσιν. (Laws 1, 631 b 3–6)

There is good reason for Cretan laws to be held in the highest possible esteem among all the Greeks. For they are correct, because they make those using them happy, for they provide all the good things.

[T5]  ἡμῖν δὲ ἡ τῶν νόμων ὑπόθεσις ἐνταῦθα ἔβλεπεν, ὅπως ὡς εὐδαιμονέστατοι ἔσονται. (Laws 5, 743 c 5–6, esp. with 742 d 7–e 1)

The hypothesis of our laws aims at making the citizens as happy as ­pos­sible.

[T6]  τέλεον γὰρ καὶ οὐ διήμισυν δεῖν τὸν νομοθέτην εἶναι, τὸ θῆλυ μὲν ἀϕιέντα τρυϕᾶν καὶ ἀναλίσκειν διαίταις ἀτάκτως χρώμενον, τοῦ δὲ ἄρρενος ἐπιμεληθέντα, τελέως σχεδὸν εὐδαίμονος ἥμισυ βίου καταλείπειν ἀντὶ διπλασίου τῇ πόλει. (Laws 7, 806 c 3–7)

A lawgiver must be complete, and not half a lawgiver; letting the female indulge in luxury and expense and follow disorderly pursuits, while supervising the male, is to leave the city with only about half of a completely happy life, instead of double that.

[T4] claims that making the citizens happy is at least sufficient for laws being correct. [T5] claims that the legislator optimizes by making the citizens as happy as possible. I think that there is good reason to see maximizing the citizens’ happiness as also a necessary condition of correctness. First, the Laws never mentions any other conditions for correctness. If there are no other sufficient conditions, then maximizing happiness is also necessary. Second, [T5] and [T6] mention maximizing happiness as the only relevant condition, and satisfying it settles the question of what laws to adopt.9 9 Cf. Laws 2, 664 a 2–3, 665 c 2–5; 3, 683 b 3–4; 9, 858 c 10–d 9; 12, 945 e 3–6, 958 a 2–3.



Is Plato a Consequentialist?

111

Third, in the rest of the Laws, no other consideration overrides maximizing happiness.10 Finally, the greatest concern about the maximizing happiness condition is that the Laws mentions other goals for laws (although not for their correctness), including achieving the best (1, 628 c 6–7; cf. 2, 664 a 3), inculcating all the virtues in the citizens (4, 705 d 2–706 a 4; 12, 962 b 4–963 a 9), and the trio of goals: making the city free, friendly to itself, and wise (3, 693 b 3–e 3; 701 d 6–e 8).11 The Athenian notes this variety of goals but claims that they ‘are not different, but the same’, as will be any others he mentions (ὁ σκοπὸς οὐχ ἕτερος ἀλλ’ ὁ αὐτός, 3, 693 b 6–c 6). Since these are clearly not all the same property or interentailing, the most plausible interpretation is that all the other goals are justified by their instrumental or constitutive contribution to one goal, and the only reasonable candidate for this goal is the citizens’ maximal happiness.12 The ‘best’, for example, just picks out the citizens’ maximal happiness (Laws 2, 662 d 1–663 a 7, with 663 e 8–664 a 3). Similarly, inculcating virtue is justified by maximizing happiness. The Laws holds a Dependency Theory of relational goods and bads, that is, of things that are good or bad for people: Dependency Thesis: All goods not identical with virtue or including it as a part are Dependent Goods. G is a Dependent Good if and only if G is good for a virtuous or just person and G is bad for an unvirtuous or unjust person. B is a Dependent Bad if and only if B is bad for a just person and B is bad for an unjust person, but is less bad for the unjust person than the corresponding Dependent Good. G is an Independent Good if and only if G is good for a person regardless of what else she possesses. B is an Independent Bad if and only if B is bad for a person regardless of what else she possesses.13 Since no good benefits without virtue and virtue is in itself a great good, the goal of making the citizens virtuous is justified by the 10  Section 4.3 discusses whether the Laws’ ‘Proportionality principle’ is justified independently of maximizing happiness. 11  Laws 3, 701 c 6–7 says that there are three such goals for the city; ‘wise’ translates ἔμϕρων at 3, 693 b 4 and νοῦν ἔχειν at 3, 701 d 9. 12  They are not interentailing, since the other goals can be achieved without achieving the best (if, for example, happiness is degreed or virtue is not sufficient for happiness). 13  Euthd. 278 e 3–282 d 3; Laws 1, 631 b 3–d 6; 2, 660 d 11–662 c 5; Meno 87 c 5–89 a 7. See C. Bobonich, Plato’s Utopia Recast [Utopia] (Oxford, 2002), 123–209.

112

Christopher Bobonich

goal of maximizing happiness (as [T6] suggests). In contrast, the goal of making the citizens virtuous does not necessitate maxi­miz­ ing the citizens’ happiness, since the virtuous can be benefited by Dependent Goods in ways going beyond their instrumental value in developing virtue (cf. Laws 4, 705 e 3–706 a 4). A similar strategy applies to the other goals.14 [T4–6] (with 816 c 7–d 2) show that the Laws is reductionist about the city’s happiness.15 [T6] is compatible with maximizing either average or total well-­being, but since it implicitly rejects leaving women uneducated while excluding them from citizenship, it suggests something like the latter.16 We thus obtain the following fundamental criterion: H: (i) A set of laws, L, that governs city C is correct if and only if L produces the greatest possible happiness for C’s citizens, and (ii) the correctness of a set of laws is a decisive reason for enacting them.17 Before turning to H’s relation to consequentialism, let me note some clarifications. Plato, however, may not have a determinate answer to some of our contemporary questions, since so much sophisticated work has been done on maximizing principles since the nineteenth century. 1. H’s criterion is maximal happiness, but H is restricted to the happiness of C’s citizens.

14 Bobonich, Utopia, 577–8 n. 107. 15  [T4–6] also support reductionism for the Republic, since there is no reason for Plato to change his mind. 16  The distinction between maximizing average or total welfare seems obvious, but classic Utilitarians such as Jeremy Bentham and John Stuart Mill did not make it (Henry Sidgwick first drew it); see J. Driver, ‘The History of Utilitarianism’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2014 edn), https://plato.stanford.edu/archives/win2014/entries/utilitarianism-­history/, accessed 15 January 2023. Parfit’s Repugnant Conclusion would trouble Plato less than Parfit, because Plato thinks that overpopulation causes poverty, which worsens citizens’ ethical characters. In Kallipolis, overpopulation would result from more eugenically undesirable births. The excess citizens will tend to lack virtue and, given the Dependency Theory, will not have lives worth living. Thus, their addition will not increase total well-­being. 17  H also applies to practices the lawgiver encourages without legislating (Laws 7, 788 a 1–790 b 6) and her contravening the law (6, 770 d 5–e 5; Polit. 295 b 10–297 b 4). The Republic is less explicit about what the city’s happiness is, but still accepts H.



Is Plato a Consequentialist?

113

2. Does H concern expected or actual consequences? If actual consequences, then laws that normally produce bad outcomes and are enacted with unjust intentions might, via deviant caus­ al­ity, produce the greatest happiness. If expected consequences, then the laws with the highest expected value might aberrantly produce disaster. On either version, human epistemic limitations (Laws 4, 709 a 1–d 7) make it impossible to be certain that one’s chosen L is correct. Plato does not address these issues and might have thought that, at least for knowledge­ able legislators, they typically would not diverge greatly. But this does not call into question Plato’s commitment to H, and similar issues recur in contemporary moral philosophy.18 3. H is a principle for the correctness of sets of laws in the way that act-­consequentialism is a criterion of morally right action. Neither is in itself a decision procedure. But Plato does intend Magnesian legislators to use it as a decision procedure.19 Deliberation about whether to adopt a given law, L, should be understood as comparing the outcome of retaining the current laws and that of supplementing them with L. 4. H explicitly specifies maximizing citizens’ happiness as the fundamental evaluative property that makes laws correct and implicitly specifies an action-­guiding principle for le­gis­la­tors. I cannot discuss here why individual legislators or citizens should follow such laws. But for the project of more generally understanding normativity in Plato, it is important that H characterizes a fundamental evaluative and normative property of institutions, not individuals, and is justified by maxi­ miz­ing all the citizens’ happiness, not that of any particular person. Throughout his career, Plato, I think, accepted the principle of rational eudaemonism, that is, for every person the ultimate end of all her rational actions is her own (greatest) happiness. More specifically, a person acts rationally if and only if her action aims at her own (greatest) happiness for its own sake and not for the sake of anything else.20 Although 18 So, ἐτέθησαν (‘have been established’, Laws 4, 715 b 4, in [T3] above) should not be seen as evidence for expected consequences. 19  Laws 7, 806 c 3–7, and since, as we shall shortly see, inculcating the virtues is justified by H, so are 4, 705 d 2–706 a 4; 12, 962 b 4–963 a 9. 20  C. Bobonich, ‘Socrates and Eudaimonia’, in D. Morrison (ed.), The Cambridge Companion to Socrates (Cambridge, 2011), 293–332. On the principle’s formulation,

114

Christopher Bobonich

this principle should play an important role in justifying the correctness criterion of laws, it does not entail it. It would have to be supplemented by principles about the evaluative standards for institutions and principles connecting the two. 5. In the Laws, correctness is a teleological notion: an activity is correct when it attains its intended goal (Laws 2, 667 d 9–e 1). This goal is internal to the activity in that it helps to constitute the activity’s identity and specifies what counts as its successful completion. In the case of laws, [T4–6] show that their correctness consists in achieving the goal of providing benefit, that is, the citizens’ maximal happiness.21 2.2.  Principle H and consequentialism Let us start with an explicit characterization of the most basic form of consequentialism, act-­consequentialism. Act-­consequentialist theories: first specify some principle for ranking overall states of affairs from best to worst from an impersonal point of view . . . then require that each agent in all cases act in such a way as to produce the highest-­ranked state of affairs that he is in a position to produce.22

This a requirement of morality, since act-­consequentialism is a theory of actions’ moral rightness. The rankings are ‘impersonal’ insofar as they are ‘agent-­neutral’, not ‘agent-­relative’, that is, they are the same for all people. Deontological views often permit bringing about something other than the best outcome and sometimes forbid it. Act-­consequentialism and H differ in several ways: the former is a criterion for actions’ moral rightness; H is a criterion for laws’ correctness.23 Thus, the sort of normativity in act-­consequentialism prima facie differs significantly from H’s. I must here largely leave

see T. Irwin, Plato’s Ethics [Ethics] (Oxford, 1995), 52–3. This allows someone to pursue an end both for its own sake and for the sake of happiness. 21  For more extended discussion of correctness in the Laws, see C.  Bobonich, ‘Plato on Legal Normativity’ [‘Normativity’], Ancient Philosophy Today, suppl. vol. 4 (2022), 24–44. 22 S.  Scheffler, The Rejection of Consequentialism (Oxford, 1982), 1. I state the definition in terms of actual consequences to make the theory’s structure clearer. 23  Although H concerns laws, it is connected with individual action, since Plato expects citizens to obey correct laws.



Is Plato a Consequentialist?

115

aside questions about the nature of normativity in Plato but will occasionally return to them.24 Act-­consequentialism and H share two fundamental features: both rank outcomes agent-­neutrally (within a target group) and require that the best one be chosen (bestness is a decisive criterion for choice). This is sufficient for H being a form of consequentialism as I understand it here. H’s agent-­neutrality, however, is not unrestricted. H’s ranking of outcomes is the same for anyone who is a lawgiver for a given city. It is not fully agent-­neutral in that it has no force for those for whom being a good lawgiver for that city is irrelevant. H establishes a norm for all lawgivers that the laws to be adopted by each are those maximizing the happiness of her own city. This principle is agent-­relative, since the ranking of outcomes differs for different cities. (The outcome in which Sparta has high well-­being and Athens has low well-­being is ranked high by Spartan lawgivers and low by Athenian lawgivers.)

3.  Objections to consequentialism in Plato Many working in ancient ethics find consequentialism to be distinctively modern and highly defective as a moral theory. So let us consider some common objections to finding consequentialist tendencies in Plato. Perhaps the most frequent is that welfarist consequentialism’s goal is bringing about overall happiness, but Plato does not think that one person can bring about happiness for another. Happiness requires having a certain character and using one’s goods correctly, and one person cannot directly bring these about in another. Consequentialism is thus only appropriate if one has a non-­Platonic conception of the good such that one person could directly produce it for another, for example, pleasure as a quale. Consequentialism, however, only requires bringing about the best outcome that is possible, so if the above claim is correct, it will not require causing happiness, but should rank highest those outcomes that are most conducive to happiness. Moreover, Plato describes what laws do in causal terms (e.g. Laws 1, 631 b 3–6) and allows external events to have psychic effects (Tim. 42 a 3–c 1).

24  See Bobonich, ‘Normativity’.

116

Christopher Bobonich

A second common objection is that Plato cannot be a consequentialist because he is a rational eudaemonist. He thinks that one’s ultimate goal is maximizing one’s own good—­not the overall good. Consequentialism, instead, holds that one benefits others because one’s goal is not one’s own well-­being, but overall well-­being. This objection, however, misunderstands both my thesis and consequentialism. I agree that Plato is a rational eudaemonist and not an act-­ consequentialist: he does not, for example, use the notion of ‘morally right action’. More importantly, H is a consequentialist account of the laws’ correctness, not of individual actions. But consequentialism is a general doctrine that is applicable to anything having consequences (for example, rules, motives, and character states), and this includes laws. Further, although many contemporary consequentialists believe that one always has most reason to act morally, this is logically independent of act-­consequentialism, which is an account of morally right action, not of what one has most reason to do. Julia Annas raises the different objection that ancient ethics, except for the Cyrenaics: never develops anything like the . . . consequentialist idea of a maximizing model of rationality. If my ethical aim is to produce a good, or the best, state of affairs, then it is only rational to produce as much as possible of it. But ancient ethics does not aim at the production of good states of affairs, and so is not tempted to think that rationality should take the form of maximizing them. Rather, what I should aim at is my living a certain way, my making the best use of goods, and acting in some ways rather than ­others. None of these things can sensibly be maximized by the agent. Why would I want to maximize my acting courageously, for example? . . . I have no need, normally, to produce as many dangerous situations as possible, in order to act bravely in them.25

But consequentialism requires bringing about the best outcome, not producing as much as possible of (or maximizing) a good or the best state of affairs (and it is hard to see what that could be). If, as many do, we think of states of affairs as particulars (possibly) exemplifying properties, it seems hard to deny this notion to Plato, even if he did not use that terminology. Thinking about alternatives, whether different courses of individual action or different sets of laws, seems to involve thinking about different ways that things 25 J. Annas, The Morality of Happiness [Morality] (Oxford, 1993), 38; cf. 274, 334, 447–8.



Is Plato a Consequentialist?

117

could be and thus about states of affairs. If states of affairs are types, then one could maximize the number of their tokens, but if a token action (say drinking this soy latte) is good, it hardly follows that its numbers or temporal or spatial extension should be maximized. What consequentialists do want to maximize is the goodness of the chosen state of affairs.26 Nor are consequentialists committed to maximizing everything, and even if acting courageously contributes to well-­being, they are not committed to maximizing the number of dangerous situations calling for courage. This would be self-­defeating, since manufacturing such situations (or rashly plunging into them) involves ­performing bad actions and having bad dispositions (cf. Arist., NE 10. 7, 1177b4–15). And why deny that an outcome in which a person is more virtuous or engages in more virtuous activity is, ceteris paribus, better? It seems entirely reasonable to talk of making the best possible use of my goods, so why cannot my use of goods be optimized? It is especially hard to accept Annas’s claim that the rationality of maximizing the good has no place in ancient ethics. Although I must be brief, I think that there is strong evidence that the rationality of maximizing the good is fundamental to Plato’s practical philosophy. Intrapersonally, the Republic asks whether the just life is happier than the unjust one and assumes the rationality of choosing the happier option, that is, the greater good. This is, for the Republic, what it is to be practically rational.27 Its argument that the just person’s life is happier than the tyrant’s justifies both the intrapersonal judgement that I should lead a just life and the interpersonal judgement that the just person is happier than the tyrant.28 26  Philippa Foot objects more radically that we should reject outcomes as consequentialism understands them. But Foot herself, as well as many deontologists, gives some place to maximizing the good: once other moral duties are satisfied, Foot accepts that the virtue of benevolence urges saving more rather than fewer lives because that is the better overall outcome (‘Utilitarianism and the Virtues’, Mind, 94 (1985), 196–209 at 206). On Foot, see S. Scheffler, ‘Agent-­Centered Restrictions, Rationality, and the Virtues’, Mind, 94 (1985), 409–19. 27  The reasoning part’s job is ruling and deciding what is best for the whole soul (4, 441 d 11–e 5, 442 c 4–8) and the other parts’ job is to support enacting this decision. 28  I cannot defend here every claim I make about the Republic and the Laws but shall try to make my assumptions clear and to defend what is uncommon. Plato’s first answer to the question about the comparative goodness of just and unjust lives is more than ordinal because it assumes a non-­arbitrary zero point of well-­being: justice profits and injustice does not (Rep. 4, 444 e 6–445 a 4). This answer is

118

Christopher Bobonich

The Laws also assumes the rationality of seeking the happiest pos­sible life and claims that the most just life is happiest (2, 661 a 4–663 c 6) and so makes intrapersonal and interpersonal comparisons of well-­being. I am not claiming that Plato has a worked-­out theory of good maximization that answers every question that contemporary philo­sophers or rational choice theorists might ask. He lacks, for example, a theory of probability and modern mathematical techniques. He does not develop a clear and precise account of aggregating goods and bads within lives or of how to make intrapersonal and interpersonal comparisons of well-­being. We can ask whether he uses cardinal or ordinal comparisons of well-­being.29 Yet even ordinal comparisons allow ranking of lives and outcomes such that we can identify the life or outcome that is better than all others.30 We can also ask whether Plato thinks that comparisons of goodness require a common metric or a qualitatively homogeneous property of goodness.31 But as we have seen, Plato accepts the possibility of intrapersonal and interpersonal comparisons of well-­being and comparisons of outcomes’ goodness. In every case, he holds that incomplete and Book 9 returns to the issue. The first argument’s conclusion at 580 a 8–c 9 is ordinal. The second and third arguments explicitly concern pleasure, not happiness. But if they are relevant to our question, it is reasonable to assume that if ordinal and cardinal judgements about pleasure are possible, they are also possible about happiness. The second conclusion is ordinal (582 e 8–583 a 4), but 581 d 9–e 4 goes beyond ordinality by suggesting that the just person lives much more pleasantly. The third conclusion at 587 a 8–d 1 is cardinal. 29  Ordinal comparisons are frequent. The cardinal comparison at Rep. 9, 587 b 12–588 a 10 might seem unserious, but the humour may concern not the possibility of cardinality but rather the method of calculating it. Similarly, Plato’s discussion of the ‘Nuptial Number’ (Rep. 8, 545 d 5–547 c 9) has humorous aspects, but he may well have thought that appropriate procreative times could, in principle, be fairly accurately calculated. [T6]’s cardinality seems entirely serious. 30  This assumes transitivity and the ranking of relevant items pairwise such that one is better than the other or they are equally good. For convenience, I leave aside ties. See J. Broome, Weighing Goods (Cambridge, Mass., 1991); R. Chang, Making Comparisons Count (New York, 2015); N. Hsieh and H. Andersson, ‘Incommensurable Values’, in E.  N.  Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2021 edn), https://plato.stanford.edu/archives/fall2021/entries/value-­incommensurable/, accessed 16 January 2023. 31  E.g. T. Irwin, ‘Review of M. Nussbaum, The Fragility of Goodness’, Journal of Philosophy, 85 (1988), 376–83 at 380; M. Nussbaum, ‘The Speech of Alcibiades: A Reading of the Symposium’, in id., The Fragility of Goodness (Cambridge, 1986), 165–99 at 179–80; A. Price, ‘Martha Nussbaum’s Symposium’, Ancient Philosophy, 11 (1991), 285–99.



Is Plato a Consequentialist?

119

reason requires maximizing the relevant good. This commitment is a central part of his ethical and political theory and theory of practical rationality. Two other objections claim that consequentialism’s structure makes it unacceptable to Plato. The first is that consequentialism denies that actions are good or bad in themselves and holds that they are only instrumentally valuable in producing outcomes.32 Some philosophers do understand act-­consequentialism this way, such as Elizabeth Anscombe, Dale Dorsey, and G. E. Moore in Ethics. But many philosophers—­such John Broome, G. E. Moore in Principia Ethica, Douglas Portmore, Sam Scheffler, and Bernard Williams—­ build the act into the consequences and thus evaluate it as part of the outcome.33 Since consequentialists typically aggregate goods and maximize the overall good, it seems odd to exclude the action from consideration or insist that it can only have instrumental value. Even the simplest form of consequentialism, hedonistic act-­ consequentialism, should allow that the action can be pleasant or painful. But again, what is important is that this criticism is directed against act-­consequentialism. In Plato’s case, the action of enactment typically involves few people and occurs over a short time, while the subsequent consequences involve many people over a long time. Since the value of the latter is much greater, the issue of whether to count the action as part of the outcome is not prac­ tic­al­ly important. The second objection is that consequentialism must be understood as a teleological rather than deontological ethical theory as John Rawls characterizes them. According to Rawls, teleological theories define the good independently of the right and define the right as that which maximizes the good. Deontological theories are those 32  M. McCabe, ‘Out of the Labyrinth: Plato’s Attack on Consequentialism’, in C. Gill (ed.), Virtue, Norms, and Objectivity (Oxford, 2005), 189–215 at 190. 33  G. E. M. Anscombe, ‘Modern Moral Philosophy’, Philosophy, 33 (1958), 1–16; Broome, Weighing Goods, 3–4; D. Dorsey, ‘Consequences’, in D. Portmore (ed.), The Oxford Handbook of Consequentialism [Consequentialism] (New York, 2020), 92–112; G. E. Moore, Ethics (London, 1965), 77–80, 97; G. E. Moore, Principia Ethica, ed. by T.  Baldwin, rev. edn (Cambridge, 1993), 24–5, 147, 177–9; D.  Portmore, ‘Introduction’, in id. (ed.), Consequentialism (New York, 2020), 1–21; Scheffler, The Rejection of Consequentialism, 1–2; J. Smart and B. Williams, Utilitarianism For and Against [Utilitarianism] (Cambridge, 1987), 81–9. T. Hurka, British Ethical Theorists from Sidgwick to Ewing [British Ethical Theorists] (Oxford, 2014), 165, details the acceptance of building in the action by C.  D.  Broad, A.  C.  Ewing, and Hastings Rashdall.

120

Christopher Bobonich

that are not teleological.34 The objection is that as a teleo­logic­al theory, consequentialism must rank outcomes independently of considerations of rightness and that this is unacceptable to Plato. I agree that Plato thinks that laws’ outcomes are ranked higher, ceteris paribus, insofar as they produce more virtuous citizens. But many contemporary philosophers reject Rawls’s distinction and its implicit characterization of consequentialism. They allow that various moral or ethical features can help constitute the goodness of outcomes; for example, since keeping a promise is an in­trin­sic­ al­ly goodmaking property of an action, the goodness of such actions helps constitute the goodness of outcomes including instances of promise keeping. I think that the latter are correct and that consequentialism is best understood as having the defining features of ranking outcomes agent-­neutrally and requiring that the best one be chosen. Consider a theory that counts killing as a badmaking feature. On the individual level, it thus requires, ceteris paribus, minimizing killings. This gives a place to moral or ethical value in a distinctively consequentialist way, that is, as a feature contributing to the outcome’s overall goodness and thus part of what’s to be maximized. This theory thus requires me to kill one to prevent five killings. Since it characterizes the good at least partially in terms of the right, it is not teleological for Rawls and would thus be deontological. But this seems to be the wrong answer. I close this section with two especially important objections. The first is that although Plato says that the goal of the city’s laws is maxi­ miz­ing the city’s happiness, he implicitly accepts side constraints.35 Consider a doctor. Medicine’s ultimate end is health, but we would not expect her to force me at gunpoint to jog. We take for granted such implicit side constraints on her action to attain her ultimate end. Should this not also be true of the legislator? But, first, in the doctor’s case, the side constraints are largely supplied by the law. Since H is a criterion for evaluating laws, we should not take current laws as side constraints. Second, Plato might not agree that the arts qua arts are subject to such side constraints. The Statesman claims:

34 J. Rawls, A Theory of Justice [Theory of Justice], rev. edn (Cambridge, Mass., 1999), 21–4, 35–6. 35  Sarah Broadie mentioned such a possibility to me without endorsing it.



Is Plato a Consequentialist?

121

[T7]  τοὺς ἰατροὺς δὲ οὐχ ἥκιστα νενομίκαμεν, ἐάντε ἑκόντας ἐάντε ἄκοντας ἡμᾶς ἰῶνται, τέμνοντες ἢ κάοντες ἤ τινα ἄλλην ἀλγηδόνα προσάπτοντες . . . πάντως οὐδὲν ἧττον ἰατρούς ϕαμεν, ἕωσπερ ἂν ἐπιστατοῦντες τέχνῃ . . . ἂν μόνον ἐπ’ ἀγαθῷ τῷ τῶν σωμάτων, βελτίω ποιοῦντες ἐκ χειρόνων . . . ταύτῃ θήσομεν, ὡς οἶμαι, καὶ οὐκ ἄλλῃ, τοῦτον ὅρον ὀρθὸν εἶναι μόνον ἰατρικῆς καὶ ἄλλης ἡστινοσοῦν ἀρχῆς. (Polit. 293 b 1–c 3)36

we should believe in them [sc. doctors] whether they cure us with or without our consent, by cutting or burning or applying some other painful treatment . . . no less do we call them ‘doctors’ so long as they govern us by art . . . if only each . . . acts for our bodies’ good, making them better than they were . . . this is the only correct criterion of medicine and of any other sort of rule whatsoever.

Arts such as medicine, however, are under political science’s control, which governs them in accordance with H. But there is no higher science limiting political science’s pursuit of its ultimate end. Plato’s position is also strengthened by the Dependency Theory, since for the unvirtuous their Dependent Bads are less bad for them than the corresponding Dependent Goods. The second objection is that Plato might not fully realize and accept all the consequences of his commitments. H uses the cri­ter­ion of the greatest happiness for a given city and is thus city-­relative. But there is no room for further relativization in this criterion. It is the same ranking for any citizen within a given city and is thus citizen-­neutral. But this sort of agent-­neutrality entails that all agent-­centred constraints (absolutist prohibitions, for example, against harming) are excluded.37 Since Plato did not explicitly formulate a conception of agent-­neutrality, he might not have seen this implication. I consider this issue in Section 4.3.

4.  Evidence for and against consequentialism in Plato 4.1. Demandingness As we have seen, some of Plato’s claims, taken literally, commit him to a form of consequentialism. In this section, I consider his position on some of consequentialism’s most distinctive and controversial 36  For the Greek text, I have used E. A. Duke, W. F. Hicken, W. S. M. Nicoll, D. B. Robinson, and J. C. G. Strachan (eds.) Platonis Opera, vol. i (Oxford, 1995). 37 Scheffler, The Rejection of Consequentialism, 5, 22–6.

122

Christopher Bobonich

features. Two inconsistent propositions can, of course, entail the same conclusion. But if Plato embraces these features, his acceptance of consequentialism seems to be the best ex­plan­ation. One of consequentialism’s most notorious divergences from common sense is that it is highly demanding. This is seen most clearly and has received the most attention with respect to act-­ consequentialism, but it is true to a greater or lesser extent of any form of consequentialism that in some way requires maximizing the impersonal good.38 Producing the best outcome typically requires great expenditures of time and energy and severely restricts agents’ pursuit of personal projects and relationships. Peter Singer, for example, requires people to sacrifice resources until their well-­ being equals that of the worst-­off.39 There are two contemporary responses to this worry: ‘Extremism’ and ‘Moderation’.40 ‘Extremists’, such as Shelly Kagan, accept that consequentialism is extremely demanding, but deny its un­rea­son­able­ness: morality just is that demanding. So, what is Plato’s position? First, he accepts along with Extremists that H requires laws that demand great reallocations of individual resources. The Republic’s auxiliaries and guardians must forgo the projects and relationships that ordinary people have and devote enormous amounts of time and energy to promoting the common good. The extent of these renunciations shows that Plato accepts consequentialism’s rigorous demand to optimize. Some Moderate consequentialists simply deny that consequentialism makes such extreme demands. They argue, for example, that because of the difficulty of determining our actions’ effects in distant places we do better by devoting our resources to our near and dear.41 Plato dis­ agrees: he does not think that we know so little about what is good for others or our actions’ consequences that concentrating on our near and dear optimizes. Other Moderate consequentialists accept that consequentialism’s demands are extreme and so are sometimes unreasonable. They modify consequentialism by not requiring bringing about the best 38  For more, see T. Mulgan, The Demands of Consequentialism (Oxford, 2001). 39  P.  Singer, ‘Famine, Affluence, and Morality’, Philosophy and Public Affairs, 1 (1972), 229–43. 40  Mulgan introduces these terms at The Demands of Consequentialism, 25. 41 Mulgan, The Demands of Consequentialism, 31–7; H. Sidgwick, The Methods of Ethics, 7th edn [Methods] (Indianapolis, 1981), 256–9, 430–9.



Is Plato a Consequentialist?

123

overall outcome and build in moral permissions to agents (‘agent-­ centred options’) to perform, within limits, suboptimal acts. (‘Agent-­ centred constraints’ which are found in deontological theories sometimes prohibit performing the best option.) We might think that Plato’s Laws endorses such Moderation in conceding that a second-­best city is to be founded, that is, one accepting private property and families (Laws 5, 739 a 3–5; the first city at 739 a 3–5 is not Kallipolis, but one where the community of property and women extends throughout the entire city). But this would be a mistake. Contemporary Moderates accept suboptimization because they think that optimization sacrifices personal projects or relationships that have a distinct value that morality must accommodate within itself.42 Plato is not permitting Magnesian lawgivers or citizens to suboptimize; rather, he claims that the best outcome of full communism is not psychologically possible: it is too demanding for the birth and education of Magnesians (739 a 4–740 a 2). He accepts that full communism would be better overall and for individuals, but thinks that citizens’ non-­rational motivations are too strong for them to act accordingly.43 Interestingly, Plato does not recommend, as ‘Government House’ consequentialists do, concealing the fact that more impersonal schemes are ideally superior because people attain better actual outcomes by sticking to partialist motivations.44 Passages such as 739 a 4–740 a 2 are part of the citizens’ education, and so they will have some grasp of how they fall short of full practical rationality. But [T1] (cf. Rep. 5, 465 d 3–466 c 4) suggests that Plato denies that H mandates great sacrifices in well-­being (see Section 6). He differs from contemporary Extremists because he thinks that most ordinary projects and relationships lack value, while Kallipolis provides its citizens with truly valuable projects and relationships. An example is the ‘community of pleasure and pain’ which involves 42  E.g. S. Keller, Partiality (Princeton, 2013), 46. 43 This psychological incapacity seems widespread (5, 739 d 6–8) and long-­ lasting. Since future Magnesians will be much superior, we should (but do not) find references to pathways to more communal schemes if they are possible. 44 B.  Williams, Ethics and the Limits of Philosophy (Cambridge, Mass., 1985), 108–9. This is Williams’s famous description of Henry Sidgwick’s suggestion that it might maximize overall if a utilitarian elite used the greatest happiness principle as a decision procedure while encouraging the masses to stick with their traditional deontological morality because most people tend to go wrong in calculating consequences (Methods, 489–92).

124

Christopher Bobonich

the citizens finding many, if not all, of the same things good and bad for themselves (Rep. 5, 462 a 2–e 3).45 This idea’s fullest real­ iza­tion (which Plato rejects) entails that no one sacrifices her happiness to bring about the best overall outcome, since everyone’s happiness is identical with the summation of all individual goods and bads (which 462 a 2–e 3 suggests is how to understand the city’s well-­being).46 So the auxiliaries and guardians, by replacing their old private projects and relationships with this new communal relationship, have increased their own well-­being. But this account faces some problems that have not been fully appreciated. One way of describing this situation is that each participating citizen makes others’ goods and bads part of her own well-­being. If the least intelligent and virtuous auxiliary adds the well-­being of her more virtuous comrades and of the philosopher rulers (whose well-­being includes the good of contemplation which she neither understands nor values) to her own well-­being, she will benefit greatly. But this seems to be a bonus that she does not deserve. After all, it seems clear that Plato would reject the idea that an unvirtuous, but community-­minded person could become happy by sharing in her virtuous fellows’ happiness. If a virtuous philosopher adds to her own well-­being the goods and bads of all her fellow citizens, which she does not value highly, this still seems to benefit her as long as the sum is positive. But this threatens to swamp the goods and bads that depend on her own psychic condition. (Although Plato’s ideal unity is a simple summation of all goods and bads, one could differentially weight the added well-­being of others.) Further, if some of her fellow citizens have nega­tive well-­ being, she would lose by adding them. Does Plato think that in Kallipolis every citizen attains or can attain her maximal happiness? And if he does, does H retain much interest? Does consequentialism have any bite without a conflict of interests? The major ancient ethical thinkers, especially those such as Plato and the Stoics, who accept a providential god, are more likely than contemporaries to hold something like a harmony of interests view.47 Some think that in the Republic Plato believes that 45  Plato has no reason to restrict what is shared to pleasures and pains rather than goods and bads. 464 d 6–e 2 supports the stronger reading. 46  Note that this supports reductionism. 47  A. C. Ewing, T. H. Green, and John McTaggart also hold something like the harmony view; see D. Brink, Perfectionism and the Common Good (Oxford, 2007), 60 and Hurka, British Ethical Theorists, 137, 167.



Is Plato a Consequentialist?

125

the maximal well-­being of each citizen is compossible with the maximal well-­being of every other citizen or even that different citizens’ well-­beings never conflict. As we shall see in Sections 6.3 and 6.4, such claims have great difficulty in avoiding counter-­ examples. But it is at least plausible that in the Republic Plato does not think that such conflicts are frequent or that, if they do occur, making some better off typically requires making others vastly worse off.48 But as we have seen, he remains committed to the basic consequentialist idea that laws should maximize overall happiness, even if this requires making some citizens worse off (Rep. 4, 420 b 3–c 3 and Section 2.1). Here he mentions no limitations on how far individuals’ well-­being should be sacrificed for greater overall happiness (although neither does he assert that there are no such limitations). I return to this issue in Sections 4.3 and 6. It is im­port­ant to note that even he did accept the compossibility thesis at the level of the city this does not entail that even knowledgeable le­gis­la­tors will always be able to know how to achieve it. So it may not be uncommon that they enact laws that they realize will require some to have less happiness than they might.49 As we shall see in Section 6, such compossibility may hold at the level of the universe. 4.2.  The do/allow distinction The do/allow distinction is a philosophical newcomer. A series of 1970s papers offering intuitively compelling cases undermining this common-­sense distinction started the debate. Since the 1990s, there has been a flood of literature on doing/allowing drawing on contemporary work in action theory, causation, and responsibility. But even this distinction’s grandfather—­the doctrine of Double Effect—­had a non-­ancient origin in Saint Thomas Aquinas.50 48 C. D. C. Reeve, Philosopher-­Kings (Princeton, 1988), argues that each citizen is happier in Kallipolis than outside it (33–41, 205–8) and may hold that each is max­ imal­ly happy (231–4). 49  Given Plato’s rational eudaemonism, this is awkward. But citizens who might be happier under another set of laws could still rationally support such laws, since the better outcome might not be accessible. 50  Aristotle’s ‘mixed actions’ are a more distant ancestor. See  F.  Woollard and F. Howard-­Snyder, ‘Doing vs. Allowing Harm’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2016 edn), https://plato.stanford.edu/archives/ win2016/entries/doing-­allowing/, accessed 16 January 2023.

126

Christopher Bobonich

Although it would be anachronistic to assign Plato an exact place in these modern debates, they have illuminating aspects. Modern consequentialists’ commitment to producing the greatest good leads most of them towards understanding ‘consequence’ broadly. Few act-­consequentialists, for example, have independent theories of responsibility restricting an action’s consequences to those for which the agent is morally responsible. Instead, one popular version identifies my action’s consequences with whatever would happen if I performed it (which is much broader than its causal effects).51 This does not distinguish doing from allowing.52 This has two important upshots. First, it provides a highly nega­ tive evaluation of many things that common sense regards as permissible. On this conception, allowing is just as bad as doing: not sending a life-­saving $5,000 charitable donation is just as bad as killing someone to gain $5,000. (This is especially problematic for Moderates accepting agent-­centred options: they permit refraining from substantial life-­saving donations but do not want to permit killing for the same gain.) On the other hand, it requires actions that common sense forbids. I am required, for example, to kill one to prevent five killings. Consequentialists could allow that, even apart from their consequences, killings are, say, four times worse than allowings of deaths. But this does not help much: I am forbidden to kill one to prevent three deaths, but I am still required to kill one to prevent five deaths and to kill one to prevent two killings. Section 4.3 discusses Plato on doing bads; here I consider allowing bads. Plato never formulated a precise conception of consequences and so may never have seen the exact problem about the do/allow distinction. (This should not disqualify him from being a consequentialist: if it did, we should also disqualify Jeremy Bentham and John Stuart Mill.) We are thus left with some indeterminacy in Plato’s views, and I shall not try to specify his position on all pos­ sible Trolley cases. But the most significant contribution of the work on the do/allow distinction was showing that the distinction between causing (for which people are held responsible) and omitting (for which they are not) largely rested on implicit views about what was normal, and that these views were not coherent. Whatever one’s ultimate position, Singer’s work on allowing deaths among the 51  D. Dorsey, ‘Consequences’, in D. Portmore (ed.), Consequentialism, 92–112. 52  Agent-­centred restrictions and evaluator-­relative theories go further in accommodating the do/allow distinction.



Is Plato a Consequentialist?

127

global poor made us aware of things previously ignored. Leaving aside technicalities, consequentialism tends to unsettle our in­tu­ itions about what is normal and to push us towards seeing ourselves as morally assessable for whatever we can affect. There are thus significant affinities between Plato’s views and consequentialism insofar as he expands the sphere of ethical concern to include more of what we can affect. Consequentialism’s denial or narrowing of the do/allow distinction subjects to moral assessment areas of life where deontological views grant people a wide range of permissible actions. They grant people, for example, great discretion over the resources they devote to projects and relationships. Consequentialism is pervasive and applies to all actions and omissions, and since it requires optimizing, it sharply restricts discretionary activities. In this respect, there are significant similarities between consequentialism and the Republic’s and the Laws’ efforts to radically reduce the private. The Republic holds that the lawgiver should aim at making the city one (Rep. 5, 462 a 3–b 2). What unifies Kallipolis is the community of pleasure and pain which involves citizens seeing others’ well-­being as part of their own, while what the good lawgiver opposes is the ‘privatization’ (ἰδίωσις) of judgements about good and bad that draw the city apart (462 b 8). The Laws’ commitment to unity and opposition to privatization is, if anything, stronger: [T8]  τοῦτ’ οὖν εἴτε που νῦν ἔστιν εἴτ’ ἔσται ποτέ—­κοινὰς μὲν γυναῖκας, κοινοὺς δὲ εἶναι παῖδας, κοινὰ δὲ χρήματα σύμπαντα—­καὶ πάσῃ μηχανῇ τὸ λεγόμενον ἴδιον πανταχόθεν ἐκ τοῦ βίου ἅπαν ἐξῄρηται, μεμηχάνηται δ’ εἰς τὸ δυνατὸν καὶ τὰ ϕύσει ἴδια κοινὰ ἁμῇ γέ πῃ γεγονέναι, οἷον ὄμματα καὶ ὦτα καὶ χεῖρας κοινὰ μὲν ὁρᾶν δοκεῖν καὶ ἀκούειν καὶ πράττειν, ἐπαινεῖν τ’ αὖ καὶ ψέγειν καθ’ ἓν ὅτι μάλιστα σύμπαντας ἐπὶ τοῖς αὐτοῖς χαίροντας καὶ λυπουμένους, καὶ κατὰ δύναμιν οἵτινες νόμοι μίαν ὅτι μάλιστα πόλιν ἀπεργάζονται, τούτων ὑπερβολῇ πρὸς ἀρετὴν οὐδείς ποτε ὅρον ἄλλον θέμενος ὀρθότερον οὐδὲ βελτίω θήσεται. (Laws 5, 739 c 3–d 6)

Whether this exists anywhere now or will ever exist—­that women, children, and every sort of property are common—­if by means of every device what is called ‘private’ has been entirely removed from all aspects of life, if insofar as possible a way has been devised to make even the things that are naturally private somehow common, in the way eyes, ears, and hands seem to see, hear, and do things in common, if, again, everyone praises and blames as one, as much as possible pleased and pained at the same things, if as far as possible the laws are the sort that make the city most of all one, no one will ever find a more correct

128

Christopher Bobonich or better definition than this of what constitutes the extreme of virtue for laws.

‘Privatization’ involves retaining unshared judgements of good and bad that engender conflicts with one’s fellows. Including ­others’ well-­being within one’s own well-­being gives citizens less reason to grant each other spheres of unregulated private activity. Prudence gives me a reason for regulating my activities to maximize my well-­ being. If your well-­being is part of mine, I have a direct reason for regulating you. Similarly, the idea that your unhappiness is not my concern unless I have caused it loses force. We need not detail the myriad ways that Kallipolis (especially for the upper two classes) and Magnesia (to a lesser extent) require their citizens to act for the common good. I consider some ex­amples shortly, but let us note a contemporary contrast. We find Athenian democratic ideology’s claim to grant citizens freedom to ‘live as you like’ in Pericles’ Funeral Oration: [T9]  ἐλευθέρως δὲ τά τε πρὸς τὸ κοινὸν πολιτεύομεν καὶ ἐς τὴν πρὸς ἀλλήλους τῶν καθ’ ἡμέραν ἐπιτηδευμάτων ὑποψίαν, οὐ δι’ ὀργῆς τὸν πέλας, εἰ καθ’ ἡδονήν τι δρᾷ, ἔχοντες, οὐδὲ ἀζημίους μέν, λυπηρὰς δὲ τῇ ὄψει ἀχθηδόνας προστιθέμενοι. ἀνεπαχθῶς δὲ τὰ ἴδια προσομιλοῦντες. (Thucydides 2. 37. 2–3)53

A spirit of freedom governs our conduct, not only in public affairs but also, in managing the small tensions of everyday life, we are not angry at our neighbours’ choice of pleasures, nor cast aspersions that may hurt even if they do not harm. We associate privately in this easy-­going spirit.54

Although Athenian citizens had no right of non-­interference against government, Athens’s consistent practice was that if personal conduct did not: materially harm others, violate another citizen’s household, or affect community obligations, it was the democracy’s . . . principle not to legislate personal conduct. The freedom to live as one likes is reflected in the total absence of laws governing the personal conduct of private citizens.55 53  For the Greek, I rely on Thucidydes, Historiae, vol. i, ed. H. Stuart-­Jones and J. E. Powell (Oxford, 1942). 54  Translation, with modification, from J. Mynott (trans.), Thucydides: The War of the Peloponnesians and the Athenians (Cambridge, 2013). Cf. Arist., Pol. 6. 2, 1317a40–b14; A. Laks, ‘Private Matters in Plato’s Laws’, in C. Horn (ed.), Platon: Gesetze/Nomoi (Sankt Augustin, 2014), 165–88. 55 R.  Wallace, ‘The Legal Regulation of Private Conduct at Athens: Two Controversies on Freedom’, Etica & Politica/Ethics & Politics, 9 (2007), 155–71 at 161.



Is Plato a Consequentialist?

129

In Kallipolis, the auxiliaries and the guardians, at least until past reproductive age, only have sex during the procreative matches the rulers arrange. So, even their most private activities are strictly regulated for the common good. Kallipolis does not require citizens to devote resources to the global poor, but auxiliaries’ and guardians’ sacrifice of resources for the common good is as extensive as that endorsed by the most Extreme consequentialist. In both cases, these requirements outrage the common-­sense in­tu­ition that, in areas of life where individuals are not harming others, they should be able ‘to live as they like’. In Magnesia, citizens are expected to watch for lawbreaking and be eager to inform: [T10] τίμιος μὲν δὴ καὶ ὁ μηδὲν ἀδικῶν, ὁ δὲ μηδ’ ἐπιτρέπων τοῖς ἀδικοῦσιν ἀδικεῖν πλέον ἢ διπλασίας τιμῆς ἄξιος ἐκείνου· ὁ μὲν γὰρ ἑνός, ὁ δὲ πολλῶν ἀντάξιος ἑτέρων, μηνύων τὴν τῶν ἄλλων τοῖς ἄρχουσιν ἀδικίαν. ὁ δὲ καὶ συγκολάζων εἰς δύναμιν τοῖς ἄρχουσιν, ὁ μέγας ἀνὴρ ἐν πόλει καὶ τέλειος οὗτος ἀναγορευέσθω, νικηϕόρος ἀρετῇ. (Laws 5, 730 d 2–7)

He who does no injustice is honourable, but he who does not allow unjust people to do injustice is more than twice as honourable. For the former counts as one, but the latter by reporting others’ injustice to the rulers counts as many. But he who helps the rulers punish must be declared the great and perfect man in the city, and victorious in virtue.

In Athens those alleging that they acted from public spirit and not from private enmities when bringing public suits tended to be suspected of being busybodies or sycophants.56 Given Magnesian education’s excellence, it is disgraceful that any citizens are lawbreakers, but this makes it vital to find them and impose curative punishment or to separate them from the city (Laws 9, 853 b 4–7). Nevertheless, it is distasteful to think that virtue’s perfection consists in monitoring and informing upon one’s fellow citizens and assisting in their punishment. The Gorgias claims that punishing justly is unenviable (468 e 6–469 a 5) and so, apparently, not a significant good (although insofar as it is a virtuous action, it should benefit the punisher). The Laws calls a man ‘great’ because of the amount of injustice he prevents. So, at least one kind of character perfection is instead understood, at least 56 Aeschin., In Tim. 1; D. Allen, ‘Changing the Authoritative Voice: Lycurgus’ Against Leocrates’, Classical Antiquity, 19 (2000), 5–33 at 16.

130

Christopher Bobonich

in  large part, in terms of the magnitude of bad consequences ­prevented. Kallipolis goes further in eradicating the private from the two upper classes’ lives than Magnesia, but the latter’s invasiveness is still striking. Young couples are instructed to take thought to prod­uce the best children for the city (Laws 6, 783 d 8–e 1) and are supervised by inspectors who freely enter their homes, reprove their errors, and report any young couple found paying attention to other tasks (784 a 1–b 1; 784 b 6–c 5). 4.3.  Doing bad things A notorious issue for consequentialism is that it sometimes requires doing something bad to prevent a worse outcome. Ethical common sense and deontological views often prohibit such actions by imposing ‘agent-­centred constraints’. These are especially challenging for consequentialists, since they cannot be justified, as agent-­centred options can, by the value of the agent’s projects and relationships. Such cases are more common at the level of individual actions but can still arise for laws. Before considering whether H requires doing bad things, let us distinguish, among cases in which I am required to do something seriously overall bad for a person, those where doing bad is necessitated by third-­party wrongdoing and those where it is not. A case of the latter is removing aid from one victim of a natural disaster to save ten victims; of the former, Williams’s famous case of Jim and the Indians.57 Some non-­consequentialists reject doing bad in both cases, but more reject it in the latter. The bad outcome, they say, is others’ responsibility, and I am morally forbidden to do bad to prevent it. Consequentialism’s requirement to do bad in such cases is often said to undermine the agent’s integrity. But before turning to cases, we should remember David Sedley’s warning against expecting the history of philosophy to happen too quickly. The full range of bad things that a consequentialist might be required to do has only recently been explored. Neither Bentham nor Mill considers such cases. A. C. Ewing in a 1927 article was the 57  This is Williams’ case in which an innocent bystander, Jim, is given the choice by a member of a South American death squad of killing one innocent Indian or else he will kill twenty Indians. Williams uses it to argue that consequentialism undermines agents’ integrity. See Smart and Williams, Utilitarianism, 100–18.



Is Plato a Consequentialist?

131

first to raise what is now the standard objection that utilitarianism might require punishing an innocent person.58 [T7] accepts that statesmen acting knowledgeably to benefit the city may justly employ violence and justifies this claim by the ana­ logy of doctors using painful means: the patient suffers a pro tanto harm that is outweighed by overall benefit (Polit. 293 a 6–e 6; cf. 308 e 4–309 a 6). For the statesman, the harsh means—­death, exile, dishonour, and enslavement—­are typically punishments, and in Plato’s penology punishment is primarily intended to benefit the punished.59 So here too, overall benefit outweighs harm.60 The doctor analogy does not support inflicting overall harm on one to prevent more harms (or produce more benefits) for others. The best-­known cases of bad means in Plato—­the ‘Noble Lie’ (Rep. 3, 414 b 7–415 d 5) and rigging Kallipolis’ procreative lotteries (5, 459 c 3–461 e 4)—are not clear cases of doing overall harm either. Book 3 justifies the rulers’ lying whenever it benefits the city (Rep. 3, 389 b 2–10). This is broad enough to include lying to someone to benefit others. But in the Noble Lie’s case, citizens come to believe truly that they should give greater weight to their fellow citizens’ well-­being at the cost of believing falsely that citizens were reared underground. Given the value of believing ethical truths and that the only cost for citizens is the false belief about their rearing, the lie should make them better off.61 There certainly could be cases in which the ruler’s action, although bad for the one lied to, benefits others overall. But Plato does not consider them here.62 Similarly, 58 Hurka, British Ethical Theorists, 165. 59 E.g. Gorg. 476 a 3–478 e 5; Laws 9, 854 d 5–e 1; T. Saunders, Plato’s Penal Code [Penal Code] (New York, 1991), 349–56. 60  Excluding afterlife punishments, death benefits incurables because their lives are not worth living (Rep. 10, 610 d 5–7). Laws 9, 862 d 1–e 1 recommends using as punishments anything effective in reforming character, even pleasures, honours, or gifts. Here too, Plato violates an ordinary ethical norm that wrongdoers deserve something bad to bring about good outcomes. 61  Plato argues (Rep. 2, 382 a 44–c 2) that the goodness of avoiding citizens having false ethical beliefs is much greater than the badness of the individual’s lie (the action). Here the non-­instrumental value of the action and the outcome are combined. 62  Socrates claims that god is always truthful (Rep. 2, 382 d 6–383 a 1) because he meets neither condition justifying lying: (1) in telling edifying stories about an­tiquity in ignorance of the truth, or (2) against enemies or against friends who are about to act badly because of madness or foolishness (Rep. 2, 382 c 7–d 4). God does not meet (2) because the foolish and mad are not dear to god and god does not worry about enemies (382 e 3–4). But foolishness or madness cannot entail in­cur­ abil­ity, since human rulers aid such people. Either god observes a deontological

132

Christopher Bobonich

in eugenic lying (Rep. 5, 459 c 9–e 4), the citizen lied to has a false belief about the lottery’s randomness but avoids the resentment that she would have had if she knew why she was not selected. What about expelling those over 10 from Kallipolis’ territory (Rep. 7, 540 e 4–541 a 7) or its exposure of children (Rep. 5, 459 d 8–e 4, 461 c 2–5)?63 In expulsion, it is implausible that all original inhabitants are ethically unimprovable (even if they could not attain the standard of those reared in Kallipolis), so it seems that some would have benefited by joining Kallipolis. Given Plato’s low estimation of ordinary Greeks’ well-­being and acceptance of the Dependency Theory, he would think that such expulsions are much less bad for the inhabitants than we (or ordinary Greeks) would. Perhaps they do not suffer a major bad (or even an overall bad). Those exposed are the deformed (Rep. 5, 460 c 3) and those born from less desirable parents (Rep. 5, 459 d 9–10, 460 c 2, 461 b 8–c 5). Plato might have uncharitably thought that deformed children could not have lives worth living (cf. Crito 47 d 8–e 6, Rep. 3, 409 e 4–410 a 4), but it is implausible to believe this of all the other children. But perhaps both expulsion and exposure are consistent with H because those affected are not citizens. Citizenship might not attach at birth (Theaet. 160 e 6–161 a 4), but this reply seems arbitrary. The last case is one in which something resembling a deontological principle is violated for overall benefit. Specifically, a prin­ ciple of justice is violated, and some are overall disadvantaged to benefit the city. Laws Book 6 states a principle of political justice. There are two kinds of equality principles (757 a 1–758 a 2). The one assigns honours, including political office, by lot; the other is the truest and best equality and Zeus’ judgement. This principle prohibition against lying or the badness of god having a falsehood in words outweighs any possible benefit. The latter seems preferable, although it would have to be reconciled with god punishing incurables for deterrence and curables for reform. Cf. N. Baima and T. Paytas, ‘True in Word and Deed: Plato on the Impossibility of Divine Deception’, Journal of the History of Philosophy, 58 (2020), 192–214. 63  G. van N. Viljoen, ‘Plato and Aristotle on the Exposure of Infants at Athens’, Acta Classica, 2 (1959), 58–69, argues that such children are raised secretly, not exposed (Rep. 5, 460 c 1–5, Tim. 19 a 1–5). S.  Halliwell, Plato: Republic V (Liverpool, 1998), on 460 c 3–5, argues for exposure. Magnesia may lack exposure, since it is never mentioned. Laws 5, 740 b 11–741 a 4 recommends exporting excess citizens to colonies. If they are similarly governed, these citizens might not be ­disadvantaged.



Is Plato a Consequentialist?

133

of justice constitutes true statesmanship and assigns honours and political offices in direct proportion to virtue (757 a 5–d 6):64 Proportionality: A distribution of political offices to persons is just if and only if it distributes offices to persons in direct proportion to each person’s virtue. But the city must sometimes, instead, employ a principle of simple equality that gives all citizens an equal chance independently of virtue by assigning offices by lot: [T11]  ἀναγκαῖόν γε μὴν καὶ τούτοις παρωνυμίοισί ποτε προσχρήσασθαι πόλιν ἅπασαν, εἰ μέλλει στάσεων ἑαυτῇ μὴ προσκοινωνήσειν κατά τι μέρος—­ τὸ γὰρ ἐπιεικὲς καὶ σύγγνωμον τοῦ τελέου καὶ ἀκριβοῦς παρὰ δίκην τὴν ὀρθήν ἐστιν παρατεθραυμένον, ὅταν γίγνηται. (Laws 6, 757 d 6–e 2)

Nevertheless, the whole city must sometimes use even these terms [‘equal’ and ‘unequal’] less precisely if it is to avoid faction in some of its parts. For equity and forgiveness, when they occur, are always infringements of the perfection and exactness of strict justice.65

What is the relation between Proportionality and H’s goal of making the city happy? Is this distributive principle merely instrumental to the city’s happiness, so that its normative force derives entirely from the goal of maximizing happiness? It seems that the city’s happiness must at least account for most of the principle’s normative force. The Laws often assigns honours on a forward-­looking basis, and this seems especially appropriate for political offices.66 We can honour the virtuous in many ways, but the primary reason for assigning someone political office is that she will promote the city’s happiness (cf. Rep. 3, 412 d 9–e 2). Does this account for all of this principle’s force? Violating this principle deprives some citizens of what they deserve and is good for them (since virtuous political activity contributes to a citizen’s well-­being). The lot’s use is only better because some citizens undermine Proportionality and prevent, by their wrongdoing, a 64  F. Harvey, ‘Two Kinds of Equality’, Classica et Mediaevalia, 26 (1965), 101–46; R. Kamtekar, ‘Social Justice and Happiness in the Republic: Plato’s Two Principles’, History of Political Thought, 22 (2001), 189–220; J. Ober, Political Dissent in Democratic Athens (Princeton, 1998) at 279–80. 65  T. Saunders, ‘Epieikeia: Plato and the Controversial Virtue of the Greeks’, in F. Lisi (ed.), Plato’s Laws and its Historical Significance (Sankt Augustin, 2001), 65–94. 66 E.g. Laws 1, 648 b 7–c 5 and 5, 730 d 2–731 a 5. Laws 6, 759 b 4–7 is another exception to Proportionality.

Christopher Bobonich

134

better overall outcome. This is, at last, a case where Plato deprives some citizens of a significant good to which a principle of justice entitles them because third-­party wrongdoing makes another overall outcome better. The willingness to make such trade-­offs is a significant affinity with consequentialism. The last sign of consequentialism or its absence is absolutism about some actions. Such exceptionless principles are sometimes cited to show that Plato and Aristotle reject consequentialism. As an example, consider Magnesia’s parricide law: [T12]  ᾧ γὰρ μόνῳ οὐδ’ ἀμυνομένῳ θάνατον . . . παρέξει νόμος οὐδεὶς κτεῖναι τὸν πατέρα ἢ μητέρα . . . ἀλλ’ ὑπομείναντα τὰ πάντα πάσχειν πρίν τι δρᾶν τοιοῦτον νομοθετήσει. (Laws 9, 869 b 7–c 4)

In this case alone, no law will permit killing a father or mother to defend oneself against death . . . but will legislate that one must endure, suffering everything rather than do such a deed.

This law bans killing a parent even in self-­preservation. It does not ban killing an insane parent to prevent her from killing your other parent (and then committing suicide), or killing all other citizens. That blanket prohibitions of action types are prone to coun­ter­ exam­ple is a lesson the early dialogues taught and the Statesman (294 a 10–b 6) reinforced. (There are also well-­known consequentialist reasons for giving people exceptionless rules: most may be so bad at identifying warranted exceptions that outcomes are better if they follow rules.) In his charge to the law guardians (for whom preserving the legal system might seem an absolute injunction), the Athenian says that, in certain circumstances, they must allow the city’s destruction—­presumably including their parents (Laws 6, 770 d 5–e 5). I am not claiming that Plato rejects all exceptionless principles inconsistent with good maximization, but the evidence for such a principle must be strong. Section 5 considers one candidate. 4.4. Impartiality The clearest difference between H and modern consequentialisms is that H is restricted to the citizens’ happiness, while modern ­consequentialisms take account of (at least) all humans. The most important issues about impartiality for Plato concern women, slaves, and foreigners.



Is Plato a Consequentialist?

135

For women, what is essential is that in both Kallipolis and Magnesia they are citizens and the laws are intended to promote their virtue and happiness just as they do for men. The strongest passage is [T6], which suggests that women are on average capable of pretty much the same degree of virtue and happiness as men. But some have argued for greater pessimism: 1. The Republic and the Laws disparage women often enough to undermine the more optimistic passages.67 2. Republic 5 holds that although many women are better at many arts than many men, the best performer in any art is always male (455 c 5–e 4). 3. The Timaeus holds that women are reincarnated men who have lived badly (42 b 3–d 2, 90 e 1–91 a 1) and that the soul’s original and best state is male (41 d 4–42 b 2). 4. Some argue that the Laws excludes women from almost all political offices because they are inferior in virtue. I cannot enter these debates here, but I think that (1)–(4) are best interpreted so as to be consistent with (a) many women being in Kallipolis’ two upper classes, and (b) educating women in Magnesia more or less doubles the city’s happiness. If the Timaeus holds that women can only attain a vastly inferior form of virtue, it is inconsistent with the Republic and the Laws. Slaves’ and foreigners’ situations differ greatly: since they are not citizens, neither’s well-­being counts towards the laws’ goal. Unfortunately, neither the Republic nor the Laws tries to justify slavery or explain whether the slave benefits.68 The Republic provides few details about slavery.69 Republic 9, 590 c 7–d 6 comes closest 67 E.g. Rep. 3, 387 e 10–388 a 3; 5, 469 d 4–e 2; Laws 6, 781 a 11–b 3. For Plato on women, see E. Blair, Plato’s Dialectic on Women (New York, 2012); C. Harry and R. Polansky, ‘Plato on Women’s Natural Ability: Revisiting Republic V and Timaeus 41 e 3–44 d 2 and 86 b 11–92 c 3’, Apeiron, 49 (2016), 261–80; R. Kamtekar, ‘Dis­ tinction without a Difference? Plato on “Genos” vs. “Race” ’, in J. Ward and T. Lott (eds.), Philosophers on Race: Critical Essays (Oxford, 2002), 1–13; T.  Saunders, ‘Plato on Women in the Laws’, in A. Powell (ed.), The Greek World (London, 1995), 591–609; G. Vlastos, ‘Was Plato a Feminist?’, in id., Studies in Greek Philosophy, 2 vols. (Princeton, 1995), ii. 133–43. 68  Rep. 9, 590 c 1–591 a 3 concerns the producers, not chattel slaves. But Aristotle’s justification of slavery may draw on it. 69  G. Vlastos, ‘Does Slavery Exist in Plato’s Republic?’, in id., Platonic Studies, 2nd edn (Princeton, 1981), 140–6 and G. Vlastos, ‘Slavery in Plato’s Thought’, in id., Platonic Studies, 147–63 argues conclusively for slavery in Kallipolis.

136

Christopher Bobonich

and could be extended (although Plato does not do so) to suggest the justification that the slave’s own reason cannot direct his life and that he is benefited by his master’s direction. (As we shall see, it is unlikely that this rationale could justify Magnesia’s practices.) Kallipolis is composed of three classes, and since slaves are not in any, they are not part of the city. The community of pleasure and pain that makes a city one does not include slaves (Rep. 5, 462 d 8–e 2), which is lucky for the citizens, since they will not have to say, ‘Mine is enslaved’. In Magnesia, the laws aim at fostering friendship throughout the city, but slaves are excluded from this, since slaves and masters cannot be friends (Laws 6, 757 a 1–2). Kallipolis does not enslave other Greeks, and so its slaves are barbarians (Rep. 5, 469 b 5–c 6). The Laws remarks that slaves from different nations are more docile (6, 777 c 8–d 3) but does not explicitly restrict slavery to barbarians. In the Laws, the slave’s fundamental characterization is as ‘a piece of property’ (τὸ κτῆμα, 6, 776 b 5–777 c 5). Magnesian laws, however, show some recognition of the slave as a member of a shared human religious community: killing a slave incurs pollution, while killing a non-­human animal does not.70 Plato also recognizes (although not legally) that slaves can be treated with hybris and unjustly and impiously (777 d 2–e 6). Treating them correctly (and presumably justly) is more for the masters’ own sake, but also partly for theirs (777 d 2–3). It is plausible that they can be so treated because of their human status, but Plato does not say this or explain how their humanity grounds these ethical facts.71 The Athenian recommends, but does not institutionalize, sowing ‘virtue’s seed’ (ἀρετῆς ἔκϕυσιν) in slaves, which seems akin to the metics’ moderation discussed below (777 d 2–e 2). But Plato does not suggest that slaves can develop genuine virtue or a close approximation. The rest of the Laws does not take up this idea: it plays no role where we might expect it (e.g. in manumission), and Plato does not recommend that masters foster such seeds’ growth. 70 G.  Morrow, Plato’s Cretan City (Princeton, 1960), 50–1, 120–1; Saunders, Penal Code, 65–6. 71  According to Book 10’s theology (Section 6.2), every human is part of the All and has the telos of contributing to the All’s virtue and happiness. Neither claim nor their conjunction entails that slaves can be treated unjustly. The theological claims are arguably true of animals, while the point about injustice is not. The claim that ordinary slaves are capable of sufficient ethical development so as not to warrant demotion in the next incarnation is more promising.



Is Plato a Consequentialist?

137

The legal system protects the slave only insofar as particular citizens extend her good will: 1. The slave has no legal recourse against her master. If a master kills his slave, the law orders purification but lacks enforcement mechanisms. A slave who kills a free man, even in self-­defence, is executed.72 2. The Athenian law against hybris emphasizes the victims’ weakness and included slaves; the corresponding Magnesian law does not.73 3. For some offences committed by slaves, they are handed over to the victim or his relatives to punish however they wish.74 With respect to foreigners, let us begin with metics.75 Magnesian citizens have a sufficient art occupying their time, that is, preserving the city (Laws 8, 846 d 3–7). This excludes them from practising skilled crafts (846 d 2–3). But pursuing trade or commerce fosters money-­loving which tends to grow insatiable (11, 918 d 4–9). Citizens thus cannot practise crafts or manage slaves who do (1, 644 a 2–5; 5, 741 e 1–6, 743 d 2–6; cf. Rep. 6, 495 c 7–496 a 10; 9, 590 c 1–5).76 Crafts must be assigned to metics or foreigners, since, the Athenian frankly remarks, their corruption would not be ‘a great loss’ to the city (μεγάλη λύμη, Laws 11, 919 c 4–6; cf. 920 a 3–4). To obtain their limited-­term residence, metics must be moderate (Laws 8, 850 a 2–8). This is ‘ordinary’ (δημώδης) moderation that occurs naturally in children and animals, is not accompanied by wisdom (ϕρόνησις), involves continence, and is worthless apart from the other virtues (4, 710 a 5–b 2). But there is no serious effort to further improve metics during their stay. Plato is simply concerned to prevent their corrupting of Magnesians and does not try to justify their treatment by arguing that, for example, they are better off in Magnesia than at home. With respect to other foreigners, what is most relevant is the cities’ war policies. The Republic bans enslaving Greek captives and places restrictions on warfare against Greeks (5, 469 b 5–471 c 2). 72 G. Morrow, Plato’s Law of Slavery in its Relation to Greek Law [Law] (New York, 1976), 120–33; Saunders, Penal Code, 220, 228–30, 239–41. Slaves in Magnesia, unlike Athens, had no right of asylum; see Morrow, Law, 55. 73 Morrow, Law, 51; Saunders, Penal Code, 241, 269–71. 74 Saunders, Penal Code, 228, 265, 346. 75 Saunders, Penal Code, 334–48. 76  Any manual craft and not just trade corrupts (Laws 5, 741 e 1–6 and 743 d 2–6).

138

Christopher Bobonich

War against Greeks is really faction, because the Greek race is akin and kindred to itself; other Greeks are friends, and conflict with them is ‘correction’ (σωϕρονιοῦσιν, 470 b 4–c 3, 471 a 6–7). But we should not overestimate Socrates’ concern for other Greeks. He advocates inflicting pain upon the innocent citizens of enemy cities so that they seek peace (471 b 2–5). He counsels refraining from inflicting certain bads but gives no content to the notion of correction. Perhaps defeat by Kallipolis would sometimes moderate other cities, but this is highly contingent, and Kallipolis does not impose political changes on other cities. Nor does Kallipolis seek alliances out of friendship. The only alliance mentioned is Socrates’ ruthless proposal that when warring against two cities, Kallipolis agree with one to attack the other and give it the victory spoils (4, 422 d 2–7). Barbarians, however, are alien and foreign and nat­ ural enemies (5, 470 c 5–7). Conflict with them is war and expresses hatred (470 c 5–d 2). The harsh measures rejected for Greeks are accepted when fighting barbarians (470 a 5–471 c 2). One can imagine ways to justify such differences ethically, but Socrates does not try. The Laws differs significantly. In Book 3, the Athenian recounts two idealized histories to show that good constitutions mix mon­ archy and democracy (693 d 2–694 a 2). Cyrus’ Persia is one ex­ample, and it is presented as ethically equal to Athens of the Persian wars (another well-­mixed constitution). And despite recounting Athens’ history during these wars, there is no mention of natural enmity or endorsements of different warfighting standards.77 For the differences considered, the decisive question for Plato is whether the groups’ members are citizens; if so, correct laws should aim at their well-­being. The Republic does not ask who should be a citizen. But especially if we are sceptical of the prod­ucers’ capacity for virtue and happiness, the question of a criterion for citizenship becomes pressing. The Statesman makes the cap­acity for genuine virtue the criterion of citizenship in a good city, and the Laws agrees: it excludes those performing the producers’ functions from citizenship and requires that the laws aim at instilling all the virtues in all the citizens.78 77 The Laws often pairs Greeks and barbarians as equally having some specific good (10, 886 a 44–5) or bad (9, 870 a 2–b 2) trait. 78 Bobonich, Utopia, 409–19.



Is Plato a Consequentialist?

139

The Laws does not try to show that all non-­citizens, including all foreigners, lack the capacity for virtue. (Plato could not have believed this; see, Laws 12, 951 b 4–c 4.) It shows some recognition of the ethical implications of the slave’s human status and is more inclined to consider ethical factors relevant in dealing with foreigners. Human status and the capacity for virtue are linked, but Plato does not develop a theory of their relation or a theory of how to treat slaves and non-­citizens. But while discussing religious festivals, the Athenian observes that they should honour Pluto because he is the best for the race (γένος) of men, since union is not better than dissolution for the soul and body (8, 828 d 1–6). To arrange festivals appropriately, the citizens must think: [T13]  δεῖ δὲ αὐτήν, καθάπερ ἕνα ἄνθρωπον, ζῆν εὖ· τοῖς δὲ εὐδαιμόνως ζῶσιν ὑπάρχειν ἀνάγκη πρῶτον τὸ μήθ’ ἑαυτοὺς ἀδικεῖν μήτε ὑϕ’ ἑτέρων αὐτοὺς ἀδικεῖσθαι. (Laws 8, 828 d 10–829 a 2)

that the city, just like an individual, must live well. For those who are to live happily, it is necessary first not to do injustice to each other or to suffer injustice from others.79

It is significant, as Section 6 shows, that the Athenian does not emphasize the gods’ special favour for Magnesia, but their concern for ‘the race of men’, and highlights the soul’s afterlife. Magnesia is prohibited from doing injustice to others, and the context of suffering injustice from others shows that Plato is thinking of foreign relations.80 Certainly, the Laws in these passages does not adopt an impartial concern for all humans, but it is significant that it does not endorse the Republic’s harsh views of barbarians and requires that justice govern Magnesia’s relations with other cities. Finally, some scholars have seen the Republic as holding that individual just action must (at least for philosophers) always max­ imal­ly promote virtue in the world. John Cooper holds that: knowing the [Form of the] good, what [the philosopher] wants is to advance the reign of rational order in the world as a whole, so far as . . .  he can . . . He recognizes a single criterion of choice: What, given the

79  I accept K. Schöpsdau’s construal of the Greek ad loc. in Nomoi (Gesetze): Platon: Werke, vol. ix (Göttingen, 1994–2011). 80  Kallipolis’ rulers know how the city will best live (Rep. 4, 428 d 3) with itself and other cities, but there is no mention of acting justly towards other cities.

140

Christopher Bobonich

c­ ircumstances, will be most likely to maximize the total amount of rational order in the world as a whole?81

This view is a form of act-­consequentialism: the criterion for ranking is that the best outcome maximizes the world’s total amount of rational order. Both justice and rationality require this action. Indeed, it is a strong form of consequentialism, since, for philo­sophers, this principle is both a criterion of just action and a decision procedure. Sarah Broadie has a related view. She thinks that the upshot of an individual’s reason functioning properly is a just action. When reason functions properly, its possessor: must seek to produce the psychic unity which is justice equally in . . . [a]s many beings (men and communities) capable of justice as he can affect directly or indirectly so as to promote justice within them. Naturally his own soul would be his most readily available ground of operation, but if at any time his range does not include others, this can only be for contingent reasons.82

The normativity involved here is ethical and rational: this is what must be sought insofar as one is just and rational. This too looks like a form of act-­consequentialism: there is an agent-­neutral outcome ranking, and the agent is required to bring about the outcome containing the most justice. I do not think, however, that either view is correct. Philosophers will obey commands to rule, since both they and these commands are just (Rep. 7, 520 e 1). But it is a necessary condition of these commands being just that Kallipolis educated these philosophers (520 b 5–c 1). Where philosophers arise spontaneously, they would justly decline ruling, since they have received no philosophical training to pay back (ὀϕεῖλον, 520 b 1–5). Yet if they were to rule, they would bring about a happier city than if they do not. It thus seems that the principle of individual just action does not require this sort of impartial maximization. One might object that a city not educating philosophers would reject their rule, so they could not bring about better outcomes. But even if this were true, it is irrelevant, since Plato does not mention this consideration.

81  J. Cooper, ‘The Psychology of Justice in Plato’, American Philosophical Quarterly, 14 (1977), 151–7 at 155–6. 82  S. Waterlow [Broadie], ‘The Good of Others in Plato’s Republic’, Proceedings of the Aristotelian Society, n.s. 73 (1972–3), 19–36 at 31–2; cf. 27–8.



Is Plato a Consequentialist?

141

Socrates rejects truth-­telling and giving back what is owed as a definition of justice (Rep. 1, 331 c 1–d 5), but because paying back debts is sometimes unjust, it does not follow that it is never just (331 c 3–5). Since Plato accepts that whatever is F is F in virtue of the very same thing (10, 596 a 5–8), paying back debts never by itself makes anything just. But this does not show that paying back debts is irrelevant to an action’s justice. ‘Bright colour’ is an un­accept­ able definition of the fine (5, 476 b 5), but some of Kandinsky’s paintings would not be fine if their colours were dull, so it seems that their fineness is partially constituted by (or at least supervenes on) their colours. Plato thus might be able to justify laws requiring citizens to have differential concern for their fellow citizens by claiming that the benefits they have received from them make preferentially benefiting them just. This is not to suggest that this justifies every instance of preferring fellow citizens’ well-­being (for example, slavery), but it might justify some. Broadie’s and Cooper’s interpretations may not be correct, but they point to a deep issue. Although its exact nature is not clear, individual just action has ethical and rational normativity. Plato is likely to have had considerable sympathy for the view that the world is ethically harmonious in the sense that if a group of people are as virtuous as possible, then each person in that group can achieve her maximal well-­being. If being virtuous is interentailing with being rational, then the world is also rationally harmonious. Even if we reject exceptionless ethical and rational harmony for individuals, the possibility of conflict between cities, each acting in accordance with H, seems much greater. For individuals, they can share great non-­competitive goods, such as knowledge and virtue, through cooperative activity. It is much less clear, as the Republic’s war policies suggest, that if each city acts in accordance with H, their maximal well-­being is compossible. One might argue that insofar as each city optimally achieves H, its citizens will become more virtuous and so will not come into conflict. But even if this is so, people’s ethical condition in many cities is far from optimal. Even if the maximal well-­being of fully virtuous cities were compossible, this does not entail that the same holds for imperfectly virtuous cities. It is hard to see how Kallipolis’ war policies are compatible with the maximal well-­being of barbarian cities (without, for example, implausibly assuming that all barbarians are incurably unjust and thus not harmed by death and Dependent Bads.) But

142

Christopher Bobonich

imperfectly virtuous Greek cities might come into conflict with Kallipolis, and although Kallipolis’ practices for fighting Greeks are far less inhumane than those for barbarians, they can still inflict great damage on Greek cities’ innocent populations. The possibility that two cities, each acting perfectly rationally and correctly according to H, could come into serious conflict is a prospect that I think Plato would find dismaying.83 Although theodicies are notoriously capable of accommodating (more or less plausibly) the inevitability of great bads, there is also at least a prima facie tension between accepting providential gods and allowing that cities acting in accordance with ethical and rational principles can ex­peri­ence such conflict. Section 6 returns to this issue.

5.  The harm principle We have seen that Plato’s basic principle for justifying laws, H, is consequentialist in significant respects. But the elenchus with Polemarchus in Republic Book 1 ends with the conclusion that a just person never harms (βλάπτειν) anyone (335 e 5–6). I shall call this ‘the Harm principle’. This argument has attracted surprisingly little attention despite its striking and important conclusion.84 It should, however, be taken more seriously than it usually is, since a careful analysis shows that it is grounded in one of Plato’s basic metaphysical commitments about causation, the Principle of Causal Synonymy: the cause of F must itself be F.85 Seeing this dependence lets us better understand why ‘harm’ in this principle is understood in the very restricted sense of ‘make unjust’ and does not apply to any other sort of reduction in well-­being. The basic idea is that ‘just’ (δίκαιος) is equivalent to ‘good’ (ἀγαθός).86 This equivalence combined with the weaker but related principle that the cause of an 83  Such conflicts typically require certain contingent circumstances to obtain, for example, scarcity, and it is unlikely that H generates conflicts in all possible circumstances. But it is still disturbing if compossibility is likely to fail. 84  I provide a reconstruction with commentary on this argument in ‘A Reconstruction and Analysis of the Republic’s Harm Principle (Rep. 1, 335 a 6–336 a 10)’, which is on my Academia webpage: https://stanford.academia.edu/ChrisBobonich. 85  See G. Fine, ‘Forms as Causes: Plato and Aristotle’, in id., Plato on Knowledge and Forms (New York, 2003), 350–96; D. Sedley, ‘Platonic Causes’, Phronesis, 43 (1998), 114–32. 86  Plato has no adjective corresponding to ἀρετή as δίκαιος corresponds to δικαιοσύνη. ἀγαθός plays this role, e.g. Rep. 1, 347 b 7; Laws 10, 899 b 5.



Is Plato a Consequentialist?

143

F effect cannot itself be opposite-F (Phaedo 101 a 9–b 2) gives us this restriction. This conclusion is important to us for several reasons. First, this argument establishes an exceptionless prohibition on harming, and ‘harm’ is not defined as failing to produce the best outcome, but rather means ‘make unjust’. The just person is not required to minimize overall harms but is prohibited herself from ever harming another. This is an agent-­centred constraint. Nor is there any other apparent necessary connection between not harming and bringing about the best outcome. They may usually go together, but it seems that there could be cases in which maximizing the city’s happiness required laws that might make some people unjust.87 This exceptionless prohibition of harming is thus in prin­ciple inconsistent with consequentialism. Second, although Plato does not explicitly draw this conclusion, the argument can easily be adapted to establish a complementary requirement that the just person benefit others in the sense of making them more just. (There are some issues about its exact formulation, but as a first approximation, take the ‘Benefit principle’ to be: If X’s justice causally affects Y, then X’s justice makes Y more just.) Some scholars have argued that Book 4’s account of justice (Rep. 443 c 9–444 a 2) is empty without a specification of the good. The Harm constraint and the Benefit principle, if Plato accepts it, give significant content to justice even without an account of the good. While most would agree that Plato’s ethical views have some sort of metaphysical grounding, it is surprising that such a substantive ethical claim depends so directly on Plato’s metaphysics of causation. Third, insofar as it rests on universal metaphysical principles, the Harm principle has a strong impartialist tendency. Heat, after all, cannot chill anyone, no matter whether they are men or women, slaves or free, Greeks or barbarians. The possibility that the sorts of actions that might make people unjust may vary somewhat from group to group does not abolish this constraint. Plato does not tell us how to resolve the potential conflict between the Harm principle and H (or between the Harm principle and the Benefit principle). 87  They might, for example, require putting some auxiliaries into the reliably corrupting circumstances of going abroad and engaging in financial dealings with foreigners to forestall invasion; cf. Rep. 3, 416 c 5–417 b 9. Plotinus rejects making some worse even to improve the kosmos (Enn. 3. 2. 12, 7–13).

144

Christopher Bobonich

In Section 6, I shall consider the role of these principles in the Laws’ theology. 6. The Laws’ theology 6.1. The petteia-player I must be brief, but the Laws’ theology is essential to our story and raises new questions. Book 10’s prelude to the impiety law describes the petteia-player’s ordering of the All:88 [T14]  τῷ τοῦ παντὸς ἐπιμελουμένῳ πρὸς τὴν σωτηρίαν καὶ ἀρετὴν τοῦ ὅλου πάντ’ ἐστὶ συντεταγμένα, ὧν καὶ τὸ μέρος εἰς δύναμιν ἕκαστον τὸ προσῆκον πάσχει καὶ ποιεῖ. τούτοις δ’ εἰσὶν ἄρχοντες προστεταγμένοι ἑκάστοις ἐπὶ τὸ σμι­κρό­ τατον ἀεὶ πάθης καὶ πράξεως, εἰς μερισμὸν τὸν ἔσχατον τέλος ἀπειργασμένοι· [I] ὧν ἓν καὶ τὸ σόν, ὦ σχέτλιε, μόριον εἰς τὸ πᾶν συντείνει βλέπον ἀεί, καίπερ πάνσμικρον ὄν, σὲ δὲ λέληθεν περὶ τοῦτο αὐτὸ ὡς γένεσις ἕνεκα ἐκείνου γίγνεται πᾶσα, ὅπως ᾖ τῷ τοῦ παντὸς βίῳ ὑπάρχουσα εὐδαίμων οὐσία, οὐχ ἕνεκα σοῦ γιγνομένη, σὺ δ’ ἕνεκα ἐκείνου. [II] πᾶς γὰρ ἰατρὸς καὶ πᾶς ἔντεχνος δημιουργὸς παντὸς μὲν ἕνεκα πάντα ἐργάζεται, πρὸς τὸ κοινῇ συντεῖνον βέλτιστον μέρος μὴν ἕνεκα ὅλου καὶ οὐχ ὅλον μέρους ἕνεκα ἀπεργάζεται· [III] σὺ δὲ ἀγανακτεῖς, ἀγνοῶν ὅπῃ τὸ περὶ σὲ ἄριστον τῷ παντὶ συμβαίνει καὶ σοὶ κατὰ δύναμιν τὴν τῆς κοινῆς γενέσεως. (Laws 10, 903 b 4–d 3)89

He who cares for the All has put all things in order for the whole’s salvation and virtue, and each part of it suffers and does what is befitting, as far it can. For each of these, rulers have been set up for every occasion over their suffering and activity down to the smallest aspect, and they have achieved their end to the last detail. [I] And one of these is your part, stubborn one, and it always strives and looks towards the All, although it is entirely small. But this very thing has escaped your notice, that all gen­er­ation comes into being for the sake of this: that a happy existence might belong to the All’s life—­it does not come into being for your sake, but you for its sake. [II] For every doctor and every skilled craftsman does everything for the sake of a whole, he creates a part that strives for what is best in common, for the whole’s sake, and not the whole for the part’s sake. [III] But you complain, ignorant of how what concerns you turns out to be best for the All and for you, by the power of common generation.

88  I identify Book 10’s transcendent god, the ‘petteia-player’ (903 d 5–6), with the Timaeus’ Demiurge and the All with its Living Creature. 89  Hereafter in this section Stephanus numbers refer to Laws 10.



Is Plato a Consequentialist?

145

In ordering the All, the petteia-player aims at the All’s salvation and virtue (903 b 5, 904 b 4–5) or its having a happy life (903 c 4–5). This is what is best for the All (903 c 7, 903 d 1), that is, the All’s maximally good overall condition (including both states and activities). The petteia-player acts for this (903 b 5) and for the sake of this in all his actions (903 c 3–d 4), and it is the auxiliary gods’ end (903 b 8). So, in each of the petteia-player’s actions concerning the All, he (1) acts for the sake of the All’s best condition, (2) pursues it for its own sake, not for the sake of anything else, and (3) whatever he pursues, he pursues for its sake.90 Platonic teleology is intentional, so this telos applies to each part and its actions. God’s telos in ordering the All and its parts might differ from the individual’s own telos. Nevertheless, god’s telos must have some positive normative status, and I return below to what motivates people to pursue it. This passage commits Plato to: A1 For each human, the petteia-player has established for all her actions the telos of the All’s best condition, and A2 The Coincidence: For each human, the same outcome is overall best for her and for the All.91 [I] and [II] are evidence for A1. [I] describes the telos as the All’s happiness and [II] and [III] understand this in an optimizing sense that is naturally construed as the All’s greatest happiness. [III] is evidence for A2. Moreover, all Magnesians are educated to accept A1 and A2. A1 is clearly a form of consequentialism. The All’s best condition provides a fully agent-­neutral end: it is the same not only for every Magnesian, but for every human (and for every part of the All, including the stars). Even before specifying its exact normativity for humans, we should not underestimate this development’s significance. Plato here for the first time explicitly establishes a single goal for humans as such and eliminates the possibility that 90  ‘For the sake of’ includes both instrumental and constitutive means (Tim. 69 a 1). Book 10 does not say that the petteia-player acts for the sake of making the All finest and non-­relationally best (Tim. 30 a 2–b 6), but this is consistent with A1 and A2 below. 91  Here is a first approximation of A2 restricted to people: for all people (p) and all times (t), (1) there is a unique outcome O1 after t such that O1 is the best outcome for p, and (2) there is a unique outcome G1 after t such that G1 is the All’s best outcome, and (3) O1 is part of G1. I assume uniqueness to simplify exposition.

146

Christopher Bobonich

the ethical and rational standards applying to different individuals and groups might conflict. This single goal is the All’s best condition, including that of all humans. To see the novelty of this, consider Richard Kraut’s rejection of similarities between Plato and Aristotle and modern teleologists: neither Plato nor Aristotle are cosmopolitans . . . it never enters their heads that everyone should maximize the good, whether the intended beneficiaries . . . are fellow citizens or not . . . they always assume that a good person benefits fellow citizens.92

The next section addresses three questions. 1. What explains the Coincidence? 2. What are the relations between A1, the Coincidence, and rational eudaemonism? 3. What difference does accepting A1 and the Coincidence make to individuals? It is illuminating to see the similarities between Book 10 and modern consequentialisms (for example, a conception of practical rationality as maximizing the good and acceptance of a strong form of impartiality). But Plato’s view must be understood in light of his other commitments, including two not standardly shared by modern consequentialists, that is, to a providential theology and personal immortality. So, we must consider some theological preliminaries. If A1 is a form of consequentialism, what about the Harm prin­ ciple? 904 b 1–3 suggests something resembling the Benefit prin­ ciple, but neither the Timaeus nor the Laws states or implies the Harm principle for either humans or gods. For both dialogues, we must ask whether a god ever harms either by causing an overall character worsening or, more generally, an overall decrease in well-­ being. The Timaeus asserts that it is not lawful for god, because he is the best, to do anything except the finest (30 a 6–7), but this does not rule out that harming might sometimes be finest. Other relevant passages also fall well short of asserting the Harm principle (Tim. 29 a 5–6, 29 e 1–3, 30 b 3–6, 53 b 5–6, 68 e 1–6). Laws 10, 903 b 7–c 1 makes the question of whether gods ever harm especially salient: if the gods’ supervision has fully succeeded, 92  R. Kraut, ‘Review of G. Santas, Goodness and Justice’, Ancient Philosophy, 25 (2005), 446–63 at 450. Kraut’s position has changed, but this view remains widespread.



Is Plato a Consequentialist?

147

is it not likely that they committed (or allowed) some harms (in both senses), given the All’s present condition? Neither the Timaeus nor the Laws confronts this question, but their silence about the Harm principle may be intentional. Plato might have thought (or hoped) that the gods observed it, but Book 10 makes A1 fundamental. Some issues that Plato does not discuss have a long afterlife. Perhaps the most fundamental is whether the All is unimprovably good. If the petteia-player, unlike the Demiurge on traditional interpretations, is unconstrained by his materials, unimprovability is more likely but is not entailed.93 The most pressing question is whether the All would be worse off if anyone were ever better than she actually is or performed fewer vicious actions or less vicious actions.94 Andrea Nightingale, appealing to free will’s value, answers yes: a universe including free will contains bad parts but is overall better.95 But unimprovability requires not only that some badness is better for the All, but that no diminution of badness could be better. Whether freedom and responsibility (being aitios, 904 b 6–d 4) require ‘liberty of indifference’ or determination by the good in a way that does not require (and may be inconsistent with) such liberty, why could not more (or all) people be free and responsible, but better?96 Book 10 does not clearly exclude this ­possibility.97 Other arguments for unimprovability appeal to (i) our inability to know whether any local diminution has worse cosmic consequences, and (ii) the idea that the All is an organic unity, so that an All with some bad parts is better in itself than one with only good

93  R. Mohr, ‘The Platonic Theodicy: Laws 899–905’ [‘Theodicy’], in id., Gods and Forms in Plato (Las Vegas, 2005), 197–201, accepts such unconstraint. 94  Yes: Mohr, ‘Theodicy’. No: S. Broadie, ‘Theological Sidelights from Plato’s Timaeus’ [‘Theological’], Proceedings of the Aristotelian Society, suppl. vol. 82 (2008), 1–17 at 13–16; G. Carone, Plato’s Cosmology and its Ethical Dimensions (Cambridge, 2005), 184. 95  A. Nightingale, ‘Historiography and Cosmology in Plato’s Laws’ [‘Historiography’], Ancient Philosophy, 19 (1999), 299–326 at 318–19. 96 D. Speak, The Problem of Evil (Cambridge, 2015), 19–46. 97 903 d 1–3 asserts the compossibility of what is best for an individual and the All, not its actuality. 905 b 1–7 does not say that diminishing badness makes the All worse (as it should, if this kosmos, with its current amount of badness, is the best possible kosmos) but that the petteia-player ameliorates the badness the unjust contribute. Theaet. 176 a 5–177 a 8 at most asserts the necessity of the All containing some badness, not the necessity of its actual amount.

148

Christopher Bobonich

parts.98 (i) undermines arguments against unimprovability but is not evidence for it. The Athenian never refers to such arguments and does not allude to the causal complexity needed to support (i) or the aesthetic examples used to support (ii).99 Further, there is no plausible explanation for the organic unity that (ii) posits. The petteiaplayer subjects humans to a transmigration system that I discuss shortly. But no injustice is far better than punished injustice (Laws 9, 853 b 1–c 3), and even if this system benefits the All because it instantiates justice, it would still do so if more people were just; it would simply apply to better people. Given the ubi­quity of Plato’s categorical insistence on being just, he would have acutely felt (ii)’s unintuitiveness. Since the prelude states A1 and encourages virtue, it is exceedingly unlikely that the Athenian would suggest that every actual injustice is better for the All without scrupulously explaining why citizens should not infer that injustice is acceptable.100 David Sedley argues that the Demiurge makes the All in accordance with craft rationality.101 Craft rationality is a form of practical rationality that maximizes the good: an art aims at bringing about what is best for what it is set over (Rep. 1, 345 d 1–5), and so the universe is ‘best’ (Tim. 30 b 6). This feature of craft rationality takes a consequentialist form in A1: the agent-­neutral telos for all humans is the All’s best condition. But to identify the product’s best condition, we must supplement craft rationality with deliberative practical rationality. According to it, what is best for a human is happiness, and Laws 10, 903 c 1–5 applies this to the All. Here again, a conception of practical rationality as maximizing the good is fundamental to Plato. A further implication of Plato’s conception of practical rationality may be relevant. A classic argument for utilitarianism is that we 98  I mention only the theodicies in the Laws’ literature. On (i), see M. Tooley, The Problem of Evil (Cambridge, 2019); for (ii), see Mohr, ‘Theodicy’. 99  (ii) requires a strong version of organic unity: for every bad part of the All, the All would be overall worse off if it were replaced by a similar part that is better in some respect. 100  Polit. 308 c 1–7 claims that every constructive science uses only good parts; cf. 309 e 10–14. The Timaeus’ claim that non-­human animals originated from unvirtuous humans does not show that the All’s completeness requires injustice. It is pos­ sible that many have culpably become unjust and that the Demiurge, foreseeing this (76 d 6–e 3), brought about some good from it, although their being virtuous would have been better. In that case, animal kinds could have been instantiated in other ways (for example, starting from a third-­rate psychic mixture). 101 Sedley, Creationism, 107–13.



Is Plato a Consequentialist?

149

should transfer the principle of individual rationality to the group. For individuals, rationality requires transferring goods and bads across times to maximize the good. For the group, rationality requires transferring goods and bads across persons to maximize the good. Plato sympathizes with this line of thought, since a city attains its greatest unity when each citizen identifies her own well-­ being with the aggregation of all citizens’ goods and bads (Section 4.1). Although the All is a single animal, it is also a composite of animals. It is thus plausible for its well-­being to be an aggregation of the well-­beings of its parts and of the World Soul (perhaps differently weighted). John Rawls famously objects that this sort of argument ignores the ‘separateness of persons’.102 Transferring goods and bads within a single person’s life to maximize overall good is rational, but transfers across different people raise fairness issues, since the winners and losers differ. This objection has less force against Plato, since virtue is by far the most significant good, and it is within each person’s control. So, the goods and bads that could be transferred across different people are much less significant goods and bads for Plato than for Rawls (although we would still have to worry about, for example, distributing educational opportunities). The petteia-player constructs a system in which better souls automatically move to better places and company, while worse souls do the opposite in successive incarnations (903 d 3–905 c 4).103 Each soul remains responsible for its character (904 b 8–c 3; cf. Rep. 10, 617 e 1–5). Previous eschatological myths’ external rewards and punishments are mentioned only fleetingly (904 d 1–4), and what is stressed instead is that better and worse souls experience what like souls do to and suffer from each other (904 a 6–905 b 7). I shall call this the ‘Transmigration principle’ or ‘Transmigration’. Transmigration clearly resembles Proportionality. Both distribute items on the basis of virtue: the former distributes environments to souls, the latter offices to citizens. Section 4.3 examined Proportionality’s relation to H’s goal of maximizing well-­being. The petteia-player aims at the All’s virtue and happiness, that is, its good condition, and establishes Transmigration for this purpose (903 e 3–4; cf. 903 b 7–c 1). 102 Rawls, Theory of Justice, 21–4. 103 Saunders, Penal Code, 202–7.

150

Christopher Bobonich

Transmigration’s clearest normative force derives from the All’s happiness, and this principle cannot come apart from the All’s happiness as Proportionality could come apart from H’s goal. The petteia-player ensures that opposing Transmigration, unlike opposing Proportionality, is fruitless. But how does Transmigration ­promote well-­being? For the good, their well-­being may increase, inter alia, by (a) avoiding others’ wrongdoing, (b) cooperating with other good people in virtuous activities, and (c) gaining in virtue and knowledge (5, 728 b 2–c 2; 9, 854 b 6–c 3).104 What about bad people? Dependent Goods are worse for the bad than the cor­res­ pond­ing Dependent Bads, so the bads they suffer from each other actually benefit them. But association with the bad may worsen their characters. If this is intended as a punishment, it is a spectacular violation of the Harm principle and hard to reconcile with Magnesia’s penology or maximizing the All’s happiness.105 Perhaps the best we can do is emphasize that the unjust are responsible for their worsening (904 b 8–c 3) and that some may be chastened by experiencing injustice and make small ethical improvement (Rep. 2, 358 e 4–359 b 2; 10, 619 d 1–7). The Athenian is not optimistic about ethical progress after the descent begins (905 a 1–c 7, with 5, 728 b 2–c 6) but does not entirely exclude it. 6.2.  Why does the Coincidence obtain? The Coincidence entails that there are two perspectives for evalu­ at­ing outcomes: that of the individual’s good and that of the All’s good. The Athenian’s references to the Coincidence (903 c 6–d 3, 905 c 1–5) recognize the individual’s existence as a distinct part and the individual’s viewpoint, which is concerned with what is best for herself. The Athenian never hints that this viewpoint is mistaken or irrational because it is partial and never suggests that a 104  Laws 12, 959 b 5–c 3 does not exclude post-­mortem improvement, but only that the living can greatly aid the dead. Laws 10, 903 d 3–e 1, 904 c 1–905 b 1, and Tim. 92 c 1–3 make ethical improvement or degeneration integral to post-­mortem existence. 105 The Laws’ justification for punishment is curing the wrongdoer (9, 862 c 6–863 a 2; 11, 933 e 7–934 b 3). Executing incurables does not harm them, since their well-­being is negative and their punishment may deter others. Retribution is rejected (11, 934 a 5–b 3), but some retributive myths (9, 872 c 7–873 a 7) are used to frighten the otherwise unpersuadable, since obeying the law through fear is better than crime.



Is Plato a Consequentialist?

151

rational person should transcend it and promote instead the All’s best condition.106 The Athenian seems to move directly from accepting the part’s individual existence to accepting the rationality of its concern with its own happiness. This is reminiscent of Sidgwick’s famous argument for rational egoism which is one-­half of the ‘Dualism of Practical Reason’. Sidgwick moves directly from my independent existence to the rationality of my special concern for my own good.107 For Sidgwick, such concern excludes any interest in others’ well-­being for their own sake, while Plato’s rational eudaemonism does not. Despite this difference, the similarity between Plato and Sidgwick is striking. This suggests that there is a deep connection between the fact of individuality and rational eudaemonism. But the complete lack of an argument by either shows the difficulty of working out that intuition.108 The Coincidence obtains because of the power of ‘common ­gen­er­ation’. Unexplained here, this refers to the Timaeus’ idea that human souls originate from a less pure form of the same ‘stuff’ as the World Soul (41 d 4–e 1), although they do not form a ‘group’ soul with it or perish by absorption into it.109 Broadie thus claims that humans and their souls are not parts of the All. But Book 10 insists that they are, although it also stresses that our souls retain their individual identities throughout all their transformations (903 d 3–905 c 4).110 Given their common origin, human souls and 106  Cf. Section 4.4. If Plato were recommending abandoning rational eudaemonism, then, after stating A1, he would not keep inviting listeners to evaluate cosmic arrangements from the perspective of their own happiness, 903 d 1–3 and 905 c 1–4. Note especially 903 c 6–d 1, where the craftsman creates a part that strives for what is best in common. ‘In common’ (κοινῇ) stresses that an outcome is shared by two parties (10, 897 e 6, 900 d 5; 11, 916 b 6); it is not referring de dicto to the All’s best condition but to what is best for both the part and the whole. This is immediately picked up by the reference to ‘common generation’ (κοινῆς γενέσεως) at 903 d 2–3. In ordering the All, god brings about for it everything that reason recommends. That he aims at its happiness strongly supports rational eudaemonism. 107 Sidgwick, Methods, 497–9. 108 D. Phillips, Sidgwickian Ethics (Oxford, 2011), 114–51. Besides mere separate existence, Plato can appeal to the facts that each individual is rational and that her soul is capable of being in a good condition which she can recognize as good for her. 109  Phileb. 30 a 5–8 suggests that human souls originated from the World Soul, but Plato’s conclusion there does not require precision. 110  The person: 903 c 1–5, 903 d 1–3, 904 e 4–905 b 1; the soul: 903 d 3–e 1, 904 b 3–8. Tim. 30 c 5–31 a 8 suggests that humans and other animals are the kosmos’s parts, which Plotinus rejects because he thinks it makes the All the author of our

152

Christopher Bobonich

the All’s soul have similar natures and thus similar goods. The All’s soul is purely rational, while our souls’ rational ordering is their original and best condition (Tim. 41 d 4–8 and 90 d 1–7). The most important goods for both humans and the All, virtue and know­ ledge, are non-­competitive and, for humans, best pursued cooperatively (the All is self-­sufficient, Tim. 34 b 4–9). These claims do not entail the Coincidence but do facilitate it. Understanding the Coincidence requires understanding the All’s happiness. Plato gives no explicit characterization, but considering Transmigration’s motivation helps: [T15]  καὶ τὸ μὲν ὠϕελεῖν ἀεὶ πεϕυκός, ὅσον ἀγαθὸν ψυχῆς, διενοήθη, τὸ δὲ κακὸν βλάπτειν· ταῦτα πάντα συνιδών, ἐμηχανήσατο ποῦ κείμενον ἕκαστον τῶν μερῶν νικῶσαν ἀρετήν, ἡττωμένην δὲ κακίαν, ἐν τῷ παντὶ παρέχοι μάλιστ’ ἂν καὶ ῥᾷστα καὶ ἄριστα. (Laws 10, 904 b 1–6) [The petteia-player] understood that whatever is good in soul is always naturally beneficial and what is bad harmful; seeing all this, he somehow devised a position for each of the parts so that, in the easiest and best way, virtue would be victorious and vice defeated in the All.

This goal should be the same as the whole’s salvation and virtue (903 b 4–7) and must be part of the petteia-player’s overall goal, which is the All’s greatest happiness (903 c 4–5). Good souls characteristically benefit others by helping them progress in virtue (9, 854 b 8–c 3). Transmigration thus advances the All’s virtue by helping its parts become more virtuous. Since human virtue is not plausibly merely instrumental to the distinct state of the All’s virtue, the parts’ virtue should partially constitute the All’s virtue. (This claim is independent of whether A1 is consequentialist, and the All’s virtue need not be a simple summation of its parts’ virtue.) Similarly, the parts’ happiness should partially constitute the All’s

actions (Enn. 3. 1. 4, 1–24). The All is an unusual animal, since it contains within its body and soul the bodies and souls of all other animals and non-­transcendent gods (Tim. 36 d 9–e 3); see D.  Baltzly, ‘Is Plato’s Timaeus Panentheistic?’, Sophia, 49 (2010), 193–215. We might think that it is not really one thing, since it violates Aristotle’s principle that a substance cannot have other substances as parts (Metaph. Ζ. 13, 1039a3–6). It possesses significant unity, however, because it has a single bios or life (903 c 4–5). Plotinus accepts that non-­bodily substances can have substances as parts (Enn. 6. 5. 10, 50–2).



Is Plato a Consequentialist?

153

happiness.111 These constitution relations entail that improvements in human virtue and happiness directly contribute to the All’s virtue and happiness. This, of course, does not entail the Coincidence either but does again facilitate it. But the All is a god, and is it not intolerable that a god’s virtue and well-­being are partially constituted by ours (cf. Euthph. 13 b 4–c 10)? 1. Since each part’s telos is the All’s good condition, our characters and actions must affect it. The All’s oddly composite nature renders the Euthyphro intuition inapt and makes it reasonable that its parts’ good condition partially constitutes the All’s good condition. Their good condition, however, does not partially constitute the transcendent petteia-player’s good condition, and the Euthyphro’s god is more like it than the All. 2. The common generation of the All and its parts suggests that they share similar properties (cf. 899 d 3–4, 906 a 2–c 6). Thus, seeing the parts as purely instrumental is to ignore their genu­ ine goodness. The All is certainly aware of its parts (Tim. 36 d 8–37 c 5), and valuing them instrumentally would be a failure of the appropriate friendship (cf. Rep. 4, 442 c 4–d 2, 443 d 3–6; 9, 589 b 1–6; Tim. 29 e 1–4, 30 a 2–3). 3. The claim that a whole cannot be in good condition unless its parts are is essential to refuting the impiety that the gods neglect humans (902 a 1–903 a 3). But it is hardly consoling to learn that the All does not neglect you because it uses you instrumentally. The examples of craftsmen’s care for their products’ small parts (an army’s soldiers, a wall’s small stones, 902 d 7–e 3) show that the small parts’ good condition is partially constitutive of the whole’s good condition, not merely instrumental to it. If this is to be consoling to me, this must to involve caring about my good for my own sake. 6.3.  Rational eudaemonism, A1, and the Coincidence A1 claims that each part’s telos is the All’s best condition; rational eudaemonism claims that each person’s telos is her own happiness. 111  If the parts’ virtue partially constitutes the All’s virtue and the All’s virtue partially constitutes the All’s happiness, then (whether or not constitution is necessarily transitive) the parts’ virtue should partially constitute the All’s happiness. If so, the parts’ happiness should partially constitute the All’s happiness.

154

Christopher Bobonich

Since I have argued that Plato endorses rational eudaemonism, we must consider whether it can accommodate A1. We might see A1 as a limit case of non-­instrumentally valuing others’ well-­being. Section 4.1 examined the strategy of valuing others’ well-­being by counting it as an additive part of one’s own and found serious worries. It would be even less plausible here. Every human’s happiness would be more or less equal and equivalent to the All’s, and there is no clear sensible way of discounting. Similarly, although both A1 and rational eudaemonism specify my telos, we should not identify my best condition with the All’s best condition. (Since my virtue is vastly inferior to the All’s, my happiness should be too, 906 b 1–4.) A plausible first step to accommodation is noting that the Coincidence entails that the best outcome for me is always part of the All’s best outcome. So, the course of action recommended by rational eudaemonism and that required by my telos are always the same.112 But there are at least two important worries about this. First, although we have seen that some of Plato’s views make the Coincidence more likely, he offers no argument for it, and apparent counterexamples seem intuitively compelling.113 Consider Jane, who could be a highly virtuous and accomplished philosopher but also has superb business skills. If the rulers send her to business school instead of the Academy, she can generate a surplus sufficient for philosophically educating ten deserving students, one of whom will make a great philosophical advance benefiting vast numbers of future philosophers.114 If the All is benefited by the virtue and happiness of its parts, those gaining in philosophical wisdom and the All seem better off, but Jane does not. Second, leaving aside counterexamples, the Coincidence apparently guarantees that one who lives as well as she can achieves her telos even if she never thinks of the All.115 This response seems 112  The ‘course of action required by my telos’ refers to the course of action that is part of the All’s best outcome. For convenience, I shall say that by so acting I achieve my telos, although I am causally and constitutively responsible only for a tiny part of the All. 113  Cf. D. Brink, Perfectionism and the Common Good (Oxford, 2007), 60–9, 120–4. 114  We cannot reply that Jane’s sacrifice constitutes a compensatory gain in virtue. Even if Plato thought that such sacrifices compensate for intellectual deficits (which is dubious), Jane’s involvement with commerce may have worsened her character, so that her surplus must be involuntarily taxed away: she need not be an effective altruist. 115  We must also ask what makes it true that people with unvirtuous conscious goals strive for the All’s happiness. First, if everyone recollects to some degree,



Is Plato a Consequentialist?

155

disappointing. Further, despite the immense ambition of claiming that a life of optimal well-­being optimally contributes to the All’s well-­being, Plato’s conception of this life is quietist in many respects and does not propose radically new content for virtuous activity.116 Although the Laws takes some steps towards widening other-­ concern’s scope, they are small (and infirm with respect to slaves). But the claim about the human telos is not presented as an interesting theoretical fact. It is a central part of Magnesian education, and Plato must have intended it to shape how reflective Magnesians think of themselves.

6.4.  Book 10’s theology and our self-­understanding Although Book 10’s theology does not radically change the virtuous person’s actions, the content of the telos that the petteia-player establishes for each of the All’s parts and its application to all humans are radically innovative.117 Also new is the way that A1 and the Coincidence encourage the person to see herself and her virtue. She should come to realize the following claims: 1. My telos was established by god, and my original constitution enables me to attain it. This original constitution includes a tendency to strive to fulfil my telos. 2. All humans and gods (except for the petteia-player) are living parts of a single divine animal, the All. 3. My telos is the All’s best condition, and I fulfil my telos by optimally contributing to the All’s good condition. Independently of A1 and the Coincidence, Magnesians are expected to accept the Dependency Theory.

this should facilitate grasping fundamental ethical and theological truths. Second, however misguided one’s beliefs or desires, each human soul consists of psychic motions that are distorted versions of the rational motions that originally entirely constituted it. Reason naturally tends to seek basic truths and the actual good. It is not clear, however, that these considerations are a sufficient answer. 116  Except for a few, it lacks the strong contemplative component found in philo­ sophers’ virtue in the Phaedo and the Republic but is closer to that found among the virtuous citizens of the Statesman’s good city. 117  Precursors occur in Gorgias 507 e 6–508 a 8 and the Euthyphro, if C. Taylor, ‘The End of the Euthyphro’, Phronesis, 27 (1982), 109–18 is right.

156

Christopher Bobonich

4. The outcome that is best for me is the outcome in which I fulfil my telos (from (3) and A2).118 5. The outcome that is best for me is the outcome in which I live as virtuously as possible (Laws 2, 662 c 6–663 b 6, 664 b 7–8). 6. The outcome in which I fulfil my telos is the outcome in which I live as virtuously as possible (from (4) and (5)). 7. My virtue and happiness partially constitute the All’s virtue and happiness (see Section 6.2). 8. I fulfil my telos by optimally contributing to the All’s good condition (3) and optimally contributing to the All’s good condition coincides with my best state (A2, (3), (4)). I am not claiming that my happiness is defined as fulfilling my telos or that fulfilling my telos makes something best for me.119 But seeing my best state in teleological terms deepens my understanding of it: I grasp that I attain my best state when I fulfil my telos. This has significant implications for my understanding of my relation to the divine. God wills the All’s best state because it is best, and since my good condition is partially constitutive of the All’s good condition, he wills the former for its own sake.120 9. By grasping A1, I see that my innate constitution includes a tendency to strive for the All’s best state. By also grasping A2, I see this as an innate tendency to strive for what is best for me. This, too, deepens my understanding of my good condition and allows me to see living virtuously and happily as living in ­accordance with my innate constitution. Other dialogues partially ­anticipate

118  Although virtue is vastly the greatest good and I am always better off being virtuous and acting virtuously, attaining my best condition depends in part on the petteia-player, since in my next incarnation he can allot me more Dependent Goods and more opportunities to progress in knowledge and virtue. This qualification applies elsewhere, but I shall not keep mentioning it. 119  Although it is metaphysically impossible, god (or the bad analogue) could have given me the telos (or quasi-telos) of striving for the All’s worst state or realizing non-­relationally bad properties. Doing so does not seem good for me. 120  This requires something like the principle that if I rationally value X ’s good condition for its own sake and Y’s good condition helps constitute X ’s good condition, then it is rational for me to value Y’s good condition for its own sake. If one is sceptical of this principle, we can supplement it with the argument at the end of Section 6.3.



Is Plato a Consequentialist?

157

this idea, but it plays a fundamental role in the Laws and the Timaeus, which claim that providential deities produce my nature.121 10. A1 at least suggests that I should value the All’s best condition for its own sake and, since this is partially constituted by the good condition of all other humans, that I should value their good condition for its own sake. Plato’s statement of A1 and Book 10’s emotive descriptions of the All (cf. Tim. 92 c 4–9) seem intended to encourage this attitude. Up to (10), we have seen nothing that is inconsistent with the individual taking her own happiness as her ultimate end and seeing it as constituted by living virtuously and using Dependent Goods correctly. Does A1 give us a new motivation and, if so, what is its content? Here, I think, any answer must be somewhat speculative. Plato may not have fully worked out his late theology’s and cosmology’s implications for his ethics. The minimalist position is that I realize that my best condition is compossible with the best condition of the All and other humans, but I value for its own sake my own happiness and not the All’s good condition. The maximalist solution is that I abandon rational eudaemonism and take as my ultimate end the All’s good condition.122 Book 10’s theology seems intended to change my self-­understanding: minimalism ignores this and gives no practical significance to my being a part of the All, while I have argued that maximalism’s abandonment of rational eudaemonism is not text­ual­ly supported. Given what we have seen, it is reasonable to think that we are supposed to come to see the All’s good condition as good for its own sake. How is this possible and what motivates it? Modifying rational eudaemonism by making my ultimate end a combination of my happiness and the All’s best condition seems unpromising since combining them additively is very unattractive. 121  Tim. 90 c 6–d 7; cf. Rep. 9, 588 b 10–592 a 4 and 10, 611 b 9–612 a 6. As in the case of one’s telos, we should not accept that realizing my nature is good for me as a fundamental principle. We need to ask whether the realization is genuinely good. 122  If Plato were to take this route, he would come very close to a cosmic version of welfarist, agent-­neutral consequentialism. Although consequentialism is not a meta-­ethical theory or a theory of moral motivation, few contemporary consequentialists think that its normative force or the typical motivation for acting in accordance with it is eudaemonic. Many think it is, instead, that of rationality. If Plato abandoned rational eudaemonism, it is plausible that he would see rationality as central to the normative force of one’s telos.

158

Christopher Bobonich

We may be able to better implement this basic idea by adopting a strategy used to argue that rational eudaemonism is compatible with valuing virtue for its own sake.123 To be a rational eudaemonist, I must think that a person’s rationally required ultimate end is what is best for her because it is best for her. Let us consider the suggestion that the person count as part of what is best for her the disposition to act to advance the All’s best condition for its own sake. Living virtuously is, as we have seen, necessary for contributing optimally to the All’s best condition, and seeing it as choiceworthy because you value the All’s best condition for its own sake can be part of what it is to be happy in the same way that seeing virtue as good for its own sake can be part of what it is to be happy. This gives the person a new way of seeing her virtue and a new and additional motivation for being virtuous. I am not suggesting that she no longer see virtue as good in itself for herself. First, as I have noted, the argument for the Dependency Thesis has no such theological premise. Next, the World Soul is not part of any other living creature, and the rational order of its soul’s circles is a fundamental part of what is best for it (Tim. 30 b 1–c 1, 36 e 5–37 e 5, 41 c 6–42 a 1, 89 e 3–90 d 7). It is reasonable that this is true for any creature capable of such order. Becoming virtuous is establishing a similar rational order in our own souls which is thus good in itself for us.124 We contribute to the All’s best condition by adding to it our own perfection, whose goodness does not entirely depend on our parthood. This is, I think, a promising strategy for accommodating valuing the All’s best condition for its own sake within rational eudaemonism. But we must still ask what justifies such valuing of the All’s best condition. As we have seen, it is not sufficient that god commands me to act for its sake or has given me innate tendencies to do

123  See J. D’Souza, ‘The Self-­Absorption Objection to Neo-­Aristotelian Virtue Ethics’ [‘Self-­Absorption’], American Catholic Philosophical Quarterly, 92 (2018), 641–68; C.  Toner, ‘The Self-­Centredness Objection to Virtue Ethics’ [‘Self-­ Centredness’], Philosophy, 81 (2006), 595–617. I offer a version of the ‘reconceptualization’ strategy; the other ‘two-­levels’ strategy posits two distinct evaluative systems and seems less plausible for Plato (although it is sometimes hard to distinguish the two strategies). 124  The World Soul need not have irrational aspects for our virtuous condition to be similar to it.



Is Plato a Consequentialist?

159

so.125 For similar reasons, the mere fact that I am part of a whole does not obviously make it the case that seeking its best condition is non-­instrumentally good for me, since it could be a bad whole. But there is some plausibility in this claim as long as it is a case of valuing a genuinely good whole as it is in the All’s case. Here I have only sketched the outlines of the ethical implications of Book 10’s theology, and many questions remain. First, one might deny that you can value something for its own sake when your ultimate justification for your choice is that it is best for you.126 Second, there is the related worry that trying to accommodate valu­ing something such as the good condition of others for its own sake within a eudaemonist framework leads to valuational schizophrenia.127 Third, critics of eudaemonism will still find the attempted accommodation too self-­centred. Finally, some other worries are especially pressing for Plato. First, in considering whether acting in accordance with my nature is good for me and whether I should care for the good of the whole of which I am a part, we have ultimately appealed to the idea that the item in question is really good. For Plato, this is ultimately a matter of whether the thing in question is non-­relationally good. If this is right, then there seems to be pressure to hold that it is rationally required to maximize its overall amount. If, after all, it simply makes the world a better place, full stop, why should it matter that a given quantity of it is located within me rather than you? It may be possible to argue that my independent existence still gives me a special reason to seek my own good, understood as the greatest non-­relational goodness of my own life, but that would require argument. The last worry is perhaps the most pressing. Plato relies on the strong assumption that each person’s best condition and the All’s best condition are compossible, and we have seen that there are convincing counterexamples to this claim. More generally, the cost in time and energy of attaining very high levels of perfection is great, and it is hard to see why diverting some of these resources to the perfection of others could never maximize the overall good. 125  Rep. 7, 520 a 6–b 5 shows that returning benefits is sometimes just, but Plato does not justify our concern with the All by an argument from gratitude. 126 R. Kraut, Aristotle on the Human Good (Princeton, 1999), 211–37. 127  D’Souza, ‘Self-­Absorption’, 645–66; Toner, ‘Self-­Centeredness’, 595–601.

160

Christopher Bobonich

Plato’s quietism and his assumption of compossibility require more defence than he provides. Nevertheless, this strategy of incorporating concern for the All’s best condition within one’s own happiness still gives a central role to a conception of practical rationality as maximizing the good. My ultimate justification for choice is maximizing my own happiness, and what I value in itself is maximally contributing to the All’s best state. I shall close by considering one of the prelude’s basic messages in which Plato’s intentions are clearer. Is learning that we are part of god’s design reassuring (at least to virtuous people)? As Thomas Nagel observes, such a thought is not consoling, unless we share in god’s design ‘in a way in which chickens do not share in the glory of coq au vin’.128 One’s good condition is not merely instrumental to the All’s good condition, but this is not sufficient to give one’s life and virtue significance, since one’s own good condition might be an insignificant part of the All’s good condition. Julia Annas observes that in late Stoicism the idea that I am ‘only a part’ of the universe was invoked to persuade me that my own viewpoint is unimportant and that nothing I do matters.129 Andrea Nightingale characterizes the Laws similarly: From the [All’s] perspective, the individual all but disappears from sight. No matter how he lives—­be it virtuously or viciously—­his contribution to the whole will be the same. He may wonder . . . why his own actions matter at all . . . his soul will be conveyed towards other souls of his ilk in his next incarnation, but this can hardly serve to make his present life more meaningful or important . . . the individual is forced to confront his own trivi­al­ity.130

I think that this gets Book 10’s message wrong. In justifying the third theological proposition—­that the gods cannot be bribed—­the Athenian remarks: [T16]  ἐπειδὴ γὰρ συγκεχωρήκαμεν ἡμῖν αὐτοῖς εἶναι μὲν τὸν οὐρανὸν πολλῶν μεστὸν ἀγαθῶν, εἶναι δὲ καὶ τῶν ἐναντίων, πλειόνων δὲ τῶν μή, μάχη δή, ϕαμέν, ἀθάνατός ἐσθ’ ἡ τοιαύτη καὶ ϕυλακῆς θαυμαστῆς δεομένη, σύμμα­ χοι δὲ ἡμῖν θεοί τε ἅμα καὶ δαίμονες, ἡμεῖς δ’ αὖ κτῆμα θεῶν καὶ δαιμόνων· ϕθείρει δὲ ἡμᾶς ἀδικία καὶ ὕβρις μετὰ ἀϕροσύνης, σῴζει δὲ δικαιοσύνη καὶ  σωϕροσύνη μετὰ ϕρονήσεως, ἐν ταῖς τῶν θεῶν ἐμψύχοις οἰκοῦσαι 128  T. Nagel, ‘The Absurd’, in id., Mortal Questions (Cambridge, 2012), 11–23 at 16. 129 Annas, Morality, 174–6. 130  Nightingale, ‘Historiography’, 323. The Laws’ occasional pessimism (4, 709 a 1–c 3; 7, 803 c 2–804 c 1) is not its predominant view.



Is Plato a Consequentialist?

161

δυνάμεσιν, βραχὺ δέ τι καὶ τῇδε ἄν τις τῶν τοιούτων ἐνοικοῦν ἡμῖν σαϕὲς ἴδοι. (906 a 2–b 4)

for since we have agreed among ourselves that heaven is full of many good things, and of the opposite, and that there are more of those that are not good, we say that battle of this sort is immortal and that it requires extraordinary vigilance, and that the gods and the daimones are our allies, while we in turn are the property of the gods and the daimones. Injustice and hybris, together with foolishness, destroy us, while justice and moderation, together with wisdom, save us— which dwell in the gods’ psychic powers, but someone could clearly see a small portion of such things also dwelling here in us.

Living virtuously is thus to be allied with the gods in a cosmic battle between goodness and badness. The military metaphor emphasizes that allies share a common goal, and this supports the above account of what Book 10’s theology adds to virtuous person’s motivation: the end that we share is the All’s best condition. The prelude’s theology reassures its listener that the All is supportive of virtue. The Athenian’s first argument establishes soul’s priority to body, and he takes this to show that justice is not conventional but is part of the original constitution of things (890 d 1–9, 892 a 2–c 7). As the prelude progresses, we learn that the All itself is not merely bodily (although the bodily has value properties, 906 b 9–c 6) but is a superlatively virtuous god (899 b 2–7, 900 c 9–d 3, 900 d 5–e 4, 901 a 7–9, Tim. 29 e 1). By grasping A1 and the Coincidence, we learn the petteia-player, the All itself, and every human share, in some way, a common goal. We already knew that virtue was best for us, but now see that it has an unexpected dignity and value. Many of us have, with good ­reason, told our students that ancient Greek philosophers were concerned that lives be virtuous and good, but that interest in meaningfulness is distinctively modern. But in stating the atheists’ position, the Athenian emphasizes some of the same features that some contemporaries think reveal life’s meaninglessness. According to the atheists, the world is fundamentally a chance bodily product, while values are merely our invention. The world is uninterested in us and our well-­being. We are trivial and accidental products of blind forces, and our hopes of achieving things of value are mere illusions. What Book 10’s theology shows us is that just people and their virtue not only play a role in god’s plan, but that this role is

162

Christopher Bobonich

significant in its aim and its nature. We see that we have, in some way, the very same goal that the petteia-player and the All do. Not only that, but we pursue this end by engaging in the very same activity that they do, that is, virtuous activity. Finally, we see that our best activity is itself a part of the virtue and happiness of the immortal god that is the All. Stanford University

BIBLIOGR A PH Y Allen, D., ‘Changing the Authoritative Voice: Lycurgus’ Against Leocrates’, Classical Antiquity, 19 (2000), 5–33. Annas, J., An Introduction to Plato’s Republic (Oxford, 1981). Annas, J., The Morality of Happiness [Morality] (Oxford, 1993). Anscombe, G. E. M., ‘Modern Moral Philosophy’, Philosophy, 33 (1958), 1–16. Baima, N. and Paytas, T., ‘True in Word and Deed: Plato on the Impossibility of Divine Deception’, Journal of the History of Philosophy, 58 (2020), 192–214. Baltzly, D., ‘Is Plato’s Timaeus Panentheistic?’, Sophia, 49 (2010), 193–215. Blair, E., Plato’s Dialectic on Women (New York, 2012). Bobonich, C., ‘Plato on Legal Normativity’ [‘Normativity’], Ancient Philosophy Today, suppl. vol. 4 (2022), 24–44. Bobonich, C., Plato’s Utopia Recast [Utopia] (Oxford, 2002). Bobonich, C., ‘A Reconstruction and Analysis of the Republic’s Harm Principle (Rep. 1, 335 a 6–336 a 10)’, https://stanford.academia.edu/ChrisBobonich. Bobonich, C., ‘Socrates and Eudaimonia’, in D.  Morrison (ed.), The Cambridge Companion to Socrates (Cambridge, 2011), 293–332. Brink, D., Perfectionism and the Common Good (Oxford, 2007). Broadie, S., ‘Theological Sidelights from Plato’s Timaeus’ [‘Theological’], Proceedings of the Aristotelian Society, suppl. vol. 82 (2008), 1–17. Broome, J., Weighing Goods (Cambridge, Mass., 1991). Cantu, G., ‘Individual and Polis in Plato’s Republic’ [‘Individual and Polis’], Polis, 28 (2011), 90–107. Carone, G., Plato’s Cosmology and its Ethical Dimensions (Cambridge, 2005). Chang, R., Making Comparisons Count (New York, 2015). Cooper, J. (ed.), Plato: Complete Works (Indianapolis, 1997). Cooper, J., ‘The Psychology of Justice in Plato’, American Philosophical Quarterly, 14 (1977), 151–7.



Is Plato a Consequentialist?

163

Des Places, É. and Diès, A., Platon, Œuvres Complètes, vols. xi–­xii (Paris, 1951–6). Dorsey, D., ‘Consequences’, in D. Portmore (ed.), Consequentialism (New York, 2020), 92–112. Driver, J., ‘The History of Utilitarianism’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2014 edn), https://plato.stanford. edu/archives/win2014/entries/utilitarianism-history/, accessed 15 January 2023. D’Souza, J., ‘The Self-Absorption Objection to Neo-Aristotelian Virtue Ethics’ [‘Self-Absorption’], American Catholic Philosophical Quarterly, 92 (2018), 641–68. Duke, E. A., Hicken, W. F., Nicoll, W. S. M., Robinson, D. B., and Strachan, J. C. G. (eds.), Platonis Opera, vol. i: Euthyphro, Apologia, Crito, Phaedo, Cratylus, Theaetetus, Sophista, Politicus (Oxford, 1995). England, E., The Laws of Plato, 2 vols. (Manchester, 1921). Fine, G., ‘Forms as Causes: Plato and Aristotle’, in id., Plato on Knowledge and Forms: Selected Essays (New York, 2003), 350–96. Foot, P., ‘Utilitarianism and the Virtues’, Mind, 94 (1985), 196–209. Halliwell, S., Plato: Republic V (Liverpool, 1998). Hammerton, M., ‘Relativized Rankings’, in D.  Portmore (ed.), Conse­ quentialism (New York, 2020), 46–66. Harry, C. and Polansky, R., ‘Plato on Women’s Natural Ability: Revisiting Republic V and Timaeus 41 e 3–44 d 2 and 86 b 1–92 c 3’, Apeiron, 49 (2016), 261–80. Harvey, F., ‘Two Kinds of Equality’, Classica et Mediaevalia, 26 (1965), 101–46. Hsieh, N. and Andersson, H., ‘Incommensurable Values’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2021 edn), https:// plato.stanford.edu/archives/fall2021/entries/value-incommensurable/, accessed 16 January 2023. Hurka, T., British Ethical Theorists from Sidgwick to Ewing [British Ethical Theorists] (Oxford, 2014). Irwin, T., Plato’s Ethics (Oxford, 1995). Irwin, T., ‘Review of M. Nussbaum, The Fragility of Goodness’, Journal of Philosophy, 85 (1988), 376–83. Kamtekar, R., ‘Distinction without a Difference? Plato on “Genos” vs. “Race” ’, in J. Ward and T. Lott (eds.), Philosophers on Race: Critical Essays (Oxford, 2002), 1–13. Kamtekar, R., ‘Social Justice and Happiness in the Republic: Plato’s Two Principles’, History of Political Thought, 22 (2001), 189–220. Keller, S., Partiality (Princeton, 2013). Kraut, R., Aristotle on the Human Good (Princeton, 1999). Kraut, R., ‘Review of G. Santas, Goodness and Justice’, Ancient Philosophy, 25 (2005), 446–63.

164

Christopher Bobonich

Laks, A., ‘Private Matters in Plato’s Laws’, in C.  Horn (ed.), Platon: Gesetze/Nomoi (Sankt Augustin, 2013), 165–88. Mayhew, R., Plato Laws 10 (Oxford, 2008). McCabe, M., ‘Out of the Labyrinth: Plato’s Attack on Consequentialism’, in C. Gill (ed.), Virtue, Norms, and Objectivity (Oxford, 2005), 189–215. Mohr, R., ‘The Platonic Theodicy: Laws 899–905’ [‘Theodicy’], in id., Gods and Forms in Plato (Las Vegas, 2005), 197–201. Moore, G. E., Ethics (London, 1965). Moore, G. E., Principia Ethica, ed. by T. Baldwin, rev. edn (Cambridge, 1993). Morrison, D., ‘The Happiness of the City and the Happiness of the Individual in Plato’s Republic’ [‘Happiness’], Ancient Philosophy, 21 (2001), 1–24. Morrow, G., Plato’s Cretan City (Princeton, 1960). Morrow, G., Plato’s Law of Slavery in its Relation to Greek Law [Law] (New York, 1976). Mulgan, T., The Demands of Consequentialism (Oxford, 2001). Mynott, J. (trans.), Thucydides: The War of the Peloponnesians and the Athenians (Cambridge, 2013). Nagel, T., ‘The Absurd’, in id., Mortal Questions (Cambridge, 2012), 11–23. Nightingale, A., ‘Historiography and Cosmology in Plato’s Laws’ [‘Historiography’], Ancient Philosophy, 19 (1999), 299–326. Nussbaum, M., ‘The Speech of Alcibiades: A Reading of the Symposium’, in id., The Fragility of Goodness (Cambridge, 1986), 165–99. Ober, J., Political Dissent in Democratic Athens (Princeton, 1998). Phillips, D., Sidgwickian Ethics (Oxford, 2011). Portmore, D., ‘Introduction’, in id. (ed.), Consequentialism, 1–21. Portmore, D. (ed.), The Oxford Handbook of Consequentialism [Consequen­ tialism] (New York, 2020). Price, A., ‘Martha Nussbaum’s Symposium’, Ancient Philosophy, 11 (1991), 285–99. Rawls, J., A Theory of Justice [Theory of Justice], rev. edn. (Cambridge, Mass., 1999). Reeve, C. D. C., Philosopher-Kings (Princeton, 1988). Reeve, C. D. C., Plato Laws (Indianapolis, 2022). Saunders, T., ‘Epieikeia: Plato and the Controversial Virtue of the Greeks’, in F. Lisi (ed.), Plato’s Laws and its Historical Significance (Sankt Augustin, 2001), 65–94. Saunders, T., ‘Plato on Women in the Laws’, in A. Powell (ed.), The Greek World (London, 1995), 591–609. Saunders, T., Plato’s Penal Code [Penal Code] (New York, 1991). Scheffler, S., ‘Agent-Centered Restrictions, Rationality, and the Virtues’, Mind, 94 (1985), 409–19. Scheffler, S., The Rejection of Consequentialism (Oxford, 1982).



Is Plato a Consequentialist?

165

Schofield, M., Plato: Political Philosophy [Political Philosophy] (Oxford, 2006). Schöpsdau, K., Nomoi (Gesetze): Platon: Werke, vol. ix (Göttingen, 1994–2011). Scott, D., ‘The Republic’, in G. Fine (ed.), The Oxford Handbook of Plato (Oxford, 2019), 208–27. Sedley, D., Creationism and its Critics in Antiquity (Berkeley, 2007). Sedley, D., ‘Platonic Causes’, Phronesis, 43 (1998), 114–32. Sidgwick, H., The Methods of Ethics [Methods], 7th edn. (Indianapolis, 1981). Singer, P., ‘Famine, Affluence, and Morality’, Philosophy and Public Affairs, 1 (1972), 229–43. Slings, S., Platonis Rempublicam (Oxford, 2003). Smart, J. and Williams, B., Utilitarianism For and Against [Utilitarianism] (Cambridge, 1987). Speak, D., The Problem of Evil (Cambridge, 2015). Taylor, C., ‘The End of the Euthyphro’, Phronesis, 27 (1982), 109–18. Toner, C., ‘The Self-Centredness Objection to Virtue Ethics’ [‘SelfCenteredness’], Philosophy, 81 (2006), 595–617. Tooley, M., The Problem of Evil (Cambridge, 2019). Viljoen, G.  van N., ‘Plato and Aristotle on the Exposure of Infants at Athens’, Acta Classica, 2 (1959), 58–69. Vlastos, G., ‘Does Slavery Exist in Plato’s Republic?’, in id., Platonic Studies, 2nd edn (Princeton, 1981), 140–6. Vlastos, G., ‘Slavery in Plato’s Thought’, in id., Platonic Studies, 2nd edn (Princeton, 1981), 147–63. Vlastos, G., ‘The Theory of Social Justice in the Polis in the Republic’, in id., Studies in Greek Philosophy, 2 vols. (Princeton, 1995), ii, 69–103. Vlastos, G., ‘Was Plato a Feminist?’, in id., Studies in Greek Philosophy, 2 vols. (Princeton, 1995), ii. 133–43. Wallace, R., ‘The Legal Regulation of Private Conduct at Athens: Two Controversies on Freedom’, Etica & Politica/Ethics & Politics, 9 (2007), 155–71. Warren, J., ‘Comparing Lives in Plato, Laws 5’, Phronesis, 58 (2013), 319–46. Waterlow [Broadie], S., ‘The Good of Others in Plato’s Republic’, Proceedings of the Aristotelian Society, n.s. 73 (1972–3), 19–36. Williams, B., Ethics and the Limits of Philosophy (Cambridge, Mass., 1985). Williams, B., Moral Luck (Cambridge, 1981). Woollard, F. and Howard-Snyder, F., ‘Doing vs. Allowing Harm’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2016 edn), https://plato.stanford.edu/archives/win2016/entries/doing-allowing/, accessed 16 January 2023.

ARISTOTLE’S ARGUMENT FOR THE NECESSITY OF WHAT WE UNDERSTAND joshua mendelsohn 1.  We think we only understand necessities? Aristotle famously introduces unqualified understanding or epistēmē haplōs, the central intellectual achievement he will be concerned with in Posterior Analytics 1, as follows: [1] Ἐπίστασθαι δὲ οἰόμεθ’ ἕκαστον ἁπλῶς, ἀλλὰ μὴ τὸν σοϕιστικὸν τρόπον τὸν κατὰ συμβεβηκός, ὅταν τήν τ’ αἰτίαν οἰώμεθα γινώσκειν δι’ ἣν τὸ πρᾶγμά ἐστιν, ὅτι ἐκείνου αἰτία ἐστί, καὶ μὴ ἐνδέχεσθαι τοῦτ’ ἄλλως ἔχειν. (Post. An. 1. 2, 71b9–12) We think we understand [ἐπίστασθαι] something without qualification (and not in the sophistical way, incidentally) whenever we think we know the explanation because of which the thing is, that it is its explanation, and also that it cannot be otherwise.1

On the interpretation favoured by most commentators, Aristotle is claiming that this intellectual achievement requires not only grasping some fact together with its explanation; the fact whose ex­plan­ation

Few of my colleagues, mentors, and friends have not heard me give this paper or discussed the ideas in it with me at some stage, and every one of these conversations has improved the paper itself or the thinking behind it. At the risk of causing offence, I should single out Marko Malink and Victor Caston, who both provided extensive comments on the paper at different stages, Martha Nussbaum, Gabriel Lear, Michael Kremer, Agnes Callard, David Bronstein, Arnold Brooks, Justin Vlasits, audiences at the 2017 Australasian Association of Philosophy Conference, the November 2019 Medieval Philosophical Gathering at the Humboldt University of Berlin, the October 2021 Ancient Philosophy Society at Northwestern University, students in my Fall 2021 graduate seminar, faculty and students who frequented the University of Chicago’s Ancient Greek and Roman Philosophy Workshop, and, for much more than just proofreading, Bronwyn Mendelsohn. 1  For the text of the Prior and Posterior Analytics I rely on W.  D.  Ross (ed.), Aristotle’s Prior and Posterior Analytics [Prior and Posterior Analytics] (Oxford, 1949). Translations from the Posterior Analytics are based on J. Barnes, Aristotle: Posterior Analytics, 2nd edn [Posterior Analytics, 2nd edn] (Oxford, 1993), with some modifications. Other translations are my own, unless marked.

Joshua Mendelsohn, Aristotle’s Argument for the Necessity of What We Understand In: Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Joshua Mendelsohn 2023. DOI: 10.1093/oso/9780192885180.003.0005

168

Joshua Mendelsohn

we grasp must itself be necessary.2 Aristotle frequently repeats the claim that what we have epistēmē of—­the epistēton or object of understanding—­is necessary or ‘cannot be otherwise’,3 and this claim serves as an important premiss in several arguments he gives in Posterior Analytics 1.4 Aristotle’s reasons for holding it, however, are not easy to discern. Aristotle presents his characterization of what it is to understand in [1] as a point of general agreement, something ‘we think’ (οἰόμεθ’, 2  See J. Barnes, Posterior Analytics, 2nd edn, 90–1; R. D. McKirahan, Principles and Proofs (Princeton, 1992), 23; W.  Detel, Aristoteles: Analytica posteriora [Aristoteles], 2 vols. (Berlin, 1993), ii. 37–8; G.  Fine, ‘Aristotle on Knowledge’, Elenchos, 14 (2010), 121–55, reprinted in ead., Essays in Ancient Epistemology (Oxford 2021), 221–42 at 223; D. Bronstein, Aristotle on Knowledge and Learning: The Posterior Analytics [Aristotle on Knowledge and Learning] (Oxford, 2016), 36; and M.  Burnyeat, ‘Aristotle on Understanding Knowledge’, in E.  Berti (ed.), Aristotle on Science: The Posterior Analytics (Padua, 1981), 97–139 at 98; but see also L. Angioni, ‘Aristotle’s Definition of Scientific Knowledge’, Logical Analysis and History of Philosophy 19 (2016), 140–66 for an unorthodox interpretation according to which Aristotle only means that a cause of a scientifically known fact is necessarily that cause and no other. I will assume the orthodox interpretation here. Two other ambiguities about this passage deserve brief mention. First, the scope of γινώσκειν is ambiguous in [1], which makes it unclear whether the last clause says only that the things of which we have epistēmē are in fact necessary or that we know that they are. Most interpreters favour the latter, but I won’t take a stand on this issue here. My interest is in the claim that it is of necessities, which follows on all of the orthodox interpretations listed above. Second, Aristotle’s term πρᾶγμα is general enough to cover not just propositional knowledge, and so not just ‘facts’ known, but also e.g. knowing the eclipse. See discussion in Bronstein, Aristotle on Knowledge and Learning, 55 and, on this issue in the context of Post. An. 1. 33, see G. Fine, ‘Aristotle’s Two Worlds: Knowledge and Belief in Posterior Analytics 1. 33’ [‘Aristotle’s Two Worlds’], Proceedings of the Aristotelian Society, 110 (2010), 323–46, reprinted in ead., Essays in Ancient Epistemology, 243–61 at 247–8. Aristotle is at least saying that we understand necessary facts or states of affairs, whether or not he also thinks that we know some other kind of ‘necessity’, and my focus here will be on the claim about factual knowledge. See Detel, Aristoteles, ii. 38 and Angioni, ‘Aristotle’s Definition of Scientific Knowledge’, 142–5 for reasons to take Aristotle to be talking exclusively about knowledge of facts in [1]. 3 See Post. An. 1. 4, 73a22 and 1. 33, 88b31, 89a10 for ‘necessary’ (ἀναγκαῖον). ‘Cannot be otherwise’ renders a number of closely related expressions: μὴ ἐνδέχεσθαι . . . ἄλλως ἔχειν (1. 2, 71b12), ἀδύνατον ἄλλως ἔχειν (1. 2, 71b15–16; 1. 4, 73a21; 1. 33, 89a7), οὐ δυνατὸν ἄλλως ἔχειν (1. 6, 74b6), οὐκ ἐνδέχεται ἄλλως ἔχειν (1. 33, 88b32). Expressions rendered by ‘necessary’ and ‘cannot be otherwise’ appear to function as ways of saying the same thing in a technical and a non-­technical register: See 1. 2, 71b9–16; 1. 4, 73a21–3; and 1. 33, 88b30–89a10. 4  In particular at Post. An. 1. 4, 73a21, where it is used to argue that the premisses of demonstrations must hold per se; at 1. 6, 74b5–6, where it is employed in order to establish that the principles of demonstrations must also be necessary; and at 1. 33, 88b30–1, where this conclusion is then used to argue that epistēmē and doxa have distinct objects. I discuss these passages below.



The Necessity of What We Understand

169

Post. An. 1. 2, 71b9), or, as he says elsewhere, that ‘we all presume’ (πάντες . . . ὑπολαμβάνομεν).5 This presumption is correct, he tells us, since all people take themselves to satisfy the characterization in [1] when they have epistēmē, and people who really have epistēmē truly do satisfy it, concluding (ὥστε, 71b15) that ‘that of which we have unqualified understanding cannot be otherwise’ (οὗ ἁπλῶς ἔστιν ἐπιστήμη, τοῦτ᾿ ἀδύνατον ἄλλως ἔχειν, 71b15–16). While this might serve to elicit assent from those who already accept this claim, perhaps without having reflected upon their acceptance of it, this statement will obviously not convince anyone who doubts that epistēmē is of necessities. It takes as a premiss that people who have epistēmē really are in the condition described in [1], which is to say inter alia that what people understand is a necessity. As an argument for the claim that epistēmē is of necessities, it is plainly circular. Given this, we might wonder whether Aristotle is simply reporting a feature of the way the verb ἐπίστασθαι is used in Greek. As Jonathan Barnes points out, however, Aristotle’s claim does not capture how this verb functions any more than it describes how ‘know’ functions in English.6 There is, in both languages, the ‘epistemic’ use of modal language, as when in English we say that something ‘must’ be so in order to express that we take what we know to imply it.7 But to say that what we know is ‘necessary’ in this sense is to say only that any fact known is implied by the sum total of our knowledge.8 This is clearly not all that Aristotle means: in elaborating the claim that epistēmē is of necessities, he adds that they are ‘eternal’ (ἀίδια) and ‘subject neither to generation nor corruption’ (ἀγένητα καὶ ἄϕθαρτα, NE 6. 3, 1139b24; cf. Post. An. 1. 8, 75b24–30). There is little reason to think that ordinary Greek speakers would have assented to such characterizations of what 5  NE 6. 3, 1139b19–20. See also Post. An. 1. 2, 71b9, 13–15 and 1. 33, 89a6–9. 6 Barnes, Posterior Analytics, 2nd edn, 91. Cf. R. Bolton, ‘Science and Scientific Inquiry in Aristotle: A Platonic Provenance’ [‘Science and Scientific Inquiry’], in C. Shields (ed.), The Oxford Handbook of Aristotle (Oxford, 2012), 46–59 at 46–7, Fine, ‘Aristotle’s Two Worlds’, 251. 7  For examples of δεῖ used to express epistemic necessity, see E. Ruiz Yamuza, Tres verbos que significan ‘deber’ en griego antiguo (Zaragoza, 2008), 107–13. On the use of modal predicates to express epistemic necessities in ancient Greek see also S. Danesi, C. Johnson, and J. Barðdal, ‘Where Does the Modality of Ancient Greek Modal Verbs Come from?’, Journal of Greek Linguistics, 18 (May 2018), 45–92. 8  On this, see also A. Kratzer, ‘What “Must” and “Can” Must and Can Mean’, Linguistics and Philosophy, 1 (1977), 337–55.

170

Joshua Mendelsohn

they knew any more readily than speakers of English will.9 Why, then, does Aristotle hold that the object of epistēmē is a necessity—­ let alone present this as a point of general agreement? Scholars who do not put Aristotle’s claim down to confusion10 often point out that Aristotle is not attempting to capture every way that epistēmē and its cognates were legitimately used.11 Aristotle only means to be talking about what we aim for in science and other systematic endeavours. This requires a principled understanding of some topic or field;12 it is the type of knowledge we ascribe when we praise people as knowledgeable, rather than when we say that someone knows how to get home. ‘Scientific knowledge’, ‘dis­cip­ lin­ary mastery’, and ‘understanding’ have, with good reason, been suggested as alternative translations,13 and I will speak here of ‘understanding’ or ‘scientific understanding’14 to distinguish Aristotle’s topic from other sorts of knowledge.15 9  R.  Pasnau, ‘Epistemology Idealized’, Mind, 122 (2013), 987–1021 at 991 remarks: ‘No conversation with an ordinary Athenian, no matter how one-­sided, could plausibly have elicited the result that knowledge concerns a proposition that is necessary and universal.’ But see J. Hintikka, ‘Time, Truth and Knowledge in Ancient Greek Philosophy’, American Philosophical Quarterly, 4 (1967), 1–14 at 7 for a different view. 10  For a pessimistic appraisal, see T. Ebert, ‘Review of Mignucci, L’argomenta­ zione dimostrativa in Aristotele and Barnes, Aristotle: Posterior Analytics’ [‘Review of Mignucci’], Archiv für Geschichte der Philosophie, 62 (1980), 85–91 at 89–90, who takes Aristotle’s theory of science to rest on a scope fallacy. Ebert takes Aristotle to slide from the correct but mundane observation that if S  knows  p, it necessarily follows that p is true, to the erroneous thesis that if S knows p, then p is ne­ces­sar­ily true. 11  See Barnes, Posterior Analytics, 2nd edn, 91, as well as Detel, Aristoteles, ii. 54, M.  Mignucci, L’argomentazione dimostrativa in Aristotele (Padua, 1975), 16–17, Bolton, ‘Science and Scientific Inquiry’, 52 and C.  C.  W.  Taylor, ‘Aristotle’s Epistemology’, in S. Everson (ed.), Epistemology (Cambridge, 1990), 116–42 at 116. 12  See Burnyeat, ‘Aristotle on Understanding Knowledge’, 106, and J. H. Lesher, ‘On Aristotelian ἐπιστήμη as “Understanding” ’ [‘Ἐπιστήμη as “Understanding” ’], Ancient Philosophy, 21 (2001), 45–55 at 50. 13  See Ross, Prior and Posterior Analytics, for ‘scientific knowledge’; Burnyeat, ‘Aristotle on Understanding Knowledge’, for ‘understanding’; and Lesher, ‘Ἐπιστήμη as “Understanding” ’, for ‘disciplinary mastery’. Fine, ‘Aristotle’s Two Worlds’, 246, takes Aristotle to be talking about ‘High-­Level Knowledge’, at least in Post. An. 1. 33, but Fine, ‘Aristotle on Knowledge’, 225–6 is ambivalent about how to take Aristotle elsewhere. 14  I use the modifier ‘scientific’ only for emphasis; in what follows, the terms ‘understanding’ and ‘scientific understanding’ should be taken to be synonymous. 15  In translating epistēmē as ‘understanding’, I do not mean to endorse Burnyeat’s thesis that Aristotelian epistēmē should be thought of as understanding rather than knowledge (Burnyeat, ‘Aristotle on Understanding Knowledge’, 102). As Fine has



The Necessity of What We Understand

171

If it is only understanding in this narrow sense that Aristotle takes to be of necessities,16 then Aristotle’s claim does not rule out other types of knowledge or understanding having contingencies as their objects. Hence, it need not be taken to defy ordinary language. When Aristotle cites this as a point of general agreement, he may be indicating agreement regarding this specific kind of know­ ledge between himself and his philosophical peers.17 It would then be understandable that Aristotle moves over the claim swiftly. Even if we grant all this, however, we are still far from understanding the motivation for this view. Why should Aristotle—­or any philosopher—­take necessities to be what we understand in this specific sense? What is it about epistēmē in the sense at issue which restricts its objects to necessities?18 pointed out, Aristotle’s close association of epistēmē with explanation does not show that he is not using it to describe a kind of justified true belief. He might be understood as giving an account of justification which requires grasping an appropriate explanation, and an account of knowledge which requires this sort of justification (Fine, ‘Aristotle on Knowledge’, 232). In that case, we could say that epistēmē is a kind of knowledge tantamount to understanding. For further recent discussion of this issue with a focus on Plato, see W. Schwab, ‘Explanation in the Epistemology of the Meno’, Oxford Studies in Ancient Philosophy, 48 (2015), 1–36 and id., ‘Understanding epistēmē in Plato’s Republic’, Oxford Studies in Ancient Philosophy, 51 (2016), 41–85. 16  This is presumably part of what Aristotle means when he specifies that he is  ‘speaking in a precise way’ (ἀκριβολογεῖσθαι, NE 6. 3, 1139b19) or referring to epistēmē in an ‘unqualified’ sense (ἁπλῶς, Post. An. 1. 2, 71b9). 17  Proposals vary regarding who in particular Aristotle might have taken himself to agree with: Plato (Bolton, ‘Science and Scientific Inquiry’); ‘the Academy’ (S. Broadie (pers. comm.) and C. Rowe (trans.), Aristotle: Nicomachean Ethics [Nicomachean Ethics] (New York, 2002), 365); or Aristotle’s own school (J. Barnes, Aristotle’s Posterior Analytics, 1st edn [Posterior Analytics, 1st edn] (Oxford, 1975), 97). Ebert, ‘Review of Mignucci’, 90, Barnes, Posterior Analytics, 2nd edn, 91, and Burnyeat, ‘Aristotle on Understanding Knowledge’, 108 n. 23, on the other hand, reject the interpretive hypothesis that Aristotle means to restrict his claim to some group of philosophical peers. 18 This is a question which we must ask even if, like Robert Pasnau, we take Aristotle to be describing the ‘ideal limit of human inquiry’, rather than a cognitive state that Aristotle thinks he or perhaps anyone has actually achieved (Pasnau, ‘Epistemology Idealized’, 994). If Pasnau’s hypothesis is correct, then it is perhaps easier to see why Aristotle takes epistēmē to require a grasp of an explanation, since it is plausible to think that the best cognitive grasp of reality as a whole would include knowing not just facts but understanding the reason why these facts hold (Pasnau, ‘Epistemology Idealized’, 995). Pasnau does not, however, explain why Aristotle requires epistēmē to be only of necessities, and the supposition that Aristotle is talking about an ideal cognitive state does little to explain this. Why should the ideal type of knowledge consist in only knowing a certain type of fact (viz. necessities)? It seems, on the face of it, equally plausible to say that the ideal

172

Joshua Mendelsohn

In Myles Burnyeat’s view, the fact that epistēmē in the relevant sense is scientific understanding explains this immediately. Since, according to Burnyeat, science is concerned to explain ‘general regularities’ and such regularities are ‘lawlike’, the objects of scientific knowledge are ‘necessary connections’.19 If this is Aristotle’s intended argument, it is nowhere made explicit. Further, the view that science is concerned only with the general and necessary is less common now than it was when Burnyeat was writing; it is, in any case, not self-­evident.20 Aristotle’s own views regarding the place of particulars in scientific know­ledge are not straightforward,21 but even if we grant that Aristotle takes only generalizations to be susceptible to scientific ex­plan­ation, that still does not explain why he should take scientific understanding to be only of necessities. For, famously, Aristotle distinguishes two kinds of scientific generalization: those which occur merely ‘for the most part’ (ὡς ἐπὶ τὸ πολύ) and those which occur ‘of necessity’ (ἐξ ἀνάγκης).22 For Aristotle, then, scientific generalizations are not to be equated with ‘necessary connections’: there are also explicable regularities which permit of exceptions. Hence, even if Burnyeat is right that Aristotle takes science to deal only with generalizations, this does not explain why he restricts scientific understanding to necessities.23 aim of inquiry should be to know the world in all of its contingent detail. For this reading, see also Taylor, ‘Aristotle’s Epistemology’, 122, and discussion in Fine, ‘Aristotle on Knowledge’, 224–6. 19  ‘What gets explained in the sciences . . . is general regularities and connections: lawlike regularities in the modern jargon, necessary connections in Aristotle’s’ (Burnyeat, ‘Aristotle on Understanding Knowledge’, 109). 20  This is stressed in N.  Cartwright, ‘Why Trust Science?’, Proceedings of the Aristotelian Society, 120 (2020), 237–52. Burnyeat’s points of reference seem to be proponents of the Deductive-­Nomological model defended in the middle of the century by Carl Hempel and those building on this account (see especially C. G. Hempel, ‘Aspects of Scientific Explanation’, in id., Aspects of Scientific Explanation and Other Essays in the Philosophy of Science (New York, 1965), 331–469; M. Friedman, ‘Explanation and Scientific Understanding’, Journal of Philosophy, 71 (1974), 5–19). 21  Aristotle consistently treats scientific knowledge of particulars as secondary in the Post. An. (see especially 1. 8, 75b24–6; 1.31, 87b30–88a2), but he never denies that facts about particulars can be explained, and 87b39–88a2 presupposes that they can be, at least in some cases. He also seems to take a different view in Metaph. Μ. 10, 1087a15–21, and possibly Pr. An. 2. 21, 67a27–67b5. 22  Metaph. Ε. 2, 1026b27–35. See also Post. An. 1. 30, 87b19–25; 1. 32, 88b7–8; and Pr. An. 1. 13, 32b18–22. 23 This criticism could be extended to L.  P.  Gerson, Ancient Epistemology (Cambridge, 2009), 67–9 and B.  van Fraassen, ‘A Re-­Examination of Aristotle’s Philosophy of Science’, Dialogue, 19 (1980), 20–45 at 25–8, who both emphasize



The Necessity of What We Understand

173

Robert Bolton takes a different approach to this problem. He argues that both Aristotle’s claim and his argument for the claim are carried over from Plato.24 While he does not deny that understanding must, for Aristotle, be accompanied by scientific ex­plan­ ation, he differs from Burnyeat in taking the ‘root idea’ (‘Science and Scientific Inquiry’, 50) motivating Aristotle’s various assertions about epistēmē and its objects to be a Platonic view that knowledge possesses ‘a certain, strong reliability’ and is such as to ‘not ever rationally let you down’ (53). In Bolton’s view, this motivates the restriction of scientific knowledge to necessities in the following way: the only sort of knowledge which is absolutely reliable is, for Plato and Aristotle, a grasp of a thing’s essence, or some knowledge that derives in an appropriate way from this grasp. But Aristotle takes all of the facts that what is explained in science is a non-­incidental regularity in their respective expositions of Aristotle’s view. 24 F.  Solmsen, Die Entwicklung der Aristotelischen Logik und Rhetorik (Berlin, 1929), 143, also argues that the theory of demonstration, together with its requirement that epistēmē be of necessities, was an expression of Aristotle’s early Platonism and that the theory was only later refined into a more distinctive view in the Prior Analytics. This interpretation is criticized in Ross, Prior and Posterior Analytics, 6–22, but Ross does not address Solmsen’s specific claim that Aristotle’s association of epistēmē with necessity in the Posterior Analytics is a Platonic holdover. I will not here try to establish whether Plato held that epistēmē or some other kind of know­ ledge was only of necessities. In any case, the evidence Bolton cites is not decisive. As he admits (Bolton, ‘Science and Scientific Inquiry’, 48), Plato tends not to use the word ‘necessary’ (ἀναγκαῖον) to describe the special character of knowledge of what a thing is, preferring to describe the object of this type of knowledge as what ‘is’, in a technical sense that is opposed to what ‘becomes’ (for example at Tim. 27 d 5–28 a 1). Even if this is in some sense a precursor to Aristotle’s notion of necessity, Aristotle’s claim is that the objects of scientific knowledge have a particular modal status; Plato’s claim in the Timaeus is that they have a kind of being that is excluded from becoming. While Plato’s distinction between objects of epistēmē and objects of doxa in Republic 5 has traditionally been interpreted as expressing a close relative of Aristotle’s view that epistēmē is of a necessity, Gail Fine has argued that Aristotle ‘restricts epistēmē to what’s necessary, whereas Plato does not do so’ (Fine, ‘Aristotle on Knowledge’, 232; see also ead., ‘Knowledge and Belief in Republic V’, Archiv für Geschichte der Philosophie, 60 (1978), 121–39). In fact, Aristotle’s talk of necessity in epistemic contexts more closely echoes language used by Parmenides, who takes the object of the favoured path of inquiry to be bound by ‘powerful Necessity’ (κρατερή . . . Ἀνάγκη, Parmenides D 8. 35–6 in the edition of A. Laks and G. W. Most, Early Greek Philosophy (Cambridge, MA, 2016); cf. D 6. 2). For interpretations on which Parmenides holds a thesis about knowledge similar to that of Aristotle as I interpret him, see H. White, What Is What-­Is? A Study of Parmenides’ Poem (New York, 2005) and especially J.  Palmer, Parmenides and Presocratic Philosophy (Oxford, 2009).

174

Joshua Mendelsohn

about a thing’s essence to be necessary. Hence, anything that we know with the strong reliability characteristic of epistēmē must be a necessity (‘Science and Scientific Inquiry’, 51–2). Even if we grant that Aristotle is committed to the premisses of this argument, I do not think this adequately captures the way that he argues. Bolton does not point to a passage where Aristotle presents this line of reasoning,25 and in fact, the way that Aristotle introduces the notion of essential predication into his discussion of understanding and its objects suggests a different and in­com­pat­ ible order of explanation. In the Posterior Analytics, essentialist notions are principally theorized in 1. 4, where Aristotle defines what it means for predications to hold ‘in themselves’ or per se (καθ’ αὑτά) and ‘universally’ (καθόλου) as certain types of essential predication.26 He prefaces this discussion with the following remark: [2] Ἐπεὶ δ’ ἀδύνατον ἄλλως ἔχειν οὗ ἔστιν ἐπιστήμη ἁπλῶς, ἀναγκαῖον ἂν εἴη τὸ ἐπιστητὸν τὸ κατὰ τὴν ἀποδεικτικὴν ἐπιστήμην· . . . ἐξ ἀναγκαίων ἄρα συλλογισμός ἐστιν ἡ ἀπόδειξις. ληπτέον ἄρα ἐκ τίνων καὶ ποίων αἱ ἀποδείξεις εἰσίν. πρῶτον δὲ διορίσωμεν τί λέγομεν τὸ κατὰ παντὸς καὶ τί τὸ καθ’ αὑτὸ καὶ τί τὸ καθόλου. (Post. An. 1. 4, 73a21–7)

Since it is impossible for that of which there is understanding simpliciter to be otherwise, what is understandable in virtue of demonstrative understanding will be necessary. . . . A demonstration, then, is a deduction which proceeds from necessities. We must see, then, from what items, i.e. from what kind of items, demonstrations proceed. First let us define what we mean by ‘of every case’, by ‘in itself’, and by ‘universally’.

This is not the remark we would expect if Aristotle’s intention were to argue that epistēmē must be of necessities because it consists in a grasp of essences. Instead, Aristotle says that we must consider these essentialist notions because scientific knowledge is of ne­ces­ sities. When, in Posterior Analytics 1. 6, he goes on to argue that the 25  Bolton, ‘Science and Scientific Inquiry’, 51–2, takes Aristotle to be following Plato in making these claims. 26  Aristotle introduces four senses of per se, but connects only two of these with the character of the object of understanding, at least explicitly (73b16–17; see, however, M. Ferejohn, The Origins of Aristotelian Science (New Haven, 1991), 109–31 and id., Formal Causes (New York, 2013), 91–4 on the relevance of the other senses). In these first two senses, predications hold per se when the predicate occurs ‘in the account saying what [the subject] is’ (ἐν τῷ λόγῳ τῷ λέγοντι τί ἐστιν) or, conversely, when the subject occurs in the predicate’s essence-­specifying account (73a34–b3). ‘Universal’ (καθόλου) is then defined in this chapter as being per se and said of all cases (73b26–7).



The Necessity of What We Understand

175

premisses of demonstrations must, in fact, have the status of per se predications, he explicitly calls on the necessity of what we understand as a premiss: [3] Εἰ οὖν ἐστιν ἡ ἀποδεικτικὴ ἐπιστήμη ἐξ ἀναγκαίων ἀρχῶν (ὃ γὰρ ἐπίσταται, οὐ δυνατὸν ἄλλως ἔχειν), τὰ δὲ καθ’ αὑτὰ ὑπάρχοντα ἀναγκαῖα τοῖς πράγμασιν . . ., ϕανερὸν ὅτι ἐκ τοιούτων τινῶν ἂν εἴη ὁ ἀποδεικτικὸς συλλογισμός· ἅπαν γὰρ ἢ οὕτως ὑπάρχει ἢ κατὰ συμβεβηκός, τὰ δὲ συμβεβηκότα οὐκ ἀναγκαῖα.

Ἢ δὴ οὕτω λεκτέον, ἢ ἀρχὴν θεμένοις ὅτι ἡ ἀπόδειξις ἀναγκαίων ἐστί, καὶ εἰ ἀποδέδεικται, οὐχ οἷόν τ’ ἄλλως ἔχειν· ἐξ ἀναγκαίων ἄρα δεῖ εἶναι τὸν συλλογισμόν. (Post. An. 1. 6, 74b5–15) If demonstrative understanding proceeds from necessary principles (for what we know cannot be otherwise), and what holds of an object in itself is necessary . . . then it is clear that demonstrative deductions will proceed from certain items of this sort [viz. per se predications]; for everything holds either in this way or incidentally, and what is incidental is not necessary. We must either argue like this or else posit as a principle that demonstration is of necessities, i.e. that if something has been demonstrated it cannot be otherwise—­the deduction, therefore, must proceed from necessities.27

In this passage, Aristotle is assuming a strict dichotomy between per se predications, which are necessary, and incidental predications, which are not. On the basis of this assumption (whose problems need not concern us here),28 he presents two paths of argument to the conclusion that the premisses of demonstrations are per se predications: either we argue from the claim that the premisses of demonstrations (what demonstrations are ‘from’) are necessary 27  Here I read ἀναγκαίων at 74b14 with the OCT against Barnes’s reading of ἀναγκαίον. This gives a clearer contrast between the two alternatives (we argue either from the premiss that a demonstration is from necessities or from the premiss that it is of necessities), but the sense does not depend on this choice, since Aristotle makes explicit that he means that ‘if something has been demonstrated, it cannot be otherwise’ (εἰ ἀποδέδεικται, οὐχ οἷόν τ’ ἄλλως ἔχειν, 74b14–15). 28  Among other problems, the so-­called ‘common axioms’, like the principle of non-­contradiction, are apparently necessities in Aristotle’s view, but not necessities grounded in the essence of any given thing. On this issue, see R. Bolton, ‘Aristotle on Essence and Necessity’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, 13 (1997), 113–38, esp. 117–19; M. Peramatzis, ‘Aristotle on How Essence Grounds Necessity’, in D. Bronstein, T. Johansen, and M. Peramatzis (eds.), Aristotelian Metaphysics, Ancient & Modern (Oxford, forthcoming), and M. Peramatzis, ‘Aristotle on Knowledge & Belief: APo. I. 33’ [‘Aristotle on Knowledge & Belief’] (unpublished), esp. 15–16.

176

Joshua Mendelsohn

truths or we argue from the claim that the conclusions of demonstrations (what they are ‘of’) are necessary.29 In neither case is his strategy to argue that what we understand is a necessity on the basis that it is a grasp of the essence of something; rather, he claims that we must argue for the essentiality of demonstrative premisses on the basis of the necessity of demonstrative premisses or conclusions. The same pattern persists throughout the Posterior Analytics: Where we might expect to find an argument, we instead find Aristotle assuming the necessity of what we understand as a premiss. This is true in particular of 1. 33, which is devoted to clarifying the claim that understanding is of necessities while opinion (doxa) is of contingencies.30 Opinion is of contingent truths, Aristotle says, because opinion and understanding (in its demonstrative and non-­ demonstrative varieties) are the only cognitive states that are of truths,31 but understanding (of both these sorts) is of necessities (88b33–7). Clearly, again, the necessity of the object of understanding is a premiss rather than a conclusion. The bulk of the chapter is then devoted to explaining ‘how it is possible to opine and understand the same thing’ (πῶς . . . ἔστι τὸ αὐτὸ δοξάσαι καὶ ἐπίστασθαι, 89a11), given that understanding is of necessities, while opinion is of contingencies. Aristotle’s answer to this question, which has been interpreted in a variety of ways, need not concern us here.32 The point is that this whole discussion is predicated on the claim that understanding is of necessities. The closest Aristotle provides to an argument for this is that it ‘agrees with how things appear’ (ὁμολογούμενον . . . τοῖς ϕαινομένοις, 89a4–5) because: 29  Cf. Barnes, Posterior Analytics, 2nd edn, 126. 30  See J. Moss and W. Schwab, ‘The Birth of Belief’, Journal of the History of Philosophy, 57 (2019), 1–32, esp. 6–7, for the translation of doxa as ‘opinion’ here. 31  This is what I take him to mean by saying that they are the only states that are ‘true’ (ἀληθές) at Post. An. 1. 33, 89a2. There is a difficulty in reconciling this with his view at NE 6. 3, 1139b15–17, where he appears to recognize a wider variety of true cognitive states, which I will not attempt to resolve here. 32  On Fine’s preferred reading (Fine, ‘Aristotle’s Two Worlds’, 261), Aristotle argues for a ‘Two Worlds Theory’ according to which we can have doxa only of propositions of the form ‘it is contingent that . . . ’ and epistēmē only of propositions of the form ‘it is necessary that . . . ’; they are ‘of’ the same thing only in that the subjects of these statements can be the same. B.  Morison, ‘Aristotle on the Distinction between What Is Understood and What Is Believed’ (unpublished), and Peramatzis, ‘Aristotle on Knowledge & Belief’ both defend a view on which we can have epistēmē of propositions that do not include an explicit necessity operator so long as we understand the proposition as being rendered necessary by essential facts. See also L. Angioni, ‘Knowledge and Opinion about the Same Thing in APo A-­33’, Dois Pontos, 10 (2013), 255–90.



The Necessity of What We Understand

177

[4] πρὸς δὲ τούτοις (i) οὐδεὶς οἴεται δοξάζειν, ὅταν οἴηται ἀδύνατον ἄλλως ἔχειν, ἀλλ’ ἐπίστασθαι· (ii) ἀλλ’ ὅταν εἶναι μὲν οὕτως, οὐ μὴν ἀλλὰ καὶ ἄλλως οὐδὲν κωλύειν, τότε δοξάζειν, ὡς τοῦ μὲν τοιούτου δόξαν οὖσαν, τοῦ δ’ ἀναγκαίου ἐπιστήμην. (Post. An. 1. 33, 89a6–10)

In addition, (i) no one thinks they have an opinion in relation to something when they think that something cannot be otherwise; rather, they think they understand it. (ii) On the other hand, when [people think] that something is so but nothing prevents it from being otherwise, then [people think] they have an opinion, since opinion is of the former sort of thing, while understanding is of necessities.33

On the one hand, Aristotle says, echoing the language of [1], p ­ eople think they have epistēmē rather than doxa when they think that something cannot be otherwise. He adds that people think they have opinion rather than understanding whenever they take something to be contingent. Aristotle accepts these ‘appearances’, as Gail Fine notes, but this acceptance hardly amounts to a defence of the position that understanding is of necessities.34 It may be tempting to conclude that Aristotle never really attempts to defend his claim that epistēmē is of necessities beyond these appeals to consensus. This, however, would be a mistake. We do find one explicit, albeit brief argument for the claim that the object of understanding is a necessity in the corpus which is not clearly question-­ begging or merely an appeal to consensus. It occurs in NE 6. 3, in the course of Aristotle’s discussion of the intellectual virtues: [5] πάντες γὰρ ὑπολαμβάνομεν, ὃ ἐπιστάμεθα, μηδ’ ἐνδέχεσθαι ἄλλως ἔχειν· τὰ δ’ ἐνδεχόμενα ἄλλως, ὅταν ἔξω τοῦ θεωρεῖν γένηται, λανθάνει εἰ ἔστιν ἢ μή. ἐξ ἀνάγκης ἄρα ἐστὶ τὸ ἐπιστητόν. (NE 6. 3, 1139b19–23)

We all think that what we understand cannot be otherwise. With what can be otherwise, we do not know whether it is or not whenever it goes out of view. Therefore, the object of understanding is of necessity.35

33  I have numbered the sentences (i) and (ii) for the purposes of exposition. 34  Fine, ‘Aristotle’s Two Worlds’, 259–60. [4] (i) is, explicitly, another appeal to consensus; depending on whether we take the final clause of [4] (ii) to fall within the scope of οἴηται, [4] (ii) is either another appeal to consensus or an argument that this appearance is correct, since understanding really is of necessities. In the latter case, the restriction of understanding to necessities serves again as a premiss rather than a conclusion. 35  For the text of the NE I rely on I. Bywater (ed.), Aristotelis Ethica Nicomachea (Oxford, 1894).

178

Joshua Mendelsohn

This passage is seldom a point of focus in the literature, probably owing to its brevity. As Bolton reads this passage, Aristotle’s ‘thesis that epistêmê is of necessary truths is defended on the ground that epistêmê is something you should be able to reliably count on even apart from continued observation of the state of affairs in question to see that it does not change’ (‘Science and Scientific Inquiry’, 53). Aristotle’s first premiss, on this reading, is that epistēmē is something we can ‘count on’—a state that we can continue to possess and employ—­whether or not we observe the state of affairs we know. Given that he draws the conclusion that understanding is of necessities, Aristotle must be assuming, on this reading, that if epistēmē were of contingencies, then it could not be reliable in this way, since we would then need to ‘observe’ the relevant state of affairs in order to see whether it holds. Aristotle says nothing here about an object of understanding being something that you can ‘count on’, however. Instead, he talks about the circumstances under which we cease to have knowledge of contingencies. Even if the idea that epistēmē is reliable in some way motivates Aristotle’s premiss in [5]—and I will argue that it does—­it is not accurate to gloss what Aristotle claims in [5] in this way. More to the point, Bolton does not explain on what grounds Aristotle might hold the extra premiss needed to make his argument valid. Why, that is, should knowledge of contingencies be unreliable unless we engage in some kind of ongoing observation of the state of affairs in question? At least on the face of it, this is implausible: it seems that I can know, for example, the contingent fact that Socrates died by drinking hemlock. This requires observation neither for its acquisition nor for its continued retention. It is my object in this paper to elucidate Aristotle’s reasoning in [5], and thus to explain the argument Aristotle actually gives for his claim that understanding is of necessities. While I agree with Bolton that the reliability of epistēmē is one of the central ideas underlying Aristotle’s argument in [5], I maintain that there is another, equally basic element of Aristotle’s conception of epistēmē which we must take into account in order to understand his reason­ ing. This is the idea that epistēmē is a relative, and thus depends on the existence of its object. These two features of epistēmē are outlined in Categories 7–8, which places epistēmē in the category of relatives and classifies it as a permanent state. Text [5] presupposes that epistēmē is both a relative and a permanent state, and in effect



The Necessity of What We Understand

179

argues that it can have the features characteristic of both of these classes only if it is of what is necessary. Aristotle’s view may thus be understood as an attempt to reconcile two theses that emerge from his analysis of understanding in the Categories. I proceed as follows. First (Section 2), I examine Aristotle’s claim that understanding is a ‘relative’ (πρός τι) in Categories 7. Then (Section 3), I consider the grounds on which Aristotle cat­egor­izes understanding as a ‘state’ (ἕξις) in Categories 8. I argue that these two characterizations lead to a tension. In Section 4, I explain why one tempting way to resolve the tension is not available to Aristotle. In Section 5, I show how the assumption that understanding is both a relative and a permanent state underlies Aristotle’s argument that the object of understanding is a necessity in [5]. Objections to the argument are considered in Section 6, and I say something about the upshot for how we should understand Aristotle’s linguistic remarks and his debt to his predecessors in closing.

2.  Understanding as a relative Understanding is assigned to the category of relative in Categories 7, the category which includes ‘all such things as are said to be just what they are, of other things, or in some other way in relation to something else’,36 as with the larger (6a38), the double (a39), and master and slave (b29–30). Understanding passes this test for being a relative: just as a larger thing is said to be larger than something, and a double the double of something, so too ‘understanding is understanding of something’ (ἡ ἐπιστήμη τινὸς ἐπιστήμη, b5).37

36  Cat. 7, 6a36–7: τὰ τοιαῦτα λέγεται, ὅσα αὐτὰ ἅπερ ἐστὶν ἑτέρων εἶναι λέγεται ἢ ὁπωσοῦν ἄλλως πρὸς ἕτερον. For the text of the Categories I rely on L. Minio-­Paluello (ed.), Aristotelis Categoriae et liber De Interpretatione (Oxford, 1949). Translations of the Categories are my own, but I have consulted J. L. Ackrill (trans.), Aristotle: Categories and De Interpretatione [Categories and De Interpretatione] (Oxford, 1975) and sometimes follow his translation closely. 37 Typically, the qualification is given by a genitive expression, but Aristotle gives no indication that the correlative must always occur in the genitive: His use of ὁπωσοῦν ἄλλως at 6b7–8 in fact suggests that he means to allow other grammatical cases or prepositional constructions. Occasionally he also uses examples with the dative: a similar thing is a relative because it is said to be similar to something else (τινί, 6b9). Philoponus explicitly allows the correlative to be in the dative (In Cat. 106. 8–11 Busse).

180

Joshua Mendelsohn

Aristotle makes clear that relatives are not substances.38 Like all beings outside the category of substance, relatives depend on an underlying subject which they exist ‘in’ (ἐν), and this may be one respect in which relatives exist in relation to something else.39 In the case of understanding, Aristotle takes the relevant subject to be an animal capable of understanding (5, 3a4–5; 7, 7a37) or, more precisely, that animal’s soul (2, 1b1–2). The distinctive type of dependency that characterizes relatives, however, is not their inherence in subjects but their dependency on correlatives, beings that they are said to be ‘of’, ‘than’, or ‘otherwise in relation to’. To call something the larger, for instance, is in Aristotle’s view ipso facto to call it larger than something else. In correspondence to this gram­mat­ ical fact, Aristotle sees a metaphysical reality: ‘relatives are those things for which to be is the same as to be related to something in a certain way’.40 This statement can be understood in two ways. On the one hand, Aristotle might mean that for a relative to exist is for it to be related 38 8a16–18. Aristotle speaks here of primary substances, but the context makes clear that he wishes to deny also that secondary substances are relatives. I will not take a stand here on whether it is best to view relatives as relational objects (M. Duncombe, Ancient Relativity: Plato, Aristotle, Stoics and Sceptics [Ancient Relativity] (Oxford, 2020)) or as relational properties (D. Yates and A. Marmodoro, ‘Introduction: The Metaphysics of Relations’, in A.  Marmodoro and D.  Yates (eds.), The Metaphysics of Relations (Oxford, 2016), 1–18): I maintain only that relatives are the kind of item that exists ‘in’ a substance, whatever those turn out to be. If relatives are properties, however, they will on my interpretation need to include particular qualities (so, not just being larger or understanding in general, but the particular being-­larger of a larger squirrel and the particular understanding of a particular student, etc.). 39  Cf. F. Morales, ‘Relational Attributes in Aristotle’, Phronesis, 38 (1994), 255– 74 at 257–258, 261; and P. M. Hood, Aristotle on the Category of Relation (Lanham, 2004), 7–8. 40  Cat. 7, 8a31–2: τὰ πρός τι οἷς τὸ εἶναι ταὐτόν ἐστι τῷ πρός τί πως ἔχειν. The relationship of this definition to the one given at 6a36–7 has been discussed since an­tiquity. I remain neutral here on whether, as most ancient and many modern commentators contend, the difference between linguistic usage and metaphysics is primarily what is at issue when Aristotle provides his revised definition (for this view, see Hood, Aristotle on the Category of Relation, 39; Morales, ‘Relational Attributes in Aristotle’, 260; Ackrill, Categories and De Interpretatione, 101), or whether Aristotle is making a different distinction and merely clarifying en passant that questions about relatives are questions of a metaphysical nature (as argued in D. Sedley, ‘Aristotelian Relatives’, in M. Canto-­Sperber and P. Pellegrin (eds.), Le style de la pensée: Recueil de textes en hommage à Jacques Brunschwig (Paris, 2002), 324–52). On either reading, Aristotle’s considered view is that being a relative is, when we are speaking in the strictest sense, a matter of metaphysics and not only language.



The Necessity of What We Understand

181

to something in a certain way. Alternatively, we might take Aristotle to mean that for some subject to be R (where R is a relative) is for that subject to be related to something else in a certain way. Aristotle’s examples fit better with the second option. It is not the case that for a slave to exist is for that slave to be related in a certain way to a master: a slave does not cease to exist when liberated; rather, that person simply ceases to be a slave. It is, however, plaus­ible that for a person to be a slave is for that person to be related in a certain way to another person who is a master. I will take Aristotle to hold, in general, that for any relative R with correlative C, for some subject SR to be R requires SR to bear an appropriate relationship to some SC that is C. We might notice that, in a case like master and slave, the converse also holds: not only does something being a slave require something else to be a master; it is also true that something can be a master only if something else is a slave. Aristotle asks whether this holds in general by introducing the notion of being ‘sim­ul­tan­eous in nature’ (ἅμα τῇ ϕύσει): [6] Δοκεῖ δὲ τὰ πρός τι ἅμα τῇ ϕύσει εἶναι. καὶ ἐπὶ μὲν τῶν πλείστων ἀληθές ἐστιν· ἅμα γὰρ διπλάσιόν τέ ἐστι καὶ ἥμισυ, καὶ ἡμίσεος ὄντος διπλάσιόν ἐστιν, καὶ δούλου ὄντος δεσπότης ἐστίν· ὁμοίως δὲ τούτοις καὶ τὰ ἄλλα. καὶ συναναιρεῖ δὲ ταῦτα ἄλληλα· μὴ γὰρ ὄντος διπλασίου οὐκ ἔστιν ἥμισυ, καὶ ἡμίσεος μὴ ὄντος οὐκ ἔστι διπλάσιον· ὡσαύτως δὲ καὶ ἐπὶ τῶν ἄλλων ὅσα τοιαῦτα. (Cat. 7, 7b15–22)

Relatives seem to be simultaneous in nature. And in most cases, this is true: at the same time there is a double, there is a half, and when there is a half, there is a double, and when there is a slave, there is a master. Likewise with the others. And they also are eliminated together with each other: when there is no double, there is no half, and when there is no half, there is no double, and likewise in all other cases of this sort.41

Aristotle uses the term ‘simultaneous in nature’ (I will write ‘sim­ ul­tan­eous’ for short) for a type of bidirectional dependence that holds between a relative R and its correlative C. On the in­ter­pret­ ation of relatives I have offered, this condition comes to the following: for any time t, (i) if something is R at t, then something is C at t, and conversely (ii) if something is C at t, then something is R at t. The second sentence states an immediate corollary: in order 41 Cf. Cat. 13, 14b27–32.

182

Joshua Mendelsohn

for R and C to be simultaneous, it must be the case that (iii) if all Cs cease to be C, then all Rs cease to be R, and also (iv) if all Rs cease to be R, then all Cs cease to be C. Aristotle states in [6] that most relative-­correlative pairs are simultaneous, indicating that some are not.42 This is consistent with what we have seen so far. For although Aristotle holds that relatives as such depend on their correlatives in the manner described in (i) and, consequently, (iii), he does not claim that the mere existence of the correlative is in general sufficient for a relative to be what it is. While something’s being larger is sufficient for something else to be smaller, and vice versa,43 the inherence of a relative in its subject may in other cases require more than the bearer of its correlative continuing to be such. It may also depend on the bearer of the relative having further, non-­relational properties.44 Having the relative attribute R might, in other words, only in part be a matter of there being something that is C, so that the persistence of the correlative is necessary, but not sufficient for the persistence of the relative. In that case, while the relative will still depend on its correlative, the correlative will not depend on the relative in a symmetrical way, so that (ii) and consequently (iv) fail to hold.45 42  Cat. 7, 7b22; but see n. 45 below. Duncombe, Ancient Relativity, 106–12, takes there to be two conditions discussed in [6], ‘simultaneity in nature’ and an unnamed relation of temporal concurrence. He takes Aristotle to introduce the former relation without further explaining it and takes [6] from καὶ ἐπὶ μὲν τῶν πλείστων ἀληθές ἐστιν onwards to describe this unnamed relation. It would be strange for Aristotle to introduce a relation, not discuss it, and then immediately discuss a different, unnamed relation. I think it is more plausible to take [6] to concern a single relation called ‘simultaneity in nature’. In this respect my reading is closer to Hood, Aristotle on the Category of Relation, 34. My reading, however, agrees with Duncombe’s on the substantive point that a correlative does not always exist at the same time as the relative (see Ancient Relativity, 112). 43 Cf. Metaph. Ι. 6, 1057a1–2. 44  Cf. Morales, ‘Relational Attributes in Aristotle’, 257–9. 45 Simplicius has a different interpretation of [6]. He takes all relatives to be simultaneous. In order to make this fit with the text, he points to Aristotle’s use of the word ‘seems’ (δοκεῖν) in stating his conclusions at 7b15 and 24 and claims on this basis that the counter-­examples Aristotle presents are not meant as genuine counter-­ examples to the simultaneity of relatives (Simpl., In Cat. 193. 33–4 Kalbfleisch). Aristotle’s conclusion that ‘in most cases it is true’ (ἐπὶ μὲν τῶν πλείστων ἀληθές ἐστιν, 7b15–16, emphasis added) that relatives are simultaneous is, however, unhappy on Simplicius’ reading, since this carries the conversational implicature that there are some cases in which it is not true. If Aristotle took it to be true in all cases, we would not expect him to say merely that it was in most cases true. I will thus work on the



The Necessity of What We Understand

183

2.1.  Understanding is not simultaneous with its object Aristotle takes simultaneity to fail, in particular, in the case of understanding and its object: [7] οὐκ ἐπὶ πάντων δὲ τῶν πρός τι ἀληθὲς δοκεῖ τὸ ἅμα τῇ ϕύσει εἶναι· τὸ γὰρ ἐπιστητὸν τῆς ἐπιστήμης πρότερον ἂν δόξειεν εἶναι . . . τὸ μὲν ἐπιστητὸν ἀναιρεθὲν συναναιρεῖ τὴν ἐπιστήμην, ἡ δὲ ἐπιστήμη τὸ ἐπιστητὸν οὐ συναναιρεῖ· ἐπιστητοῦ γὰρ μὴ ὄντος οὐκ ἔστιν ἐπιστήμη—­οὐδενὸς γὰρ ἔτι ἔσται ἐπιστήμη—­ἐπιστήμης δὲ μὴ οὔσης οὐδὲν κωλύει ἐπιστητὸν εἶναι. (Cat. 7, 7b22–31)

It is not held to be true in all cases of relatives that they are simultaneous in nature: the object of understanding might be held to be prior to understanding. . . . When the object of understanding is eliminated, understanding is eliminated with it, but when understanding is ­eliminated, the object of understanding is not eliminated with it. For if there is no object of understanding, there will be no understanding—­ there will be nothing for understanding to be of—­but if there is no understanding, there is nothing to prevent there still being an object of understanding.

Aristotle states, as we should expect, that understanding and its object satisfy condition (i) of simultaneity. Whenever someone’s soul is in a condition of understanding, then something else, the correlative of that understanding, is an object of understanding. That implies that understanding also satisfies (iii): for if, at any time, the correlative is no longer an object of understanding, then at that time the state of the soul will no longer be a state of understanding it; otherwise we would have a violation of (i). Understanding and the object of understanding do not, however, constitute a simultaneous relative-­correlative pair, because they fail condition (ii) and, by the same token, condition (iv). The fact that something is an object of understanding does not, for Aristotle, imply that anything actually understands it: he says that there is ‘nothing to prevent there still being an object of understanding’ (οὐδὲν κωλύει ἐπιστητὸν εἶναι), even if there is no understanding of it (7b30–1 [7]). Hence, we cannot reason, as in the case above, that the destruction of the relative, the understanding, assumption that Aristotle means to endorse these as genuine counter-­examples. The use of ‘seems’ (δοκεῖ) can be explained, as Simplicius himself notes (189. 27–9), in other ways: as an expression of uncertainty or, more plausibly, as expressing that it is widely (but, Aristotle thinks, falsely) believed that all relatives are simultaneous.

184

Joshua Mendelsohn

would bring about an elimination of the object of understanding. Instead, an object of understanding can both pre- and post-­exist understanding of it. Aristotle illustrates how the object of understanding can exist before understanding of it with the example of ‘squaring the ­circle’ (ὁ τοῦ κύκλου τετραγωνισμός), which he assumes, at least for the sake of argument, to be possible but undemonstrated (7b31–2). The theorem is, then, an object of understanding in the sense that it is the type of thing which can be understood, but understanding of it does not yet exist, because it has not been demonstrated.46 Aristotle provides another case to illustrate a different way that the simultaneity condition fails for understanding and its object: [8] ἔτι ζῴου μὲν ἀναιρεθέντος οὐκ ἔστιν ἐπιστήμη, τῶν δ’ ἐπιστητῶν πολλὰ ἐνδέχεται εἶναι. (Cat. 7, 7b33–5) Again, if an animal ceases to exist, its understanding will not exist, but many of the objects of its understanding may still exist.

This sentence illustrates the failure of condition (iv) of sim­ul­tan­eity for understanding and its object. Aristotle has us consider what takes place upon the death of an animal that understands something. The animal’s death is sufficient for its understanding perishing, because, as noted above, understanding is dependent on the animal as well as being dependent on its object: understanding can only exist ‘in’ that animal’s soul, and hence for as long as that ­animal is alive.47 Yet the death or other psychic harm to an animal with understanding need induce no change in the object of its understanding, the worldly thing that it understands. Nor do these cease to be objects of understanding when the animal dies: they remain intelligible, ready to be understood by others, even though they are no longer actually understood by that animal.48 46  Cf. Philop., In Cat. 121. 15–16 Busse. 47  I take no stand here on whether Aristotle takes the soul to be immortal. If he does, Aristotle could maintain that the soul ceases to exist as a subject for understand­ ing when the animal dies, while continuing to exist in some other way. On the immortality of the soul in Aristotle, see S. Menn, ‘Aristotle’s Definition of Soul and the Programme of the De Anima’, Oxford Studies in Ancient Philosophy, 22 (2002), 83–139 at 87. 48  On an alternative reading suggested by the translation of Ackrill, Categories and De Interpretatione, 21, Aristotle is considering a scenario in which the entire genus of animal perishes, rather than one particular animal (cf. Cat. 13, 15a6–7). In this case, Aristotle’s point is basically the same but more emphatic: even if every



The Necessity of What We Understand

185

Understanding thus fails to be simultaneous with its correlative, because an object of understanding remains an object of understanding even after the one who understands it—­and thus the understanding of it—­is eliminated.49 Unlike the double and the half, then, understanding and the object of understanding are not on a par as relatives. The object of understanding can both pre-­exist it and post-­exist it. Aristotle elaborates on this point in his treatment of relatives in Metaphysics Δ. 15, where he draws a distinction between a relative that is ‘rela­ tive because that which it is itself is said to be that very thing of [sc. in relation to] something else’50 and that which is only relative because ‘something else is said to be relative to it’.51 Relatives of the first type, which Aristotle calls relatives per se (καθ’ αὑτά, Metaph.

animal were to go out of existence, many of the things that these animals understood would remain things that could be understood. 49  Simplicius, in line with his view that all relatives are simultaneous (see n. 45), takes this to be a merely apparent counter-­example to the simultaneity of relatives. He remarks: ‘one should compare what is potential with what is potential, and what is actual with what is actual, and in this way say that relatives are simultaneous [sc., in all cases]’ (ἔδει γὰρ τὸ μὲν δυνάμει πρὸς τὸ δυνάμει παραβάλλειν, τὸ δὲ ἐνεργείᾳ πρὸς τὸ ἐνεργείᾳ, καὶ οὕτως ἅμα λέγειν τὰ πρός τι, Simpl., In Cat. 196. 28–9 Kalbfleisch, trans. Fleet). Simplicius’ remark is embedded in a complex discussion of the views of Philo and Diodorus, which it would exceed the scope of this paper to consider, but his idea seems to be that we can distinguish between an actual and a potential object of understanding, and that the actual object of understanding is the correlative of actual understanding, while the potential object of understanding is the correlative of potential understanding. Both of these relative–­correlative pairs are sim­ul­tan­eous. Even if all animals cease to exist, Simplicius holds, only actual objects of understanding are eliminated, not potential objects of understanding. Since the actually understanding animals and the potential objects of understanding do not form a relative–­correlative pair, this is not a genuine counter-­example to the sim­ul­tan­eity of relatives. However, it is not clear why elimination of all of the animals eliminates only actual understanding and not also potential understanding, since the capacity of all of these animals to understand is also, presumably, thereby eliminated. Simplicius bolsters his point by appealing to the understanding ‘in the unmoved cause’ (ἐν τῷ ἀκινήτῳ αἰτίῳ, 194. 22), but he himself seems to admit that this reading is difficult to square with the text when he claims that Aristotle ‘sets this out better and more systematically in the Metaphysics’ (βέλτιον δὲ αὐτὸς καὶ πραγματειωδέστερον ἐν τοῖς Μετὰ τὰ ϕυσικὰ περὶ τούτων διατάττεται, 194. 3–4) and that the Categories’ treatment of relatives serves only ‘to exercise the minds of his r­ eaders in anticipation’ (προκεκινῆσθαι ἤδη τὴν διάνοιαν τῶν ἀκροατῶν, 194. 10). 50  πρός τι τῷ ὅπερ ἐστὶν ἄλλου λέγεσθαι αὐτὸ ὅ ἐστιν, Metaph. Δ. 15, 1021a27–8; see Jaeger’s note in the apparatus criticus on αὐτὸ ὅ ἐστιν in id. (ed.), Aristotelis Metaphysica (Oxford, 1957), which I use for the Greek text of the Metaphysics. 51  τῷ ἄλλο πρὸς ἐκεῖνο, Metaph. Δ. 15, 1021a28–9. Cf. Metaph. Ι. 6, 1056b36–1057a1.

186

Joshua Mendelsohn

Ι. 6, 1056b34),52 are those that are principally theorized in Categories 7: relatives which are ‘just what they are’ (αὐτὰ ἅπερ ἐστίν) by being of something else (6a36–7). The object of understanding is given as an example of the latter type of relative in Metaphysics Δ. 15 as well, along with the object of thought (διανοητόν, 1021a31) and the object of measurement (μετρητόν, 1021a29; cf. Metaph. Ι. 6, 1057a7–8). These are called relatives only by courtesy of having something else that is essentially relative being relative to them.53 These items have the superficial features of a relative (an object of understanding is said to be the object of understanding ‘of’ something understood or capable of being understood), but they lack the metaphysical dependence on another characteristic of per se relatives, just as a measurable object does not depend in any real way on its being measured.54 The main claims of Categories 7 and Metaphysics Δ. 15 as they regard understanding are, therefore, the following: (1) understanding is a relative per se, (2) as such, part of what it is for someone’s soul to be in a condition of understanding is to bear an appropriate relation to the object of understanding, and so (3) someone having understanding implies the existence of something that is the object of their understanding; (4) the object of understanding is also a relative, however only by courtesy of understanding being relative to it, and so (5) the existence of the object of understanding does not imply that there is any understanding of it. For our purposes, the most important of these is (3), which I will call the dependency principle. 2.2.  What are the objects of understanding? In order to clarify the import of the dependency principle, we must ask what type of entity Aristotle takes an object of understanding to be, and what precisely it means for this type of entity to exist. We might assume that Aristotle is talking about objects in the sense of primary substances—­things like Socrates, the moon, or a plant. Aristotle’s dependency principle would then state that scientific understanding of Socrates, for instance, requires Socrates to exist. 52  I am not claiming that this coincides with what Aristotle calls relatives καθ’ ἑαυτά at Metaph. Δ. 15, 1021b3–4. 53 T.  Kiefer, Aristotle’s Theory of Knowledge (London, 2007), 29, calls these ‘rela­tive relatively’. 54  Metaph. Ι. 6, 1057a9–12. Cf. Kiefer, Aristotle’s Theory of Knowledge, 31–2.



The Necessity of What We Understand

187

However, Aristotle’s usage suggests that he has in mind objects with predicative structure, things that consequently may be said to hold or fail to hold.55 We have already seen one example of this kind. The squaring of the circle is not an ‘object’ in the sense of a primary substance, as a circle perhaps is, but rather a predicative entity that may be said to exist just in case a certain mathematical object, the square, has the property that it can be constructed with area equal to that of a given circle. Aristotle also uses the phrase ‘object of understanding’ to refer to entities with predicative structure throughout the Posterior Analytics, where the objects of understanding are what one grasps when one grasps the conclusion56 or, sometimes, a premiss of a demonstration.57 The premisses and conclusions of demonstrations are sentences, paradigmatically subject–predicate sentences of the form ‘P holds of S ’. Thus, what one grasps in knowing a premiss or conclusion of a demonstration is presumably something with the structure corresponding to a predicative sentence rather than a substance. I will take Aristotle to mean ‘object’ in this sense when he claims that the objects of understanding are ‘necessary’.58 I will not attempt 55 At Top. 4. 4, 125b4, Aristotle entertains, counterfactually, that a man or a soul might be the object of understanding. The counterfactual context means we should not place too much weight on this, however: Aristotle may be talking about the types of things that others might treat as objects of understanding rather than the types of things that he thinks are properly described as such. Together with Cat. 7, 7b31, these are the only places I have found where Aristotle gives an explicit ex­ample of the type of thing he means by ‘object of understanding’ (ἐπιστητόν). 56  ἐπιστητόν is used to refer to what one grasps when one grasps a conclusion of a demonstration at Post. An. 1. 4, 73a22; 1. 24, 86a6–7; and 1. 33, 88b30. 57  Post. An. 1. 4, 73b16–18. This may be what Aristotle calls ‘non-­demonstrative’ (ἀναπόδεικτος) understanding at Post. An. 1. 33, 88b36. Without the qualification, Aristotle usually means the understanding of a demonstrative conclusion, but see Bronstein, Aristotle on Knowledge and Learning, 51–61. Since demonstrative understanding is not tied to the grasp of any particular token demonstration but rather the ability to produce a given type of demonstration (e.g. the ability to demonstrate that triangles have their characteristic angle sum), what one grasps corresponds to the conclusion or premiss of a given type of demonstration. 58  Fine, ‘Aristotle’s Two Worlds’, 246–8 argues for a similar thesis with respect to the objects of epistēmē in Post. An. 1. 33, adducing as evidence that Aristotle ex­pli­ cit­ly refers to doxa as being ‘of a proposition’ (προτάσεως, Post. An. 1. 33, 89a2–4). In the Categories, Aristotle does occasionally describe a theory or a body of know­ledge like grammar (γραμματική) or the arts (μουσική) as what our understanding is ‘of’, as at Cat. 8, 11a29–31, but this does not make them objects of understanding, at least not in the sense which is here at issue. Aristotle tends to avoid the word ἐπιστητόν for these bodies of knowledge like grammar and the arts, and prefers to call them ‘sciences’ (ἐπιστῆμαι, Cat. 8, 11a25–32). This is also his preferred word for bodies of

188

Joshua Mendelsohn

here to specify their nature fully.59 For our purposes, the important points are only these: first, ‘objects’ in this sense have distinct parts corresponding to a subject (like ‘the circle’) and a predicate (like ‘being squarable’, i.e. having the property that a square of equal area can be constructed using a straight edge and compass) and are responsible for the truth of the corresponding predicative sentences. Second, Aristotle does not conceive of the relationship between the subject- and predicate-­entities in a predicatively structured entity statically.60 Rather, he seems to think of a predicatively structured entity as something that may exist at one time but not at another. In particular, it exists when the entity corresponding to the predicate holds of the entity corresponding to the subject, and it does not exist when the entity corresponding to the predicate fails to hold of the entity corresponding to the subject.61 Thus, for the object of understanding to depend on its object means that it depends on the ongoing obtaining of the state of affairs understood. In light of this, the meaning of the dependency principle that we extracted from Categories 7 can be further specified. If a person understands that S is P, then that person’s psychic condition can only count as understanding at those times when S is, in fact, P. If ever S is not P, the object of understanding will fail to ‘exist’, and, being a relative, that person’s condition will no longer count as one of understanding. 2.3.  Is the dependency principle specific to scientific understanding? Another question which will be important for comprehending Aristotle’s argument is the following: to what extent does the knowledge like harmony, medicine, geometry, and arithmetic in the Post. An. (1. 10, 76b16–25; 1. 13, 78b32–79a16). Thus, when Aristotle claims that the object of understanding is a necessity, he probably means that it is a predicatively structured entity which is a necessity, not that the object of understanding is a body of know­ledge or a theory which is in some sense ‘necessary’ (although, to be sure, it follows that the sentences which make up such a theory will express necessary truths). 59  For a careful attempt to do so, see P. Crivelli, Aristotle on Truth (Cambridge, 2004), 45–76. 60  This is not the same claim as that objects in this sense change. I am only claiming that Aristotle takes objects of understanding to belong to an ontological cat­ egory whose members do not constitutively hold or fail to hold once and for all. See further below and Crivelli, Aristotle on Truth, 183–97. 61 Cf. Metaph. Θ. 8, 1050a13–15.



The Necessity of What We Understand

189

dependency principle reflect a special feature of scientific understanding? Does Aristotle take something similar to hold of other types of knowledge and cognitive states? If the reading offered so far is correct, we should expect Aristotle to hold a version of the dependency principle for any factive mental state.62 Further, since Aristotle’s grounds for holding understanding to be a relative are fairly abstract (we say that understanding is understanding of something), we should expect him to be willing to classify any object-­directed cognitive state as a relative.63 Aristotle’s texts indicate that he would indeed be willing to extend this principle widely. In Categories 5, he holds that even true belief depends on its objects in the manner of a relative: [9] ὁ γὰρ αὐτὸς λόγος ἀληθής τε καὶ ψευδὴς εἶναι δοκεῖ, οἷον εἰ ἀληθὴς εἴη ὁ λόγος τὸ καθῆσθαί τινα, ἀναστάντος αὐτοῦ ὁ αὐτὸς οὗτος ψευδὴς ἔσται· ὡσαύτως δὲ καὶ ἐπὶ τῆς δόξης· εἰ γάρ τις ἀληθῶς δοξάζοι τὸ καθῆσθαί τινα, ἀναστάντος αὐτοῦ ψευδῶς δοξάσει τὴν αὐτὴν ἔχων περὶ αὐτοῦ δόξαν. (Cat. 5, 4a23–8) The same sentence seems to be both true and false, for example, if the sentence that someone is sitting is true, then the same sentence will be false when they get up. Likewise with beliefs: if someone truly believes that someone is sitting, they will have a false belief about them if they have not changed their mind when the person gets up.

Where a contemporary philosopher might take there to be a differ­ ent belief corresponding to the assertion that Socrates is seated now, when he is, and the later assertion that he is seated, when he is not,64 Aristotle prefers to think of there being a single belief that is made true when Socrates is sitting and false when he is not.65 This implies that one may acquire a false belief in one of two rather different ways. First, one may acquire a false belief by changing one’s mind about something true: if, at first, I truly believe that Socrates is sitting and then for whatever reason change my mind 62 Whether other cognitive states are simultaneous with their objects is, of course, a different question, and one I will not broach here. 63 As we have seen, Aristotle explicitly classifies perception (Cat. 2, 6b2) and thought (διάνοια, Metaph. Δ. 15, 1021a31) as relatives; he also lists sight (ὄψις, 1021a33–b1). 64 See the comparison of Aristotle with W.  V.  O.  Quine in Hintikka, ‘Time, Truth and Knowledge in Ancient Greek Philosophy’, 2. See also J.  Hintikka, ‘Time, Truth and Knowledge in Aristotle and Other Greek Philosophers’, Time and Necessity: Studies in Aristotle’s Theory of Modality (Oxford, 1973), 62–92. 65  Aristotle also emphasizes the sameness of the belief at Metaph. Θ. 10, 1051b13–14.

190

Joshua Mendelsohn

while Socrates stays in his seat, I thereby acquire a false belief and thus lose my true belief. Call this way of losing a true belief a ­‘primary loss’. I can also, on this view, lose a true belief in a rather different way. Even if I do not change my mind, Aristotle holds, the belief that Socrates is sitting will become a false belief if Socrates gets up. This is the way of losing a true belief and acquiring a false one that Aristotle is discussing in [9]. What is responsible for the acquisition of a false belief in this case is not explained by any psychological change in me, the believer (I have not reconsidered things, been persuaded otherwise, etc.). Instead, it is the fact that the truthmaker that previously secured the truth of my belief (the complex consisting of the subject Socrates and the quality of being seated) has ceased to exist.66 The believer and the belief undergo in this case a mere ‘Cambridge’ change,67 but the true belief is nevertheless lost on account of the same belief68 losing its object and thus becoming false. I will call this a ‘secondary loss’.69 Now, since belief is not factive, these sorts of changes do not, as in the case of understanding, bring about the loss of the belief itself: the belief simply goes from being a true one to a false one. On the other hand, [9] gives us reason to think that true belief depends on the ongoing existence of its object, where ‘existence’ amounts to the ongoing obtaining of a state of affairs, in just the same way as understanding. Should the relevant state of affairs cease to obtain, the psychic condition will, for that reason, cease to count as a condition of that kind (as a true belief or a condition of understanding), without any real change needing to occur in the knower or believer. In this respect understanding is on a par with true belief, and Aristotle ought to extend similar reasoning to any factive cognitive state. However, understanding is also crucially different from true belief, in a way that stands in tension with this requirement. Let us turn to this now. 66 Cf. DA 3. 3, 428b8–9 with Crivelli, Aristotle on Truth, 62–71. 67  Aristotle goes on below to explain that the reception of truth-­values at different times does not constitute a change in his strict sense, since only substances can undergo changes, strictly speaking: see Cat. 5, 4a28–b1. On this, see also Crivelli, Aristotle on Truth, 183–9. 68 Cf. Cat. 4a34–b2 in addition to [9]. 69  With the distinction between primary and secondary loss, compare also Post. An. 1. 33, 89a4–5 with Fine, ‘Aristotle’s Two Worlds’, 249.



The Necessity of What We Understand

191

3.  Understanding as a state The chapter following Aristotle’s discussion of understanding as a relative, Categories 8, places understanding in the category of qual­ ity. Aristotle distinguishes, as two species of quality, ‘state and con­ dition’ (ἕξις καὶ διάθεσις, 8b27), giving ‘instances of understanding and virtues’ (αἵ τε ἐπιστῆμαι καὶ αἱ ἀρεταί, 8b29) as examples of ‘states’. Hence, in addition to being categorized as a relative, understanding is also classified in the category of quality, in particular the type of quality Aristotle calls a hexis or ‘state’. There is evidence that this ‘doubling up’ on the category of understanding is conscious and deliberate on Aristotle’s part. Aristotle notes explicitly in Categories 7 that there are states in the category of relative,70 and given the preceding analysis of what it means to call understanding a relative, it should not surprise us to find that it belongs to the category of quality as well. Aristotle’s conception of quality is broad: a quality is anything that can be predicated of a subject ‘to say what sort of thing it is’.71 In this broad sense of ‘quality’,72 qualities with a relational component are unremarkable. To call a vehicle roadworthy, for instance, is at least in part to say that it is deemed acceptable for use on roads by some country’s road authority; hence, to say that it bears the relation of being officially approved for road travel to the relevant institution or officials within it. In general, there is no reason why saying that something stands in a certain relation cannot be a way of qualifying it, and hence no reason why a relative cannot also be a quality. There is thus no real tension between Aristotle’s claim that understanding is a relative and his claim that understanding is a quality,

70  Cat. 7, 6b2. We need not take Aristotle here to be claiming that all states, or all conditions, are relatives (for this reading, see O. Harari, ‘The Unity of Aristotle’s Category of Relatives’, Classical Quarterly, n. s., 61 (2011), 521–37). Another piece of evidence that Aristotle is aware and untroubled by understanding occupying two categories is Cat. 8, 11a37–8. I discuss this passage below as part of [11]. 71  Cat. 8, 8b25: ποιότητα δὲ λέγω καθ’ ἣν ποιοί τινες λέγονται. For an excellent treatment of Aristotle’s category of quality more broadly, see P. Studtmann, ‘Aristotle’s Category of Quality: A Regimented Interpretation’ [‘Aristotle’s Category of Quality’], Apeiron, 36 (2003), 205–27. 72  Aristotle discusses a narrower notion of quality at Cat. 8, 9b13–32, which comes close to what he elsewhere in Cat. 8 calls a hexis. On this, see D. S. Hutchinson, The Virtues of Aristotle (London, 1986), 14.

192

Joshua Mendelsohn

properly understood.73 A tension does, however, develop as Aristotle goes on to explain what it means for understanding to be the specific type of quality he calls a hexis or ‘state’.74 Aristotle distinguishes states from more superficial qualities of a subject which he calls ‘conditions’ (διαθέσεις).75 These, like a person’s blushing or being angry, tend to be short-­lived and are easily gained and lost without other significant changes in their bearers. A state differs in that it is more ‘stable’ (μονιμώτερον) and ‘long-­lasting’ (πολυχρονιώτερον, Cat. 8, 8b28). Aristotle goes on to explain why understanding is a state rather than a condition: [10] ἥ τε γὰρ ἐπιστήμη δοκεῖ τῶν παραμονίμων εἶναι καὶ δυσκινήτων, ἐὰν καὶ μετρίως τις ἐπιστήμην λάβῃ, ἐάνπερ μὴ μεγάλη μεταβολὴ γένηται ὑπὸ νόσου ἢ ἄλλου τινὸς τοιούτου. (Cat. 8, 8b29–32). Understanding seems to be something very abiding and steady whenever someone has even a moderate grasp of their understanding, so long as no great change comes about by illness or something else of this sort.

Aristotle takes understanding to be a state rather than a condition, because it is ‘abiding’ (παραμόνιμος) and ‘steady’ (δυσκίνητος).76 These terms, while no doubt intended to align with the terms ‘stable’ (μονιμώτερον) and ‘long-­lasting’ (πολυχρονιώτερον), are not simply synonyms for them. This language, especially the term ‘abiding’, echoes the terminology and imagery used to describe the value of epistēmē as compared with doxa in the Meno.77 This suggests that they are intended as normative descriptions of understanding. They 73  This may come as a surprise to readers who view the categories as an exclusive taxonomy. For a persuasive case that the textual evidence does not support taking Aristotle’s categories to be an exclusive taxonomy, see D.  Morrison, ‘The Taxonomical Interpretation of Aristotle’s Categories’, in A. Preus and J. P. Anton (eds.), Essays in Ancient Greek Philosophy, vol. v: Aristotle’s Ontology (New York, 1992), 19–46. 74  With this term I refer specifically to the notion of hexis Aristotle develops in Categories 8. A different notion may be at play in the Metaphysics: see Hutchinson, The Virtues of Aristotle, 8–20. 75  Aristotle uses the term διάθεσις in a number of other ways in Categories 8; for discussion see Studtmann, ‘Aristotle’s Category of Quality’. 76  Cat. 8, 8b30. Cf. 9a5, 9a9–10, 8b34–7. 77 Compare παραμονίμων at Cat. 8, 8b30 with forms of παραμενεῖν at Meno 97 d 10, 97 e 4, 97 e 7, and 98 a 1–2, and compare δυσκίνητος at Cat. 8, 8b30 with the tether imagery, contrasted with the imagery of something running away. μονιμώτερον at Cat. 8, 8b28 also parallels μόνιμοι at Meno 98 a 6. With πολυχρονιώτερον, compare Meno 97 e 7 and especially 98 a 1.



The Necessity of What We Understand

193

explain why it meets the definition of a hexis and, at the same time, in what way this renders it valuable. Aristotle thinks that understanding is a valuable quality in part because it is the kind of thing that sticks by you and is available when you need it, much like, in Plato’s metaphor, a slave or living statue that isn’t liable to run away.78 Now, Aristotle thinks that understanding can only have this valu­able feature if it is a fundamentally different kind of quality from, for example, being hot or cold, qualities which someone might gain or lose by stepping outside.79 So long as someone has a ‘moderate’ handle on their own understanding,80 it is liable to loss only in the face of more radical changes. Aristotle does not specify the class of things that he takes to be  capable of erasing understanding, but rather only gives one ex­ample, illness. The example is, however, telling. The pertinent feature of illness in this context cannot be that it is acquired from without, since then Aristotle’s claim that we only lose understanding through something like illness would rule out loss of understanding by mental deterioration naturally occurring in old age, which Aristotle does appear to recognize.81 Instead, the point of adverting 78  Meno 97 e 7. Cf. Bolton, ‘Science and Scientific Inquiry’, 49. 79  Cat. 8, 8b34–7. Cf. Hutchinson, The Virtues of Aristotle, 19–20, who argues that the core idea here is that of being well entrenched and resistant to change, whereas the longevity of understanding serves as evidence for this. I think Hutchinson is right about this, but I would add, following Bolton, that the well-­ entrenchedness and resistance to change are in turn grounded in the conception of understanding as a reliable state. I would also add that being well entrenched is not the only source of understanding’s stability here: the character of its object is, as we will see, equally important. 80  It is not entirely clear what Aristotle means by ‘moderate grasp of one’s understanding’ (μετρίως τις ἐπιστήμην λάβῃ; the phrase could also be translated as ‘a moderate grasp of a science’). See, however, NE 7. 3, 1147a20–2, where Aristotle is committed to the view that understanding can be defective if it has only been acquired recently. I take it he has in mind the sort of shaky ‘understanding’ someone might have by, e.g., reading a physics textbook once without doing the exercises, as compared with the sort of understanding a student acquires who has pored over the same textbook in a physics class. Aristotle’s claim may be that understanding becomes ‘steady’ in the manner under discussion here so long as it is subject to sufficient reflection, inculcation, or drill (cf. Meno 85 c 9–d 1). Interestingly, if this is correct, then Aristotle’s claim concerns not only expert understanding, since experts presumably have more than just a ‘moderate’ grasp of their understanding. 81 See Mem. 1, 450b1–8; also De long. vit. 2, 465a19–23 and GA 784b30–2. At DA 1. 4, 408b19–28, Aristotle claims that cognitive decline in old age comes about as a result of the destruction of organs required for reasoning and contemplating, not as a result of the deterioration of reasoning or contemplating itself, which he takes to

194

Joshua Mendelsohn

to illness as a paradigm of the sort of change that could cause one to lose understanding is presumably to emphasize that such a change cannot come about via any bodily change that is compatible with the animal’s normal, healthy functioning. While intoxication and passions might inhibit the exercise of understanding,82 only a change that causes some sort of harm or represents some sort of degeneration can bring about a loss of understanding.83 Categories 8 thus yields what I will call a durability principle: If  S understands O, then S continues to understand O so long as she ­experiences no detrimental changes to the constitution of her cognitive faculties.

3.1.  The tension between the two principles At this point we may begin to perceive a tension with Aristotle’s claim in Categories 7 that understanding is a relative. That discussion gave the impression that, far from being a secure possession, understanding is altogether precarious: one counts as having understanding only when, in addition to one’s faculties being in order, the object of one’s understanding continues for its part to be such as one understands it to be. How is the requirement that understanding be steady and abiding in such a way as to generally preclude being lost related to the claim that understanding depends, like true belief, on the continued existence of its object? The point can be focused by noting that the durability principle rules out loss of understanding corresponding to both the primary and secondary loss of true belief. On the one hand, the durability principle expresses the fact that understanding, in the sense at issue when Aristotle classifies it as a state, is deeply ingrained within be unaffected. It does not follow that Aristotle takes all mental deterioration to result from an external source: the required cognitive machinery might be such as to decline by its own nature. 82 See NE 7. 3, esp. 1147a10–24 and 1147b8–17. I discuss this passage below in Section 4. 83  We might today think of a brain injury or a degenerative condition. Cf. NE 3. 5, 1114a25–8; 7. 1, 1145a31; and 7. 5, 1149a4–12. R.  Bodéüs, Aristote: Catégories [Catégories] (Paris, 2001), 133, taking epistēmē here broadly to include also practical knowledge, suggests that Aristotle might have added ‘bestial’ (θηριώδη) affections, which he discusses as an impediment to knowledge at Nicomachean Ethics 7. 5, 1149a6–8. But in that passage, bestial affections are not described as leading to loss of knowledge; rather, Aristotle invokes bestial affections there to explain why some people never acquire certain types of knowledge to begin with.



The Necessity of What We Understand

195

one’s web of beliefs, such that one is ‘incapable of being persuaded otherwise’ (ἀμετάπειστον, Post. An. 1. 2, 72b3–4) regarding what one understands. One cannot, in other words, lose a piece of understanding by being persuaded that what one understands is false. Yet it is important to see that this is not the only type of loss of understanding Aristotle rules out when he claims that an ‘illness or something else of this sort’ (Cat. 8, 8b32 [10]) would be required to erase understanding. For another event that surely would not count as an illness or something of that sort would be a change in the object of understanding leading to the loss of scientific understanding. That is, if it were possible that someone could at one point in time have scientific understanding and at another point fail to have scientific understanding, where the only difference between these two times was that the object of that person’s scientific understanding had undergone some change incompatible with their continuing to understand it, that would be incompatible with their scientific understanding being stable in the requisite sense. As Aristotle puts it in Metaphysics Ζ. 15, ‘understanding cannot sometimes be under­standing and sometimes be ignorance; rather, it is opinion that is like this’.84 I take Aristotle to mean that whatever cognitive state accounts for our understanding, this cannot be the type of state that sometimes counts as understanding and sometimes counts as mere igno­ rance.85 If there are certain intrinsic features of my psyche that are at one time sufficient for me counting as having scientific understanding, then they must always suffice for understanding, at least until death or some cognitive misfortune befalls me. To put it differently, according to the durability principle, understanding is not like true belief, where a mere change in the fact of Socrates sitting might erase one’s mental state. And yet, insofar as understanding is a relative, it is just like true belief in depending on its object. This is a tension but not, I will maintain, a contradiction. In the  remainder of this paper, I will argue that the claim that 84  Metaph. Ζ. 15, 1039b32–4: οὐδ’ ἐπιστήμην ὁτὲ μὲν ἐπιστήμην ὁτὲ δ’ ἄγνοιαν εἶναι, ἀλλὰ δόξα τὸ τοιοῦτόν ἐστιν. Cf. Top. 5. 3, 131b21–3. 85  Aristotle’s view is most straightforward if we suppose that he takes understanding to be a species of belief. This is how he is read by Moss and Schwab, ‘The Birth of Belief’ and Fine, ‘Aristotle on Knowledge’. We can then say that a primary loss of understanding occurs when a primary loss of the relevant true belief occurs, and a secondary loss of understanding occurs when a secondary loss of the relevant true belief occurs. I will not, however, presuppose any view on whether Aristotle thinks that understanding is a type of belief in this paper.

196

Joshua Mendelsohn

­ nderstanding is only of necessities serves to resolve this tension u and render consistent the claim that understanding is a relative and that it is a state. First, however, it is necessary to explore another way that someone might take Aristotle to resolve this tension, and to show why this solution will not work. Doing so will serve to illustrate the depth of the problem generated by the dependency and durability principles and point the way to an alternative resolution which, I will argue, is the one Aristotle actually pursues in Nicomachean Ethics 6. 3.

4.  A tempting solution One might try to alleviate the tension between the dependency and durability of understanding by pointing out that Aristotle treats understanding as an intellectual virtue (NE 6. 3, 1139b16–17).86 In the sense in which it denotes a virtue, understanding is acquired only with significant expenditure of time and effort, since it requires a deep assimilation of specialized knowledge (Phys. 7. 3, 247b17– 18; NE 2. 1, 1103a15–17; 7. 3, 1147a22). Once it is acquired, the possesssor then has distinctive scientific abilities, depending on what specific type of understanding is acquired. These include the ability to construct demonstrations (6. 3, 1139b31) to teach the rele­ vant science (1139b25) and, more generally, the abilities associated with mastery of a complex network of explanatory connections pertaining to a particular scientific domain.87 86  Aristotle does not explicitly call epistēmē a virtue there, but he is taken this way by D.  Bronstein, ‘Aristotle’s Virtue Epistemology’, in S.  Hetherington and N.  D.  Smith (eds.), What the Ancients Offer to Contemporary Epistemology (New York, 2020), 165, and Bolton, ‘Science and Scientific Inquiry’, 53. R. A. Gauthier and J.  Y.  Jolif, L’éthique à Nicomaque: Introduction, traduction et commentaire [L’éthique à Nicomaque], 2 vols. (Leuven, 1959), ii. 163 stresses that Aristotle would deny that epistēmē is the virtue of the ἐπιστημονικόν, since he takes the more complete virtue of σοϕία to be the best condition of this part of the soul. I agree, but we can distinguish between the state that is the best condition of the ἐπιστημονικόν (the virtue of this part of the soul), and the states that represent the various ways it may be excellent (the virtues of this part of the soul). This observation does not, therefore, deprive epistēmē of its status as an excellence of this part of the soul. Thanks to Mike Coxhead for discussion on this point. 87  I am not maintaining that the object of the state is only ever a whole science. The object of understanding can be a single proposition or state of affairs, but it needs to be understood in an appropriate explanatory context, and in certain cases



The Necessity of What We Understand

197

Like other virtues, understanding may be inhibited or prevented from being manifested properly. Revellers at a symposium might, say, drink so much as to be unable to piece together an explanation and thus in this sense fail to have understanding at a certain time. Aristotle will deny that the cognitive virtue the revellers have is at any point lost when this occurs (Phys. 7. 3, 247b13–16). Rather, as in the practical case, the possession of the virtue is not enough to ensure that one always exercises it when it is called for.88 That the drunk botanist cannot explain why broad-­leaved plants shed their leaves would mean, on Aristotle’s analysis, that she is too drunk to ‘employ’ or ‘make use of’ (χρῆσθαι, Phys. 7. 3, 247b16; cf. NE 7. 3, 1147a12) the understanding that she has. And so, while he will grant that there is a sense in which such a person does not at that time understand,89 he also maintains that there is another sense in which the person still has understanding but fails to make use of it. Now Aristotle surely holds that it is understanding in the former sense, which is not jeopardized by a lapse in memory or a tem­por­ ary inhibition, that is steady and abiding in the sense at issue in Categories 8. His point is that only a more severe sort of memory loss, one which comes from years of letting one’s understanding languish, or a repeated and persistent impediment—­alcoholism, perhaps—­could cause loss of understanding in this sense.90 While our grasp of one part of the scientific edifice we comprehend in possessing a virtue might be easily lost, our grasp of the edifice as a whole cannot be. Someone who did suddenly lose their grasp of the edifice, without any mitigating factors, would thereby have been shown not to have had a reasonable grasp of what they under-

this explanatory context may include most or all of the science. On this, see further Lesher, ‘Ἐπιστήμη as “Understanding” ’. 88 Compare NE 7. 3, 1147a13–14. For some reasons to think that Aristotle’s points there are not restricted to practical knowledge, see B. Morison, ‘Colloquium 2: An Aristotelian Distinction between Two Types of Knowledge’ [‘Two Types of Knowledge’], Proceedings of the Boston Area Colloquium of Ancient Philosophy, 27 (2012), 29–63. 89 Cf. NE 7. 3, 1147a13: ‘in a way [the drunk person] has and does not have [understanding]’ (ἔχειν πως καὶ μὴ ἔχειν). 90 Cf. Cat. 8, 9a1–3, where Aristotle claims that a quality which typically represents a temporary condition may become a permanent state if it is had for a long enough time. Aristotle may be thinking of the way that a person who is not just temporarily sick but constantly falling ill could be said to have a ‘sickly disposition’. For this reading, see Kiefer, Aristotle’s Theory of Knowledge, 19.

198

Joshua Mendelsohn

stood91 to begin with. Aristotle, therefore, does not mean to deny that someone might over a brief period learn or forget a particular fact at the periphery of their web of knowledge. The durability principle only denies that such a change constitutes a change to the person’s understanding in the sense of the possession of an intellectual virtue. That may seem enough to resolve the tension between Aristotle’s claims that understanding is a relative and that it is an abiding structural quality (a ‘state’). For Aristotle distinguishes two senses of epistēmē, the word I have been translating as ‘understanding’, one denoting the hexis in virtue of which one is able to exercise certain cognitive capacities, and another denoting the condition one is in when everything is in place to exercise them.92 What abides and resists change is, for Aristotle, only understanding in the former sense. It may thus seem natural to suppose, conversely, that it is only understanding in the latter sense that has an object and thus is a relative. If this were correct, then the problem we have been dealing with would turn out to have been merely lexical: it is only because the word ‘understanding’ is used in one sense to denote a state of capacity and in another sense for its deployment that we end up listing ‘epistēmē’ in two categories. That would be no more problematic than the fact that we use ‘healthy’ to describe both a condition of the body and the things conducive to that condition.93 No more than in this case, so this response goes, should one expect the properties of ‘understanding’ in the respective categories to be consistent. A response of this sort is encouraged by the traditional (but, I will argue, incorrect) reading of a passage at the end of Categories 8, in which Aristotle explains why ‘we should not be disturbed lest 91  This is another way we might gloss μετρίως . . . ἐπιστήμην λάβῃ (Cat. 8, 8b31). 92  Whether or not we identify this hexis with knowledge in second potentiality as described at DA 2. 5, 417b27–8 (an issue on which I will take no stand here), this distinction should not be conflated with the distinction between first and second potentiality or the distinction between first and second actuality. One can be in a state where one is free to exercise scientific abilities without actually exercising them, so this type of epistēmē is not the same as epistēmē in second actuality (cf. 417a28–9). But if this type of epistēmē is epistēmē in first actuality or second potentiality, then that would leave epistēmē as a hexis in the role of first potentiality, which is clearly not right: epistēmē as a virtue is already an acquired intellectual achievement, whereas knowledge in first potentiality is not (cf. 417a27, 417b31–2). 93 Cf. Top. 1. 15, 107b6–12.



The Necessity of What We Understand

199

someone should say that though we proposed to discuss quality, we are counting in many relatives (since states and conditions are relatives)’.94 He says: [11] σχεδὸν γὰρ ἐπὶ πάντων τῶν τοιούτων τὰ γένη πρός τι λέγεται, τῶν δὲ καθ’ ἕκαστα οὐδέν· ἡ μὲν γὰρ ἐπιστήμη, γένος οὖσα, αὐτὸ ὅπερ ἐστὶν ἑτέρου λέγεται—τινὸς γὰρ ἐπιστήμη λέγεται—τῶν δὲ καθ’ ἕκαστα οὐδὲν αὐτὸ ὅπερ ἐστὶν ἑτέρου λέγεται, οἷον ἡ γραμματικὴ οὐ λέγεται τινὸς γραμματικὴ οὐδ’ ἡ μουσικὴ τινὸς μουσική, ἀλλ’ εἰ ἄρα κατὰ τὸ γένος καὶ αὗται πρός τι λέγεται· οἷον ἡ γραμματικὴ λέγεται τινὸς ἐπιστήμη, οὐ τινὸς γραμματική, καὶ ἡ μουσικὴ τινὸς ἐπιστήμη, οὐ τινὸς μουσική· ὥστε αἱ καθ’ ἕκαστα οὐκ εἰσὶ τῶν πρός τι. λεγόμεθα δὲ ποιοὶ ταῖς καθ’ ἕκαστα· ταύτας γὰρ καὶ ἔχομεν—ἐπιστήμονες γὰρ λεγόμεθα τῷ ἔχειν τῶν καθ’ ἕκαστα ἐπιστημῶν τινά—­ὥστε αὗται ἂν καὶ ποιότητες εἴησαν αἱ καθ’ ἕκαστα, καθ’ ἅς ποτε καὶ ποιοὶ λεγόμεθα· αὗται δὲ οὐκ εἰσὶ τῶν πρός τι. (Cat. 8, 11a23–36) For in almost all of these cases, the genus is said to be a relative, but none of the specific types is. For understanding, a genus, is called just what it is, of something else (it is called understanding of something); but none of the specific types is called just what it is, of something else. For example, grammar is not said to be grammar of something, nor music, music of something. Thus, the specific types are not relatives. But we are said to be qualified with the specific types, since we have them (it is because we have some particular type of understanding that we are said to understand). Hence these—­the specific types, in virtue of which we are on occasion said to be qualified—­would indeed be qualities; but these are not relatives.

Aristotle makes a distinction here between the categorial status of specific types of understanding like music and grammar and understanding as a ‘genus’ (γένος), that is, understanding as the kind encompassing all of these specific types of understanding.95 94  Cat. 8, 11a20–3: οὐ δεῖ δὲ ταράττεσθαι μή τις ἡμᾶς ϕήσῃ ὑπὲρ ποιότητος τὴν πρόθεσιν ποιησαμένους πολλὰ τῶν πρός τι συγκαταριθμεῖσθαι· τὰς γὰρ ἕξεις καὶ τὰς διαθέσεις τῶν πρός τι εἶναι. 95  Minio-­Paluello, Aristotelis Categoriae et liber De Interpretatione, Praefatio, n. 1, and M. Frede, ‘The Title, Unity and Authenticity of Aristotle’s Categories’ [‘Title, Unity and Authenticity’], in id., Essays in Ancient Philosophy (Minneapolis, 1987), 11–28 at 13, hold lines 11b10–16, which immediately precede this passage, to be suspect, and Bodéüs, Catégories, 50, transposes these lines to just after 11a38. M. Frede, ‘Title, Unity and Authenticity’, 13–17, argues for extending suspicion to a passage including [11] (specifically, to 11a20–38), on grounds of style and content. I will make no attempt to address stylistic issues here. However, the reasons that Frede gives for doubting the authenticity of this passage on the basis of the doctrine it espouses are not convincing. Frede notes that, in a different passage (Cat. 7, 8a13–8b24), Aristotle goes to pains to avoid the conclusion that the same item is a

200

Joshua Mendelsohn

He points out that the peculiarity of being in two categories does not apply to music, grammar, etc. This is because these are qual­ ities alone, not relatives. To be schooled in (or ‘have’) grammar or music is not to have grammar or music ‘of’ or ‘than’ something else. We may be tempted to infer from this that Aristotle holds that understanding as a genus is, conversely, not a state. This would then give him a tidy solution to the puzzle: all understanding is either understanding as a genus or one of its species, and the genus is only a relative (not a quality), while the species are only qualities (not relatives).96 It would be from there a small step to attribute to Aristotle the analogous claim that understanding as a cognitive virtue is a state only (and not a relative), whereas the sense in which understanding is a relative refers only to understanding in the sense of the exercise of our capacity to understand. He would then avoid the claim that understanding in the very same sense is both a state and a quality. The problem for this response is that Aristotle does not deny that understanding as a genus is a quality in [11]. He only affirms that it is a relative and denies that its species are relatives. In fact, relative and a substance, and so finds it surprising that Aristotle should be willing to allow the same item to be both a relative and a quality here. There is, however, an independent reason for Aristotle to wish to avoid the conclusion that relatives are substances: relatives are posterior in nature to beings in the other non-­substantial categories (Metaph. Ν. 1, 1088a24), while substances are prior to them in nature (NE 1. 6, 1096a21). Hence, we need not take Aristotle’s desire to avoid the conclusion that some substances are relatives as the outcome of a general aversion on his part to assigning the same item to multiple categories. Rather, Aristotle may hold this view so as to avoid violating the antisymmetry of priority in nature. Frede’s other reason for taking the content of this passage to be at odds with Aristotelian doctrine relies on the categories being interpreted as highest genera, and as such being mutually exclusive. But as Frede himself notes (13), the categories are only described as highest genera in the Categories at 11b15, in a part of the text generally agreed to be suspect. On this, see further Bodéüs, Catégories, 141, and Morrison, ‘The Taxonomical Interpretation of Aristotle’s Categories’. My approach here will be to proceed under the assumption that 11a20–38 is authentic and to argue that this passage is consistent with my reading. The passage is not, however, required to establish my claim that Aristotle takes understanding to be both a quality and a relative: this is already claimed at 6b3 and 8b29. Since, however, I suspect the passage is authentic, I will explain how, properly understood, it is consistent with my reading and may be taken to provide further details of Aristotle’s position on the categorial status of understanding. 96 Elias endorses this interpretation very explicitly; see Elias, In Cat. 238. 8–10 Busse, with H. Taieb, ‘Classifying Knowledge and Cognates: On Aristotle’s Categories, 8, 11a20–38 and its Early Reception’ [‘Classifying Knowledge and Cognates’], Documenti e studi sulla tradizione filosofica medievale, 27 (2016), 85–106 at 98–9.



The Necessity of What We Understand

201

his criterion for being a quality requires that he include in this ­cat­egory understanding as a genus. For the criterion that Aristotle uses to argue that the particular types of understanding are qual­ ities is that we qualify people with them (Cat. 8, 11a32–6): We say someone is ‘musical’ (μουσικός) or ‘literate’ (γραμματικός). We also qualify people with understanding generally; indeed, ‘understanding’ (ἐπιστήμη) is Aristotle’s example of something ‘in’ but not ‘said of’ a subject (2, 1b1).97 The fact that such an attribution stands in need of semantic supplementation by a correlative does not imply that it is not used to qualify people, and hence does not imply that it is not a quality. This is one reason to reject the reading that takes Aristotle to be solving the puzzle proposed in [11] by denying that understanding as a genus is a quality. Another reason to reject this reading is that it makes it very difficult to understand what Aristotle says next. He goes on: ‘Moreover, if the same thing really is a quality and a rela­tive, there is nothing absurd in its being counted in both the genera’.98 This makes little sense if Aristotle has just been arguing that nothing is really both a quality and a relative. Why go to the trouble of providing that argument if there is anyway nothing absurd in something occupying both categories?99 What this remark makes clear is, rather, that Aristotle’s purpose in [11] is more modest: he only endeavours to clarify which items it is that belong to both categories, not to argue that there aren’t any. His point is only that such cases are less pervasive than we might first have thought, since the species of understanding are only in the category of

97 Cf. Cat. 8, 10b2. Porphyry (In Cat. 140. 20 Busse) denies that the term ‘understanding’ is ever used to qualify someone with a particular type of understanding like grammar or music, but he does not justify his claim. Olympiodorus (In Cat. 129. 28 Busse) attempts to defend this claim by asserting that it is impossible for any one person to know everything, but this is clearly beside the point. When we say that someone has ‘understanding’ without further specification, we are not saying this person knows everything. On this, see further Taieb, ‘Classifying Knowledge and Cognates’, 97. 98  ἔτι εἰ τυγχάνει τὸ αὐτὸ ποιὸν καὶ πρός τι ὄν, οὐδὲν ἄτοπον ἐν ἀμϕοτέροις τοῖς γένεσιν αὐτὸ καταριθμεῖσθαι (Cat. 8, 11a37–8). 99  As Porphyry reads him (In Cat 140. 24–141. 5 Busse), Aristotle is offering an alternative, incompatible solution to the puzzle at 11a37–8, but it is hard to see a further solution in Aristotle’s flat assertion that there is nothing absurd in the same thing being counted in both genera. See further Taieb, ‘Classifying Knowledge and Cognates’, 96–100 on Porphyry’s interpretation and its problems.

202

Joshua Mendelsohn

­ uality.100 Understanding as a genus, however, still occupies the cat­ q egory of relative as well as being a certain kind of quality (a state). This means that even if we restrict our attention to understanding as a genus, we still face the problem discussed in Section 3.1: since it is a state, understanding in this sense ought to be capable of being lost only with harm to its possessor, but as a relative, this state of understanding ought to be liable to expire on account of its object, without any intrinsic change in its possessor. The same is true of understanding in the sense of an intellectual virtue: as a virtue, it is a state and therefore stable in the way that a state is required to be. It is, however, the type of virtue that relies on an appropriate relation to something external to the knower, and thus also a relative.101 There is thus a real and not merely lexical tension between the characterizations of understanding that emerge from Cat. 7 and 8 respectively. Both of these are motivated by plausible intuitions about scientific understanding. On the one hand, Aristotle wishes to pay heed to the fact that we regularly take our understanding to be stable in a way that it only could be if we did not have to reckon with our understanding changing on account of factors outside us. On the other hand, Aristotle takes the grammar of epistēmē at face value, as reflective of a metaphysical reality in which understanding, even in the statal sense, is essentially of something. His 100 Simplicius has a similar view. As he interprets the text, ‘Aristotle did not mean that the genera were not qualities’ (οὐκ εἶπεν τὰ γένη μὴ εἶναι ποιότητας); in­stead, Aristotle thinks that ‘even if [the] state and condition [of understanding] are said to be relative, this is not true of all states and conditions, but only the generic’ (εἰ καὶ εἴρηται πρός τι ἡ ἕξις καὶ ἡ διάθεσις, οὐ πᾶσα ἔχει τοῦτο, ἀλλ’ ἡ γενικὴ μόνον, Simplicius In Cat. 293. 22–5 Kalbfleisch, trans. Fleet modified). This interpretation also allows us to address another point that leads Frede to doubt the authenticity of 11a20–38. Frede, ‘Title, Unity and Authenticity’, 13, complains that 11a37–8 ‘contributes nothing to solving the difficulty raised in 11a20–22’. I agree with this judgement, but on my reading the sentence nevertheless has a clear function. Aristotle does not accept that there is a problem with certain items falling in both the cat­egory of quality and the category of relative; he only addresses the problem that ‘many’ (πολλά, 11a21) relatives end up in the category of quality. The function of the sentence at 11a37–8 is to remind the reader that nothing about the notion of quality or relative requires the two categories to be fully disjoint. Its function is thus to clarify what has and what has not been shown in the preceding lines. 101  See also Phys. 7. 3, 247b2–3. The context there makes it clear that Aristotle means to include epistēmē in the sense of an intellectual virtue (see 247b9–10, 246b20–247a2). For some reasons to think the doctrine of this text is not so at odds with Aristotle’s discussions elsewhere as some have thought, see Harari, ‘The Unity of Aristotle’s Category of Relatives’.



The Necessity of What We Understand

203

metaphysics of relatives requires that the perishing of the object of understanding would bring in its wake the perishing of any understanding of it. This way of putting things, however, suggests a solution to the dilemma. For—­and this is the key observation—­nothing about the considerations motivating the dependency principle requires Aristotle to hold that the object of understanding ever does actually perish. His remarks in Categories 7 about what is entailed by the object of understanding perishing need not be taken to show that he takes this to be possible; they may be taken to have the character of a per impossibile thought experiment designed to illustrate how understanding depends on its object by having us consider the consequences of a scenario that could never actually occur. Now, if it is not possible for the object of understanding ever to actually perish, then the core claims of Categories 7 and 8 regarding understanding will fall short of a contradiction, although the logical space between these claims is narrow indeed. Dependency requires only that, were it possible for the object of understanding to perish, understanding would perish with it. It entails no commitment to the object of understanding actually being capable of perishing. If it is not, then the dependency of understanding on its object will be compatible with the claim that scientific understanding is never such as to be lost on account of changes in the world. Thus, if the object of understanding cannot, in fact, perish, then the dependency principle and the durability principle generate no contradiction. Seeing this is key to understanding Aristotle’s argument for the necessity of the object of understanding in Nicomachean Ethics 6. 3. Let us turn to this now.

5.  Aristotle’s argument for the necessity of what we understand Nicomachean Ethics 6 discusses the intellectual virtues, among which Aristotle counts scientific understanding,102 placing it alongside craft (technē ), practical wisdom (phronēsis), theoretical wisdom (sophia), and insight (nous). He announces his intention to specify what understanding is, ‘if one is to be precise about the matter’ (εἰ δεῖ ἀκριβολογεῖσθαι, NE 6. 3, 1139b18–19). He starts, however, by 102  See n. 86 above.

204

Joshua Mendelsohn

discussing the character of ‘what we understand’ (ὅ ἐπιστάμεθα, 1139b20), that is, of the object of understanding. It is here, I claim, that Aristotle gives an argument for the necessity of the object of understanding, drawing on the durability and dependency of understanding as tacit premisses. Here again is the key passage: [12] (i) πάντες γὰρ ὑπολαμβάνομεν, ὃ ἐπιστάμεθα, μηδ’ ἐνδέχεσθαι ἄλλως ἔχειν· (ii) τὰ δ’ ἐνδεχόμενα ἄλλως, ὅταν ἔξω τοῦ θεωρεῖν γένηται, λανθάνει εἰ ἔστιν ἢ μή. (iii) ἐξ ἀνάγκης ἄρα ἐστὶ τὸ ἐπιστητόν. (NE 6. 3, 1139b19–23) (i) We all think that what we understand cannot be otherwise. (ii) With what can be otherwise, we are not aware whether it is so or not whenever it goes out of view (ἔξω τοῦ θεωρεῖν). (iii) Therefore, the object of understanding is of necessity.103

Sentence (i) states the conclusion Aristotle intends to establish in this passage: ‘What we understand cannot be otherwise’. As in other passages where he makes this claim, Aristotle notes the widespread acceptance of this claim, but whereas elsewhere he seems content to rely on consensus, here he presents an argument,104 albeit a highly compressed one. Sentence (ii) gives the only explicit premiss. In sentence (iii) he proceeds without further ado to draw the conclusion stated in sentence (i), rephrasing it as the claim that what we know is ‘of necessity’ (ἐξ ἀνάγκης, 1139b22). Three other passages parallel the language of Nicomachean Ethics 6. 3, 1139b19–23, but none of them is arguing for precisely this claim. The first is Posterior Analytics 1. 6, 74b32–6, where Aristotle argues that the middle term of a demonstration that provides epistēmē cannot ‘perish’ (ϕθαρείη), on pain of the demonstration ceasing to impart understanding. I discuss this passage below. The second passage is Metaphysics Ζ. 15, 1040a2–5, where Aristotle argues that there are no definitions of individual perceptible substances or demonstrations of facts about them. He argues that because any purported definition of an individual perceptible substance would not be necessary, no definition of an individual perceptible substance can be an object of understanding. This argument too relies on the claim that what we understand is necessary, using it to establish further conclusions. Finally, there is Topics 5. 3, 131b21–3, which presents an argument for the conclusion that 103  The division into sentences (i), (ii), and (iii) is for ease of exposition. 104  Note the ἄρα at 1139b23.



The Necessity of What We Understand

205

no object of perception is, in a certain technical sense, ‘properly assigned’ (καλῶς κείμενον) to a subject. This argument draws on similar considerations and will be relevant in analysing the argument here, but it does not aim to establish a conclusion about our knowledge or understanding directly; its conclusion is about the status of a certain type of predication in dialectic.105 It will thus pay to analyse [12] closely. The only explicit premiss of the argument states what might seem a queer claim about contingencies: if something is a contingency, then we are not aware whether it is so or not (λανθάνει εἰ ἔστιν ἢ μή) whenever it goes out of view (ἔξω τοῦ θεωρεῖν). The first question to be settled concerns the use of ἔστιν here. Given that Aristotle’s conclusion concerns our knowledge of necessities, and these elsewhere must be taken to be necessary facts or states of affairs, I will take the sense of ἔστιν in the premiss to be veridical. This reading is warranted by the context, since Aristotle indicates that he intends to be talking about the same notion of scientific knowledge that he discusses in the Analytics (NE 6. 3, 1139b32) and, as I argued above, the objects of scientific knowledge at least include states of affairs or propositions there. My central justification for this reading, however, is that it allows us to make good sense of Aristotle’s argument, as I will endeavour to show. Supposing, then, that ἔστιν is to be taken veridically, [12] (ii) comes to the following: If p is a contingent state of affairs, then we are not aware whether it is the case that p when p is out of view (ἔξω τοῦ θεωρεῖν)

Equivalently, replacing the internal conditional with its contra­ posi­tive: If p is a contingent state of affairs, then: when we are aware whether it is the case that p, p is not out of view (ἔξω τοῦ θεωρεῖν)

We should note an important fact that this paraphrase reveals. Aristotle is not denying knowledge of contingencies outright. He is 105  Metaph. Δ. 5, 1015b6–9 does contain an argument for the claim that what we understand is a necessity, but this argument relies on the premiss that what we understand is the conclusion of a demonstration from necessary principles, a claim which Aristotle attempts to establish in turn from the premiss that what we understand is necessary in Post. An. 1. 6, as we have seen (see the discussion of text [3] in Section 1 above).

206

Joshua Mendelsohn

only making a claim about the conditions under which such know­ ledge could occur. His claim in [12] (ii) presupposes that know­ledge of contingencies would require conscious awareness of them (the negation of λανθάνει εἰ ἔστιν ἢ μή) and that such awareness, in turn, would require them to be in some sense ‘theorized’ or ‘in view’. Below I will argue that Aristotle has reason to think the possession of scientific understanding must be possible even when not ‘in view’ in the relevant sense, and so scientific understanding cannot be of a contingency. It is, however, open to Aristotle to maintain that some less demanding type of knowledge, or even a type of know­ ledge that is equally demanding but not in such a way as to rule out these conditions, is of contingencies. In motto form, then, (ii) of [12] says that contingencies are ‘known only when theorized’ (KOWT, where ‘theorize’ is intended as a placeholder transliteration for θεωρεῖν, which I have so far rendered prejudicially). What does Aristotle mean by this, and how does this support his conclusion that understanding is of ne­ces­sities? Let us consider these questions in turn. 5.1.  Why does Aristotle hold that contingencies are KOWT? Aristotle sometimes uses θεωρεῖν to mean ‘observe’,106 and this meaning is intelligible even if the relevant objects are contingent states of affairs. I might, for instance, be said to observe the contingency that Socrates is sitting when I look at Socrates in a seated position and recognize that he is in such a position. One possibility, then, is that Aristotle is talking about what happens when a contingency ceases to be observed. His claim is that when we cease to observe a contingent state of affairs, we cease to be aware of it, and hence cease to know it. A reading like this is encouraged by many translations.107 It is also supported by a parallel passage in Topics 5. 3, where Aristotle writes: 106  See the references in H. Bonitz, Index Aristotelicus (Berlin, 1870), 328a4–40. 107  ‘outside our view’ (Broadie and Rowe, Nicomachean Ethics, trans. Rowe, 178), ‘outside our observation’ (W.  D.  Ross (trans.), ‘Nicomachean Ethics’, in id., The Complete Works of Aristotle: The Revised Oxford Translationi (Oxford, 1984), ii. 1798), ‘beyond our observation’ (R. Crisp, Aristotle: Nicomachean Ethics [Nicomachean Ethics, trans. Crisp] (Cambridge, 2004), 105), ‘cessons de regarder’ (Gauthier and Jolif, L’éthique à Nicomaque, ii. 163). T. Irwin (trans.), Nicomachean Ethics [Nicomachean



The Necessity of What We Understand

207

[13] Ἔπειτ’ ἀνασκευάζοντα μὲν εἰ τοιοῦτο ἀποδέδωκε τὸ ἴδιον, ὃ ϕανερὸν μὴ ἔστιν ἄλλως ὑπάρχον ἢ αἰσθήσει· οὐ γὰρ ἔσται καλῶς κείμενον τὸ ἴδιον. ἅπαν γὰρ τὸ αἰσθητὸν ἔξω γινόμενον τῆς αἰσθήσεως ἄδηλον γίνεται· ἀϕανὲς γάρ ἐστιν εἰ ἔτι ὑπάρχει, διὰ τὸ τῇ αἰσθήσει μόνον γνωρίζεσθαι. (Top. 5. 3, 131b19–23). For destructive criticism, see whether the idion is of such a sort that it is not evident whether it holds except by perception. For in that case, the idion will not be properly assigned. For all objects of perception take on an unclear status when they go outside perception, since it is not evident whether they still hold, on account of their being known only in perception.108

Aristotle is discussing what it takes for an idion, a counter-­predicating but non-­essential term (1. 5, 102a18–19), to have a certain favourable status in dialectical which he calls being ‘properly assigned’ (καλῶς κείμενον). The issue of whether the idion is ‘properly assigned’ is distinct, for Aristotle, from the issue of whether the idion holds of a given subject at all (5. 4, 132a22–4). In this context, Aristotle assumes that the purpose of assigning an idion to the subject is to render that subject ‘more comprehensible’ (γνωριμώτερον).109 This rules out, on the one hand, predicating the more obscure of the less obscure, as when ‘most similar to the soul’ (ὁμοιότατον ψυχῇ) is predicated of the subject ‘fire’ (πυρός, 5. 2, 129b9–13). The requirement also, however, disqualifies true predications that parties of the debate are in no position to verify, even if true (129b14–17). In [13], Aristotle claims that an idion which needs to be verified by means of perception is dialectically inappropriate in this way. He does not mean that no perceptible property can be a properly assigned idion, as he carefully clarifies: he takes the fact that surfaces are coloured to be a perceptible feature of them110 but not to Ethics, trans. Irwin], 2nd edn (Indianapolis, 1999), 88, gives a similar translation and adds what is in my view a correct parenthetical gloss: ‘whenever what admits of being otherwise escapes observation, we do not notice whether it is or is not, [and hence we do not know about it]’. 108 Cf. Metaph. Ζ. 10, 1036a5–7. For the text of the Topics, I employ the edition of W.D. Ross (ed.), Aristotelis Topica et Sophistici Elenchi (Oxford, 1958). 109  Top. 5. 3, 131a17. Cf. 5. 2, 129b1–5, 13–14, 22–3; 5. 3, 131a1. 110  αἰσθητόν, Top. 5. 3, 131b31–2. In order to avoid a contradiction, we must interpret αἰσθητόν here more broadly than at 131b21. In this line it evidently refers to any perceptible feature of a thing (whether knowledge of that perceptible feature depends on its being perceived), while at 131b21 it is used in a narrower sense to refer to properties which can be known to hold only by being perceived.

208

Joshua Mendelsohn

be dialectically inappropriate, since in this case the predicate ‘obviously belongs to its subject of necessity’ (ἐξ ἀνάγκης ὑπάρχον δῆλόν ἐστιν, 5. 3, 131b32). The example he offers of an improperly assigned perceptible property is ‘the brightest star which revolves around the earth [as an] idion of the sun’.111 Even if the sun always in fact goes around the earth and in fact always is the brightest star to do so (let us suppose with Aristotle that these things are both so), ‘goes around the earth’ (and so the compound property which includes this) is not properly assigned to the sun as an idion. This is because it is the type of property that is ‘known by perception’ (τῇ αἰσθήσει γνωρίζεται, 131b27) or ‘made clear to perception’ (τῇ αἰσθήσει ϕανερόν, 131b31). What Aristotle means, I take it, is that knowing whether the sun has this property at a given time requires perceptual verification: we need to perceive the movement of the sun at t  in order to know that it is moving at t, and so we cannot do this when the sun sets ‘on account of our lack of perception at that time’ (διὰ τὸ τὴν αἴσθσιν τότε ἀπολείπειν ἡμᾶς, 131b29–30). We might wonder whether this is really so, even in Aristotle’s view (shouldn’t a sufficiently developed theory of astronomy allow us to know that the sun moves around the earth all the time?). We needn’t place too much pressure on the example. Aristotle makes clear that the type of case he is trying to illustrate is one where our knowledge of p at t  depends on perceiving p at t. Let’s call this perception-­dependent knowledge. His view is that at least some knowledge is like this, and that while we can have such knowledge, we have it only when we are actually perceiving the object of our knowledge.112 Assuming that, for example, our basis for knowing the sun moves around the earth is just our perception of it doing so, we cease to know this at those times when we are unable to perceive the sun. One way to read [12] is to assume that Aristotle thinks know­ ledge of any contingency is perception-­dependent. Contingencies, according to this reading, are like ‘objects of perception’ (αἰσθητά, in the sense that αἰσθητόν is used at Top. 5. 3, 131b21) in that know­ ledge of a contingent state of affairs requires that contingent state of affairs to be perceptually present to the knower. While I think that this reading cannot be decisively ruled out, it has the unwelcome 111  Top. 5. 3, 131b25–6: ἡλίου ἴδιον ἄστρον ϕερόμενον ὑπερ γῆς τὸ λαμπρότατον. 112 Cf. Pr. An. 2. 21, 67a39–b3.



The Necessity of What We Understand

209

consequence of committing Aristotle to the view that all contingencies can be perceived, or else that non-­perceptible contingencies cannot in any sense be known. Aristotle never makes any claim of this sort so far as I am aware, and it is unfortunate if this assumption should be required to make sense of his views about the necessity of understanding. After all, some contingencies—­for example, facts about what someone is thinking at some time and place—­are not in any obvious sense things we know perceptually. Yet it is hard to see why Aristotle would want to deny that we can know them, at least in some mundane sense. An alternative is offered by C. D. C Reeve, who holds that we should resist assimilating Aristotle’s point in [12] to his point in [13]. He proposes a different way of understanding the phrase ἔξω τοῦ θεωρεῖν, and a very different way of taking Aristotle’s point in [12]. As Reeve reads him, Aristotle is not making a statement about contingency per se in this passage, but rather about the epistemic status of theorems of natural science which hold only for the most part. He uses the expression ἔξω τοῦ θεωρεῖν to describe theorems of natural sciences as opposed to those of ‘rigorous theoretical sciences’. As he puts it, Aristotle’s ‘thought’ in [12] (ii) ‘is that because theorems of natural science hold for the most part and so do not constitute strictly theoretical scientific knowledge, we cannot know whether they hold of unobserved cases’.113 Reeve consequently recommends translating ἔξω τοῦ θεωρεῖν as ‘whenever they fall ­outside theoretical knowledge’ (129). Aristotle’s point, on Reeve’s reading, is that whereas a demonstration in a rigorous science like mathematics allows us to know that all triangles everywhere, for example, have their characteristic angle sum, a demonstration in a natural science can at best allow us to know that something holds of those cases observed so far. Leaving aside whether Aristotle holds that demonstrations in natural science apply only to observed cases,114 there are two problems 113 C. D. C. Reeve, Aristotle on Practical Wisdom: Nicomachean Ethics VI [Aristotle on Practical Wisdom] (Cambridge, Mass., 2013), 129. 114 See Post. An. 1. 1, 71a34–b3, where Aristotle places weight on the claim that we do know unobserved instances of a fact we have demonstrated. Reeve might reply that Aristotle means to restrict his claim to demonstrations in rigorous sciences, but Aristotle does not say as much. In any case, the evidence Reeve adduces (Post. An. 1. 8, 75b24–30) does not show that Aristotle is committed to this claim.

210

Joshua Mendelsohn

with this reading: Aristotle’s argument turns, on Reeve’s reading, on the difference between demonstrations in different sciences, but Aristotle does not mention demonstration or rigorous as opposed to non-­rigorous sciences in [12]. In order to find this point in [12], Reeve needs to take Aristotle to be using a number of terms in restricted or unusual ways.115 The more serious problem for this reading, however, is Aristotle’s use of γένηται with ὅταν in [12] (ii). On Reeve’s reading, Aristotle’s point is that theorems of natural science always fall outside theoretical sciences, for which reason they are always restricted to observed cases. They do not sometimes fall within theoretical science on his view, and certainly they do not come to be (γένηται) outside theoretical science. Reeve’s reading would require Aristotle to say that theorems of natural science are restricted in this way because they (always) fall ἔξω τοῦ θεωρεῖν, but [12] (ii) cannot be translated in this way. For these reasons, I think that a reading closer to the original interpretation is preferable. We can, however, develop a reading along these lines without committing Aristotle to the questionable view that knowledge of contingencies depends on occurrent sensory perception. We noted that Aristotle has good grounds for taking not just scientific understanding but any type of knowledge that has an object to be a relative, and thus to depend on that object as a correlative (even where the object does not reciprocally depend upon it). Consider, then, the consequences for knowledge that has a contingent proposition p as its object. If at some point I know p, then, at least at that time, p must be true on Aristotle’s view.116 Aristotle holds, however, that a contingent state of affairs is capable of ceasing to be, in particular when the predicate ceases to be ‘combined’ 115  τὸ ἐπιστητόν (and ὃ ἐπιστάμεθα) needs to be understood to refer specifically to the object of knowledge in rigorous natural sciences, while τὰ δ’ ἐνδεχόμενα ἄλλως are taken to refer specifically to the type of contingency proven in natural sciences. Aristotle gives no indication of intending the latter restriction, and, while he makes clear that he is speaking about scientific knowledge in a precise way in this passage, he says nothing about any restriction to theoretical sciences. Likewise, Reeve wants to get out of the words λανθάνει εἰ ἔστιν ἢ μή (‘we don’t know whether it is so or not’) the thought that something cannot be known to hold of a case of a given generalization not yet observed. This would, to say the least, be a very opaque way for Aristotle to make that claim. 116 See Post. An. 1. 2, 71b25. Aristotle’s statement is about understanding there, but I take it that all forms of knowledge for Aristotle are at least factive. See further Fine, ‘Aristotle on Knowledge’, 228.



The Necessity of What We Understand

211

(συγκεῖσθαι, Metaph. Θ. 10, 1051b12) with the subject. This has the consequence, he notes, that ‘the same account and the same belief about contingent things come to be at one time true and at another time false, and it is possible for it to sometimes indicate the truth (ἀληθεύειν) and sometimes represent things falsely (ψεύδεσθαι)’.117 Given our analysis of Cat. 7, we can understand why. If a known fact or state of affairs ceases to exist, then there will no longer be anything for my knowledge to be ‘of ’.118 Given that contingent states of affairs can perish, then, my beliefs regarding contingencies are liable to fall out of sync with the object of my knowledge, even if (or precisely because) I don’t change my mind about anything.119 In the terminology developed above, knowledge of contingencies is liable to secondary loss. Now, this does not imply that knowledge of contingencies is impossible. What it does imply, however, is that the security of this type of knowledge will depend on the vigilance of the knower. Someone who has this knowledge and wishes to keep it will need to be poised to immediately update her cognitive state so as to match the changes in this contingent state of affairs, coming to hold p to be false, should the object of her knowledge that p perish, coming to hold it to be true, should the relevant state of affairs once again come into being. This is what I take Aristotle’s point to be in [12] (ii). He is not attempting to exclude all knowledge of contingencies, but rather only to articulate the conditions under which this type of know­ ledge is retained. In order to be guaranteed to remain knowledge, knowledge of contingencies requires constant attention to the thing known, specifically a type of attention that makes us notice when the relevant fact changes and thus change our beliefs. Visual observation will fit the bill, at least in cases where the state of affairs is visually perceptible. If, for instance, I am closely observing Socrates, then I will inevitably notice and thus come to know if it ceases to 117  περὶ μὲν οὖν τὰ ἐνδεχόμενα ἡ αὐτὴ γίγνεται ψευδὴς καὶ ἀληθὴς δόξα καὶ ὁ λόγος ὁ αὐτός, καὶ ἐνδέχεται ὁτὲ μὲν ἀληθεύειν ὁτὲ δὲ ψεύδεσθαι (Metaph. Θ. 10, 1051b13–15). This does not imply that the state of affairs itself is a bearer of truth: see D. Charles and M. Peramatzis, ‘Aristotle on Truth-­Bearers’, Oxford Studies in Ancient Philosophy, 50 (2016), 101–41. In other respects my interpretation of this sentence agrees with that of Crivelli, Aristotle on Truth, 60–1. 118  Cat. 7, 7b29–30. 119 Cf. DA 3. 3, 428b8–9.

212

Joshua Mendelsohn

be the case that Socrates is sitting. (I take ‘observe’ here to be a success term, so that observing Socrates entails knowing how things genuinely stand with Socrates’ visually perceptible features. Riders like ‘so long as I am not hallucinating, or a brain in a vat, etc’. are, therefore, unnecessary.) We needn’t, however, commit Aristotle to the view that perception is the only way that we can keep our beliefs regarding the contingent up to date. Aristotle uses θεωρεῖν to refer to a variety of different intellectual and perceptual activities.120 Here, it is likely that it functions as a catch-­all for the various activities we engage in that would ensure we notice changes in contingent states of affairs. θεωρεῖν might, in particular, be intended to include various sorts of purely or partially non-­perceptual forms of attention that can serve to keep our beliefs in sync with non-­perceptual contingencies, such as the introspection required to track whether I am currently thinking or whether I am currently sleepy, etc. Regardless of whether that is so, Aristotle’s claim in [12] (ii) will be that we are guaranteed to keep knowledge concerning the contingent only if we engage in a certain kind of active attending to that contingency, so as to neutralize the risk of secondary loss. Here it is important to bear in mind that scientific understanding, epistēmē, is only one type of knowledge, and so even if Aristotle allows knowledge of contingencies of some kind, it does not follow that he allows epistēmē of contingencies. In fact, Aristotle intends to leverage this conclusion to show that epistēmē differs from other kinds of knowledge in not permitting contingencies as its objects. Let us turn to this issue now. 5.2.  How does the claim that contingencies are KOWT establish that understanding is of necessities? On one reading, proposed by Jaakko Hintikka, Aristotle’s reason­ ing in [12] relies on the assumption that knowledge must always remain knowledge, since ‘ “false knowledge”—even merely some­ times false knowledge—­struck the Greeks, as it is likely to strike us

120 See Bonitz, Index Aristotelicus, 328a4–b56, esp. the references attached to 328a54–5, where Bonitz takes it to function to distinguish an activity from a capacity associated with epistēmē.



The Necessity of What We Understand

213

today, as a misnomer’.121 Knowledge of contingencies, however, is bound at some point to cease to be knowledge on Hintikka’s view, since there must come some time when the contingency fails to hold, and we inevitably will (in my terminology) suffer a secondary loss of knowledge at that time. There can, consequently, be no know­ ledge of contingencies. Since understanding is a kind of knowledge,122 understanding cannot be of contingencies. This reading has a number of problems. First, as noted above, Aristotle’s formulation in [12] (ii) strongly suggests, although it does not imply, that we can have knowledge of contingencies, albeit only with appropriate vigilance. If Aristotle’s conclusion is that we, after all, cannot have any type of knowledge of contingencies, then his conclusion contradicts a strong suggestion of his premisses, which is an awkward result. Second, if the considerations of Section 5.1 are sound, then Aristotle does not think that known contingencies will inevitably cease to hold and hence cease to be known, only that such knowledge can be subject to secondary loss. There is no reason to think that one’s knowledge will, therefore, sometimes be false.123 Third, Hintikka provides scant evidence that Aristotle holds the view that all knowledge must remain knowledge at all times, and this view is not very plausible. As we have seen, Aristotle seems to think that our knowledge lasts at most as long as we do, and in cases of cognitive decline he allows that it might not even last that long. As far as I can see, Aristotle does not think that the fact that knowledge of contingencies would require constant attention rules it out as genuine knowledge. Aristotle’s warrant for drawing this conclusion derives rather from the distinctive feature he takes ­scientific understanding to have as compared with other kinds of knowledge. In particular, whereas Aristotle holds that some other 121 Hintikka, ‘Time, Truth and Knowledge in Aristotle and Other Greek Philosophers’, 75. 122  See n. 15 above. 123  Hintikka is no doubt assuming what has become known as the ‘principle of plenitude’, the principle that truth at all times implies and is implied by necessary truth: see J. Hintikka, ‘Necessity, Universality, and Time in Aristotle’, in J. Barnes, M. Schofield, and R. Sorabji (eds.), Articles on Aristotle, vol. iii: Metaphysics (London, 1979), 108–24 at 111, for a classical formulation. See J. Barnes, ‘The Principle of Plenitude’, Journal of Hellenic Studies (1997), 183–6; and, especially, L.  Judson, ‘Eternity and Necessity in De caelo 1. 12: A Discussion of Sarah Waterlow, Passage and Possibility: A Study of Aristotle’s Modal Concepts’ [‘Eternity and Necessity’], Oxford Studies in Ancient Philosophy, 1 (1983), 217–55 for powerful arguments that Aristotle did not accept the thesis, at least in full generality.

214

Joshua Mendelsohn

types of knowledge are sustained only by certain mental or perceptual acts, scientific understanding cannot be like this: If we have understanding of p, then it is not the case that we understand p only when we actively attend to p. It should at this point be no surprise why I take Aristotle to be committed to this. The durability principle says that understanding is a type of knowledge we retain so long as we undergo no cognitive harm or decline. But failing, for example, to keep an eye on Socrates to see whether he has risen from his seat is certainly not a disqualifier of this sort. Ceasing to watch Socrates is typically a harmless procedure. In general, no type of active attending ought to be required in order to keep understanding if simply coming to no harm is sufficient to keep it. So, scientific understanding cannot be such that we only understand what we understand when we actively attend to or ‘theorize’ it.124 It will be noted that this is precisely the premiss Aristotle requires to render his argument in [12] valid. Aristotle’s argument may thus be represented as follows (writing, again, KOWT for ‘Known Only When Theorized’): (1) Everything that can be otherwise is such that it is KOWT [explicit premiss] (2) (but no object of scientific understanding is such that it is KOWT). [suppressed premiss] (3) Therefore, no object of scientific understanding can be other­wise. [conclusion] As we have seen, Aristotle endorses (1) on the basis of the dependency principle. (2) is a paraphrase of the claim just discussed, and I have argued it follows from Aristotle’s durability principle. Both the idea that understanding is stable from Categories 8 and the idea that it is dependent from Categories 7 are thus needed to secure Aristotle’s conclusion. Because scientific understanding is a relative, it depends on its correlative, the object of scientific understanding, 124 If this is Aristotle’s idea, then Hintikka, ‘Time, Truth and Knowledge in Ancient Greek Philosophy’ is exactly wrong to claim that Aristotle is assuming that ‘the highest forms of knowledge [are] somehow analogous to immediate observation’. It is not clear what role this premiss plays in Hintikka’s own reconstruction, but as I am reading him, Aristotle’s point is that the highest kinds of (theoretical) knowledge do not depend on immediate observation as other types of knowledge might.



The Necessity of What We Understand

215

holding.125 The status of scientific understanding as a relative thus explains why, if the object of understanding were a state of affairs that may cease to hold, we would be liable to lose this understanding when what we understand goes ‘out of view’ (ἔξω τοῦ θεωρεῖν). It does not, however, explain why this result is unacceptable, and hence it also does not explain why Aristotle takes himself to be warranted in rejecting the possibility of a contingent object of understanding. This is provided by the durability prin­ciple, which supplies (2).

6. Objections At this point an objection to Aristotle’s argument might be raised. Suppose that there is some state of affairs that is true from the time that it is learned by some knower S to the end of S’s life, but which is not true at all times simpliciter (it is false either before it is learned by S or after S dies, or both). In this case, S’s knowledge will satisfy the dependency condition, since, by stipulation, p is true whenever S knows it. It will apparently also satisfy the dur­abil­ity condition, since there is no time during her life when S ceases to understand p, and therefore, a fortiori, S continues to know p for as long as her cognitive machinery remains intact. Apparently, then, knowledge of this type of contingency would satisfy the durability and the dependency principles. If that is right, then the fact that understanding is durable and dependent does not rule out its having contingent objects, and something is wrong with Aristotle’s argument (at least as I have interpreted it). Another closely related objection is as follows. Let us grant, leaving aside these objections, that Aristotle establishes that whatever we know in the sense of epistēmē haplōs is true at all times. We might still worry here that Aristotle is playing fast and loose with temporal and modal notions, for the conclusion that he wishes to draw is not merely that the object of understanding is always true, but that it is necessarily true. 125 Kiefer, Aristotle’s Theory of Knowledge, 12–40 likewise highlights the im­port­ ance of the fact that epistēmē (which he translates as ‘knowledge’) is both a relative and a hexis, but he does not emphasize the tension between these claims, and he does not discuss the role they play in Aristotle’s argument that scientific understanding is of necessities.

216

Joshua Mendelsohn

I think Aristotle could respond to these objections. Seeing how will highlight the source of the properly modal status of understanding’s objects, about which I have so far had fairly little to say. 6.1.  Objection 1: Durability without eternal truth There are a few ways that the type of scenario sketched in the first objection might be envisioned, and for the purposes of responding to this objection it will help to separate them. First, what is known might be something which is true at all times when known by  S because it refers in some way to S’s cognitive state. For instance,  S might know the proposition ‘I am alive’, ‘I understand geometry’, or even ‘my cognitive faculties are intact’.126 These propositions, if true, are clearly contingent truths (S might not have been alive; S might not have known geometry, etc.), but we can easily imagine a case where they are known all the time during a person’s life, once learned. Let’s call this the ‘problem of introspective truth’. Second, we might consider cases where the thing known has nothing to do with S’s own cognitive state, but nevertheless turns out to be true at all the times when S  knows it. We might imagine, for instance, that Socrates, either due to a temporary disability or as some sort of long joke, remains seated from the moment that  S learns he is sitting and gets up only after S dies. Let’s call this the problem of luckily persisting knowledge, not to be confused with problems of epistemic luck discussed by contemporary ­epis­tem­olo­gists.127 Finally, we can imagine a rather different case of this sort. Suppose there is some proposition which becomes true at some point in the world’s history and remains true ever after, for instance, that the world is established as an ordered kosmos by a divine craftsman at some point in time. In this case too, if someone learns this fact once 126  Readers concerned about the indexical content introduced by personal pronouns may replace them with their own name, here and throughout, to see that nothing hangs on this. 127  When post-­Gettier epistemologists discuss epistemic luck, they are typically concerned with the possibility that a belief is acquired in a lucky way, not, as we are, with the possibility that a belief in a contingent state of affairs happens to remain true. M. Fricker, ‘The Value of Knowledge and the Test of Time’, Philosophical Aesthetics and the Sciences of Art: Royal Institute of Philosophy Supplement, 64 (2009), 121–38, esp. 128–9, however, argues that epistemologists ought to be more concerned with the diachronic stability of knowledge, an issue she traces back to the Meno.



The Necessity of What We Understand

217

it has become true, they will never lose it on account of a change in its truth, since (by stipulation) it always remains true after becoming true. Yet the proposition is not an eternal truth; there is a time at which it was false. Call this the ‘problem of Timaean possibilities’.128 Consider first the problem of luckily persisting knowledge. While these cases clearly satisfy the dependency principle, they do not satisfy the durability principle on the most plausible way of understanding it. Aristotle’s durability principle, as I formulated it above, says: If S  understands  O, then S continues to understand O  so long as she experiences no detrimental changes to the constitution of her cognitive faculties. There is, however, an ambiguity in this formulation as it stands, since the modal force of ‘continues’ is not specified. On a narrow reading, we might take this to mean simply: If S understands p at t, then, for every t ¢  after t, S understands p at t ¢ (so long as she experiences no detrimental changes to the constitution of her cognitive faculties between t  and t ¢ ). However, this is not the only way to understand this principle. We might also take the condition to be modally robust, as follows: If S understands p at t, then necessarily, for every t ¢ after t, S understands p at t  (so long as she experiences no detrimental changes to the constitution of her cognitive faculties between t  and t ¢ ).129 On the first formulation, which I will call the ‘non-­modal dur­abil­ ity principle’, the cases of luckily persisting belief stand. On the second, they do not. Even if Socrates in fact remains seated until S’s death, the fact that he could have ceased to sit without injuring S’s psyche means that the object of her knowledge could have perished without any detrimental change to her. Given the dependency 128  I draw the moniker ‘Timaean’ from Judson, ‘Eternity and Necessity’, 285. As we will see, it is no coincidence that we must reach for a non-­Aristotelian example. 129 To avoid any ambiguity of scope, we can write this condition formally as (∀t)(Uspt → (∀t' > t) ☐ [¬Dstt' → Uspt' ]), whereUxyz means x understands y at time z, and Dxyz means x experiences a relevant detrimental change between time y and time z. The necessity, that is, takes narrow scope and governs the internal ­conditional.

218

Joshua Mendelsohn

principle, this means that S might have ceased to know it—­and not on account of any change in her, but rather just because Socrates got up. Hence S’s knowledge is not, in this case, in fact durable on the modal reading of the durability principle, since it is subject to a possible secondary loss (even if no actual one occurs), whereas the modal durability principle requires scientific understanding to be counterfactually and not just actually stable. The same response will not, in general, work for cases of introspective truth. Suppose the proposition I know is that my cognitive faculties are functioning well, and that as a matter of fact I continue to know this throughout my life. In this case, if the proposition I know ceases to be true, I ipso facto incur a detrimental change to my cognitive condition, since for that known proposition to cease to be true just is for it to cease to be the case that my cognitive faculties are functioning well. It follows that a secondary loss, that is, a loss where the thing known changes without any change in me, cannot possibly occur. This is admittedly a more difficult case. I think Aristotle’s best response would be to claim that we cannot rule out a primary loss of this type of belief. Whereas a belief in a fundamental scientific fact might come to be so deeply ingrained in my belief system that nothing could persuade me to renounce it, it is difficult to see how the same could be true in the case of some fact about the contingent condition of my own psyche. I might, for instance, be misled by a particularly cunning and manipulative sophist who convinces me that I am in cognitive decline when in fact I am not. Such a scenario, far-­fetched as it is, does seem a genuine possibility, and there seems no reason to assume that I must actually be cognitively injured in any such scenario. The same is plausible, mutatis mutandis, for other cases where the proposition I know is implied by the condition that the I undergo no cognitive detriment. Given these possibilities, this type of contingency also could cease to be known even in cases where the knower comes to no harm. If that is so, the modal durability principle rules out understanding of contingent introspective truths. If Aristotle holds the modal durability principle, then, he is not vulnerable to the first class of counter-­examples, and at least has a serviceable reply to the second. The first piece of evidence in favour of a modal understanding of the durability principle is the language Aristotle uses to describe the durability of understanding. He says that understanding is ‘steady’ (δυσκίνητος) and ‘abiding’



The Necessity of What We Understand

219

(παραμόνιμος). This language indicates more than just an actual persistence. As we saw, the reference is to Meno,130 where these terms are used to describe the reliability of knowledge compared with true belief, on the model of an abiding slave. Now, an abiding slave is not one who just in fact sticks around, perhaps because the costs of running away turn out by chance to be too great. An abiding slave is one who would not run away in some appropriate range of counterfactual circumstances. Similarly, for understanding to be steady and abiding means that it not only actually remains understanding, but that it would do so at least in those circumstances that do not include a destructive change in the soul, whose condition is the source of this guarantee. Further evidence that Aristotle intends the principle to be understood in this way may be garnered from an argument Aristotle gives in Posterior Analytics 1. 6: [14] ἔτι εἴ τις μὴ οἶδε νῦν ἔχων τὸν λόγον καὶ σῳζόμενος, σῳζομένου τοῦ πράγματος, μὴ ἐπιλελησμένος, οὐδὲ πρότερον ᾔδει. ϕθαρείη δ’ ἂν τὸ μέσον, εἰ μὴ ἀναγκαῖον, ὥστε ἕξει μὲν τὸν λόγον σῳζόμενος σῳζομένου τοῦ πράγματος, οὐκ οἶδε δέ. (Post. An. 1. 6, 74b32–6) (i) Again, if someone does not know something now, although he possesses the account and is preserved, and the object is preserved, and he has not forgotten, then he did not know it earlier either. (ii) But the middle term might perish if it is not necessary, so that he will retain the account and the object will be preserved, but he will not have knowledge.131

Here Aristotle gives an argument that is structurally similar to the argument of [12], but for a different conclusion. In this chapter, Aristotle is assuming that scientific understanding of p requires grasping a demonstration whose conclusion is p.132 Taking as a premiss that what we understand is a necessary truth, Aristotle argues for a thesis concerning the character of the demonstration by which we have scientific understanding. Not only must the thing we understand by means of demonstration (its conclusion) be 130  See n. 77 above. 131  The division into (i) and (ii) is my own, for ease of exposition. 132  Given that [12] is embedded within a summary of the theory of understanding in the Posterior Analytics, Aristotle probably takes this for granted when he formulates [12] as well, but his argument there as I have reconstructed it does not depend on any particular assumption about the connection between understanding and demonstration.

220

Joshua Mendelsohn

­ ecessary; the ‘middle term’ (μέσον) of the demonstration must n also be necessary. By the middle term of a demonstration being ‘necessary’ (ἀναγκαῖον), Aristotle seems to mean that it constitutes a necessary ‘link’ between the subject and predicate that makes the premisses of the demonstration true. Aristotle argues in [14] that if the middle term could ‘perish’ (ϕθαρείη), causing one or both of the premisses to go from being true to being false, it would be possible for someone to continue to remember a demonstration without coming to any harm, and without the object of scientific understanding changing in the manner countenanced in [12], and yet for what they remember to cease to be a sound argument for what they know.133 Aristotle holds that this cannot occur and that, consequently, the premisses of a demonstration must be necessary truths. What he is assuming is that demonstrations are reliably sufficient for retaining understanding, modulo certain disqualifying conditions that he specifies. He states this condition in [14] (i), in contrapositive form. Where D is a demonstration for p, Aristotle says that: If, between t  and  t ¢, S possesses D,  p does not perish, S does not forget [D or p], and S experiences no detrimental change and yet S does not understand p at t ¢, then  S also did not under­ stand p at t. It is easy to see that this is equivalent to the following: If D is a demonstration for p, then, if   S understands p by possessing D, S continues to understand p so long as S experiences no cognitively detrimental change, and does not forget [D or p], and it does not cease to be the case that p. In other words, remembering a demonstration is supposed to suffice for the continued possession of demonstrative scientific know­ ledge, given certain additional provisos made explicit in [14].134 Since a demonstration only imparts knowledge of its conclusion if, 133  Cf. Barnes, Posterior Analytics, 2nd edn, 128. 134  In light of the foregoing, we can see that the conditions that S does not forget and that p does not cease to be true are unnecessary on Aristotle’s view, since he thinks that if S has understanding of p , she is guaranteed not to forget it and it is guaranteed not to change. In sentence (i) of [14], Aristotle is presumably stating the principle in its most general form, even though he takes some of these conditions always to be satisfied. His theory of knowledge as relatives still entails that one would cease to have know­ledge if p were to cease to be true, and that S would cease to know if she were to forget.



The Necessity of What We Understand

221

at a minimum, its premisses are true (Post. An. 1. 2, 71b25), it ­follows that one will cease to grasp a demonstration that p if the premisses of the demonstration become false. The person still ‘will possess the account’ (ἕξει . . . τὸν λόγον), but the account will cease to constitute a demonstration; consequently, given the assumption that understanding p requires demonstration that p, S will cease to understand p. Hence, the premisses of a demonstration must also, like the fact understood, not change their truth values.135 What is of greatest relevance here is the remark that Aristotle goes on to make next. He says: [15] εἰ δὲ μὴ ἔϕθαρται, ἐνδέχεται δὲ ϕθαρῆναι, τὸ συμβαῖνον ἂν εἴη δυνατὸν καὶ ἐνδεχόμενον. ἀλλ’ ἔστιν ἀδύνατον οὕτως ἔχοντα εἰδέναι. (Post. An. 1. 6, 74b36–9) And if, although the middle term has not perished, it is possible for it to perish, the result can occur and is possible; but it is impossible to have knowledge under such conditions.

Aristotle is considering an objection parallel to the one under consideration, but about the middle term of a demonstration rather than the fact thereby understood (which corresponds to its conclusion). What if the middle term could perish, so that the premisses could cease to be true, but it never actually does perish, and so the premisses never actually cease to be true? Can a person in such a case be said to have scientific understanding on the basis of a proof from contingent premisses? Aristotle answers in the negative. The reason he gives is that it would still be possible for the scenario envisaged to occur (viz. for one or both premisses to become false without any of the other defeaters to knowledge occurring). It is, however, ‘impossible to have knowledge under such conditions’ (ἀδύνατον οὕτως ἔχοντα εἰδέναι, 74b38–9) in Aristotle’s view. In other words, the very possibility of the premisses switching truth value, and not just the actual occurrence of this at some time, is incompatible with the guarantee of continued understanding that the grasp of a demonstration is meant to provide. Demonstrations must be such as to necessarily guarantee knowledge, modulo the disqualifiers he lists in [14] (i). If we translate this reply to the case of the object of scientific understanding, the response would go like this: suppose someone understands a contingency, and suppose that this contingency 135 Cf. Metaph. Ζ. 15, 1039b32–1040a2.

222

Joshua Mendelsohn

never actually ceases to obtain. If the fact is contingent, then it is still the case that it could cease to obtain. Further, there is no reason to suppose that this counterfactual occurrence would need to entail any harm to the knower or a corresponding change in the knower’s mental state. The mere possibility that it could switch in truth value without any harm to the person who understands it is incompatible with the condition that epistēmē is guaranteed to abide so long as the knower comes to no harm. Hence, the object of scientific know­ ledge must not only, as a matter of fact, remain true for as long as S is in the appropriate condition; it must necessarily remain so during this time. Admittedly, Aristotle does not actually consider this objection to the argument in [12], and his language there emphasizes temporal continuity rather than counterfactual possibility.136 Nevertheless, given that Aristotle offers [15] as a supplement to the argument in [14], it seems not unlikely that he would be willing to provide a supplement to the argument in [12] along the same lines. If that is so, then there is good reason to suppose that Aristotle would endorse the modal durability principle. Not only does scientific understanding require knowledge to be actually retained where the knower is uninjured; he takes it to be incompatible with any circumstance which could result in its loss where the knower is uninjured.137 There remains the problem of Timaean possibilities. The reply given to the other cases is not applicable here, since the problem in this case is not that the state of affairs known could cease to hold (even if it never does). The problem is that this type of knowable object is not eternal (because it comes to be only after some particular time t ), and this holds even if we suppose it can never pos­ sibly cease to be true after someone comes to know it. In De caelo 1, the prospect of such possibilities is at issue, since Aristotle is concerned with whether the kosmos has a beginning, and the fact that the kosmos is imperishable seems to him directly relevant to this

136 Note ὅταν at 1139b21. Cf. ὁτέ in Metaph. Ζ. 15, 1039b33. 137  Notice that if Aristotle would endorse this argument, then it warrants attributing the modal durability principle with scope as formalized in n. 129 above. The parallel claim is that it would disqualify S’s knowing p now if at some future time she could cease to know it without undergoing any disqualifying changes; in other words, if she knows it now, then at all future times, it is necessary that she knows it if she doesn’t undergo such changes.



The Necessity of What We Understand

223

question. Aristotle argues against the view that the kosmos has a beginning on the very abstract grounds that: [16] τὸ δὲ ϕάναι μηδὲν κωλύειν γινόμενόν τι ἄϕθαρτον εἶναι καὶ ἀγένητον ὂν ϕθαρῆναι, ἅπαξ ὑπαρχούσης τῷ μὲν τῆς γενέσεως τῷ δὲ τῆς ϕθορᾶς, ἀναιρεῖν ἐστι τῶν δεδομένων τι. (De caelo 1. 12, 283a4–7) to say that nothing prevents something subject to generation from being imperishable, and something that is not subject to generation from perishing, so long as the coming-­to-­be, in the one case, and the perishing, on the other, happen only once, is to remove one of the givens.138

Aristotle is denying two things here: (1) that something could come to be once, and subsequently be imperishable, and (2) that something which is not subject to generation could perish. He holds that there can only be things whose duration of existence is unlimited in both directions or whose duration is limited on both sides; there cannot, he claims, be things whose duration of existence is limited in one direction only. I will not consider his argument here. What concerns us is that Aristotle denies (1). The context suggests that Aristotle means to include predicative beings like states of affairs among things which come to be and perish.139 Supposing, as elsewhere, that for a state of affairs to ‘come to be’ is for it to come to be so, and for it to be ‘imperishable’ at least implies that it could not cease to be so, this amounts to a denial of Timaean possibilities. While I will not attempt to treat this case fully here, [16] shows that Aristotle may have independent reasons for ruling out contingencies with this temporal profile.140 6.2.  Objection 2: Eternal truth without necessity I have argued so far that the object of understanding must be an eternal truth. The second objection is that Aristotle’s argument, 138  The text here follows P. Moraux (ed.), Aristote: Du ciel [Du ciel] (Paris, 1965). Translation modified from S. Legatt, Aristotle: On the Heavens: Books I & II [On the Heavens] (Oxford, 1995). 139 See De caelo 1. 12, esp. 281a30–3, 281b15–17. 140  Judson, ‘Eternity and Necessity’, 235–41, argues that Aristotle’s claim makes most sense if understood to apply only to ‘natural’ possibilities, but Aristotle does not qualify his claim in this way. C. J. F. Williams, ‘Aristotle and Corruptibility’, Religious Studies, 1 (1965), 95–107 at 212 n. 8, stresses that the claim should be understood as a very general one.

224

Joshua Mendelsohn

even if it does establish that the objects of scientific understanding are eternally true, does not establish them as necessary truths. A possible response to this objection is to claim that Aristotle has in mind a notion of necessity for which eternal truth is sufficient, or even which just means ‘true at all times’, when he claims that the object of scientific understanding is ‘of necessity’ (ἐξ ἀνάγκης) at NE 6. 3, 1139b22. A proponent of this response need not maintain that this is Aristotle’s only sense of necessity.141 Rather, one need only maintain that this is one sense in which Aristotle uses ‘ne­ces­sar­ily’, and in fact the sense he employs in [12]. The clearest ­support for such a view is a passage in De generatione et cor­ ruptione 2.11:142 [17] τὸ γὰρ ἐξ ἀνάγκης καὶ ἀεὶ ἅμα· ὃ γὰρ εἶναι ἀνάγκη οὐχ οἷόν τε μὴ εἶναι· ὥστ’ εἰ ἔστιν ἐξ ἀνάγκης, ἀίδιόν ἐστι, καὶ εἰ ἀίδιον, ἐξ ἀνάγκης. (GC 2. 11, 337b35–338a2) For ‘necessarily’ and ‘always’ go together (since what necessarily is cannot not be), so that if it is necessarily, it is eternal, and if it is eternal, it is ne­ces­sar­ily.143

Aside from the controversies over whether Aristotle recognizes this as even one sense of necessity,144 a downside of this reading is that Aristotle would not be establishing his conclusion in a way that supports the type of necessity he takes the object of understanding to have in the Posterior Analytics. For there, as we have seen, Aristotle takes the necessity of knowledge to be of a spe­cif­ic­al­ly essentialist kind. This objection is not fatal: perhaps Aristotle has independent reasons for thinking that the objects of understanding are essentialist 141  Hintikka, ‘Necessity, Universality, and Time in Aristotle’, and S. Waterlow, Passage and Possibility: A Study of Aristotle’s Modal Concepts (Oxford, 1982), both take Aristotle to operate always with a conception of necessity for which truth at all times is sufficient. 142  That Aristotle goes on immediately to infer that the object of understanding is ‘eternal’ (ἀίδια, NE 6. 3, 1139b24) and ‘subject neither to generation nor corruption’ (ἀγένητα καὶ ἄϕθαρτα, 1139b24) does suggest that he associates the relevant notion of necessity with eternality, but he only commits himself to eternality being a necessary condition here, not a sufficient one. Aristotle’s inference, therefore, neither strongly supports nor strongly weighs against the proposal that eternal truth is sufficient for being necessary in the sense at issue in [12]. 143  The text here follows C. Mugler (ed.), Aristote: De la génération et de la corruption [De la génération et de la corruption] (Paris, 1966), and the translation C. J. F. Williams, Aristotle’s De generatione et corruptione (Oxford, 1982). 144  See the references in n. 123 above.



The Necessity of What We Understand

225

necessities specifically, and perhaps his arguments there require only the premiss that understanding is necessary in this broad sense of omnitemporality. Such a reading would need, of course, to be substantiated by a close reading of the Posterior Analytics. Fortunately, the foregoing discussion gives us the resources to mount a reply on Aristotle’s behalf which does not depend on any such assumption. For, as we have seen, the modal durability prin­ciple implies that the object of understanding is such that it ne­ces­sar­ily remains true. Now, that a fact necessarily remains true is, of course, not the same as for it to be necessarily true. Yet it is not easy to come up with examples of contingencies that necessarily remain true, and Aristotle at any rate seems to take his notion of contingency to rule out any such examples at Metaphysics Θ. 10, where he claims statements of contingencies can come to be true and can come to be false.145 One possibility is that Aristotle simply does not distinguish being necessary from necessarily remaining true.146 Alternatively, Aristotle might be working from the assumption that being ne­ces­sar­ily true is the only reason a fact could necessarily remain true. In that case, he could justify his conclusion by a sort of inference to the best explan­ation: he might say that nothing could explain the necessary permanence of a known truth other than that fact’s necessity, and so infer the necessity of the object of understanding from its necessarily permanent truth. Admittedly, this is a speculative interpretation, but it seems to me the strongest defence available to Aristotle.

7.  Concluding remarks I have argued that Nicomachean Ethics 6. 3 gives an argument for the claim that scientific understanding is of a necessity, and that we 145  περὶ μὲν οὖν τὰ ἐνδεχόμενα ἡ αὐτὴ γίγνεται ψευδὴς καὶ ἀληθὴς δόξα καὶ ὁ λόγος ὁ αὐτός, καὶ ἐνδέχεται ὁτὲ μὲν ἀληθεύειν ὁτὲ δὲ ψεύδεσθαι (Metaph. Θ. 10, 1051b13–15). Contemporary philosophers might consider propositions about the past, like the fact that there was a battle on the Nile under Ramses III, to be examples of contingent propositions that necessarily remain true. While Aristotle’s views about this kind of case cannot be settled fully here, it is at least not clear that Aristotle would agree, and there is some suggestion that he treats such past occurrences simply as necessities (at least from the perspective of the present): See De caelo 1. 12, 283b12–14; NE 6. 2, 1139b7–9; Rhet. 3. 17, 1418a1–5; and the notoriously difficult statement at De int. 9, 19a2–4. 146  For a view with this consequence, see Crivelli, Aristotle on Truth, 21.

226

Joshua Mendelsohn

should read the argument in [12] as tacitly drawing on premisses spelled out more fully in Categories 7–8. The durability and dependency principles of these chapters generate a tension, but not a ­contradiction. What is, in fact, contradictory is not the combination of durability and dependency, but the triad that (1) some object of understanding may cease to hold, (2) understanding depends on its object’s continuing to hold, and (3) understanding is retained so long as the person with understanding comes to no harm. Aristotle does not wish to relinquish either (2) or (3), since he wishes to maintain that understanding has a place both in the category of relatives and in the subcategory of qualities he calls states. Instead, he rejects (1), the assumption that we ever have understanding of something that may cease to hold, and infers that it is a necessity. I  do not claim to have defended Aristotle’s argument against all possible objections in this paper, but I do hope to have shown that Aristotle has replies available to the most pressing ones and that his argument is more subtle and has greater staying power than it may seem. Like Bolton, I have emphasized the importance of the idea that scientific understanding is diachronically reliable. Aristotle’s reason­ing in [12] is bound to seem questionable if we take Aristotle to be defending his claim solely on the basis of its reliability, however, as Bolton in effect does when he takes Aristotle to defend his conclusion ‘on the ground that epistēmē is something you should be able to reliably count on even apart from continued observation of the state of affairs in question’.147 This presentation of Aristotle’s argument raises the question: why should we think that such continued observation is required for any type of knowledge, scientific or otherwise? To understand this, we need to take account of the character of knowledge as a relative. Conversely, while Hintikka is correct to emphasize the importance of the view that the truth may change in such a way as to undermine our knowledge,148 this also cannot explain Aristotle’s thesis about understanding on its own, since Aristotle does not take this to be a problem for all types of knowledge, but only for knowledge with the special type of stability understanding is supposed to have. It is Aristotle’s desire to combine these ideas which generates his view. 147  Bolton, ‘Science and Scientific Inquiry’, 53. 148  Hintikka, ‘Time, Truth and Knowledge in Ancient Greek Philosophy’, 7.



The Necessity of What We Understand

227

Both the idea that understanding is of something and the idea that a certain notion of understanding is distinctively stable are plausibly things that a wide variety of speakers might assent to, even if they do not understand them precisely in the way that Aristotle does in the Categories. If that is right, then Aristotle’s statement that the necessity of what we understand is something we all presume might be intended to say that this is something many people in effect presume, or are committed to, given certain features of their pre-­theoretic conception of understanding in the context of science. We needn’t take him to mean that ordinary speakers, or even his philosophical peers, would have actually drawn the inference. I have not attempted to ascertain whether, in Aristotle’s overall picture, he intends for explanatory or essential connections to ground the necessities that we understand. What we can say is that he does not argue that we understand essential or explanatory connections and therefore necessities; if he argues in either direction, it is the reverse. While I have also not traced the lineage of these ideas to past thinkers, these results cast doubt on any in­ter­pret­ ation which sees Aristotle’s view simply as an inheritance from the tradition. For even if the ideas underlying his argument have currency in the tradition, Aristotle elaborates and precisifies them using his own philosophical machinery, particularly the theory of relatives and states in Categories 7–8. In this respect, at least, he presents an argument for his position that his predecessors could not have given. Loyola University Chicago

BIBLIOGR A PH Y Ackrill, J. L. (trans.), Aristotle: Categories and De Interpretatione [Categories and De Interpretatione] (Oxford, 1975). Angioni, L., ‘Aristotle’s Definition of Scientific Knowledge’, Logical Analysis and History of Philosophy, 19 (2016), 140–66. Angioni, L., ‘Knowledge and Opinion about the Same Thing in APo A-33’, Dois Pontos, 10 (2013), 255–90. Barnes, J., Aristotle: Posterior Analytics, 2nd edn [Posterior Analytics, 2nd edn] (Oxford, 1993). Barnes, J., Aristotle's Posterior Analytics, 1st edn [Posterior Analytics, 1st edn] (Oxford, 1975).

228

Joshua Mendelsohn

Barnes, J., ‘The Principle of Plenitude’, Journal of Hellenic Studies, 97 (1977), 183–6. Bodéüs, R., Aristote: Catégories [Catégories] (Paris, 2001). Bolton, R., ‘Aristotle on Essence and Necessity’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, 13 (1997), 113–38. Bolton, R., ‘Science and Scientific Inquiry in Aristotle: A Platonic Provenance’ [‘Science and Scientific Inquiry’], in C. Shields (ed.), The Oxford Handbook of Aristotle (Oxford, 2012), 46–59. Bonitz, H., Index Aristotelicus (Berlin, 1870). Broadie, S. (comm.) and Rowe, C. (trans.), Aristotle: Nicomachean Ethics [Nicomachean Ethics] (New York, 2002). Bronstein, D., Aristotle on Knowledge and Learning: The Posterior Analytics [Aristotle on Knowledge and Learning] (Oxford, 2016). Bronstein, D., ‘Aristotle’s Virtue Epistemology’, in S. Hetherington and N. D. Smith (eds.), What the Ancients Offer to Contemporary Epistemology (New York, 2020), 157–77. Burnet, I., Platonis Opera, vol. iii (Oxford, 1903). Burnet, I., Platonis Opera, vol. iv (Oxford, 1902). Burnyeat, M., ‘Aristotle on Understanding Knowledge’, in E. Berti (ed.), Aristotle on Science: The Posterior Analytics (Padua, 1981), 97–139. Busse, A. (ed.), Eliae in Porphyrii Isagogen et Aristotelis Categorias Commentaria, Commentaria in Aristotelem Graeca, vol. xviii.1 (Berlin, 1900). Busse, A. (ed.), Olympiodori Prologomena et in Categorias commentarium, Commentaria in Aristotelem Graeca, vol. xii.1 (Berlin, 1902). Busse, A. (ed.), Philoponi (olim Ammonii) in Aristotelis Categorias ­commentarium, Commentaria in Aristotelem Graeca, vol. xiii.1 (Berlin, 1907). Busse, A. (ed.), Porphyrii Isagoge et in Aristotelis Categorias commentar­ ium, Commentaria in Aristotelem Graeca, vol. iv.1 (Berlin, 1887). Bywater, I. (ed.), Aristotelis Ethica Nicomachea (Oxford, 1894). Cartwright, N., ‘Why Trust Science?’, Proceedings of the Aristotelian Society, 120 (2020), 237–52. Charles, D. and Peramatzis, M., ‘Aristotle on Truth-Bearers’, Oxford Studies in Ancient Philosophy, 50 (2016), 101–41. Crisp, R., Aristotle: Nicomachean Ethics [Nicomachean Ethics, trans. Crisp] (Cambridge, 2004). Crivelli, P., Aristotle on Truth (Cambridge, 2004). Danesi, S., Johnson, C., and Barðdal, J., ‘Where Does the Modality of Ancient Greek Modal Verbs Come from?’, Journal of Greek Linguistics, 18 (May 2018), 45–92. Detel, W., Aristoteles: Analytica posteriora [Aristoteles], vol. ii (Berlin, 1993). Duncombe, M., Ancient Relativity: Plato, Aristotle, Stoics and Sceptics [Ancient Relativity] (Oxford, 2020).



The Necessity of What We Understand

229

Ebert, T., ‘Review of Mignucci, L’argomentazione dimostrativa in Aristotele and Barnes, Aristotle: Posterior Analytics’ [‘Review of Mignucci’], Archiv für Geschichte der Philosophie, 62 (1980), 85–91. Ferejohn, M., Formal Causes (New York, 2013). Ferejohn, M., The Origins of Aristotelian Science (New Haven, 1991). Fine, G., ‘Aristotle on Knowledge’, Elenchos, 14 (2010), 121–55, reprinted in ead., Essays in Ancient Epistemology, 221–42. Fine, G., ‘Aristotle’s Two Worlds: Knowledge and Belief in Posterior Analytics 1. 33’ [‘Aristotle’s Two Worlds’], Proceedings of the Aristotelian Society, 110 (2010), 323–46, reprinted in ead., Essays in Ancient Epistemology, 243–61. Fine, G., Essays in Ancient Epistemology (Oxford, 2021). Fine, G., ‘Knowledge and Belief in Republic V’, Archiv für Geschichte der Philosophie, 60 (1978), 121–39. Fleet, B., Simplicius: On Aristotle’s ‘Categories 7–8’ [Simplicius, trans. Fleet] (Ithaca, 2002). Frede, M., ‘The Title, Unity and Authenticity of Aristotle’s Categories’ [‘Title, Unity and Authenticity’], in id., Essays in Ancient Philosophy (Minneapolis, 1987), 11–28. Fricker, M., ‘The Value of Knowledge and the Test of Time’, Philosophical Aesthetics and the Sciences of Art: Royal Institute of Philosophy Supplement, 64 (2009), 121–38. Friedman, M., ‘Explanation and Scientific Understanding’, Journal of Philosophy, 71 (1974), 5–19. Gauthier, R. A. and Jolif, J. Y., L’éthique à Nicomaque: Introduction, tra­ duction et commentaire [L’éthique à Nicomaque], 2 vols. (Leuven, 1959). Gerson, L. P., Ancient Epistemology (Cambridge, 2009). Harari, O., ‘The Unity of Aristotle’s Category of Relatives’, Classical Quarterly, n.s., 61 (2011), 521–37. Hempel, C. G., ‘Aspects of Scientific Explanation’, in id., Aspects of Scientific Explanation and Other Essays in the Philosophy of Science (New York, 1965), 331–469. Hintikka, J., ‘Necessity, Universality, and Time in Aristotle’, in J. Barnes, M.  Schofield, and R.  Sorabji (eds.), Articles on Aristotle, vol. iii: Metaphysics (London, 1979), 108–24. Hintikka, J., ‘Time, Truth and Knowledge in Ancient Greek Philosophy’, American Philosophical Quarterly, 4 (1967), 1–14. Hintikka, J., ‘Time, Truth and Knowledge in Aristotle and Other Greek Philosophers’, Time and Necessity: Studies in Aristotle’s Theory of Modality (Oxford, 1973), 62–92. Hood, P. M., Aristotle on the Category of Relation (Lanham, 2004). Hutchinson, D. S., The Virtues of Aristotle (London, 1986). Irwin, T. (trans.), Nicomachean Ethics [Nicomachean Ethics, trans. Irwin], 2nd edn (Indianapolis, 1999).

230

Joshua Mendelsohn

Jaeger, W. (ed.), Aristotelis Metaphysica (Oxford, 1957). Judson, L., ‘Eternity and Necessity in De caelo 1. 12: A Discussion of Sarah Waterlow, Passage and Possibility: A Study of Aristotle’s Modal Concepts’ [‘Eternity and Necessity’], Oxford Studies in Ancient Philosophy, 1 (1983), 217–55. Kalbfleisch, C. (ed.), Simplicii in Aristotelis Categorias commentarium, Commentaria in Aristotelem Graeca, vol. viii (Berlin, 1907). Kiefer, T., Aristotle’s Theory of Knowledge (London, 2007). Kratzer, A., ‘What “Must” and “Can” Must and Can Mean’, Linguistics and Philosophy, 1 (1977), 337–55. Laks, A. and Most, G.  W., Early Greek Philosophy (Cambridge, MA, 2016). Legatt, S., Aristotle: On the Heavens: Books I & II [On the Heavens] (Oxford, 1995). Lesher, J. H., ‘On Aristotelian ἐπιστήμη as “Understanding” ’ [‘Ἐπιστήμη as “Understanding” ’], Ancient Philosophy, 21 (2001), 45–55. McKirahan, R. D., Principles and Proofs (Princeton, 1992). Menn, S., ‘Aristotle’s Definition of Soul and the Programme of the De Anima’, Oxford Studies in Ancient Philosophy, 22 (2002), 83–139. Mignucci, M., L’argomentazione dimostrativa in Aristotele (Padua, 1975). Mignucci, M., ‘Aristotle’s Definitions of Relatives in “Cat.” 7’, Phronesis, 31 (1986), 101–27. Minio-Paluello, L. (ed.), Aristotelis Categoriae et liber De Interpretatione (Oxford, 1949). Morales, F., ‘Relational Attributes in Aristotle’, Phronesis, 38 (1994), 255–74. Moraux, P. (ed.), Aristote: Du ciel [Du ciel] (Paris, 1965). Morison, B., ‘Aristotle on the Distinction between What Is Understood and What Is Believed’ (unpublished). Morison, B., ‘Colloquium 2: An Aristotelian Distinction between Two Types of Knowledge’ [‘Two Types of Knowledge’], Proceedings of the Boston Area Colloquium of Ancient Philosophy, 27 (2012), 29–63. Morrison, D., ‘The Taxonomical Interpretation of Aristotle’s Categories’, in A. Preus and J. P. Anton (eds.), Essays in Ancient Greek Philosophy, vol. v: Aristotle’s Ontology (New York, 1992), 19–46. Moss, J. and Schwab, W., ‘The Birth of Belief’, Journal of the History of Philosophy, 57 (2019), 1–32. Mugler, C. (ed.), Aristote: De la génération et de la corruption [De la génération et de la corruption] (Paris, 1966). Palmer, J., Parmenides and Presocratic Philosophy (Oxford, 2009). Pasnau, R., ‘Epistemology Idealized’, Mind 122 (2013), 987–1021. Peramatzis, M., ‘Aristotle on How Essence Grounds Necessity’, in D.  Bronstein, T.  Johansen, and M.  Peramatzis (eds.), Aristotelian Metaphysics, Ancient & Modern (Oxford, forthcoming).



The Necessity of What We Understand

231

Peramatzis, M., ‘Aristotle on Knowledge & Belief: APo. I. 33’ [‘Aristotle on Knowledge & Belief’] (unpublished). Reeve, C. D. C., Aristotle on Practical Wisdom: Nicomachean Ethics VI [Aristotle on Practical Wisdom] (Cambridge, Mass., 2013). Ross, W. D. (ed.), Aristotelis Ars Rhetorica (Oxford, 1959). Ross, W. D. (ed.), Aristotelis De Anima (Oxford, 1956). Ross, W. D. (ed.), Aristotelis Topica et Sophistici Elenchi (Oxford, 1958). Ross, W.  D. (ed.), Aristotle’s Prior and Posterior Analytics [Prior and Posterior Analytics] (Oxford, 1949). Ross, W. D. (trans.), ‘Nicomachean Ethics’, in id., The Complete Works of Aristotle: The Revised Oxford Translation, vol. ii (Oxford, 1984), 1729–867. Ross, W. D. (ed.), Parva Naturalia (Oxford, 1955). Ruiz Yamuza, E., Tres verbos que significan ‘deber’ en griego antiguo (Zaragoza, 2008). Schwab, W., ‘Explanation in the Epistemology of the Meno’, Oxford Studies in Ancient Philosophy, 48 (2015), 1–36. Schwab, W., ‘Understanding epistēmē in Plato’s Republic’, Oxford Studies in Ancient Philosophy, 51 (2016), 41–85. Sedley, D., ‘Aristotelian Relatives’, in M. Canto-Sperber and P. Pellegrin (eds.), Le style de la pensée: Recueil de textes en hommage à Jacques Brunschwig (Paris, 2002), 324–52. Solmsen, F., Die Entwicklung der Aristotelischen Logik und Rhetorik (Berlin, 1929). Studtmann, P., ‘Aristotle’s Category of Quality: A Regimented Interpretation’ [‘Aristotle’s Category of Quality’], Apeiron, 36 (2003), 205–27. Taieb, H., ‘Classifying Knowledge and Cognates: On Aristotle’s Categories, 8, 11a20–38 and its Early Reception’ [‘Classifying Knowledge and Cognates’], Documenti e studi sulla tradizione filosofica medievale, 27 (2016), 85–106. Taylor, C. C. W., ‘Aristotle’s Epistemology’, in S. Everson (ed.), Epistemology (Cambridge, 1990), 116–42. van Fraassen, B., ‘A Re-Examination of Aristotle’s Philosophy of Science’, Dialogue, 19 (1980), 20–45. Waterlow, S., Passage and Possibility: A Study of Aristotle’s Modal Concepts (Oxford, 1982). White, H., What Is What-Is? A Study of Parmenides’ Poem (New York, 2005). Williams, C. J. F., ‘Aristotle and Corruptibility’, Religious Studies, 1 (1965), 95–107. Williams, C. J. F., Aristotle’s De generatione et corruptione (Oxford, 1982). Yates, D. and Marmodoro, A., ‘Introduction: The Metaphysics of Relations’, in A. Marmodoro and D. Yates (eds.), The Metaphysics of Relations (Oxford, 2016), 1–18.

ARISTOTLE ON DIGESTION, SELF-­MOTION, AND THE ETERNITY OF THE UNIVERSE A Discussion of Physics 8. 6 and De Somno wei wang 1. Introduction Self-­motion is a crucial focal point in Aristotle’s natural philosophy and his engagement with Plato’s relevant theories. It is where two sets of fundamental questions on the nature of motion and change converge. The first set of questions concern the agency of mover. Plato’s view, as presented in the Phaedrus and Laws 10, is that a particular motion or change cannot be caused by an infinite series of moved movers, but there has to be a first, self-­moving mover: because all inanimate bodies are ultimately moved by an animate body, and an animate body in turn by its soul, Plato identifies such a self-­mover as soul (Phdr. 245 c 2–246 a 2; Laws 10, 894 b 8–896 b 2). In Physics 8. 5, Aristotle similarly argues that particular motions, at least some of them, are caused ultimately by something that moves itself (256a4–21).1 However, he then goes on to argue that because nothing can move itself in its entirety, there is always one part of the self-­mover that remains unmoved insofar as it causes the other part to move (257a33–258b9). This coheres with the theory of soul we find in De anima, and especially in 1. 3 (passim) and 1. 4, 408a34–b18, For helpful comments and criticism of various versions of this paper I would like to thank John Cooper, Benjamin Morison, Wei Cheng, and two anonymous referees of this journal. I thank Victor Caston for his detailed criticism and helpful suggestions, from which I learned immensely. My special thanks go to Christian Wildberg, who, with his invaluable support and encouragement, oversaw the entire project of which the current paper is a part. Portions of an earlier draft of the paper appeared in Chinese in Fudan Journal (Social Sciences), 316 (2021), 45–53. 1  Aristotle in Physics 8. 5 allows for the possibility that some motions are caused ultimately by unmoved movers that are external to the bodies that they immediately move; hence not all motions can be traced to something that moves itself. For a defence of this interpretation, see D. Blyth, ‘Heavenly Soul in Aristotle’, Apeiron, 48 (2015), 427–65 at 430–6. Wei Wang, Aristotle on Digestion, Self-­Motion, and the Eternity of the Universe: A Discussion of Physics 8. 6 and De somno In: Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Wei Wang 2023. DOI: 10.1093/oso/9780192885180.003.0006

234

Wei Wang

where Aristotle argues that soul, as such, is not susceptible to any kind of change at all.2 Admittedly, Aristotle never calls the unmoved part of a self-­mover ‘soul’ in Physics 8,3 but it is evident from his discussion of animal self-­motion in Physics 8. 6 that he takes the first mover of animal self-­motion to be its soul, which is unmoved per se, but moved per accidens by itself.4 More importantly, after he has offered his own definition of motion and defended it against his predecessors’ accounts in Physics 3. 1–2, Aristotle argues in 3. 3 that the fulfilment of the mover’s active potential is one and the same motion as the fulfilment of the passive potential of the thing moved. Hence, when an agent of motion or change acts on a patient, its action is nothing other than the motion undergone by the patient. This means that the model according to which an internal motion or change in the agent causes the motion or change in the patient is incorrect, and he thus rules out, from the point of view of causal agency, the possibility of there being a self-­moving 2  For a detailed analysis of Aristotle’s project in De anima and, in particular, how and why he thinks that soul is not moved, see S. Menn, ‘Aristotle’s Definition of Soul and the Programme of the De anima’ [‘Definition’], Oxford Studies in Ancient Philosophy, 22 (2002), 83–139. 3  The only place Aristotle mentions soul in Physics 8 is in a dialectical context, where he describes the theory of ‘those [i.e. the Platonists] who make [self-­moving] soul the cause of motion’ as follows: ‘Moreover to these we may add those who make soul the cause of motion; for they say that things that undergo motion have as their first principle that which moves itself; and when animals and all living things move themselves, the motion is motion in respect of place’ (8. 9, 265b32–266a1: ἔτι δὲ παρὰ τούτους οἱ τὴν ψυχὴν αἰτίαν ποιοῦντες κινήσεως· τὸ γὰρ αὐτὸ αὑτὸ κινοῦν ἀρχὴν εἶναί ϕασιν τῶν κινουμένων, κινεῖ δὲ τὸ ζῷον καὶ πᾶν τὸ ἔμψυχον τὴν κατὰ τόπον αὑτὸ κίνησιν). For the Greek text of the Physics, I use that of W. D. Ross, Aristotle’s Physics: A Revised Text with Introduction and Commentary [Physics] (Oxford, 1936). All translations of the Physics are by R. Hardie and R. Gaye, in J. Barnes (ed.), The Complete Works of Aristotle: The Revised Oxford Translation [Complete Works], vol. i (Princeton, 1984), with slight modifications. 4  Phys. 8. 6, 259b16–22: ‘But in all these cases [of self-­motion] the first mover— ­i.e. the cause of the animal’s being itself moved by itself—­is nonetheless moved per accidens: that is to say, the body changes its place, so that that which is in the body changes its place also and moves itself by leverage. From these [considerations] we may confidently conclude that if a thing belongs to the class of unmoved things that move themselves per accidens, it is impossible that it should cause continuous motion’ (ἐν πᾶσι δὲ τούτοις κινεῖται τὸ κινοῦν πρῶτον καὶ τὸ αἴτιον τοῦ αὐτὸ ἑαυτὸ κινεῖν ὑϕ’ αὑτοῦ, κατὰ συμβεβηκὸς μέντοι· μεταβάλλει γὰρ τὸν τόπον τὸ σῶμα, ὥστε καὶ τὸ ἐν τῷ σώματι ὂν καὶ τῇ μοχλείᾳ κινοῦν ἑαυτό. ἐξ ὧν ἔστιν πιστεῦσαι ὅτι εἴ τί ἐστι τῶν ἀκινήτων μὲν κινούντων δὲ καὶ αὑτὰ κατὰ συμβεβηκός, ἀδύνατον συνεχῆ κίνησιν κινεῖν). The soul is unmoved insofar as it causes motion but can be moved insofar as it is placed in its body, which fact is incidental to its causing motion; hence the soul is unmoved per se but moved by itself per accidens. Cf. the ‘pilot in the ship’ passage from DA 2. 1, 413a8–10.



Digestion, Self-Motion, and Eternity

235

first mover.5 So, while Aristotle also allows that there are self-­movers, he dis­agrees with Plato as to what the real first mover is and how it causes motion: while a Platonic soul moves other things by its own self-­motion, its Aristotelian counterpart does so as a per se motionless agent. The second set of questions, which again are closely connected with the problem of self-­motion, concerns the eternity of motion and the universe. The context in which Socrates brings up self-­ motion in the Phaedrus is an argument for the immortality of soul. He argues for the immortality of soul, both the divine and the human, on the assumption that whatever is always in motion is immortal (245 c 2–5): while body is that which moves and is moved by something else, and so has motion intermittently, soul by its very being is that which has motion because of itself, and so has motion eternally; therefore, every soul is eternal (245 e 2–246 a 2). In the midst of this argument, Socrates connects the immortality of self-­movers to the continuous motion and existence of the universe: if self-­ movers, as the only source of motion there is, were capable of being destroyed or generated, then ‘all heaven and every­thing that has been started up would collapse, come to a stop, and never have a cause to start moving again’ (245 d 8–e 2: πάντα τε οὐρανὸν πᾶσάν τε γένεσιν συμπεσοῦσαν στῆναι καὶ μήποτε αὖθις ἔχειν ὅθεν κινηθέντα γενήσεται).6 Similarly in Laws 10, after describing the source (ἀρχή) of all motions as self-­motion (895 a 1–3), the Athenian goes on to argue that if the universe (τὰ πάντα, 895 a 6) were to somehow ­coalesce and come to a stop, as most cosmogonists dare to hold, ‘self-­generating motion’ (τὴν αὐτὴν ἑαυτὴν δήπου κινοῦσαν [sc. κίνησιν], 895 b 1) would be the first to arise and be the source of all transmitted motions (895 a 6–b 7).7 Thus, the self-­moving soul both guarantees the continuity and eternity of cosmic motion and, in a hypothetical

5  For an analysis of the argument in Physics 3. 3 and its implications, see U. Coope, ‘Aristotle’s Account of Agency in Physics III. 3’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, 22 (2005), 201–27. 6  For the Greek text of both the Phaedrus and the Laws, I use Burnet’s OCT edition (Clarendon, 1901). For the English translation of the Phaedrus, I use that by A.  Nehamas and P.  Woodruff in J.  M.  Cooper (ed.), Plato, Complete Works (Indianapolis, 1997). I follow Nehamas and Woodruff’s adoption of γένεσιν instead of γῆν εἰς ἓν at 245 e 1. 7  For the connection between this particular argument and Aristotle’s discussion of self-­motion in Physics 8. 2 and 8. 6, see n. 13 below.

236

Wei Wang

total cosmic rest, would be the first cause and trigger of all other motions. Aristotle’s task in Physics 8 is to argue for the eternity of motion and to prove the existence and explain the function of an eternal and numerically single prime mover of the universe—­God.8 So it comes as no surprise that his discussion there shares thematic simi­ lar­ity and internal reference to the discussions we find in the Phaedrus and the Laws. Beyond criticizing, as we have seen, the Platonic self-­mover as a first cause of motion in Physics 8. 5,9 Aristotle also examines the case of self-­movers with regard to the eternity of motion and the uniqueness of the prime mover. It is in this context that the difficulty we are concerned with in 8. 6, the problem of animal self-­motion and motion ex nihilo, first arises.

2.  Physics 8. 6: The difficulty and its context The difficulty is a long-­standing one, and it concerns a particular part of Physics 8. 6. Aristotle seems to deny outright in 259b1–20 that animals are self-­movers by saying that they are externally moved. This apparent denial would be problematic in two ways: first, it contradicts the well-­known doctrine shared by Plato and Aristotle that animate motion or change is due to the living being itself or, more precisely, its soul.10 Second, it conflicts with Aristotle’s statement in the very same chapter that there exist animate self-­movers (259b2–3) and that they have within them prin­ciples that are 8  There is some disagreement over which of the two should have precedence. If Aristotle establishes the eternity of motion in order to posit the existence of a single mover, Physics 8 would be theology. This is, for example, Thomas Aquinas’ take on Physics 8 (see his commentary on Aristotle’s Physics at section §965). But if the aim in analysing the single unmoved mover is to explain the eternity of motion, Physics 8 would be natural philosophy (see e.g. Ross, Physics, 85). I don’t take a position on this issue. 9 It is important to note that Aristotle’s criticism of Platonic self-­movers in Physics 8. 5 is not restricted to the first mover of the universe. It is a general claim about the agency of mover that the first mover of any series of motions has to be itself unmoved. It leaves open the question whether all series terminate in a single unmoved mover or not, and for Aristotle immortality does not follow from the fact that a mover is unmoved insofar as it causes motion (e.g. Socrates’ soul). 10  See e.g. Phdr. 245 c 2–246 a 2, discussed above, and DA 2. 2, 413a21–5. Despite their different views on how soul causes motion, they both think that soul is a principle of motion that is internal to its body.



Digestion, Self-Motion, and Eternity

237

unmoved and unaffected by anything else insofar as they cause motion (259a27–b1; 259b17–20). A lot of ink has been spilled over this difficulty and the more general problem of self-­motion in recent years.11 Now briefly, regarding the context of the passage and why Aristotle raises the issue of animal self-­motion in Physics 8. 6, Aristotle’s aim in the first two chapters of Physics 8 is to argue for the continuity and eternity of motion. In 8. 1, he argues that there is always motion before any temporally finite motion and that there is always an extended temporal continuum before and after a given temporal point, so both motion and time are infinite and continuous.12 In 8. 2, he considers three possible counter-­arguments against this conclusion: (1) that all motions are motions towards a contrary, and therefore cannot go on ad infinitum; (2) that the start of motion from rest or vice versa is impossible if the cause itself is always in motion; and (3) that animate (ἔμψυχα) self-­movers setting themselves in motion from absolute rest might lead to positing motion ex nihilo for the universe. All three counter-­arguments are answered in 8. 2 and then examined again in the latter half of the book. I quote the third counter-­argument here almost in its entirety: πολὺ δὲ μάλιστα τὸ τοιοῦτον ἐπὶ τῶν ἐμψύχων εἶναι ϕανερόν· οὐδεμιᾶς γὰρ ἐν ἡμῖν ἐνούσης κινήσεως ἐνίοτε, ἀλλ’ ἡσυχάζοντες ὅμως κινούμεθά ποτε, καὶ ἐγγίγνεται ἐν ἡμῖν ἐξ ἡμῶν αὐτῶν ἀρχὴ κινήσεως, κἂν μηθὲν ἔξωθεν κινήσῃ. . . . τὸ δὲ ζῷον αὐτό ϕαμεν ἑαυτὸ κινεῖν. ὥστ’ εἴπερ ἠρεμεῖ ποτὲ πάμπαν, ἐν ἀκινήτῳ κίνησις ἂν γίγνοιτο ἐξ αὑτοῦ καὶ οὐκ ἔξωθεν. εἰ δ’ ἐν ζῴῳ τοῦτο δυνατὸν γενέσθαι, τί κωλύει τὸ αὐτὸ συμβῆναι καὶ κατὰ τὸ πᾶν; εἰ γὰρ ἐν μικρῷ κόσμῳ γίγνεται, καὶ ἐν μεγάλῳ· καὶ εἰ ἐν τῷ κόσμῳ, κἀν τῷ ἀπείρῳ, εἴπερ ἐνδέχεται κινεῖσθαι τὸ ἄπειρον καὶ ἠρεμεῖν ὅλον. (8. 2, 252b17–28) 11  D. Furley, ‘Self-­Movers’, in G. E. R. Lloyd and G. E. L. Owen (eds.), Aristotle on the Mind and on the Senses (Cambridge 1978), 165–79; M. Nussbaum, Aristotle’s De motu animalium [De motu] (Princeton, 1978); M.  L.  Gill, ‘Aristotle on Self-­ Motion’ [‘Self-­Motion’], in M. L. Gill and J. G. Lennox (eds.), Self-­Motion: From Aristotle to Newton (Princeton 1994), 15–34; S.  S.  Meyer, ‘Self-­Movement and External Causation’ [‘Self-­Movement’], in M. L. Gill and J. G. Lennox (eds.), Self-­ Motion: From Aristotle to Newton (Princeton 1994), 65–80; B.  Morison, ‘Self-­ Motion in Physics VIII’ [‘Self-­Motion’], in A. Laks and M. Rashed (eds.), Aristote et le mouvement des animaux: Dix études sur le De motu animalium (Villeneuve d’Ascq, 2004), 67–79. 12  The main views that Aristotle argues against here are (1) Anaxagorean: that there is an infinite period of total rest before motion (250b24–6); (2) Empedoclean: that there is some period of total rest between the cycles of love and strife (250b26–9); and (3) Platonic: that the cosmos and time have a beginning (251b17–19).

238

Wei Wang

The fact is evident above all in the case of animate beings; for it sometimes happens that there is no motion in us and we are quite still, and that never­ the­less we are then at some moment set in motion, that is to say, it sometimes happens that we produce a beginning of motion in ourselves spontaneously without anything having set us in motion from without. . . .  The animal, we say, moves itself; therefore, if an animal is ever in a state of absolute rest, we have a motionless thing in which motion can be produced from the thing itself, and not from without. Now if this can occur in an animal, why should not the same be true also of the universe [τὸ πάν]? If it can occur in a small world, it could also occur in a great one; and if it can occur in the world, it could also occur in the infinite, that is, if the infinite could as a whole possibly be in motion or at rest.

Aristotle responds to this counter-­argument in two places: first immediately at Physics 8. 2, 253a7–21 and then at 8. 6, 259b1–20 (both quoted below).13 The overall aim of Physics 8. 6 is to prove the eternity and uniqueness of the first unmoved mover of the universe. After two a priori arguments, in which the eternity and continuity of motion are taken as given, Aristotle in 259a20–b20 turns to considerations concerning the causal mechanism at work within movers, especially those animate beings that appear, as a matter of experience, to be self-­movers. The discussion takes the form of a response to the third counter-­argument’s suggestion that the behaviour of the animate movers offers a model of motion ex nihilo, which might be applicable to the bigger ‘cosmos’, the universe, and so render the first motion (i.e. the motion of the first heaven) non-­ eternal. The implication is that if the first motion were non-­ eternal, the prime mover of the universe need not be eternal or at least eternally acting either. It is this implication that is the reason why Aristotle returns to the topic of animal self-­motion in Physics 8. 6.

13  Solmsen is the first to relate the third counter-­argument to Laws 10, 895 a 6–b 7, a passage we have discussed above. See  F.  Solmsen, ‘Plato’s First Mover in the Eighth Book of Aristotle’s Physics’, in R. B. Palmer and R. Hammerton-­Kelly (eds.), Philomathes: Studies and Essays in the Humanities in Memory of Philip Merlan (The Hague, 1971), 171–82 at 172–3. It is noteworthy that, for Plato in the Laws, the fact that the self-­moving soul can initiate motion from total cosmic rest is an indication of its priority and power, whereas, for Aristotle in Physics 8, it is an indication that it is not capable of causing eternal motion.



Digestion, Self-Motion, and Eternity

239

3.  The passage (Physics 8. 6, 259b1–20) I now quote his response in Physics 8. 6 in its entirety and offer a preliminary discussion: (1) ὁρῶμεν δὲ καὶ ϕανερῶς ὄντα τοιαῦτα ἃ κινεῖ αὐτὰ ἑαυτά, οἷον τὸ τῶν ἐμψύχων καὶ τὸ τῶν ζῴων γένος, ταῦτα δὲ καὶ δόξαν παρεῖχε μή ποτε ἐνδέχεται κίνησιν ἐγγίγνεσθαι μὴ οὖσαν ὅλως, διὰ τὸ ἐν τούτοις ὁρᾶν ἡμᾶς τοῦτο συμβαῖνον (ἀκίνητα γάρ ποτε ὄντα κινεῖται πάλιν, ὡς δοκεῖ), τοῦτο δὴ δεῖ λαβεῖν, ὅτι μίαν κίνησιν αὑτὰ κινεῖ, καὶ ὅτι ταύτην οὐ κυρίως· οὐ γὰρ ἐξ αὐτοῦ τὸ αἴτιον, (2) ἀλλ’ ἔνεισιν ἄλλαι κινήσεις ϕυσικαὶ τοῖς ζῴοις, ἃς οὐ κινοῦνται δι’ αὑτῶν, οἷον αὔξησις ϕθίσις ἀναπνοή, ἃς κινεῖται τῶν ζῴων ἕκαστον ἠρεμοῦν καὶ οὐ κινούμενον τὴν ὑϕ’ αὑτοῦ κίνησιν. τούτου δ’ αἴτιον τὸ περιέχον καὶ πολλὰ τῶν εἰσιόντων, οἷον ἐνίων ἡ τροϕή· πεττομένης μὲν γὰρ καθεύδουσιν, διακρινομένης δ’ ἐγείρονται καὶ κινοῦσιν ἑαυτούς, τῆς πρώτης ἀρχῆς ἔξωθεν οὔσης, (3) διὸ οὐκ ἀεὶ κινοῦνται συνεχῶς ὑϕ’ αὑτῶν· ἄλλο γὰρ τὸ κινοῦν, αὐτὸ κινούμενον καὶ μεταβάλλον πρὸς ἕκαστον τῶν κινούντων ἑαυτά. ἐν πᾶσι δὲ τούτοις κινεῖται τὸ κινοῦν πρῶτον καὶ τὸ αἴτιον τοῦ αὐτὸ ἑαυτὸ κινεῖν ὑϕ’ αὑτοῦ, κατὰ συμβεβηκὸς μέντοι· μεταβάλλει γὰρ τὸν τόπον τὸ σῶμα, ὥστε καὶ τὸ ἐν τῷ σώματι ὂν καὶ τῇ μοχλείᾳ κινοῦν ἑαυτό. (259b1–20) (1) In fact, we see clearly that there are things of such a sort as to move themselves, for example the genus of the animate things and that of animals. They suggest the thought that perhaps it is possible for motion to come to be in a thing without having been in existence at all before, because we see this actually occurring in them [sc. animals]: when they are not in motion, sometimes they move themselves again, as it seems. We must grasp the fact, therefore, that they move themselves only with one kind of motion, and not with that motion, strictly speaking, for the cause is not from itself; (2) there are other natural motions in animals in which they are not self-­ movers, for example increase, decrease, and respiration, with which each kind of animal is moved when they are at rest and not being moved by the motion by which they move themselves. But the cause of this is the surroundings and the many things that come in. In some cases, the cause is food: when it is being digested, animals sleep, and when it is being separated,14 they awaken and move themselves, the original starting point (viz. the food ingested) being from outside. (3) That is why animals are not always moved continuously by themselves: it is something else that moves them, being itself in motion and changing as it comes into relation with each thing that moves itself. But in all these cases [of self-­motion] the 14  That is, being separated into the more corporeal and the purer sort of blood that is suitable for nourishment of the organs and other constituent parts of the body. For this interpretation of what Aristotle is referring to here by διακρινομένης, see Section 6 below. Most translators incorrectly render this as ‘distributing’, namely, of nourishment from the heart to the various parts of the body.

240

Wei Wang

first mover—­i.e. the cause of the animal’s being itself moved by itself—­is nonetheless moved per accidens, that is to say, the body changes its place, so that that which is in the body changes its place also and moves itself by leverage.15

The behaviour of animals does not pose a threat to the eternity of the motion of the universe, Aristotle explains: animals move themselves only with one kind of motion,16 and they do not move themselves with this kind of motion ‘kuriōs’, or strictly speaking: what causes the motion (τὸ αἴτιον, 259b7–8) is not from within the animal itself. Hence, animals, strictly speaking, do not initiate their own motion and rest; therefore, they cannot be the model for motion ex nihilo, whose application to the universe and especially to the first heaven would make its motion non-­eternal. That something like this has to be the general thrust of the argument no one disputes; however, as will become clear in Section 4, there is much disagreement as to exactly how the argument works, especially how part (2) hangs together with the rest. This has important ramifications for the question of whether Aristotle believes that animals are ­genuine self-­movers or not, because his argument builds on the claim that ‘animals do not move themselves with that motion kuriōs’ (259b7). What does he mean in saying this?

4.  Three interpretations and their shortcomings Three lines of interpretation have been suggested. First, the simple reading, suggested by Hardie and Gaye and Ross, is to take part (2) as explanatory of part (1).17 This reading connects locomotion 15  Although the translation is largely based on that of Hardie and Gaye (Barnes, Complete Works, i, 433–4), I often depart from their version. 16  Locomotion, as we gather from the parallel passage at Phys. 8. 2, 253a14–15. For the relationship between the 8. 2 passage and the 8. 6 passage, see Section 4 below. 17  See Hardie and Gaye’s translation (I refer here to the unrevised version before Barnes’s edition (W. D. Ross (ed.), The Works of Aristotle, vol. ii: Physica, trans. by R.  Hardie and R.  Gaye (Oxford, 1930)): ‘the cause of it is not derived from the animal itself: it is connected with other natural motions in animals, which they do not experience through their own instrumentality’ (259b7–9). Ross in his paraphrase (Ross, Physics, 442) suggests the same reading: he takes the ‘this’ in ‘the cause of this’ (259b11: τούτου δ’ αἴτιον) and the motion that ‘the first principle’ (τῆς πρώτης ἀρχῆς, 259b13–14) governs to refer to all kinds of animal motion including locomotion. See Morison’s summary of the ‘orthodox interpretation’ in Morison, ‘Self-­ Motion’, 75–6. See also Nussbaum, De motu, 119, on Hardie and Gaye’s translation.



Digestion, Self-Motion, and Eternity

241

to the ‘other natural motions’ (ἄλλαι κινήσεις ϕυσικαί, 259b8) such as increase, decrease, respiration, and digestion, which are caused externally by ‘the surroundings and the many things that come into [the animal]’ (τὸ περιέχον καὶ πολλὰ τῶν εἰσιόντων, 259b11–12). These externally caused motions are taken as an intermediary that explains why locomotion is also externally caused and animals are not self-­movers at all. There are two ways in which such a connection between locomotion and the other motions might be spelled out.18 First, locomotion may be thought of as directly caused by these other motions, and so indirectly by things external. This is less obvious in the 8. 6 passage than in its parallel passage at 8. 2, 253a7–21, where Aristotle claims that some of the externally caused motions set the intellect or desire in motion, and intellect or desire in turn moves the animal as a whole: μάλιστα δ’ ἂν δόξειεν τὸ τρίτον ἔχειν ἀπορίαν, ὡς ἐγγιγνομένης οὐκ ἐνούσης πρότερον κινήσεως, τὸ συμβαῖνον ἐπὶ τῶν ἐμψύχων· ἠρεμοῦν γὰρ πρότερον μετὰ ταῦτα βαδίζει, κινήσαντος τῶν ἔξωθεν οὐδενός, ὡς δοκεῖ. τοῦτο δ’ ἐστὶ ψεῦδος. ὁρῶμεν γὰρ ἀεί τι κινούμενον ἐν τῷ ζῴῳ τῶν συμϕύτων· τούτου δὲ τῆς κινήσεως οὐκ αὐτὸ τὸ ζῷον αἴτιον, ἀλλὰ τὸ περιέχον ἴσως. αὐτὸ δέ ϕαμεν αὑτὸ κινεῖν οὐ πᾶσαν κίνησιν, ἀλλὰ τὴν κατὰ τόπον. οὐδὲν οὖν κωλύει, μᾶλλον δ’ ἴσως ἀναγκαῖον, ἐν μὲν τῷ σώματι πολλὰς ἐγγίγνεσθαι κινήσεις ὑπὸ τοῦ περιέχοντος, τούτων δ’ ἐνίας τὴν διάνοιαν ἢ τὴν ὄρεξιν κινεῖν, ἐκείνην δὲ τὸ ὅλον ἤδη ζῷον κινεῖν, οἷον συμβαίνει περὶ τοὺς ὕπνους· αἰσθητικῆς μὲν γὰρ οὐδεμιᾶς ἐνούσης κινήσεως, ἐνούσης μέντοι τινός, ἐγείρεται τὰ ζῷα πάλιν. ἀλλὰ γὰρ ϕανερὸν ἔσται καὶ περὶ τούτων ἐκ τῶν ἑπομένων. (253a7–21) The third objection may be thought to present more difficulty than the others, namely, that which alleges that motion arises in things in which it did not exist before, and adduces in proof the case of animate things; thus, an animal is first at rest and afterwards walks, not having been set in motion apparently by anything from without. This, however, is false; for we observe that there is always some part of the animal’s organism in motion, and the cause of the motion of this part is not the animal itself but, it may be, the surroundings. Moreover, we say that the animal itself originates not all of its motions but its locomotion. So nothing prevents it from being the case—­or, rather, perhaps it must be the case—­that many motions are produced in the body by its surroundings, and some of these set in motion the intellect or desire, and this again then sets the whole animal in motion. This is what happens when animals are asleep: though 18  Given the nature of translation and paraphrase, it is unclear which way Hardie and Gaye or Ross opts for.

242

Wei Wang

there is then no perceptive motion in them, there is some motion that causes them to wake up again. But we will leave this point also to be elucidated at a later stage in our discussion.

As we can see from the end of the passage, the discussion here is presented as a sketch, looking forward to what we find in Physics 8. 6, 259b1–20. The Physics 8. 6 passage contains a more detailed description of what the externally caused motions are, and it identifies food as a possible external cause of an animal’s nutritive and perceptive activities. Moreover, it states more explicitly that the animals do not move themselves in locomotion, strictly speaking. The Physics 8. 2 passage is less explicit about this, and in general more tentative in its tone (e.g. ἴσως, 253a13), but by saying that the intellect or desire ‘moves the animal as a whole’ (τὸ ὅλον ζῷον κινεῖν, 253a18), Aristotle can only mean locomotion. (On this final point, see e.g. DA 3. 10, 433a17–21.) Second, locomotion may be put on a par with the other natural motions or changes. Thus, just as in the case of these other natural changes, the cause is not to be found in the animals themselves, but rather the cause is external, so in the case of locomotion the cause is external as well.19 When Aristotle claims that animals are not, strictly speaking, self-­movers even in their own locomotion because the cause is not from itself, he means that the cause is external just as the causes of the other natural motions are external, and indeed when he says ‘the first principle being thus originally from outside’ (τῆς πρώτης ἀρχῆς ἔξωθεν οὔσης, 259b13–14), he means by ‘the first principle’ that of both the other motions and locomotion. One way or the other, this reading, through the intermediary of the other natural motions, makes animal locomotion externally caused, and it leaves no room for animals to be genuine self-­movers. Thus, when Aristotle says that animals do not move themselves with locomotion kuriōs, he means that animals might appear to be self-­movers in their locomotion, but in reality, they are only externally moved. The major shortcoming this reading faces, therefore, is that it renders nonsense what Aristotle says in the very same chapter, namely, that there exist animate self-­movers (259b2–3) and that they have

19  Note the strong parallelism in construction between οὐ γὰρ ἐξ αὐτοῦ τὸ αἴτιον (259b7–8), ἃς οὐ κινοῦνται δι’ αὑτῶν (b8–9), and καὶ οὐ κινούμενον τὴν ὑϕ’ αὑτοῦ κίνησιν (b10–11). I thank Victor Caston for bringing up this parallel.



Digestion, Self-Motion, and Eternity

243

within them principles that are unmoved and unaffected by anything else insofar as they cause motion or change (259a27–b1; 259b16–20). Two subtler ways of interpreting the passage, which underscore the contrast between locomotion and the other natural changes, have been suggested more recently.20 Aristotle either (a) doesn’t explain at all the grounds for his claim that animals are not self-­movers strictly speaking, or (b) only very briefly, and so makes two separate points in parts (1) and (2). Thus, (a) Nussbaum thinks that the reason why animals are not self-­movers, strictly speaking, is ‘undeveloped’.21 The real reason why animals are not self-­movers, strictly speaking, even in their locomotion, according to Nussbaum’s Aristotle, lies in the fact that they are moved locally by the external objects of thought or desire, which point is only developed in De anima and De motu animalium.22 Nussbaum’s view is shared by Furley, among others. Whereas Nussbaum is not explicit about how external objects explain animal motion and how exactly to understand the ‘not strictly speaking’ claim, Furley interprets it in a weaker sense in which the animals can be said to move themselves. According to  Furley’s interpretation, which may be called the ‘intentional reading’, when Aristotle says that animals move themselves only with one kind of motion, and with this ‘not strictly speaking’ or ou kuriōs, he means that an animal’s self-­locomotion is, strictly speaking, caused by the intentional object, which is outside the animal in question, but loosely speaking, because of the ambivalent status of intentional objects, which under some description can be said to be

20  These interpretations emphasize the contrastive nature of the ἀλλά (‘but’) at 259b8 between part (1) and (2), which both Hardie and Gaye’s translation and Ross’s paraphrase omit. 21  See Nussbaum, De motu, 119: ‘In fact, Aristotle seems to be making two sep­ ar­ate points [at Phys. 8. 6, 259b1–16]: (1) Local motion is the only genuine self-­ motion; but even this is not strictly self-­motion, since it depends on external aition. (2) In the case of the other natural motions (growth, decay, respiration, etc.), there is no reason to think any is genuine self-­motion. . . . (2) is the same point already made in chapter 2 [i.e. Physics 8. 2]; it will be repeated and developed in MA chapter 11, . . . (1) is obscure and undeveloped’ (emphasis original). 22  See Nussbaum, De motu, 19–20, in a continuation of the passage quoted in n. 21 above: ‘What is the external aition? And in what sense, given that there is this external aition, can animals be said to be self-­moving? The Physics provides no further answers. Its arguments have gaps that can be filled only by an adequate account of animal motion and its relationship to external goals and external ne­ces­ sities. The MA undertakes to provide such an account, grounded in the specific data of the preceding biological works’ (emphasis mine).

244

Wei Wang

internal to the mind, animals can be said to move themselves.23 I respond to Furley’s in­ter­pret­ation in Section 7 below.24 (b) Morison, who follows Nussbaum’s division of points but interprets them very differently,25 offers an interpretation of the passage that may be called the ‘bipartite reading’. Morison thinks that the reason why according to Aristotle animals are not self-­movers, strictly speaking, is stated very succinctly in the clause ‘for the cause is not from itself’ (οὐ γὰρ ἐξ αὐτοῦ τὸ αἴτιον, 259b7–8): according to Aristotle’s account of self-­movers in Physics 8. 5, the cause of self-­motion is not the animal as a composite whole, but only a part of it is, that is, the unmoved part—­its soul. So, strictly speaking, animal as a soul–­body composite does not move itself, because there is one part of the composite that does not cause motion. Further, the mover within a composite animal, its soul, is a self-­mover not per se but only per accidens, because, although it per se moves the moved part (the body), it undergoes that motion only in that and because it is situated within the moved part, which fact is incidental to its being a mover (Morison, ‘Self-­Motion’, 73–5). Thus, when Aristotle says ‘the cause [of the self-­motion] is not from itself’, his claim is not that the cause is from outside the animal, as Hardie and Gaye, Ross, Nussbaum, and Furley think, but that the cause is not from the animal as a whole; rather, the cause is from only part of it—­its soul (‘Self-­Motion’, 74). Interestingly, Aristotle’s claim here, thus interpreted, looks like a separate criticism of the Platonic model of the self-­moving soul. A Platonic soul would, in Aristotle’s terminology, be a self-­mover per se, so there will be no further cause that explains its motion, and 23  See Furley, ‘Self-­Movers’, 170, 179 n.14, but especially 176–7: ‘An animal is correctly described as a self-­mover, because when it moves, its soul moves its body, and the external cause of its motion (the ὀρεκτόν) is a cause of motion only because it is “seen” as such by a faculty of the soul. There must be an external object, however, and hence the movement of an animal does not provide an example of a totally autonomous beginning of motion’ (emphasis mine). 24 Gill, in her account of Aristotle’s conception of self-­motion, distinguishes between (1) the inherent characteristics of animal motion such as its sort, direction, and goal and (2) the start and end of animal motion, which require ‘trigger’ and ‘impetus’, which are found in the organism’s environment. Given her endorsement of Furley’s solution, she apparently thinks what triggers animal motion is the external objects of desire (see Gill, ‘Self-­Movers’, 15–6, and especially 15 n. 1, where she also cites Nussbaum, De motu, 107–42). 25 For Nussbaum’s division of points, see n. 21 above. Cf. Morison, ‘Self-­ Motion’, 72. Morison proposes a similar change of punctuation—­a full stop before ἀλλά at 259b8—as does Nussbaum (De motu, 119). Both of them also translate ἀλλά as ‘but’, departing from Hardie and Gaye and Ross.



Digestion, Self-Motion, and Eternity

245

it will cause motion eternally. An Aristotelian soul, by contrast, is only a self-­mover per accidens, and the ensouled animal not, strictly speaking, a self-­mover. So according to Morison’s interpretation of the passage, by saying that ‘the cause is not from itself’, Aristotle is just making this anti-­Platonic point. There is no doubt, as I have explained in Section 1 above, that Aristotle does hold this view against Plato. But still, one may ask, why does he need to make this point here, and in such a telegraphic way? The advantage of Morison’s interpretation, as we can see, is that it preserves Aristotle’s commitment to the view that animals are genuine self-­movers. However, it comes at the cost of misconstruing the flow of the argument. The reason why ‘it is necessary to grasp’ (259b6) the fact that animals move themselves only with self-­ locomotion and do not move with that motion, strictly speaking, is not only that it rules out the per accidens self-­movers as causes of the continuous motion of the universe. More importantly, if we read the text more closely,26 the fact is meant to respond to the problem that motivates the passage, as I outlined in Section 2, that the behaviour of animals in their self-­motion may lead us to posit motion ex nihilo for the universe. Yet if the fact necessary for us to grasp, as Morison would have it, is that the animal soul moves itself only per accidens, it is difficult to see how this rules out animal self-­motion as a model for positing motion ex nihilo for the universe. It is not on the face of it impossible for the animal soul to move its body from absolute rest and itself by leverage. Indeed, the contrast between Plato and Aristotle’s views on soul seems to invite the thought that whereas a Platonic soul eternally moves itself, an Aristotelian soul only does so intermittently because of the in­tern­al hindrance of its body, so it has to set its body in motion from rest time and time again. Therefore, in order to respond to the motivating problem of the passage, animal self-­motion has to be, in one way or another, qualified externally, and it is precisely this point that Morison aims to deny (‘Self-­Motion’, 76). 5.  A new interpretation and its evidence in Physics 8 Three ways of understanding the ou kuriōs claim at 259b8 are now on the table. (1) According to the simple reading, Aristotle means 26  Note that the δή at 259b6 picks up the puzzle of motion ex nihilo at b3–6.

246

Wei Wang

that animals only appear to be self-­movers in their locomotion, but in reality they are not: just as the other internal natural motions, locomotion is caused externally. (2) According to the intentional reading, Aristotle means that animals are, strictly speaking, moved by external objects of thought and desire, but loosely speaking, because these objects need to be perceived and desired by the soul, animals can be said to be moved internally by their souls. Thus, the externality of the other natural motions has nothing to do with an animal being externally caused in its locomotion. (3) According to the bipartite reading, Aristotle means that, strictly speaking, the animal as a soul-­body composite does not move itself: it is a part of the composite that moves the other part and itself by leverage. Since soul is internal to the animal, Aristotle’s ou kuriōs claim here has nothing to do with external causation. I suggest a fourth way of interpreting the ou kuriōs claim, which may be called the ‘waking reading’. Although animals in their locomotion do decide where they would like to go—­in this and similar respects they qualify as self-­movers—­according to Aristotle they make such decisions only when they are awake and conscious, and their being conscious or not is not a decision that they themselves make: ‘the cause’ of an animal being able to move at all ‘is not from the animal itself’. It is precisely because locomotion is causally contingent on being awake that it does not belong to animals simpliciter and hence not kuriōs.27 It is true that Aristotle does contrast animal locomotion with other animate motions and changes that are externally caused. Nonetheless, according to Aristotle, locomotion is causally contingent on some of these externally caused changes, and in particular, falling asleep and waking up. This is confirmed by the parallel passage in Physics 8. 2 (quoted more fully above):

27  Gill, as noted earlier, distinguishes between (1) the inherent characteristics of animal motion such as its sort, direction, and goal and (2) the start and end of animal motion. She argues that the inherent characteristics are due to an animal’s internal ‘active potency’, whereas the start and end of animal motion are due to external objects of desire (‘Self-­Movers’, 15). She further argues that the motion the animal undergoes with respect to its sort and direction and goal counts as self-­ motion, whereas the active capacity of desire counts as unmoved mover (‘Self-­ Movers’, 16–24). I agree with Gill that these inherent characteristics are internally caused, but not that the start and end of animal motion are due to external objects of desire, i.e. intentional objects.



Digestion, Self-Motion, and Eternity

247

οὐδὲν οὖν κωλύει . . . τούτων [sc. πολλῶν κινήσεων] δ’ ἐνίας τὴν διάνοιαν ἢ τὴν ὄρεξιν κινεῖν, ἐκείνην δὲ τὸ ὅλον ἤδη ζῷον κινεῖν, οἷον συμβαίνει περὶ τοὺς ὕπνους· αἰσθητικῆς μὲν γὰρ οὐδεμιᾶς ἐνούσης κινήσεως, ἐνούσης μέντοι τινός, ἐγείρεται τὰ ζῷα πάλιν. (253a15–20) So nothing prevents it from being the case that . . . some of these [many externally caused changes in the animal] set in motion [κινεῖν] the intellect or desire, and this again then sets the whole animal in motion [κινεῖν]. This is what happens when animals are asleep: though there is then no perceptive change in them, there is some change that causes them to wake up again.

In the Physics 8. 6 passage, Aristotle again outlines this causal relationship, with special emphasis on its externality (quoted more fully above): τούτου δ’ αἴτιον τὸ περιέχον καὶ πολλὰ τῶν εἰσιόντων, οἷον ἐνίων ἡ τροϕή· πεττομένης μὲν γὰρ καθεύδουσιν, διακρινομένης δ’ ἐγείρονται καὶ κινοῦσιν ἑαυτούς, τῆς πρώτης ἀρχῆς ἔξωθεν οὔσης, διὸ οὐκ ἀεὶ κινοῦνται συνεχῶς ὑϕ’ αὑτῶν. (259b11–15) But the cause of this [sc. sort of change] is the surroundings and the many things that come in. In some cases [ἐνίων] the cause is food: when it is being digested, animals sleep, and when it is being separated, they awaken and move themselves, the original starting point being from outside [τῆς πρώτης ἀρχῆς ἔξωθεν οὔσης]. That is why animals are not always moved continuously by themselves.

According to the example given in this passage, food which enters an animal body from without is the first and external cause of some changes (ἐνίων) in the animal body. What sort of change does food cause? According to Aristotle’s statement here, it seems that the entrance of food into the animal body, which gives rise to the nutritive activities of the digestion (πέψις, lit. ‘cooking up’) of the ingested food and the separation (διάκρισις) of the product—­blood—­into its purer and less pure parts, causes the animal in question to go to sleep and wake up and be able to self-­move. Food is here a sunaition (lit. ‘co-­cause’) or, in other words, a condition sine qua non of the animal’s motion and rest because it causes both sleep, which is an impediment to consciousness and intentional behaviour, and waking, which is the removal of such impediment.28 Therefore, 28  For other examples of such sunaitia, see (1) DA 2. 4, 416a9–18, where fire is described as a sunaition (a14) of digestion, and (2) Metaph. Δ. 5, 1015a20–2, where breathing and food are said to be ‘necessary’ because they are sunaitia (a21) and conditions sine qua non of life. See also n. 38 below on material causation.

248

Wei Wang

food, a proximate cause of an animal’s nutritive ac­tiv­ities, is a sunaition of the animal’s perceptive and locomotive states precisely because the perceptive and locomotive activities of an animal are contingent on periods of wakefulness. The waking reading differs from the simple reading in that it contrasts locomotion with the other motions and changes going on in an animal that are caused by external things directly, and thus leaves room for the animal to be a self-­mover in some qualified way, that is, when it is awake and is able to exercise its locomotive cap­ abil­ity. It differs from the intentional reading and the bipartite reading in that it still causally connects the two kinds of motions and understands food as a condition sine qua non for animal locomotion. Thus, an animal fails to be self-­mover kuriōs in its locomotion, neither because it is moved by its external object of thought and desire nor because it is a soul-­body composite that doesn’t move itself in its entirety, but because its perception and locomotion are contingent on intermittent periods of wakefulness, which depend on the ingestion of food and its digestion. Ever since Simplicius (In Phys. 1258. 4–­30 Diels), interpreters have noticed the correlation between digestion and locomotion. This comes as no surprise, as it is after all what both passages literally say. Simplicius identifies food as the cause of sleep and waking (19–20) and as the reason why animals do not move themselves continuously (29–30). Thus, he interprets the ou kuriōs claim as having to do with the fact that locomotion is causally related to the other motions in animals that are externally caused (12–17). The fact that he explains ‘strictly’ with ‘in every way’ (πάντῃ, 14) shows that he thinks, as I do, that animals fail to be self-­movers because their locomotion is not solely a matter of their internal condition, but they are self-­movers in some qualified way. What is surprising though, to a modern reader, is that Simplicius does not seem sensitive to the conflict between internal and external causation his interpret­ ation might imply. Just like Aristotle, Simplicius talks about animal ‘self-­motion’ (24) and the external origin or archē (15 and 25) of animal motion in close vicinity. And because he describes food generically as the archē (15) of animal motion and the cause or aitia (20), his position on the conflict would seem rather close to the simple reading: animals only appear to be self-­movers in their locomotion, but in reality they are not. In my version of the ‘waking reading’, by contrast, food is specified only as a sunaition and



Digestion, Self-Motion, and Eternity

249

condition sine qua non of animal locomotion, and this leaves conceptual room for animal motion to have its first origin in soul, on the one hand, and be contingent on food and its digestion, on the other.29 My interpretation of the two passages in Physics 8. 6 can be further corroborated by Aristotle’s remarks on sleep and digestion in De somno et vigilia (henceforth De somno).30

6.  De somno on sleep and digestion If what is referred to by Aristotle as his On Nourishment (Περὶ τροϕῆς) had survived, it would be a better place to go for his theory of digestion.31 Unfortunately it didn’t, and given what we do have, De somno contains Aristotle’s most detailed account.32 That said, my focus on De somno is not entirely contingent upon textual trad­ition, 29 In the contemporary literature, S.  S.  Meyer offers a similar, Simplician account of the passage (Meyer, ‘Self-­Movement’, 68–71). She correctly points out, against Furley, that the ou kuriōs claim doesn’t concern the external objects of motion, but digestion. She also correctly identifies Aristotle’s motivation in making the claim, i.e. his worry that animal self-­motion might be taken as a model for motion ex nihilo (‘Self-­Movement’, 70). But she doesn’t give a precise account of the relationship between the changes due to digestion and the intentional states that cause locomotion, because her real interest lies in understanding the role external objects of desire play in voluntary action and motion. According to Meyer, an external object of desire only counts as an accidental cause of the agent’s action, because it doesn’t ‘reliably precipitate’ a desire in the agent (‘Self-­Movement’, 79), whereas it is the agent’s character that ‘naturally produces’ certain results (‘Self-­ Movement’, 78). 30  Graham and McKirahan mention De somno 3 in connection with the Physics 8. 6 passage, respectively at D. Graham, Aristotle Physics Book VIII (Oxford 1999), 111, and R.  McKirahan, Simplicius: On Aristotle Physics 8. 6–10 (Duckworth 2001), 161. But no one, so far as I know, has discussed De somno in detail in connection with the problem of self-­motion in Physics 8. 2 and 8. 6. 31 See De somno 3, 456b5–6. Aristotle refers to On Nourishment at the following places: Meteor. 4. 3, 381b13; DA 2. 4, 416b30–1; PA 4. 4, 678a16–19; and GA 5. 4, 784b2–3. Diogenes Laertius’ catalogue, however, doesn’t include anything close. 32  Recent scholarship on De somno concentrates on the problem of its unity. Accepting Nuyens’s developmental thesis concerning Aristotle’s psychology, Drossaart Lulofs discerns two discontinuous parts in De somno, the first part of the treatise (before 455b13) representing Aristotle’s ‘late’, hylomorphic account of the soul and the second part (after 455b13) representing Aristotle’s ‘middle’, instrumentalist account. See H. J. Drossaart Lulofs, Aristotelis De insomniis et De divinatione per somnum (Leyden 1947), xvi–­xx. Cf. M.  Lowe, ‘Aristotle’s De somno and his Theory of Causes’, Phronesis, 23 (1978), 279–91, and J. Wiesner, ‘The Unity of the De somno and the Physiological Explanation of Sleep in Aristotle’ [‘Unity’], in

250

Wei Wang

for De somno occupies a special place in Aristotle’s animal psychology. On the one hand, as Aristotle spares no effort in ­pointing out in the first chapter of this treatise, sleep and waking are affections (πάθη) of the perceptive soul.33 Therefore, De somno certainly belongs to a description of the perceptive part of the animal soul. On the other hand, somewhat surprisingly, he devotes a major part of the treatise to the discussion of nutritive activities, and especially digestion. The reason for this apparent mismatch lies in sleep and waking being the result of the causal interaction between the nutritive and the perceptive processes: when the perceptive soul, or its organ, is acted on per accidens by the nutritive activities going on in the animal body, sleep or waking arises as an affection of the perceptive part. This is exactly what we need in order to understand the Physics 8. 6 passage. Now, which perceptive part is acted on in the case of sleep and waking? Aristotle tells us in De somno 2, 455a12–27 that it is not a special part such as that of vision, but ‘the ruling perceptive organ’ (τὸ κύριον αἰσθητήριον, 455a21), that is, ‘that which has the power of touch’ (τῷ ἁπτικῷ, 455a22–3), a part that all animals have in common. Aristotle locates this part around the heart (περὶ τὴν καρδίαν, 456a4) or its analogue in bloodless animals, from which not only sense-­perception, but also locomotion, originates (455b34–456a23; cf. MA 9, 702b20–5). Aristotle’s next task in De somno, after locating the part to which sleep and waking belong, is to find the condition and cause of their G.  E.  R.  Lloyd and G.  E.  L.  Owen (eds.), Aristotle on the Mind and the Senses (Cambridge 1978), 241–80. Everson, on the other hand, argues for the traditionalist position, especially against Lowe, who thinks that 455b13–34 is an insertion from an earlier draft, that De somno offers a coherent account of sleep which combines material and teleological explanations. See S. Everson, ‘The “De somno” and Aristotle’s Explanation of Sleep’ [‘Explanation’], Classical Quarterly, n. s., 57 (2007), 502–20. Whatever the outcome of the debate, there is no doubt, as I will show below, that Aristotle believes that sleep and waking are the affections of animal as a perceptive being, and that they come about through the functioning of animal as a nutritive being. 33  Or, to put it differently, they are the affections of the bodily organs which correspond to the perceptive soul. Here I remain agnostic on the question where in Aristotle’s philosophical career to locate De somno. It may be true, as according to Menn’s refined developmentalist account (see Menn, ‘Definition’, 86–90), that De somno represents an earlier view according to which the soul is still ‘moved via its body’ (see De somno 1, 454a8–10). In translating and discussing the relevant passages, however, I remain neutral as to whether ‘the perceptive part’ is a psychic part or a bodily part or both, because it doesn’t affect my overall purpose.



Digestion, Self-Motion, and Eternity

251

occurrence (3, 456a30–2, 458a25–6). Aristotle thinks, presumably on the basis of the everyday experience that we get drowsy after a meal, that sleep and waking are caused by the process of nutrition (456b18–20). Much of De somno 3 consists of Aristotle’s description of the nutritive process, and in particular how it induces sleep and waking. According to Aristotle’s description, the entrance of food into the animal body gives rise to a series of nutritive activities consisting mainly of two processes: (1) the digestion (πέψις, lit. ‘cooking up’) of food into blood and waste and (2) the separation (διάκρισις) of blood into finer and denser blood. According to Aristotle’s theory of animal digestion elsewhere, external food first enters the mouth, which breaks it down to facilitate good digestion (PA 2. 3, 650a8–10); then it travels through the oesophagus into the stomach, which is the receptacle for undigested food and the first organ of digestion. From there, food that has been cooked and ­broken down in the stomach and liver is first turned into blood in the veins and then moved towards the heart, which is the proper origin (ἀρχή) and receptacle for blood.34 The blood thus coming out of the heart is hot, and it has to be cooled in the brain and sep­ ar­ated out in the heart before it can be used to sustain and replenish the organs and the other organic parts of the body (457b29–458a19). It is Aristotle’s view that the first of the two processes, digestion, causes an animal first to be drowsy and then to fall asleep, whereas the second, separation, causes an animal to wake up and be conscious. According to Aristotle’s physiology, processed food produces hot ‘evaporation’ (ἀναθυμίασις), which first travels to the upper region, in particular the head, and causes the animal to nod and be drowsy; then, when the cooled evaporation returns to the lower and interior region, and in particular the heart, it causes the

34  De somno 3, 456a32–b6: ‘It is manifest that . . . when the food enters from outside the parts fitted for its reception, the evaporation enters into the veins, and there, undergoing a change, is converted into blood, and makes its way to their origin [i.e. the heart]. All this has been dealt with in On Nourishment.’ (ϕανερὸν δὴ ὅτι . . . τῆς μὲν οὖν θύραθεν τροϕῆς εἰσιούσης εἰς τοὺς δεκτικοὺς τόπους γίγνεται ἡ ἀναθυμίασις εἰς τὰς ϕλέβας, ἐκεῖ δὲ μεταβάλλουσα ἐξαιματοῦται καὶ πορεύεται ἐπὶ τὴν ἀρχήν. εἴρηται δὲ περὶ τούτων ἐν τοῖς Περὶ τροϕῆς). Cf. PA 3. 4 665b9–666b1. For the translation of De somno, I use that by J. I. Beare in Barnes (ed.), Complete Works, 721–35, with my own modifications. For the Greek text, I use that of W.  D.  Ross, Aristotle’s Parva Naturalia: A Revised Text with Introduction and Commentary (Oxford 1955).

252

Wei Wang

animal to fall asleep.35 Then, a person awakes when ‘the digestion is completed’ (πεϕθῇ) and the finer and the heavier blood ‘have been separated’ (διακριθῇ).36 In sum, the cause of sleep and waking consists in the ebb and flow of the corporeal fluids and solids, first as food and then as blood and waste, in a mass around the primary perceptive organ (τὸ πρῶτον αἰσθητήριον), the heart.37 Food is not only the cause of an animal’s nutritive process, but because it affects the heart, it is also the cause of the animal’s two perceptive affections: sleep and waking. This echoes with what Aristotle claims in Physics 8. 6 (quoted above): Thus in some cases the cause is food: when it is being digested [πεττομένης] animals sleep, and when it is being separated [διακρινομένης] they awaken 35  De somno 3, 456b17–28: ‘As we observed above, sleep is not just any impotence of the perceptive faculty, but this affection is one which arises from the evaporation involving food. The matter evaporated must be driven onwards to a certain point, then turn back and change like a tide race. Now, in every animal the hot naturally tends to move upwards, but when it has reached the parts above, it turns back again, and moves downwards in a mass. This explains why sleeps come from food; for the matter, both the liquid and the corporeal, which is borne upwards in a mass, is then of considerable quantity. When, therefore, this comes to a stand, it weighs a person down and causes him to nod, but when it has actually sunk downwards, and by its return has repulsed the hot, sleep comes on, and the animal is presently asleep. (ἀλλὰ γάρ, ὥσπερ εἴπομεν, οὐκ ἔστιν ὁ ὕπνος ἀδυναμία πᾶσα τοῦ αἰσθητικοῦ, ἀλλ’ ἐκ τῆς περὶ τὴν τροϕὴν ἀναθυμιάσεως γίγνεται τὸ πάθος τοῦτο· ἀνάγκη γὰρ τὸ ἀναθυμιώμενον μέχρι του ὠθεῖσθαι, εἶτ’ ἀντιστρέϕειν καὶ μεταβάλλειν καθάπερ εὔριπον. τὸ δὲ θερμὸν ἑκάστου τῶν ζῴων πρὸς τὸ ἄνω πέϕυκε ϕέρεσθαι· ὅταν δ’ ἐν τοῖς ἄνω τόποις γένηται, ἀθρόον πάλιν ἀντιστρέϕει καὶ καταϕέρεται. διὸ μάλιστα γίγνονται ὕπνοι ἀπὸ τῆς τροϕῆς· ἀθρόον γὰρ πολὺ τό τε ὑγρὸν καὶ τὸ σωματῶδες ἀναϕέρεται. ἱστάμενον μὲν οὖν βαρύνει καὶ ποιεῖ νυστάζειν· ὅταν δὲ ῥέψῃ κάτω καὶ ἀντιστρέψαν ἀπώσῃ τὸ θερμόν, τότε γίγνεται ὁ ὕπνος καὶ τὸ ζῷον καθεύδει). Cf. De somno 3, 457a33–b6, 458a26–8, and PA 2. 7, 653a10–19. There are disputes, both ancient and modern, over whether Aristotle thinks sleep is the result of the heart being warmed or being cooled, because his own statement is obscure and can be interpreted in either way. For a survey of these disputes, see Wiesner, ‘Unity’, 251–71. I do not take a position on the issue because it does not affect my conclusion. 36  De somno 3, 458a10–12: ‘An animal awakes when the digestion of the food is completed, when the heat, which had been forced together in large quantity within a small compass from the surrounding part, has once more prevailed, and when the more corporeal and the purer blood have been separated’ (ἐγείρεται δ’ ὅταν πεϕθῇ καὶ κρατήσῃ ἡ συνεωσμένη θερμότης ἐν ὀλίγῳ πολλὴ ἐκ τοῦ περιεστῶτος, καὶ διακριθῇ τό τε σωματωδέστερον αἷμα καὶ τὸ καθαρώτερον). 37  De somno 3, 458a25–8: ‘We have now stated the cause of sleep, viz. that it consists in the return of the corporeal [matter], borne upward by the connatural heat, in a mass upon the primary perceptive organ’ (τί μὲν οὖν τὸ αἴτιον τοῦ καθεύδειν εἴρηται, ὅτι ἡ [ὑπὸ] τοῦ σωματώδους τοῦ ἀναϕερομένου ὑπὸ τοῦ συμϕύτου θερμοῦ ἀντιπερίστασις ἀθρόως ἐπὶ τὸ πρῶτον αἰσθητήριον).



Digestion, Self-Motion, and Eternity

253

and move themselves, the original starting point being from outside. (259b8–14)

To sum up the discussion on De somno, I have shown that Aristotle thinks that it is the food that enters from without, together with the nutritive activities of digestion (πέψις) and separation, that acts on the perceptive part of an animal, and that waking and sleep are the result of the perceptive part being acted on.38

7.  Temporal continuity and the ou kuriōs claim in Physics 8. 6 It may appear paradoxical to claim that what causes an animal’s self-­motion is not from itself; yet on the basis of the above analysis of De somno, we can see that, strictly speaking, the external cause (i.e. the food) and the internal cause (i.e. the perceptive soul) are not responsible for the same effect: what food causes, through an animal’s nutritive activities, is the start, the end, and the duration of the animal’s perceptive and locomotive activities, whereas the perceptive soul determines what these activities are. In other words, an animal can only perceive some object X and move itself from place A to place B (where X is situated) when it is awake,39 so food 38  In Aristotelian terms, the sense-­perceptive capability is what differentiates animals from plants; both have the nutritive capability as a common, defining feature. In this sense, one may define the animal kind as ‘a being capable not only of nourishment, but also of sense-­perception’, the former being a genus, the latter a differentia. Given that the definitional pair genus and differentia are sometimes characterized as ‘matter’ and ‘form’ (e.g. Metaphysics Η. 6), one may be tempted to think that the way an animal’s nutritive parts affect its perceptive functioning should be characterized as a kind of material causation, as opposed to formal caus­ ation, which is due to the animal soul as such. Thus, the two kinds of causation operate on different levels upon the same thing: the animal is a self-­mover on the formal level, but is externally determined on the material level, proximately by its nutritive parts and remotely by food and other lifeless things. On Aristotle’s discussion, in the latter half of Metaphysics Z, concerning the relation between the parts of the matter-­form composite and the parts of the genus-­species definition, see Menn, ‘Metaphysics Ζ 10–16 and the Argument-­ Structure of Metaphysics Ζ’, Oxford Studies in Ancient Philosophy, 21 (2001), 83–134 at 133–4, n. 53. For an interpretation of Aristotle’s explanation of sleep and waking offered in De somno as material explanation, see Everson, ‘Explanation’, 512–20. 39  On sleepwalking, see De somno 2, 456a24–7 and De Insomniis 2. Aristotle thinks that our dreams are due to the qualitative changes in our sense organs that continue after they have ceased to actively perceive, just as a projectile still travels forward after the hand ceases to actively move it (De Insomn. 2, 459b3–23). So, when an animal sleepwalks due to its dreams, it does so insofar as it is not strictly asleep,

254

Wei Wang

is a condition sine qua non of the animal’s perception of X and its locomotion to X. But the fact that it perceives X and not something else is caused by the presence of the object and the functioning of the perceptive soul; similarly, the fact that it moves itself from A to B and not in some other direction is caused by its perceiving and desiring X at B. Therefore, while the animal is not a self-­mover kuriōs so far as independence from external sunaitia (co-­causes) is concerned, it is a self-­mover with respect to particular motions and objects. Aristotle spells out this important distinction in the last sentence of the Physics 8. 6 passage: ‘that is why animals are not always moved continuously by themselves’ (διὸ οὐκ ἀεὶ κινοῦνται συνεχῶς ὑϕ’ αὑτῶν, 259b14–15). This is to say, as long as the animals are in a temporally continuous episode of being conscious, they are self-­movers, but that they are in such episodes is caused by something external, namely food. This is supported by the physiological theory in De somno 3 that food, with its digestion and distribution, punctuates an animal’s daily activities into temporally discrete chunks. To return to the puzzling sentence in Physics 8. 6, ‘animals do not move themselves with locomotion kuriōs’: they are self-­movers only when they are awake, but whether they are awake or not is determined by ‘the other motions’—the nutritive activities—­which are in turn caused by ‘the surroundings and the many things that come in’, namely food. A kuriōs self-­mover would have to cause its own motion simply by itself, and not in one way by itself but in another way externally.40 Thus, an animal may fail to be a self-­mover kuriōs; yet it can still move itself in a qualified way, for example, insofar as the direction and the speed of its motion are concerned. Crucially, the reason why Aristotle uses kuriōs to pick out all causal aspects instead of the relevant aspect in which the animal is a self-­mover lies in the motivating problem he faces when making the distinction: a real self-­mover is one that can initiate motion ex nihilo, and the animal fails to meet that standard because it is not causally independent in all aspects, whereas the real mover of the universe, God, does meet the standard, as it is causally independent in that way. because its sense organs are still affected by the object that had been perceived when it was awake. 40  See Bonitz s.v. κυρίως, 416b7–10.



Digestion, Self-Motion, and Eternity

255

At this point one may easily see the deficiency of the intentional reading. According to this reading, when Aristotle says that ‘animals move themselves only with one kind of motion, and with this ou kuriōs’, he means that an animal’s self-­locomotion is, strictly speaking, caused by its intentional object, which is outside the ­animal in question. But loosely speaking, because the intentional objects can be said to be internal to the mind under some description, animals can be said to move themselves. Leaving aside the question whether this interpretation is supported by De anima and De motu animalium or not, in Physics 8 at least it has no basis. Although the parallel passage in Physics 8. 2 may tempt one towards such an interpretation,41 Aristotle clarifies in Physics 8. 6 that the externally caused changes and activities in the animal body—­some of which in turn move the intellect (διάνοια) or desire (ὄρεξις)—are not the perception of the external objects, as the intentional reading would require, but such mundane changes and activities as increase, decrease, breathing, and digestion, which are all nutritive rather than perceptive. Besides, Aristotle’s claim about the con­ tinu­ity of animal self-­motion at 259b14–15 implies that, at least in some way, such continuity is not interrupted by the occurrence of different intentional objects.

8.  The broader context I have shown that, in order to understand the particular argument concerning self-­motion in Physics 8. 6, one has to distinguish two aspects of an animal’s locomotion: the temporal extension of the motion which is externally caused, on the one hand, and the goal and direction in which the animal remains a self-­mover, on the other. More broadly, to understand Aristotle’s arguments in Physics 8 concerning the first unmoved mover and the non-­eternal unmoved movers, it is important to distinguish between the being and nature of some substance and the temporal extension of its being, that is,

41  Phys. 8. 2, 253a15–17 (quoted more fully above): ‘Many motions are produced in the body by its surroundings (ὑπὸ τοῦ περιέχοντος), and some of these set in motion the intellect (διάνοια) or desire (ὄρεξις), and this again then sets the whole animal in motion.’

256

Wei Wang

its existence.42 Insofar as a substance has a certain being and nature, it tends to fulfil this nature, but because its body or matter is susceptible to external causes, it does not fulfil this nature eternally or even continuously.43 The continuity within single episodes and the discontinuity between consecutive episodes are due to ‘the surroundings’ (τὸ περιέχον)—and in the case of eternity, due to the eternal first motion that contains (περιέχει) all motions (De caelo 2. 1, esp. 284a5–11) and to the first unmoved mover. That these non-­eternal unmoved movers cannot cause eternal motion is what Aristotle claims in the immediately ensuing paragraph in Physics 8. 6. Unless we follow the interpretation outlined above, he seems to have digressed from arguing that the non-­eternal movers cannot be the model of a non-­eternal universe—­it cannot be the case that the universe, to which there is nothing external, operates like an intermittent self-­mover, the intermittence of whose motion is due to external causes—­to arguing that they cannot cause an eternal effect. Surely they cannot, but how does this follow as a 42  One might cite De long. vit. 5 and argue that because each living species has a natural lifespan, its nature should incorporate how long it can be expected to live. But this is not true. According to Aristotle’s account there, the length of life of each individual is due to how much and how warm the moisture inside the body is as compared with the environment, and this varies across different kinds of living beings, different species within the same kind, and different individuals within the same species but having different habitats. Cf. GA 4. 8, 777b6–8, where Aristotle claims that the length of life is associated with the proximity of the way the animal is mixed elementally to the way the surrounding (περιέχοντα) air is mixed. Note also that in GC 2. 10 and GA 4. 8, the life cycles of living things are said to be caused remotely by the approaching and retreating sun. 43  In a number of passages, Aristotle makes the interesting claim that the con­ tinu­ity of motion or activity is laborious (ἐπίπονος) for a mover or an agent that is in the relevant sense a form-­matter composite. See e.g. Metaph. Θ. 8, 1050b24–8: ‘And [the heavenly bodies] do not tire in this activity [i.e. locomotion]; for movement is not for them, as it is for the perishables, connected with the potentiality for op­pos­ ites, so that the continuity of the movement should be laborious; for it is substance in the sense of matter and potentiality, not actuality, that causes this’ (trans. Ross; οὐδὲ κάμνει τοῦτο δρῶντα· οὐ γὰρ περὶ τὴν δύναμιν τῆς ἀντιϕάσεως αὐτοῖς, οἷον τοῖς ϕθαρτοῖς, ἡ κίνησις, ὥστε ἐπίπονον εἶναι τὴν συνέχειαν τῆς κινήσεως· ἡ γὰρ οὐσία ὕλη καὶ δύναμις οὖσα, οὐκ ἐνέργεια, αἰτία τούτου, text from W.  D.  Ross (ed.), Aristotle’s Metaphysics: A Revised Text with Introduction and Commentary, 2 vols. (Oxford 1957)). Cf. Metaph. Λ. 9, 1074b28–9; Phys. 8. 10, 267b2–4; and De somno 1, 454a24–9. The point is presumably that because the matter of a form-­matter composite, as what underlies its actual form, is potentially the privation of that form, when the composite mover acts on something else qua formal and actual, it is reciprocally acted upon qua material and is changed to what it potentially is, i.e. the contrary of its formal status; therefore, it cannot cause continuous motion. For Aristotle’s discussion of reciprocal action and motion, see GC 1. 6–7.



Digestion, Self-Motion, and Eternity

257

conclusion from the argument against motion ex nihilo? If this were the case, then the connecting relative pronoun ‘from which’ (ἐξ ὧν) would seem very strange indeed: ἐξ ὧν ἔστιν πιστεῦσαι ὅτι εἴ τί ἐστι τῶν ἀκινήτων μὲν κινούντων δὲ καὶ αὑτὰ κατὰ συμβεβηκός, ἀδύνατον συνεχῆ κίνησιν κινεῖν. ὥστ’ εἴπερ ἀνάγκη συνεχῶς εἶναι κίνησιν, εἶναί τι δεῖ τὸ πρῶτον κινοῦν ἀκίνητον καὶ κατὰ συμβεβηκός, εἰ μέλλει, καθάπερ εἴπομεν, ἔσεσθαι ἐν τοῖς οὖσιν ἄπαυστός τις καὶ ἀθάνατος κίνησις, καὶ μενεῖν τὸ ὂν αὐτὸ ἐν αὑτῷ καὶ ἐν τῷ αὐτῷ· τῆς γὰρ ἀρχῆς μενούσης ἀνάγκη καὶ τὸ πᾶν μένειν συνεχὲς ὂν πρὸς τὴν ἀρχήν. (Physics 8. 6, 259b20–8) From which [considerations] [ἐξ ὧν] we may confidently conclude that if a thing belongs to the class of unmoved things that move themselves per accidens, it is impossible that it should cause continuous motion. So the necessity that there should be motion continuously requires that there should be a first mover that is not moved, even per accidens, if, as we have said, there is to be in the world of things an unceasing and undying motion, and the world is to remain [permanently] self-­contained and within the same limits; for if the first principle remains, the universe must also remain, since it is continuous with the first principle.

The fact that some unmoved thing moves itself per accidens implies that it has and is in an instrumental body (259b19–20), and the fact that an animate body, as what is material and potential, is sus­cep­ tible to external causation—­as has been pointed out in the case of animal nutrition—­ explains why such a mover does not cause motion continuously. Hence, Aristotle’s qualification of animal as self-­mover also explains why its soul, as a per accidens self-­mover, does not cause continuous motion. Morison makes a crucial point at the end of his paper commenting on the paragraph: No-­one can pretend that this part of the argument [i.e. 8. 6, 259b22–31] is easy. Why should the fact that an animal soul moves itself accidentally prevent it from being the cause of undying motion? Understanding this is the principle difficulty of interpretation at this point in Aristotle’s Physics. The remarks in Physics VIII which commentators have thought pose problems for Aristotle’s theory [i.e. that the animals do not move themselves kuriōs] do not in fact: as a result, our attention has been diverted from the main points of interest in the long and complex argument of Physics VIII. (Morison, ‘Self-­Motion’, 79; emphasis original)

I hope the discussion in this section has responded to this point adequately. But even if we follow Morison’s interpretation of the

258

Wei Wang

ou kuriōs claim as a hidden criticism of Plato’s conception of soul in the Phaedrus (Section 4 above), there is still a link between the fact that soul is only a per accidens self-­mover and the fact that it doesn’t cause undying motion. A Platonic, per se self-­moving soul, as is described in Phaedrus 245 c 2–246 a 2, would be able to cause eternal motion because the cause of its motion is itself as such. An Aristotelian per accidens self-­ moving soul, by contrast, fails to cause eternal motion because the body it uses to transmit motion, both to itself and to other things, as something that is susceptible to external causation, wears out and needs regular rest and replenishment, and will not remain such a mover for eternity. Indeed, any unmoved mover that is moved per accidens, including the eternal movers of the solar, the lunar, and the planetary heavens, which have to be in some bodily thing that is per se movable, would fail to provide such continuity and eternity (Phys. 8. 6, 259b28–31). So the necessity of there being motion eternally requires that there be a first mover that is unmoved not only per se but also per accidens. At the beginning of Physics 8. 6, Aristotle describes such an unmoved mover as unmoved simpliciter (ἁπλῶς, 258b15). Such a mover is the only one that causes its effects simply by itself, independently of any external factors. It is no wonder that an animal soul does not cause its effects simply by itself and therefore that ensouled animals are not self-­movers kuriōs. Fudan University

BIBLIOGR A PH Y Barnes, J. (ed.), The Complete Works of Aristotle: The Revised Oxford Translation [Complete Works], vol. i (Princeton, 1984). Blyth, D., ‘Heavenly Soul in Aristotle’, Apeiron, 48 (2015), 427–65. Coope, U., ‘Aristotle’s Account of Agency in Physics III.3’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, 22 (2005), 201–27. Cooper, J. M. (ed.), Plato: Complete Works (Indianapolis, 1997). Drossaart Lulofs, H. J., Aristotelis De insomniis et De divinatione per somnum (Leyden, 1947). Everson, S., ‘The “De somno” and Aristotle’s Explanation of Sleep’ [‘Explanation’], Classical Quarterly, n. s., 57 (2007), 502–20.



Digestion, Self-Motion, and Eternity

259

Furley, D., ‘Self-Movers’, in G. E. R. Lloyd and G. E. L. Owen (eds.), Aristotle on the Mind and on the Senses (Cambridge, 1978), 165–79. Gill, M. L., ‘Aristotle on Self-Motion’ [‘Self-Motion’], in M. L. Gill and J. G. Lennox (eds.), Self-Motion: From Aristotle to Newton (Princeton, 1994), 15–34. Graham, D., Aristotle Physics Book VIII (Oxford, 1999). Hardie, R. and Gaye, R. (trans.), Physica, in W. D. Ross (ed.), The Works of Aristotle, vol. ii (Oxford, 1930). Lowe, M., ‘Aristotle’s De somno and his Theory of Causes’, Phronesis, 23 (1978), 279–91. McKirahan, R., Simplicius: On Aristotle Physics 8. 6–10 (London, 2001). Menn, S., ‘Aristotle’s Definition of Soul and the Programme of the De Anima’ [‘Definition’], Oxford Studies in Ancient Philosophy, 22 (2002), 83–139. Menn, S., ‘Metaphysics Ζ 10–16 and the Argument-Structure of Metaphysics Ζ’, Oxford Studies in Ancient Philosophy, 21 (2001), 83–134. Meyer, S. S., ‘Self-Movement and External Causation’ [‘Self-Movement’], in M. L. Gill and J. G. Lennox (eds.), Self-Motion: From Aristotle to Newton (Princeton, 1994), 65–80. Morison, B., ‘Self-Motion in Physics VIII’ [‘Self-Motion’], in A.  Laks and M. Rashed (eds.), Aristote et le mouvement des animaux: Dix études sur le De motu animalium (Villeneuve d’Ascq, 2004), 67–79. Nussbaum, M. C., Aristotle’s De motu animalium [De motu] (Princeton, 1978). Ross, W. D. (ed.), Aristotle’s Metaphysics: A Revised Text with Introduction and Commentary, 2 vols. (Oxford, 1957). Ross, W. D., Aristotle’s Parva Naturalia: A Revised Text with Introduction and Commentary (Oxford, 1955). Ross, W.  D., Aristotle’s Physics: A Revised Text with Introduction and Commentary [Physics] (Oxford, 1936). Solmsen, F. F., ‘Plato’s First Mover in the Eighth Book of Aristotle’s Physics’, in R. B. Palmer and R. Hammerton-Kelly (eds.), Philomathes: Studies and Essays in the Humanities in Memory of Philip Merlan (The Hague, 1971), 171–82. Wiesner, J., ‘The Unity of the De somno and the Physiological Explanation of Sleep in Aristotle’ [‘Unity’], in G. E. R. Lloyd and G. E. L. Owen (eds.), Aristotle on the Mind and the Senses (Cambridge, 1978), 241–80.

GIVING GIFTS AND MAKING FRIENDS Seneca’s De beneficiis on How to Expand One’s Sphere of Ethical Concern allison piñeros glasscock 1. Introduction The ethical ideal of the ancient Stoics is ambitious and chal­len­ ging. The person who attains this ideal—­the sage—­regards only vir­ tue and what participates in it (e.g. virtuous actions and persons) as good.1 In addition, the sage has a maximally expansive sphere of ethical concern: she cares about the good (i.e. the virtue) of everyone, not merely the good of those who are related to her by pre-­existing social bonds, such as the bonds of kinship.2 The demanding For challenging comments on a predecessor of this paper, I am grateful to George Boys-­Stones, Brad Inwood, and Christian Pfeiffer, as well as audiences at the University of Toronto, the University of Illinois at Chicago, Georgia State University, and Yale University. The current version of the paper is indebted to thorough and insightful comments from Victor Caston and two anonymous referees. As always, thank you to Juan S. Piñeros Glasscock for listening to and critiquing the argument through of all its messy stages. I am also grateful to Nana Yaa Awere for her diligent and speedy help with formatting the references. 1  See D.L. 7. 94 = SVF iii. 76 and Stob. 2. 7. 5 a, 57. 20–2 Wachsmuth–­Hense = SVF iii. 70. In what follows, and unless otherwise noted, translations and Greek and Latin texts are from A. A. Long and D. N. Sedley, The Hellenistic Philosophers [LS], 2 vols. (Cambridge: 1987), with slight modifications. References to primary sources also include a cross reference to LS, when possible. If LS does not have the passage, I provide a cross reference to J. von Arnim, Stoicorum Veterum Fragmenta [SVF], 4 vols. (Leipzig: 1903–1924). I do not provide cross references for texts from Seneca. Translations of Seneca’s De beneficiis are from M. Griffin and B. Inwood (trans.), Seneca, On Benefits (Chicago, 2011), sometimes with slight modifications. The Latin text of De beneficiis is from C. Hosius (ed.), L. Annaei Senecae, De ben­ eficiis libri VII; De clementia libri II (Leipzig, 1905). Translations from Seneca’s Epistulae morales are from M. Graver and A. A. Long (trans.), Lucius Annaeus Seneca, Letters on Ethics (Chicago, 2015), sometimes with slight modifications. The Latin text of Seneca’s Epistulae morales is from L. D. Reynolds (ed.), L. Annaei Senecae: Ad Lucilium Epistulae morales [OCT], 2 vols. (Oxford, 1965). 2  For the idea that ethical concern should include all fellow humans, see Cic., Fin. 3. 63–8 = LS 57F; Stob. 4. 27. 23, 671. 7–672. 18 = LS 57G; and Anon., In Plat. Theaet. 5. 18–6. 31 = LS 57H. Allison Piñeros Glasscock, Giving Gifts and Making Friends: Seneca’s De beneficiis on How to Expand One’s Sphere of Ethical Concern In: Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Allison Piñeros Glasscock 2023. DOI: 10.1093/oso/9780192885180.003.0007

262

Allison Piñeros Glasscock

character of the Stoic ideal raises pressing questions about the pos­ sibility and nature of ethical progress. It is generally agreed that the Stoics think non-­sages can make progress towards virtue. A more contentious issue is how they think such progress is made.3 My central claim is that the Roman Stoic Seneca offers a blue­ print for ethical progress in his work De beneficiis (On Benefits). The primary purpose of Seneca’s treatise is to provide a philosophical examination of the social practice of exchanging favours, or giving and receiving benefits, from a Stoic perspective.4 Over the course of that examination, Seneca identifies benefaction as a source of friendship (see e.g. Ben. 2. 18. 5 and 2. 21. 2). I shall argue that by linking benefaction with friendship Seneca positions the practice of giving and receiving benefits to play a (and perhaps the) central role in ethical development.5 When Seneca remarks in Ben. 1. 4. 2 that the topic of benefaction ‘more than any other binds together human society’ (quae maxime humanam societatem adligat), he does not simply mean that benefaction is very im­port­ant for maintaining 3  Much of the discussion has focused on the Stoic doctrine of oikeiōsis or ‘orien­ tation’. One of the central questions has been how the Stoics envision the agent progressing from caring for himself to caring about others. There is a vast literature on the topic. For a helpful recent discussion of oikeiōsis in Roman Stoicism, see G. Reydams-­Schils, ‘Human Bonding and oikeiōsis in Roman Stoicism’ [‘Bonding’], Oxford Studies in Ancient Philosophy, 22 (2002), 221–51. For a persuasive challenge to the traditional division of types of oikeiōsis into personal and social, as well as extensive references to the literature, see J.  Klein, ‘The Stoic Argument from oikeiōsis’ [‘Stoic Argument’], Oxford Studies in Ancient Philosophy, 50 (2016), 143–200. I return to this issue in Section 2. 4  For discussion of the cultural and social context of the treatise, see M. Griffin, Seneca on Society: A Guide to De beneficiis [Society] (Oxford, 2013); B. Inwood, ‘Politics and Paradox in Seneca’s De beneficiis’, in id., Reading Seneca: Stoic Philosophy at Rome (Oxford, 2005), 65–94; and the introduction to M. Griffin and B. Inwood’s translation of Seneca, On Benefits, 1–14. 5 It is difficult to say how novel Seneca’s proposal is in its Stoic context. As ­others have noted, Seneca regularly attributes aspects of his analysis of benefaction to other Stoics, including Chrysippus (Ben. 1. 3. 9) and Hecaton (Ben. 2. 18. 2). On the other hand, as we shall see, Seneca’s analysis poses a serious challenge to canon­ ical Stoicism. I am inclined to think that the discrepancies are not the result of a thinker gone rogue but rather the manifestation of tensions that are internal to Stoicism. For more discussion of this point, see the second half of Section 4. F.-R. Chaumartin, Le De beneficiis de Sénèque: Sa signification philosophique, politique et sociale (Paris: 1985) defends the view that De beneficiis is primarily based on a work by Hecaton (see esp. chs. 1 and 2 = pp. 31–105). See Griffin, Society, 24–5 for helpful discussion of assessments of Seneca’s sources, and cf. M.  Graver, ‘Interiority and Freedom in Seneca’s De beneficiis: Acts of Kindness and the Perfected Will’ [‘Interiority’], in M. R. Niehoff and J. Levinson (eds.) Self, Self-­ Fashioning and Individuality in Late Antiquity (Tübingen, 2020), 71–88 at 72.



Giving Gifts and Making Friends

263

a surface-­level social cohesion, complete with functioning relation­ ships, but that in benefaction participants are properly oriented towards, and come to care for, each other’s good.6 I begin with a few remarks about the Stoic ideal and about the nature of the problem that I think Seneca’s De beneficiis addresses (Section 2). In Section 3, I discuss the relationship between bene­ fac­tion and friendship. Like other Stoics, Seneca thinks that friend­ ship is a relationship between equals and that friends come to identify with one another, to regard each other as members of the same community, and to see the goods each has as being held in common. I argue that benefaction is particularly suited to the for­ mation of friendship because it is a joint activity, one that is real­ ized in the reciprocal exchange of good intentions between the giver and recipient, and because it is aimed at the good (i.e. the virtue) of the participants. Benefaction is also an equalizing activity, in that it puts its participants on a par with each other, at least with respect to their good intentions. The special nature of benefaction explains how those who participate in it come to have the qualities and atti­ tudes that the Stoics take to be distinctive of friendship. I examine two issues raised by this analysis in Section 4, one concerning the uniqueness of benefaction relative to other virtuous actions and one concerning the implications of the analysis for the self-­sufficiency of the virtuous person. In the concluding section of the paper (Section 5), I show that friendship, as conceptualized by Seneca and the Stoics, shares its core features with the Stoic ethical ideal. When you make friends with someone, you come to care for and promote their good in precisely the way required by the ideal. Thus, bene­ faction, because it results in friendship, turns out to be a powerful means of expanding one’s sphere of ethical concern.

6  Several scholars have already commented on the important role that bene­fac­tion plays in the creation of healthy social relationships and, in turn, well-­functioning societies. See K. Takaki, ‘Benefits and Communication: Semiotics in Seneca’s De beneficiis’ [‘Communication’], History of Philosophy Quarterly, 31 (2014), 293–316 at 305–7, and Griffin, Society, 74. Some have also provided helpful discussions of how engagement in benefaction might enable one to cultivate virtue (e.g. Takaki, ‘Communication’, 307–13), and Graver has noted that beneficent actions are actions that are ‘motivated primarily by the needs of others’ (‘Interiority’, 71–2). But to my knowledge no one has discussed the way in which benefaction might specifically cultivate concern for the good (i.e. the virtue) of the other.

264

Allison Piñeros Glasscock 2.  The Stoic ideal

The central question that I propose to investigate—­how do we develop concern for the good of others—­is related to the thorny issue of ‘oikeiōsis’ in Stoic ethics. ‘Oikeiōsis’, which I shall translate as ‘orientation’,7 refers to the disposition to value and so to pro­ mote one’s own good and the good of others. Traditionally, oikeiōsis is divided into two types, personal oikeiōsis (the orientation towards one’s own good) and social oikeiōsis (the orientation towards the good of others).8 It is with the latter type of oikeiōsis that this paper is concerned. Someone who is properly oriented towards others regards them as being part of the same rational community and as being, essen­ tially, like himself. In Cicero’s De finibus, the Stoic Cato says that ‘the mere fact that someone is a man makes it incumbent on another man not to regard him as alien’.9 Plutarch reports that according to Zeno’s Republic, ‘we should regard all men as our fellow citizens and local residents, and there should be one way of life and order’.10 Cato also makes clear that the properly oriented agent identifies with the entire community. From the perspective of this identifica­ tion, he sees others as parts or extensions of himself and comes to value and promote the good of the community as a whole.11 One 7  For helpful discussion of the meaning and appropriate translation of oikeiōsis and its corresponding verb form, see Klein, ‘Stoic Argument’, 149–50; B. Inwood, Ethics and Human Action in Early Stoicism [Human Action] (Oxford, 1985), 184–5; and G. Striker, ‘The Role of oikeiōsis in Stoic Ethics’, in ead., Essays on Hellenistic Epistemology and Ethics (Cambridge, 1996), 281–97 at 281. 8  Recently, this division has been challenged: see Klein, ‘Stoic Argument’, 143–200 and Reydams-­Schils, ‘Bonding’, 222–3. The dispute over whether there are two distinct types of oikeiōsis does not matter for my purposes. My focus will be on how it is that we become oriented towards the good of others; if this turns out to be one and the same—­or continuous with—­the process by which we become oriented towards our own good, so much the better for Stoicism. 9 Cic., Fin. 3. 63 = LS 57F: ut oporteat hominem ab homine ob id ipsum, quod homo sit, non alienum videri. 10 Plut., De Alex. fort. 1. 6, 329 a = LS 67A: ἀλλὰ πάντας ἀνθρώπους ἡγώμεθα δημότας καὶ πολίτας, εἷς δὲ βίος ᾖ καὶ κόσμος. 11 See Cic., Fin. 3. 63–6, esp. 3. 64 = LS 57F, together with M.  Schofield’s insightful analysis of the passage in ‘Two Stoic Approaches to Justice’, in A. Laks and M.  Schofield (eds.), Justice and Generosity: Studies in Hellenistic Social and Political Philosophy, Proceedings of the Sixth Symposium Hellenisticum (Cambridge, 1995), 191–212 at 195–9. Cf. Hierocles’ remarks on how to accomplish the expan­ sion of ethical concern in Stob. 4. 27. 23, 672. 7–673. 11 Wachsmuth–­Hense = LS 57G 6–7 in part.



Giving Gifts and Making Friends

265

result of this refocusing is that it places one’s own good and the individual good of others on a par. Since the good of human beings is their virtue, and since virtue does not come in degrees, the indi­ vidual good of all people contributes in the same way and equally to the good of the whole.12 Thus, Cato argues that ‘we are driven by nature to desire to benefit as many people as possible, especially by giving instruction and handing on the principles of prudence’.13 The idea that the good of individual persons is integrally connected to the good of the whole rational community appears in other places as well. For example, Stobaeus reports that ‘all goods are common to the virtuous. . . . Therefore, a person who benefits someone also benefits himself’ and ‘all virtuous people benefit one another’.14 The Stoic ideal of the rational community is presented as a nor­ mative ideal. Each of us should try to enter into rational commu­ nity with others. We should try to identify with the community and to promote its good, where that includes promoting the individual good of as many other community members as we can. Yet there is  an immediate difficulty with the proposal. The Stoics’ central example of how we arrive at this expansive concern for others is that of parents and their children. Cicero’s Cato claims that we are naturally suited to procreate and that nature has endowed us with a complementary impulse to love what we generate, that is, our children. Similarly, the argument continues, we humans are nat­ur­ al­ly suited to enter into community with each other, and nature has seen to it that when we associate with one another, we develop a concern for the good of the resulting community (Fin. 3. 62–4 = LS 57F). Once we realize that all humankind forms a rational community that transcends the local communities of city, state, and family, we will naturally care for all humanity, just as parents naturally care for their children and just as citizens naturally care for their country.

12  The human good lies in virtue: D.L. 7. 87 = LS 63C; Sen., Ep. 76. 8–11. Virtue does not come in degrees: D.L. 7. 127 = LS 61 I; Stob. 2. 7. 5 b 8, 65. 7 = SVF 1. 566. 13 Cic., Fin. 3. 65 = LS 57F: inpellimur autem natura, ut prodesse velimus quam plurimis in primisque docendo rationibusque prudentiae tradendis. 14  Stob. 2. 7. 11 i, 101. 21–5 = LS 60P: τὰ δ’ ἀγαθὰ πάντα κοινὰ εἶναι τῶν σπουδαίων . . . δι’ ὃ καὶ τὸν ὠϕελοῦντά τινα καὶ αὐτὸν ὠϕελεῖσθαι . . . πάντας δὲ τοὺς σπουδαίους ὠϕελεῖν ἀλλήλους.

266

Allison Piñeros Glasscock

The argument gains what plausibility it has from its opening moves. It is easy to see that most parents do, in fact, develop an intense concern for their children. So, at least in this case, nature has seen to it that we love what we generate. But it is much less clear that nature has made a corresponding provision in the case of communities—­whether local or cosmic. Indeed, one might reason­ ably doubt whether it is even possible to care for one’s community and its members in the way that parents care for their children. In the case of children, there is a direct connection between the asso­ ciation of the parents and the generation of the child. This makes it easy for the parents to see the child as belonging to them. The connection between community members and their communities is much less direct. While it may be true that I help bring the brotherhood of man into existence by associating with others, it is difficult to specify what precisely my contribution to the gen­er­ation of that community is. As a result, it may also be difficult for me to take ownership of this community and to see it and its members as natural objects of my concern in the way that my child is a natural object of concern for me. The worry I just raised has an ancient counterpart. In an an­onym­ ous commentary on Plato’s Theaetetus, the author complains: ᾠκειώμε|θα γὰρ τοῖς ὁμοειδέσι· | μᾶλλον μέντοι ὠι|κείωται το̣[ῖς ἑα]υ̣το̣ ῦ πολίται[ς· ἐπιτείνε]|ται γὰρ καὶ ἀ[νίετ]α[ι] ἡ | οἰκείωσις· ὅ̣[σοι το]ίν̣ υ̣ ν̣ ̣ | ἀπὸ τῆς οἰκε[ι]ώσεως | εἰσάγουσι τὴν δι[κ]αι[ο]|σύνην, εἰ μὲν λέγου|σιν ἴσην αὐτοῦ τε πρὸς | α̣ὑτὸν καὶ πρ̣ὸ̣ς̣ [τὸν ἔ]|[σ]χατον Μυσῶν τεθέν|τος μὲν τούτου σώζε|ται ἡ δικαιοσ[ύ]νη, οὐ | συγχωρεῖται [δ]ὲ [εἶ]|ν̣αι ἴσην· π̣αρὰ γὰ[ρ τὴν] ἐνάργειάν ἐστιν [κ]α[ὶ] τὴν συναίσθησιν. (5. 18–36 Diels-­Schubart = LS 57H) We are oriented towards members of the same species. But one is more oriented towards one’s own citizens. For orientation varies in its intensity. So those people [the Stoics] who derive justice from orientation, if, on the one hand, they are saying that a man’s orientation to himself is equal to his orientation to the most distant Mysian, their assumption preserves justice; on the other hand, no one agrees with them that the orientation is equal. That is contrary to plain fact and one’s self-­awareness.

The basic critique is that it is simply false to claim that we nat­ur­al­ly care just as much for those who are at a relational distance from us as for those who are connected to us by blood or by close social relation­ ships. The anonymous commentator does not argue that we have no connection to people in the former group but does insist that what­ ever connection we have to them is mitigated by their distance from us. The ethical concern we have for others ‘varies in its intensity’.



Giving Gifts and Making Friends

267

The Stoics were not unaware of this difficulty, and some texts appear to address it directly. For example, Hierocles famously compares the agent’s situation to that of being at the centre of a set of concentric circles: ὅλως γὰρ ἕκαστος ἡμῶν οἷον κύκλοις πολλοῖς περιγέγραπται, τοῖς μὲν σμικροτέροις, τοῖς δὲ μείζοσι, καὶ τοῖς μὲν περιέχουσι, τοῖς δὲ περιεχομένοις, κατὰ τὰς διαϕόρους καὶ ἀνίσους πρὸς ἀλλήλους σχέσεις. . . . κατὰ τὸν ἐντεταμένον ἐστὶ περὶ τὴν δέουσαν ἑκάστων χρῆσιν τὸ ἐπισυνάγειν πως τοὺς κύκλους ὡς ἐπὶ τὸ κέντρον καὶ τῇ σπουδῇ μεταϕέρειν ἀεὶ τοὺς ἐκ τῶν περιεχόντων εἰς τοὺς περιεχομένους. . . . ἀϕαιρήσεται μὲν γάρ τι τῆς εὐνοίας τὸ καθ’ αἷμα διάστημα πλέον ὄν· ἡμῖν δ’ ὅμως σπουδαστέα περὶ τὴν ἐξομοίωσίν ἐστιν. (Stob. 4. 27. 23, 671. 7–672. 18 Wachsmuth–­ Hense = LS 57G 1, 5–6 in part) Each one of us is as it were entirely encompassed by many circles, some smaller, others larger, the latter enclosing the former on the basis of their different and unequal dispositions relative to each other. . . . [I]t is the task of a well-­tempered man, in his proper treatment of each group, to draw the circles together somehow towards the centre, and to keep zealously transferring those from the enclosing circles into the enclosed ones. . . . For although the greater distance in blood will remove some affection, we must still try hard to assimilate them.

Hierocles’ depiction of our social situation acknowledges that we are born into a world in which many people are at a social remove from us, and it concedes that our concern for others decreases as relational distance increases. However, Hierocles also thinks that we can expand our spheres of ethical concern by working to trans­ fer people from the outer circles inwards. Unfortunately, Hierocles’ own recommendations for how to accomplish the transfer do not inspire much confidence. His main suggestion is that we begin by treating the people in each circle as if they were actually one circle closer to us. Thus, we are to treat ­cousins as brothers, more distant relatives as cousins, neighbours as distant relatives, and so on. This approach is to be supported by the terms of address we use for others. In addition to treating ­cousins as brothers, we should actually call cousins brothers (and similarly for the other members of each circle). It is certainly possible that adopting this kind of behaviour could help increase concern for others, but it is doubtful that the practice would get us very far.15 If I were faced with a choice between saving my actual brother and a 15  See J. Annas, The Morality of Happiness (Oxford, 1993), 268–7 for a similar objection.

268

Allison Piñeros Glasscock

cousin whom I am calling my brother, would I really approach the situation as though the choice is between saving one or the other of my actual brothers? The Stoics’ anonymous critic makes a similar remark: εἰ δὲ καὶ α̣[ὐτ]οὶ ϕήσου|σ̣ι ̣ ἐπιτεί[ν]εσ̣ θα[ι] τὴν | οἰκεί̣ ω ̣ [σιν, ἔσ]τ̣αι μὲν | ϕιλανθρ[ωπί]α, ἐλέγξου|σι δὲ τ[ούτους α]ἱ περιστάσεις̣ ̣ [ναυαγῶ]ν, ὅ|που ἀνά̣γ̣ [κη μό]νο̣ν | σώζεσθαι τὸ̣ν̣ ἕτ̣ε|ρον αὐτῶν. (Anon., In Plat. Theaet. 6. 17–25 Diels-­Schubart = LS 57H)16 If, on the other hand, they themselves should say that orientation can be intensified, we may grant the existence of philanthropy, but the situations of shipwrecked sailors will refute them, where it is inevitable that only one of two survive.

The example of the shipwrecked sailors suggests that there are serious constraints on the expansion of the sphere of ethical con­ cern, constraints that cannot be overcome even by the most dili­ gent rethinking (and relabelling) of one’s social relationships. Recognizing this problem sharpens the question with which I began. If we want to know how to make progress towards the Stoic ethical ideal, we will need to know whether and to what extent ethical concern for others can be cultivated towards those at a relational distance from us. It is important to note that this ques­ tion is intermediate between two related questions. The first of these is the question of how we develop ethical concern for those close to us, for example, blood relatives. I shall set aside this ques­ tion because from the perspective of the Stoic project it seems less pressing. The Stoics have good empirical evidence for the claim that we do, in fact, care for the good of those close to us (though we may be confused about what their good actually is), and for most of us this concern develops naturally, without the investment of much (or any) conscious effort. The second related question is how we develop concern for all human beings. This is a question of how 16  Accepting LS’s supplement [ναυαγῶ]ν. In this passage the commentator draws on an example that the Stoics themselves use. Cicero (Off. 3. 90) reports a case that he attributes to Hecaton, in which two sages are shipwrecked and must decide which of them gets to live. The Stoics argue that the sages will not make this decision on the basis of self-­interest but rather on the basis of what is most in accordance with nature (and if nature is silent on this point, they will simply draw lots). The com­ mentator’s remark turns the Stoics’ own example against them by appealing to what will seem (to many) the more intuitive analysis of the scenario (cf. LS ii. 349 ad 57H, line 18). I am grateful to an anonymous referee for the reference.



Giving Gifts and Making Friends

269

we attain the most expansive version of the ethical ideal, and in Section 5 I shall make a few, more speculative remarks about how the view I propose can address this question. But my main focus here will be the intermediate question of how we first move beyond our concern for those closest to us to develop ethical concern for individuals with whom we are not closely connected. How do we come to see them as belonging to the same rational community as us and thus to regard their good as integral to our own? Do the Stoics offer a method of achieving this aim that can overcome the con­ straints that self-­interest and lack of blood ties place on our affections?17 My proposal is that Seneca’s De beneficiis provides the material for just such a method.

3.  Benefaction, joint activity, and friendship Seneca’s goal in De beneficiis is twofold. First, he aims to provide a Stoic analysis of the practice of gift exchange. Second, he aims to give his audience—­ non-­ sages striving to become virtuous (henceforth, progressors)—concrete guidance for how to engage in bene­fac­tion. That Seneca’s take on benefaction is Stoic emerges clearly from his definition of a beneficium or ‘benefit’ in Book 1 of the treatise: Quid est ergo beneficium? Benevola actio tribuens gaudium capiensque tribuendo in id, quod facit, prona et sponte sua parata. Itaque non, quid fiat aut quid detur, refert, sed qua mente, quia beneficium non in eo, quod 17  In this section, I have emphasized the apparent chasm between ethical concern for those closely related to us and for those at a relational distance. However, one might argue that the gap can be bridged without introducing a wholly new method for the development of ethical concern. A parent who has come to appreciate what is truly good for her child may then behave more justly towards everyone by refusing to act in ways that gain her child an unfair advantage over everyone else. (Thanks to an anonymous referee for the objection and the example.) Notice, however, that although the parent would then be behaving more justly towards a broader group of people, she is not thereby acting out of concern for the good of those others. Her actions are still motivated by a concern for her child’s best interest (she wants to cultivate virtue in her child, not vice). We still need to explain how the parent comes to see the good of others as an object of her ethical concern, and this process is likely to be complicated by the fact that a central motivation for the parent’s concern for her child is absent in the case of those at a relational remove. The parent is invested in the child’s well-­being because the child is her own, but in what sense do the ­people around her belong to her?

270

Allison Piñeros Glasscock

fit aut datur, consistit, sed in ipso dantis aut facientis animo. Magnum autem esse inter ista discrimen vel ex hoc intellegas licet, quod beneficium utique bonum est, id autem, quod fit aut datur, nec bonum nec malum est. (Ben. 1. 6. 1–2) So what is a benefit? It is a well-­intentioned action that confers joy and in so doing derives joy, inclined towards and willingly prepared for doing what it does. And so it matters not what is done or what is given, but with what attitude, since the benefit consists not in what is done or given but rather in the intention of the giver or agent. You can see how big a differ­ ence there is between them by reflecting that a benefit is unconditionally good, while what is done or given is neither good nor bad.

This is an application of the familiar Stoic point that the only good things are virtue and what participates in it.18 From the perspective of this theory of value, the things typically exchanged in an instance of benefaction (e.g. money) are not genuinely good and thus not genuinely beneficial. The only thing that is eligible to have real value is the intention or mind of the giver, and that only to the extent that it manifests the agent’s virtue.19 The link between benefaction and virtue connects to another key aspect of Seneca’s analysis. In keeping with orthodox Stoicism, Seneca holds that only sages can engage in genuine benefaction. That is because only sages know (scire) how to give and receive benefits.20 Nevertheless, Seneca allows that progressors can engage in an imperfect form of benefaction. In Epistulae morales 81. 13, discussing recipients, he claims that ‘the foolish person does also render thanks to the extent of his knowledge and ability. He is defi­ cient in knowledge rather than willingness.’21 Similarly, in De beneficiis 5. 4. 3–5. 1, Seneca argues that someone who owes a benefit to Socrates or Diogenes would be outdone (vincar) in benefits (since virtuous people like these do not need any of the material goods that the recipient might try to give them in return). But that the 18  See n.1 above. 19  Cf. M. Lentano, ‘De beneficiis’, in G. Damschen and A. Heil (eds.), Brill’s Companion to Seneca: Philosopher and Dramatist (Leiden, 2014), 201–6 at 202. Throughout the text, Seneca refers both to good intention (bona voluntas) and to the good mind (bonus animus). For my purposes, the distinction between the two is not important. 20 See Ep. 81. 10–14. Cf. Ben. 5. 12. 5–7 and Plut. Comm. not. 21, 1068 d = SVF 3. 672. 21 Sen. Ep. 81. 13: Stultus quoque, utcumque scit et quemadmodum potest, ­referat; scientia illi potius quam voluntas desit.



Giving Gifts and Making Friends

271

non-­sage can be outdone in benefits implies that they can, neverthe­ less, engage in benefaction (and they can even do so with those who are already virtuous; cf. Ben. 7. 12. 1). The point is somewhat complicated by Seneca’s remarks a few sections later. At De beneficiis 5. 12. 5–7, Seneca says that bad ­people cannot give or receive benefits. The reason for this is that ‘no one can give what he does not have; such a person lacks the willingness to benefit’.22 But here I think Seneca is making a tech­ nical point. In the same passage, he says that ‘bad and good are incompatible and cannot be combined’.23 It follows that a bad per­ son cannot have good intentions, for that would require the com­ bin­ation of two incompatible things. Thus, in order to have a truly good intention one must oneself be good or virtuous. However, it is consistent with this that even a (technically) bad person might want very much to exchange good intentions with others, despite the fact that he currently doesn’t have what it takes and that he might take active steps to acquire this ability. Seneca clearly thinks his readers can work to refocus their attention from the material things exchanged in benefaction to the attitudes and intentions that motivate such exchanges. More specifically, the progressor who internalizes and acts on Seneca’s advice about how best to give and receive benefits will begin to think and act in ways that resemble the sage’s thought and action (cf. Ben. 7. 2. 1).24 Although the pro­ gressor’s acts of beneficence will not qualify as acts of virtue, his performance of this proper function will, nevertheless, approximate the sage’s.25 22  Ben. 5. 12. 7: nemo potest, quod non habet, dare; hic bene faciendi voluntate caret. 23  Ben. 5. 12. 5: mala bonaque dissentiunt nec in unum eunt. 24  Takaki provides an illuminating account of how progressor engagement in different types of benefaction (benefits that are necessary, those that are useful, and those that are pleasant) might prepare one for the real business of exchanging genu­ inely good intentions (‘Communication’, 307–10). I am not entirely persuaded by Takaki’s claim that these types of benefaction form a hierarchy through which the progressor must ascend in order to attain genuinely good intentions. (It’s true that Seneca introduces the three types of benefits in order, but that may merely be to indicate the relative importance of the material goods involved in each type.) However, Takaki’s discussion offers helpful insight into how engagement in bene­ fac­tion might draw one’s attention to ‘the very worthiness of, and difficulty in achieving, sagely virtues’ (309). 25  Seneca refers to benefaction as both a proper function (officium (Ben. 5. 1. 1) corresponds to the Greek technical term kathēkon) and a right action (recte factum (Ben. 1. 5. 3) corresponds to the Greek technical term katorthōma). As a proper

272

Allison Piñeros Glasscock

We can already see that the practice of benefaction has implica­ tions for ethical development. The determined progressor can engage in it with an eye to cultivating virtue in herself. But the practice has a more specific application as well. It can be used to generate con­ cern for the good of others, and it accomplishes this via the forma­ tion and cultivation of friendship (amicitia). Seneca repeatedly links benefaction with friendship. In Book 2, he calls benefaction ‘that most sacred bond, from which friendship arises’ (sacratissi­ mum ius, ex quo amicitia oritur, Ben. 2. 18. 5–6). Later, he cautions potential recipients not to accept benefits from just anyone, since to accept a benefit is to ‘enter into friendship, a relationship that binds similar people together’ (in amicitiam, quae similes iungit, Ben. 2. 21. 2). He also notes that one of the defining features of benefac­ tion is that, in it, the giver transcends the bounds of professional duty to enter into a more personal relationship with the recipient. Thus, for example, we owe gratitude to doctors and teachers who have shown us special regard because ‘from being a doctor or a teacher they turn into a friend’ (ex medico et praeceptore in ami­ cum transeunt, Ben. 6. 16. 1). There is a simple explanation for why Seneca thinks that bene­ fac­tion gives rise to friendship: benefaction is personalized (Ben. 6. 7. 2). Thus, in the discussion of gratitude owed to doctors and teachers quoted above, Seneca emphasizes that it is the personal atten­ tion bestowed by one’s doctor or teacher that makes them eligible to receive gratitude from us and also transforms them into friends.26 function, benefaction is among the activities that reason tells us to engage in. Even non-­sages can perform proper functions. But only the sage can transform the per­ formance of a proper function into a right action (see Ben. 7. 17. 1 and cf. the remarks in Cic., Fin. 3. 58–9 = LS 59F and Stob. 2. 7. 8, 85. 12–86. 4 Wachsmuth–­ Hense = LS 59B). For helpful discussion of Seneca’s pedagogic strategy in De beneficiis and his categorization of benefaction as both proper function and right action, see Griffin, Society, ch. 7, esp. 125–32. Jula Wildberger also offers illuminating dis­ cussions of how Seneca envisions progressors engaging in analogues of virtuous action in ‘Care of the Self and Social Bonding in Seneca: Recruiting Readers for a Global Network of Progressor Friends’, Vita Latina, 197 (2018), 117–30 and in ‘Amicitia and Eros: Seneca’s Adaptation of a Stoic Concept of Friendship for Roman Men in Progress’ [‘Amicitia’], in G. M. Müller and F. M. Zini (eds.), Philosophie in Rom—­Römische Philosophie?: Kultur-, Literatur-, und Philosophiegeschichtliche Perspektiven (Berlin, 2017), 387–425, esp. 395–7. 26  Seneca’s position on the personal nature of benefaction and friendship distin­ guishes him from earlier Stoics (and perhaps from Stoics more generally). See G. Lesses, ‘Austere Friends: The Stoics and Friendship’ [‘Austere Friends’], Apeiron, 26 (1993), 57–75, for a discussion of the impersonal nature of Stoic friendship.



Giving Gifts and Making Friends

273

However, Seneca’s analysis of benefaction also secures an even stronger link between benefaction and friendship. The Stoic view of friendship emphasizes three central features:27 first, a friendship is a community of those who are similar to one another; second, they identify with each other; finally, they regard themselves as having goods in common. Thus, Diogenes Laertius reports that ‘[t]hey [sc. the Stoics] say that friendship exists only among the virtuous, on account of their similarity. They describe friendship as a cer­ tain sharing [or ‘community’, κοινωνίαν τινά] of life’s wherewithal, since we treat our friends as we treat ourselves.’28 Diogenes Laertius also reports that when someone asked Zeno what a friend is, he replied, ‘Another me’ (ἄλλος . . . ἐγώ; D.L. 7. 23 = SVF 1. 324).29 And from Stobaeus we learn that all virtuous people (a class that includes all friends) have ‘all good things in common’30 and ‘live harmoniously with one another because they agree about matters concerning life’.31 The same features appear in Seneca’s account of friendship. We saw above that Seneca thinks that a friendship is a relationship between people who are similar to one another (Ben. 2. 21. 2). In  Epistulae morales, Seneca claims that friendship involves two minds entering into a community that is based on this similarity: it arises ‘where equal willingness draws minds into a companionship of honourable intentions’ (cum animos in societatem honesta cupiendi par voluntas trahit, 6. 3). Within this community, the two friends identify with each other, with the result that each friend regards what is good for the other as also being good for him or herself. Thus, Seneca concludes the passage just quoted by noting that true 27 This is a necessarily brief overview of Stoic friendship. For more detailed discussions, see M. Graver, Stoicism and Emotion [Emotion] (Chicago, 2007), ch. 8; A.  A.  Long, ‘Friendship and Friends in the Stoic Theory of the Good Life’, in D. Caluori (ed.), Thinking about Friendship: Historical and Contemporary Philosophical Perspectives (London, 2012), 218–39; and G. Reydams-­Schils, ‘From Self-­Sufficiency to Human Bonding’, in ead., The Roman Stoics: Self, Responsibility, and Affection (Chicago, 2005), 53–82. Wildberger, ‘Amicitia’, 389–95 is also helpful. 28  D.L. 7. 124 = LS 67P: λέγουσι δὲ καὶ τὴν ϕιλίαν ἐν μόνοις τοῖς σπουδαίοις εἶναι, διὰ τὴν ὁμοιότητα· ϕασὶ δ´ αὐτὴν κοινωνίαν τινὰ εἶναι τῶν κατὰ τὸν βίον, χρωμένων ἡμῶν τοῖς ϕίλοις ὡς ἑαυτοῖς. 29  See also Cic., Fin. 3. 70 = SVF 3. 348 and Sen. Ep. 109. 15; cf. Ben. 2. 15. 1. 30  Stob. 2. 7. 11 i, 101. 21–2 Wachsmuth–­Hense = LS 60P: τὰ δ’ ἀγαθὰ πάντα κοινὰ εἶναι τῶν σπουδαίων. 31  Stob. 2. 7. 11 b, 94. 3–4 = SVF 3. 625: ὁμονοεῖν ἀλλήλοις διὰ τὸ συμϕωνεῖν ἐν τοῖς κατὰ τὸν βίον (translation is my own). See also Stob. 2. 7. 11 c, 94. 21–4 = SVF 3. 98.

274

Allison Piñeros Glasscock

friends ‘know that everything they have is held in common’ (sciunt enim ipsos omnia habere communia, 6. 3). In Epistulae morales 109. 15, he also says that ‘it is in accordance with nature to embrace one’s friends and to take as much pleasure in their action as if it were one’s own’.32 My proposal is that engagement in benefaction leads to precisely these core features of friendship, and that it does so because (1) benefaction is a joint activity, that is, it is an activity with multiple participants working in relation to one another to bring it about, and because (2) it is an activity that is intrinsically oriented towards the good.33 My evidence for the claim that benefaction is a joint activity comes from Seneca’s account of what benefaction consists in. In Book 5 of De beneficiis, Seneca, rejecting the possibility that one can bene­ fit oneself, argues that ‘ “benefit” must be included among those terms that require a second person. Certain acts, though they are honourable, supremely noble, and of consummate virtue, cannot happen without another person’.34 Seneca’s central point is that benefaction must be directed at a person who is not oneself. However, he is also making a stronger claim. One of the reasons that Seneca gives for the classification of benefaction as essentially other-­directed is that benefaction involves both giving and receiving. Seneca claims that in order for these two acts to remain distinct (as they must if there is to be genuine giving), they must be performed by non-­identical parties (5. 10. 1–2). That is, bene­fac­tion requires not just the presence but the participation of a second person (more on the nature of this participation below). It is worth noting that 32  Praeterea secundum naturam est et amicos conplecti et amicorum actu ut suo proprioque laetari, with actu (with the MSS) for Reynolds’s auctu. I will return to the significance of this choice at the end of this section. 33  I elaborate further on the notion of a joint or shared activity below. My think­ ing about this issue has been aided by A. S. Roth, ‘Shared Agency’, in E. N. Zalta (ed.) The Stanford Encyclopedia of Philosophy (Summer 2017 edn), https://plato. stanford.edu/archives/sum2017/entries/shared-­agency/, accessed 11 February 2022, especially the idea that ‘shared activity is distinguished from a mere aggregation of individual acts by a structure of appropriately related participatory intentions across different individuals’ (§3). On my view, it is the fact that the giver and re­cipi­ ent have intentions that are directed towards each other that enables them to share the activity of benefaction. 34  Ben. 5. 10. 4: cum inter ea sit habendum beneficium, quae secundam personam desiderant. quaedam, cum sint honesta, pulcherrima, summae virtutis, nisi in altero non habent locum.



Giving Gifts and Making Friends

275

this claim is actually implied by one plausible in­ter­pret­ation of Seneca’s Book 1 definition of benefaction as ‘a well-­intentioned action that confers joy and in so doing derives joy’ ([b]enevola actio tribuens gaudium capiensque tribuendo, 1. 6. 1). The present participles tribuens and capiens may indicate imperfective or sim­ ple aspect. If the participles indicate the simple aspect, Seneca is defining benefaction in terms of what it in fact accomplishes. On this interpretation, benefaction must culminate in the mutual conferring of joy; anything that falls short of this does not count as genuine benefaction at all. This is just what we would expect, given Seneca’s stark claim in Book 5. However, the passage is not decisive. As I noted, the participles might also be understood to have the imperfective aspect. In that case, to say that benefaction confers and derives joy might mean that benefaction intrinsically aims at so doing or wants to accomplish this (i.e. this might be a use of the conative imperfect). Clearly, this interpretation does not entail that benefaction only takes place when good intentions are reciprocated. On the other hand, it does not rule out this possibility either. It might be true both that bene­ fac­tion intrinsically aims at the reciprocation of joy and that it does not take place unless joy is reciprocated. Two additional passages provide further evidence for this account of benefiting, as well as greater insight into what it implies. In Book 2, Seneca addresses the question of whether someone who merely accepts a benefit with a reciprocal good intention has adequately returned it. Underlying this question is the worry that Seneca’s account of benefaction, which focuses entirely on the intentions of the giver and recipient, cheapens it. It is easy to have good feelings towards someone who has done you a good turn, much harder to return the favour by giving up one’s own time or resources. Seneca’s response is as follows: Quotiens, quod proposuit, quisque consequitur, capit operis sui fructum. Qui beneficium dat, quid proponit? prodesse ei, cui dat, et voluptati esse. Si, quod voluit, effecit pervenitque ad me animus eius ac mutuo gaudio adfecit, tulit, quod petit. Non enim in vicem aliquid sibi reddi voluit; aut non fuit beneficium, sed negotiatio. Bene navigavit, qui quem destinavit portum tenuit; teli iactus certae manus peregit officium, si petita percussit; beneficium qui dat, vult excipi grate: habet, quod voluit, si bene acceptum est. (Ben. 2. 31. 2–3)

276

Allison Piñeros Glasscock

Whenever someone achieves his intent, he gets the fruit of his labours. What is the intention of the person who gives a benefit? To be useful to the recipient and to give him pleasure. If he achieved this objective and if his intention got through to me and we felt mutual joy, then he got what he was aiming at. For he did not want to be given something in exchange; other­ wise, it was not a benefit but a business deal. The man who gets to the port he was aiming for has had a successful sailing. A spear thrown by an expert hand does the job if it strikes its target. The giver of a benefit wants it to be received with gratitude; if it is received properly, he has attained his aim.

Seneca’s response turns on two central claims. First, he insists that to achieve one’s aim is a reward in itself. That is why a person ‘gets the fruit of his labours’ simply by achieving what he set out to do. The examples of the spear hitting the target and the navigator arriv­ ing at his chosen port illustrate this feature of achievement. Second, Seneca claims that the giver’s aim is to have his benefit well received by the recipient. The good reception of a benefit is here glossed variously as the feeling of pleasure (voluptas) or joy (gaudium) or the benefit’s being gratefully received (excipi grate). A little later it is described as the proffering of ‘good intentions in return for good intentions’ (bono animo bonum optulit, 2. 34. 1).35 Like sailing or spear-­throwing, benefiting is a single activity. However, unlike the former activities, successful benefaction requires contributions from two distinct parties.36 It takes place only when (1) the giver directs good intentions towards the re­cipi­ent and when (2) the recipient, feeling joy or pleasure, reciprocates those good intentions in the form of gratitude.37 35  These varied descriptions of the good reception of a benefit are perfectly com­ patible with one another. In an instance of benefaction, the giver makes his good intentions known to the recipient; awareness of the giver’s good intentions sparks joy or pleasure in the recipient, and the recipient’s expression of that joy/pleasure constitutes a ‘grateful reception’ of the benefit, i.e. the reciprocation of good inten­ tions. What the recipient feels depends on whether they are a sage or a non-­sage. The sage feels the eupatheia of joy; the non-­sage pleasure. See D.L. 7. 116 = LS 65F for the distinction between joy and pleasure and Sen., Ep. 59. 2 for the allocation of joy to the sage. On joy and pleasure in Stoic ethics, see T. Brennan, ‘Stoic Moral Psychology’ [‘Psychology’], in B. Inwood (ed.), The Cambridge Companion to the Stoics (Cambridge, 2003), 257–94 at 269–71; Graver, Emotion, 52–3; and Inwood, Human Action, 173–5. 36  Sailing might also involve contributions from multiple parties (as when a captain and her crew are working together to bring the ship safely to port), but it need not. 37  For simplicity’s sake, I have been treating the joy of the recipient as being at least conceptually distinct from the good intentions conferred by the giver. However, D. Konstan, ‘The Joy of Giving: Seneca De beneficiis 1. 6. 1’, in E. K. Emilsson,



Giving Gifts and Making Friends

277

The same view of benefaction emerges shortly afterwards, when Seneca responds to the objection that someone who doesn’t return a favour by giving material goods is like someone playing a game of catch who fails to return the ball. In this case (the objection goes) there is something deficient in the game—­no matter how grateful the recipient is and no matter how much she wants to return the ball.38 Seneca responds: Nolo diutius hoc refellere; existimemus ita esse, desit aliquid lusui, non lusori; sic et in hoc, de quo disputamus, deest aliquid rei datae, cui par alia debetur, non animo, qui animum parem sibi nanctus est et, quantum in illo est, quod voluit, effecit. (Ben. 2. 32. 4) I do not want to argue this point any further. Let us suppose that it is so and that there is something deficient in the playing of the game—­but not in the player. Similarly, in the case we are discussing there is indeed a defi­ ciency with respect to the thing given, which lacks its counterpart; but there is no deficiency with respect to the mind, for it has found another mind which shares its attitude and has achieved what it intended insofar as it was able to do so.

The comparison of benefaction with a game of catch (and Seneca’s ready adoption of the analogy) further supports the idea that bene­ fac­tion is a joint activity, and I return to the image below.39 Here it A. Maravela, and M. Skoie (eds.), Paradeigmata: Studies in Honour of Øivind Andersen (Athens: 2014), 171–6, makes an excellent case for thinking that joy is not merely a response to something already given (e.g. money) but is the very thing that is given at the non-­material level. This proposal takes seriously Seneca’s claim (in the def­in­ ition at Ben. 1. 6. 1–2) that benefaction confers joy (as, at the material level, one might confer money on another). It also raises the interesting possibility that Seneca may ultimately want to identify the recipient’s joy with the good intentions being proffered by the giver: my good intentions are—­or at least become—­your joy. I will not pursue this line of thought any further here. But note that if that is right, it further strengthens the case I am making for the shared nature of benefaction. 38  The use of the analogy with games has a history in Stoicism. Epictetus (Diss. 2. 5) employs it to respond to the criticism that it is incoherent to spend one’s life pursuing and selecting various indifferents when those indifferents are (as their name implies) not good. For helpful discussion of the analogy in the context of defences of the Stoic conception of the end, see T. Brennan, The Stoic Life: Emotions, Duties, and Fate [Stoic Life] (Oxford, 2006), 146–51, and J. Klein, ‘Of Archery and Virtue: Ancient and Modern Conceptions of Value’, Philosophers’ Imprint, 14 (2014), 1–16 at 13–15. Here Seneca envisions a spin on the analogy. It is brought in, not to explain what justifies the pursuit of indifferents, but to press the worry that (pace Seneca) the game of benefaction requires the return of material goods. 39  It is true that the analogy originates from the objector, and thus one might conclude that Seneca merely entertains it for dialectical purposes. However, we have evidence that the analogy was one the Stoics relied on in other contexts; see

278

Allison Piñeros Glasscock

suffices to note that Seneca’s response once again implies that suc­ cessful benefaction requires the reciprocation of the giver’s good intentions by the recipient. His explanation for why there is no ‘deficiency with respect to the mind’ in this particular case is that the mind has achieved its aim, which was to find another mind that is its equal. That is to say, the giver has found someone who has responded gratefully to the giver’s good intentions and the recipient has found someone who has received his gratitude.40 Reciprocation of good intentions is thus introduced as a necessary condition of successful benefaction.41 If there were to be no reciprocation of good intentions, then there would be a deficiency in this instance of

n. 38 above. Seneca himself appeals to the game analogy at Ben. 2. 17. 3–5, citing Chrysippus as its source, and there is no evidence to suggest that he disowns it here, though he is eager to resist certain ways of fleshing it out. At the beginning of the text quoted above (Ben. 2. 32. 4), when Seneca makes a dialectical concession to his interlocutor (‘let us suppose that it is so’, existimemus ita esse), the supposition in question is only that the game of benefaction is ‘incomplete’ when a material thing is not returned in kind. (This is a mere supposition for Seneca, because—­as we are about to learn—­his considered view is that the game will only be rendered incom­ plete or deficient when the giver’s intention is not returned.) Thanks to Victor Caston for pressing me on this point. 40  It is an oddity of the passage that Seneca appears to switch from the perspec­ tive of the recipient (and whether his gratitude is a sufficient repayment of the benefit) to the perspective of the giver (for this observation, I am indebted to a discussion in Brad Inwood’s course on De beneficiis at Yale in the autumn of 2016). However, I think the remark is telling regardless of whether it is read from the perspective of the recipient or the giver. In either case, some kind of getting through to the other is required for the achievement of the agent’s aim. 41  Seneca occasionally says that benefits can be given even though they are not received gratefully. For example, at Ben. 5. 20. 2 he insists, ‘We should not pay regard to whether a person grieves on receiving a benefit, but to whether he ought to rejoice’ (Non est spectandum, an doleat aliquis beneficio accepto, sed an gaudere debeat. Thanks to an anonymous referee for bringing this passage to my atten­ tion.) The passages discussed above would seem to rule out precisely this kind of case. But there is an easy way to dispel the inconsistency without having to write off either set of passages. ‘Benefit’ or beneficium is an ambiguous term and can refer both to the completed activity of benefaction (which includes the grateful response of the recipient) and to the giver’s bestowing of good intentions on another, considered in isolation. I propose that beneficium is used in the latter sense here, to pick out what has happened when a giver has successfully made another aware of her good intentions, despite the fact that those intentions are not reciprocated (and so despite the fact that benefaction, in the inclusive sense, has not taken place). As Seneca puts it in Book 2, the game (lusus) of benefaction ‘can only carry on as long as both want to play’ (qui non potest, nisi consentitur, extendi, Ben. 2. 17. 5).



Giving Gifts and Making Friends

279

benefaction. In such cases the giver has attempted to benefit the recipient but has failed to achieve his aim.42 One might argue that, despite the fact that both parties make contributions to benefaction, the recipient’s contribution is merely passive. If that is the case, benefaction is not a shared activity. It belongs instead to the giver, who is acting on the recipient as a doc­ tor acts on a patient. However, the texts we have considered assign the recipient an active role in benefaction. The recipient’s feeling of joy or pleasure is not automatically triggered by the giver’s good intentions. Rather, such feelings are (or are based on) judgements that are up to the recipient.43 She decides whether she wants to com­ plete the conferral of a benefit by feeling and expressing joy and so reciprocating the giver’s good intentions.44 The volitional aspect of an agent’s response is especially clear in the case of benefaction. As Seneca himself warns, one common response to the proffering of a benefit is resentment. The recipient may initially feel an impulse to reject the benefit because they do not want to be indebted to anyone. A well-­given benefit can help to allay this reaction, but ultimately it is up to the recipient to respond appropriately.45 Can we say anything more specific about how the actions of the giver and recipient combine to create a joint activity of bene­fac­ tion? As we saw above (Ben. 2. 32. 4, cf. 2. 17. 3–5), benefaction is compared to a game of catch. The analogy is telling, for games offer an example of a joint activity with a structure similar to that of benefaction. In a game, the players often undertake different tasks (e.g. one throws the ball, while the other catches it; each 42  One might object that it is a strike against my interpretation that it conflicts with the Stoic commitment to the self-­sufficiency of the virtuous agent. This ana­ lysis of benefaction certainly does conflict with that commitment, but we should not therefore reject the analysis. I discuss the issue more fully in the second half of Section 4. 43  See Graver, Emotion, 32–4, for discussion of the relationship between emo­ tions and judgements in Stoicism. 44  For discussion of the rational character of emotion, see Brennan, Stoic Life, 90–108; id., ‘Psychology’, 265–9; Inwood, Human Action, 128–32; and B. Inwood, ‘Why Do Fools Fall in Love?’, Bulletin of the Institute of Classical Studies, 41 (1997), 55–69 at 62. 45 At Ben. 2. 17. 4–5, Seneca advises the giver to choose appropriately sized gifts so as to ward off hostile reactions from the recipient. At 2. 18. 7, he reminds the recipient that it is up to him to decide whether or not he is willing (utrum velis an non) to receive the gift. Thanks to an anonymous referee for encouraging me to develop this point.

280

Allison Piñeros Glasscock

places her own tokens on the board). Nevertheless, the players are at the same time participating in a single, shared activity: they are playing the game together. The reason for this is that the inten­ tional actions of each are directed towards the other (and any add­ ition­al game players). Each player makes her move with the aim of eliciting certain actions or reactions on the part of the other. So too in the case of benefaction. Although the giver and recipient under­ take different actions, they do so in relation to one another. The giver bestows good intentions with the aim of eliciting reciprocal good intentions from the recipient. The recipient expresses grati­ tude with the aim of returning what the giver has bestowed upon her. The giver and the recipient both share in the activity of bene­ fac­tion because they work together, via their relational actions, to bring it about. The second important feature of benefaction is that its aim is the reciprocal exchange of things that are genuinely good, namely, the good intentions of the participants. This distinguishes the activity from other types of virtuous action. In what we might call the ‘or­din­ary case’ of virtuous action, the agent selects an outcome that he deems to be in accordance with nature and does what he can to achieve it. For example, he might plan to win a political race or to dine at 7 p.m. Importantly, however, the intended outcomes (winning the race, dining at the appropriate time) are not themselves genuine goods. According to Stoic theory, they are merely preferred indifferents. These two features of benefaction make it especially suited to give rise to the core features of Stoic friendship that I described above. Because the giver and the recipient each make a contribu­ tion to the act of benefiting, they are, together, the source of a sin­ gle activity. This means that they have grounds to see themselves as a unity—­or better still as a community. The reciprocal nature of their joint activity also ensures that theirs is a community of equals. In fact when he discusses the link between benefaction and friend­ ship, Seneca often highlights the way in which the giver and re­cipi­ ent are or come to be equals. For example, in De beneficiis 2. 34. 1, he notes that benefaction involves the reciprocation of good inten­ tions ‘in a spirit of equality’ (ex aequo), which is an indication of friendship (amicitia). Elsewhere, he argues that the parties to bene­ fac­tion are equals precisely with respect to their good intentions, since good intentions do not admit of degrees (cf. Ben. 5. 4. 1–2 and Ep. 6. 6. 3). It is worth noting that this account of the relationship



Giving Gifts and Making Friends

281

between giver and recipient marks a striking departure from Aristotle’s. Aristotle begins from the idea that there is a strong asymmetry between the giver and the recipient: recipients are (qua recipients) the products of the activity of the giver.46 This meta­ physical asymmetry leads to an asymmetry in affection: givers love recipients more than recipients love givers. Seneca’s insistence on the equality of the participants in benefaction removes this prob­ lematic asymmetry, thereby making more plausible his claim that friendship (which requires equality) results from benefaction. Moreover, insofar as the giver and recipient are co-­agents, and (importantly) co-­agents in an activity that is intrinsically oriented towards the good, they will come to identify with one another and to regard the goods of each as being held in common.47 The idea that the parties to benefaction come to identify with one another finds confirmation in a passage from Seneca’s Epistulae morales. Earlier I quoted the following passage as evidence that Seneca endorses the Stoic position that friends treat each other as exten­ sions of themselves: ‘it is in accordance with nature to embrace one’s friends and to take as much pleasure in their action [actu] as if it were one’s own [ut suo proprioque]’.48 Interestingly, where I have translated ‘action’, Graver and Long have translated ‘advance­ ment’. The discrepancy arises from a textual emendation. Although all the manuscripts have ‘actu’,49 the editor of the Oxford Classical Text of the Epistulae morales, L. D. Reynolds, emends to ‘auctu’ (advancement), citing the editor of the Teubner edition, F. Haase, as the source of the emendation. Presumably, Haase and Reynolds 46 See NE 9. 7, 1168a5–10; cf. EE 7. 8, 1241b1–3. 47  The view I develop here is compatible with Graver’s interpretation of what it means for friends to have goods in common. Graver emphasizes that friends think of one another as ‘co-­agents’, with the result that they have ‘a community of motives as well as a community of interests’ (Emotion, 181–2). Like Graver, I think that friends regard one another as ‘co-­agents’. The interpretation I have proposed supplements Graver’s view by explaining how the friends’ community is grounded in the genuine co-­agency of benefaction. My interpretation is also sympathetic to Reydams-­Schils’s view that in the Roman Stoics we find an ‘upgrading’ of the sta­ tus of social relationships (‘Bonding’, 251). Reydams-­Schils argues that the upgrade is partly the result of the fact that reason turns out to have a relational aspect. I take it that the reciprocation of good intentions that constitutes benefaction provides a concrete example of this relational aspect. 48  Ep. 109. 15: Praeterea secundum naturam est et amicos conplecti et amicorum auctu ut suo proprioque laetari (trans. Graver and Long, with modifications). 49  As reported in Reynolds, OCT, vol. ii ad loc.

282

Allison Piñeros Glasscock

thought that it made better sense to endorse taking pleasure in the advancement of others, rather than taking pleasure in their action. However, on my proposal, we can make excellent sense of Seneca’s recommendation if we follow the manuscripts’ reading. In the first place, one should take pleasure in the actions of one’s friends as if they were one’s own, because, when my friend and I are engaged in the joint activity of benefaction, the action of my friend (the bene­fit­ing) is literally my own. Of course, not every action one’s friend does will be one’s own in this strong sense. Friends do not only act in relation to each other. However, once a community of agency is established, co-­agents in benefaction have a vested interest in all of each other’s actions. Seneca reminds his readers that it is important to be cautious about who one receives benefits from and gives them to, since one does not want to accidentally ally oneself with a bad person.50 On one level, this is purely a matter of optics. Being closely involved with a vicious person is not a good look. But there is a deeper problem too. Only good p ­ eople can benefit each other (since only they have the good intentions, the exchange of which comprises benefaction), and only good people can become friends. Thus, if a sage wants benefaction to actually occur—­and a friend­ ship to actually arise, it is essential for her to find a giver or recipient who is actually good. Similarly, for the progressor trying to approxi­ mate the activity and relationships of the sage, it will be necessary to seek out partners who are also committed to becoming good and acting virtuously. There is no reason to think that this interest in the character of the other is abandoned once an exchange of bene­ fits has occurred (or has seemed to occur). Indeed, there is every reason to think that the interest is heightened because more is at stake. Now, it is not just a matter of whether I should enter into a relationship with the person, but whether an existing alliance will turn out to be genuine after all. Thus, it is right for me to be pleased when I see my friend acting well, for this is additional confirmation that ours is truly a community of equal minds. Seneca’s analysis of benefaction also suggests that there is a strong sense in which co-­agents in benefaction (and, ultimately, 50 Cf. Ben 2. 21. 1, where Seneca strongly advises against allowing a male prosti­ tute to become one’s benefactor (e.g. by saving one’s life), and 2. 18. 5, quoted and discussed at the end of this section.



Giving Gifts and Making Friends

283

friends) have goods in common. As we saw earlier, the performance of benefaction is an act of virtue (or, in the case of progressors, an approximation of it). Since virtue is good, each party to bene­fac­ tion has a genuine good, insofar as he is engaged in an act of virtue. And, importantly, each agent possesses the numerically same good, that is, the activity of benefaction in which they are jointly engaged. Thus, benefaction is especially suited to the cultivation of friend­ ship, because the act that initiates the relationship—­as well as the other instances of benefaction that nourish it—­is itself a good that is literally shared by the co-­agents.51 It is helpful to contrast this strong sense of a common good with a weaker alternative. One might think that to say that friends have goods in common is just to say that they have the same kinds of goods in their souls (namely, the virtues).52 This interpretation denies that friends possess nu­mer­ic­ al­ly identical goods. However, Seneca himself seems to disavow this weaker alternative towards the end of De beneficiis. In Book 7, Seneca addresses the objection that friends cannot benefit each other since they already share their goods. As part of his response, he argues that: non enim mihi sic cum amico communia omnia sunt, quomodo cum socio, ut pars mea sit, pars illius, sed quomodo patri matrique communes liberi sunt, quibus cum duo sunt, non singuli singulos habent, sed singuli binos. (Ben. 7. 12. 1) all things are shared between me and my friend not in the way they are shared with a business partner (one part being mine, one part being his), but in the way the children are shared by the mother and the father. If they have two children, the parents do not have one each, but rather each has both.

To say that friends have goods in common merely insofar as they each possess the virtues resembles the business partner model of sharing that Seneca rejects. According to the weaker alternative, the friends are engaged in a joint venture in which each has a cer­ tain stake, represented by the virtues belonging to each friend. Just as the stock of an enterprise can be divided up among share­holders, so too the virtues present in a friendship can be allocated to each 51  Seneca holds that friendship involves repeated instances of benefaction (Ep. 9. 8–10). See also Ben. 2. 18. 5, quoted and discussed at the end of this section. 52  Lesses endorses this weaker interpretation of common goods (‘Austere Friends’, 73).

284

Allison Piñeros Glasscock

party. According to Seneca, however, this is precisely not how friends have goods in common. Rather, the relationship each friend has to the shared good is like the relationship that each member of a set of parents has to their child. The child, that is, belongs to both its parents jointly and cannot be allocated simply to one or the other. My strong interpretation of how friends have the good of benefaction in common aligns well with Seneca’s preferred analogy for what possession of common goods involves. As in the case of the child, the good that is the activity of benefaction is constituted by contributions from both parties and cannot be allocated to either party individually but only ever to both. Once a community of agency is initiated through benefaction, it is—­according to Seneca—­self-­sustaining. In fact, Seneca cites the self-­sustaining nature of benefaction as an explanation for why one should be careful about who one receives benefits from: et quidem diligentius quaerendus beneficii quam pecuniae creditor. Huic enim reddendum est, quantum accepi, et, si reddidi, solutus sum ac liber; at illi et plus solvendum est, et nihilo minus etiam relata gratia cohaere­ mus; debeo enim, cum reddidi, rursus incipere, manetque amicitia. (Ben. 2. 18. 5) in fact, I should be even more careful when seeking someone to be indebted to for a benefit than for money. The financial creditor has only to be paid back as much as I accepted, and once I pay him off, then I am free and clear. But I have a larger payment to make to the other creditor, and even after the favour has been returned, we are still linked to each other. For once I have paid him back, I must start again, and a friendship persists.

Thus, a single instance of benefaction has the potential to spark a long-­lasting relationship that has the central features of friendship, as Seneca and the Stoics conceived of it: the formation of a com­ munity of equals, the close identification with the other, and the possession of goods in common. 4.  The uniqueness of benefaction and self-­sufficiency Before turning to the connection between friendship and oikeiōsis, I  address two issues that the preceding analysis raises. The first concerns the status of benefaction relative to other types of virtuous action. The second concerns its implications for the Stoic commitment



Giving Gifts and Making Friends

285

to the self-­sufficiency of the virtuous person. I take the issues in turn. I have argued that benefaction is especially suited to the forma­ tion of friendship. Does that mean that it is the only action that can give rise to friendship? Seneca himself is unclear on this point. He claims that friendship arises from (oritur) benefaction, but that need not entail that friendship only results from benefaction (though, of course, it is consistent with such a claim). Still, a closer look reveals that benefaction is in a category apart from other virtuous actions. Consider, first, virtuous actions that are not joint, for example, the virtuous doctor’s fulfilment of his duties towards a patient. Fulfilment of such duties can certainly be an expression of the doctor’s virtue, but it will not automatically give rise to the core features of friend­ ship. For one thing, the doctor’s recognition that he has these duties towards another person presupposes that the two are in some sort of community already; his performance of them does not generate a new community. In addition, there is no reason to think that the fulfilment of these duties will directly lead both parties to identify with each other or to treat the other’s goods as belonging to both. The doctor might identify with the patient for the reason that the patient (once healthy) is, in a sense, his production, but this fact will not explain how or why the other party (the patient/product) comes to identify with the producer.53 Second, consider virtuous actions that do seem to involve shared or joint activity. Throughout De beneficiis (including some of the passages discussed above), Seneca contrasts benefaction with com­ mercial exchange and business partnerships. One might reasonably point out that at least in some cases these activities can be shared in precisely the way that I have argued benefaction is. Moreover, when undertaken by virtuous people, these shared activities will be 53  It is a strength of my view that it can explain how two people come to identify with each other in friendship and (more generally) in oikeiōsis. Others have tried to explain how we come to be properly oriented towards others by focusing on the status of one agent as the producer of the relationship. Thus, M.  W.  Blundell, ‘Parental Nature and Stoic οἰκείωσις’, Ancient Philosophy, 10 (1990), 221–42, exploring the implications of generative relationships for oikeiōsis, leans heavily on the idea that the producer comes to identify with and love his product as an instan­ tiation of himself and his virtue. According to Blundell, the producer–­product rela­ tionship explains why parents love their children, but it also explains why a virtuous person loves the people around him (their goodness is the result of his own virtuous action). However, this kind of explanation is inadequate for the reason mentioned above: it only explains one half of the explanandum. See also n. 47 above.

286

Allison Piñeros Glasscock

virtuous. It is relatively easy to explain why commercial exchanges will not result in friendship. They may be shared activities, but they lack one of the features that Seneca regards as essential to friend­ ship: equality between partners. The asymmetry of power between lender and borrower means that this relationship cannot by itself turn the participants into friends. The case of business partnerships is more difficult, since they do not require asymmetry. Nevertheless, they too are importantly different from benefaction. In benefaction the intentions of the participants are directed towards the good of the other. The giver aims to confer her good intentions on the re­cipi­ent so as to elicit reciprocal good intentions; the recipient aims to receive and return good intentions as a response to those proffered by the giver. By contrast, the aim of a business partner­ ship as such is not the goodness of the partner, but, for example, to make a return on an investment. It is certainly not inconceivable that the partners may move from caring about their shared activity and its intrinsic aim (making money) to caring about each other’s good. But the transition will not be seamless as it is in the case of benefaction, where the activity itself focuses on the good of the parties involved. In fact, it seems reasonable to think that in busi­ ness partnerships where this transfer of focus has taken place, the transfer is not due to the shared business activity but rather to the fact that the partners have taken their frequent interactions as opportunities to bestow benefits on one another, thus turning from business partners into friends.54 Thus, although we cannot rule out on textual grounds the possibility that other activities besides benefaction may result in friendship, I do think we have good philosophical reasons for thinking that benefaction is likely to be the (unique) source of friendship. And even if other friendship-­generating activities were to be identified, that would not undermine the fact that bene­fac­tion’s special nature makes it extremely well suited to the formation of friendship. The second issue my analysis raises concerns self-­sufficiency. According to orthodox Stoicism, the correctness or success of a virtuous action depends solely on the intention from which it 54  Recall that Seneca invokes a similar explanation for how doctors and teachers become the friends of their patients and pupils in Ben. 6. 16. 1, quoted at the begin­ ning of Section 3 above.



Giving Gifts and Making Friends

287

­ roceeds and not on its results. This places the virtue, and ultimately p the happiness, of the virtuous person entirely within her control.55 However, on the interpretation that I have developed, successful benefaction depends partly on the responsiveness of the recipient and, thus, is not up to the giver alone. If that is right, then there is at least one virtuous action that is not within the individual agent’s control and that appears to threaten her self-­sufficiency. This implication of my interpretation is shocking, but we should not conclude that the interpretation is, therefore, wrong. I believe that the challenge posed by benefaction is an instance of a more general problem that the Stoics face, one that I explore more fully else­ where.56 Briefly stated, the problem is that the Stoics are commit­ ted both to the view that the virtue of others is genuinely good and thus that enabling others to express virtue or become virtuous is a genuinely good thing and to the view that the virtuous person is self-­sufficient, able to attain each and every good at which he aims. Clearly, there will be occasions on which, despite one’s best efforts, the object of one’s ethical concern remains vicious; and yet it is also clear that bringing about virtue or virtuous action in others is not a matter of indifference (even of preferred indifference) for the virtuous agent. Seneca’s treatment of benefaction exposes this tension between the Stoics’ account of goodness and their well-­documented com­ mitment to self-­sufficiency. On the one hand, since benefaction aims at eliciting something that is genuinely good from another person (the reciprocation of good intentions), it should come as no surprise that Seneca makes the actual attainment of that good a condition for benefaction’s success. Seneca is simply following one key tenet of Stoicism where it leads. On the other hand, the resulting ana­ lysis of benefaction as a joint activity makes it abundantly clear that this tenet conflicts with the self-­sufficiency thesis. There is not enough evidence to establish that Seneca develops his account of benefaction precisely in order to draw out this difficulty. He might have, of course. We know that Seneca is an independent thinker,

55  For a statement of this view, see Cic., Fin. 3. 24 = LS 64H and 3. 32 = LS 59L; see also, D. Baltzly, ‘Stoicism’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2019 edn), https://plato.stanford.edu/archives/spr2019/entries/ stoicism/, accessed 4 June 2021, §5. 56  In a paper on Stoic self-­sufficiency (in progress).

288

Allison Piñeros Glasscock

ready and willing to disagree with his earlier counterparts.57 However, we also know that Seneca himself espouses a commitment to self-­ sufficiency.58 It may be that, like his predecessors, he failed to rec­ ognize the way in which this commitment conflicts with the Stoic account of what is good.

5.  Conclusion: Friendship and oikeiōsis Recall that my original question for the Stoics was this: how do we develop ethical concern for those at a relational distance? We are now in a position to appreciate the answer that Seneca’s De beneficiis offers us. The fact that benefaction is a joint activity makes it particularly (perhaps, as discussed in Section 4, uniquely) useful for the formation of friendship. Moreover, as I have argued, friends enter into community with one another. The community of friend­ ship is a small-­scale one. In the most basic case, there are only two members. However, it has the same features as the much larger rational community that is attained through social oikeiōsis. In both cases, as we have seen, community members identify with each other in a way that involves seeing the good of individual members as integral to the good of the whole and, thus, as integral to one’s own individual good. It also involves working to promote the good of others. The overlap between friendship and the Stoic ideal of the rational community means that, in a friendship, each friend has the ethical concern for the other that is the hallmark of social oikeiōsis. To make a friend is to put oneself into the ideal state (or, in the case of progressors, into an approximation of that state) with respect to that friend. Since benefaction gives rise to friendship, participation in benefaction is thus a means of working towards the Stoic ideal. The giver and recipient—­now friends—­expand their spheres of ethical concern to include each other. This account of how benefaction could figure in ethical forma­ tion is further supported by the fact that the practice of bene­fac­ tion is not constrained by pre-­existing social roles. Seneca devotes 57 See K.  M.  Vogt, ‘Seneca’, in E.  N.  Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2020 edn), https://plato.stanford.edu/archives/spr2020/entries/ seneca/, accessed 4 June 2021, §2. 1; cf. H. Hine, ‘Rome, the Cosmos, and the Emperor in Seneca’s Natural Questions’, Journal of Roman Studies, 96 (2006), 42–72 at 57. 58  See e.g. Seneca Ep. 9. 13–18 and 92. 3–4.



Giving Gifts and Making Friends

289

special attention to showing that even people who are in (otherwise) asymmetrical relationships can engage in the equalizing practice of benefaction. For example, slaves can benefit masters and children can benefit their parents (Ben. 3. 18. 1–22. 4 and 3. 29. 2–32. 6). But the implications of his arguments are far-­reaching. Benefaction—­ and so friendship—­is not restricted to those who already stand in a certain relationship to one another, much less does it require that its participants have a pre-­existing close relationship. Thus, engagement in benefaction offers a means of actively and deliber­ ately developing a disposition to care for one who is, as yet, at a relational distance. Put another way, it is a method of transferring those in Hierocles’ outer relational circles to the inner circles. But in contrast to Hierocles’ own proposal, this method does not rely on merely treating those in the outer circles as blood relatives. It is important to emphasize that the practice of benefaction is not itself the ground of the other-­regarding obligations involved in oikeiōsis. Benefaction does generate an obligation for the recipient (to reciprocate the good intentions and to make every effort to return the material goods given; see Ben. 5. 20. 2), and appreciation of this specific obligation will certainly help solidify the emerging friendship. However, as we saw in Cicero (Section 2, above), it is ultimately the fact that the other is a human being that explains why I must care for her. On the view I have advanced here, en­gaging in benefaction enables the agent to recognize the obligations she has to others by directing her attention towards what is really valuable in or about them. Just as the participants in benefaction come to see that it is the reciprocal good intentions that are the real source of value in the exchange (and not material goods), so too they come to see that it is precisely the friend’s goodness, and her mani­fest­ ation of that goodness in action, that is worthy of regard. I have focused primarily on what I have termed the intermediate question of how we come to care for those at a relational distance. In conclusion I want to briefly consider whether benefaction and friendship could get the progressor all the way to the full Stoic ideal. I said at the beginning of the paper that this ideal is ambi­ tious because it includes all people within its scope. The sage is supposed to care about the good of everyone. This kind of concern clearly goes well beyond the concern one has for individual friends. Nevertheless, I think benefaction sets the agent up for the large-­ scale project in two respects.

290

Allison Piñeros Glasscock

First, givers and recipients begin to develop the habits and ­ ispositions required to act on concern for the good of others. Of d course, some participants—­sages—­will already have these disposi­ tions. But many progressors, especially those at early stages of ethical development, will not. For them, the friendship that emerges from benefaction provides a concrete context in which to practise promoting the good of another. Someone who already has the dis­ positions (whether partially or fully formed) that are required to care for another will be much better prepared to enact such con­ cern even as its scope is enlarged.59 More importantly, as I noted above, the parties to benefaction become intimately familiar with what concern for another’s good requires (namely, promotion of the other’s virtue). This will be true even when both parties are progressors. They might be as yet incapable of acting virtuously (because they as yet lack virtue), but they can still appreciate that true benefaction consists in the exchange of good intentions and that this goodness is the sole source of the value of the exchange. But to realize that it is the friend’s goodness (or potential for it) that makes her the appropriate object of one’s ethical concern is also to realize that there is no reason to limit one’s concern to this particular friend. Anyone who is good or has the potential for goodness is an appropriate object of that concern. Thus, the practice of benefaction does prepare participants to actualize the Stoic ideal. It offers the progressor a mechanism for coming to understand why other people—­however distant—­deserve a place within that sphere, and it prepares him to act on that under­ standing, once achieved. The progressor is invited to embark on this ambitious and challenging project by simply making one friend at a time. BIBLIOGR A PH Y Annas, J., The Morality of Happiness (Oxford, 1993). Baltzly, D., ‘Stoicism’, in E.  N.  Zalta (ed.), The Stanford Encyclopedia of  Philosophy (Spring 2019 edn), https://plato.stanford.edu/archives/ spr2019/entries/stoicism/, accessed 4 June 2021. 59  For the idea that Stoic virtuous action is supported by trained dispositions, see Inwood, Human Action, 210–13 and A. A. Long, Stoic Studies (Cambridge, 1996), 168–9.



Giving Gifts and Making Friends

291

Blundell, M. W., ‘Parental Nature and Stoic οἰκείωσις’, Ancient Philosophy, 10 (1990), 221–42. Brennan, T., The Stoic Life: Emotions, Duties, and Fate [Stoic Life] (Oxford, 2006), 146–51. Brennan, T., ‘Stoic Moral Psychology’ [‘Psychology’], in B. Inwood (ed.), The Cambridge Companion to the Stoics (Cambridge, 2003), 257–94. Chaumartin, F.-R., Le De beneficiis de Sénèque: Sa signification p ­ hilosophique, politique et sociale (Paris, 1985). Graver, M., ‘Interiority and Freedom in Seneca’s De beneficiis: Acts of Kindness and the Perfected Will’ [‘Interiority’], in M.  R.  Niehoff and  J.  Levinson (eds.) Self, Self-Fashioning and Individuality in Late Antiquity (Tübingen, 2020), 71–88. Graver, M., Stoicism and Emotion [Emotion] (Chicago, 2007). Graver, M. and Long, A.  A. (trans.), Lucius Annaeus Seneca: Letters on Ethics (Chicago, 2015). Griffin, M., Seneca on Society: A Guide to De beneficiis [Society] (Oxford, 2013). Griffin, M. and Inwood, B. (trans.), Seneca, On Benefits (Chicago, 2011). Hine, H., ‘Rome, the Cosmos, and the Emperor in Seneca’s Natural Questions’, Journal of Roman Studies, 96 (2006), 42–72. Hosius, C. (ed.), L. Annaei Senecae: De beneficiis libri VII; De clementia libri II (Leipzig, 1905). Inwood, B., Ethics and Human Action in Early Stoicism [Human Action] (Oxford, 1985). Inwood, B., ‘Politics and Paradox in Seneca’s De Beneficiis’, in id., Reading Seneca: Stoic Philosophy at Rome (Oxford, 2005), 65–94. Inwood, B., ‘Why Do Fools Fall in Love?’, Bulletin of the Institute of Classical Studies, 41 (1997), 55–69. Klein, J., ‘Of Archery and Virtue: Ancient and Modern Conceptions of Value’, Philosophers’ Imprint, 14 (2014), 1–16. Klein, J., ‘The Stoic Argument from oikeiōsis’ [‘Stoic Argument’], Oxford Studies in Ancient Philosophy, 50 (2016), 143–200. Konstan, D., ‘The Joy of Giving: Seneca De beneficiis 1. 6. 1’, in E. K. Emilsson, A. Maravela, and M. Skoie (eds.), Paradeigmata: Studies in Honour of Øivind Andersen (Athens, 2014), 171–6. Lentano, M., ‘De beneficiis’, in G.  Damschen and A.  Heil (eds.), Brill’s Companion to Seneca: Philosopher and Dramatist (Leiden, 2014), 201–6. Lesses, G., ‘Austere Friends: The Stoics and Friendship’ [‘Austere Friends’], Apeiron, 26 (1993), 57–75. Long, A. A., ‘Friendship and Friends in the Stoic Theory of the Good Life’, in D.  Caluori (ed.), Thinking about Friendship: Historical and Contemporary Philosophical Perspectives (London, 2012), 218–39. Long, A. A., Stoic Studies (Cambridge, 1996).

292

Allison Piñeros Glasscock

Long, A. A., and Sedley, D. N., The Hellenistic Philosophers [LS], 2 vols. (Cambridge, 1987). Reydams-Schils, G., ‘From Self-Sufficiency to Human Bonding’, in ead.,The Roman Stoics: Self, Responsibility, and Affection (Chicago, 2005), 53–82. Reydams-Schils, G., ‘Human Bonding and oikeiōsis in Roman Stoicism’ [‘Bonding’], Oxford Studies in Ancient Philosophy, 22 (2002), 221–51. Reynolds, L. D. (ed.), L. Annaei Senecae: Ad Lucilium Epistulae morales [OCT], 2 vols. (Oxford, 1965). Roth, A.  S., ‘Shared Agency’, in E.  N.  Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2017 edn), https://plato.stanford. edu/archives/sum2017/entries/shared-agency/, accessed 11 February 2022. Schofield, M., ‘Two Stoic Approaches to Justice’, in A.  Laks and M. Schofield (eds.), Justice and Generosity: Studies in Hellenistic Social and Political Philosophy, Proceedings of the Sixth Symposium Hellenisticum (Cambridge, 1995), 191–212. Striker, G., ‘The Role of oikeiōsis in Stoic Ethics’, in ead., Essays on Hellenistic Epistemology and Ethics (Cambridge, 1996), 281–97. Takaki, K., ‘Benefits and Communication: Semiotics in Seneca’s De beneficiis’ [‘Communication’], History of Philosophy Quarterly, 31 (2014), 293–316. Vogt, K. M., ‘Seneca’, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2020 edn), https://plato.stanford.edu/archives/spr 2020/entries/seneca/, accessed 4 June 2021. von Arnim, J., Stoicorum Veterum Fragmenta [SVF], 5 vols. (Leipzig, 1903–24). Wildberger, J., ‘Amicitia and Eros: Seneca’s Adaptation of a Stoic Concept of Friendship for Roman Men in Progress’ [‘Amicitia’], in G. M. Müller and F.  M.  Zini (eds.), Philosophie in Rom—­Römische Philosophie?: Kultur-, Literatur-, und Philosophiegeschichtliche Perspektiven (Berlin, 2017), 387–425. Wildberger, J., ‘Care of the Self and Social Bonding in Seneca: Recruiting Readers for a Global Network of Progressor Friends’, Vita Latina, 197 (2018), 117–30.

HIEROCLES’ CONCENTRIC CIRCLES ralph wedgwood

The Stoics prided themselves on the systematic character of their philosophy, which they often expounded in equally systematic treatises. Almost all these treatises have been lost, but we are fortunate to have substantial fragments of some systematic work in ­ethics, probably dating from the mid-­second century ce, by the late Stoic Hierocles. In one of these fragments, Hierocles puts forward a remarkably suggestive image—­the image of the system of concentric circles that surrounds each of us. In this paper, I shall propose an account of what this image is intended to convey, solving a number of exegetical puzzles to which the passage gives rise. In articulating this account, I shall have to offer an in­ter­pret­ ation of a large part of Hierocles’ ethical theory. Admittedly, some parts of this interpretation are speculative. As I shall argue, however, if my speculations are correct, Hierocles made a significant contribution. Specifically, he developed a refined and arguably improved version of his predecessors’ solutions to one of the deepest problems that orthodox Stoics have to face. This is the problem of what, for the Stoics, makes an action one that it is rational to select—­in their terminology, what makes the action ‘appropriate’ (καθῆκον). This problem arises not only for the selections of non-­ ideal human agents like you and me, but also, in an equally challenging form, for the selections of the divine world mind: what is it about the providential unfolding of god’s agency in the world that makes it a course of action that it is rational for god to select? In his As a newcomer to the field of ancient Stoicism, I am more than usually indebted to everyone who has helped me with this paper. Specifically, I wish to thank Simon Shogry, Phil Mitsis, Chris Shields, and Catharine Edwards; audiences at the Ancient Philosophy Society in 2021 and the Central Division of the American Philosophical Association in Chicago in 2022; my commentators in Chicago, Elizabeth Asmis and J. Clerk Shaw; and, especially, the editor of this volume, Victor Caston, and the two anonymous readers from whom he obtained some extremely challenging and helpful comments.

Ralph Wedgwood, Hierocles’ Concentric Circles In: Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Ralph Wedgwood 2023. DOI: 10.1093/oso/9780192885180.003.0008

294

Ralph Wedgwood

work on solving this problem, I shall suggest, Hierocles was a ­worthy contributor to this grand tradition.

1.  Four puzzles about the circles Hierocles’ description of the circles appears in the following passage in Stobaeus’ Anthology:1 ὅλως γὰρ ἕκαστος ἡμῶν οἷον κύκλοις πολλοῖς περιγέγραπται, τοῖς μὲν σμικροτέροις, τοῖς δὲ μείζοσι, καὶ τοῖς μὲν περιέχουσι, τοῖς δὲ περιεχομένοις, κατὰ τὰς διαϕόρους καὶ ἀνίσους πρὸς ἀλλήλους σχέσεις. πρῶτος μὲν γάρ ἐστι κύκλος καὶ προσεχέστατος, ὃν αὐτός τις καθάπερ περὶ κέντρον τὴν ἑαυτοῦ γέγραπται διάνοιαν· ἐν ᾧ κύκλῳ τό τε σῶμα περιέχεται καὶ τὰ τοῦ σώματος ἕνεκα παρειλημμένα. . . . δεύτερος δὲ ἀπὸ τούτου καὶ πλέον μὲν ἀϕεστὼς τοῦ κέντρου, περιέχων δὲ τὸν πρῶτον, ἐν ᾧ τετάχαται γονεῖς ἀδελϕοὶ γυνὴ παῖδες. ὁ δ’ ἀπὸ τούτων τρίτος, ἐν ᾧ θεῖοι καὶ τηθίδες, πάπποι τε καὶ τῆθαι, καὶ ἀδελϕῶν παῖδες, ἔτι δὲ ἀνεψιοί. μεθ’ ὃν ὁ τοὺς ἄλλους περιέχων συγγενεῖς. τούτῳ δ’ ἐϕεξῆς ὁ τῶν δημοτῶν καὶ μετ’ αὐτὸν ὁ τῶν ϕυλετῶν, εἶθ’ ὁ πολιτῶν, καὶ λοιπὸν οὕτως ὁ μὲν ἀστυγειτόνων, ὁ δὲ ὁμοεθνῶν. ὁ δ’ ἐξωτάτω καὶ μέγιστος περιέχων τε πάντας τοὺς κύκλους ὁ τοῦ παντὸς ἀνθρώπων γένους. (4. 27. 23, 4. 671. 7–672. 2 Wachsmuth–­Hense = LS 57G 1–4) For, in general, each of us is as it were encompassed by many circles, some smaller, some larger, with the latter containing the former, according to their different and unequal relations to one another. (1) The first and closest circle is that which a person has drawn as if around a centre, his own mind; in this circle is contained the body and whatever is used for the sake of the body. . . . (2) After this there is a second, standing out farther from the centre but containing the first, in which are stationed one’s parents, siblings, wife, and children. (3) The third, after these, is that in which there are one’s uncles and aunts, grandfathers and grandmothers, one’s siblings’ children, and also cousins. (4) After this there is the one containing one’s other relatives. (5) Next to this is the circle of one’s neighbourhood, (6) after which that of one’s district, (7) then that of one’s fellow citizens, and so, finally, (8) that of people from neighbouring cities, and (9) that of those 1  For the excerpts from Stobaeus’ Anthology, the text cited is that of H. von Arnim (ed.), Hierokles: Ethische Elementarlehre (Papyrus 9780) nebst den bei Stobäus erhaltenen ethischen Exzerpten aus Hierokles (Berlin, 1906), as reprinted in I. Ramelli (ed.), Hierocles the Stoic: Elements of Ethics, Fragments, and Excerpts [Hierocles] (Atlanta, 2009). But I have cited them by the book, chapter, and section numbers, along with the volume, page, and line numbers from the edition of C. Wachsmuth and O. Hense (eds.), Ioannis Stobaei Anthologium, 5 vols. (Berlin, 1884–1912). (For some excerpts, Ramelli uses the numbering from the older Gaisford–­Meineke edition, but provides the Wachsmuth–­Hense numbering in parentheses.)



Hierocles’ Concentric Circles

295

of the same ethnicity. (10) The farthest out and largest one, which contains all the circles, is that of the whole race of human beings.2

In this passage, Hierocles lists ten circles that surround each of us. It appears that each of these circles is effectively a set of human beings, where the innermost circle is the unit set containing the individual in question, and whenever one circle is contained within another, the first circle is a proper subset of the second. In other words, as the circles go outwards, they become more and more inclusive, including all the human beings who are included in the inner circles, but also additional human beings as well. (Hierocles is ignoring a complication here: even in ancient Greek city states, not all members of your extended family are also residents of your neighbourhood. In order to accommodate this, Hierocles would have to switch to a different model—­perhaps a model of overlapping ellipses, where you are situated on the line between the focal points of each ellipse. Understandably, however, Hierocles prefers to keep things simple by ignoring this complication.3) The focus of the rest of the passage is ‘on how one should treat’ each of these different sets of human beings (περὶ τὴν δέουσαν ἑκάστων χρῆσιν, 4. 672. 3). He first presents his central claim metaphorically, by saying that one should ‘in some way draw the circles together as if towards the centre’ (τὸ ἐπισυνάγειν πως τοὺς κύκλους ὡς ἐπὶ τὸ κέντρον, 4. 672. 3–4). This metaphor is illuminated by the following further claims: 1. One should ‘in one’s efforts continually transfer those from the containing circles into the circles that are contained within them’ (τῇ σπουδῇ μεταϕέρειν ἀεὶ τοὺς ἐκ τῶν περιεχόντων εἰς τοὺς περιεχομένους, 4. 672. 4–6). 2. One ‘should honour those from the third circle similarly to those [in the second circle], and one’s [more remote] relatives similarly to those [in the third circle]’ (τούτοις μὲν ὁμοίως τιμητέον τοὺς ἐκ τοῦ τρίτου κύκλου, τούτοις δ’ αὖ πάλιν τοὺς συγγενεῖς, 4. 672. 14–16). 2  All translations from Greek are mine; translations from Latin will be those of other scholars, as indicated in the notes below. 3  It is possible that Hierocles is influenced by an earlier use of this image; specifically, see Seneca’s use of the image of concentric circles (Ep. 12. 6), although Seneca’s circles are extended in time (today, this month, this year, . . . my whole lifetime), rather than in a ‘space’ of social closeness and distance. Unfortunately, I shall not be able to explore the connections between Hierocles’ and Seneca’s use of this image here.

296

Ralph Wedgwood

3. ‘The greater distance in blood [of those in each containing circle, compared with those in the circle that is immediately contained within it] will take away something of our goodwill, but we should make an effort to make it similar’ (ἀϕαιρήσεται μὲν γάρ τι τῆς εὐνοίας τὸ καθ’ αἷμα διάστημα πλέον · ἡμῖν δ’ ὅμως σπουδαστέα περὶ τὴν ἐξομοίωσίν ἐστιν, 4. 672. 16–18). 4. ‘The appropriate degree would be reached if through our own initiative we cut down the distance of our relation towards each person’ (ἥκοι μὲν γὰρ ἂν εἰς τὸ μέτριον, εἰ διὰ τῆς ἡμετέρας αὐτῶν ἐνστάσεως ἐπιτεμνόμεθα τὸ μῆκος τῆς πρὸς ἕκαστον τὸ πρόσωπον σχέσεως, 4. 672. 18–673. 1). Finally (5), he also makes the following additional claim: χρὴ δ’ ἐπιμετρεῖν καὶ κατὰ τὴν τῶν προσηγοριῶν χρῆσιν, τοῦς μὲν ἀνεψιοὺς καὶ θείους καὶ τηθίδας ἀδελϕοὺς ἀποκαλοῦντας πατέρας τε καὶ μητέρας, τῶν δὲ συγγενῶν τοὺς μὲν θείους, τοὺς δὲ ἀδελϕιδοῦς, τοὺς δὲ ἁνεψιούς, ὡς ἂν καὶ τὰ τῆς ἡλικίας παρήκῃ ἕνεκα τῆς ἐν τοῖς ὀνόμασιν ἐκτενείας. οὗτος γὰρ τῆς προσρήσεως ὁ τρόπος ἅμα μὲν ἂν σημεῖον οὐκ ἀμαυρὸν εἴη τῆς οὔσης ἡμῖν σπουδῆς περὶ ἑκάστους, ἅμα δ’ ἂν ἐποτρύνοι καὶ προσεντείνοι πρὸς τὴν ὑποδεδειγμένην οἷον συνολκὴν τῶν κύκλων. (4. 673. 2–11) But one must also do more in one’s usage of modes of address, calling one’s cousins and uncles and aunts ‘brothers’ and ‘fathers’ and ‘mothers’, and among one’s other relatives calling some ‘uncles’, others ‘nephews’, and others ‘cousins’, as the occasion arises in terms of their age, for the sake of the motivating force in these names. For this manner of address would be a by no means faint sign of our efforts concerning each group, and at the same time it would heighten and intensify our motivation towards the contraction, as it were, of the circles that has been outlined.

A number of puzzles can be raised about these claims. First, there is a puzzle about the first claim (1). How can one ‘transfer’ those from the circle of one’s uncles, aunts, grandparents, and cousins into the circle of one’s parents, spouse, children, and siblings? One cannot literally change one’s grandparents into one’s parents. (Admittedly, it is possible in many countries to make one’s cousin into one’s spouse—­but in a monogamous society like Greece in the second century ce, one can only do this to one cousin at any one time.) What can Hierocles have meant by this claim? Secondly, there is a puzzle about the second (2) and third (3) claims as well. Hierocles says that one should make one’s honour and goodwill towards those in each containing circle similar to one’s



Hierocles’ Concentric Circles

297

honour and goodwill towards those in the circle that is immediately contained within it. How should we understand this? Some readers have interpreted him as meaning that one should ideally have exactly the same degree of honour and goodwill towards everyone.4 On this interpretation, the ideal is to achieve total impartiality between everyone—­complete impartiality between oneself and the ‘remotest Mysian’, as Plato puts it (Theaet. 209 b 8). This could be supported by a reading of the image of ‘transferring people’ from the containing circles into the circles that are contained within them—­if we read this image as describing a process whose end result is to transfer absolutely everyone into the innermost circle, so that one has the same degree of honour and goodwill towards every other human being as to oneself. Strictly, however, in his fifth claim (5), Hierocles seems to contra­ dict this. He does not recommend that one should call all human beings ‘brother’, ‘sister’, ‘mother’, or ‘father’: he recommends that one should call one’s uncles ‘father’ and one’s aunts ‘mother’—but one should call one’s parents’ cousins ‘uncles’ or ‘aunts’, and so on.5 So, according to his recommendations, one should still treat the third and fourth circles in different ways. This suggests a different reading of the image of transferring people from each of the containing circles into the circle that is contained within it: we should transfer those in Circle 3 into Circle 2, those in Circle 4 into Circle 3, and so on—­so that we also transfer everyone in Circle 10 into Circle 9, emptying Circle 10, but still leaving different people in different circles.6 However, if Hierocles does not advocate complete impartiality between everyone, we still need to know in what way exactly one should make one’s honour and goodwill towards those in Circle 3 similar to those in Circle 2, and so on. Is there anything more precise and illuminating that can be said about this? 4  This interpretation is defended by J. Annas, The Morality of Happiness (Oxford, 1993), 267–70. In effect, this interpretation takes Hierocles to be making broadly the same point that Peter Singer famously made with the related image of the ‘expanding circle’; see P. Singer, The Expanding Circle: Ethics, Evolution, and Moral Progress, 2nd edn (Princeton, 2011). 5  In this way, Hierocles’ recommendation is importantly different from Plato’s proposal (Rep. 5, 461 d 1–e 1) that citizens of Kallipolis should call all fellow citizens who are old enough to be their parents ‘father’ or ‘mother’, and so on. 6  This reading of ‘transferring’ people from the containing circles into the circles that are contained within them was suggested to me by an anonymous reader for this volume.

298

Ralph Wedgwood

Thirdly, what exactly is the import of the metaphor of ‘cutting down’ the distance between the circles, ‘contracting’ the circles, or ‘drawing the circles together’? If this is not a matter of obliterating all distinctions between the circles (as it were, reducing the distance between the circles to zero), what exactly is this image of ‘drawing the circles together’ meant to convey?7 Finally, there is a general puzzle about the whole passage. What, in Hierocles’ view, explains why these are the ways in which we ‘should treat’ the people in these different circles? How are these claims justified on the basis of the more fundamental principles of Stoic ethics? Unless solutions to these puzzles can be found, then this suggestive discussion is in the end disappointingly uninformative.

2.  Is Hierocles advocating complete impartiality? Some of our available evidence suggests that some of the early Stoics advocated complete impartiality between oneself and every other human being. For example, Plutarch gives the following description of the Republic of Zeno of Citium:8 καὶ μὴν ἡ πολὺ θαυμαζομένη πολιτεία τοῦ . . . Ζήνωνος εἰς ἓν τοῦτο συντείνει κεϕάλαιον, ἵνα μὴ κατὰ πόλεις μηδὲ κατὰ δήμους οἰκῶμεν, ἰδίοις ἕκαστοι διωρισμένοι δικαίοις, ἀλλὰ πάντας ἀνθρώπους ἡγώμεθα δημότας καὶ πολίτας, εἷς δὲ βίος ᾖ καὶ κόσμος, ὥσπερ ἀγέλης συννόμου νόμῳ κοινῷ συντρεϕομένης. (De Alex. fort. 1. 6, 329 a–­b = SVF i. 262 = LS 67A 1) Indeed, the much-­admired Republic of Zeno . . . is aimed at this one main point: that we ought not to make our home in cities or neighbourhoods, each marked out by separate systems of justice, but we should consider all human beings our neighbours and fellow citizens, and there should be one way of life and order, like that of a herd raised together ruled by a common law. 7  Scholars have made various suggestions here. For example, Inwood suggests that Hierocles’ ideal is that ‘one’s feeling of concern and love for all mankind . . . be rendered equal in intensity with one’s concern and love for oneself’ (B.  Inwood, ‘Hierocles: Theory and Argument in the Second Century ad’ [‘Hierocles’], Oxford Studies in Ancient Philosophy, 2 (1984), 151–83, at 181); this seems to imply—­ implausibly—­that, given a choice between saving oneself and saving all mankind, one should view both options as equally appropriate. Other scholars’ suggestions are strikingly vague; for example, according to Gloyn, ‘ultimately Stoic sages . . . consider the interests of all humanity as integral to their own’ (L. Gloyn, The Ethics of the Family in Seneca (Cambridge, 2017), 28). 8  The Greek text cited is that of H. von Arnim (ed.), Stoicorum Veterum Fragmenta [SVF], 4 vols. (Leipzig, 1905–24), and A. A. Long and D. N. Sedley (eds.), The Hellenistic Philosophers [LS] (Cambridge, 1989).



Hierocles’ Concentric Circles

299

If Plutarch’s testimony can be trusted here, then in his Republic Zeno rejected the idea that we should draw any distinction between people based on their citizenship or on the neighbourhood that they live in: we should simply regard all human beings as equal fellow citizens of the ‘cosmic city’.9 An anonymous commentator on Plato’s Theaetetus seems to criticize the Stoic conception of justice as requiring too much impartiality: ὅ̣[σοι το]ί ν̣ ̣υ̣ν̣ | ἀπὸ τῆς οἰκε[ι]ώσεως | εἰσάγουσι τὴν δι[κ]αι[ο]|σύνην, εἰ μὲν λέγου|σιν ἴσην αὐτοῦ τε πρὸς | α̣ὑτὸν καὶ πρ ̣ὸς̣ ̣ [τὸν ἔ]|[σ]χατον Μυσῶν, τεθέν|τος μὲν τούτου σώζε|ται ἡ δικαιοσ[ύ]νη, οὐ | συγχωρεῖται [δ]ὲ [εἶ]|ν̣αι ἴσην . . . εἰ δὲ καὶ α̣[ὐτ]οὶ ϕήσου|σ̣ι ̣ ἐπιτεί[ν]ε̣σθα[ι] τὴν | οἰκε̣ίω ̣ [σιν, ἔσ]τ̣αι μὲν | ϕιλανθρ[ωπί]α, ἐλέγξου|σι δὲ τ[ούτους α]ἱ περιστάσει ̣ς̣ [ναυαγῶ]ν, ὅ|που ἀνά̣γ  [̣ κη μό]νο̣ν | σώζεσθαι τὸ̣ν̣ ἕτ̣ε|ρον αὐτῶν. (Anon., In Plat. Theaet. 5. 24–6. 31 = LS 57H 3 in part, 7) Thus, if those who derive justice from oikeiōsis say that it is equal both towards oneself and towards the remotest Mysian, on this assumption just­ice is saved, but we do not agree that it is equal . . . If they say that the oikeiōsis is intensified, there will be a general love of human beings, but they will be refuted by the circumstances of shipwrecked men, where it is necessary that only one of the two be saved.

Those who ‘derive justice from oikeiōsis’ are clearly the Stoics. Oikeiōsis, as I shall explain in Section 3, is a kind of perception of something as oikeion, that is, as ‘one’s own’ or ‘belonging to one’. So, in this passage the anonymous commentator implies that some Stoics may hold that a perfectly wise agent will perceive himself as no more ‘his own’ or ‘belonging to him’ than the ‘remotest Mysian’, and that this perception will lead to a kind of impartiality in the ‘circumstances of the shipwrecked men’. This case of the ‘shipwrecked men’ seems to have been discussed in the work On Appropriate Acts of Hecato of Rhodes, a Stoic of the early first century bce, which Cicero translates as follows:10 ‘Quid? si una tabula sit, duo naufragi, iique sapientes, sibine uterque rapiat an alter cedat alteri?’ ‘Cedat vero, sed ei cuius magis intersit vel sua vel reipublicae causa vivere’.

9  This interpretation of Stoic ethics, including Hierocles’ circles, is defended in K. M. Vogt, Law, Reason, and the Cosmic City: Political Philosophy in the Early Stoa (Oxford, 2008), 103. 10  The Latin text is that of M. Winterbottom (ed.), M. Tulli Ciceronis De Officiis (Oxford, 1994); the translation is that of M. T. Griffin and E. M. Atkins, Cicero: On Duties (Cambridge, 1991), slightly modified.

300

Ralph Wedgwood

‘Quid si haec paria in utroque?’ ‘Nullum erit certamen, sed quasi sorte aut micando victus alteri cedet alter’. (Cic., Off. 3. 90) ‘Well, suppose that there is one plank and two shipwrecked men, both of them wise. Would each of them grab it for himself, or would one give in to the other?’ ‘One should give in to the other, that is, to the one whose life matters more for his own or the republic’s sake’. ‘And what if such considerations are equal for both?’ ‘There will be no contest, but one will give in to the other as if losing by lot or by a game of chance’.

This evidence might suggest that some early Stoics may have advocated complete impartiality in our treatment of other human beings—­rather like William Godwin, who notoriously argued that, given the choice between saving Archbishop Fénelon and one of one’s parents from a fatal fire, one should act impartially for the greater good and save the archbishop.11 Even if some of the early Stoics advocated this sort of impartiality, however, there are—­as I shall now argue—­compelling reasons to deny that Hierocles is doing so here. Stobaeus labels the circles passage as coming from ‘Hierocles’ How one should treat one’s relatives’ (Ἱεροκλέους ἐκ τοῦ πῶς συγγενέσι χρηστέον, 4. 671. 3 Wachsmuth–­ Hense); and it begins as an explicit sequel to ‘what has been said earlier about how one should treat one’s parents, siblings, wife, and children’ (Τοῖς εἰρημένοις περὶ γονέων χρήσεως καὶ ἀδελϕῶν γυναικός τε καὶ τέκνων ἀκόλουθον, 4. 671. 4–5). In this way, it seems that the circles passage comes from the same treatise as the other excerpts from Hierocles in Stobaeus’ Anthology, about how one should treat one’s parents (4. 25. 53), siblings (4. 27. 20), wife (4. 22a. 21–4) and children (4. 24a. 14). It seems clear that these other excerpts do not support an impartialist interpretation. This is particularly clear in the discussion of how one should treat one’s parents. Here, Hierocles argues for some extremely strong claims: for us, our parents are ‘in a way second and terrestrial gods’ (δευτέρους καὶ ἐπιγείους τινὰς θεούς, 4. 25. 53, 4. 640. 8–9); we ‘must consider ourselves as in a way ministers

11 W. Godwin, An Enquiry Concerning Political Justice (London, 1793), i. 82.



Hierocles’ Concentric Circles

301

and priests in our home as in a temple’ (νομιστέον ἑαυτοὺς καθάπερ ἐν ἱερῷ τῇ οἰκίᾳ ζακόρους τινὰς καὶ ἱερέας, 4. 642. 2–3). The rest of this excerpt describes in detail the intimate and attentive services that we must provide to our parents, like priests serving gods in a temple (4. 642. 3–644. 9). It is clear that Hierocles’ conception of the filial duty that we owe to our parents is entirely different from Godwin’s view that one should impartially sacrifice one’s father or one’s mother in order to save Archbishop Fénelon. In general, as I shall explain in Section 5 below, the excerpts from Hierocles in Stobaeus’ Anthology all provide accounts of the duties that we have towards various different beings in virtue of the role (πρόσωπον, 4. 640. 7) that they play in relation to us, that is, in virtue of their relationship with us (σχέσις, 4. 27. 23, 4. 671. 11).12 Thus, for example, he discusses our duties to the gods (1. 3. 53–4 and 2. 9. 7), our country (3. 39. 34–6), our parents (4. 25. 53), and our siblings (4. 27. 20).13 In general, different relationships and different roles give rise to different duties. For example, our parents are hierarchically positioned above us, as ‘secondary and terrestrial gods’ (4. 25. 53, 4. 640. 8–9), while our siblings stand to us in a relation of strict equality (4. 27. 20, 4. 661. 7–12). Since one’s relationship to oneself is radically different from one’s relationship to the ‘remotest Mysian’—one plays a quite different role from the remotest Mysian in relation to oneself—­it seems that, for Hierocles, one will have radically different duties towards oneself and towards the remotest

12  The idea of one’s ‘roles’ in life is central to the ethics of Epictetus, as has been demonstrated by B. E. Johnson, The Role Ethics of Epictetus: Stoicism in Ordinary Life [Role Ethics] (Lanham, 2014), especially chs. 1–3; as Johnson persuasively argues (ch. 8), Epictetus’ use of the notion of one’s ‘roles’ is importantly different from the use that Cicero makes of the idea of one’s ‘personae’ (Off. 1. 107–15). Unlike Epictetus, however, Hierocles uses the word (πρόσωπον) to refer not to the roles that are played by the agent under discussion, but to the roles that other beings play in relation to the agent (such as the roles of being the agent’s mother or fellow citizen or the like); see also the use of the word at Stob. 4. 27. 20, 4. 660. 18 and 4. 661. 9–10. 13  It is striking—­especially to anyone who approaches Hierocles’ work after reading Aristotle’s Nicomachean Ethics—­that friendship is not one of the relationships that he considers. I speculate that the reason for this is that the older Stoics defended such an idealized and austere conception of friendship—­a conception in which only sages can be each other’s friends—­that Hierocles regards true friendship as a vanishingly rare phenomenon in human life. For the older Stoic conception of friendship, see G. Lesses, ‘Austere Friends: The Stoics and Friendship’, Apeiron, 26 (1993): 57–75.

302

Ralph Wedgwood

Mysian. For this reason, it is unlikely that the circles passage is designed to advocate complete impartiality in the style of Godwin. This conclusion is also supported by another passage in earlier ancient ethical writing that has a striking similarity to Hierocles’ circles. This passage is Cicero’s list of the ‘degrees of fellowship among human beings’: Gradus autem plures sunt societatis hominum. Ut enim ab illa infinita discedatur, propior est eiusdem gentis nationis linguae, qua maxime homines coniunguntur. Interius etiam est eiusdem esse civitatis; multa enim sunt civibus inter se communia . . . Artior vero conligatio est societatis propinquorum; ab illa enim immensa societate humani generis in exiguum angustumque concluditur. Nam cum sit hoc natura commune animantium, ut habeant libidinem procreandi, prima societas in ipso coniugio est, proxima in liberis, deinde una domus, communia omnia . . . Sequuntur fratrum coniunctiones, post consobrinorum sobrinorumque . . . Sequuntur conubia et adfinitates ex quibus etiam plures propinqui. (Off. 1. 53–4) There are indeed several degrees of fellowship among men. To move from (1) the one that is unlimited [sc. the fellowship of the entire human race, 1. 50], next (2) there is a closer one of the same race, tribe, and tongue, through which men are bound strongly to one another. More intimate still is (3) that of the same city, as citizens have many things that are shared with one another . . . A tie narrower still is that of the fellowship between relations: moving from that vast fellowship of the human race we end up with a confined and limited one. For since it is by nature common to all animals that they have a drive to procreate, (4) the first fellowship exists within marriage itself,14 and (5) the next with one’s children. Then (6) there is the one house in which everything is shared . . . Next, (7) there follow bonds between brothers, and then (8) between cousins and second cousins . . .  Finally, (9) there follow marriages and those connections of marriage through which even more relations arise.

Here Cicero lists nine ‘degrees of fellowship’, which differ only slightly from the nine circles that according to Hierocles surround the innermost circle (which contains only one’s own mind and body). In view of the similarity between these two lists, it is possible that they are both influenced by a common source; this could have been one of the two Stoic sources which Cicero mentions in

14  This statement is virtually identical to one that Hierocles makes: ‘The first and most fundamental of communities is that in accordance with marriage’ (πρώτη δὲ καὶ στοιχειωδεστάτη τῶν κοινωνιῶν ἡ κατὰ τὸν γάμον, 4. 22a. 21, 4. 502. 3–4).



Hierocles’ Concentric Circles

303

De Officiis, Panaetius (1. 7) and Hecato (3. 63); or conceivably Hierocles could have been influenced directly by Cicero. As Cicero puts it, these ‘degrees of fellowship’ are relevant in the case of a ‘competition . . . as to who ought most to receive our dutiful services’ (contentio . . . quibus plurimum tribuendum sit officii, Off. 1. 58). In such a case, those who have a closer degree of fellowship with us have a stronger claim on our services. It seems plaus­ ible to read Hierocles as using his list of circles to make a broadly similar point to the one that Cicero makes with his list of ‘degrees of fellowship’. If this reading is along the right lines, then Hierocles’ point implies that the people in the inner circles have a stronger claim on our services than those in the outer circles. For these reasons, then, it seems that Hierocles is not advocating Godwin-­like impartiality in one’s treatment of all human beings but is rather arguing that different forms of treatment are appropriate towards the people in different circles. But this only intensifies the puzzles that I identified in Section 1. What can Hierocles mean by transferring people from Circle 3 into Circle 2 (and so on)? What sort of similarity is he advocating in our honour and goodwill towards the people in Circles 2 and 3? What is meant by ‘pulling the circles together towards the centre’? And what within the Stoic ethical system explains why these are the appropriate ways to treat these people?

3.  Do the circles illustrate the theory of oikeiōsis? Our fragments of Hierocles’ work derive from two sources. First, as mentioned above, there are the excerpts in Stobaeus’ Anthology. Second, there are some papyrus fragments of a work entitled The Elements of Ethics (Ἠθικὴ Στοιχείωσις). The focus of these papyrus fragments is on the Stoic theory of oikeiōsis. Some scholars have assumed that the circles passage is also concerned with this the­ory.15 As I shall argue in this section, this is a misunderstanding: the circles 15  For scholars who associate the ‘circles’ passage with the theory of oikeiōsis, see, for example: Inwood, ‘Hierocles’, 180, and M. M. McCabe, ‘Extend or Identify: Two Stoic Accounts of Altruism’, in R. Salles (ed.), Metaphysics, Soul, and Ethics in Ancient Thought: Themes from the Work of Richard Sorabji (Oxford, 2005), 413–43. In a similar way, Nussbaum associates the circles with ‘moral education’ (M.  C.  Nussbaum, The Cosmopolitan Tradition: A Noble but Flawed Ideal

304

Ralph Wedgwood

passage is not concerned with the theory of oikeiōsis (although it clearly seems designed to be consistent with that the­ory). Oikeiōsis, as mentioned above, is a kind of perception of something as ‘one’s own’ or ‘belonging to one’.16 Such perceptions occur in both non-­rational animals and in rational beings, and in both cases they involve a distinctive kind of impression (ϕαντασία). Presumably, when rational beings are in favourable conditions, their impressions of this kind count as reliable or ‘kataleptic’ impressions: in such cases, if one assents to such a kataleptic impression, one’s assent will count as a ‘grasp’ of the truth (κατάληψις); and ultimately, if one becomes a sage, this grasp would count as full-­blown know­ ledge (ἐπιστήμη). Whenever one grasps that something is in this way ‘one’s own’ or ‘belongs to one’, this puts one in a position to grasp that it is something that it is normally appropriate (καθῆκον) for one to preserve and protect. In this way, the theory of oikeiōsis is the central component of the Stoic theory of how we can come to grasp what it is appropriate for us to do. The theory also explains how we are motivated to act appropriately; this is because, according to the Stoics, assenting to the impression that an available act is appropriate necessarily counts as an impulse (ὁρμή) to perform the act. Finally, the theory of oikeiōsis also involves an account of a developmental process, whereby one starts from the most primitive impressions of how one’s own body is ‘one’s own’ and ‘belongs to one’ and gradually develops an increasingly sophisticated apprehension of what is ‘one’s own’ and ‘belongs to one’, thereby achieving an increasingly adequate grasp of what it is appropriate for one to do. Hierocles’ account of oikeiōsis in these papyrus fragments is clearly in line with this orthodox Stoic doctrine.17 First, he argues at length that all animals perceive themselves, their body parts, and (Cambridge, Mass., 2019), 78). In their hugely influential collection, Long and Sedley place the circles passage in the section on oikeiōsis (LS 57G). 16  For a detailed interpretation of the theory of oikeiōsis that is broadly along the lines that I sketch in this paragraph, see especially J. Klein, ‘The Stoic Argument from oikeiōsis’, Oxford Studies in Ancient Philosophy, 50 (2016), 143–200. 17  For illuminating analyses of these parts of Hierocles’ work, see J.-B. Gourinat, ‘La gestation de l’animal et la perception de soi’, in J.-B. Gourinat (ed.), L’ethique du stoïcien Hiéroclès [L’éthique] (Villeneuve d’Ascq, 2016), 15–46; C. Gill, ‘La continuité de la perception depuis la naissance’, in Gourinat (ed.), L’éthique, 47–62; and F. Alesse, ‘La représentation de soi et les différentes formes de l’appropriation chez Hiéroclès’, in Gourinat (ed.), L’éthique, 63–85.



Hierocles’ Concentric Circles

305

the functions of their body parts, continuously throughout their lives. On this basis, he draws the following conclusion:18 ὅθεν ὁ συλλογισμὸς οὗτος ἀναγκάζει ὁμολογεῖν ὅτι τὸ ζῷον, τὴν πρώτην αἴσθησιν ἑαυτοῦ λαβόν, εὐθὺς ᾠκειώθη πρὸς ἑαυτὸ καὶ τὴν ἑαυτοῦ σύστασιν (6. 49–53 Bastianini-­Long). Consequently, this reasoning compels us to agree that an animal, as soon as it receives its first perception of itself, immediately perceives itself and its constitution as its own and belonging to it.

As he claims, it is through this kind of perception that nature gives all animals ‘an intense love of themselves, since otherwise their preservation would be impossible’ (σϕῶν αὐτῶν . . . σϕοδρὸν ἵμερον, τῷ τὴν σωτηρίαν ἄλλως ἄπορον ὑπάρχειν, 7. 4–5 Bastianini-­Long). Unfortunately, after this point, the papyrus becomes increasingly fragmentary and hard to decipher. But it seems that there is a discussion of how one’s impression of what is ‘one’s own’ and ‘belongs to one’ later becomes more ‘clear and precise’ (τρανὴς ἡ ϕαντασία γίνεται καὶ διηκριβωμένη, 8. 54 Bastianini-­Long), and of how we come to perceive our family and our children as ‘our own’ (9. 3–7 Bastianini-­Long). In this way, it seems, we can over time develop a more adequate grasp of what it is appropriate for us to do. In general, the focus of The Elements of Ethics is on moral psych­ology, moral epistemology, and moral development. The excerpts reproduced in Stobaeus’ Anthology have a different focus—­specifically, on what kinds of acts are appropriate in various general kinds of circumstances. The context of these excerpts in the Anthology is revealing. None of these excerpts appears in chapters that focus on moral psychology, moral epis­tem­ol­ogy, or moral development. The excerpts on our duties to the gods appear in chapters on broadly theological topics (Stob. 1. 3 and 2. 9), and the excerpts on our duties to our country appear in the chapter ‘On one’s country’ (3. 39, Περὶ πατρίδος). All the other excerpts appear in chapters of Book 4 of the Anthology that are concerned with the family:

18  The Greek text is that of G. Bastianini and A. A. Long, ‘Ierocle: Elementi di Etica’, Corpus dei papiri filosofici greci e latini (CPF), 1. 1. 2 (Florence, 1992), 296– 362, cited by the numbers of the columns and lines in the papyrus fragments. This text, incorporating the corrigenda in ‘Indici 1. 1’, CPF 4. 1 (Florence, 2002), viii–­ix, is reproduced in Ramelli, Hierocles, 2–32.

306

Ralph Wedgwood



i. The excerpts on marriage appear in the chapter ‘That marriage is a most fine thing’ (4. 22a, Ὅτι κάλλιστον ὁ γάμος). ii. The excerpt on children appears in the chapter ‘That it is fine to have children’ (4. 24, Ὅτι καλὸν τὸ ἔχειν παῖδας). iii. The excerpt on our duties to our parents appears in the chapter ‘That parents must be held worthy of the appropriate honour from their children, and whether they are to be obeyed in all things’ (4. 25, Ὅτι χρὴ τοὺς γονεῖς τῆς καθηκούσης τιμῆς καταξιοῦσθαι παρὰ τῶν τέκνων, καὶ εἰ ἐν ἅπασιν αὐτοῖς πειστέον). iv. Both the excerpt on one’s duty to one’s siblings and the circles passage appear in the chapter ‘That brotherly love and the [right] disposition towards one’s relatives are a most fine thing, and that they are necessary’ (4. 27, Ὅτι κάλλιστον ἡ ϕιλαδελϕία καὶ ἡ περὶ τοὺς συγγενεῖς διάθεσις, καὶ ὅτι ἀναγκαῖοι). Thus, oikeiōsis is not the main focus of the excerpts in Stobaeus’ Anthology. Admittedly, there is a brief discussion of oikeiōsis in one of the excerpts on marriage (4. 22a. 22, 4. 502. 20–503. 10); but otherwise, these excerpts contain no other discussion of oikeiōsis. The focus is not on how one gets to have impressions of what is ‘one’s own and belongs to one’ or on how we can develop towards genuine knowledge of what it is appropriate for us to do. The focus is on the first-­order question of how it actually is appropriate for us to act. The higher-­order questions of moral psychology and moral epistemology, about how in principle one could know the answer to this first-­order question, are set to one side. In other words, Hierocles’ goal in these excerpts is to find out how it is appropriate to act (εὕρεσις . . . καθηκόντων, 4. 25. 53, 4. 644. 10–11).19 For this reason, some scholars have concluded that these excerpts come, not from The Elements of Ethics, but from a separate work, possibly entitled On Appropriate Acts (Περὶ τῶν καθηκόντων).20 19  Veillard has an interesting discussion of the Stoic idea of ‘la découverte des devoirs’; see C.  Veillard, ‘Hiéroclès, les devoirs envers la patrie et les parents’ [‘Hiéroclès’], in Gourinat (ed.), L’éthique, 105–43 at 105 n. 1. As Veillard argues, the phrase ‘εὕρεσις . . . καθηκόντων’ seems to derive from Panaetius and is imitated in Cicero’s phrase ‘inveniendi officii’ (Off. 1. 107). The phrase also appears in an excerpt that is believed to be from Arius Didymus’ Epitome of Stoic Ethics; see Stob. 2. 7. 5b. 3, 2. 62. 10–11 Wachsmuth–­Hense = SVF iii. 264, 65. 1–2. 20  For a discussion of whether the Stobaean excerpts all belong to a single work, On Appropriate Acts, see Ramelli, Hierocles, xxvi–­xxx, and Veillard, ‘Hiéroclès’, 108–11.



Hierocles’ Concentric Circles

307

Admittedly, the circles passage has seemed to some readers to describe a developmental process of transferring people from the outer circles into the inner circles and of making one’s goodwill towards those in the outer circles similar to one’s goodwill towards those in the inner circles. In Section 8 below, I shall outline an interpretation on which this language does not describe a developmental process. Instead, on my interpretation, this passage, like all the other excerpts in Stobaeus’ Anthology, is just concerned with ‘how we should treat each group’ (περὶ τὴν δέουσαν ἑκάστων χρῆσιν, 4. 27. 23, 4. 672. 3).

4.  Finding out what is appropriate To understand this project of finding out what is appropriate, it will be useful to step back from Hierocles and consider some of the fundamental principles of Stoic ethics. Here is Diogenes Laertius’ account of the Stoics’ conception of what is ‘appropriate’:21 ἔτι δὲ καθῆκόν ϕασιν εἶναι ὃ προαχθὲν εὔλογόν {τε} ἴσχει ἀπολογισμόν, οἷον τὸ ἀκόλουθον ἐν ζωῇ, ὅπερ καὶ ἐπὶ τὰ ϕυτὰ καὶ ζῷα διατείνει· ὁρᾶσθαι γὰρ κἀπὶ τούτων καθήκοντα . . . ἐνέργημα δὲ αὐτὸ εἶναι ταῖς κατὰ ϕύσιν κατασκευαῖς οἰκεῖον. (D. L. 7. 107–8 = LS 59C 1, 3 = SVF iii. 493) Furthermore, they say that something is appropriate if, when it is preferred, it has a reasonable justification, such as what follows in life, and this extends to both plants and animals; for appropriate acts are observed in them too . . . It is an activity that is proper to the constitutions that are according to nature.

In this way, the notion of an ‘appropriate act’ has a crucial connection to nature. It is also undeniable that the Stoics interpret nature in an essentially teleological fashion: in some sense, the natures of things involve purposes. So, for the Stoics, to act appropriately is in a way to fulfil certain purposes that are involved in the natures of things. 21  The text is that of T. Dorandi (ed.), Diogenes Laertius: Lives of the Eminent Philosophers [D. L.] (Cambridge, 2013). Dorandi prefers the manuscripts’ προαχθὲν to the widely accepted emendation πραχθὲν; although this certainly seems to change the sense, the difference does not matter for our purposes. For an illuminating study of the Stoic conception of appropriate acts, see J.-B. Gourinat, ‘Comment se détermine le kathekon?’, Philosophie Antique, 14 (2014), 13–40.

308

Ralph Wedgwood

However, it cannot be true that every appropriate act that a human being performs fulfils all of the purposes of that human being’s nature. For the Stoics, the purposes of our rational nature are fulfilled only by living in accordance with perfect virtue (D. L. 7. 87 = LS 63C 1).22 But not every appropriate act is done in accordance with virtue. As Cicero (Fin. 3. 59) points out, returning a deposit is an appropriate act, even if it is done by a foolish person, who lacks the kind of sage-­like knowledge that would make them virtuous. It seems, then, that appropriate acts fulfil some other aspect of our nature. Part of the difference is that this other aspect of our nature is fulfilled by external acts, which extend outside the soul’s ‘leading part’ (ἡγεμονικόν), and not just by rational thought itself. Since appropriate acts are present in non-­human animals as well as in rational beings, I shall label this aspect of our nature our ‘animal nature’.23 However, according to the Stoics, the goal of life consists not just in living in accordance with one’s own individual nature as a human being, but also in living in accordance with the nature of the whole cosmos (D. L. 7. 89 = LS 63C 5 = SVF iii. 4). This is puzzling: although interfering external factors can prevent an individual plant or animal from fulfilling its individual nature (for example, a tree might be prevented from growing when a roof is built over it), for the Stoics no interfering external factors could prevent the whole cosmos from fulfilling its nature.24 Indeed, this conception of how the cosmos always perfectly fulfils its natural purposes is one of the Stoics’ reasons for viewing the cosmos as a perfect and divine being (for example, see Sen., QNat. 1, Praef. 3 and 14). Thus, everything that ever happens—­including every inappropriate act that any human being performs—­is part of the fulfilment of the natural purposes of the cosmos as a whole. In what sense, then, is living virtuously in accordance with the nature of the cosmos while living viciously is not? 22  For a pioneering discussion of the Stoic doctrine that virtue is equivalent to following nature, see G. Striker, ‘Following Nature: A Study of Stoic Ethics’, in ead., Essays on Hellenistic Epistemology and Ethics [Essays] (Cambridge, 1996), 221–80. 23 I owe this label to Striker, ‘The Role of oikeiōsis in Stoic Ethics’, in ead., Essays, 281–97, at 295. 24  For this point, see S. Bobzien, Determinism and Freedom in Stoic Philosophy (Oxford, 1998), 30.



Hierocles’ Concentric Circles

309

This sense of ‘living in accordance with the nature of the cosmos’ seems to be given by Chrysippus’ formulation: πάλιν δ’ ἴσον ἐστὶ τὸ κατ’ ἀρετὴν ζῆν τῷ κατ’ ἐμπειρίαν τῶν ϕύσει συμβαινόντων ζῆν, ὥς ϕησι Χρύσιππος ἐν τῷ πρώτῳ Περὶ τελῶν. (D. L. 7. 87 = LS 63C 2 in part) Again, living in accordance with virtue is equivalent to living in accordance with the experience of events that occur by nature, as Chrysippus says in the first book of On Goals.

Human experience of nature reveals some of the natural purposes of the cosmos, but it leaves many of these cosmic purposes in­scrut­ able. Being virtuous fulfils those natural purposes of the cosmos that are revealed by human experience. As Chrysippus also said, someone who has progressed to the point of complete virtue always acts appropriately and never inappropriately (Stob. 4. 39. 22, 5. 906. 18–907. 1 Wachsmuth–­Hense = LS 59I, in part = SVF iii. 510). So, acting appropriately must also be part of what fulfils these natural purposes of the cosmos that are revealed by human experience. In general, then, to discover what is appropriate, we must investigate nature. As Chrysippus is reported to have said:25 Πόθεν οὖν . . . ἄρξωμαι; καὶ τίνα λάβω τοῦ καθήκοντος ἀρχὴν καὶ ὕλην τῆς ἀρετῆς, ἀϕεὶς τὴν ϕύσιν καὶ τὸ κατὰ ϕύσιν; (Plut. Comm. not. 1069 e = LS 59A = SVF iii. 491) So, where am I to start from? What am I to take as the principle of what is appropriate and the raw material of virtue if I let go of nature and what is in accordance with nature?

This investigation, the Stoics believe, will make it reasonable for us to make certain presumptions about what nature’s purposes for us are. When we act in accordance with these presumptions, we will act appropriately. One way of thinking of what is in accordance with nature is presented in the following terms by Epictetus:26 Πῶς οὖν λέγεται τῶν ἐκτός τινα κατὰ ϕύσιν καὶ παρὰ ϕύσιν; ὥσπερ ἂν εἰ ἀπόλυτοι ἦμεν. τῷ γὰρ ποδὶ κατὰ ϕύσιν εἶναι ἐρῶ τὸ καθαρῷ εἶναι, ἀλλ᾿, ἂν αὐτὸν ὡς πόδα

25  The Greek text cited is that of H. Cherniss (ed.), Plutarch: Moralia, vol. xiii. 2: Stoic Essays (Cambridge, Mass., 1976). 26  The Greek text is that of W. A. Oldfather (ed.), Epictetus: The Discourses as Reported by Arrian, the Manual, and Fragments, vol. i (Cambridge, Mass., 1925).

310

Ralph Wedgwood

λάβῃς καὶ ὡς μὴ ἀπόλυτον, καθήξει αὐτὸν καὶ εἰς πηλὸν ἐμβαίνειν καὶ ἀκάνθας πατῆσαι καὶ ἔστιν ὅτε ἀποκοπῆναι ὑπὲρ τοῦ ὅλου· εἰ δὲ μή, οὐκετι ἔσται πούς. τοιοῦτόν τι καὶ ἐϕ᾿ ἡμῶν ὑπολαβεῖν δεῖ. τί εἶ; ἄνθρωπος. εἰ μὲν ὡς ἀπόλυτον σκοπεῖς, κατὰ ϕύσιν ἐστὶ ζῆσαι μέχρι γήρως, πλουτεῖν, ὑγιαίνειν. εἰ δ᾿ ὡς ἄνθρωπον σκοπεῖς καὶ μέρος ὅλου τινός, δι᾿ ἐκεῖνο τὸ ὅλον νῦν μέν σοι νοσῆσαι καθήκει, νῦν δὲ πλεῦσαι καὶ κινδυνεῦσαι, νῦν δ᾿ ἀπορηθῆναι, πρὸ ὥρας δ᾿ ἔστιν ὅτ᾿ ἀποθανεῖν. (Diss. 2. 5. 24–5) So, what is meant by saying that some externals are in accordance with nature and some contrary to nature? This is said as if we were detached entities. For I would say that it is according to nature for the foot to be clean; but if we consider it as a foot and as not detached, it will be appropriate for it to walk in the mud and tread on thorns and sometimes to be chopped off for the sake of the whole. Otherwise, it will no longer be a foot. We should assume a similar point for us too. What are you? A human being. If you look at yourself as a detached entity, it is according to nature for you to live to be old, to be rich, to be healthy. But if you look at yourself as a human being and as part of a certain whole, it is appropriate on account of that whole for you now to get sick, now to take a dangerous voyage, now to be poor, and sometimes to die an early death.

Here Epictetus implies that if we were ‘detached entities’, it would be in accordance with our nature to live long and be healthy; but in fact, he implies, we are not detached entities of this kind. Thus, what is in truly accordance with our nature is just whatever serves the purposes of the whole of which we are a part—­which sometimes involves suffering ill health or dying young. As he explicitly says, ‘The whole is more authoritative than the part, and the city is more authoritative than the citizen’ (κυριώτερον δὲ τὸ ὅλον τοῦ μέρους καὶ ἡ πόλις τοῦ πολίτου, Diss. 2. 10. 5). So, does Epictetus completely reject the idea that good health and longevity are ‘in accordance with our nature’? Not quite: he approvingly cites Chrysippus’ claim that ‘god himself has made me prone to select’ such outcomes as health and longevity (αὐτὸς γάρ μ᾿ ὁ θεὸς ἐποίησεν τούτων ἐκλεκτικόν, Diss. 2. 6. 9–10), and that for this reason, outcomes like health and longevity are ‘more nat­ural­ly suited for getting what is in accordance with nature’ (εὐϕυεστέρων . . .  πρὸς τὸ τυγχάνειν τῶν κατὰ ϕύσιν, 2. 6. 9) or ‘more naturally suited for selection’ (πρὸς ἐκλογὴν εὐϕυεστέρων, 2. 10. 6). The thought seems to be that because health and longevity are in this way ‘more naturally suited for selection’, we should presume—­in the absence of special reasons to think that the purposes of the cosmos require



Hierocles’ Concentric Circles

311

us to get sick or to die young—­that it is appropriate for us to continue to select health and longevity.27 There are a number of concerns that can be raised about Epictetus’ conception of how we can find out what it is appropriate for us to do. First, Epictetus seems to imply that the cosmic citizen’s role is to serve the cosmic city, but it is not the role of the city to serve the citizen: the citizen is totally subservient to the city. This raises a concern about whether Epictetus’ conception makes the cosmic city into something like a totalitarian state. A second concern is that we are told very little about how we can find out what the natural purposes of the cosmos are. If we are virtually always in the dark about these cosmic purposes, why should we ever select a course of action that involves a risk of ill health or premature death—­why should we not always ‘presume’ (at least in advance of finding out what fate holds in store for us) that it is appropriate for us to try to preserve our lives and health for as long as possible? Finally, it seems that according to Epictetus what is appropriate is fundamentally a matter of our fulfilling the roles that have been assigned to us. But there are many cases in which the roles that appear to be assigned to us conflict with each other. What, if anything, can we say about how to act in these cases?28 According to the interpretation that I shall propose in the rest of this paper, Hierocles aims to adjust and supplement the picture that we have sketched so far in a way that allows it to address these concerns.

5.  Hierocles’ investigation of duties Virtually all the excerpts from Hierocles in Stobaeus’ Anthology provide explanations of claims about what it is appropriate to do on the basis of claims about nature. For example, he appeals to

27 In Gorgias (the ‘helmsman’ passage, 511 e 6–512 b 1), Plato explores the idea that vicious agents are not made ‘better off’ by the preservation of their lives, so that preserving their lives is not, in fact, the right thing to do. Strikingly, the Stoics firmly reject this idea, insisting that it is normally appropriate even for vicious agents to stay alive, while it is often appropriate for virtuous agents to end their lives (Plut. Stoic. repugn. 1039 d–­e). 28  For Epictetus’ difficulties with conflicting roles, see especially Johnson, Role Ethics, ch. 7.

312

Ralph Wedgwood

nature to explain why it is appropriate for most of us to get­ married: ἔοικε δὲ . . . παρακαλεῖν ἡμᾶς ἡ καὶ αὐτὸν τὸν σοϕὸν ἐπὶ τὸν γάμον ἐξοτρύνουσα ϕύσις, ἥ τις οὐ συναγελαστικοὺς ἡμᾶς ἀπειργάσατο μόνον, ἀλλὰ καὶ συνδυαστικούς, μετὰ τοῦ ἕν τε καὶ κοινὸν ἔργον ὑποθεῖναι τῷ συνδυασμῷ· λέγω δὲ τὴν παίδων γένεσιν καὶ βίου διεξαγωγὴν εὐσταθοῦς. (4. 22a. 22, 4. 502. 15–20 Wachsmuth–­ Hense) It seems that . . . nature, which motivates even the wise man towards marriage, summons us to it too, since it not only fashions us as social beings but also as beings who live in couples, along with setting down a single common function for the couple: I mean the generation of children and the leading of a stable life.

He then justifies appealing to nature in this way in the only passage in the excerpts in the Anthology that makes a glancing reference to the theory of oikeiōsis: δικαία δὲ διδάσκαλος ἡ ϕύσις, ὅτι τῇ παρ’ αὐτῆς κατασκευῇ σύμϕωνον τὴν ἐκλογὴν χρὴ γίνεσθαι τῶν καθηκόντων. ζῇ γοῦν ἕκαστον τῶν ζῴων ἑπομένως τῇ ἑαυτοῦ ϕυσικῇ κατασκευῇ . . . χρώμενα . . . τὰ δὲ ζῷα ϕαντασίαις τε σπώσαις ἐπὶ τὰ οἰκεῖα καὶ ἐξελαυνούσαις προθυμίαις. ἡμῖν δὲ ἡ ϕύσις ἔδωκε τὸν λόγον . . . αὐτὴν κατοψόμενον τὴν ϕύσιν, ὅπως ὡς πρός τινα σκοπὸν εὐϕεγγῆ τε καὶ ἀραρότα τεταμένος ταύτην, ἐκλεγόμενός τε τὸ σύμϕωνον αὐτῇ πᾶν καθηκόντως βιοῦντας ἡμᾶς ἀπεργάζοιτο. (4. 502. 20–503. 10) Nature is a just teacher, since our selection, if it agrees with the constitution that comes from her, will necessarily be of what is appropriate. Indeed, each of the animals lives in a way that follows its own natural constitution . . .  using impressions that draw them towards what is proper to them (οἰκεῖα) and desires that drive them on. But nature gave us reason . . . so that reason may see nature herself and, as if intent upon a well-­lit and fixed target, select what agrees with her in every case and make us live in an appropriate way.

As Hierocles makes clear, however, he is not claiming that married life is always appropriate for all human beings, since sometimes unmarried life is rationally selected, on account of some special ‘circumstance’:29 τῷ σοϕῷ προηγούμενος μέν ἐστιν ὁ μετὰ γάμου βίος, ὁ δ᾿ ἄνευ γυναικὸς κατὰ περίστασιν. (4. 502. 10–11) 29  For a similar claim that some acts are appropriate ‘independently of circumstance’ (ἄνευ περιστάσεως), while others are ‘circumstantial’ (περιστατικά), see D. L. 7. 109 (= LS 59E 3 = SVF iii. 496).



Hierocles’ Concentric Circles

313

For the wise man, married life is the favoured option, but life without a wife is so according to circumstance.

If marriage is sometimes but not always appropriate, what should we say about whether marriage is ‘in accordance with nature’ or not? I propose to interpret Hierocles as drawing a fundamental distinction between (1) the natural purposes of particular parts of the cosmos and (2) the natural purposes of the cosmos as a whole. As I noted in Section 4, interfering external factors can prevent particular parts of the cosmos from fully fulfilling their nature, although no interfering external factors can prevent the whole cosmos from fulfilling its nature. Hierocles’ view seems to be that it is always part of the natural purposes of each individual household that it should be led by a married couple: as he puts it, the household of an unmarried person is always ‘half-­complete’, while that of a married couple is ‘complete and full’ (οἶκός τε ἡμιτελὴς μὲν τῷ ὄντι ὁ τοῦ ἀγάμου, τέλειος δὲ καὶ πλήρης ὁ τοῦ γεγαμηκότος, 4. 22a. 21, 4. 502. 5–7); as he says elsewhere, a household without marriage is always incomplete (ἀτελῆ, 4. 22a. 23, 4. 503. 12–13). However, circumstances can arise in which it is in accordance with the nature of the cosmos for a particular household to be led by a single person. Moreover, in some cases, this might be something that human experience of nature can reveal—­and in cases of this sort, it would be appropriate to select a life without marriage. If this interpretation of Hierocles is correct, he can retain a stronger attachment to the idea that health and longevity are ‘in accordance with nature’ than Epictetus seems to do. Hierocles can confidently say that my health and longevity are always in accordance with my individual animal nature—­even if he cannot say that they are always in accordance with the nature of the cosmos. This fits with the way in which he states without qualification that ‘illness . . . , disability, death’ (νόσου . . . , πηρώσεως, θανάτου) are ‘contrary to nature’ (παρὰ ϕύσιν, 2. 9. 7, 2. 182. 11–13). It also fits with his discussion of oikeiōsis: as we saw in Section 3, he argues that every animal always recognizes itself and its constitution as oikeia (as ‘its own and belonging to it’). On my interpretation, the animal and its constitution are oikeia to it because their preservation is in accordance with the animal’s individual nature. As I read the excerpts from Hierocles in Stobaeus’ Anthology, then, his goal in this work is to derive a series of principles about how it is appropriate to act on the basis of this conception of nature.

314

Ralph Wedgwood

Admittedly, the surviving excerpts do not provide an account of all appropriate acts—­not even of those appropriate acts that contemporary philosophers would describe as ‘duties’. The surviving excerpts include no discussion of our negative duties of non-­maleficence (our duties not to harm the innocent) or of our duties of fidelity (our duties to keep promises and the like). We should not assume that Hierocles does not recognize such duties or that he thinks that they are derived from the duties that he explicitly discusses; indeed, he may have explained these duties in passages that have not survived. Instead, the surviving excerpts explore only certain specific kinds of duties. For example, the excerpts on our duties to the gods (Stob. 1. 3. 53–4 and 2. 9. 7), our country (3. 39. 35–6), and our parents (4. 25. 53) explore certain duties of hierarchical sub­or­din­ ation, that is, duties that we have towards beings who in Hierocles’ view have natural authority over us and whom we are duty-­bound to treat as our superiors. By contrast, the duties explored in the circles passage are both positive and reciprocal: they are duties that the members of each circle have to provide certain positive forms of service to each other in virtue of their shared membership of that circle. For example, the members of your immediate family are duty-­bound to help each other in certain ways in virtue of their membership of that family group; similarly, fellow citizens are duty-­bound to help each other in certain other ways in virtue of their shared citizenship; and so on. (It is important to see that, for Hierocles, one’s duties towards one’s country, considered as an enduring institutional body defined by its legal and social traditions, are distinct from the reciprocal duties that individual fellow citizens have towards each other. This is why these two duties are discussed in different parts of the text, with the former duties being discussed in the excerpt on ‘How one should treat one’s country’ and the latter reciprocal duties being discussed in the circles passage.) Hierocles’ account of our duties to our siblings (Stob. 4. 27. 20) is also concerned with reciprocal positive duties of the same general kind. He characterizes these fraternal duties in the following terms: ὅθεν καλῶς ἡ ϕύσις, ὡς ἂν ἐϕ’ ἃ γεννᾷ μὴ ἀγνοοῦσα, παρήγαγεν ἡμῶν ἕκαστον τρόπον τινὰ μετὰ συμμαχίας. οὐδεὶς οὖν ἐστι μόνος οὐδ’ ἀπὸ δρυὸς οὐδ’ ἀπὸ πέτρης, ἀλλ’ ἐκ γονέων καὶ μετ’ ἀδελϕῶν καὶ συγγενῶν καὶ ἄλλων οἰκείων. (4. 664. 4–9)



Hierocles’ Concentric Circles

315

Hence nature has, as though not ignorant of why it creates us, admirably brought each of us into the world with, in a way, an alliance. Thus, no one is alone or born from an oak or a rock, but from parents and with siblings and relatives and other members of the household.

As he also puts it, siblings are ‘the helpers and allies who have been provided by nature herself’ (τῶν . . . παρ’ αὐτῆς χορηγουμένων τῆς ϕύσεως . . . βοηθῶν κἀπικούρων, 4. 664. 16–18). Our duty to our siblings is to be their allies, that is, to stand in a dependable relationship in which each sibling is ready to help another when needed.30 Thus, the duties explored both in the circles passage and in the discussion of siblings are what John Rawls calls duties of mutual aid.31 This idea of duties of mutual aid is somewhat different from a simpler conception that some contemporary moral philosophers have of our duties to help other people. According to the simpler conception, the fact that an available act would benefit another person is just in itself a reason in favour of the act. The idea of mutual aid involves a further element—­the idea of a mutual relationship between the two parties. It is this richer idea of duties of mutual aid that Hierocles explores in these passages.32 Our task is to understand how Hierocles attempts to explain such duties of mutual aid on the basis of his investigations of nature.

6.  Organic parts within parts In Section 4 above, we saw how Epictetus compares your relationship to the larger whole to which you belong with the relationship between a foot and the person to whom it belongs (Diss. 2. 5. 24).

30  This idea, that those who regard each other as ‘brothers’ naturally treat each other as ‘helpers and allies’, also plays an important role in Plato’s Republic (3, 414 e 6, 416 b 2, 417 b 1), as has recently been shown by F.  C.  C.  Sheffield, ‘Moral Motivation in Plato’s Republic? Philia and the Return to the Cave’, Oxford Studies in Ancient Philosophy, 59 (2021), 79–131. Hierocles also uses the language of ‘allies’ to describe the relationship between spouses (4. 22a. 24, 4. 505. 1–2 and 4. 507. 4) and between parents and their children (4. 25. 53, 4. 641. 21). 31  See J. Rawls, A Theory of Justice (Cambridge, Mass., 1971), 114. 32  Compare how Seneca (Ben. 2. 17. 3–5) appeals to Chrysippus’ image of a ball game: we should help others like a skilled ballplayer who throws the ball in such a way that the other player can catch and return the ball to us, that is, we should help others in such a way as to create and sustain a relationship of mutual aid.

316

Ralph Wedgwood

Similar comparisons are made in many Stoic texts, as well as in Christian texts of the same period. For example:

i. Seneca says: ‘This universe that houses us is a unity, and is God; we are God’s companions, God’s limbs’ (Totum hoc, quo continemur, et unum est et deus; et socii sumus eius et membra; Ep. 92. 30).33 ii. According to Marcus Aurelius, each of us should say to themselves, ‘I am a limb of the system that is formed by all rational beings’ (μέλος εἰμὶ τοῦ ἐκ τῶν λογικῶν συστήματος, Med. 7. 13. 2).34 iii. The Apostle Paul says that Christians are ‘all baptized into one body’ (ἡμεῖς πάντες εἰς ἓν σῶμα ἐβαπτίσθημεν, 1 Cor. 12:13).35 Moreover, these Stoic and Christian texts frequently appeal to our all being limbs of a single greater organism to explain our duties of mutual aid and concern (for example, see Marc. Aur., Med. 2. 1. 4, and Paul, 1 Cor. 12:25).36 How exactly is our all being limbs of a single organism supposed to explain these duties of mutual aid? Presumably, it explains these duties because it is appropriate for all limbs of a single organism, in all their operations, to serve the same overarching purpose—­ whatever purpose is appropriate for the organism as a whole. Some contemporary ethicists may object that this explanation is incredible. Is it really plausible that the reason why it is appropriate for me to help my sister with clearing up after dinner is that I am thereby serving the purposes of the organism that is formed by all rational beings? Surely the rationale is just that this act helps my sister, not that it serves this vast corporate body?37 While I am sympathetic to this objection, it is unclear how else to interpret this recurrent theme in both Stoic and Christian thought. Moreover, as 33  The translation is that of M. Graver and A. A. Long (eds. and trans.), Seneca: Letters on Ethics (Chicago, 2015); the Latin text is cited from L. D. Reynolds (ed.), L. Annaei Senecae Ad Lucilium Epistulae morales (Oxford, 1965). 34  The Greek text is that of J. Dalfen (ed.), M. Aurelii Antonini Ad se ipsum, 2nd edn (Leipzig, 1987). 35 The Greek text is that of M.  W.  Holmes (ed.), The Greek New Testament, Society of Biblical Literature edn (Atlanta, 2010). 36  In a similar way, Cicero appeals to an analogy between human society and the parts of the body to explain our duties of non-­maleficence (Off. 3. 22). 37  This objection was forcefully put to me by an anonymous reader for this journal.



Hierocles’ Concentric Circles

317

I shall explain in the rest of this paper, Hierocles makes a small but crucial adjustment to this common theme. Combined with the conception of what is ‘in accordance with nature’ that I have ascribed to him, this adjustment at least softens the force of this objection. Specifically, Hierocles focuses on smaller wholes—­not the whole cosmos, but the household or the immediate family. These smaller wholes are themselves parts of larger wholes—­ thus creating a nested sequence of parts and wholes. For example, in his explanation of why siblings have a duty of mutual aid, he appeals to the fact that siblings are parts of the same household (οἶκος, 4. 27. 20, 4. 663. 5 Wachsmuth–Hense), while elsewhere he makes it clear that ­households are necessary parts of the city (οὔτε γὰρ πόλεις ἂν ἦσαν μὴ ὄντων οἴκων, 4. 22a. 21, 4. 502. 5); thus, siblings are parts of the household, while households are parts of the city. Similarly, in his discussion of our duty to our country, he compares citizens to fingers and their country to the hand to which the fingers belong (3. 39. 35, 3. 732. 1–8)—thereby implicitly reminding us that, just as the hand is part of the arm, which is part of the whole human being, so too our country is also part of a larger whole, such as ‘the whole race of human beings’ (τοῦ παντὸς ἀνθρώπων γένους, 4. 27. 23, 4. 672. 1–2). Specifically, Hierocles explains why siblings have a duty of mutual aid as follows: μετὰ ταῦτα δ’ ἐνθυμητέον, ὅτι τρόπον τινὰ οἱ ἀδελϕοὶ ταὐτοῦ μέρη τυγχάνουσιν, ὥσπερ οἱ ἐμοὶ ὀϕθαλμοὶ ἐμοῦ καὶ ὡσαύτως σκέλη τε καὶ χεῖρες καὶ τὰ λοιπά. καὶ γὰρ οὗτοι τοῦτον ἔχουσιν τὸν τρόπον, εἴ γε πρὸς τὸν οἶκον ἐξετάζοιντο. ὥσπερ οὖν οἱ ὀϕθαλμοὶ καὶ αἱ χεῖρες, εἴπερ ἕκαστον ἰδίαν ψυχὴν καὶ νοῦν λάβοι, περιέποι ἂν τὰ λοιπὰ πάσῃ μηχανῇ διὰ τὴν εἰρημένην κοινωνίαν, τῷ μηδ’ αὐτὰ τὸ ἴδιον ἔργον οἷάτε εἶναι παρέχειν καλῶς δίχα τῆς τῶν ἑτέρων παρουσίας· οὕτως δεῖ καὶ ἡμᾶς, ἀνθρώπους γε ὄντας καὶ ψυχὴν ὁμολογοῦντας ἔχειν, μηδὲν παριέναι σπουδῆς ὑπὲρ τοῦ δεόντως προσϕέρεσθαι τοῖς ἀδελϕοῖς. (4. 27. 20, 4. 663. 1–12) After this one should bear in mind that brothers are in a way parts of the same being, just as my eyes are parts of me, and in the same way my legs and hands and so on. For, in fact, brothers are related in this way, at least if they are judged in relation to the household. So, just as each of our eyes and hands, if it got its own soul and understanding, would take care of the other parts in every way on account of the community that has been described, since they are not even able to perform their special function well apart from the presence of the others, in the same way we, who are human beings and agree that we have a soul, ought to spare no effort in treating our brothers as we ought.

318

Ralph Wedgwood

For Hierocles, siblings are parts of the household, just as your eyes, legs, and hands are parts of you. He seems to assume that each of these parts, both of the household and of you, has an individual nature, involving an essential purpose or ‘special function’ (ἴδιον ἔργον, 4. 663. 8). Specifically, I propose that this individual nature involves two functions:

i. First, it involves what I shall call an internal function—­how it is this part’s purpose to operate internally, within itself. ii. Secondly, it involves an external function—­how it is this part’s purpose to relate to things outside it; specifically, it has the external function of ‘taking care’ (περιέποι, 4. 663. 7) of the other parts—­presumably, to help these other parts to fulfil the natural purposes of the larger wholes to which both it and these other parts belong.

This pattern of interlocking internal and external functions of parts and wholes is found both at the level of the individual human being and at the level of the individual household. The duty of mutual aid between siblings—­the duty of acting as one’s siblings’ ally—­is explained by each sibling’s external function as a part of the household. To make sense of this explanation in more detail, we need to understand what the internal function of corporate entities like households and cities might be. As we have seen in Section 3 above, according to Hierocles, every animal recognizes both itself and its constitution (σύστασις) as oikeia; as I read this, it implies that it is part of the internal function of each animal’s individual nature to preserve both itself and its constitution. Since Hierocles is comparing households and cities with whole organisms like animals, he seems to imply that it is also part of the internal function of cor­por­ate entities, like households and cities, to preserve both themselves and their constitutions. There is clear evidence that, for Hierocles, the individual nature of a city involves the purpose of the city’s preserving itself for as long as possible. As he says, in arguing that having children is one way to serve one’s city: ἀλλ’ εἰ μὲν ἦν ὀλιγοχρόνιόν τι σύστημα πόλις, ὅ τε βίος αὐτῆς κατὰ βίον ἀνθρώπου τὴν συμμετρίαν ἐλάμβανεν, οὐδὲν ἔδει διαδοχῆς. ἐπεὶ δ’ εἰς πολλὰς γενεὰς ἐξι­ κνεῖται, δαίμονι δὴ εὐδαιμονεστέρῳ χρησαμένη, καὶ εἰς μακροὺς αἰῶνας πόλις, ϕανερὸν ὡς οὐ τοῦ παρόντος ἐστοχάσθαι δεῖ μόνον, ἀλλὰ καὶ τοῦ μετέπειτα. (4. 24a. 14, 4. 605. 8–14)



Hierocles’ Concentric Circles

319

Now, if a city were a short-­lived organization and its life were com­men­sur­ ate with the life of a human being, there would be no need for succession. But since a city’s life extends for many generations—­indeed, if it enjoys a happier fate, for long ages of time—­it is clear that one must be focused not only on the present, but also on the time to come.

What about the purpose of preserving the constitution of a cor­por­ ate entity such as a city or a household? I propose that an animal’s constitution consists precisely in the proper organization and proper functioning of its organic parts.38 Thus, in effect, at least with non-­ rational animals, the preservation of an animal’s constitution is the preservation of its health.39 This explains why, as we saw in Section 5, good health is one of the Stoics’ most central examples of what is in accordance with an animal’s nature. In the case of a rational being, the proper functioning of its parts presumably also involves the proper functioning of its reason; and, for the Stoics, the proper functioning of reason consists in having genuine knowledge and virtue. So, for a virtuous rational being, the preservation of its constitution would also involve the preservation of knowledge and virtue. Thus, if a city or a household preserves its constitution, it must ensure the proper functioning of its parts. In this way, the city must take care of its citizens, protecting their lives and their health, and fostering their progress towards virtue; similarly, the household must also take care of its members, helping its members to fulfil the purposes of their individual natures. In this way, Hierocles does not simply view the parts as totally subservient to the whole, as Epictetus seems to have done (as we saw in Section 5). Instead, both parts and wholes have natural purposes that involve supporting each other. He may be gesturing in the direction of this thought when he says, in his discussion of our duties to our country, ‘the whole is nothing without the parts’ (τὸ γὰρ ὅλον δίχα τῶν μερῶν ἐστιν οὐδέν, 3. 39. 35, 3. 732. 17–18). In this way, Hierocles’ explanation of siblings’ duties of mutual aid, on the basis of their shared membership of the household or immediate family, is not obviously vulnerable to the objection that 38  Compare the use of the word for ‘constitution’ (σύστασις) in Plato, Rep. 8, 546 a 2. 39  For this conception of health, compare Plato, Rep. 4, 444 d 2–4, and Arius Didymus, Epitome of Stoic Ethics (ap. Stob. 2. 7. 5b. 4, 2. 62. 20–2 Wachsmuth–­ Hense = SVF iii. 278).

320

Ralph Wedgwood

I considered at the beginning of this section. It is plausible that when I help my sister with clearing up after dinner, the way in which I ‘take care’ of her is precisely as a fellow member of my immediate family. By treating her in this way, I am acting as her ally, cooperating with her to fulfil the purposes of the immediate family to which we both belong—­where these purposes centrally include the family’s taking care of all its members. In general, then, Hierocles postulates a system of organic parts and wholes: the finger is part of the hand, the hand of the whole human being; the sibling is part of the household, the household of the city, and so on. Each of these entities is conceived as an organic part of a larger living thing. Within this system of organic parts, the individual nature of each part involves the following purposes: (1) the internal function of preserving itself and its constitution—­ where preserving its constitution involves ensuring that its own internal parts fulfil the natural purposes of their individual natures; and (2) the external function of acting as an ally of every other part, by cooperating with that other part to help to fulfil the natural purposes of the larger wholes to which they both belong. 7.  The circles as organic parts of the universal body When Hierocles introduces the circles passage, he explains that it provides ‘the account of our relatives’, stating that ‘in a way the same thing happens’ with this account as with the accounts that he provided earlier of ‘our parents, siblings, wife, and children’ (Τοῖς εἰρημένοις περὶ γονέων χρήσεως καὶ ἀδελϕῶν γυναικός τε καὶ τέκνων ἀκόλουθόν ἐστι προσθεῖναι καὶ τὸν περὶ συγγενῶν λόγον, συμπεπονθότα μέν πως ἐκείνοις, 4. 27. 23, 4. 671. 4–6). So, we should expect his explanations of our duties of mutual aid to our relatives to be exactly parallel to those that he gave of our duties of mutual aid towards our siblings. As I shall now explain, it is plausible to read the circles passage in exactly this way. First, each circle is an organic part of a larger living whole; even the outermost circle is a part of the larger cosmos as a whole—­and according to the Stoics, the cosmos as a whole is itself also a living thing.40 Secondly, each circle contains parts that are themselves 40  See D. L. 7. 139 (= LS 47O 3, in part = SVF ii. 634): ‘Thus, the whole cosmos is a living thing both animate and rational’ (οὕτω δὴ καὶ τὸν ὅλον κόσμον, ζῷον ὄντα καὶ ἔμψυχον καὶ λογικόν).



Hierocles’ Concentric Circles

321

smaller living wholes; even the innermost circle contains your individual mind and other body parts. In this way, the larger circles are related to the smaller circles as macrocosms to microcosms; the nat­ural purposes of all of these circles are parallel to each other.41 As we have seen, the duties of mutual aid that siblings have towards each other are explained by the fact that ‘brothers are in a way parts of the same being’ (τρόπον τινὰ οἱ ἀδελϕοὶ ταὐτοῦ μέρη τυγχάνουσιν, 4. 27. 20, 4. 663. 2–3). In just the same way, the duties of fellow citizens to each other are grounded in their being organic parts of their country; the duties that every human being owes to every other human being are grounded in their being organic parts of the entire human race. More generally, each of these duties that you owe to another human being is, fundamentally, the duty of acting as an ally of that other human being, cooperating with the other human being to help to fulfil the natural purposes of all the larger circles of which you are both parts. In general, I propose that Hierocles would apply the same account to each system of circles as to other systems of organic parts and wholes. Thus, the individual nature of each circle C involves the following internal and external functions: 1. Internally, this circle C has the function of preserving itself and its constitution for as long as possible—­where preserving its constitution involves helping to ensure that every smaller circle C − contained within C fulfils the purposes of its individual nature too. 2. Externally, for every larger circle C + that contains this circle C, C has the function of cooperating with any other circles that are contained within C + to help to fulfil the purposes of C +’s individual nature as well. To put it roughly and intuitively, at every level of organic structure, there are (1) internal functions that aim at health and longevity, and (2) external functions that aim at harmonious relations of cooperation between the parts and the wholes to which those parts belong. This account goes a long way towards addressing the concerns about Epictetus’ picture that I raised at the end of Section 4. First, while Epictetus’ description made the cosmic city seem like a 41  The ultimate inspiration for this idea may well lie in the analogy between the city and the individual soul that was central to Plato, Rep. 2, 368 c 8–e 6.

322

Ralph Wedgwood

to­tali­tar­ian regime in which the citizen is totally subservient to the state, Hierocles’ description is rather different. Admittedly, it hardly makes the cosmic city seem like a liberal state, since it gives no special primacy to the welfare or autonomy of the individual citizen, but it seems to resemble a society that meets a broadly corporatist version of a communitarian ideal.42 Secondly, it suggests a much richer account of the purposes of the cosmos: since the cosmos as a whole is also a living thing, the purpose of its nature is presumably analogous, on the macro level, to the purpose of each of the circles. We are thus given clearer guidance about what the purposes of the cosmos might be than anything that we can glean from the surviving words of Epictetus. Finally, I believe that Hierocles’ account can provide the basis for a clearer account of how we should act when the roles that we have in virtue of our membership of different circles come into conflict. Let us take as our example the story of Regulus, as told by Cicero (Off. 3. 99–115), who sacrifices his life to save Rome from defeat at the hands of the Carthaginians. To keep things simple, suppose that Regulus has two options: (1) saving his own life but allowing Rome to be defeated, and (2) saving Rome from defeat but allowing himself to be killed. The innermost circle—­containing only Regulus himself—­has as its internal function that Regulus should save his own life, but the circle containing all his fellow citizens has as its internal function that he should save Rome. In this case, however, he cannot save both. How can Stoic ethics explain what Regulus should do? Unfortunately, no surviving ancient text contains an explicit answer to this question.43 We can only speculate about what some 42 For an example of such a corporatist communitarian political theory, see F. Tönnies, Community and Civil Society, ed. by J. Harris, trans. by J. Harris and M. Hollis (Cambridge, 2001). 43  Cicero (Off. 3. 21) tries to account for some cases of apparent conflict with this ‘formula’: ‘for one man to take something from another and to increase his own advantage at the cost of another’s disadvantage is more contrary to nature than death, than poverty, than pain, and than anything else that may happen to his body or external possessions’ (Detrahere igitur alteri aliquid et hominem hominis incommodo suum commodum augere magis est contra naturam quam mors, quam paupertas, quam dolor, quam cetera quae possunt aut corpori accidere aut rebus externis). But the point of Cicero’s formula is that the negative duties of non-­ maleficence (not to ‘take’ something from another person that that other person is entitled to) override any apparent reasons to increase one’s own advantage. It does not address the question that concerns us here, namely, how to adjudicate apparently conflicting duties of mutual aid.



Hierocles’ Concentric Circles

323

Stoics might say about this. For example, some readers might speculate that the Stoics could adopt a quasi-­utilitarian view in which what is appropriate is always just to maximize the total sum of the normally preferred indifferents (longevity, bodily and mental health, and so on) in the universe as a whole.44 However, as I suggested in Section 2 above, this quasi-­utilitarian view seems out of keeping with Hierocles’ apparent belief that we have special duties towards our own parents, while we have lesser duties towards the parents of strangers. So I speculate that Hierocles would favour an alternative account. Here is a speculative suggestion about what such an alternative might be. Intuitively, it seems that your saving the life of your sister makes a greater contribution to the internal function of your immediate family than your saving the life of a single stranger on the far side of the world makes to the internal function of the entire human race. If your immediate family loses your sister, that is like a human being’s losing an arm or a leg, while if the human race loses this particular stranger, that is like the human being’s losing part of a finger, at most. At the same time, if you save your sister, you are making a small contribution to the internal function of the human race (since your sister is a member of the human race), while if you save a single stranger on the far side of the world, you are not making any contribution to the internal function of your immediate family (since that distant stranger is not a member of your immediate family). However, in the account that I sketched above, each circle has an external function, as well as an internal function; and in general the outer circles are supported by the external functions of a larger number of circles than the inner circles are. For this reason, the outer circles (such as the entire human race) are in a way cosmically more important than the inner circles (such as your immediate family). This suggests the following formal model. Suppose that, for each act A, and for each circle i, there is a measure mi  (A) of how much this act A contributes to the internal functions of this circle i. Similarly, suppose that, for each circle j, there is a weight αj that captures the cosmic importance of this circle j. Then, the measure of how great a contribution an act A makes to the duties of aid that 44  This speculation is considered by T.  Brennan, ‘Stoic Moral Psychology’, in B. Inwood (ed.), The Cambridge Companion to the Stoics (Cambridge, 2003), 257–94 at 232.

324

Ralph Wedgwood

the agent has to the people in all ten circles would be: α1m1(A) + α2m2(A) + . . . + α10m10(A). This model can explain why Regulus’ duties of aid require him to sacrifice his life to save Rome. To fix ideas, suppose that if A is an act whose only significant effect is the saving of lives, then, for each circle i, mi(A) is given by the proportion of the population of this circle i whose lives are saved, and the weight αi is simply given by this circle i’s population; so, αimi(A) is given simply by the number of members of this circle i whose lives are saved.45 So, if Regulus saves his own life, the measure of his contribution to the duties of aid that he has is exactly 10 (that is, 1 for each circle, since he is a member of all ten circles). By contrast, suppose that if he saves Rome from defeat, he will thereby save no members of Circles 1–6, but he will save 5,000 people who are members of Circles 7–10. Then, the measure of his contribution to these duties of aid is 20,000 (0 for Circles 1–6, but 5,000 each for circles 7–10). Since 20,000 > 10, the act that makes the greatest contribution to these duties of aid is for Regulus to sacrifice his life to save Rome. By contrast, if you save the life of your sister, the measure of your contribution is 9 (0 for Circle 1, and 1 for each of Circles 2–10), while if you save a single stranger on the far side of the world, the measure of your contribution is 1 (0 for each of Circles 1–9, and 1 for Circle 10). Since 9 > 1, the act that makes the greatest contribution to your duties of aid is for you to save your sister rather than the distant stranger. In general, this approach can provide an account of what has long seemed one of the most puzzling features of Stoic ethics. Many scholars have formulated this puzzle as concerning ‘how the Stoic is supposed to deliberate’.46 Fundamentally, though, it is just the puzzle of how precisely to explain what makes an act appropriate, that is, one that it is rational to select. According to the in­ter­ pret­ation that I propose, the Stoic sage’s deliberation has two stages. First, the sage discovers the individual natures both of himself and of other organic parts of the cosmos (including of all 45 Arguably, a more intuitive weighting of the circles would be given by an increasing but concave function of each circle’s population (such as the square root of the population). The weighting suggested here is given only for the sake of simplicity. 46  For this formulation, see R. Barney, ‘A Puzzle in Stoic Ethics’, Oxford Studies in Ancient Philosophy, 24 (2003), 303–40, and T. Brennan, The Stoic Life: Emotions, Duties, and Fate (Oxford, 2005), 182–230.



Hierocles’ Concentric Circles

325

of the circles surrounding each person). Secondly, the sage determines which of the available acts appears to make the greatest ­contribution to the internal and external functions of his own nature—­including his external function of supporting the natural purposes of other parts of the cosmos (in the way that is appropriate for his particular position in the cosmos). The sage should presume that this act is what the purpose of the cosmos as a whole requires of him—­to the extent that human experience can shed light on the matter. This act is therefore appropriate and rational for the sage to select. In addition, this approach can also give an account of why the acts of the divine world mind are appropriate and of why the selections of this cosmic mind are rational. Since there is nothing outside the cosmos except the infinite void (D. L. 7. 140), the natural purpose of the cosmos consists purely in its internal function of preserving itself and its constitution for as long as possible—­and in the case of the cosmos, this involves preserving itself and its constitution forever.47 The constitution of the cosmos involves the proper organization and proper functioning of its organic parts. The hallmark of this proper organization and proper functioning is that each organic part of the cosmos is related to the whole as microcosm to macrocosm, and the purposes of the whole and the purposes of the parts are, at least for the most part, harmonious and mutually supporting. It is this, I suggest, that makes the cosmic plan of the divine world mind so sublimely rational. At all events, if these proposals are at least roughly correct, the Stoics are not left with nothing to say, either about how the individual Stoic sage should deliberate or about what makes nature’s providential plan so supremely rational.48

47  Admittedly, the ‘self-­preservation’ of the Stoic cosmos involves endless cycles of conflagration and renewal. But in many ways this makes the temporally limited self-­preservation of an individual human being (who ages, dies, and then returns in the next cosmic cycle) more similar to the self-­preservation of the Stoic cosmos than to that of the endlessly stable cosmos that was envisioned by Aristotle. 48  In discussing Stoicism, Cooper refers repeatedly to the ‘beauty’ of the cosmos, as if this is what makes the activity of the divine world mind so rational; for example, see J.  Cooper, Pursuits of Wisdom: Six Ways of Life in Ancient Philosophy from Socrates to Plotinus (Princeton, 2012), 170. But I doubt that this gets to the heart of the matter. No doubt the Stoics would agree that the cosmos is sublimely beautiful. But if I am right, they would not agree that this is why its activity is rational. On the contrary, for the Stoics, the cosmos is beautiful because it is rational; it is not rational because it is beautiful.

326

Ralph Wedgwood 8.  Solving the puzzles about the circles

Finally, we may return to the puzzles about the circles that I raised in Section 1. The proposals of Section 7 are designed to answer the fourth and final puzzle, that is, the question of what in Stoic ethics explains why the claims that Hierocles makes about the circles are true. But this still leaves us with the remaining three puzzles. How are we to make sense of Hierocles’ talk of ‘transferring those in the containing circles into the circles that are contained within them’? How should we make our honour and goodwill towards those in Circle 3 similar to our honour and goodwill towards those in Circle 2? What is the import of the metaphor of ‘drawing the circles together towards the centre’? The first puzzle arises because the membership of the circles seems to be determined by biological, legal, and geographical facts. Membership of the circle of my immediate family is determined by the biological facts of who my parents, siblings, and children are, and by the legal fact of who is my spouse; membership of the circle of my fellow citizens is determined by the legal facts of who is, and who is not, a citizen of the relevant country; membership of the circle of ‘those of neighbouring countries’ is determined by the geographical facts of where these countries are located. It is not plausible that Hierocles is advocating that we should change these legal, biological, or geographical facts (say, by encouraging a mass migration of people into the countries that neighbour our own).49 Nonetheless, in some way, Hierocles is recommending that you should treat your cousins as if they were your siblings, treat the citizens of neighbouring countries as if they were your fellow citizens, and so on. I propose the following interpretation: for each of these circles, there is a certain degree of goodwill (εὔνοια, 4. 27. 23, 4. 672. 17 Wachsmuth–­Hense) and effort (σπουδή, 4. 672. 5 and 4. 673. 9) that you are socially expected to invest into helping the ­people in this circle—­where the closer the circle is to the centre, the greater this degree of goodwill and effort is. However, as Hierocles sees it, what is socially expected of us is systematically less than is 49  In his discussion of our duties to our country, Hierocles argues for a strong kind of political conservatism, insisting on the importance of preserving our country’s constitution and laws (Stob. 3. 39. 36, 3. 733. 8–734. 10 Wachsmuth–­Hense). This would presumably include the laws defining citizenship. So Hierocles is surely not advocating that these laws should be reformed to become more inclusive.



Hierocles’ Concentric Circles

327

really appropriate. The degree of goodwill and effort that it is really appropriate for you to invest into helping your cousins is the degree that you are socially expected to invest into helping your siblings; and similarly for the other circles.50 Thus, if you treat your cousins appropriately, you will in this way be treating them as if they were your siblings; in this way, it is as if you have transferred your cousins from the third circle into the second. Strictly, however, the help that you provide to your cousins is aimed at fulfilling the natural purposes of the circles to which you and your cousins both belong—­ which do not include the second circle that includes only your immediate family. The second puzzle concerns how to make sense of Hierocles’ claim that one should make one’s honour and goodwill towards those in each containing circle n + 1 similar to one’s honour and goodwill towards those in the smaller circle n that is immediately contained within it. In fact, we can interpret this claim as making the same point as the metaphor of ‘drawing the circles closer together’. If you draw these two circles closer together, your treatment of these two circles is in a way more similar than if you push these circles farther apart. Thus, we can give the same solution to this puzzle as to the third puzzle, which concerns the import of the metaphor of ‘drawing the circles together’. As we have seen, Hierocles implies that this ‘contraction of the circles’ has already been outlined (4. 673. 10–11 Wachsmuth–­ Hense) in the preceding argument (4. 672. 2–673. 7). So, the image of ‘drawing the circles together’ must somehow be justified by the ideal of your treating every other person as an ally by helping them to fulfil the natural purposes of all the circles to which you both belong. Any interpretation of this metaphor needs two elements: first, some way of making sense of the metaphor of ‘the distance of our relation to each person’ (τὸ μῆκος τῆς πρὸς ἕκαστον τὸ πρόσωπον σχέσεως, 4. 672. 20–673. 1); and secondly some case that can serve as a benchmark of comparison, so that when we follow Hierocles’ recommendations this distance is less than it is in this benchmark case. 50  In this way, my interpretation is close to the common reading on which Hier­ ocles is urging us to transfer the ‘affection’ that it is normal for most human beings to feel for members of the closer circles onto members of the outer circles; for a reading of this kind, see, for example, Veillard, ‘Hiéroclès’, 106.

328

Ralph Wedgwood

I propose to interpret this metaphor of ‘distance’ and ‘closeness’ in the following way: the degree to which you are close to another person is given by the extent to which your efforts are directed towards promoting that person’s natural purposes. As we have seen, according to the Stoics, each person has a range of natural purposes: the purpose of your rational nature is to perfect your reason by achieving knowledge and virtue; the purposes of your animal nature involve the internal function of preserving your life and health and the external function of helping each other person to fulfil the purposes of all the circles to which you both belong. So, for example, suppose that Sibyl is a sage. Then, all her efforts would be entirely directed towards promoting her own natural purposes, and by this measure, she is as close to herself as it is possible to be. If Sibyl has a friend—­another sage, Isaiah—­then, some of Sibyl’s efforts would be directed towards promoting Isaiah’s natural purposes; for example, she would put effort into helping him to maintain his knowledge and virtue and to preserve his life and health. But some of her efforts would be directed towards purposes that do not centrally involve Isaiah in this way, for example, to maintain her own virtue and health or serve social circles to which Sibyl belongs but Isaiah does not. In this way, her efforts would not be entirely directed towards promoting Isaiah’s natural purposes, and so she has a greater distance from Isaiah than from herself. According to Hierocles, when we act appropriately, our distance from each person is less than in a certain case that serves as a benchmark of comparison. I propose that this benchmark of comparison is what Hierocles may have regarded as the commonest form of vice. Specifically, Hierocles may have agreed with Plato (Rep. 9, 586 a 6–b 5) that the commonest form of vice is a kind of selfish greed for external things, such as material wealth and honour. A selfish person of this kind pursues many of his own natural purposes (such as his own health and longevity), while totally ignoring the natural purposes of most other people. As a result, the selfish person at best expends no effort to promote the natural purposes of most other people and at worst actively works against their natural purposes. Thus, with the selfish person, the inner circle is relatively close, but all the other circles are far away, since the extent to which the selfish person’s efforts are directed towards promoting the nat­ ural purposes of others is so small. By contrast, if you act appropriately in all that you do, the circles are all much closer than with



Hierocles’ Concentric Circles

329

the selfish person: you act as an ally of every other person, working together to fulfil the natural purposes of all the circles to which you both belong—­and so, for every other person, your efforts are to a significant extent directed towards promoting that other person’s natural purposes. Still, on this interpretation, if you are virtuous, although the circles around you are much closer than the circles around the selfish person, the circles are still at some distance from each other. Your efforts are directed more towards promoting your own natural purposes than your sibling’s natural purposes, more towards promoting your fellow citizen’s natural purposes than a distant foreigner’s, and so on. So, the metaphor of ‘drawing the circles together’ is not the same as eliminating all distance between the circles. We do not have to take this metaphor as advocating acting impartially for the good of the world in the style of William Godwin. Hierocles’ ideal, in short, is that we should help and cooperate with every other human being—­but not indiscriminately. On the contrary, the form that our cooperation with each person should take must reflect the nature of our relationship with them—­where this relationship is constituted by our position in a network of social relations. In general, Hierocles’ surviving work supports the conclusion that his ethical thought stands out for its coherence, its philosophical insight, and its deep humanity. The fact that the Stoic tradition was still capable of inspiring ethical thought of this quality, some 450 years after the school’s foundation, is a testament to the richness and fertility of that tradition. University of Southern California

BIBLIOGR A PH Y Alesse, F., ‘La représentation de soi et les différentes formes de l’appropriation chez Hiéroclès’, in Gourinat (ed.), L’éthique, 63–85. Annas, J., The Morality of Happiness (Oxford, 1993). Arnim, H., von (ed.), Hierokles: Ethische Elementarlehre (Papyrus 9780) nebst den bei Stobäus erhaltenen ethischen Exzerpten aus Hierokles (Berlin, 1906). Arnim, H., von (ed.), Stoicorum Veterum Fragmenta [SVF], 4 vols. (Leipzig, 1903–24).

330

Ralph Wedgwood

Barney, R., ‘A Puzzle in Stoic Ethics’, Oxford Studies in Ancient Philosophy, 24 (2003), 303–40. Bastianini, G. and Long, A.  A., ‘Ierocle: Elementi di Etica’, Corpus dei papiri filosofici greci e latini [CPF], vol. i. 1. 2 (Florence, 1992). Bobzien, S., Determinism and Freedom in Stoic Philosophy (Oxford, 1998). Brennan, T., The Stoic Life: Emotions, Duties, and Fate (Oxford, 2005). Brennan, T., ‘Stoic Moral Psychology’, in B. Inwood (ed.), The Cambridge Companion to the Stoics (Cambridge, 2003), 257–94. Cherniss, H. (ed. and trans.), Plutarch: Moralia, vol. xiii. 2: Stoic Essays (Cambridge, Mass., 1976). Cooper, J., Pursuits of Wisdom: Six Ways of Life in Ancient Philosophy from Socrates to Plotinus (Princeton, 2012). Dalfen, J. (ed.), Marci Aurelii Antonini Ad se ipsum, 2nd edn (Leipzig, 1987). Dorandi, T. (ed.), Diogenes Laertius: Lives of the Eminent Philosophers (Cambridge, 2013). Gill, C., ‘La continuité de la perception depuis la naissance’, in Gourinat (ed.), L’éthique, 47–62. Gloyn, L., The Ethics of the Family in Seneca (Cambridge, 2017). Godwin, W., An Enquiry Concerning Political Justice (London, 1793). Gourinat, J.-B., ‘Comment se détermine le kathekon?’, Philosophie Antique, 14 (2014), 13–40. Gourinat, J.-B. (ed.), L’éthique du stoïcien Hiéroclès [L’éthique] (Villeneuve d’Ascq, 2016). Gourinat, J.-B., ‘La gestation de l’animal et la perception de soi’, in Gourinat (ed.), L’éthique, 15–46. Graver, M. and Long, A. A. (eds. and trans.), Seneca: Letters on Ethics (Chicago, 2015). Griffin, M. T. and Atkins, E. M. (eds.), Cicero: On Duties (Cambridge, 1991). Holmes, M.  W. (ed.), The Greek New Testament, Society of Biblical Literature edn (Atlanta, 2010). Inwood, B., ‘Hierocles: Theory and Argument in the Second Century ad’ [‘Hierocles’], Oxford Studies in Ancient Philosophy, 2 (1984), 151–83. Johnson, B.  E., The Role Ethics of Epictetus: Stoicism in Ordinary Life [Role Ethics] (Lanham, 2014). Klein, J., ‘The Stoic Argument from oikeiōsis’, Oxford Studies in Ancient Philosophy, 50 (2016), 143–200. Lesses, G., ‘Austere Friends: The Stoics and Friendship’, Apeiron, 26 (1993), 57–75. Long, A.  A. and Sedley, D.  N. (eds.), The Hellenistic Philosophers [LS] (Cambridge, 1989). McCabe, M. M., ‘Extend or Identify: Two Stoic Accounts of Altruism’, in R.  Salles (ed.), Metaphysics, Soul, and Ethics in Ancient Thought: Themes from the Work of Richard Sorabji (Oxford, 2005), 413–43.



Hierocles’ Concentric Circles

331

Nussbaum, M. C., The Cosmopolitan Tradition: A Noble but Flawed Ideal (Cambridge, Mass., 2019). Oldfather, W.  A. (ed.), Epictetus: The Discourses, the Manual, and Fragments, vol. i (Cambridge, Mass., 1925). Ramelli, I., Hierocles the Stoic: Elements of Ethics, Fragments, and Excerpts [Hierocles] (Atlanta, 2009). Rawls, J., A Theory of Justice (Cambridge, Mass., 1971). Reynolds, L. D. (ed.), L. Annaei Senecae Ad Lucilium Epistulae morales, 2 vols. (Oxford, 1965). Sheffield, F. C. C., ‘Moral Motivation in Plato’s Republic? Philia and the Return to the Cave’, Oxford Studies in Ancient Philosophy, 59 (2021), 79–131. Singer, P., The Expanding Circle: Ethics, Evolution, and Moral Progress, 2nd edn (Princeton, 2011). Striker, G., Essays on Hellenistic Epistemology and Ethics [Essays] (Cambridge, 1996). Striker, G., ‘Following Nature: A Study of Stoic Ethics’, in ead., Essays, 221–80. Striker, G., ‘The Role of oikeiōsis in Stoic Ethics’, in ead., Essays, 281–97. Tönnies, F., Community and Civil Society, ed. by J.  Harris, trans. by J. Harris and M. Hollis (Cambridge, 2001). Veillard, C., ‘Hiéroclès, les devoirs envers la patrie et les parents’ [‘Hiéroclès’], in Gourinat (ed.), L’éthique, 105–43. Vogt, K. M., Law, Reason, and the Cosmic City: Political Philosophy in the Early Stoa (Oxford, 2008). Wachsmuth, C. and Hense, O. (eds.), Ioannis Stobaei Anthologium, 5 vols. (Berlin, 1884–1912). Winterbottom, M. (ed.), M. Tulli Ciceronis De Officiis (Oxford, 1994).

ARCHAIC EPISTEMOLOGY A Discussion of Jessica Moss, Plato’s Epistemology: Being and Seeming. Oxford University Press, Oxford, 2021. pp. x + 258. matthew evans

Does Plato still have anything worthwhile to tell us today about how we ought to conduct ourselves as thinkers? Those of us who have worked our way through the various twists and turns of master­ pieces like the Meno, the Republic, and the Theaetetus might be forgiven for assuming that the answer to this question is an un­equivo­cal yes. As Jessica Moss has shown in her engaging and provocative new book, however, that assumption turns out to be a surprisingly risky one. For, over the last century or so, our thinking about how we ought to conduct ourselves as thinkers—­our epis­tem­ ol­ogy, as we might put it—­has developed in directions that seem to depart from the Platonic tradition quite dramatically.

I Consider knowledge, for example. If you have taken or taught a standard epistemology class in an Anglo-­American philosophy department during the last fifty years, you will know that—­ according to the current consensus, anyway—­knowledge is cheap. You can get it by believing what you see when you look around you, or what you remember about rising from sleep this morning, or what your friend tells you about the weather on the other side of town. But if you are reading this paper, then you will also know that, according to Plato, epistēmē—the Greek term that is frequently

I would like to thank Jessica Moss and the editor, Victor Caston, for their generous comments on earlier drafts of this paper.

Matthew Evans, Archaic Epistemology: A Discussion of Jessica Moss, Plato’s Epistemology: Being and Seeming In: Oxford Studies in Ancient Philosophy, Volume LXII (Summer 2022). Edited by: Victor Caston, Oxford University Press. © Matthew Evans 2023. DOI: 10.1093/oso/9780192885180.003.0009

334

Matthew Evans

translated as ‘knowledge’—is not at all cheap. You cannot get it by simply believing what you see, what you remember, or what some­ one tells you; you must also, at a minimum, have a firm intellectual grasp of something else, something deeper. Indications of this idea can be found already in the Meno, where Socrates famously con­ trasts epistēmē with ‘correct’ (ὀρθή) or ‘true’ (ἀληθής) doxa—­a term that is standardly translated as ‘belief’ or ‘opinion’, but that I will leave untranslated here. After convincing Meno that true doxai are no less reliable than epistēmai as guides to action, he compares them to the mythical statues of Daedalus, which could move around on their own: Τῶν ἐκείνου ποιημάτων λελυμένον μὲν ἐκτῆσθαι οὐ πολλῆς τινος ἄξιόν ἐστι τιμῆς, ὥσπερ δραπέτην ἄνθρωπον . . . δεδεμένον δὲ πολλοῦ ἄξιον· . . . πρὸς τί οὖν δὴ λέγω ταῦτα; πρὸς τὰς δόξας τὰς ἀληθεῖς. καὶ γὰρ αἱ δόξαι αἱ ἀληθεῖς, ὅσον μὲν ἂν χρόνον παραμένωσιν, καλὸν τὸ χρῆμα καὶ πάντ’ ἀγαθὰ ἐργάζονται· πολὺν δὲ χρόνον οὐκ ἐθέλουσι παραμένειν, ἀλλὰ δραπετεύουσιν ἐκ τῆς ψυχῆς τοῦ ἀνθρώπου, ὥστε οὐ πολλοῦ ἄξιαί εἰσιν, ἕως ἄν τις αὐτὰς δήσῃ αἰτίας λογισμῷ. . . . ἐπειδὰν δὲ δεθῶσιν, πρῶτον μὲν ἐπιστῆμαι γίγνονται, ἔπειτα μόνιμοι· καὶ διὰ ταῦτα δὴ τιμιώτερον ἐπιστήμη ὀρθῆς δόξης ἐστίν, καὶ διαϕέρει δεσμῷ ἐπιστήμη ὀρθῆς δόξης. (97 e 2–98 a 8) Procuring a work of [Daedalus] when it has been released is not worth a lot, like [procuring] a runaway slave . . . But when it has been chained down, it is worth a lot. . . . What am I talking about? True doxai. For true doxai, for as long a time as they remain, are a fine thing, and all they achieve is good. But they are not willing to remain for a very long time, and they run away from a person’s soul, so that they are not worth a lot until one chains them down with an account of the cause. . . . Once they are chained down, first they become epistēmai, and then they remain. And that is why epistēmē is more prized than correct doxa, and epistēmē differs from correct doxa in being chained down.1

Though there are still some philosophers who will cite this passage as an early attempt to analyse knowledge as justified true belief, it is important to notice that if they are right, then the standard of justification Socrates proposes here is so demanding that no main­ stream epistemologist today would take it seriously. The term logismos, which I have translated as ‘account’, applies only to the products of at least moderately strenuous intellectual labour—­the 1  All translations are my own, unless otherwise noted. Texts from the Meno are from Burnet’s OCT.



Archaic Epistemology

335

sort of reasoning we do when we are trying to solve problems in arithmetic, geometry, or astronomy, for example; and I assume that in this context the term aitia, which I have translated as ‘cause’, is meant to designate whatever it is that makes the relevant true doxa true: the reason why it is true, or that because of which it is true. So, if Socrates is giving an analysis of knowledge here, then in his view you could not come to know that the moon is waning, say, just by believing what you see when you look at it, or what you remember from looking at it yesterday, or what your astronomy-­obsessed col­ league down the hall tells you about it; until you have worked out for yourself what makes it true that the moon is waning—­no small task, to be sure—­you do not know that it is. Now maybe I have been corrupted by my training or my culture, but I confess that I do not find this view even remotely credible. And I doubt that I am an outlier here. Indeed, I would be amazed if I could locate even one philosopher working in the trenches of contemporary epistemology who would disagree with me about this. For better or worse, it seems obvious to all of us, or nearly all of us, that knowledge is not as hard to get as this view says it is. Suppose, then, that the current consensus is correct and know­ ledge is cheap. Should we conclude from this that Socrates has nothing worthwhile to tell us, at Meno 97 d–98 e, about how we ought to conduct ourselves as thinkers? It depends. Clearly we, as thinkers, do want to acquire knowledge of the various things we care about. But is that all we want? If so, then I believe the answer to our question is yes: Socrates has nothing worthwhile to tell us here. Either he is working with an overly narrow conception of knowledge or else—­when he is talking about epistēmē—he is talking about an intellectual achievement that is far more demanding, and far less common, than mere knowledge. One is tempted to call it ‘comprehension’, ‘understanding’, ‘insight’, or ‘expertise’, but the name is less important than the idea: to acquire epistēmē is to figure out, by way of reasoning, what it is that makes your true doxa true. But once we see that Socrates is right about at least this much—­ that there is an intellectual achievement of this more demanding, less common kind—­I think we will also see that, at least sometimes, we want it as well. Merely knowing that the moon is waning will be fine for some purposes, but for others, understanding why will be our primary aim. And in that case it looks as if he does have some­ thing worthwhile to tell us about how we ought to conduct ourselves

336

Matthew Evans

as thinkers: in some contexts, anyway, we should not be satisfied with mere knowledge; we should seek epistēmē instead. This concession by itself would not require any especially r­ ad­ical revision of our contemporary epistemological framework, however. For we could maintain that epistēmē is just a special kind of know­ ledge, where knowledge is modelled in the usual way, as consisting in a true belief that meets some further standard—­one that is dif­ ferent from justification, as Gettier showed it must be, but is not so demanding as to guarantee the truth of any belief that meets it.2 Call it the ‘standard of warrant’.3 Then we could hold that epistēmē consists in a true doxa that not only meets the standard of warrant, as any knowledge-­constituting belief does, but also meets a more selective standard, insofar as it stems from, or is otherwise based upon, a firm intellectual grasp of the reason why it is true. As a result of all this, we might persuade ourselves that we can easily accommodate Plato’s suggestion here by agreeing with him that in some cases what we want is not just true doxa that meets the stand­ ard of warrant, but true doxa that meets this other, more selective standard as well.4

II One of the most important achievements of Moss’s new book, however, is to show that this rather optimistic take on Plato’s con­ tribution to contemporary epistemology (which I myself accepted at one point) does not hold up all that well under scrutiny. It found­ ers, in particular, on the argument that concludes Book 5 of the Republic—­a crucial bit of evidence concerning Plato’s ‘middle-­ period’ convictions in this area. The most salient upshot of this argument, given our focus here, is that his notion of doxa fails to match up with our notion of belief, 2  See E. L. Gettier, ‘Is Justified True Belief Knowledge?’, Analysis, 23 (1963), 121–3, and L. Zagzebski, ‘The Inescapability of Gettier Problems’, Philosophical Quarterly, 44 (1994), 65–73. 3 On the standard of warrant, see A.  Plantinga, Warrant and Proper Function (Oxford, 1993). 4  I say that this second standard is ‘more selective’ because I assume that fewer beliefs will meet it. But I do not mean to imply that it is ‘higher’ in the sense that all beliefs that do meet it will also meet the standard of warrant. This latter claim strikes me as dubious, in fact. See Section VI below.



Archaic Epistemology

337

as we put that notion to work in our conception of knowledge—­not only the everyday, ordinary kind, but also the special kind I tried to align with epistēmē a moment ago. That is because, in the argument at the end of Republic 5, Socrates makes it relatively clear that, in his view, epistēmē does not consist in a doxa of any sort, no matter what standard it might meet (476 e 6–480 a 13). His primary point there, in fact, seems to be that epistēmē and doxa are fundamentally different from each other. Moreover, and perhaps more im­port­ant­ly, this epistemological difference evidently derives from a metaphys­ ical one: whereas the ‘power’ (δύναμις) of epistēmē is exclusively ‘set over’ (ἐπί + dative) ‘what purely is’ (τὸ εἰλικρινῶς ὄν), the power of doxa is exclusively set over what is ‘intermediate between what purely is and what in no way is’ (μεταξύ . . . τοῦ εἰλικρινῶς ὄντος καὶ τοῦ αὖ μηδαμῇ ὄντος) (477 a 6– b 10).5 Though the meaning of these last two expressions is a bit murky, especially at the beginning of the argument, it becomes tolerably clear by the end that the sort of thing that is ‘intermediate between what purely is and what in no way is’ is the sort of thing that is beautiful in some respect only if it is ugly in another, large in some respect only if it is small in another, and so on (479 a–­b); by contrast, the sort of thing that ‘purely is’ is the sort of thing that is ‘always the same in every respect’ (ἀεὶ κατὰ ταὐτὰ ὡσαύτως, 479 e 6–7). The details here are complicated and controversial, to be sure, but I am inclined to agree with Moss that Plato takes things of the ‘intermediate’ sort to be perceptibles, such as plants, animals, and artefacts, and things of the ‘pure’ sort to be imperceptibles, such as numbers or forms. Following her, I will refer to this interpretation as the ‘distinct objects’ (DO) reading. I will also assume, for the remainder of this paper, that Moss is right: the textual evidence, both here in the Republic and throughout Plato’s other so-­called ‘middle-­period’ dialogues, makes the DO reading basically unavoidable.6 From this reading alone it seems to follow, more or less directly, that my earlier attempt to translate Plato’s views into our contem­ porary epistemological framework is bound to fail. For if he holds that epistēmē is exclusively set over imperceptibles and doxa is 5 All Republic quotations are taken from Slings’s OCT. 6  The canonical challenge to this assumption is to be found in the work of Gail Fine. See, in particular, G. Fine, ‘Knowledge and Belief in Republic V’, Archiv für Geschichte der Philosophie, 60 (1978), 121–39.

338

Matthew Evans

exclusively set over perceptibles, then—­as Moss emphasizes (82–5)— he cannot also hold that epistēmē is just a special sort of doxa. On the other hand, it does not follow that every such attempt to trans­ late his views into our framework is bound to fail. That will depend on how we answer a number of pressing exegetical questions, including this one: What is the relationship Socrates envisions between each power and ‘what it accomplishes’ (ὃ ἀπεργάζεται, 477 d 2)?

Here I am inclined to follow Moss once again and tentatively assume that the accomplishments of both the power of epistēmē and the power of doxa are to be understood as psychological attitudes of a fairly familiar sort—­those that carry a commitment to the truth of their (broadly propositional) contents.7 I will call such attitudes judgements. On this assumption, then, any given accomplishment of either your power of doxa or your power of epistēmē is a particu­ lar judgement of yours: that attacking Sparta is wise, say, or that the smallest even number is prime. Thus we will need to draw a distinction, for each thinker, between her doxa-judgements (or D-­judgements) on the one hand, and her epistēmē-judgements (or E-­judgements) on the other. After we have reflected on this distinction for a while, we might start to entertain the possibility that although these two powers are exclusively set over different things, their accomplishments—­the judgements that result from their exercise—­are not. On a view of this sort, a particular D-­judgement might be set over an im­per­ cept­ible thing such as a form, and a particular E-­judgement might be set over a perceptible one such as a ship. This would at least allow Plato to hold that D-­judgements and E-­judgements might be set over some of the same things.8 As Moss points out, however, Plato only rarely deploys the power/accomplishment distinction, 7  See Moss’s remarks at 79–85, 138, 206, and 235. I use the term ‘broadly’ here because the contents of these attitudes need not be understood as propositions in the technical sense associated with the work of Frege and Russell. They need only be the sort of things we are talking about when we talk about what Pericles told the crowd, or what Xenophon believed about the campaign, or what Socrates said about the fear of death. It is uncontroversial that there are such things, even if their nature remains a source of philosophical puzzlement. 8  For an admirably clear attempt to pursue this strategy, see N. D. Smith, ‘Plato on Knowledge as a Power’ [‘Knowledge’], Journal of the History of Philosophy, 38 (2000), 145–68.



Archaic Epistemology

339

and never to support a view of this kind (80); when he does deploy it, at 477 c–478 a, his point seems to be that E-­judgements (the accomplishments of epistēmē) are guaranteed to be true, whereas D-­judgements (the accomplishments of doxa) are not. That is why I think it makes better sense to assume, with Moss, that the accom­ plishments of these powers, no less than the powers themselves, are exclusively set over different things. What does Plato mean, though, when he has Socrates say that a given judgement, or a given power of judgement, is ‘set over’ a given thing, or a given kind of thing? The most obvious answer to this question is the one that Moss herself chooses to defend: for a given judgement to be ‘set over’ a given thing is for that judgement to be about that thing (81).9 On this interpretation, Plato holds that a certain judgement is ‘set over’ Diotima, for example, just in case it is about Diotima—­that Diotima is its topic or its subject matter. In what follows I will call this the topic-­based DO reading. The alter­ native answer, which has its defenders as well, would be something like this: for a given judgement to be ‘set over’ a given thing is for that judgement to be grounded upon an acquaintance with that thing.10 On this interpretation, Plato holds that a certain judgement of yours is ‘set over’ Diotima just in case that judgement is grounded upon your acquaintance with Diotima herself—­and not merely upon your hearing some bit of testimony about her, seeing a por­ trait of her, imagining her, or whatever. Since it is possible for a judgement of yours to be about a given thing without being grounded upon an acquaintance with that thing, this interpretation is clearly distinct from the first one. I will call it the acquaintance-­based DO 9  Later on in the book, she winds up having to weaken this view slightly, in order to handle a thorny textual problem that arises at the end of Republic 6 (see Section IV below). She also emphasizes, in her discussion of Plato’s concept of doxa, that the notion of ‘aboutness’ she is working with here is a bit narrower than the one we are sometimes willing to use in our characterizations of other people’s judgements (193–4). While we might concede, for example, that Aristotle formed judgements about H2O, given that he formed judgements about water, the notion of ‘aboutness’ Moss has in mind here does not allow for this: the things our judgements are ‘about’, in the sense that is relevant to her understanding of Plato’s view, cannot be hidden from us in this way. So I will grant, for the purposes of this paper, that her preferred sense of ‘about’ is the correct one. 10  This is meant only as a rough sketch of the general sort of view advanced by interpreters such as J.  Szaif, ‘Doxa and Epistēmē as Modes of Acquaintance in Republic V ’, Études Platoniciennes, 4 (2007), 253–72, and Smith, ‘Knowledge’. As will become clear later, in Section VI, it stands in need of further refinement.

340

Matthew Evans

reading. Assuming that we must adopt exactly one of these inter­ pretations, then, which of them should we adopt, and why?

III One immediate concern about the topic-­based DO reading is that it will not allow us to make decent sense of what Socrates says in the Meno passage I quoted above about the relationship between epistēmē and true doxa. In that passage, remember, a true doxa of yours is said to ‘become’ epistēmē when you ‘chain it down by figur­ ing out the cause’ of its truth. These remarks at least suggest that, according to Plato, every E-­judgement is just a certain sort of D-­judgement—­one that is both true and ‘chained down’ by the relevant sort of explanatory reasoning. This would spell disaster for every DO reading, given our assumption that D-­judgements and E-­judgements are exclusively set over different things, but—­as Moss and others have emphasized—­Socrates’ remarks there can be given an alternative interpretation.11 Here is how Katja Vogt glosses it: once the true [doxa] has become stable (which might involve that the rele­ vant reasoning is in place), a piece of [epistēmē] replaces what was earlier a true [doxa]. This stable attitude, so the proposal goes, is genuinely differ­ ent. That is, it is not the case that an attitude gets changed in one of its properties, becoming stable. Instead, through stabilization a new attitude— [epistēmē]—is generated. (Belief, 14)

Could this proposal save the DO reading—­and, more specifically, the topic-­based version of it? Moss seems to think so: concerning the familiar case of the double square at Meno 82 b–87 c, for ex­ample, she suggests that (according to this proposal) the epistēmē that replaces one’s true doxa ‘will not be about particular percept­ ible squares (although it will of course inform [one’s D-]judge­ ments about them). Instead it will be about the form of the square itself’ (216). Therefore, she thinks, ‘one ceases to have doxa on this topic, for the resulting epistēmē is not a kind of doxa’ (216 n. 15).

11  See, for example, L. Gerson, Ancient Epistemology (Cambridge, 2009), 28–30, and K. Vogt, Belief and Truth: A Skeptic Reading of Plato [Belief] (Oxford, 2012), 13–14.



Archaic Epistemology

341

But Vogt’s proposal, as Moss understands it anyway, strikes me as a bit of a stretch. According to the ‘statue of Daedalus’ meta­ phor, remember, your true doxa is not removed or eliminated when you figure out what makes it true; on the contrary, it ‘becomes’ epistēmē when you do this, precisely because you have thereby pre­ vented it from being removed or eliminated. If you were trying to ‘replace’ a particular statue of Daedalus with something ‘genuinely different’, in other words, your first move would presumably not be to chain that statue down exactly where it is. Maybe this is the wrong way to understand Vogt’s proposal, however. Maybe what she means is not that in these cases your true doxa is simply replaced, root and branch, by something entirely different, but rather that one and the same truth-­apt content—­that Nicias is cowardly, say, or that piety is a part of justice—­goes from being the (true) content of a D-­judgement of yours to being the (true) content of an E-­judgement of yours: in short, it is the atti­ tude that changes, not the content. If this is, in fact, Vogt’s proposal, then I think she is right to claim that it remains a live exegetical option for us. More importantly, given our purposes here, it also allows us to retain the DO reading. For if we adopt it, then we could insist on Plato’s behalf that although no E-­judgement is itself a D-­judgement, the content of every E-­judgement is, or could be, the content of some D-­judgement. Indeed, the familiar reference to ‘figuring out the cause’ at Meno 97 d–98 a provides what looks like a relatively straightforward rationale for just such a reading: whereas every E-­judgement is grounded upon a special sort of intellectual grasp of what makes its true content true, every D-­judgement, without exception, is not. So Vogt’s proposal (assuming that this is the correct understand­ ing of it) might well save the DO reading. And yet, at the same time, it would clearly doom the topic-­based version of the DO read­ ing. For recall that, on this reading, Plato holds that it is impossible for anyone’s E-­judgement to be about the very same thing that someone else’s D-­judgement is about. If Vogt’s proposal is correct, however, and Plato grants that the content of my D-­judgement could be the content of your E-­judgement, then this reading is plainly mistaken: if the content of our two judgements is that the side of a square twice the size of this square is the diagonal of this square, for example, then neither of our judgements is any less about this square than the other one is. Once we grant Vogt’s proposal,

342

Matthew Evans

then, we can save the DO reading—­but only at the cost of aban­ doning the topic-­based version of it.

IV As Moss stresses at various points, however, her primary aim in the book is not to show that the topic-­based DO reading can be fruit­ fully applied across the board to all the dialogues in which Socrates attempts to distinguish epistēmē from true doxa (3, 207). Her pri­ mary aim, rather, is to show that the topic-­based DO reading can be fruitfully applied to what she calls the ‘Two Worlds’ dialogues—­ those that demonstrate a clear commitment on Plato’s part to the ontological distinction between perceptibles and forms. The Meno is not among those dialogues, in her view, and I am inclined to agree. So even if the topic-­based DO reading cannot do justice to Meno 97 a–98 d, as I have argued it cannot, Moss’s core project remains untouched by this. The deeper and more pressing ques­ tion for her at this point is whether there are any passages in dia­ logues like the Phaedo, the Republic, and the Timaeus that will generate similarly intractable problems. And I am afraid that there are some. The first one that jumps to my mind is from the Phaedo, where Socrates is recounting to his gathered friends how and why he originally developed his theory of forms: ἔρχομαι γὰρ δὴ ἐπιχειρῶν σοι ἐπιδείξασθαι τῆς αἰτίας τὸ εἶδος ὃ πεπραγμάτευμαι, . . .  ὑποθέμενος εἶναί τι καλὸν αὐτὸ καθ’ αὑτὸ καὶ ἀγαθὸν καὶ μέγα καὶ τἆλλα πάντα· . . . ϕαίνεται γάρ μοι, εἴ τί ἐστιν ἄλλο καλὸν πλὴν αὐτὸ τὸ καλόν, οὐδὲ δι’ ἓν ἄλλο καλὸν εἶναι ἢ διότι μετέχει ἐκείνου τοῦ καλοῦ· καὶ πάντα δὴ οὕτως λέγω. (100 b 3–c 6) I’m going to try to show you the kind of cause I have concerned myself with, . . . hypothesizing that there is a beautiful, itself by itself, and a good, and a large, and all the rest. . . . for it appears to me that if there is anything beautiful other than the beautiful itself, it is beautiful for no other reason than that it participates in that beautiful [itself]. And I say the same with all of them.12

12  The text is from Strachan’s OCT.



Archaic Epistemology

343

This last phrase (‘and I say the same with all of them’) seems to indicate that, as far as he is concerned, the basic structure of his theory will hold not only for the form of the beautiful—­which I will refer to as ‘the Beautiful’—but also for all the other forms: if anything is large besides the Large, for example, then (in some sense) it is large because (and only because) it participates in the Large; and so on with the rest of the forms. Suppose, then, that the content of one of Socrates’ judgements is that the Acropolis is tall. Given what he says here, we can infer that the content of another one of his judgements is that the Acropolis participates in the Tall. Now is this second judgement a D-­judgement or an E-­judgement? Since it is about both a perceptible (the Acropolis) and a form (the Tall), defenders of the topic-­based DO reading cannot classify it only as a D-­judgement or only as an E-­judgement. Nor can they classify it as both; for, as we have seen, every DO reading entails that there can be no D-­judgement that is also an E-­judgement. So the topic-­based DO reading simply cannot handle judgements of the sort Socrates avows and recommends to others in this critically important part of the Phaedo. Moss seems to be aware of the problem here, since she raises it (or something close to it) right after acknowledging the robust explanatory role that forms are said to play in the Phaedo passage above, and in similar passages elsewhere: Does [Phaedo 100 b–­c, for example] entail that there is epistēmē of percep­ tibles after all, insofar as they are explained by Forms? It does not. We can reconcile an account of epistēmē as explanation-­grasping with [the topic-­ based DO reading, as long as we maintain that] Plato thinks that genuine explanatory connections only hold between Forms themselves; the per­ ceptible realm is too messy to admit of genuine explanation. (116)

But this response will not work. Plato clearly does think that we can arrive at certain form-­centred explanations for why perceptibles are the way they are—­why beautiful things are beautiful, large things are large, and so on; that is a crucial point of his argument in the Phaedo, as we have seen. In this limited sense, anyway, he clearly does think that there are ‘genuine explanatory connections’ between forms and perceptibles. Maybe there is some other sense in which he does not think this, as Moss suggests, but that is beside the point: once he allows that we can make the sort of explanatory judgements Socrates explicitly recommends to his friends in the

344

Matthew Evans

Phaedo—­that snow is cold because it participates in the Cold, for example—­the topic-­based DO reading runs into trouble. Nor is the Phaedo the only ‘Two Worlds’ dialogue where this problem arises. Even in the middle books of the Republic, where we might expect the topic-­based DO reading to be most at home, we find many passages that seem to tell against it. Here is one, from Book 7, where Socrates imagines what he would say to all the recently trained philosophers in order to justify the requirement that they give up their life of contemplation, return to the cave, and take up a life of worldly politics: ἄμεινόν τε καὶ τελεώτερον ἐκείνων πεπαιδευμένους καὶ μᾶλλον δυνατοὺς ἀμϕοτέρων μετέχειν. καταβατέον οὖν ἐν μέρει ἑκάστῳ εἰς τὴν τῶν ἄλλων συνοίκησιν καὶ συνεθιστέον τὰ σκοτεινὰ θεάσασθαι· συνεθιζόμενοι γὰρ μυρίῳ βέλτιον ὄψεσθε τῶν ἐκεῖ καὶ γνώσεσθε ἕκαστα τὰ εἴδωλα ἅττα ἐστὶ καὶ ὧν, διὰ τὸ τἀληθῆ ἑωρακέναι καλῶν τε καὶ δικαίων καὶ ἀγαθῶν πέρι. (520 b 7–c 5) You are better and more fully educated than the rest, and are more able to share in both [types of life]. So each of you in turn must go down to the common dwelling place of the others and get used to seeing in the dark. For when you are used to it, you’ll see far better than they do, and because you have seen the truth about what is beautiful, just, and good, you will know each of the images for what it is, and [you will know] those things of which they are the images.

In this last sentence, the verb I have translated as ‘know’ is gignōskein, which Socrates uses repeatedly throughout the argu­ ment at the end of Book 5 (476 e–480 a); and as he uses it there, it clearly stands to epistēmē as doxazein does to doxa. So here, in Book 7, Socrates seems to be saying, or at the very least implying, that it is possible for a philosopher ruler’s E-­judgements to be about not only perceptible things (‘the images’) but also imperceptible ones (‘those things of which [the images] are the images’)—which is plainly inconsistent with the topic-­based DO reading. Moss tries to solve this problem by pointing out, correctly, that in this context gignōskein need not mean what it means in the argu­ ment at the end of Book 5. Here in Book 7 it could mean ‘recognize’ or ‘distinguish’, for example, neither of which must be understood as designating an E-­judgement (125). But again, I worry that this reply is beside the point. Even if gignōskein has a relatively deflated meaning in this context, Socrates assigns both perceptibles and imperceptibles as its conjoined direct objects at 520 c 4. This indicates



Archaic Epistemology

345

that the philosopher ruler does (or at least could) make a judge­ ment of some kind—­whatever kind that may be—­to the effect that this perceptible thing is an image of that im­per­cept­ible one. So the very same problem will arise once again here.

V An even more troublesome passage turns up near the end of Book 6, where Glaucon and Adeimantus press Socrates to tell them what he thinks the Good is, given his recent insistence that it is ‘the most important thing [for the philosopher ruler] to study’ (μέγιστον μάθημα, 505 a 2). He demurs, claiming (A) that he does not have the relevant epistēmē, and (B) that no doxa, even if true, is worth much of anything in comparison with epistēmē (506 b–­d). The clear implication of (A) and (B) together, given the conversational con­ text, is that he does have some doxai about the Good—­doxai that, for some reason or other, he is unwilling to share with his conversa­ tion partners. This implication becomes explicit a few lines later, when he begs off with the following remark: αὐτὸ μὲν τί ποτ’ ἐστὶ τἀγαθὸν ἐάσωμεν τὸ νῦν εἶναι· πλέον γάρ μοι ϕαίνεται ἢ κατὰ τὴν παροῦσαν ὁρμὴν ἐϕικέσθαι τοῦ γε δοκοῦντος ἐμοὶ τὰ νῦν. ὃς δὲ ἔκγονός τε τοῦ ἀγαθοῦ ϕαίνεται καὶ ὁμοιότατος ἐκείνῳ, λέγειν ἐθέλω. (506 d 7–e 3) let us set aside for now what the good itself is. For it appears to me that get­ ting to my own doxai about this is too much to ask of our present effort. But I am willing to say what is apparently a child of the good and most like it.

Here Socrates makes it quite clear that, by his own lights, he has some doxai about the Good. But the topic-­based DO reading entails that, according to Plato, there can be no D-­judgements about any form, including the Good. So this passage provides still further evidence against that reading. Moss’s reaction to this evidence is a bit flat-­footed: ‘I conclude that in this passage Plato is loosening his terminology’ (191). Since this passage follows so closely upon Socrates’ emphatic reassertion of the epistēmē/doxa contrast at  506  c–­d, however, I do not find this conclusion especially ­convincing. But suppose she is right, and Socrates does not really mean what he seems to mean here. Still, there is another, more serious prob­ lem lurking in the background of this one. That is because, directly

346

Matthew Evans

after this passage, and all the way to 509 c, Socrates speaks to Glaucon and Adeimantus at length about the Good—­proposing and developing an elaborate comparison between it and the sun. This entire stretch of text poses what I take to be a major challenge to any version of the DO reading, not just the one Moss defends. It is perhaps best put in the form of a paradox about how Socrates would understand what he himself is doing in this stretch of text: (1) He is expressing some judgements about the Good at 507 a–509 c. (2) If he is expressing any judgements about anything any­ where, then they are either D-­judgements or E-­judgements. (3) The judgements he is expressing about the Good at 507 a– 509 c are neither D-­judgements nor E-­judgements. These three claims are jointly inconsistent, so we need to reject at least one of them. But which one should it be? My own sense is that, of the three, the only one we cannot rea­ sonably reject is (1). After all, it is (or ought to be) uncontroversial that one of the judgements Socrates expresses at 507 a–509 c is that the sun is a child of the Good. And Moss, if I am interpreting her correctly, would not disagree.13 She wants to pursue a different strategy, which is to accept both (1) and (3), and therefore to reject (2). Here is what she says: Commentators often refer to [507 a–509 c] as expressing Socrates’ doxai about the Good . . .  But [the passage quoted above, 506 d 7–e 3] shows that this is wrong: Socrates explicitly refuses to state any doxai about the Good. How then should we characterize his thoughts? They are clearly thoughts about the Good. They are not epistēmē, which he denies having, nor are they doxai about the Good, which he refuses to offer. They must then be something in between. This points us toward dianoia, and indeed they fit the description [of dianoia] very well. (187, emphasis original)

13  At one point she suggests that there are some interpreters who would want to reject something like (1), given their broader readings of Plato’s epistemology and philosophy of mind (182). The two she cites in this connection are Gerson, Ancient Epistemology, 40–1, and R. Woolf, ‘Plato and the Norms of Thought’, Mind, 122 (2013), 171–216. She says that she finds their ‘general line of argument compel­ ling’—which might be thought to indicate that she would want to reject (1) as well. But I believe it is clear from what she goes on to say a few pages later, in the passage I am going to quote in a moment, that she would not want to reject (1).



Archaic Epistemology

347

It is not clear to me why Moss believes that if these judgements are neither D-­judgements nor E-­judgements, then they must be ‘some­ thing in between’. Nor is it clear to me that there could be something in between, given the argument at the end of Book 5. For in that argument Socrates repeatedly (and crucially) assumes that if a given judgement is not an accomplishment of either the power of doxa or the power of epistēmē, then it is an accomplishment of the power of agnōsia (‘ignorance’)—which his later judgements about the Good, by his own lights, plainly are not.14 In other words, he seems to leave no conceptual room for a viable third alternative, despite what Moss suggests here. I will set this difficulty aside, however, and simply grant that Plato would allow for such a third alternative. What could it be? As the last sentence of the excerpt I quoted above indicates, Moss thinks it ought to be dianoia—­the psychological attitude Socrates assigns to the third subsection of the divided line, between pistis (‘trust’) and noēsis (‘comprehension’).15 This proposal faces a num­ ber of difficulties, and Moss is conscientious about flagging and confronting each of them in turn (181–90). But there is one in par­ ticular that I take to be especially severe, and it is the last one she discusses. Simply put, the problem is this: according to the story of the divided line, anything that dianoia is ‘set over’ is distinct from anything that epistēmē (or noēsis) is ‘set over’. So if epistēmē is set over the Good, as Moss maintains, then dianoia is not; at best it is set over what she calls an ‘abstract image’ of the Good (189). On the topic-­based DO reading, however, a given judgement is not about a given thing unless it is set over that thing. Therefore, if this reading is correct, then there are no dianoia-judgements about the Good after all—­and Moss’s proposal collapses. She is aware of this problem, but thinks she can handle it without excessive strain: the distinction between [what dianoia is set over and what epistēmē is set over] seems to conflict with [the reading] I have defended in this book, for it opens a gap between what a cognitive kind is [set] over and what it is about. Is this a problem for my accounts of epistēmē and doxa? I think it is not, for two reasons. If Socrates’ thoughts about the Good are really about the Good, despite being ‘[set] over’ an abstract image of the Good, this 14  Consider his reasoning at 477 a–­b and 478 d–­e, for example. 15 On the divided line, see 509 d–511 e; and for some evidence that Socrates ul­tim­ate­ly identifies noēsis—­as he discusses it here—­with epistēmē, see 533 d–534 a.

348

Matthew Evans

shows what is distinctive about dianoia. Socrates recognizes that he is working with mere images . . . This ability to recognize an image as an image allows for a gap between the different kinds of objects: between what cognition is [set] over, and what it is about . . . Such a gap is precluded in doxa, whose signature failing is the failure to recognize that its objects are only images . . . It is also precluded in epistēmē, which fully grasps its objects and has no need for attention to anything else. (190)

Evidently Moss is willing to concede that, at least in the case of dianoia, the topic-­based DO reading will not work: for dianoiajudgements can be about things that they are not set over, includ­ ing (most importantly) the Good. Yet at the same time she is clearly not willing to concede that D-­judgements or E-­judgements are like dianoia-judgements in this respect—­that they too could be about things that they are not set over. But why not? If there is  nothing in the nature of either the ‘set over’ relation or the (now admittedly distinct) ‘about’ relation that would explain why a D-­judgement could not be about the Beautiful or why an E-­judgement could not be about the Piraeus, then what would explain this? Her answer, apparently, is that there is a critical difference between dianoia, on the one hand, and doxa and epistēmē, on the other. What makes dianoia special, she thinks, is the role Socrates reserves for it in his account of geometrical reasoning at 510 d–511 a: when a geometer contemplates the problem of the double square, for example, he conjures up in his mathematical mind various images of the Square and the Diagonal, and as a result he makes certain judgements—­dianoia-judgements—­that are set over those images. As long as he recognizes them as images, however, his dianoiajudgements are not about them; they are about the things of which they are images: the Square and the Diagonal (182–8). It is pre­ cisely his ability to recognize these images as images, she believes, that ‘allows for a gap’ between what his judgements are set over and what they are about. But the same explanation will not work in the case of either D-­judgements or E-­judgements, she thinks, and for different reasons in each case: there can be no gap between what D-­judgements are set over and what they are about, in her view, because D-­judgements always involve a ‘failure to recognize that [the things they are set over] are only images’; and there can be no gap between what E-­judgements are set over and what they are about because E-­judgements ‘have no need for attention to’ the



Archaic Epistemology

349

images of the things they are set over in addition to those things themselves.16 This is why D-­judgements and E-­judgements, but not dianoia-judgements, are guaranteed to be set over whatever they are about—­despite the fact that being set over something turns out to be different, in the end, from being about that thing. I find this answer unsatisfying, however. Recall that, according to Moss’s (newly restricted version of the) topic-­based DO read­ ing, the only kind of judgements anyone could ever make about perceptible things, such as Sparta, are D-­judgements. But if this is right, then we cannot accept her answer here unless we are ready to agree that, according to Plato, no one, not even the philosopher ruler, could ever make any judgement about Sparta—­the judge­ ment that it is courageous, say—­unless he thereby exhibits a ‘failure to recognize’ that Sparta is (at best) only an ‘image’ of the Courageous. And I do not see how we can reasonably agree to this.17 Nor do I see how we can safely assume that, if the philosopher-­ruler really has ‘no need’ to make any E-­judgements about things like Sparta—­a claim that I myself remain sceptical of—­then it is impossible for him to do so. On balance, then, I doubt that Moss’s solution to the paradox above is successful. Her solution, remember, is to accept (1) and (3), and therefore to reject (2). But, as we have discovered, (2) turns out to be harder to reject than we might have supposed at first. So, since we cannot reasonably reject (1), maybe we—­the defenders of the DO reading—­ could try to reject (3) instead. Here, again, is (3): The judgements Socrates expresses about the Good at 507 a–509 c are neither D-­judgements nor E-­judgements.

Moss holds that (3) is irresistible, given his remarks at 506 b–­e. But I am not so sure about this. When we look again carefully at the language he uses in that passage, I think we find some room to 16  It is worth emphasizing that the claims about doxa and epistēmē that appear in the ‘because’ clauses here do not suddenly show up, for the first time, in the para­ graph I quoted above from page 190; on the contrary, Moss supports them with a wide variety of textual evidence that I cannot address properly within the confines of this paper. See, in particular, her detailed discussion of each in chapters 5 and 8. 17  Because of space limitations, I cannot engage as closely as I would like with Moss’s comprehensive analysis of Plato’s concept of doxa, which takes up the chap­ ters 6–8 of her book. I will only point out that if her analysis is correct, then—­as I  have just argued—­the consequences for Plato’s overall view are dire, given her topic-­based DO reading.

350

Matthew Evans

manoeuvre. Notice, first, that the question Adeimantus raises about the Good right at the beginning of that passage, at 506 b 2–4, is a very specific one: he wants Socrates to tell him ‘whether . . . the Good is  epistēmē, or pleasure, or something else besides these’ (πότερον ἐπιστήμην τὸ ἀγαθόν . . . εἶναι ἢ ἡδονήν, ἢ ἄλλο τι παρὰ ταῦτα). So, when Socrates disavows epistēmē at 506 c–­d and withholds doxa at 506 d–­e, the E-­judgements and D-­judgements he is refer­ ring to there might be restricted, in their subject matter, to the spe­ cific question Adeimantus has just raised. This possibility becomes even more salient at 506 d 5–e 2, the passage I quoted earlier, because what Socrates actually says there is that he will not try to tell Glaucon and Adeimantus ‘what the good itself is’ (τί ποτ’ ἐστὶ τἀγαθόν), and that he will try to tell them instead ‘what is appar­ ently a child of the good and most like it’ (ὃς δὲ ἔκγονός τε τοῦ ἀγαθοῦ ϕαίνεται καὶ ὁμοιότατος ἐκείνῳ). This suggests that, in his view, there is a critical distinction to be observed between two questions, both of which are plainly about the Good: the first is what the Good is, and the second is what is a child of the Good and most like it. Once we draw this distinction ourselves, we are in an excellent position to challenge (3). For we now have a textually well-­founded reason to think that the judgements Socrates expresses at 507 a–509 c are about what is a child of the Good and most like it, whereas both the E-­judgements he disavows at 506 c–­d and the D-­judgements he withholds at 506 d–­e, are about what the Good is. Thus, the ‘dis­ avowal and withholding’ evidence of 506 c–­e no longer supports (3), and we are free to reject it. Because this solution allows us to retain both (2) and the DO reading, I think it is plainly preferable to the one Moss goes for. But is it a solution that she too might adopt, without compromis­ ing her (restricted) topic-­based version of the DO reading? Here, again, I am not optimistic. For recall that one of the judgements Socrates expresses at 507 a–509 c is that the sun is a child of the Good—­which is yet another classic example of a judgement that is both about a perceptible and about a form. On the (restricted) topic-­based DO reading, however, any such judgement must be neither a D-­judgement nor an E-­judgement, for the reasons I dis­ cussed in Section IV. So her reading will eventually require her to reject (2) in any case, and my alternative solution will be of no help to her.



Archaic Epistemology

351

VI In light of the many and various problems I have raised for the topic-­based version of the DO reading, I think we should seriously consider the possibility of abandoning it in favour of the alterna­ tive version I introduced at the end of Section II: the acquaintance-­ based version. On this reading, remember, what it is for a given judgement to be ‘set over’ a given thing is for that judgement to be grounded upon an acquaintance with that thing. If this reading allows us to avoid the various problems I have raised for Moss’s topic-­based version and does not generate any new and similarly severe problems of its own, then I think we should adopt this ­version instead—­assuming, of course, that we must adopt some version of the DO reading, which Moss has persuaded me we must. Consider first the case from the Meno. Here we arrived at the con­ clusion that, in order for any DO reading to handle this passage, it would need to allow for one and the same truth-­apt content to go from being the content of a D-­judgement of yours (and not the con­ tent of an E-­judgement of yours) at one time to being the content of an E-­judgement of yours (and not the content of a D-­judgement of yours) at another time. As I argued in Section I, the topic-­based ver­ sion of the DO reading cannot allow for this. But I think the acquaintance-­based version clearly can. Suppose, for example, that the content of one of Lampito’s judgements is this: The diagonal of any given square is the side of another square that is twice the size of the first one.

Now suppose that she made this judgement just a moment ago by looking at a square figure on a blackboard and then estimating the relative size of another square figure with its side on the diagonal of that one. Since her judgement is grounded upon her acquaint­ ance with this perceptible square and not upon her acquaintance with what it is to be a square, the acquaintance-­based DO reading will hold that it is a D-­judgement, not an E-­judgement. But now suppose that, at some later point in her education, Lampito acquires a firm intellectual grasp of what it is to be a square and then reconsiders the issue. She eventually makes another judge­ ment with the same content as her earlier one, but this time by simply contemplating what it is to be a square. Since this new judgement (unlike the old one) is grounded upon her acquaintance

352

Matthew Evans

with the nature of a square, and not upon her acquaintance with any perceptible square, the acquaintance-­based DO reading holds that it is an E-­judgement, not a D-­judgement. Problem solved. Can a similar story be told about the sort of theoretical judge­ ments Socrates commends to his friends, and to us, in the Phaedo? Suppose that Xanthippe combines her visual experience of the landscape with her husband’s testimony to arrive at the judgement that Mount Olympus participates in the Beautiful. In that case, the acquaintance-­based DO reading would hold that hers is a D-­judgement, not an E-­judgement, even though it is about both a perceptible (Mount Olympus) and a form (the Beautiful). That is because it is clearly grounded upon her acquaintance with Olympus and with her husband, and not upon her acquaintance with the Beautiful. But now suppose that Diotima, who has a firm intellec­ tual grasp of the Beautiful, were to make a judgement with the same content. Here I think it is no longer obvious what we should say. Should we insist that her judgement is grounded upon her acquaintance with the form, or should we grant that it is (or might be) grounded upon her acquaintance with the mountain? What we cannot say, of course, is that it is grounded upon both: for in that case the acquaintance-­based DO reading would classify this judge­ ment as both a D-­judgement and an E-­judgement—­which is incon­ sistent with every DO reading, as we have seen.18 Better to stipulate, on Plato’s behalf, that a given judgement of yours is grounded upon your acquaintance with a given thing if, and only if, your acquaintance with that thing is the primary causal factor in the mak­ ing of your judgement. Assuming that there can be at most one primary causal factor in the making of any given judgement, this would rule out the possibility that Diotima’s judgement is grounded both upon her acquaintance with Olympus and upon her acquaint­ ance with the Beautiful—­and thus would allow us to sidestep the problem. If Diotima’s acquaintance with the (imperceptible) form is the primary causal factor in the making of her judgement, then it will be an E-­judgement, not a D-­judgement; and if her acquaint­ ance with the (perceptible) mountain is the primary causal factor,

18  Recall that, according to every DO reading, there can be no D-­judgement that is also an E-­judgement, because D-­judgements and E-­judgements are set over mutually exclusive sets of things. See Section II above.



Archaic Epistemology

353

then it will be a D-­judgement and not an E-­judgement. But there is no possibility that it will be both.19 What we seem to have discovered, then, is that the acquaintance-­ based version of the DO reading does not generate the same sort of intractable problems that Moss’s topic-­based version does. So why not adopt this alternative version instead? One possible reason, I suppose, would be that it makes a bad fit with the rest of our text­ ual evidence. But I do not believe that it does. In an early part of her book, Moss helpfully works through all the various linguistic devices Plato uses to differentiate the ‘objects’ of D-­judgements from the ‘objects’ of E-­judgements, and, as far as I can tell, all (or nearly all) of these devices could be used to pick out objects of dependence (that is, the things our acquaintance with which grounds our judgements) just as easily as they could be used to pick out objects of reference (that is, the things our judgements are about) (26–42).20 Consider, for example, the most important of these devices, which (as we have seen) is the preposition epi (‘set over’). When epi is used with the dative, as it is here, it could be used to characterize something as a mere object of reference; but it could also be used to characterize something as an object of dependence, as in the familiar expression eph’ hēmin (‘up to us’).21 On the latter interpretation, Plato’s thought here would presumably be that whereas our D-­judgements are grounded upon our acquaintance with perceptibles, our E-­ judgements are grounded upon our acquaintance with forms. So if we adopt this interpretation, we should immediately infer that the acquaintance-­based DO reading makes an even better fit with our textual evidence than the topic-­ based one does. It is when I reach this stage of reflection on the issue that I find myself most at a loss as to why Moss chooses to defend the topic-­ based version of the DO reading instead. It is not as though she is  unaware of the alternative. In fact, she raises the question of 19  See Smith, ‘Knowledge’, 153–63, for a broadly similar solution to a broadly similar problem. 20  In this section of the book, Moss is making what she calls her ‘prima facie case’ in favour of the DO reading. But the case she makes here, if I understand it cor­ rectly, does not tell in favour of any particular version of the DO reading, and so does not help us decide between the topic-­based version of it and the acquaintance-­ based one. 21  See LSJ, s.v. ἐπί B. I. 1. g.

354

Matthew Evans

switching sides—­or at least staying neutral—­very early on, in the first chapter, and her answer is worth quoting here in full: Why not . . . defend or at least leave open [the acquaintance-­based DO] reading, insisting only that epistēmē and doxa each bear some special rela­ tion to an ontological correlate, and proceed to develop an account of Plato’s epistemology from there? I have a number of reasons for experi­ menting with [the topic-­based] DO reading. First, I believe, and hope to convince the reader, that this is very much the most obvious way to read Plato, and also the way that he was read for about two millennia. We sim­ ply never see him nor his older interpreters working to draw the kind of distinctions on which [acquaintance-­based] readings rely, nor showing signs of noticing problems with the [topic-­based] view. Second, . . . I will argue . . . that both Plato’s characterization of how powers in general are related to objects, and also his characterization of how cognitive kinds share significant properties with their objects, point us toward an objects-­ based epistemology on which the objects of each cognitive kind are sharply distinct. Third, I will try to show in the remainder of the book that read­ ing Plato this way yields compelling accounts of both epistēmē and doxa, ones not only consistent with the texts but also with Plato’s wider philosoph­ ical projects. In short, I think the only reason to prefer [an acquaintance-­ based DO] reading to a [topic-­based] one is the thought that the latter cannot be made to work—­but I hope to show that it can. (26)

The first reason she gives here, as I understand it, has two compo­ nents: (1) that the topic-­based DO reading is both more obvious and more historically popular than the acquaintance-­based one; and (2) that Plato never explicitly draws the distinction I have been using to differentiate these two readings from each other, that is, the dis­ tinction between objects of reference, on the one hand, and objects of dependence, on the other.22 But I do not believe that either of these claims is fully responsive to the question she has just asked. After all, I myself am inclined to accept both (1) and (2), and yet my belief in the superiority of the acquaintance-­based DO reading remains unshaken. If I were unable to examine the rele­vant pas­ sages on my own, then I might be ready to give (1) greater weight; in my experience, however, the most obvious readings of ancient texts usually turn out to be wrong, and the vast majority of inter­ preters who have gone before us are, in many cases, shockingly 22 Why does Moss treat this as a single reason with two distinct components rather than as two distinct reasons? I suspect the answer is that, in her view, (2) is the best explanation for (1).



Archaic Epistemology

355

insensitive to the philological, historical, exegetical, and philosophical complications that have inspired the best work in the scholarly tradition to which Moss and I belong—­a tradition that first emerged only about seventy-­five years ago. And, as I said, I am inclined to accept (2) as well; but I fail to see why this, by itself, would support the topic-­based DO reading over the acquaintance-­based one. If we are going to make a decision between these two readings and not reject the decision itself as poorly framed, then I think we must first assume that Plato is alive to the distinction that sets these two readings apart from each other in the first place. The real question here, then, is not whether he explicitly draws the distinction between objects of reference and objects of dependence, but whether he nonetheless observes it somehow—­whether, in his reasoning, he treats the two things differently. It is only if he does so, I think, that we are entitled to demand a decision between the two readings we have been discussing here. Now my argument up to this point has been, in effect, that in a wide range of passages he does observe this distinction and that in these passages we find powerful evidence against the topic-­based DO reading. Moss might want to challenge the latter part of my argument by showing that I have overstated the force of this evidence or that I have ignored other evidence in favour of that reading. But I do not understand how she could hope to settle the dispute between us by insisting (correctly) that Plato never explicitly draws the distinction that gives life and meaning to the dispute itself. The second reason she gives here looks to me like a red herring. After all, both readings attribute to Plato ‘an objects-­based epis­ tem­ol­ogy on which the objects of each cognitive kind are sharply distinct’. So the decision between the two readings does not turn on this; it turns on the issue of what it is for a thing (or a certain kind of thing) to be, in the relevant sense, the ‘object’ of a judge­ ment (or a certain kind of judgement). According to the topic-­ based version, it is for that judgement to be about that thing; and according to the acquaintance-­based version, it is for that judge­ ment to be grounded upon an acquaintance with that thing. As far as I can see, then, the claim Moss makes here does not tell in favour of or against either of these two alternatives. Her third reason, I take it, is just the claim that the topic-­based DO reading can be shown to make good sense of the various pas­ sages in Plato’s works where he deploys his distinction between

356

Matthew Evans

doxa and epistēmē. Unsurprisingly, given everything I have said so far, I very much doubt that it can. As I have argued, there are many passages—­including several in the ‘Two Worlds’ dialogues—­that confound the topic-­based DO reading and render it borderline untenable. I say ‘borderline’ because, if every reasonable alterna­ tive to this reading were found to generate equally severe problems, then this one might still be the best of a bad lot. And yet, as Moss herself seems to acknowledge, the acquaintance-­based DO reading is a reasonable alternative, and, unless I am missing something, it does not generate any especially severe problems at all. Maybe this is not enough to establish that the topic-­based DO reading ‘cannot be made to work’, as Moss puts it; but I think it is enough to estab­ lish that it cannot be made to work well—­and that should be suffi­ cient to tip the scales against it. Would Moss truly disagree with me about any of this, however? When I started working on this paper, I had assumed that she would, but after rereading the paragraph above several times, my confidence began to ebb. Nowhere does she say outright that she believes the topic-­based DO reading is correct, or even that she will try to show that it is. In her very first declarative sentence, in fact, she describes herself not as being persuaded by it, or as com­ mitting herself to it, or as endorsing it, but as ‘experimenting’ with it. Moreover, as I said earlier, I am inclined to agree with her that it is ‘the most obvious way to read Plato, and the way he was read for about two millennia’, and I could certainly imagine myself trying to show—­for fun, profit, or simple exercise—­that it ‘yields compel­ ling accounts of epistēmē and doxa’ and can ‘be made to work’. Yet at the same time I definitely do not accept it. Maybe Moss, by using this language, is signalling to us that she does not accept it either.

VII Suppose, then, that we decide to reject the topic-­based DO reading in favour of the acquaintance-­based one. In that case, we will be ready to ask our original question once again: does Plato have any­ thing worthwhile to tell us, today, about how we ought to conduct ourselves as thinkers? To answer this question effectively, we first need to make sure we understand how Plato’s distinction between



Archaic Epistemology

357

epistēmē and doxa maps onto our contemporary distinction between knowledge and belief—­assuming, again, that the acquaintance-­based DO reading is correct. This mapping will not be straightforward, as I believe Moss’s work has shown. For nearly all of us take it for granted that to have knowledge is just to have a certain kind of true belief—­one that meets the standard of warrant.23 On either DO reading, however, Plato holds that to have epistēmē is not to have any kind of true doxa, no matter what standard it might meet. So adopting any DO reading will compel us to draw one of two con­ clusions: either Plato’s view of knowledge and belief is rad­ic­al­ly at odds with ours, and fruitful conversation between us is not poss­ ible, or else his talk of epistēmē and doxa is semantically disjoined from our talk of knowledge and belief. In what follows I will assume that the latter conclusion is the correct one, if only to keep the lines of communication open between us and him.24 How, then, should we translate Plato’s talk of epistēmē and doxa into the conceptual framework that we are working with today? Though our notion of belief is clearly different from his notion of doxa, it is not clearly different from another notion that I (follow­ ing Moss) assigned to him in Section II: the notion of judgement—­ where this is understood, as before, as a psychological attitude by which we commit ourselves to the truth of (something like) a prop­ osition. In fact, this notion of judgement and our notion of belief

23 Here I do not mean to be ruling out of court the increasingly popular ‘knowledge-­first’ approach associated with the work of Timothy Williamson. On this approach, it is a mistake to suppose, as most mainstream epistemologists do, that knowledge can be analysed as a kind of true belief. But even Williamson would not deny that there is a constitutive relation between knowledge and true belief; he just holds that the more familiar direction of analysis should be reversed: (mere) true belief should be analysed as kind of (failed) knowledge. See, for example, T. Williamson, Knowledge and its Limits (Oxford, 2000), 47. So it is not as though there is some recognizable group of ‘knowledge-­first’ epistemologists who would be ready to embrace the idea—­which the DO reading attributes to Plato—­that know­ ledge and belief have completely different objects. If I had been more willing to sacrifice clarity for accuracy, I would have written the sentence above like this: ‘For nearly all of us take it for granted that there is some constitutive relation between knowledge and true belief.’ 24  To some extent, I suppose, this is a matter of philosophical taste. But I admit I am suspicious of any readers of Plato who, when confronted with an unforced choice between taking him to be comprehensible and taking him to be incompre­ hensible, are for some reason drawn to the latter option. One hopes that they respond differently to their neighbours and fellow citizens.

358

Matthew Evans

look close enough, at least on the surface, to be treated as equiva­ lent.25 So let us grant for the moment that they are. Once we do grant this, I think we will see that nearly all cases of knowledge are going to be, by Plato’s lights, cases of (true and war­ ranted) D-­judgement, rather than cases of E-­judgement. That is because almost all of our judgements, including our true and war­ ranted ones, are grounded upon on our acquaintance with some­ thing, or someone, in our perceptual environment. Not all of them, to be sure; I imagine that many (if not most) of our true and war­ ranted mathematical judgements, for example, will turn out to be E-­judgements. But the important point is that, on this reading, there is a significant amount of overlap between knowledge and doxa, on the one hand, and knowledge and epistēmē, on the other, even though there is no overlap whatsoever between epistēmē and doxa. Knowledge and epistēmē will differ in this respect, then, but will they differ in any other respect as well? Certainly there will be lots of cases of knowledge that are not cases of E-­judgement, for the reasons I gave a moment ago, but what about the converse? Will there be any cases of E-­judgement that are not cases of knowledge? Surprisingly, perhaps, I think there will be. To see why, consider the following slight variation on Alvin Goldman’s famous ‘fake barn county’ scenario: someone (Artemisia, say) is driving down the road, turns to her left, sees a building that looks like a barn, and makes the judgement that the building over there is a barn. Now suppose this judgement is true, but only because of luck: the hun­ dreds of other structures she has driven by so far and will drive by before she leaves the county look (from the road) no less real than this one does, but are all, in fact, fake.26 In that event, I assume, Artemisia’s judgement would not meet the standard of warrant and therefore would not be a case of knowledge. But that would still be so, presumably, even if she had been a barn expert, in­tim­ ate­ly acquainted with what it is for any building—­including that one there—­to be a barn, and even if her judgement had been grounded upon her acquaintance with what it is for any building to be 25  As Moss points out (227–9), Plato seems to have this notion in mind when he uses doxa terminology in later dialogues, such as the Theaetetus (189 e–190 a) and the Sophist (263 e–264 a). 26 See  A.  Goldman, ‘Discrimination and Perceptual Knowledge’, Journal of Philosophy, 73 (1976), 771–91 at 772–3.



Archaic Epistemology

359

a barn. So, although her judgement would not be a case of know­ ledge, it would be, nonetheless, an accomplishment of epistēmē. This scenario raises a number of troubling questions, however, and I want to address two of them in particular here. The first one concerns the status of epistēmē as a power that Plato takes to be ‘infallible’ (ἀναμάρτητον, 477 e 7–478 a 2). Remember that, accord­ ing to my description of the scenario, Artemisia’s E-­judgement (that the building over there is a barn) is true, but only as a matter of luck: if she had turned to her right instead of her left, or if she had looked up from the road a few minutes earlier, she would have made the same judgement—­same content, same ground—­but it would have been false. Is this story really consistent with Plato’s idea that the power of epistēmē is infallible? It is tempting to think not. After all, what this scenario seems to have shown is that Artemisia’s judgement could easily have been mistaken. But if her judgement was the accomplishment of an infallible power, we might think, then it could not easily have been mistaken. Only by admitting this much luck into the power of epistēmē, I think, would Plato be able to accept the conceptual possibility of our original scenario. The alternative response would be to deny that if Artemisia’s belief could easily have been mistaken, then it is the accomplish­ ment of a fallible power, not an infallible one. But this response would be open to us, I take it, only if we were able to offer Plato an account of infallibility according to which Artemisia’s E-­judgement, despite being the accomplishment of an infallible power, could eas­ ily have been mistaken. Here I think our best option would be to offer him a deflationary account, according to which both know­ ledge and epistēmē will turn out to be infallible, and uncontrover­ sially so. On an account of this sort, both knowledge and epistēmē are infallible in the following sense: it is impossible for a mistaken judgement either to constitute knowledge or to be an accomplish­ ment of epistēmē. In that sense, at least, an accomplishment of epistēmē could not in fact be mistaken, even if it could easily have been mistaken. There is a significant downside to this response as well, of course. For, on this approach, whether any given judgement is an accomplishment of the power of epistēmē will be, to some extent, a matter of luck. That is because, on this account, Artemisia’s belief is an E-­judgement, even though it could easily have been mistaken;

360

Matthew Evans

but if it had been mistaken, then it would not have been an E-­judgement. In that sense, anyway, the status of any given judge­ ment as an E-­judgement is going to depend at least partly on chance. I admit this is not easy to accept; but the only obvious alternative, as far as I can tell, would be to deny the conceptual possibility of the original scenario—­which has its own problems, as we have seen. So admitting some degree of luck into the power of epistēmē might end up being the more reasonable option. In any case, we are now in a better position to address the last question I want to raise here, which is this: given how much we care about acquiring knowledge of the things that matter to us, and given how different knowledge and epistēmē turn out to be, why should we care about acquiring epistēmē—assuming that we already have knowledge? What added benefit, if any, would we get from it? The answer I imagine Plato would give is a variation on the one Socrates himself gives in the Meno to a similar question about epistēmē and true doxa. As I suggested earlier, in Section I, he would probably maintain that acquiring a firm intellectual grasp of what makes your true judgement true not only satisfies a certain hunger for explanation that mere knowledge leaves unfulfilled, but also constitutes a worthwhile intellectual achievement in its own right. What is more, it does appear to provide greater stability for your judgement, just as Socrates suggests it does: if your judge­ ment that this or that historical figure is pious, say, is grounded upon your intellectual acquaintance with the Pious, then presum­ ably you would be less likely to change your mind about this under questioning than you would have been if your judgement had been grounded upon a stranger’s testimony instead. Finally, epistēmē— because of its explanatory depth—­is unlike mere knowledge inso­ far as it is guaranteed to make some contribution to a unified, systematic understanding of the relevant subject matter.27 Mere knowledge might be enough to put you in touch with the facts; but epistēmē will put you in touch with the underlying structure that sup­ ports the facts. Epistēmē, unlike mere knowledge, gets to the heart of things.

27  As Moss puts it, epistēmē (unlike mere knowledge) requires ‘a deep grasp of ultimate reality’ (234).



Archaic Epistemology

361

VIII I have left out of discussion here many other questions about Plato’s understanding of epistēmē and doxa that Moss herself raises, and that many other readers would no doubt raise as well. But I hope I have at least provided a sketch of how I believe he might answer some of the more important ones—­and thus how he might still have something worthwhile to contribute to our own epistemological reflections today. To whatever extent I have been successful in this effort, however, I owe it to Jessica Moss’s fiery new book. Every serious interpreter of ancient philosophy should read it, and read it with care. It is a radical challenge to a variety of interpretive assumptions that have become commonplace among us, all of which ought to be unearthed, examined, and either defended or cast aside. I may not agree with Moss about which of these two options is the better one in any given case, but I am grateful to her for pressing the issue upon us as forcefully and skil­ fully as she has. The University of Texas at Austin

BIBLIOGR A PH Y Fine, G., ‘Knowledge and Belief in Republic V’, Archiv für Geschichte der Philosophie, 60 (1978), 121–39. Gerson, L. P., Ancient Epistemology (Cambridge, 2009). Gettier, E. L., ‘Is Justified True Belief Knowledge?’, Analysis, 23 (1963), 121–3. Goldman, A. I., ‘Discrimination and Perceptual Knowledge’, Journal of Philosophy, 73 (1976), 771–91. Plantinga, A., Warrant and Proper Function (Oxford, 1993). Smith, N. D., ‘Plato on Knowledge as a Power’ [‘Knowledge’], Journal of the History of Philosophy, 38 (2000), 145–68. Szaif, J., ‘Doxa and Epistēmē as Modes of Acquaintance in Republic V’, Études Platoniciennes, 4 (2007), 253–72. Vogt, K., Belief and Truth: A Skeptic Reading of Plato [Belief] (Oxford, 2012). Williamson, T., Knowledge and its Limits (Oxford, 2000). Woolf, R., ‘Plato and the Norms of Thought’, Mind, 122 (2013), 171–216. Zagzebski, L., ‘The Inescapability of Gettier Problems’, Philosophical Quarterly, 44 (1994), 65–73.

IN DE X L OCORU M Aeschines In Timarchum 42: 65 n. 43 Aeschylus Persians 397: 28 n. 80 Anaximander, 12 DK A 10: 27–28 A 10. 7–10: 13 A 23: 14–15 A 24: 14–15 A 26: 19 n. 57 B 1: 13–14, 18–19 Anaximenes, 13 DK A 5. 3–6: 15 Anon., In Platonis Theaetetum commentarium, ed. Diels and Schubart 5. 18–36: 267 5. 18–6. 31: 261 n. 2 5. 24–6. 31: 299 6. 17–25: 268 Antiphon Against the Stepmother 13.8: 5 n. 12 Aristophanes Birds 198: 5 n. 12 Aristotle Categories 1b1: 200–201 1b1–2: 180 3a4–5: 180 4a23–8: 189 4a28–b1: 190 n. 67 4a34–b2: 190 n. 68 6a36–7: 180 n. 40, 185–186 6b2: 189 n. 63, 191 n. 70 6b3: 199 n. 95

6b5: 179 6b7–8: 179 n. 37 6b9: 179 n. 37 6a29–30: 179 6a36–7: 179 n. 36 6a38: 179 6a39: 179 7a37: 180 7b15: 182 n. 45 7b15–16: 182 n. 45 7b15–22: 181 7b22: 182 n. 42 7b22–31: 183 7b24: 182 n. 45 7b29–30: 211 n. 118 7b30–1: 183–184 7b31: 187 n. 55 7b31–2: 184 7b33–5: 184 8a13–b24: 199 n. 95 8a16–18: 180 n. 38 8a31–2: 180 n. 40 8b25: 191 n. 71 8b27: 191 8b28: 191–192, 192 n. 77 8b29: 191, 199 n. 95 8b29–32: 192 8b30: 192 nn. 76–7 8b31: 198 n. 91 8b32: 194–195 8b34–7: 192 n. 76, 193 n. 79 9a1–3: 197 n. 90 9a5: 192 n. 76 9a9–10: 192 n. 76 9b13–32: 191 n. 72 10b2: 201 n. 97 11a20–2: 202 n. 100 11a20–3: 199 n. 94 11a20–38: 199 n. 95, 202 n. 100 11a23–36: 199 11a25–32: 187 n. 58 11a29–31: 187 n. 58 11a32–6: 200–201 11a37–8: 191 n. 70, 201 nn. 98–100 11b10–16: 199 n. 95 11b15: 199 n. 95

364 14b27–32: 181 n. 41 15a6–7: 184 n. 48 De anima 408a34–b18: 233–235 408b19–28: 193 n. 81 413a8–10: 234 n. 4 413a21–5: 236 n. 10 416a9–18: 247 n. 28 416b30–1: 249 n. 31 417a27: 198 n. 92 417b27–8: 198 n. 92 417b31–2: 198 n. 92 428b8–9: 190 n. 66, 211 n. 119 433a17–21: 242 De caelo 281a30–3: 223 n. 139 281b15–17: 223 n. 139 283a4–7: 223 283b12–14: 225 n. 145 284a5–11: 256–257 De generatione animalium 777b6–8: 256 n. 42 784b2–3: 249 n. 31 784b30–2: 193 n. 81 De generatione et corruptione 337b35–338a2: 224 De interpretatione 19a2–4: 225 n. 145 De longitudine et brevitate vitae 465a19–23: 193 n. 81 De memoria et reminiscentia 450b1–8: 193 n. 81 De motu animalium 702b20–5: 250 De partibus animalium 650a8–10: 250–251 653a10–19: 252 n. 35 665b9–666b1: 251 n. 34 678a16–19: 249 n. 31 De somno et vigilia 454a8–10: 250 n. 33 454a24–9: 256 n. 43 455a12–27: 250 455b13–34: 249 n. 32 455b34–456a23: 250 456a24–7: 253 n. 39 456a30–2: 250–251 456a32–b6: 251 n. 34 456b5–6: 249 n. 31 456b17–28: 252 n. 35 456b18–20: 250–251 457a33–b6: 252 n. 35 457b29–458a19: 250–251

Index Locorum 458a10–12: 252 n. 36 458a25–6: 250–251 458a25–8: 252 n. 37 458a26–8: 252 n. 35 459b3–23: 253 n. 39 Eudemian Ethics 1241b1–3: 281 n. 46 Metaphysics 983b6–18: 22 n. 63 984a5–7: 15 n. 45 984a5–11: 22 n. 63 1005b23–5: 1 n. 1 1012a24–b22: 1 n. 1 1015a20–2: 247 n. 28 1015b6–9: 205 n. 105 1021a27–8: 185 n. 50 1021a29: 185–186 1021a31: 185–186, 189 n. 63 1021a33–b1: 189 n. 63 1021b3–4: 186 n. 52 1026b27–35: 172 n. 22 1036a5–7: 207 n. 108 1039a3–6: 151 n. 110 1039b32–4: 195 n. 84 1039b32–1040a2: 221 n. 135 1039b33: 222 n. 136 1040a2–5: 204–205 1050a13–15: 188 n. 61 1050b24–8: 256 n. 43 1051b12: 210–211 1051b13–14: 189 n. 65 1051b13–15: 211 n. 117, 225 n. 145 1056b34: 185–186 1056b36–1057a1: 185 n. 51 1057a1–2: 182 n. 43 1057a7–8: 185–186 1057a9–12: 186 n. 54 1074b28–9: 256 n. 43 1087a15–21: 172 n. 21 1088a24: 199 n. 95 Meteorologica 381b13: 249 n. 31 Nicomachean Ethics 1096a21: 199 n. 95 1103a15–17: 196 1114a25–8: 194 n. 83 1139b7–9: 225 n. 145 1139b15–17: 176 n. 31 1139b16–17: 196 1139b18–19: 203–204 1139b19: 171 n. 16 1139b19–20: 169 n. 5 1139b19–23: 177, 204–205



Index Locorum

1139b20: 203–204 1139b21: 222 n. 136 1139b22: 204, 223–224 1139b24: 169–170, 224 n. 142 1139b25: 196 1139b31: 196 1139b32: 205 1145a31: 194 n. 83 1147a10–24: 194 n. 82 1147a12: 197 1147a13: 197 n. 89 1147a13–14: 197 n. 88 1147a20–2: 193 n. 80 1147a22: 196 1147b8–17: 194 n. 82 1149a4–12: 194 n. 83 1168a5–10: 281 n. 46 1177b4–15: 117 Physics 185b19–25: 1 n. 1 188a27–31: 17 n. 49 203b6–28: 13 n. 35 246b20–247a2: 202 n. 101 247b2–3: 202 n. 101 247b9–10: 202 n. 101 247b13–16: 197 247b16: 197 247b17–18: 196 250b24–6: 237 n. 12 250b26–9: 237 n. 12 251b17–19: 237 n. 12 252b17–28: 237–238 253a7–21: 238, 240–242 253a13: 242 253a14–15: 240 n. 16 253a15–17: 255 n. 41 253a15–20: 247 253a18: 242 256a4–21: 233–235 257a33–258b9: 233–235 258b15: 258 259a20–b20: 238 259a27–b1: 236–237, 242–243 259b1–16: 243 n. 21 259b1–20: 236–240, 242 259b2–3: 236–237, 242–243 259b3–6: 235 n. 6 259b6: 245 259b7–8: 240, 242 n. 19, 244 259b7–9: 240 n. 17 259b8: 240–241, 243 n. 20, 245–246 259b8–14: 252–253 259b10–11: 242 n. 19

259b11: 240 n. 17 259b11–12: 240–241 259b11–15: 247 259b13–14: 240 n. 17, 242–243 259b14–15: 254–255 259b16–20: 242–243 259b16–22: 234 n. 4 259b17–20: 236–237 259b19–20: 257 259b20–8: 257 259b22–31: 257 259b28–31: 258 265b32–266a1: 234 n. 3 267b2–4: 256 n. 43 Politics 1255b16–17: 8 n. 21 Posterior Analytics 71a34–b3: 209 n. 114 71b9: 168–169, 171 n. 16 71b9–12: 167 71b9–16: 168 n. 3 71b12: 168 n. 3 71b13–15: 169 n. 5 71b15: 168–169 71b15–16: 168 n. 3, 168–169 71b25: 210 n. 116, 220–221 72b3–4: 194–195 73a21: 168 nn. 3–4 73a21–3: 168 n. 3 73a21–7: 174 73a22: 168 n. 3, 187 n. 56 73a34–b3: 174 n. 26 73b16–17: 174 n. 26, 187 n. 57 73b26–7: 174 n. 26 74b5–6: 168 n. 4 74b5–15: 175 74b6: 168 n. 3 74b14: 175 n. 27 74b14–15: 175 n. 27 74b32–6: 204–205, 219 74b36–9: 221 74b38–9: 221 75b24–6: 172 n. 21 75b24–30: 169–170, 209 n. 114 76b16–25: 187 n. 58 78b32–79a16: 187 n. 58 86a6–7: 187 n. 56 87b19–25: 172 n. 22 87b30–88a2: 172 n. 21 87b39–88a2: 172 n. 21 88b7–8: 172 n. 22 88b30: 187 n. 56 88b30–1: 168 n. 4

365

366 88b30–89a10: 168 n. 3 88b31: 168 n. 3 88b32: 168 n. 3 88b33–7: 176 88b36: 187 n. 57 89a2: 176 n. 31 89a2–4: 187 n. 58 89a4–5: 176, 190 n. 69 89a6–9: 169 n. 5 89a6–10: 177 89a7: 168 n. 3 89a10: 168 n. 3 89a11: 176 Prior Analytics 32b18–22: 172 n. 22 67a27–67b5: 172 n. 21 67a39–b3: 208 n. 112 Rhetoric 1407b14–15: 4 n. 8 1418a1–5: 225 n. 145 Topics 102a18–19: 207 107b6–12: 198 n. 93 125b4: 187 n. 55 129b1–5: 207 n. 110 129b9–13: 207 129b13–14: 207 n. 110 129b14–17: 207 129b22–3: 207 n. 110 131a1: 207 n. 110 131a17: 207 n. 110 131b19–23: 207 131b21: 207 n. 110, 208–209 131b21–3: 195 n. 84, 204–205 131b25–6: 208 n. 111 131b29–30: 207–208 131b31: 207–208 131b31–2: 207 n. 110 131b32: 207–208 132a22–4: 207 159b30–3: 1 n. 1 Cicero De finibus 3. 24: 287 n. 55 3. 32: 287 n. 55 3. 58–9: 271 n. 25 3. 59: 308 3. 62–4: 265 3. 63: 264 n. 9 3. 63–6: 264 n. 11 3. 63–8: 261 n. 2 3. 65: 265 n. 13 3. 70: 273 n. 29

Index Locorum De officiis 1. 7: 302–303 1. 53–4: 302 1. 58: 303 1. 107–15: 301 n. 12, 306 n. 19 3. 21: 322 n. 43 3. 22: 316 n. 36 3. 63: 302–303 3. 90: 268 n. 16, 299–300 3. 99–115: 322 Demosthenes Against Timocrates 28. 3: 28 n. 80 Diels, H. and Kranz, W. (eds.), Die Fragmente der Vorsokratiker, 6th edn 12 A 10: 27–28 12 A 10. 7–10: 13 12 A 23: 14–15 12 A 24: 14–15 12 A 26: 19 n. 57 12 B 1: 13–14, 18–19 13 A 5. 3–6: 15 22 B 1: 1–3, 4, 7, 23–24, 29, 29 n. 83 22 B 2: 23–24 22 B 7: 29 n. 82 22 B 9: 9 n. 25 22 B 10: 7, 26–28, 36–37 22 B 11: 9 n. 25 22 B 12: 10 n. 27 22 B 13: 9 n. 25 22 B 30: 6 n. 18, 22, 23, 25, 26, 27–28 22 B 31: 23, 32 22 B 31. 2–3: 32 22 B 31. 9–10: 32 22 B 32: 25–26 22 B 41: 7, 25–26 22 B 50: 5–7, 23–26 22 B 51: 9, 11–12, 34–36 22 B 53: 29 n. 82 22 B 57: 8 22 B 61: 9, 11–12 22 B 67: 25–26 22 B 76: 10–11, 12, 22, 32, 32 n. 87 22 B 76b–c: 10, 32 22 B 80: 18, 18–19, 29 n. 82 22 B 84a: 5 n. 11 22 B 90: 22, 28 n. 81 22 B 123: 5 n. 17 22 B 126: 10–12 28 B 8. 53–61: 19 n. 55



Index Locorum

Diogenes Laertius 2. 3: 15 n. 43 7. 23: 273 7. 87: 265 n. 12, 308, 309 7. 89: 308 7. 94: 261 n. 1 7. 107–8: 307 7. 109: 312 n. 29 7. 116: 276 n. 35 7. 124: 273 n. 28 7. 127: 265 n. 12 7. 139: 320 n. 40 7. 140: 325 Elias In Aristotelis Categorias commentarium, ed. Busse 238. 8–10: 200 n. 96 Epictetus Dissertationes 2. 5: 277 n. 38 2. 5. 24: 315–316 2. 5. 24–5: 309–310 2. 6. 9–10: 310–311 2. 10. 5: 310 2. 10. 6: 310–311 Gorgias Encomium of Helen 19: 65 n. 42 Heraclitus, 22 DK B 1: 1–3, 4, 7, 23–24, 29, 29 n. 83 B 2: 23–24 B 7: 29 n. 82 B 9: 9 n. 25 B 10: 7, 26–28, 36–37 B 11: 9 n. 25 B 12: 10 n. 27 B 13: 9 n. 25 B 30: 6 n. 18, 22, 23, 25, 26, 27–28 B 31: 23, 32 B 31. 2–3: 32 B 31. 9–10: 32 B 32: 25–26 B 41: 7, 25–26 B 50: 5–7, 23–26 B 51: 9, 11–12, 34–36 B 53: 29 n. 82 B 57: 8 B 61: 9, 11–12 B 67: 25–26 B 76: 10–11, 12, 22, 32, 32 n. 87

367

B 76b–c: 10, 32 B 80: 18, 18–19, 29 n. 82 B 84a: 5 n. 11 B 90: 22, 28 n. 81 B 123: 5 n. 17 B 126: 10–12 Herodotus Histories 1. 64: 28 n. 80 Hierocles Elements of Ethics, ed. Bastianini and Long 6. 49–53: 305 7. 4–5: 305 8. 54: 305 9. 3–7: 305 Hippolytus Refutatio omnium haeresium 9. 9. 1: 25 n. 76 9. 9. 2: 9 n. 26 9. 9. 3: 4 n. 8 9. 10. 5: 9 n. 24 Homer Iliad 4. 105–26: 36 n. 90 10. 127: 5 n. 15 Odyssey 10. 302–6: 5 n. 17 10. 549: 5 n. 15 11. 2: 5 n. 14 14. 3: 5 n. 14 Laks, A. and Most, G. (eds.), Early Greek Philosophy Parmenides D 8. 35–6: 173 n. 24 Long and Sedley (eds.), The Hellenistic Philosophers 47O 3: 320 n. 40 57F: 261 n. 2, 264 n. 9, 264 n. 11, 265, 265 n. 13 57G: 261 n. 2, 303 n. 15 57G 1–4: 294–295 57G 5–6: 267 57G 6–7: 264 n. 11 57H: 261 n. 2, 266, 268 57H 3: 299 59: 271 n. 25 59A: 309

368

Index Locorum

59C 1: 307 59C 3: 307 59E 3: 312 n. 29 59F: 271 n. 25 59L: 287 n. 55 60P: 265 n. 14, 273 n. 30 61I: 265 n. 12 63C: 265 n. 12 63C 1: 308 63C 2: 309 63C 5: 308 64H: 287 n. 55 65F: 276 n. 35 67A: 264 n. 9 67A 1: 298 67P: 273 n. 28 Marcus Aurelius Meditations 2. 1. 4: 316 7. 13. 2: 316 Olympiodorus In Categorias Commentarium, ed. Busse 129. 28: 201 n. 97 Parmenides, 28 DK B 8. 53–61: 19 n. 55 Parmenides, LM D 8. 35–6: 173 n. 24 Paul 1. Corinthians 12:13: 316 12: 25: 316 Philoponus, John In Aristotelis Categorias commentarium, ed. Busse 106. 8–11: 179 n. 37 121. 15–16: 184 n. 46 Pindar Isthmian Odes 4. 50: 5 n. 17 Plato Charmides 168 b–169 a: 86 n. 7 Crito 47 d 8–e 6: 132 Euthydemus 278 e 3–282 d 3: 111 n. 13

Euthyphro 7 b 2–c 8: 70 n. 61 13 b 4–c 10: 153 Gorgias 468 e 6–469 a 5: 129–130 476 a 3–478 e 5: 131 n. 59 507 e 6–508 a 8: 155 n. 117 511 e 6–612 b 1: 311 n. 27 Laws 628 c 6–7: 110–111 631 b 3–6: 110, 115 631 b 3–d 6: 111 n. 13 644 a 2–5: 137 648 b 7–c 5: 133 n. 66 660 d 11–662 c 5: 111 n. 13 661 a 4–663 c 6: 117–118 662 c 6–663 b 6: 156 662 d 1–663 a 7: 111 663 e 8–664 a 3: 111 664 a 2–3: 110 n. 9 664 a 3: 110–111 664 b 7–8: 156 665 c 2–5: 110 n. 9 667 d 9–e 1: 114 683 b 3–4: 110 n. 9 693 b 3–e 3: 111 693 b 4: 111 n. 11 693 b 6–c 6: 111 693 d 2–694 a 2: 138 701 c 6–7: 111 n. 11 701 d 6–e 8: 111 701 d 9: 111 n. 11 705 d 2–706 a 4: 111, 113 n. 19 705 e 3–706 a 4: 111–112 709 a 1–c 3: 160 n. 130 709 a 1–d 7: 113 710 a 5–b 2: 137 715 a 7–b 6: 109–110 715 b 4: 113 n. 18 728 b 2–c 2: 150 728 b 2–c 6: 150 730 d 2–7: 129 730 d 2–731 a 5: 133 n. 66 739 a 3–5: 122–123 739 a 4–740 a 2: 123 739 c 3–d 6: 127–128 739 d 6–8: 123 n. 43 740 b 11–741 a 4: 132 n. 63 741 e 1–6: 137 742 d 7–e 1: 110 743 c 5–6: 110 743 d 2–6: 137 757 a 1–2: 135–136

757 a 1–758 a 2: 132–133 757 a 5–d 6: 132–133 757 d 6–e 2: 133 759 b 4–7: 133 n. 66 770 d 5–e 5: 112 n. 17, 134 776 b 5–777 c 5: 136 777 c 8–d 3: 135–136 777 d 2–3: 136 777 d 2–e 2: 136 777 d 2–e 6: 136 781 a 11–b 3: 135 n. 67 783 d 8–e 1: 130 784 a 1–b 1: 130 784 b 6–c 5: 130 788 a 1–790 b 6: 112 n. 17 803 c 2–804 c 1: 160 n. 130 806 c 3–7: 110, 113 n. 19 816 c 7–d 2: 112 828 d 1–6: 139 828 d 10–829 a 2: 139 846 d 2–3: 137 846 d 3–7: 137 850 a 2–8: 137 853 b 1–3: 147–148 853 b 4–7: 129–130 854 b 6–c 3: 150, 152–153 854 d 5–e 1: 131 n. 59 858 c 10–d 9: 110 n. 9 862 c 6–863 a 2: 150 n. 105 869 b 7–c 4: 134 870 a 2–b 2: 138 n. 77 872 c 7–873 a 7: 150 n. 105 875 a 2–c 2: 110 875 a 4: 110 886 a 44–5: 138 n. 77 890 d 1–9: 161 892 a 2–c 7: 161 893 e 6: 151 n. 106 894 b 8–896 b 2: 233–235 895 a 1–3: 235–236 895 a 6: 235–236 895 a 6–b 7: 235–236, 238 n. 13 895 b 1: 235–236 899 b 2–7: 161 899 b 5: 142 n. 86 899 d 3–4: 153 900 c 9–d 3: 161 900 d 5: 151 n. 106 900 d 5–e 4: 161 901 a 7–9: 161 902 a 1–903 a 3: 153 902 d 7–e 3: 153 903 b 4–7: 152–153

Index Locorum 903 b 4–d 3: 144 903 b 5: 145 903 b 7–c 1: 146–147, 149 903 b 8: 145 903 c 1–5: 148, 151 n. 110 903 c 3–d 4: 145 903 c 4–5: 145, 151 n. 110, 152–153, 903 c 6–d 1: 151 n. 106 903 c 6–d 3: 150–151 903 c 7: 145 903 d 1: 145 903 d 1–3: 147 n. 97, 151 n. 106, 151 n. 110 903 d 2–3: 151 n. 106 903 d 3–e 1: 150 n. 104, 151 n. 110 903 d 3–905 c 4: 149, 151–152 903 d 5–6: 144 n. 88 903 e 3–4: 149 904 a 6–905 b 7: 149 904 b 1–3: 146 904 b 1–6: 152 904 b 3–8: 151 n. 110 904 b 4–5: 145 904 b 6–d 4: 147 904 b 8–c 3: 149–150 904 c 1–905 b 1: 150 n. 104 904 d 1–4: 149 904 e 4–905 b 1: 151 n. 110 905 a 1–c 7: 150 905 c 1–4: 151 n. 106 905 c 1–5: 150–151 906 a 2–b 4: 160–161 906 a 2–c 6: 153 906 b 1–4: 153–154 906 b 9–c 6: 161 916 b 6: 151 n. 106 918 d 4–9: 137 919 c 4–6: 137 920 a 3–4: 137 923 b 2–7: 110 925 e 7–926 a 3: 110 933 e 7–934 b 3: 150 n. 105 934 a 5–b 3: 150 n. 105 945 e 3–6: 110 n. 9 951 b 4–c 4: 139 958 a 2–3: 110 n. 9 959 b 5–c 3: 150 n. 104 962 b 4–963 a 9: 111, 113 n. 19 Meno 82 b–87 c: 340 85 c 9–d 1: 193 n. 80 87 c 5–89 a 7: 111 n. 13

369

370

Index Locorum

97 a–98 d: 342 97 d–98 a: 341 97 d–98 e: 335 97 d 10: 192 n. 77 97 e 2–98 a 8: 334 97 e 4: 192 n. 77 97 e 7: 192 nn. 77–8 98 a 1–2: 192 n. 77 98 a 6: 192 n. 77 Parmenides 164 d 2: 90 n. 16 Phaedo 66 b 3–4: 90 n. 16 78 d: 86 n. 7 79 c 2–8: 62 n. 35 81 d 2: 66–68 82 d 9–83 a 1: 66 n. 46 82 e 3: 90 n. 16 100 b–c: 343 100 b 3–c 6: 342 101 a 9–b 2: 142–143 Phaedrus 245 c 2–5: 235–236 245 c 2–246 a 2: 233–235, 236 n. 10, 257–258 245 d 8–e 2: 235–236 245 e 2–246 a 2: 235–236 Philebus 30 a 5–8: 151 n. 109 35 b 1–d 3: 64 n. 39 39 b 7: 66–68 40 a 6–9: 67 n. 48 40 a 6–12: 75–76 40 a 9: 66–68 40 a 9–12: 75–76 40 e 22–3: 74 n. 71 41 a 5–42 c 5: 77 n. 79 41 a 7–42 c 3: 53–54 41 e 9–42 a 3: 77 n. 78 42 a 5–9: 76–77 42 b 8–c 3: 76–77 51 c: 86 n. 7 Politicus 293 a 6–e 6: 131 293 b 1–c 3: 121 294 a 10–b 6: 134 295 b 10–297 b 4: 112 n. 17 308 c 1–7: 148 n. 100 308 e 4–309 a 6: 131 309 e 10–14: 148 n. 100 Protagoras 313 c 7–314 b 4: 79–80 352 b 4–c 2: 45

352 b 7: 64–65 352 b 7–8: 74 n. 71 352 c 2: 49, 64–65 352 c 3–6: 46 352 d 6–7: 42 n. 2 352 d 8–e 1: 42 n. 2 352 e 1: 64–65 352 e 2: 64–65 353 a 1: 64–65 353 a 5: 64–65 353 c 1–8: 64–65 353 c 6–7: 64–65 353 c 6–d 4: 52–53 354 a 4–b 5: 53 n. 18 355 a 1–5: 46 355 a 7–b 1: 46 n. 11 355 b 2–3: 46 n. 12 355 b 3: 49 n. 13, 64 n. 37 355 d 5–6: 65–66 356 a 5–c 3: 52 356 c 5: 65–66 356 c 5–6: 49 n. 13 356 c 5–8: 63–64 356 c 5–357 b 3: 47–48, 49 n. 13 356 d 4: 44 n. 10 356 d 4–7: 63–64 356 d 5: 49 356 d 5–7: 62 n. 34 356 d 8–e 1: 59–60 356 d 8–e 2: 70 n. 60 357 b 2–3: 44 n. 10, 50–52, 69 n. 59, 70 n. 60 357 b 5–6: 71 n. 62 357 c 3–4: 59 n. 26 358 b 6–360 e 5: 53–54 358 b 7–c 1: 61 Republic 331 c 1–d 5: 141 331 c 3–5: 141 335 a 6–336 a 10: 142 n. 84 335 e 5–6: 142–143 345 d 1–5: 148 347 b 7: 142 n. 86 358 e 4–359 b 2: 150 368 c 8–e 6: 321 n. 41 382 a 1–2: 66–68 382 a 44–c 2: 131 n. 61 382 c 7–d 4: 131 n. 62 382 d 6–383 a 1: 131 n. 62 382 e 3–4: 131 n. 62 387 e 10–388 a 3: 135 n. 67 389 b 2–10: 131–132 409 e 4–410 a 4: 132

412 d 9–e 2: 133 414 b 7–415 d 5: 131–132 414 e 6: 315 n. 30 416 b 2: 315 n. 30 416 c 5–417 b 9: 143 n. 87 417 b 1: 315 n. 30 420 b 3–7: 108 420 b 3–c 3: 125 420 b 4–8: 106 420 b 7–8: 108 420 d 2–5: 107 420 e 1–7: 108 420 e 1–421 a 3: 106–107 421 b 5–c 5: 108 421 b 6: 107 n. 4 422 d 2–7: 137–138 428 d 3: 139 n. 80 438 a–d: 86 n. 7 441 d 11–e 5: 117 n. 27 442 c 4–8: 117 n. 27 442 c 4–d 2: 153 443 c 9–444 a 2: 143 443 d 3–6: 153 444 d 2–4: 319 n. 39 444 e 6–445 a 4: 117 n. 28 455 c 5–e 4: 135 459 c 3–461 e 4: 131–132 459 c 9–e 4: 131–132 459 d 8–e 4: 132 459 d 9–10: 132 460 c 1–5: 132 n. 63 460 c 2: 132 460 c 3: 132 461 b 8–c 5: 132 461 c 2–5: 132 461 d 1–e 1: 297 n. 5 462 a 2–b 6: 109 462 a 2–e 3: 123–124 462 a 3–b 2: 127 462 b 8: 127 462 d 8–e 2: 135–136 464 d 6–e 2: 124 n. 45 465 d 3–466 c 4: 123–124 469 b 5–c 6: 135–136 469 b 5–471 c 2: 137–138 469 d 4–e 2: 135 n. 67 470 a 5–471 c 2: 137–138 470 b 4–c 3: 137–138 470 c 5–7: 137–138 470 c 5–d 2: 137–138 471 a 6–7: 137–138 471 b 2–5: 137–138 476 b 5: 141

Index Locorum 476 e–480 a: 344 476 e 6–480 a 13: 336–337 477 a–b: 347 n. 14 477 a 6–b 10: 336–337 477 c–478 a: 338–339 477 d 2: 338 477 e 7–478 a 2: 359 478 d–e: 347 n. 14 479 a–b: 336–337 479 e 6–7: 336–337 495 c 7–496 a 10: 137 505 a 2: 345 506 b–d: 345 506 b 2–4: 349–350 506 c–d: 345, 349–350 506 c–e: 349–350 506 d–e: 349–350 506 d 5–e 2: 349–350 506 d 7–e 3: 345–346 507 a–509 c: 346, 349–350 509 d–511 e: 347 n. 15 510 d–511 a: 348–349 515 c 4–516 a 3: 69 n. 57 515 c 4–516 c 6: 71–72 519 e 1–520 a 2: 106–107 520 a 6–b 5: 159 n. 125 520 b 1–5: 140 520 b 5–c 1: 140 520 b 7–c 5: 344 520 c 4: 344–345 520 e 1: 140 533 d–534 a: 347 n. 15 540 e 4–541 a 7: 132 541 a 4–7: 107 n. 4 545 d 5–547 c 9: 117 n. 28 546 a 2: 319 n. 38 580 a 8–c 9: 117 n. 28 581 d 9–e 4: 117 n. 28 582 e 8–583 a 4: 117 n. 28 584 a 7–10: 66–68 586 a 6–b 5: 328–329 587 a 8–d 1: 117 n. 28 587 b 12–588 a 10: 117 n. 28 588 b 10–592 a 4: 157 n. 121 589 b 1–6: 153 590 c 1–5: 137 590 c 1–591 a 3: 135 n. 68 590 c 7–d 6: 135–136 595 b 5–7: 72 n. 63, 79–80 596 a 5–8: 141 599 a 1–3: 66–68 599 a 2: 66–68 601 a 4–b 4: 66–68

371

372 601 a 4–b 7: 76 n. 76 602 b 1–4: 69 n. 56 610 d 5–7: 131 n. 60 611 b 9–612 a 6: 157 n. 121 617 e 1–5: 149 619 d 1–7: 150 Sophist 216 b 1: 92 nn. 18 and 19 219 a 8: 92 n. 20 219 c 2: 92 n. 20 219 d 5: 92 n. 20 220 a 6: 92 n. 20 220 a 7: 92 n. 20 220 e 6: 92 n. 20 222 d 6: 92 n. 20 222 e 3: 92 n. 20 223 c 6: 92 n. 20 223 c 9: 92 n. 20 225 c 2: 92 n. 20 226 c 11: 92 n. 20 226 e 1: 92 n. 20 226 e 5: 92 n. 20 227 c 7: 92 n. 20 227 c 8: 92 n. 20 227 d 13: 92 n. 20 228 c 1: 92 nn. 18 and 19 229 c 2: 92 n. 20 230 a 9: 92 n. 20 234 b 2: 92 n. 20 234 b 3: 92 n. 20 234 e 1: 66–68 235 a 6: 92 nn. 18 and 19 235 d 1: 92 n. 20 235 d 6–236 c 7: 66–68 236 a 4–6: 69 n. 55 236 b 4–6: 69 n. 56 236 c 6: 92 n. 20 236 d 2: 92 n. 20 238 e 2: 92 nn. 18 and 19 239 a 10: 92 n. 20 240 d 1–2: 66–68 246 b 8: 92 n. 20 246 c 9: 92 n. 20 248 a 4: 92 n. 20 249 d 1: 92 n. 20 251 e 10: 92 nn. 18 and 19 252 a 7: 92 n. 20 253 d 1: 92 n. 20 253 d 1–3: 5 n. 16 254 b 8–d 5: 93–95 254 c 3: 92 n. 20 254 d 4–255 e 2: 85–86 254 d 14–255 a 2: 93–95

Index Locorum 255 b 1: 92–93 255 b 3: 92–93 255 c 6: 92 n. 20 255 c 13–14: 85–86, 96–98 255 c 9–d 7: 92–93 255 c 9–e 2: 85–86, 88–89, 93–95, 99–100 255 d 1: 85–86, 93, 96–98 255 d 3–4: 96 255 d 4: 91–97, 99–100 255 d 4–6: 87–88, 90–91, 93, 95–100 255 d 4–7: 96 255 d 5: 91 n. 17 255 d 5–6: 96 255 d 6–7: 85–86, 93, 96–98 255 d 9–e 1: 92–93, 96 255 e 1: 92 n. 20 255 e 3–6: 98–99 255 e 5: 92–93 256 a 1: 92–93 256 a 7: 92–93 256 d 9: 92–93 256 e 3: 92–93 256 e 6: 92 n. 20 258 a 11–b 1: 85–86 258 c 4: 92 n. 20 258 d 6: 92 n. 20 259 a 6: 92 nn. 18 and 19 259 a 8: 92 nn. 18 and 19 259 e 6: 92 n. 20 260 d 3: 92 nn. 18 and 19 260 d 5: 92 nn. 18 and 19 260 d 7: 92 nn. 18, 19 and 20 260 d 8: 92 nn. 18 and 19 261 d 1: 92 n. 20 263 e–264 a: 358 n. 25 264 c 1: 92 n. 20 264 c 4: 92 n. 20 264 c 12–d 1: 66–68 265 a 8: 92 n. 20 266 b 9–10: 66–68 266 c 4: 92–93 266 d 6: 92 n. 20 266 e 5: 92 n. 20 267 d 6: 92 n. 20 Symposium 211 a–b: 86 n. 7 Theaetetus 152 d: 86 n. 7 156 e–157 b: 86 n. 7 160 b: 86 n. 7 160 e 6–161 a 4: 132 176 a 5–177 a 8: 147 n. 97



Index Locorum

189 e–190 a: 358 n. 25 209 b 8: 296–297 Timaeus 19 a 1–5: 132 n. 63 27 d 5–28 a 1: 173 n. 24 29 a 5–6: 146 29 e 1: 161 29 e 1–3: 146 29 e 1–4: 153 30 a 2–3: 153 30 a 2–b 6: 145 n. 90 30 a 6–7: 146 30 b 1–c 1: 158 30 b 3–6: 146 30 b 6: 148 30 c 5–31 a 8: 151 n. 110 34 b 4–9: 151–152 36 d 8–37 c 5: 153 36 d 9–e 3: 151 n. 110 36 e 5–37 e 5: 158 41 c 6–42 a 1: 158 41 d 4–8: 151–152 41 d 4–e 1: 151–152 41 d 4–42 b 2: 135 42 a 3–c 1: 115 42 b 3–d 2: 135 46 a 2: 66–68 51 c: 86 n. 7 53 b 5–6: 146 68 e 1–6: 146 69 a 1: 145 n. 90 71 a 3–5: 67 n. 48 71 a 3–d 4: 66–68 76 d 6–e 3: 148 n. 100 89 e 3–90 d 7: 158 90 c 6–d 7: 157 n. 121 90 d 1–7: 151–152 90 e 1–91 a 1: 135 92 c 1–3: 150 n. 104 92 c 4–9: 157 Plotinus Enneads 3. 1. 4, 1–24: 151 n. 110 3. 2. 12, 7–13: 143 n. 87 6. 5. 10, 50–2: 151 n. 110 Plutarch De Alexandri magni fortuna aut virtute 329 a: 264 n. 10 329 a–b: 298 De communibus notitiis 1068 d: 270 n. 20

373

1069 e: 309 De Stoicorum repugnantiis 1039 d–e: 311 n. 27 [Plutarch] Stromateis 2: 13 n. 36 Porphyry In Aristotelis Categorias commentarium, ed. Busse 140. 20: 201 n. 97 140. 24–141. 5: 201 n. 99 Seneca De beneficiis 1. 3. 9: 262 n. 5 1. 4. 2: 262–263 1. 5. 3: 271 n. 25 1. 6. 1: 274–275 1. 6. 1–2: 269–270, 277, 276 n. 37 2. 15. 1: 273 n. 29 2. 17. 3–5: 277 n. 39, 279–280, 315 n. 32 2. 17. 4–5: 279 n. 45 315 n. 32 2. 17. 5: 278 n. 41 2. 18. 2: 262 n. 5 2. 18. 5: 262–263, 282 n. 50, 283 n. 51, 284 2. 18. 5–6: 272 2. 18. 7: 279 n. 45 2. 21. 2: 262–263, 272, 273–274, 282 n. 50 2. 31. 2–3: 275–276 2. 32. 4: 277, 277 n. 39, 279–280 2. 34. 1: 276, 280–281 3. 18. 1–22. 4: 288–289 3. 29. 2–32. 6: 288–289 5. 1. 1: 271 n. 25 5. 4. 1–2: 280–281 5. 4. 3–5. 1: 270–271 5. 10 1–2: 274–275 5. 10. 4: 274 n. 34 5. 12. 5–7: 270 n. 20, 271 5. 20. 2: 278 n. 41, 289 6. 7. 2: 272 6. 16. 1: 272, 286 n. 54 7. 2. 1: 271 7. 12. 1: 270–271, 283 7. 17. 1: 271 n. 25 Epistulae morales ad Lucilium 6. 3: 273–274 6. 6. 3: 280–281 9. 8–10: 283, 283 n. 51

374

Index Locorum

9. 13–18: 288 n. 58 12. 6: 295 n. 3 59. 2: 276 n. 35 76. 8–11: 265 n. 12 81. 10–14: 270 n. 20 81. 13: 270–271 92. 3–4: 288 n. 58 92. 30: 316 109. 15: 273 n. 29, 273–274, 281 n. 48 Naturales quaestiones 1, Praef. 3: 308 1, Praef. 14: 308 Sextus Empiricus Adversus mathematicos 7. 132–3: 4 n. 8 Simplicius In Aristotelis Categorias commentarium, ed. Kalbfleisch 189. 27–9: 182 n. 45 193. 33–4: 182 n. 45 194. 3–4: 185 n. 49 194. 10: 185 n. 49 194. 22: 185 n. 49 196. 28–9: 185 n. 49 293. 22–5: 202 n. 100 In Aristotelis Physica commentaria, ed. Diels 24. 13–16: 13 n. 35 24. 26: 15 n. 43 24. 26–25. 1: 15 n. 44 150. 24–5: 13 n. 37 1258. 4–30: 248–249 Stobaeus 1. 3. 53–4: 301, 314 2. 7. 5a: 261 n. 1 2. 7. 5b. 3: 306 n. 19 2. 7. 5b. 4: 319 n. 39 2. 7. 5b. 8: 265 n. 12 2. 7. 8: 271 n. 25 2. 7. 11b: 273 n. 31 2. 7. 11c: 273 n. 31 2. 7. 11i: 265 n. 14, 273 n. 30

2. 9. 7: 301, 313–314 3. 39. 34–6: 301 3. 39. 35: 317, 319 3. 39. 35–6: 314 3. 39. 36: 326 n. 49 4. 22a: 306 4. 22a. 21: 302 n. 14, 313, 317 4. 22a. 21–4: 300 4. 22a. 22: 306, 312 4. 22a. 23: 313 4. 22a. 24: 315 n. 30 4. 24: 306 4. 24a. 14: 300, 318–319 4. 25: 306 4. 25. 53: 300, 300–301, 301, 301–302, 306, 314, 315 n. 30 4. 27: 306 4. 27. 20: 300–302, 314, 317, 321 4. 27. 23: 261 n. 2, 264 n. 11, 267, 294–295, 301, 307, 317, 320, 326–327 4. 39. 22: 309 Stoicorum veterum fragmenta, ed. von Arnim i. 262: 298 i. 324: 273 i. 566: 265 n. 12 ii. 634: 320 n. 40 iii. 4: 308 iii. 70: 261 n. 1 iii. 76: 261 n. 1 iii. 98: 273 n. 31 iii. 264: 306 n. 19 iii. 278: 319 n. 39 iii. 348: 273 n. 29 iii. 491: 309 iii. 493: 307 iii. 496: 312 n. 29 iii. 510: 309 iii. 625: 273 n. 31 iii. 672: 270 n. 20 Thucydides 2. 37. 2–3: 128

Notes for Contributors to Oxford Studies in Ancient Philosophy 1. Articles should be submitted with double line-spacing throughout. At the stage of initial (but not final) submission footnotes may be given in small type at the foot of the page. Page dimensions should be A4 or standard American quarto (8½ × 11ʺ ), and ample margins (minimum 1¼ʺ or 32 mm) should be left. 2. Submissions should be made as an anonymized PDF file attached to an e-mail sent to the Editor. Authors are asked to supply an accurate word-count (a) for the main text, and (b) for the notes. The e-mail which serves as a covering letter should come from the address to be used for correspondence on the submission. A postal address should also be provided. If necessary, arrangements for alternative means of submission may be made with the Editor. Authors should note that the version first submitted will be the one adjudicated; unsolicited revised versions cannot be accepted during the adjudication process. The remaining instructions apply to the final version sent for publication, and need not be rigidly adhered to in a first submission. 3. In the finalized version, the text should be double-spaced and in the same typesize throughout, including displayed quotations and notes. Notes should be numbered consecutively, and may be supplied as either footnotes or endnotes. Any acknowledgements should be placed in a final note attached to the last word of the article. Wherever possible, references to primary sources should be built into the text. 4. Use of Greek and Latin. Relatively familiar Greek terms such as psuchē and polis (but not whole phrases and sentences) may be used in transliteration or likewise a few, isolated terms where translation would be prejudicial. Wherever possible, Greek and Latin should not be used in the main text of an article in ways which would impede comprehension by those without knowledge of the languages; for example, where appropriate, the original texts should be accompanied by a translation. For further details or instructions, please consult the Editor. Greek must be supplied in an accurate form, with all diacritics in place. Please indicate whether the Greek is in a Unicode font or if not, the software used to input it (e.g. GreekKeys, Linguist’s Software) to facilitate file conversion. 5. For citations of Greek and Latin authors, house style should be followed. This can be checked in any recent issue of OSAP with the help of the Index Locorum. The most exact reference possible should normally be employed, especially if a text is quoted or discussed in detail: for example, line references for Plato (not just Stephanus page and letter) and Aristotle (not just Bekker page and column).

376

Notes for Contributors

6. In references to books, the first time the book is referred to give the initial(s) and surname of the author (first names are not usually required), and the place and date of publication; where you are abbreviating the title in subsequent citations, give the abbreviation in square brackets, thus: T. Brickhouse and N. Smith, Socrates on Trial [Trial ] (Princeton, 1981), 91–4. Give the volume-number and date of periodicals, and include the full page-extent of articles (including chapters of books): D. W. Graham, ‘Symmetry in the Empedoclean Cycle’ [‘Symmetry’], Classical Quarterly, ns 38 (1988), 297–312 at 301–4. G. Vlastos, ‘A Metaphysical Paradox’ [‘Metaphysical’], in G. Vlastos, Platonic Studies, 2nd edn. (Princeton, 1981), 43–57 at 52. Where the same book or article is referred to on subsequent occasions, usually the most convenient style will be an abbreviated ­reference: Brickhouse and Smith, Trial, 28–9. Do not use the author-and-date style of reference. 7. Authors are asked to supply in addition, at the end of the article, a full list of the bibliographical entries cited, alphabetically ordered by (first) author’s surname. Except that the author’s surname should come first, these entries should be identical in form to the first occurrence of each in the article, including where appropriate the indication of abbreviated title: Graham, D. W., ‘Symmetry in the Empedoclean Cycle’ [‘Symmetry’], Classical Quarterly, ns 38 (1988), 297­­–312. 8. If there are any unusual conventions contributors are encouraged to include a covering note for the copy-editor and/or printer. Please say whether you are using single and double quotation marks for different purposes (otherwise the Press will employ its standard single quotation marks throughout, using double only for quotations within quotations). 9. Authors should send a copy of the final version of their paper in electronic form by attachment to an e-mail. The final version should be provided as a Microsoft Word file (or Rich Text Format file), accompanied by a note of the system (not just the font) used for producing Greek characters (see point 4 above). This file must be accompanied by a second file, a copy in PDF format of the final version, to which it must correspond exactly. If necessary, arrangements for alternative means of submission may be made with the Editor. With final submission authors should also send, in a separate file, a brief abstract and a list of approximately ten keywords.