Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy 0198866739, 9780198866732

Citation preview

Deterininisin, Freedoin,

and Moral Responsibility Essays in Ancient Philosophy SUSANNE BOBZI EN

OXFORD UN I VERS ITY P R ESS

OXFORD UNIVERSITY PRESS

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, scholarsh ip, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Susanne Bobzien 202 1

The moral rights of the author have been asserted First Edition published in 2021 Impression: I 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: 20209460 11 ISBN 978-0-19- 886673-2 DOI: 10.1093/oso/9780198866732.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CRO 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.

Contents

Preface Notes on Original Publication Abbreviations

ix

xi xiii

Introduction 1. The Inadvertent Conception and Late Birth of the Free-Will

13

Problem: Aristotle to Alexander of Aphrodisias

2. Choice and Moral Responsibility in Nicomachean Ethics 3.1-5

51

3. Aristotle's Nicomachean Ethics ll 13b7-8 and Free Choice

77

4. Found in Translation: Aristotle's Nicomachean Ethics 3.5, 1113 b7-8 and Its Reception 5. Moral Responsibility and Moral Development in Epicurus Philosophy

93 ,

128

6. Did Epicurus Discover the Free-Will Problem?

152

7. Stoic Conceptions of Freedom and Their Relation to Ethics: Early Stoics, Epictetus, Late Stoics

194

8. Early Stoic Determinism: The Merging of Teleology and Universal Causation

217

9. Chrysippus' Theory of Causes

253

Bibliography Index Locorum General Index

291 301 313

Preface The nine essays in this volume were written over a period of close to twenty years. They were not envisaged, let alone planned, to be published together as a book, but, with hindsight, they hold together rather well and adequately present my overall view and interpretations of determinism, freedom, and moral responsibility in antiquity, in both its philosophical and historical aspects. The nine chapters in this volume comprise all the essays I have written on these topics to date, with the exception of book reviews, encyclopedia articles, replies to papers, and an afterword. The chapters are organized in four parts, mostly according to the chronology of the ancient philosophers whose positions are their subject matter, from Aristotle via Epicurus to the Stoics, with an overarching first chapter that covers the period from Aristotle to Alexander of Aphrodisias, the early third-century  Peripatetic commentator. Within each part, the order is from the general to the specific, from the easily read to detailed scholarly analysis. Chapters 2, 5, and 7 (and 8) can be used as an introduction to the views of Aristotle, Epicurus, and the Stoics, respectively. All chapters can be studied independently of each other. This said, reading the first section, or the first four sections, of Chapter 1 may facilitate the understanding of the later chapters. Chapter 1 also provides a useful introduction to my philosophical and historical approach to the philosophers discussed in the volume. No knowledge of Greek or Latin is required for reading the main text of the chapters. Greek and Latin passages are translated. The texts in their original language are placed in footnotes for consultation, and all Greek expressions are transliterated in the main text, except in rare instances where matters of translation are themselves the topic, as is the case—solely—in Chapter 4. Some minor changes have been made to the essays, mostly for the convenience of the reader. Thus, I have added Greek and Latin texts in starred footnotes in a number of places where only the English was originally given. Conversely, I have added English translations in a number of places where only the Greek or Latin original was present. I have also translated some untranslated contemporary language elements into English. Typos and similar minor errors have been corrected, and, in places—especially in the earlier essays—I have made the English more idiomatic. Where on rereading a phrase or sentence appeared to me ambiguous, I disambiguated. The cryptically short last two paragraphs of Chapter 9 have been expanded. Chapter 8, section 8.4, which had to be shortened for its first publication due to a word limit, has been restored roughly to its original

x



length. Throughout the volume, some elements of presentation, such as layout, numbering, citations, and the like have been unified. As far as possible, the original footnote numbers have been retained. Where Greek or Latin texts etc. have been added in new footnotes, these footnotes are starred rather than numbered. Square brackets in a translation indicate a phrase supplied by context. Angled brackets in text and translation are used to indicate a textual emendation. A list of abbreviations used for the ancient texts is given on pages xiii–xv. Since 1997, when the first of the essays in this volume was originally published, there has been a vast number of publications on questions of determinism, freedom, and moral responsibility in antiquity. This is not the place to list them all, nor will I here respond to individual papers and books. (It is gratifying to see that the majority of responses to the essays collected here have agreed with my main theses.) In this volume, I restrict myself to mentioning a few recent booklength studies and collections that will provide the reader with additional information about the developments in late ancient philosophy and in areas that are covered neither in these essays nor in my book Determinism and Freedom in Stoic Philosophy. Precursors to eight of the papers collected here were first presented on the occasion of a lecture I had been invited to give or as a conference contribution, and I am grateful to those who issued the invitations. (The ninth, Chapter 9, is a— substantial—offshoot from my book Determinism.) Those who helped form the final products with probing questions and learned comments at these occasions, and sometimes before or after, on walks, on trains, in cafes, in emails, are acknowledged in each paper, although I have doubtless overlooked some people here and there and collectively thank them at this point. For this collection as a whole, I am first of all most grateful to Peter Momtchiloff, who encouraged publication of these essays as a volume and who has provided vital assistance at various points during its coming into being. I am also indebted to the eminently competent Paul Scade, who was a great help in putting these essays—which displayed a large variety of layouts and editing conventions—into a consistent form, and assisted with various issues ranging from punctuation to idiomaticity. Thanks go also to Chiara Martini for most adeptly and reliably helping in getting the Greek transliterated or untransliterated as the case required, as well as for assistance with restoring texts from PDFs and for proofreading against the original papers. Nathan A. Schatz has been a steady source of sustenance, a paragon of patience, and much more.

Notes on Original Publication The essays of which the nine chapters in this volume are revised versions were first published in the following places: 1. Phronesis 43 (1998), 133–75. 2. R. Polansky (ed.), The Cambridge Companion to Aristotle’s Nicomachean Ethics (Cambridge: Cambridge University Press, 2014), 81–109. © Ronald Polansky 2014, published by Cambridge University Press. 3. P. Destrée, R. Salles, and M. Zingano (eds), What Is up to Us? Studies on Causality and Responsibility in Ancient Philosophy (Sankt Augustin: Academia Verlag, 2014). 4. Oxford Studies in Ancient Philosophy 45 (2013), 103–48. 5. B. Reis and S. Haffmans (eds), The Virtuous Life in Greek Ethics (Cambridge: Cambridge University Press, 2006), 206–29. © Cambridge University Press 2006. 6. Oxford Studies in Ancient Philosophy 19 (2000), 287–337. 7. R. Sorabji (ed.), Aristotle and After (London: Institute of Classical Studies, 1997), 71–89. By permission of Oxford University Press on behalf of the Institute of Classical Studies. 8. Revue de métaphysique et de morale 48(4) (2005), 489–516. 9. K. Ierodiakonou (ed.), Topics in Stoic Philosophy (Oxford: Oxford University Press, 1999), 196–242. The author is grateful to the publishing houses who gave permission to use material that is in their copyright.

Abbreviations Aetius Plac. Alcinuous, Didasc. Alex. An. Pr. Alex. Fat. Alex. Mant. Alex. Mixt. Alex. Quaest. Alex. Top. Ammonius, Int. Anon. EN Apuleius, Plat. Aquinas, Sent.Eth. Arist. An. Pr. Arist. Athen.Const. Arist. DA Arist. EE Arist. EN Arist. Int. Arist. Metaph. Arist. Phys. Arist. Rhet. Arist. Soph.Elen. Arist. Top. Aspasius, EN Augustine, De civ. D. Augustine, Lib. arb. Boethius, Int. Calc. Tim. Cic. Acad. Cic. Div. Cic. Fat. Cic. Fin. Cic. Mur. Cic. Nat. D. Cic. Top. Clement, Strom. DL Epic. Ep. Hdt. Epic. Ep. Men.

Aetius De Placita Philosophorum Alcinous, Didascalicos Alexander of Aphrodisias, Commentary on Aristotle’s Prior Analytics Alexander of Aphrodisias, On Fate Alexander of Aphrodisias, Mantissa Alexander of Aphrodisias, On Mixture Alexander of Aphrodisias, Quaestiones Alexander of Aphrodisias, Commentary on Aristotle’s Topics Ammonius, Commentary on Aristotle’s de Interpretatione anonymous, Commentary on Aristotle’s Nicomachean Ethics Apuleius, De Platone et dogmate eius Thomas Aquinas, Sententia libri Ethicorum Aristotle, Prior Analytics Aristotle, Constitution of the Athenians Aristotle, De Anima Aristotle, Eudemian Ethics Aristotle, Nicomachean Ethics Aristotle, De Interpretatione Aristotle, Metaphysics Aristotle, Physics Aristotle, Rhetoric Aristotle, Sophistical Refutations Aristotle, Topics Aspasius, Commentary on Aristotle’s Nicomachean Ethics Augustine of Hippo, De civitate Dei Augustine of Hippo, De libero arbitrio Boethius, Commentary on Aristotle’s de Interpretatione Calcidius, On Plato’s Timaeus Cicero, Academica Cicero, On Divination Cicero, On Fate Cicero, On Ends Cicero, Pro Murena Cicero, On the Nature of the Gods Cicero, Topics Clement of Alexandria, Stromata Diogenes Laertius, Lives of Eminent Philosophers Epicurus, Letter to Herodotus Epicurus, Letter to Menoeceus

xiv



Epic. Ep. Pyth. Epic. Nat. Epic. Sent. Epict. Diss. Epict. Ench. Epiphanius, Adv. Haer. Eusebius, Praep. Ev. Galen, Alim. fac. Galen, Caus. cont. Galen Caus. puls. Galen, PHP Galen, Plen. Galen, Syn. puls. [Gal.] Def. med. Gellius, NA Hippolytus, Ref. Josephus, Ant. Justin, Apol. Lucr. Marc. Aur. Nemesius, Nat. hom. Olymp. Phd. Origen, Cels. Origen, Evang. Ioannis Origen, Orat. Origen, Princ. Philo, Prob. Philodemus, Ind. Stoic. Philodemus, Ir. Philodemus, Piet. Philodemus, Sign. Plot., Enn. Plut. Comm. Not. Plut. Quaest. Conv. Plut. Quomodo adul. Plut. Soll. An. Plut. Stoic. Rep. [Plut.] Epit. [Plut.] Fat. Porph. Marc. SE M SE PH Sen. Ben. Sen. Ep. Sen. QNat. Sen. Vit. Beat.

Epicurus, Letter to Pythocles Epicurus, On Nature Epicurus, Vatican Sayings Epictetus, Discourses Epictetus, Handbook Epiphanius, Adversus Haereses Eusebius, Praeparatio Envangelica Galen, De alimentorum facultatibus Galen, De Causis Contentivus Galen, De Causis Pulsuum Galen, On the Doctrines of Hippocrates and Plato Galen, De Plenitudine Galen, Synopsis Librorum Suorum, Sexdecim, de Pulsibus [Galen], Medical Definitions Aulus Gellius, Attic Nights Hippolytus of Rome, Refutation of All Heresies Josephus, Antiquities of the Jews Justin, Apology Lucretius, On the Nature of Things Marcus Aurelius, Meditation Nemesius, On the Nature of Man Olympiodorus, Commentary on Plato’s Phaedo Origen, Against Celsus Origen, Commentary on the Gospel of John Origen, De Oratione Origen, De Principiis Philo of Alexandria, Quod omnis probus liber sit Philodemus, Index Stoicorum Philodemus, De Ira Philodemus, De Pietate Philodemus, De Signis Plotinus, Enneads Plutarch, Against the Stoics on Common Conceptions Plutarch, Quaestiones convivales Plutarch, Quomodo adulescens poetas audire debeat Plutarch, De sollertia animalium Plutarch, On Stoic Self-Contradictions [Plutarch], Epitome of the Opinions of the Philosophers [Plutarch], De Fato Porphyry, Letter to Marcella Sextus Empiricus, Against the Mathematicians Sextus Empiricus, Outlines of Pyrrhonism Seneca, On Benefits Seneca, Letters Seneca, Natural Questions Seneca, On the Happy Life

 Simplicius, Cat. Simplicius, Ench. Stobaeus, Ecl. Syrianus, Met. Tacitus, Ann. Tatian, Ad. Gr. Tertullian, An. Tertullian, Apol.

Simplicius, Commentary on Aristotle’s Categories Simplicius, Commentary on the Encheiridion Stobaeus, Eclogues Syrianus, Commentary on Aristotle’s Metaphysics Tacitus, Annals Tatian, Oratio ad Graecos Tertullian, De Anima Tertullian, Apology

xv

Introduction The cluster of problems around freedom, determinism, and moral responsibility is one of those themes in philosophy that are fascinating in both their complexity and their seemingly direct relevance to human life. Historians of ideas often assume that in Western philosophy this cluster of problems was the subject of an ongoing discourse from antiquity to the present day. This is, however, an illusion. Much of my research on ancient theories of determinism and freedom is devoted to showing that what commonly counts as this problem cluster today (often labelled as ‘the problem of free will and determinism’) is noticeably distinct from the issues that the ancients discussed—at least prior to the second century . It is true that one main component of the ancient discussion concerned the question of how moral accountability can be consistently combined with certain causal factors that impact human behaviour. However, it is not true that the ancient problems involved the questions of the compatibility of causal determinism with either our ability to do otherwise or a human faculty of a free will. Instead, we encounter questions about human autonomous agency and its compatibility with preceding causes, external or internal; with external impediments; with divine predetermination and theological questions; with physical theories like atomism and continuum theory, and with sciences more generally; with elements that determine character development from childhood—nature and nurture; with epistemic features such as ignorance of circumstances; with necessity and modal theories generally; with folk theories of fatalism; and also with questions of how human autonomous agency is related to moral development, to virtue and wisdom, to blame and praise. All these questions were discussed without reference to freedom to do otherwise or a faculty of the will—at least in Classical and Hellenistic philosophy. This volume of essays considers all of these questions to some extent. Chapter 1 follows up the question of when the philosophical problems of the compatibility of universal causal determinism with freedom to do otherwise and with free will emerged. To that end, I work through an array of ancient philosophical sources from Aristotle to Alexander of Aphrodisias, the early thirdcentury  Peripatetic and commentator on Aristotle. I start with a number of philosophical distinctions that provide a set of tools for the enquiry (and that are also expedient for some of the later chapters in this collection). Most important are the distinctions between (i) indeterminist freedom (freedom to do otherwise, freedom of decision, and freedom of the will); (ii) freedom from predetermination;

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0001

2

, ,   

and (iii) freedom that can coexist with full causal determination (freedom from external compulsion, from external and internal causal determination). I subsume under the name of the ‘free-will problem’ the three problems of the compatibility of determinism with the three kinds of indeterminist freedom. The results of the investigation are significant. It turns out that what is often described in terms of the ‘discovery’ of the problem of causal determinism and freedom of decision in Greek philosophy in fact arises from an accidental combination and mixing up of Aristotelian and Stoic thought in later antiquity. More precisely, it appears that the free-will problem is inadvertently conceived as the result of a (mis-)interpretation of Aristotle’s philosophy of deliberate choice and action—as he introduces it in his ethics—in the light of the early Stoic theory of determinism and moral responsibility. The (con-)fusion originates with the beginnings of Aristotle scholarship, at the latest in the early second century , possibly a little before that, and undergoes several metamorphoses. A range of ideas are absorbed at various points into this development: from the late Stoics, especially Epictetus and possibly Musonius; from some anonymous Middle Platonists; from some later Peripatetics; and, it seems, also from some early Christian writers. Eventually, the development arrives at a full notion of freedom of decision and an exposition of the ‘free-will problem’ in Alexander of Aphrodisias’ On Fate and in the Mantissa ascribed to him. The late birth of the free-will problem is thus dated in the second century . Chapter 1 provides a framework for the remainder of the volume, both historically, by covering the entire development from Aristotle to Alexander, and conceptually, in that it introduces a number of basic notions that are made use of repeatedly in later chapters. There are some pre-Aristotelian texts that consider the relation between moral accountability, on the one hand, and necessitation, fate, and external or internal force, on the other. These include—prominently—Gorgias’ Encomium of Helen and various passages in Plato’s oeuvre, in particular in the Republic. None of these texts displays any awareness of the contemporary problem of causal (event) determinism and freedom to do otherwise, or even of either of these concepts. For moral responsibility in the Encomium, Rachel Barney (Barney 2016, §2) is a good source of information. For Plato, Hecht 2014 is useful: Hecht comes to the conclusion, explicitly stated, that Plato did not have a notion of freedom to do otherwise, and he argues convincingly that Plato’s freedom is freedom from constraint (Hecht 2014, 204–6). Serious sustained philosophical theory building and analysis of some of the problems of determinism and freedom starts with Aristotle (see Chapter 1, section 4). Aristotle is often considered to be the first philosopher who discussed the free-will problem. In particular, he is thought to connect free will to moral responsibility in his Nicomachean Ethics and Eudemian Ethics. Three essays in this volume are dedicated to moral responsibility and the question of free will in Aristotle.



3

Nicomachean Ethics Book 3, chapters 1–5 (EN 3.1–5), has been the topic of innumerable publications that connect this part of the Ethics with free will, since in it Aristotle considers the topics of responsibility, choice, and determination of character. The question of the overall purpose of this part of the Nicomachean Ethics is, however, generally left unresolved. Chapter 2 provides an answer to this question and in doing so produces a unified interpretation of EN 3.1–5. This interpretation makes it clear that Aristotle’s theory of responsibility is not based on a notion of free decision or free will, and that there is no notion of free will in the theory at all. The primary focus of the chapter is an exposition of Aristotle’s theory of what makes us responsible for our actions and our character. In this context, I remark at the relevant places on issues of freedom-to-do-otherwise, free choice, and free will, and how they are wrongly attributed to Aristotle. The chapter starts with some preliminary observations about praise, blame, and responsibility. It then sets out in detail how all the key notions of EN 3.1–5 are interrelated. In this way the carefully structured complexity of Aristotle’s theory of action, character development, and moral responsibility becomes apparent. The setting out of these interconnections makes it then possible to provide a comprehensive interpretation of the purpose of the passage. It emerges that its primary purpose is to explain how agents are responsible for their actions not just insofar as these are actions of this kind or that, but also insofar as they are noble or base: this is the step from responsibility generally to moral responsibility. Agents are responsible for their actions qua noble or base because, typically via choice, their character dispositions are a causal factor of those actions. The chapter illustrates the various different ways in which Aristotle believes that agents can be causes of their actions by reference to his four basic types of agents: the virtuous, the vicious, the strongwilled, and the weak-willed. An important second purpose of EN 3.1–5 is to explain how agents can be held responsible for certain consequences of their actions, in particular for their character dispositions insofar as these are noble or base, that is, virtues or vices. Again, it is shown that no notion of free will is involved in Aristotle’s explanation. These two purposes are not the only ones that Aristotle pursues in the passage. They are, however, the purposes Aristotle himself indicates in its first sentence and summarizes in its last paragraph, and those that give the passage a systematic unity. Aristotle thus sets the scene with the discussion of several philosophical problems: how moral responsibility and external force relate to action; how ignorance of certain factors in action affects moral responsibility; and what effect character determination by nature and education (or the lack thereof ) have on moral responsibility. There is, above all, one famous sentence in the Nicomachean Ethics Book 3, chapter 5, about that which is in our power or up to us (eph’hēmin), that has markedly shaped scholarly and popular opinion alike with regard to Aristotle’s— presumed—theory of free will, and which has repeatedly been used to support the assumption that Aristotle discussed free choice or free will. This goes against the

4

, ,   

grain of what I am convinced to be the truth about Aristotle and ancient Classical and Hellenistic philosophy. So in Chapter 3 of this volume (which falls into the category of my work that I call ‘setting things right once and for all—I hope’) I offer an in-depth textual analysis of that sentence and its linguistic context, with a view to free will and free choice. The result is that there is no acceptable interpretation in which the sentence provides any evidence that Aristotle discussed freedom of choice or freedom of the will. Most assertions that it does tend to be based on an odd mistranslation of the Greek. Thus the sentence that is sometimes adduced as the main piece of evidence for the claim that Aristotle was an indeterminist with respect to both choosing (prohairesis) and acting (praxeis, prattein) turns out to be no evidence for this claim at all. The fairly widespread mistranslation noted in Chapter 3, which has led generations of philosophers to endorse the incorrect view that the Nicomachean Ethics shows that Aristotle had a notion of free will or a notion of free choice, is most peculiar. Its peculiarity had mystified me since I was teaching the Nicomachean Ethics for ‘Finals’ (the Classics final examination) in Oxford in the early 1990s. Its stubborn recurrence even in the late twentieth and early twenty-first centuries strongly suggested that there had to be more to it than just a simple error. So I decided to follow the journey of this one sentence from when it left Aristotle’s ‘pen’ to the present time. This led to Chapter 4 (which falls in the category of my work that I call ‘academic detective stories’). It is a companion essay to Chapter 3 which investigates the convoluted history of the reception of Aristotle’s Nicomachean Ethics 3.5, lines 1113b7–8. It becomes apparent that this sentence has not only markedly shaped scholarly and general opinion alike with regard to Aristotle’s theory of free will, but it has also taken on a curious life of its own. The short first part of the essay examines the two lines and their immediate context, this time with a focus on the text and its meaning rather than the philosophical questions it raises. (There is, nonetheless, some unavoidable overlap in this first part with Chapter 3.)¹ The much more extensive second part of the chapter explores the reception of the sentence. This includes its presence in present-day popular culture, partially for entertainment but also because it is a symptom of how, in the era of ‘virtual wisdom’ and social media, a philosophically significant blunder, once it has gained a foothold, is almost impossible to reverse in the common consciousness. More seriously, the investigation follows a number of unforeseen and exhilarating twists and turns along the route from Aristotle to the contemporary engagement with ancient philosophy via later ancient, Byzantine, Arabic, Latin Medieval, Renaissance, and Victorian scholarship. The error is revealed to be the result of a mismatch in textual transmission—at the meeting point of Latin, Arabic, and Greek manuscripts—between Aristotle’s text and a ¹ It is Melissa Lane to whom I owe the insight that what I presented as one paper at Princeton in 2014 were actually two papers, companions to each other, but best divided by their separate purposes.



5

commentary on that text. Once in existence, this error influenced the understanding of Aristotle’s view on free will by leading to reinterpretations of the—faulty—textual reading from a nineteenth-century perspective and in the process imbuing it with a good dose of Christian dogma. And for reasons that the reader is left to puzzle out for themselves, even in the scholarly exacting and philosophically unforgiving climate of analytic ancient philosophy, the error lives on. Chapters 5 to 9 are devoted to issues of determinism, freedom, and moral philosophy in Hellenistic philosophy. One of the main obstacles any scholar of that period encounters is the lack of surviving sources. (The reader should not kid themselves about Aristotle in this respect, though. Only about a fifth of his work survives, and a good bit of this not fully worked out.) We have only a small percentage of what the Epicureans and Stoics wrote, and most of that in secondary, tertiary, or quaternary sources, including many that either lack the background for understanding the original texts or have an intention to distort them or take them out of context, so as to be better able to criticize them. This presents a number of methodological challenges. The next philosopher after Aristotle who has been credited with either discovering or introducing the free-will problem is Epicurus. We are fortunate to have a summary of his philosophical theory in his letters to Menoeceus, Pythocles, and Herodotus, and in Lucretius’ poem On Nature. However, most of Epicurus’ works are lost, an exception being the Papyrus with Book 25 of his On Nature, which is highly relevant to the topics of this volume but in many places impenetrable, largely because it is in such lacunose form. Epicurus and his followers are atomists and, as a result, the Epicurean theory and discussion of issues of determinism, freedom, causation, and moral responsibility are all grounded in Epicurus’ atomism, including his introduction of the notorious swerves. Chapters 5 and 6 in this volume are devoted to Epicurus and the Epicureans. Chapter 5 discusses moral responsibility and moral development, leaving for Chapter 6 questions of the role of the swerve in Epicurean discussion of freedom and determinism. The reliance of Epicurus’ doctrine on atomism requires him to provide detailed explanations of how causation, freedom, moral development, moral responsibility, and their relations to physics and ethics are to be understood at the micro level in terms of atoms. The result is in many respects a surprisingly modern theory of human psychological and moral development. It is combined with an equally fascinating forward-looking notion of responsibility, which is not concerned with praising or blaming individuals for past actions, but lays the grounds on which everyone is to use the past and present in order to improve their future behaviour. This theory has much to tell to contemporary philosophers and moral psychologists. It rejects the well-trodden backwards-looking perspective of moral responsibility on non-virtuous past behaviour as a basis for justified blame and punishment: in its place it puts a fresh future-orientated perspective in

6

, ,   

which the fact that adult human beings are morally responsible agents is focused on primarily as a basis for future self-improvement of ourselves and others. There is sufficient evidence to show that Epicurus had a notion of moral responsibility for actions that is based on the agent’s causal responsibility, as opposed to the agent’s ability to act or choose otherwise. He considered it a necessary condition for praising or blaming an agent for an action that it was the agent and not something else that brought the action about. As a result, the central question of moral responsibility is whether the agent was the, or a, cause of the action, or whether the agent was forced to act by something else. The main factors for ‘something else’ that Epicurus considers are what we would call nature (one’s hereditary make-up) and nurture (upbringing and teaching). Epicurus holds the view that actions can be attributed to agents from the point onwards when it is in their actions that the agent, qua moral being, manifests themself as a rational being that has beliefs that go beyond their (unquestioned) hereditary and environmentally induced beliefs. As a result, unlike in theories that tie moral responsibility to the ability to do otherwise, the question of moral development becomes all-important. Consequently, the chapter collects and discusses the evidence for Epicurus’ views on moral development, asking the questions of (i) how humans become moral beings and (ii) how humans can become better persons. Epicurus envisaged a complex web of hereditary and environmental factors as shaping the moral aspect of a human being. It appears that he explained this development both at the macro level of human behaviour and in detail with his theory of atoms. Of particular interest are his views concerning how, atomically, we acquire beliefs and how we can change our beliefs, in particular how we can replace false beliefs with true ones. This atomic-level explanation sheds important light on Epicurean ethics, both the importance it places on teaching and the importance it places on lifelong learning and on the use of repetitive exercises for learning. One significant result is that Epicurean ethics does not have the function of developing or justifying a moral system that allows for the effective allocation of praise and blame. Rather the function of ethics is to give everyone a chance to morally improve. The most frequently discussed Epicurean issue regarding human freedom is that of the role played by the swerves in their theory. There is no first-hand testimony from Epicurus on this point. Swerves are atomic minimal random motions. There is sufficient indirect evidence that Epicurus introduced the swerves, among other things, as a necessary condition for the existence of free volitions. Beyond that, the function of the swerves for freedom is something of a mystery. On the face of it, the introduction of random motions in order to secure the freedom of human volitions seems dubious rather than enlightening. If the motions are random, how can the volitions be the agent’s volitions? A multitude of suggestions have been offered as to how exactly the swerves could accomplish



7

this. When this essay was written, there were two main camps: the prevalent viewpoint required at least one swerve in every free volition, usually to warrant that the agent has two-sided freedom (Chapter 1, section 2); the other camp required at least one swerve in the development of human agents that makes them agents whose volitions are neither internally nor externally forced. These factions are roughly aligned with the views that Epicurus’ notion of free action required a faculty of two-sided freedom and that Epicurus’ notion of free action required autonomous rational agents. Chapter 6 aims to settle for good at least some of the questions regarding the role of the swerves and Epicurus’ notion of freedom. The chapter traces three interlocking questions: (i) Did Epicurus deal with the kind of free-will problem with which he is traditionally associated, i.e. did he discuss free choice or moral responsibility grounded on free choice? (ii) To save such free choice, did Epicurus stipulate that the swerves were involved in individual decision processes? (iii) If not free choice, what then was the problem that the swerves were meant to solve? There is no direct evidence for a positive answer to (i) or (ii). Rather, the evidence suggests that, for Epicurus, actions are fully determined by the agent’s mental disposition at the outset of their action, and that Epicurean moral responsibility presupposes not free choice but that a person is unforced and causally responsible for their action. Regarding (iii), this means that on the basis of these results and the passage that connects the swerves to free volition, it becomes plausible that Epicurus’ problem was this: how can human agency be non-necessary in an atomistically explained ordered universe? The context of his argument is provided by other atomistic theories. It appears that Epicurus’ swerves were intended to explain the non-necessity of agency without undermining Epicurus’ atomistic explanation of the order in the universe, and that they did so by making the mental dispositions of adults nonnecessary. Two alternative conjectures are offered to account for how the swerves could have achieved this. Epicurus also seems to have connected logic with determinism and argued that a logical principle similar to a semantic version of the Law of Excluded Middle, if accepted, would lead to the necessity and fatedness of all events, and that therefore it should be rejected. This is discussed in Bobzien 1998, 75–86.² The last three chapters are given over to the Stoics. The first concerns freedom and responsibility, the second teleology and causation, and the last is a detailed study of the theory of causes advanced by Chrysippus, third head of the Stoa, who refined and elaborated the early Stoic theory of determinism and defended it against objections by other schools. Chapter 7 provides an introduction to the Stoic theories of freedom and what is up to us from the beginning to the end of Stoic philosophy in antiquity.

² On this topic see now also Bown 2016.

8

, ,   

It investigates how, from the early Stoics via Epictetus to the late Stoics of the second century , freedom and responsibility are connected in Stoic philosophy with ethics. In this context, the most important distinction is that between the Stoic notion of that which depends on us (eph’hēmin) and their notion of freedom (eleutheria). That which depends on us is always connected with human action and intention in Stoic philosophy, and is located within the sphere of Stoic psychology. For every human being, there are things that are up to them. Freedom, by contrast, is a notion that originates from political theory, is contrasted with slavery, and is, from there, then introduced into Stoic ethics. For the Stoics, freedom is a character disposition—more specifically a virtue—and thus can be attributed only to wise people. The confusion of these two quite distinct concepts and their roles in Stoic philosophy has wreaked much havoc in twentieth-century literature on Stoic freedom (which the essay untangles and invalidates in the process). The chapter traces the relations between that which depends on us, freedom, and ethics through three stages of Stoic philosophy. (i) For the early Stoics, that which depends on us is not a philosophical term but is used as a natural language expression which denotes our behaviour, including our actions and our refrainings from action, as well as our assent and intention that lead to the actions. Freedom (eleutheria) is only mentioned in the context of politics and ethics, and rarely so even there. It appears to have referred to a virtuous person’s power to do what they want to do, which is the same as what they should do. No source connects the two notions. (ii) For Epictetus, both freedom and that which depends on us are notions that are central to his philosophy. We witness a change of focus from that which depends on us (to eph’hēmin) to the things that depend on us (ta eph’hēmin); the range of these things is more restricted, comprising our assent, intention, and refraining from action. Unlike actions, these three are internal to agents. Epictetus discusses the things that depend on us mainly in the context of an already established moral theory. Epictetus’ notion of freedom is similar to that of the early Stoics, but it typically appears in company with a group of terms connected with tranquillity of the mind. Importantly, in his ethics Epictetus connects the things that are up to us and freedom, something not known for the early Stoics. The discussion has progressed to an emphasis on internal acts (assent, intention) and has been moved to ethics. This notwithstanding, Epictetus’ theory seems, unsurprisingly, to be fully compatible with early Stoic determinism. (iii) In the second century  the debate shifts further. The late Stoics are faced with Middle Platonists and some Peripatetics proposing roughly libertarian two-sided notions of what is up to us that corrode the earlier discrepancy between what is up to us and freedom. Adapting to the new situation, the late Stoics developed new arguments, adjusted to the new debate, and defended their own theory as superior. This late Stoic theory is still in all major points compatible with early Stoic determinism.



9

Chapter 7 complements Chapter 1 in several ways. Rather than focusing on the development of the free-will problem, it traces the development of the associated debates in which the ancient Stoics were involved through the centuries, showing how topics and interests shifted and, relatedly, how the conceptual framework within which the Stoics elaborated and defended their theories altered. As Epicurean theory of freedom and responsibility is deeply influenced by and based on their physics, so too is Stoic theory. While the Epicureans are atomists, the Stoics are continuum theorists. There are no smallest parts of matter, place, or time, and whatever happens is caused by efficient causes that at least partially coexist with their effect. Where the Epicureans have gods who are external to the world and who do not care about human affairs, the Stoics have a divinity that is immanent to the universe and that governs every element of what happens in it— past, present, and future. As a consequence, the distinctive feature of early Stoic determinism is its combination of teleological and apparently mechanical elements. More broadly, the question they face is how to combine theology and science. Chapter 8 pursues the essential question of how teleological elements and efficient causation were merged in Stoic determinism. At the very beginning of their school, the Stoics do not present their determinism as causal determinism, that is, in terms of universal efficient causation. Rather, it is introduced in Stoic cosmology with a strong teleological element in the context of a theory of natural motions. It is a theory built on a fundamental distinction between a global and an inner-worldly or ‘local’ perspective on events, in which one active principle, variously called Nature (phusis), Cause (Aitia), and Reason (logos), is the global principle that determines all inner-worldly ‘local’ events. Yet, in order to explain his determinism, Chrysippus also employed his conception of efficient causality as that which connects inner-worldly causes and their effects. It can be shown that he maintained a universal causal determinism in the modern sense of the term, if presented in a slightly different way, using counterfactual relations rather than causal laws. The Stoic notions of efficient causes as corporeal and responsible for both sustenance and change, and of effects as incorporeal and predicates, are indispensable for a full understanding of this theory. The teleological and seemingly mechanical elements of early Stoic determinism were brought together in Chrysippus’ conception of fate (heimarmenē). It is here that Chrysippus places elements of rationality in every cause. (The notion of fate obtained its pivotal role in ancient theories of determinism and freedom only as a response to the Stoic theory.) In our surviving texts, the exact workings of this placement of teleological elements into apparently mechanical causes are presented by means of biological and psychological analogies, a method often employed by the early Stoics. The chapter pursues these analogies and makes suggestions about how they might translate into Stoic cosmology. It emerges that the Stoic theory of Nature as world seed and their theory of Nature as world soul and rational agent, which is contrasted with the individual seeds, souls, and agents within Nature, offer a

10

, ,   

fascinating solution to the question of how science and theology, in particular the issue of divine predetermination, can be combined consistently within one theory of the world. In the early Stoic view, we do not face a choice between theological explanation and scientific explanation of the world. Rather, these are two complementary explanations of the same thing. The last chapter, Chapter 9, is a companion piece to Chapter 6 of my book Determinism and Freedom in Stoic Philosophy (Bobzien 1998), and it can be read in conjunction with it, to good effect. But it also stands on its own. Both chapters are concerned with the interpretation of a long and famous passage in Cicero’s De Fato on Chrysippus’ defence of the compatibility of fate in the forms of causal determinism and moral responsibility. The chapter in Determinism and Freedom investigates how the passage, and especially the cylinder analogy, increases our understanding of Stoic compatibilism, describes the specific problem of moral responsibility and determinism that Chrysippus and his opponents encountered, and shows how it differs from problems that are today subsumed under labels such as ‘the free-will problem’. Chapter 9 of the present volume has as its main purpose the laying to rest of a recalcitrant misconstruction of a distinction between two kinds of causes (auxiliary causes and principal causes) that Chrysippus introduces in the context of defending his view. (Another essay in the category of my work that I call ‘setting things right once and for all—I hope’.) This misinterpretation takes the two types of causes to be cooperating in one instance of causation. However, painstaking textual analysis, together with the use of some parallel passages, makes it clear that Chrysippus’ distinction is instead one of the characteristics of efficient antecedent causes and is situated in the context of a debate about the predetermination of human action, possibly among other things. In principle, such antecedent causes could be thought of as principal and self-sufficient causes that necessitate their effect. This is how Chrysippus’ opponents in the debate characterized them in their argument against the Stoic theory of fate. Alternatively, antecedent causes can be thought of as auxiliary causes that start or trigger an instance of causation and that, in order to produce the effect, must cooperate with another cause, namely the object at which the effect occurs. In this case, the antecedent causes are neither main causes of the effect nor do they necessitate the effect, since whether the effect occurs or what effect occurs depends crucially on the properties of the object at which it occurs, which functions as the second and more substantial cause of the effect. This alternative is Chrysippus’ view, and the chapter shows that the point of Chrysippus’ distinction of the two types of causes is to mark it out against his opponents’ assumption that the antecedent cause is the sole cause of the effect and necessitates it. Thus Chrysippus does not assume that causes of the first type exist within the world, and it seems that such causes played no role in Stoic theory of efficient causation. The above-mentioned recalcitrant misinterpretation of the text can thus be laid to rest. The chapter pursues a second purpose, which is to set out and systematize the



11

evidence we have for an early Stoic, and in particular Chrysippean, theory of causes. The most fundamental distinction of causes is that between causes of change and causes of states. The active principle or pneuma in an object is the cause of its present properties, and in this sense the cause of the state it is in and of its function. Another distinction is that between causes whose presence at least partially precedes that of their effects, and causes that are concurrent with their effects. Where there is an antecedent cause, there always seems to be cooperation of these two kinds. The overall results of the chapter do not greatly affect the interpretation of Chrysippus’ cylinder and cone analogy for determinism and freedom (for which see Bobzien 1998, chapter 6). Rather, they provide a more stable and more consistent base from which to understand the analogy, which tallies with what else we know about Chrysippus’ theory of causation and, more widely, with his cosmology and physics. It also nicely supports what we know about Chrysippus’ theory of causal determinism and fate from other sources. As Chapter 1 illustrates, the development of theories on determinism, freedom, and moral responsibility does not end with the Stoics, or even with the Stoics discussed in Chapters 1 and 7. Thus, Seneca has his own view on the issues, similar to, but far from identical with, those of the other Stoics here discussed. A valuable recent overview of Seneca’s position and how it relates to earlier Stoics can be found in Jula Wildberger’s extensive study (Wildberger 2006). For the early commentators on the Nicomachean Ethics, there exists now a volume edited by Antonina Alberti and Bob Sharples, which much improves our understanding of these commentaries (Alberti and Sharples 1998). The fifth chapter, by Alberti herself, treats the topics of the voluntary and of choice. The development of the notions of the will and of free will in later antiquity is set out in Michael Frede’s posthumously published Sather Lectures (Frede 2011). Frede broadly agrees with the philosophico-historical interpretation put forward in the first essay of this volume, in particular with the absence of a will in Aristotle and the early Stoics, the emphasis on freedom as a moral quality in Epictetus, and the Platonist and Peripatetic responses to the Stoic theory of free will and its compatibility with divine providence and causal determinism, as well as the fact that Alexander of Aphrodisias was the first to adopt a notion of free will as the ability to choose to act or not to act, given identical circumstances (in his chapters 2–6). As I write elsewhere, ‘The achievement of these chapters lies in the way Frede carefully joins them together and uses them as a basis for some substantive criticism and rewriting of the history of free-will regarding late antique Pagan and Christian authors, in particular Plotinus, Origen and Augustine’, and ‘[Frede’s] Chapter 7 traces early Christian notions of a will, with emphasis on Origen, whose view Frede argues to be at base Epictetan, and, where it differs from the Stoics, to rely on Platonist elements. Chapter 8 contends that Plotinus’ theories of god and freewill were not based on Judaeo-Christian thought, but rather strongly impacted by Stoic philosophy, with some Platonist elements. Chapter 9 argues that Augustine’s

12

, ,   

notion of free-will is based on later Stoic theory to an even greater degree than those of Origen and the early apologists.’³ So this small but important book provides a welcome and easily accessible sequel to the present volume. It is the separability of an incorporeal soul from the body that brings the theories and debates on freedom to a new level. Aristotle’s hylomorphism and Epicurean and Stoic materialism do not permit a human soul to survive the destruction or demise of the human body. This is not so with Platonist and Christian philosophy. Here, Ursula Coope’s excellent recent study on freedom and responsibility in Neoplatonist philosophy provides both overview and indepth analysis (Coope 2020). It ties in Neoplatonism with the Pre-Plotinian historical development, including in particular Plato and the Middle Platonists. It provides a learned philosophical study of the various elements of freedom and responsibility, introducing the reader to the specifically Neoplatonist notions of self-determination, self-constitution, self-making, and self-reflexivity that move the discussions and debates to new levels, from Plotinus, via Porphyry and Iamblichus, to Proclus, Damascius, and Simplicius. Let me finally also mention two recent multi-authored collections of papers on freedom, determinism, causality, and moral responsibility—Destrée, Salles, and De Zingano 2014 and Masi and Maso 2013—the topics of the contributions in which range from Democritus to Simplicius. As should thus have become clear, the essays in the present volume introduce and examine only a small part of the rich trove of ancient thoughts and theories—as well as their historical developments—on questions of determinism, freedom, and moral responsibility.

³ Bobzien 2012.

1 The Inadvertent Conception and Late Birth of the Free-Will Problem Aristotle to Alexander of Aphrodisias

1. Various Problems of Freedom and Determinism I will start with a number of distinctions vital to the subsequent discussion of ancient philosophical theories. These distinctions are kept rough and schematic. They are left deliberately vague in certain respects, because the ancient theories whose understanding they are intended to further are themselves stubbornly vague in those respects. The first distinction is of different kinds of freedom. I distinguish three kinds of indeterminist freedom: 1. Freedom to do otherwise. I am free to do otherwise if, being the same agent, with the same desires and beliefs, and being in the same circumstances, it is possible for me to do or not to do something, in the sense that it is not fully causally determined whether or not I do it. 2. Freedom of decision. A subtype of freedom to do otherwise. I am free in my decision if, being the same agent with the same desires and beliefs, and being in the same circumstances, it is possible for me to decide between alternative courses of action, in the sense that it is not fully causally determined which way I decide. 1 differs from 2 in that it leaves it undecided in which way it is possible for the agent to do or not to do something. 3. Freedom of the will. A subtype of freedom of decision. I act from free will if I am in the possession of a will, i.e. a specific part or faculty of the soul by means of which I can decide between alternative courses of action independently of my desires and beliefs, in the sense that it is not fully causally determined in which way I decide. 1 differs from 3 in that the latter This essay first appeared as ‘The Inadvertent Conception and Late Birth of the Free-Will Problem’ in Phronesis 43 (1998), 133–75, and is reproduced here with permission. The essay was reprinted, with minor corrections, in R. Sorabji (ed.), Aristotle Re-Interpreted (London: Bloomsbury, 2016), 125–60. This is the revised version of a paper I gave to a seminar of the Institute of Classical Studies of the University of London in 1994, and I would like to thank the audience for helpful criticism in the discussion and Michael Frede for a probing question. I am especially grateful to Bob Sharples, who not only put me on to the topic and invited me to give the paper, but also generously sent me several pages of detailed comments.

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0002

14

, ,    postulates a specific causally independent faculty or part of the soul which functions as a ‘decision-making faculty’.

Proponents of any of the kinds of indeterminist freedom may be called indeterminist libertarians. From these three types of indeterminist freedom must be distinguished what I call ‘un-predeterminist’ freedom. 4. Un-predeterminist freedom. I have un-predeterminist freedom of action/ choice if (i) there are no causes prior to my action/choice which determine whether or not I perform/choose a certain course of action, and (ii) in the same circumstances, if I have the same desires and beliefs, I would always do/choose the same thing. Un-predeterminist freedom guarantees the agents’ autonomy in the sense that nothing except the agents themselves is causally responsible for whether they act, or for which way they decide. Un-predeterminist freedom requires a theory of causation that is not (just) a theory of event-causation (i.e. a theory which considers both causes and effects as events). For instance, un-predeterminist freedom would work with a concept of causality which considers things or objects (material or immaterial) as causes, and events, movements, or changes as effects. Such a conception of causation is common in antiquity. Indeterminist freedom always requires the absence of predetermining causal factors, but in addition allows for different decisions of the same agent in the same circumstances. In the interpretation of ancient texts, indeterminist freedom is often confounded with un-predeterminist freedom. From both these types of freedom must be distinguished the following ones, which are compatible with both indeterminism and ‘un-predeterminism’: 5. Freedom from force and compulsion. I am free in my actions/choices in this sense if I am neither externally nor internally forced or compelled when I act/choose. This does not preclude that my actions/choices may be fully causally determined by external and internal factors. 6. Freedom from determination by external causal factors. Agents are free from external causal factors in their actions/choices if the same external situation or circumstances will not necessarily always elicit the same (re-)action or choice of different agents, or of the same agent but with different desires or beliefs. 7. Freedom from determination by (external and) certain internal causal factors. In my actions/choices, I am free from certain internal factors (e.g. my desires) if having the same such internal factors will not necessarily always elicit in me the same action/choice.

      - 

15

The last two types of freedom (6 and 7) differ from freedom from force, etc. (5) in that the latter only rules out force, compulsion, and necessitation, whereas 6 and 7 also rule out full causal determination, as is the case, for example, in a theory based on nothing but universal regularity of the respective causal factors. The list of types of freedoms 1 to 7 is evidently neither exhaustive nor exclusive. Note that the only proper Greek term for freedom is eleutheria, and that our evidence suggests that eleutheria played no role in the discussion of determinism and moral responsibility up to the second century . In particular, the term eleutheria is not involved in the development of the concept of freedom to do otherwise. Rather, it is the conceptual development of the phrase eph’hēmin that is pertinent here, and this has an altogether different history.¹ It is the notion of autonomous agency and of not being determined by something else (freedoms type 5–7), which in philosophical discussions in later antiquity becomes connected with eleutheria. Next, there are two categorically different conceptions of moral responsibility, one grounded on the autonomy of the agent, the other on the ability of the agent to do otherwise. The first (MR1) considers it a necessary condition for praising or blaming an agent for an action that it was the agent and not something else that was causally responsible for whether the action occurred. The contrast is between self-determination and other-determination to act. Actions or choices can be attributed to the agent because it is in them that agents, qua rational or moral beings, manifest themselves. Some thinkers consider the un-predeterminedness of an action/choice as a necessary condition for autonomy, and consequently for the attribution of moral appraisal. The second idea of moral responsibility (MR2) considers it a prerequisite for blaming or praising an agent for an action that the agent could have done otherwise. This idea is often connected with the agents’ sentiments or beliefs that they could have done otherwise, as well as the agents’ feelings of guilt or regret. Some philosophers consider the indeterminedness of an action/choice as a necessary condition for the guarantee that the agent could have done otherwise. The concepts of indeterminist freedom of an agent (see above, this section) gain importance at the point at which moral appraisal is connected with the idea that, at the very same time, the same agent, with the same beliefs and desires, could have done otherwise. Depending on what conception of moral responsibility an ancient determinist philosopher has, they will encounter different philosophical problems. With an autonomy-based concept of moral responsibility, they tend to face the problem of the compatibility of autonomy and determinism: how can I, the agent, be held responsible for my actions/choices, if everything, including my actions/choices, is ¹ I argue for the importance of realizing the very different philosophical functions of the terms eleutheria and eph’hēmin in Chapter 7 of this volume (originally published in Bobzien 1997, 71–89).

16

, ,   

determined, predetermined, or necessitated by god, fate, providence, necessity, or various other external and/or internal causal factors? This is the problem that the early Stoics, among others, faced. With a concept of moral responsibility based on a concept of freedom to do otherwise, determinists tend to face a very different kind of difficulty: the problem of the compatibility of freedom to do otherwise and determinism. In accordance with the threefold distinction of indeterminist freedom, three problems can be distinguished: • The problem of the compatibility of freedom to do otherwise and determinism • The problem of the compatibility of freedom of decision and determinism • The problem of the compatibility of freedom of the will and determinism All three problems are often referred to as ‘the free-will problem’, although only the third actually involves a notion of a free will. The label ‘free-will problem’ is also sometimes used for the problem of the compatibility of autonomy and determinism, namely when the agent is thought to have a faculty of the will and it is by means of this faculty that the agent decides between different courses of actions. Quite often it is taken to be ‘understood’, and is hence left completely unclear, what an author means when talking about ‘free will’ and ‘the free-will problem’. In the following I reserve the expression ‘free will’ for the kind of freedom I called ‘freedom of the will’ above. To avoid confusion, I use the phrase ‘free-will problem’ sparingly, and for the above-mentioned three problems only. Modern philosophers tend to concentrate on physical or causal determinism based on principles of the kind ‘same causes, same effects’ or ‘like causes, like effects’, and the prevalent types of free-will problem are those of the compatibility of universal causal determinism with freedom to do otherwise or freedom of decision. Few philosophers nowadays would postulate a faculty of the will. The earliest unambiguous evidence for the awareness of any kind of ‘free-will problem’ occurs in Alexander of Aphrodisias. It resembles the problems modern philosophers discuss in that it is concerned with a theory of universal causal determinism which contains a principle of the type ‘same causes, same effects’, and in that it involves a concept of indeterminist freedom without invoking a concept of the will. It is the ‘discovery’ of this kind of problem with which I am primarily concerned in this chapter. The historical treatment of the question of freedom and determinism is exacerbated by the fact that almost all key terms and phrases used to describe the problems involved are hopelessly vague or ambiguous. This is no different in Greek and Latin than in English. Many phrases and statements in philosophical texts before Alexander are—at least at first sight—compatible with an interpretation that they concern indeterminist freedom. However, there is a conspicuous

      - 

17

absence of any unambiguous account of indeterminist freedom, and of any philosophical problems that would involve such a concept.² I have therefore adopted the strategy of denying the awareness of a concept of indeterminist freedom and of the free-will problem (in any of its manifestations) as long as there is neither textual evidence nor any philosophical reason for assuming the opposite. On the following pages I shall first present the situation as we find it in Alexander; then sketch the development that leads to that state of the discussion; and finally interpret the problems presented by Alexander and some related philosophers in the light of the development that led up to them.

2. The State of the Debate in Alexander In Alexander’s treatise On Fate we are presented with a kind of stalemate situation between two philosophical positions: the Stoic compatibilist determinist one and Alexander’s Peripatetic and—seemingly—libertarian one. These positions are characterized by their stand (i) on causal determinism and (ii) on that which depends on us (to eph’hēmin). The expression ‘depending on us’ is central to much of the debate: both parties are agreed that moral appraisal for an action presupposes that the action depends on the agent, or is eph’autōi. The Stoic compatibilist position is orthodox and stands in the Chrysippean tradition. Like Chrysippus, these later Stoics are concerned with the compatibility of universal causal determinism with moral responsibility based on the idea of autonomy (MR1). They maintain that everything is fated and define fate in terms of a network of causes. They hold that there is no change without a cause and that every change and every event has preceding causes (Fat. 191.30–192.14). The most remarkable element of their determinism is the formulation of a causal principle whose function it is to back up their basic assumption that there is no change without cause (Fat. 192.22). This principle is not recorded for any earlier Stoics. It states that in the same circumstances the same cause will necessarily bring about exactly the same effect: . . . that it is impossible that, when all the same circumstances around the cause and that of which it is a cause are present, things should sometimes not happen in a certain way and sometimes should so happen. (Alex. Fat. 192.22–4)³

² Except perhaps the problem in Aristotle, EN 3.5 (1114a3–1114b25), but even that is doubtful. ³ Τὸ ἀδύνατον εἶναι, τῶν αὐτῶν ἁπάντων περιεστηκότων περί τε τὸ αἴτιον καὶ ᾧ ἐστιν αἴτιον, ὅτε μὲν δὴ μὴ οὑτωσί πως συμβαίνειν, ὅτε δὲ οὕτως. Cf. Nemesius, Nat. hom. 105.18–21 (Morani), Alex. Fat. 176.21–2; 181.21–5; 185.7–9; Mant. 174.2–7; see also §8.2 of Bobzien 1998.

18

, ,   

Universal causal determinism is thus guaranteed; both in the Stoic sense, in which causes are bodies which actively bring about their incorporeal effects, and in the common modern sense that the same cause in the same circumstances brings about the same effect, where both cause and effect are understood as events. The Stoics in Alexander argue for the compatibility of this physical theory with moral responsibility by means of their concept of what depends on us. They define that which depends on us as that which happens through us (di’ hēmōn), i.e. that which is the result of impulse and assent, and in which the nature of the agent manifests itself.⁴ We are thus the main causal factor of our actions and can consequently be held morally responsible for them. The opposing Peripatetic position which Alexander puts forward is less clear. Instead of a uniform stand, there is a variety of views, alternating and occasionally fused, a point to which I return later. But there is evidence for a position that proposes freedom to do otherwise and which resembles, up to a point, modern notions of freedom of decision. For example, we find the account: ‘depending on us’ is predicated of the things over which we have in us the power of also choosing the opposite. (Alex. Fat. 181.5)*

and the explicit requirement that this choosing has to be independent of preceding causes: we have this power of choosing the opposite and not everything that we choose has predetermining causes, because of which it is not possible for us not to choose this. (Alex. Fat. 180.26–8)**

Thus universal causal predeterminism is rejected. Moreover, it seems that the causes that are rejected include not only the external preceding circumstances, but also the agents’ disposition or character and reason (Fat. 171.11–17; 199.27–200.7): agents can act against their dispositions or character and reason, and are thus causally independent from them. This concept of what depends on us combines two features: the first is that of non-predeterminism, i.e. the freedom from previous states of the world, including those concerning the agent. There are no causes prior to our choosing by which it is predetermined what we choose. The second is the agent’s freedom to do (i.e. choose) otherwise.

⁴ Cf. Alex. Fat. ch. 13, Nemesius, Nat. hom. 105–6 and see below, section 12. Despite its simple form, the definition is highly technical. For a detailed discussion see §§8.1 and 8.4 of Bobzien 1998. * (ἀλλ᾽ὅτι μὲν) τὸ ἐφ᾽ἡμῖν ἐπὶ τούτων κατηγορεῖται, ὧν ἐν ἡμῖν ἡ ἐξουσία τοῦ ἑλέσθαι καὶ τὰ ἀντικείμενα. ** (ὡς) δύνασθαι διαιρεῖσθαι τὸ ἀντικείμενον καὶ μὴ πᾶν ὃ αἱρούμεθα ἔχειν προκαταβεβλημένας αἰτίας, δι᾽ἃς οὐχ οἷόν τε ἡμᾶς μὴ τοῦτο αἱρεῖσθαι.

      - 

19

It should be plain that the positions of the two parties in the debate are incompatible. The Stoics maintain that every change in the world is causally determined by preceding causes. The same cause under the same circumstances will necessarily bring about the same effect. The Peripatetic claim is that there are some changes in the world that are causally undetermined, and among these are the things that depend on us. In the very same situation, we, the very same causes (causes understood as corporeal entities), could choose one time one way, another time another way, undetermined in our decision by external and internal causal factors. It is essential to see that there is no solution to this conflict: causal determinism and partial causal indeterminism are mutually exclusive. However, although the text implies awareness of a problem of the compatibility of freedom and determinism, the discussion seems to have focused on a different problem: the question is not ‘which is the correct concept of freedom?’ (It is telling that there is no word for ‘freedom’ used in the debate; whether or what concepts of freedom are involved in the opposed theories has to be inferred from the context. The unfortunate custom of translating the Greek phrase eph’hēmin by ‘free will’ or cognates of ‘free’ simply begs the question.) The question of the debate in Alexander is rather: ‘which is the right concept of what depends on us?’, i.e. ‘which concept provides a sufficient condition for the possibility of moral appraisal?’ And here two very dissimilar underlying theories collide. To understand the nature of this controversy, we have to realize that the two parties work with two fundamentally different conceptions of what depends on us. For this we need to make explicit the different ways in which the phrase ‘depending on someone’, or rather the Greek ‘epi’ with dativus personae, could be understood. The phrase apparently could denote both what I have named a ‘one-sided, causative’ concept of what depends on someone, and what I call a ‘two-sided, potestative’ concept of what depends on someone. The two-sided, potestative version is well-attested (cf., e.g., LSJ epi, I. 1g). It refers to a power for alternative kinds of behaviour; it depends on me whether something happens (or will happen). When I call this kind of depending on us ‘two-sided’, I mean that if something x depends on us, then not x depends on us as well. Thus, in the two-sided, potestative understanding, ‘up to us’ would be a good translation of the Greek expression: for example, if walking is up to me, so is not walking, and vice versa. In this understanding of ‘what depends on us’, the class of things that depend on us includes unrealized possibilities. For example, when at a certain time walking depends on me, then not walking depends on me as well. But I will be able only either to walk or not to walk at that time. Hence either one or the other will remain an unrealized possibility. Note that the two-sided, potestative eph’hēmin itself entails neither determinism nor indeterminism. A reading compatible with determinism (and indeterminism) is this: walking depends on me at a certain time if at that time I have the

20

, ,   

general two-sided capacity for walking—even if in the specific situation it is fully causally determined that I will (or that I will not) walk.⁵ But, importantly, the two-sided, potestative eph’hēmin can also be understood as indeterminist in the following way: at a certain time walking depends on me if at that time it is causally undetermined whether or not I (will) walk, and it depends on my free decision whether or not I (will) walk. When the expression is understood as two-sided, potestative in this way, the ‘we’ (‘us’) in eph’hēmin takes on an interesting role: the ‘us’ in, for example, ‘walking depends on us’ is given the status of an active decision maker. We decide whether or not we walk. Instead of a general capacity had at a certain time, in this case there is a power for undetermined deciding between, and initiating, courses of action. This is a very different kind of capacity. And in this case, if something depends on me then I have the indeterminist freedom to do and not to do it (cf. section 11). Things are quite different again in the case of the one-sided, causative eph’hēmin. When I call the phrase ‘one-sided’, I understand this to entail that if something x depends on us, then not x does not depend on us; and by ‘causative’ I refer to the fact that the prepositional phrase in ‘x depends on y’ refers to that which is the cause or reason of x. In this case a translation like ‘attributable to us’ may be preferable. If at a certain time my walking is attributable to me, then it is not the case that my not walking is attributable to me as well. For in the assumed situation my not walking does not obtain at all. Here, the natural understanding of ‘x depends on us’ is that it expresses who has the causal responsibility for the thing or action in question. ‘The walking is attributable to you’ translates into ‘You are causally responsible for your walking’. The ‘we’ in eph’hēmin now expresses the cause of what happens and depends on us. The one-sided, causative eph’hēmin, too, can be used in the description of an indeterminist as well as a determinist system. However, whereas the two-sided eph’hēmin can be used to express an element of undeterminedness, by implying that we, qua decision makers, can decide freely between alternative options, the one-sided ‘depending on us’ cannot be so used. Its function is to help to distinguish between different types of ‘causes’ of events, not to imply the possibility of freedom to do otherwise.⁶ The one-sided, causative concept of what depends on us ⁵ A related reading compatible with determinism (and indeterminism) is this: walking depends on me at a certain time if at that time I have the general two-sided capacity for walking and nothing (or nothing external) forces me to walk or prevents me from walking—even if in the specific situation it is fully causally determined whether or not I will walk. This adds to the previous reading my freedom (of type 5 above), i.e. the additional requirement of the agent’s being neither hindered from nor forced to follow up either alternative. Still differently, a two-sided, potestative concept of eph’hēmin that is neutral towards determinism and indeterminism can also be used for action types without reference to a specific time. So walking may be said to be the sort of thing that is generally up to human beings. ⁶ Matters are complicated by the fact that when the one-sided, causative eph’hēmin is employed, it usually refers to the bringing about of something by way of using a general two-sided capacity (see below, section 7). Thus the one-sided conception and the two-sided conception that expresses a general twosided capacity come very close, and the context does not always allow one to decide which one is at issue.

      - 

21

is not a concept of any kind of freedom, but of a particular kind of causal dependency. However, it presupposes a certain kind of freedom: freedom (type 6) from being externally determined to act; or freedom (type 5) from being in any way forced to act and prevented from acting. Note that these concepts of freedom are not the same as this concept of what depends on us. An action depends on me if (in some way) I bear causal responsibility for it and am in this sense its originator. It is in order for this to be possible that I must not be compelled to act or prevented from acting, i.e. that I must be free from external or from necessitating influences. Depending on which conception of eph’hēmin a philosopher works with, the concept of moral responsibility will differ. In the case of the one-sided, causative eph’hēmin moral responsibility is attached to someone if they are—in some sense—the main causal originator and thus autonomous (MR1). This is the position of the Stoics in Alexander. In the case of the indeterminist two-sided, potestative eph’hēmin, moral responsibility is attached to someone if they are free to do otherwise; if they are not fully causally determined in their decision between alternative courses of action (MR2). This appears to be Alexander’s own position, according to the passages cited above. The case which understands the two-sided, potestative eph’hēmin as a general capacity, and which is neutral regarding determinism or indeterminism, is usually linked with a concept of moral responsibility based on the agent’s autonomy (MR1): moral responsibility is attached to my making use of the two-sided capacity, because it is through the use of this capacity that I, qua rational or moral person, become the originator of the action.

3. The One-Sided, Causative Conception of That Which Depends on Us Next I shall trace the philosophical development that led up to the antagonistic views in Alexander, with emphasis on the question of how the indeterminist twosided, potestative concept of what depends on us entered the debate. But let me begin with some remarks on the development of the one-sided, causative concept. This concept has a long history which reaches at least from the third century  to the third century . However, I have not found a matching philosophical definition or account before the second century .⁷ ⁷ In his discussion of the problem of the compatibility of necessitarianism and fatalism with moral responsibility Epicurus can be shown to have used throughout a one-sided, causative concept of what depends on us. (See Chapter 5 of this volume, originally printed as Bobzien 2000, 287–337.) The standard phrase used by him and his followers is par’hēmas gignesthai (‘to happen because of us’) rather than eph’hēmin einai—a linguistic point which provides additional support for my claim. Questions of freedom to do otherwise and freedom of decision were, it seems, not discussed.

22

, ,   

The early Stoics, in particular Chrysippus, clearly did not have an indeterminist two-sided conception of what depends on us.⁸ There are reasons internal to the Stoic system which help explain this. They regarded the mind as corporeal, and as unitary: a person’s character, dispositions, beliefs, and desires are all reducible to what impressions a person, or that person’s mind, gives assent to. (For the Stoics volitions or desires are a kind of beliefs.) The Stoics operated with a model of a person, or of agency, in which neither a person’s character or dispositions nor a person’s volitions can be severed from that person. Rather, they identified the person with the entire nature of the individual’s mind, including character, dispositions, and all. In this model an action is voluntary, and causally attributable to the person, if it is the result of the mind assenting to impulse-prompting impressions (phantasiai hormētikai), i.e. impressions of something as desirable. Whether or not assent is given to the impressions depends on the individual nature of the agent’s mind. What makes an agent morally responsible is that the agent, and not something else, causes the action. There is in this model no space for free will (i.e. for a decision-making faculty that is causally independent of the mind’s individual nature). For the fact that I act in accordance with the overall nature of my mind is considered a prerequisite for attributing the action to me, whereas free will takes the detachment of the decision-making faculty from the rest of the person as a necessary condition. The concept of an internally undetermined decision made by the agent is thus ill-fitting in the Stoic conceptual framework. And so, accordingly, is the free-will problem. It follows that the libertarian condition for the attribution of moral responsibility, that we have freedom to do otherwise, would have made little sense to Chrysippus. We can see that the Stoic account of that which depends on us in Alexander is an attempt to capture exactly this early Stoic concept of the causal responsibility of rational agents: an action is said to depend on us if it happens through us. The expression ‘happens through us’ is explicated as happening by (hupo) the human being, as a result of impulse and assent, and in accordance with the human being’s individual nature (Alex. Fat. ch. 13, esp. 181.18–21, 182.11–16; Nemesius, Nat. hom. 105–6).

4. The Two-Sided, Potestative eph’hēmin: Aristotle Things are less straightforward on the side of the indeterminist two-sided, potestative eph’hēmin, and with the indeterminist concept of freedom. As I have said above, the first full and unambiguous statement of freedom to do otherwise seems to occur in Alexander’s On Fate, and in the Mantissa. Where then do Alexander’s ⁸ Cf. §6.3.5 of Bobzien 1998. We do not know with certainty which Greek expressions Chrysippus used, but it is likely that ‘epi’ with dativus personae was among them.

      - 

23

indeterminist concept of what depends on us and his theory of freedom to do otherwise originate? In the Mantissa this theory and concept are attributed to Aristotle.⁹ Of course, this must not be taken literally. But it is worthwhile taking up the hint and asking: where in Aristotle’s extant writings do we find his view on that which depends on someone (‘epi’ with dativus personae)? The answer is: in the Nicomachean Ethics (EN) 3.2 and 3.3, on deliberation and deliberate choice (prohairesis), and 3.5, on the question of whether we are morally responsible for our actions and our virtues and vices; and in the parallel sections in the Eudemian Ethics (EE) 2.6 (1223a1–9) and 2.10. In Nicomachean Ethics 3.3 Aristotle argues that deliberate choice is deliberate desire of those things that depend on us (EN 1113a10–11). The things that depend on us are the things that we can bring about, as opposed to events brought about by nature, necessity, or chance, and also as opposed to those things only people other than us could bring about (EN 1112a21–33). They are in the first instance actions (EN 1112a31, 34). In Nicomachean Ethics 3.5 we learn that, besides actions, virtues and vices also depend on us (EN 1113b3–1115a3). Note the relation between deliberate choice and the things that depend on us: deliberate choice is choice of the things that depend on us, i.e. in the first instance of actions (EN 1113a10–11). That is, we deliberate about and choose between possible courses of actions. The choice we make (prohairesis) is itself not one of the things that depend on us, and the idea that it is would have been quite alien to Aristotle’s thinking. There is, however, one factor in Aristotle’s concept of what depends on us that we also have in Alexander: its ‘two-sidedness’. In Nicomachean Ethics 3.5 we learn that if doing something depends on us, then not doing that same thing also depends on us, and vice versa (EN 1113b7–8, cf. EE 1226a27–8), and this relational property of the concept is preserved in later Peripatetic philosophy. But Aristotle’s concept of what depends us does not entail indeterminism. We have no reason to assume that he has anything more in mind than that the things that depends on us are those which on a generic level it is possible for us to do and not to do, given that we are not externally prevented from doing them. In the two Ethics, all the concept of what depends on us does is give the general range of courses of action from which we can choose. The concept is independent of (and prior to) Aristotle’s concept of deliberate choice, and of any mental capacity we have. It is taken as a basic concept, undefined and generally understood, by means of which the scope of the objects of deliberate choice is determined. Thus Aristotle’s remarks on that which depends on us are a far shot from Alexander’s definitions. In none of the passages does Aristotle give a philosophical definition of that which depends on us; nor is he concerned with fate or causal

⁹ Cf. the titles of chs 22 and 23; see also Alex. Fat. ch. 39.

24

, ,   

determinism; and certainly there is no mention of freedom to act or choose otherwise, circumstances and agent being the same.

5. Early Commentaries on the Nicomachean Ethics Let us consider next the extant early commentaries on the Nicomachean Ethics, i.e. the commentary by Aspasius (who wrote in the first half of the second century ) and the Anonymous on Books 2 to 5 (who presumably wrote in the second half of the second century).¹⁰ The first thing to notice is that both commentators introduce the topic of fate where they comment on the Aristotle passage that states that deliberate choice is of the things that depend on us, and in which Aristotle lists as types of causes: necessity, nature, and chance.¹¹ Aspasius contrasts a thing’s being fated and a thing’s being determined by necessity with its depending on us (EN 74.10–13) and a little later contrasts ‘depending on us’ with ‘being necessitated’ (katēnagkastai, EN 76.11–14).¹² This suggests that he understands something’s depending on us as presupposing freedom from force or compulsion—just like Aristotle and the Stoics. The Anonymous has a different concept of fate: fate is subordinated to nature, and it is not untransgressable (aparabatos). He thus expressly rejects a property commonly linked with fate by the Stoics and many others, and his view of fate closely resembles that of Alexander (cf. Fat. ch. 6 and Mant. 186). A second point of interest is that both commentators introduce into the present context the idea of something’s being able to be or happen otherwise: Aspasius presents the Aristotelian account of what is necessary as what cannot be otherwise (EN 71.27*). The Anonymous writes that the actions we deliberate about (and, that is, the things that depend on us) are things which can be done in this way and otherwise (EN 149.30–1**). Again, the Anonymous is one step ahead, talking about acting otherwise. The formulations, especially that of the Anonymous, lend themselves in principle to an indeterminist concept of freedom to do otherwise, as we found it in Alexander. But there are no signs that either commentator took them that way. As in Aristotle, there is no philosophical account of that which depends on us. As in Aristotle, too, deliberate choice is not one of the things that depend on us but is of the things that depend on us. The things that depend on us are actions and ¹⁰ Cf. Moraux 1984, 226–7, 324–7; or perhaps later? See Sharples 1990, introduction. ¹¹ Aspasius, EN 74.10–14; Anon. EN 150.1–4. It may have been Theophrastus who introduced fate into this context; cf. Stobaeus, Ecl. 1.89.2–5. The earliest clearly two-sided concept of eph’hēmin in the fate debate that I have found so far is in Josephus, Ant. 13.172. I have been unable to trace any connection between Josephus and the positions discussed in this chapter. ¹² This is the standard Stoic contrast, as we usually find it in Epictetus. * ἀναγκαῖον γὰρ λέγεται τὸ μὴ ἐνδεχόμενον ἄλλως ἔχειν. ** ἀλλὰ περὶ τούτων ἃ καὶ ἄλλως καὶ οὕτως ἐνδέχεται πραχθῆναι.

      - 

25

virtues and vices. And, finally, it depends on us to do and not to do things, not to choose and not to choose things, as Alexander has it. You may say, well, that was to be expected; after all, this is what Aristotle says and we are here dealing with commentaries on Aristotle. This is true. However, in the Paraphrase of Book 3 of the Nicomachean Ethics, ascribed variously to some Heliodorus and others, in at least one place we have ‘helesthai kai mē’ instead of Aristotle’s original ‘prattein kai mē’ (52.25–7); and in the Mantissa ch. 22 we read, marked out as Aristotle’s view, that deliberate choice depends on us (eph’hēmin hē prohairesis, Mant. 169.38). So at some point Aristotle’s remarks on the things that depend on us must have been understood as being primarily about choice and only secondarily about actions (cf. also Ammonius, Int. 242.24–5 and below, section 10). We do not know when the Paraphrase was written, but it is very likely to date after Alexander. Thus, in both commentators the concept of what depends on us seems wholly compatible with both causal determinism and causal indeterminism. They both include elements which we find in the same context in Alexander but not in Aristotle. They both introduce formulations of the ‘it could be/happen otherwise’ kind and they both connect Aristotle’s Nicomachean Ethics 3.3, on deliberate choice and what depends on us, with fate—something Alexander did as well (e.g. Fat. 180).

6. Middle Platonists on Contingency and That Which Depends on Us For a closer connection between Aristotle’s works and Alexander concerning that which depends on us, we need to look elsewhere, viz. at the texts of the Middle Platonists.¹³ The Middle Platonists—like virtually all philosophical schools, sects, and currents in the first and second centuries —had developed their own position on fate and that which depends on us. Their theory of fate is based on a handful of passages from Plato and is influenced in many aspects by the Stoics. We find variants of the Middle Platonist theory, for example, in Alcinous’ Handbook of Platonism, in the treatise On Fate by [Plutarch], in Nemesius’ On Human Nature, and in Calcidius’ commentary on Plato’s Timaeus.¹⁴ [Plutarch]’s, Calcidius’, and Nemesius’ passages—although at variance in many details—go back to some elaborate Platonist theory of fate which may stem from

¹³ I use the term ‘Middle Platonist’ to refer to those Platonists commonly classified in this way, but do not maintain that there was any unified Middle Platonist school. ¹⁴ Alcinous, Didasc. ch. 26, Nemesius, Nat. hom. e.g. 110.125–6, Calcidius, Tim. 142–87, [Plut.] Fat. passim, Apuleius, Plat. 1.12, perhaps an echo in Alex. Mant. ch. 25, 183.

26

, ,   

the first half of the second century,¹⁵ but which should in any event at least in part precede Alexander. It is likely that there were written versions of this theory. It appears that such written versions of the common source encompassed first a part on fate in which—among other things—a distinction was drawn between things that are included (periechein) in fate, which are all things, and those things that are fated, or in accordance with fate (kath’ heimarmenēn), which are all those things that are necessary. In addition, it contained a section on the things included in fate but not fated; a section on providence; a critique of Stoic doctrine on fate; and a discussion of fallacies concerned with determinism. For my present purposes, what is relevant is the section on things included in fate but not fated. In all three sources this section differs from the rest of the Middle Platonist theory in that (i) the only traces from Plato are a couple of examples tacked on in the section on chance; (ii) the passage is clearly based on a whole range of texts from Aristotle, which all dealt with ‘that which is not necessary’.¹⁶ It looks as if someone has taken a list of types of things that are not necessary, perhaps from Aristotle EN 3.3.1112a31–3, perhaps from some later, ‘updated’, list,¹⁷ and has then worked his way through the works of Aristotle, picking out and systematizing the relevant sections. The passage draws from Aristotle’s Physics, Metaphysics, Nicomachean Ethics, De Interpretatione, and perhaps from the Categories. The list of things not necessary is of interest for two reasons: first, it includes that which depends on us (to eph’hēmin). Second, we do not simply have a presentation or co-ordination of bits from Aristotle, but a systematization in which a couple of distinctions and terms are also added which we do not find, or do not find used in that way, in Aristotle. Regarding the origin of this passage, we may assume that it was compiled in the second century  at the latest. It may well be earlier. As I said above, the only bits from Plato in it are two examples tacked on to the section on chance. There is thus no reason to think that the author of the common source of the Middle Platonist doctrine of fate is the originator of the passage. The rest of the Middle Platonist theory of fate stands without it, and vice versa. The original author of this bit of ‘Middle Platonist’ theory could equally well be a Peripatetic or a Platonist—if indeed such a distinction made sense at the time. We may say that the author was an Aristotle scholar. The passage appears to have employed the following classificatory scheme: the most general term is the possible (to dunaton). It encompasses both the necessary ¹⁵ So Moraux 1984, 495–6, following Gercke and others. There is a problem in that the texts that report from this theory are either clearly later (Nemesius, Calcidius) or cannot be dated with any certainty (Alcinous, [Plutarch]). It is also clear that over time the theory underwent extensive step-bystep development, and many elements of the theory as reported by Calcidius and Nemesius are certainly later than the second century. ¹⁶ Nemesius, Nat. hom. 103–4, [Plut.] Fat. 570f–572f, Calcidius, Tim. 155–6. ¹⁷ See e.g. Theophrastus in Stobaeus, Ecl. 1.89.2–5; Alex. Fat. 211.1–4, Mant. ch. 25; Nemesius, Nat. hom. 112.13–15 for such lists.

      - 

27

(to anagkaion) and the contingent (to endechomenon): the necessary is determined as the possible the opposite of which is impossible; the contingent as the possible the opposite of which is also possible (dunaton hou kai to antikeimenon dunaton).¹⁸ This is the distinction between one-sided and two-sided possibility as we find it in Aristotle’s De Interpretatione 12 and 13 (Int. 22b36ff; 23a15–16), although Aristotle did not consistently use the term ‘contingent’ (endechomenon) for two-sided possibility in the way our source seems to have done. That which depends on us is then characterized as a subclass of the contingent.¹⁹ That is, the section connects explicitly a distinction and an account from Aristotle’s modal theory (that of two-sided possibility) with the question of moral responsibility as it comes up in Aristotle’s ethics. Is all this not a bit far-fetched? you may object. Surely the Stoics and the Megarics or Dialecticians had already connected the problem of determinism with modal logic, for instance in the Mower Argument and the Master Argument.²⁰ This is true, and given that the same Middle Platonist source discussed these very fallacies and criticized the Stoic theory of fate,²¹ the trigger for connecting Aristotelian modal theory and Aristotelian ethics may well have come from there. You may also object that the connection between the concept of contingency, qua two-sided possibility, and of what depends on us is just a commonplace. However, for instance the early Stoics, it seems, had neither a definition of what depends on us nor a term for contingency: they talked about two-sided possibility in terms of what is true but not-necessary and what is false but possible.²² Moreover, the Stoics had a one-sided concept of what depends on us—and they were by far not the only ones (see section 12). Be that as it may, the best way to find out whether Aristotle’s De Interpretatione 12 and 13 were used for the development of a concept of what depends on us is to look at the passages and the ancient commentaries on them. In De Interpretatione 13 Aristotle connects the concept of capacity (dunamis) with that of two-sided possibility, stating that not all capacities are two-sided, or capacities of opposites (antikeimena), although rational capacities are, like that of human beings for actions such as walking (Int. 13.22b36–23a6; cf. Metaph. 9.1046b1–2, 4–7; 1048a2–3, 8–9; 1050b30–4). Thus here Aristotle links the concept of two-sided possibility with the capacities for opposites, and draws the connection to human rational capacities. All the signs are that what Aristotle has in mind is a general capacity and not a capacity involving freedom to do otherwise. (For instance, he

¹⁸ Nemesius, Nat. hom. 103.20–1, [Plut.] Fat. 571b. ¹⁹ [Plut.] Fat. 571c–d, Nemesius, Nat. hom. 114.21–4. ²⁰ For the Mower Argument, see e.g. DL 7.25, Ammonius, Int. 131, 25–32; for the Master Argument, Epict. Diss. 2.19. ²¹ [Plut.] Fat. 574e, Calcidius, Tim. 160–1. Note also that Ammonius discusses the Mower Argument in his commentary on De Interpretatione, see previous note. ²² Cf. Cic. Fat. 13, Plut. Stoic. Rep. 1055e.

28

, ,   

mentions the fact that there are non-rational capacities for opposites, e.g. Int. 13.23a3–4.*) Ammonius, in his commentary on this very passage of the De Interpretatione, states that in the case of us human beings, whose rational capacities are two-sided (Int. 242.19–20), we are master . . . of our deliberate choice, and it depends on us to do or not to do any of the things that happen in accordance with deliberate choice. (Ammonius, Int. 242.24–7)

Ammonius here connects that which depends on us with Aristotle’s two-sided possibility (the contingent), via the rational capacities of which Aristotle speaks, interpreting these in particular as power of deliberate choices.²³ But deliberate choice and that which depends on us were the topic of Nicomachean Ethics 3.3. The next line of our passage (Int. 242.27–8) is also of interest: there the question is raised as to whether the capacities of the gods correspond to one-sided or twosided possibility. And then we learn: Alexander asked himself the same question. Now, if Alexander asked himself that question, it seems very likely indeed that what led up to the question, i.e. the above-quoted passage in the commentary just beforehand, stems also from Alexander, that is—presumably—from his (lost) commentary on the De Interpretatione.²⁴ Thus it appears that both our Aristotle scholar from the Middle Platonist common source and Alexander connected that which depends on us with a passage on modal logic from Aristotle’s De Interpretatione 13. But let us return to our Aristotle scholar and the subordination of that which depends on us to the contingent in the Middle Platonist common source. There we find three types of the contingent: one part of the contingent is ‘for the most part’, one ‘for the lesser part’, and one ‘in equal parts’.²⁵ Those for the most part and for the lesser part are characterized as opposites. For example, if for the most part the weather is hot in August, it is cold, or not-hot, in August for the lesser part.²⁶ On the other hand, the ‘in equal parts’ is that which depends on us, as for instance walking and not walking, and in general acting and not acting.²⁷ * ἔνια μέντοι δύναται καὶ τῶν κατὰ τὰς ἀλόγους δυνάμεις ἅμα τὰ ἀντικείμενα. ²³ Aristotle himself introduced deliberate choice (but not that which depends on us) into this context in Metaph. 9.1048a10–15, but in a very different way than Ammonius: his point is that, circumstances permitting, we will necessarily realize that side of our two-sided capacities for which we have a desire resulting from deliberate choice. ²⁴ See Moraux 1984, 363. ²⁵ ἐπὶ τὸ πολύ, ἐπ᾽ἔλαττον, ἐπ᾽ἴσον; Nemesius, Nat. hom. 104, 1–2, [Plut.] Fat. 571c, cf. Calcidius, Tim. 156. ²⁶ Nemesius, Nat. hom. 104.2–4, [Plut.] Fat. 571c. [Plutarch] maintains that both are subordinated to nature (ibid.). ²⁷ [Plut.] Fat. 571c–d, Nemesius, Nat. hom. 104.4–5 in connection with 114.19–22, cf. below, pp. 29–30.

      - 

29

In Aristotle we find neither this threefold distinction of the contingent nor the category of what is ‘in equal parts’. However, there can be no doubt that this third category is derived from Aristotle’s Int. 9.18a39–b9, and is meant to pick up what Aristotle calls ‘as it happens’ (hopoter’etuchen) there.²⁸ In chapter 9 of the De Interpretatione Aristotle investigates whether or in what way the Principle of Bivalence holds for future propositions. One of his problems is that, if all propositions that state something about the future are already true or false now, this fact could somehow entail that all future events are predetermined already now.²⁹ Aristotle contrasts the ‘as it happens’ with necessity and explains it as ‘it is no more thus than not thus (ouden mallon), nor will it be’ (Int. 18b9, cf. 19a18), and it is about these (the things ‘as it happens’) that one deliberates (Int. 18b31, cf. 19a9). Hence, I assume, our Aristotle scholar—or some earlier Aristotle exegete— simply reasoned as follows: the ‘as it happens’ must be part of the contingent, since according to Aristotle it is not necessary. An expression, parallel to ‘for the most part’, was then coined for this subtype of contingent, namely ‘in equal parts’, based on Aristotle’s phrase ‘no more thus than not thus’. Since Aristotle says that the ‘as it happens’ is concerned with deliberation and action, and this is—according to Aristotle himself—the sphere of that which depends on us, the Aristotle scholar concluded that the ‘as it happens’ must be that which depends on us. This identification of the ‘in equal parts’ with that which depends on us suggests that, unlike the ‘for the most part’ and the ‘for the lesser part’, the ‘in equal parts’ was not given a statistical interpretation. The phrases ‘for the most part’ and ‘for the lesser part’ express probability in the sense that if it is, say, hot in August 95 per cent of all years (of all days?), then it is ‘for the most part hot’ in August. But our Aristotle scholar cannot have understood the statement ‘walking is “in equal parts” ’ to mean ‘people walk 50 per cent of the time’, or ‘people walk 50 per cent of the time that is relevant to walking’. Rather, the idea must have been that in any situation of possible walking it is no more likely that the person walks than not—quite independently of how much people statistically actually walk. In Nemesius we are twice given a definition of the ‘in equal parts’; it is ‘that of which we are capable of * both it and its opposite’ (Nemesius, Nat. hom. 104.6–7; 114.21–2**). In the second passage this definition is followed by the explicit identification of the ‘in equal parts’ with what depends on us, and it is illustrated with examples (Nat. hom. 114.24–115.2). Note the similarity of the definition with, on the one hand, Aristotle’s things that ‘are at the same time capable of opposites’*** from De Interpretatione 13 (23a3–4, see above, this section), and with Alexander’s definition of that which depends on us as ‘that of ²⁸ Cf. e.g. [Plut.] Fat. 571c: τὸ δὲ ὡς ἐπίσης καὶ ὀπότερον ἔτυχεν. See also Ammonius, Int. 143.1–7, Alex. An. Pr. 163.21–9, Fat. 174.30–175.4. ²⁹ For the controversy among scholars over this passage, see e.g. Frede 1985, 31–87. * Angled brackets in text and translation are used to indicate a textual emendation. ** αὐτό τε δυνάμεθα καὶ τὸ ἀντικείμενον αὐτῷ. *** δύναται ἅμα τὰ ἀντιεείμενα.

30

, ,   

which we have the power of choosing also its opposite’**** (Fat. 181.5–6), on the other (see above, section 2). As an interim result we can state: we have in this Middle Platonist common source a concept of what depends on us which seems to be the outcome of bringing together and systematizing three bits of Aristotelian doctrine: • The things that depend on us, as those we deliberate about and from which we choose, from Nicomachean Ethics 3.3; • The concept of two-sided possibility, or the contingent, and its relation to the two-sided capacities of rational beings, from De Interpretatione 13 and Metaphysics 9; • The problem of future contingents, and the idea of things that can equally happen and not happen, from De Interpretatione 9. The resulting concept is captured in the account of what is ‘in equal parts’—which is identified with what depends on us—as: ‘that of which we are capable of both it and its opposite’.³⁰

7. The Philosophical Relevance of the Link between the Nicomachean Ethics and the De Interpretatione Why have I spent so much time on the development of a concept of what depends on us in the context of Aristotle’s De Interpretatione? The reason is this: my above question was when and where that concept was first understood as indeterminist, and as implying the possibility that the same person, with the same beliefs and desires, in the same circumstances, does otherwise. The obvious question now is whether the eph’hēmin of our Aristotle scholar was taken to imply indeterminist freedom. The definition ‘that of which we are capable of both it and its opposite’ itself is of no help: it is just as ambiguous in this respect as were Aristotle’s original phrases ‘what depends on him to do and not to’ (cf. EE 1223a7–8, 1226b30–1) and ‘being at the same time capable of opposites’ (Int. 23a3–4). Take walking as an example: walking depends on me because I am capable of both walking and the opposite, i.e. not walking. This can mean that I have the general two-sided capacity of walking and of not walking—which would be compatible with determinism. Alternatively, it can be **** ὧν ἐν ἡμῖν ἡ ἐξουσία τοῦ ἑλέσθαι καὶ τὰ ἀντικείμενα. ³⁰ We do not know who first brought together the Nicomachean Ethics and the De Interpretatione in this context. We do know that Aspasius wrote a commentary not only on the former, but also on the latter, which is however lost; cf. Moraux 1984, 231.

      - 

31

understood as implying that there are no preceding causes which sufficiently determine that I walk or determine that I do not walk. This understanding would be incompatible with predeterminism. Or it can be understood as implying that it is causally undetermined whether or not I walk. This would be incompatible with determinism. Perhaps the Middle Platonist identification of what depends on us with what happens ‘in equal parts’ was a step in the direction of undetermined choice. For if the ‘in equal parts’ is understood as tied to individual situations such that, for example, in every situation of possible walking it is no more likely than not that I walk (see section 6), then the capacity expressed in the account ‘capable of both it and its opposite’ will also be tied to an individual situation. That is, something depends on us at a certain time if at that time we are capable of doing both it and its opposite. But this is still ambiguous between (i) having at that time a general capacity to walk and being at that time (ii) un-predetermined or (iii) causally undetermined in our walking.³¹ The vagueness in the definition of that which depends on us seems to be resolved in favour of an un-predeterministic or indeterministic concept once the connection of that which depends on us with the problem of the truth values of future propositions is fully taken into account. For in De Interpretatione 9 the problem of truth values of future propositions is connected with the question of the undeterminedness of the future, more precisely, the undeterminedness of whether something will happen at a particular future time: it is not yet determined now whether there will be a sea battle tomorrow. Here pairs of propositions about the occurring of future events are at issue, and the occurring is indexed to a particular time in the future: the occurring of a sea battle tomorrow versus the absence of the occurring of a sea battle tomorrow. Plainly, the question here is not whether a sea battle (or anything else) has a general capacity of occurring, and whether it has that capacity now. The question is whether tomorrow a sea battle will or will not take place. That is, starting out from one and the same situation, viz. the present one, it is assumed that something could or could not obtain at some later time; and it is at present undetermined whether or not it will obtain. Thus here we have expressly one necessary condition for indeterminist freedom to do otherwise: exactly the same antecedent situation is combined with the possibility of two opposed states of affairs obtaining in some later situation in such a way that the antecedent situation leaves it undetermined which of the later states will obtain. However, in Aristotle’s De Interpretatione this is merely a matter of logic: Aristotle does not consider whether the present situation is causally responsible for what happens in the future. A fortiori, he does not ask whether

³¹ The same ambiguity is connected with ἅμα in Aristotle’s phrase δύναται ἅμα τὰ ἀντικείμενα from De Interpretatione 13.23a3–4.

32

, ,   

definite truth-bivalence of future propositions would entail that human decisions are causally predetermined.

8. Alcinous and Ammonius Did our Aristotle scholar, or the Middle Platonists—when drawing on De Interpretatione 9 and identifying what depends on us with what happens ‘to equal parts’—make this step from the logical undeterminedness of the future by the present to the causal undeterminedness of human decisions and/or actions? We do not know. However, remnants of such a thought can perhaps be detected in chapter 26—on fate, possibility, and that which depends on us—in the Handbook of Platonism of the Middle Platonist Alcinous. This passage displays many similarities to those in Nemesius, Calcidius, and [Plutarch], but is sufficiently distinct to suggest that it belonged to a slightly different tradition.³² Alcinous works with a two-sided, potestative concept of what depends on us which is closer in its formulation to Aristotle than the one in Nemesius. Thus he writes: ‘the soul has no master and acting or not acting depends on it’* (Didasc. 179.10–11; cf. Aristotle, EE 1226b30–1). Of interest to us now is what we find shortly afterwards: The nature of the possible falls somehow between the true and the false, and being by nature undetermined, that which depends on us uses it (i.e. the possible) as a vehicle. Whatever happens as a result of our choosing will be either true or false. (Didasc. 179. 20–3)³³

and then again: The possible . . . is undetermined, and it takes on truth or not depending on the inclining in either direction of that which depends on us. (Didasc. 179.31–3)³⁴

This can be read as the abbreviated version of an argument in which the idea of undeterminedness as found in De Interpretatione 9 is applied to the concept of what depends on us. The link with the De Interpretatione is suggested by the

³² Cf. e.g. Dillon 1993, 160–4. Alcinous flourished sometime between the first and third century . I assume the second half of the second century or the early third century as a likely date, but he may be even later: Göransson 1995. * ἀδέσποτον οὖν ἡ ψυχὴ καὶ ἐπ᾽αὐτῇ μὲν τὸ πρᾶξαι ἢ μή. ³³ ἡ δὲ τοῦ δυνατοῦ φύσις πέπτωκε μέν πως μεταξὺ τοῦ τε ἀληθοῦς καὶ τοῦ ψεύδους, ἀορίστῳ δὲ ὂντι αὐτῷ τῇ φύσει ὥσπερ ἐποχεῖται τὸ ἐφ᾽ἡμῖν. ὃ δ᾽ ἂν ἑλομένων ἡμῶν γένηται, τοῦτο ἔσται ἢ ἀληθὲς ἢ ψεῦδος. ³⁴ Τὸ δὲ δυνατὸν . . . ἀορισταῖνον δὲ τῷ ἐφ᾽ ἡμῖν κατὰ τὴν ἐφ᾽ ὁπότερον ῥοπὴν λαμβάνει τὸ ἀληθεύειν ἢ μή. As Whittaker 1990, 134 n. 424 (in his comments in the Budé edition) has pointed out, the whole section reflects Aristotelian thought and terminology.

      - 

33

repeated reference to truth and falsehood, which is entirely absent in the relevant chapters of the Nicomachean and Eudemian Ethics. The argument could have run like this: future contingents are not yet true or false now. It is so far undetermined whether the event announced in them will or will not happen. If we choose that it happens, it will happen. If we choose that it does not happen, it will not happen. Once we have chosen, circumstances permitting, the corresponding propositions will be either true or false. This implies that at least up to the point of the decision it is not predetermined which way we choose.³⁵ A similar thought can be detected in Ammonius at the beginning of his comments on De Interpretatione 9. There he maintains that the logical principle that states the indefiniteness of the truth values of contrary pairs of propositions about the future is necessary for ethics. It is needed so that it can depend on us whether we choose or do not choose and perform or do not perform certain actions. Ammonius’ argumentation in this passage implies that prior to our decision it is not yet fully determined which way we decide (see Ammonius, Int. 130.23–33, quoted below). So, two passages, one of them presumably earlier than Alexander, suggest that an indeterminist two-sided concept of what depends on us was developed in the context of the exegesis of Aristotle’s De Interpretatione 9. However, two things should be noted. First, in both passages no mention is made of causation: that is, we do not know whether the indeterminedness was understood as the absence of causal factors—as opposed to the ‘necessitation’ of the future by logical determinism which leaves it open in which way the future is determined by the past or present. However, we know that some Stoics, presumably from the second century, discussed Aristotle’s De Interpretatione 9.³⁶ Since in Stoic philosophy the connection was drawn between logical and causal determinism, it is thus likely that the link with causation was made at least by some philosophers in the second century. Second, even if there is no predetermination (causal or otherwise) of a human choice or action, within the context of ancient theory of causation which allows for objects to be causes this need not imply that the choice or action is causally undetermined. (That is, it could be freedom of type 4.) For example, if the assumption is that there is a core of personality or moral character in each person, then even if the decision is not predetermined, it may be determined at the time of the decision by the fact that the person is such and such a person. That is, the person would still in identical circumstances always go for the same action.

³⁵ The two passages quoted are most remarkable also for the fact that in them eph’hēmin does not refer to the things that depend on us, but to some kind of decision-making faculty. ³⁶ Cf. Alex. Fat. 177.7–14; Boethius, Int. 2.208.1–3. In the Alexander passage the example of the sea battle draws the link to De Interpretatione 9. The terminology and theory are Stoic, although presumably not early Stoic, and Alexander connects the argument with the theory of fate.

34

, ,   

9. From the Middle Platonists to Alexander of Aphrodisias I leave it open whether the Middle Platonists understood their two-sided concept of eph’hēmin as implying undeterminedness or non-predetermination by causal factors. My concern is rather whether Alexander was familiar with the secondcentury theory of eph’hēmin as found in some of the Middle Platonist texts on fate, and whether it is likely that his concepts (which I sketched in section 2) could have developed from it. There is some evidence that this was so. To begin with, although we do not have the distinction of things contingent into ‘for the most part’, ‘for the lesser part’, and ‘in equal parts’ in his treatise On Fate, we do find it in his commentaries on the Prior Analytics and the Topics, and in the Mantissa ch. 22. In his comments on Aristotle, Top. 112b1, the threefold distinction occurs preceded by the distinction between necessity and contingency (Alex. Top. 177.19–27), just as in Nemesius (Nat. hom. 103–4) and [Plutarch] (Fat. 571b–d). In his comments on Aristotle, An. Pr. 32b8, the threefold distinction is introduced by ‘one meaning of “the contingent” is the following; to this belong also the things that come to be in accordance with deliberate choice’ (Alex. An. Pr. 162.31–2). So Alexander, too, connects the threefold distinction with Aristotle’s concept of deliberate choice. However, unlike our Aristotle scholar from the Middle Platonist texts, he does not equate things that happen in accordance with deliberate choice with the things ‘in equal parts’. Rather, he classifies them among things for the most part (to epi to pleiston, cf. also Alex. An. Pr. 169.6–9; 270.23–5). For things ‘in equal parts’ we obtain the example of Socrates’ taking a walk in the evening, and his talking to some particular person. This fact, together with the absence of the threefold distinction in his On Fate, and of that which depends on us in his commentaries, suggests to me (i) that Alexander draws not from the Middle Platonist texts, but from a source they, like him, drew from,³⁷ and (ii) that he did not connect the threefold distinction with the question of determinism and freedom at all. In the above passages on things for the most part he is not concerned with acts of choice, but with the relation between one’s choices and their realization:³⁸ far more often than not one actually does what one has chosen to do. Alexander’s two examples for what happens ‘in equal parts’, on the other hand, invite a statistical interpretation. They are both cases in which one cannot easily say that Socrates does them more often than not, or less often than not.³⁹

³⁷ On this point I agree with Bob Sharples; see Sharples 2001. ³⁸ See also Sharples, ibid. ³⁹ The story undergoes a still different twist in the Mantissa ch. 22, which seems not to present Alexander’s own view as held in On Fate, but an alternative Peripatetic one. Here a twofold distinction of what happens for the most part and what happens for the lesser part comes up in the context of the discussion of that which depends on us. The things in accordance with deliberate choice are partly subordinated to what happens for the most part (when one acts ‘in character’, as it were), partly to what happens for the lesser part, and each is connected with its own type of eph’hēmin (see also below,

      - 

35

Thus Alexander, though acquainted with the concept of the ‘in equal parts’, seems not to have made the connection between it and that which depends on us. Nonetheless, it seems likely that he knew the account of what depends on us which in Nemesius was identified with that of the ‘in equal parts’; that he understood it as implying indeterminism; and that it was a precursor of his own concept. If we trust Alexander’s own words, in his On Fate ch. 26 (196, 24ff.) he presents one of a number of arguments of his opponents which were meant to criticize ‘that that which depends on us is such as the common conception of human beings takes it to be’. The main point of the argument is the claim that a two-sided eph’hēmin would preclude virtues and vices depending on us, since at the time when we are virtuous we are not capable of acting viciously, and vice versa. The argument begins: If, they say, those things depend on us of which we are capable of also the opposites . . . ⁴⁰

This is almost exactly the definition of what depends on us as we find it in the Middle Platonist texts. What shall we make of this? It seems to me that the most natural conjecture would run somewhat as follows. These critics of the ‘Middle Platonist’ two-sided concept of what depends on us were most probably the Stoics Alexander criticizes most in his book, i.e. orthodox Stoics who in the second century  held a theory of fate similar to that of Chrysippus’. We know that the Stoic doctrine of fate had been the subject of criticism by the Middle Platonists in their treatise(s) on fate.⁴¹ These Middle Platonist critics, as we have seen, had adopted a two-sided, potestative concept of what depends on us, based on Aristotle’s writings—although this was not their only concept of what depends on us (see below, section 12). So it is likely that some second-century Stoics in turn criticized this two-sided concept, and that this is what we find in the Alexander passage. However, we cannot rule out with certainty that Alexander’s opponents in this chapter are not Stoics but ‘dissident’ Peripatetics. If the argument Alexander presents is Stoic, we can infer that the concept of eph’hēmin was at least un-predeterminist, if not indeterminist. For only then is it incompatible with the Stoic theory of fate and would give the Stoics reasonable grounds to reject it. In that case there would be evidence that, before Alexander, an un-predeterminist two-sided concept of what depends on us was discussed among Middle Platonists (or Peripatetics) and Stoics. If the argument was part of a dispute internal to the Peripatetic school, the criticism of the two-sided, potestative

section 12). This time the connection is between one’s character, etc. and one’s choices, not between one’s choices and one’s actions. ⁴⁰ Εἰ, φασίν, ταῦτά ἐστιν ἐφ᾽ἡμῖν, ὧν καὶ τὰ ἀντικείμενα δυνάμεθα . . . ⁴¹ [Plut.] Fat. 574e–f, Calcidius, Tim. 160–1.

36

, ,   

eph’hēmin need not have had anything to do with the question of determinism. It could merely have been a way of pointing out that the definition does not harmonize with Aristotle’s own claim that virtues and vices depend on us. In any event, we can be reasonably certain that Alexander understood the criticized account of what depends on us as not only un-predeterminist, but indeterminist. For his criticism of the argument (which extends to ch. 29 of his On Fate) is one of the passages where he undoubtedly defends an indeterminist concept of what depends on us (Fat. ch. 29, 199.29–200.7, see section 12).⁴² Thus Alexander interpreted the ‘Middle Platonist’ concept of what depends on us as indeterminist. However, in Alexander’s standard accounts of what depends on us there are two additional elements that are absent in the ‘Middle Platonist’ account: ‘depending on us’ is predicated of the things over which we have in us the power of also choosing the opposite. (Fat. 181.5–6)*

Thus, the two new features are: (i) the element of choosing or not choosing (helesthai) to perform an action, instead of simply acting and not acting; (ii) the introduction of a power (exousia) which the individual on whom something depends possesses.⁴³ Both features are significant in that they reflect important developments of the understanding of moral responsibility and its relation to freedom in later antiquity.

10. The Element of Choice in the Accounts of What Depends on Us The first important innovation in the account, which we witness not only in Alexander but also in some later authors, is that from action to choice. In ⁴² Could Nemesius’ source have taken this account from Alexander, rather than the other way about? This is chronologically possible, but unlikely. For in Nemesius the account occurs twice as an account of the ‘in equal parts’ (which suggests it was a technical definition), and in one of these cases what happens ‘in equal parts’ is then equated with what depends on us. Moreover, it would be rather odd if the one account of what depends on us which Nemesius picked from Alexander is taken from an argument by Alexander’s opponents. * τὸ ἐφ’ ἡμῖν ἐπὶ τούτων κατηγορεῖται, ὧν ἐν ἡμῖν ἡ ἐξουσία τοῦ ἑλέσθαι καὶ τὰ ἀντικείμενα. ⁴³ As far as I am aware, in Alexander the two-sided ‘Middle Platonist’ account of what depends on us (i.e. the simple account, without reference to choosing and/or power) occurs only in the opponents’ argument in ch. 26. Of the four closest passages one contains exousia (Fat. 180.2), one the verb ‘to choose’ (Mant. 174.32); the remaining two (Mant. 170.1; 171.24) are from ch. 22 of the Mantissa which in any event presents a position quite different from Alexander’s in his On Fate and Mantissa ch. 23; of these Mant. 171.24 does not provide an account either. However, the fact that the two first-mentioned passages are so close to the ‘Middle Platonist’ one, and each only adds one of Alexander’s additional features, makes it the more likely that Alexander’s accounts are a development of the more basic ‘Middle Platonist’ one that resulted from Aristotle exegesis.

      - 

37

Alexander’s accounts it manifests itself in the change from ‘the power of doing opposites’ to ‘the power of choosing opposites’. We find variations of the formulation with ‘to choose’ (usually haireisthai) many times over in his On Fate.⁴⁴ We also find such formulations in chapters 22 and 23 of the Mantissa.⁴⁵ Similar accounts are preserved in Ammonius’ On De Interpretatione (130.30–2, on Arist. Int. 9, quoted below), in Boethius’ On De Interpretatione 2 (203, on Arist. Int. 9),⁴⁶ in the later paraphrase of the Nicomachean Ethics 52.25–7, a passage to which I referred earlier, and in Nemesius, Nat. hom. 115.22–7, a passage whose origin I assume to be later than Alexander’s On Fate. A comparable explanation is preserved in Calcidius, Tim. 151. There were three main philosophical theories concerned with human choice available to second- and third-century philosophers, all of which are possible influence factors, one deriving from Aristotle, another from Epictetus, a third from Plato. I suspect that a combination of them is responsible for the introduction of ‘choice’ into the accounts of what depends on us. None of the three positions was originally concerned with freedom of decision or indeterminist freedom in general. First, I have traced above the adaptation of Aristotle’s theory of deliberate choice (prohairesis) from his ethics into the debate over fate: his concept of deliberate choice was connected with that which depends on us by the commentators, in [Plutarch]’s On Fate (571d), by Nemesius, and by Alexander (e.g. Fat. ch. 12, Mant. ch. 22). For Aristotle, deliberate choice is what distinguishes human, rational agency from animal action. Its characteristic feature is that it is a certain appetitive state of the soul which results from deliberation about possible courses of action. Whether we deliberate well, and what the outcome of our deliberation is, depends on our character or settled dispositions. There is no evidence that Aristotle maintained that the same agent in the same circumstances could come up with a different choice (prohairesis). Moral responsibility is grounded on the fact that the agents are the beginning (archē) of their actions (and dispositions). Second, Epictetus, spelling out parts of early Stoic philosophy, restricts that which depends on us to certain ‘mental events’ or movements of the soul. Only the use of our impressions, that is giving assent to them or withholding it, depends on us, since these are the only things not subordinate to external force or hindrances. Assenting to impulse-prompting impressions (phantasiai hormētikai), i.e. impressions of something as desirable or to be avoided, is choosing a course of action. The realization of what we have chosen to do does not depend on us, insofar as it is always possible that it is thwarted by external hindrances. The stress in Epictetus is ⁴⁴ See e.g. Fat. 180.26–8; 181.5–6, 13–14; 184.18–19. ⁴⁵ In ch. 23, which seems to present Alexander’s view, there is no definition of that which depends on us with ‘to choose’, but choosing plays a major role passim; e.g. Mant. 174.9–12; 175.23–5. In ch. 22 prohaireisthai is used instead of haireisthai (Mant. 171.22–4; 172.10–12). ⁴⁶ Ex libero arbitrio, ut quod possum et velle et non velle, an velim hoc antequam fiat incertum est.

38

, ,   

on the points that it is oneself who chooses, and that one is not necessitated (katēnagkasthai) in one’s choice. To what impressions we give assent depends on our dispositions (or prohairesis, see section 11). The question of whether the same person in the same circumstances could choose otherwise is not addressed. I believe that—in harmony with the orthodox Stoic view—Epictetus’ answer would have been ‘no’.⁴⁷ If a person wants to act in a different manner than they do, they have to change their disposition or prohairesis, i.e. that factor on the basis of which they make individual choices. Epictetus emphasizes that moral accountability is—primarily—connected with the use of our impressions (e.g. Diss. 1.12.34) rather than with our actions. We are morally responsible because it is in our assenting and choosing that our character and dispositions are reflected (MR1). The influence of Epictetus on philosophers and intellectuals in later antiquity was immense, and at the beginning of the third century various elements of his philosophy had been absorbed into the general philosophical discussion, including Christian and Platonist thought.⁴⁸ Third, Plato may have provided a further motive for the change from action to choice. As I mentioned above, the Middle Platonist philosophers arranged their doctrine of fate around a number of passages from Plato. One of them comes from the Myth of Er in Book 10 of the Republic. There, the souls, before they are born again, have to choose a life, and in that context they are told that the consequences of their choice, whether good or bad, will be their responsibility, and that they cannot blame god: aitia helomenou: theos anaitios (Rep. 617E). For Plato, in this passage, the question was not one of freedom of decision. His concern was that the human soul and not someone else—in particular not god—is responsible for the choice (MR1). From the second century onwards, mainly in Platonist texts, the above quote from Plato occurs so regularly that we can infer that it, and with it parts of the Myth of Er, were a central element of the Platonist theory of fate.⁴⁹ In some texts that present the Middle Platonist theory of hypothetical fate,⁵⁰ Plato’s theory undergoes a significant development. In Alcinous (Didasc. ch. 26, 179.8–13) Plato’s formerly ‘pre-natal’ choice of a life is presented as including the choice of individual actions in one’s life, and it has become dependent on the soul whether or not to act. In Nemesius the term prohairesis has entered the interpretation of Plato’s statement: now the individual choices (prohairesis) and some of the actions in accordance with choice (kata prohairesin) depend on us (Nat. hom. ⁴⁷ I argue this point more fully in Bobzien 1998, §7.1. ⁴⁸ Epictetus’ philosophy was known e.g. to Dio Chrysostom, Philostratus, Lucian, Celsus, Marcus Aurelius, Gellius, Galen, Origen, Gregory of Nazianzus, and Augustine. ⁴⁹ Cf. e.g. Calcidius, Tim. 154, Hippolytus, Ref. 19.19 (Doxogr. Graec. 569.19–22), Nemesius, Nat. hom. 110.7–9, Maximus of Tyre 41.5a, Justin, Apol. 44, Porphyry apud Stobaeus, Ecl. 2.164; see also Tacitus, Ann. 6.22. ⁵⁰ This theory, the earliest traces of which are preserved in Tacitus, Ann. 6.22 and Plut. Quaest. Conv. 740c, maintains that certain human activities are not fated but caused by the person, whereas the consequences of these activities are fated. Cf. e.g. Den Boeft 1970, 28–34.

      - 

39

110.5–9; cf. 109). Neither text suggests that prohairesis or haireisthai refers to freedom of decision. Rather, the importance of the introduction of individual choices lies in the fact that it is in their choices that people manifest themselves qua rational or moral beings: my choices, since determined by nothing but myself, reflect who I am. This is why I am morally responsible for what I choose (MR1). It is in order to ensure this that choices have been exempted from the predetermination by fate (freedom of type 4). In the parallel passage in Calcidius (Tim. 151), on the other hand, we find in this context a statement that connects that which depends on us with the motions of the soul only, and which invokes a notion that comes close to that of freedom of decision: in it the fact that the choice between opposite motions of the soul is in our power is used to explain why these motions depend on us. Still—as in Plato—the choice is one between good or bad.⁵¹ The Middle Platonist interpretations of Plato, with their focus on individual choices of actions, are likely to reflect the general focus on choices and mental events which seems to have started at the time of Epictetus or a little earlier.⁵² Returning to Alexander and the other authors who add choice into the account of that which depends on us, we can note two things: first, the introduction of choice appears to result from a combination of the three possible influence factors, Plato, Aristotle, and Epictetus. Second, the motivation for adding choice into the account appears not to have been an attempt to express freedom of decision. Alexander knows and uses both a concept of prohairesis of the Epictetan/ Platonic type, as moral choice (Fat. 169.12), and the Aristotelian one of deliberate choice (in the majority of places, e.g. Fat. 180; 194–5; 212). How exactly Alexander thought these concepts of choice linked up with ‘to choose’ in his account of ‘power to choose opposites’ is uncertain.⁵³ Mostly the expressions appear to be understood as non-moral and as the result of deliberation, i.e. in the Aristotelian sense. However, there seems to have also been a distinctly non-Aristotelian element involved in the introduction of choosing into the accounts of what depends on us. In Alexander, the accounts containing the verb ‘to choose’ (haireisthai) are

⁵¹ Collocati autem in alterutram partem (i.e. good or bad) meriti praecessio animum nostrarum motus est iudiciumque et consensus earum et appetitus vel declinatio, quae sunt in nobis posita, quoniam tam horum quam eorum quae his contraria sunt optio penes nos est. ⁵² On the one hand, from the second century onwards sources that discuss determinism seem generally to concentrate more on mental activities like thinking, deliberating, assenting, and choosing. For instance, the Chaldaeans listed such mental states and events among the things they claimed were predetermined by the stars (e.g. Gellius, NA 14.1.23, cf. Nemesius, Nat. hom. 104.18–21). These explicit mentions of the predetermination of the motions of the soul may have triggered their explicit exemption from external causal predetermination or force on the side of the ‘libertarians’. On the other hand, there is the importance of the problem of choice (prohairesis) between good and evil in early Christian theory, Platonism, and Gnosticism, in particular in the context of the question of the origin of evil. ⁵³ The terms in the accounts are hairesis/haireisthai, not prohairesis/prohaireisthai, but Alexander also uses haireisthai to refer to Aristotle’s deliberate choice (Fat. ch. 11).

40

, ,   

apparently not regarded as a substitute for those containing the verb ‘to act’ (prattein), but rather as a supplement. Not only do we find both kinds of accounts several times, we also regularly find choosing and acting co-ordinated in one phrase or account.⁵⁴ We find the same juxtaposition in Nemesius (Nat. hom. 115.22–28; 116.3–5) and in Ammonius (Int. 130.30–32). These latter authors provide a reason why action as well as choice is considered: action presupposes choice, and praise and blame concern both action and choice: both are culpable (Nemesius, Nat. hom. 115.27–28, Ammonius, Int. 130.32–3); moreover, sometimes we are prevented from realizing our choices (Nemesius, Nat. hom. 116.3–5). This suggests that the switch from action to choice, or rather the addition of choice to action, was motivated by a change of focus regarding what is of primary moral relevance: choices rather than actions. Here Stoic, and in particular Epictetan, thought appears to have been influential, possibly via the Middle Platonist reinterpretation of Plato’s Myth of Er. This may be the most promising conjecture of why in Alexander the account of eph’hēmin so frequently includes the term ‘choice’. Alexander states, for instance, in a similar vein, ‘the assessment of morally right action is made not only from the things that are done, but much rather from the disposition and capacity from which it is done’ (Alex. Fat. 206.16–18). Thus it seems that the origin of the term ‘to choose’ in the account of eph’hēmin is non-Peripatetic, although Alexander then generally interprets it in the Aristotelian sense, as choice that is the result of deliberation and not as fundamental moral choice. Taking the various points together, it seems that the initial grounds for the inclusion of choice in the accounts of what depends on us in Alexander, Ammonius, Boethius, Heliodorus, and Nemesius are unlikely to have been the quest for an indeterminist concept of freedom of decision (as opposed to freedom of action), or the question of whether people are causally indetermined in their choices between alternatives. Rather it is the recognition of choice as the specific activity through which human rational beings can have an influence in the world, and accordingly, to which moral appraisal is to be attached. (The issue was autonomy rather than freedom to do otherwise.) This is perhaps further corroborated by the fact that in several of the passages in Alexander that are most clearly indeterminist (see next section) the version of the account with ‘to act’, and not the one with ‘to choose’, occurs.

11. The Term exousia in the Accounts of What Depends on Us On the other hand, the second change in Alexander’s account—from ‘being capable of doing and not doing something’ to ‘having the power (exousia) of ⁵⁴ Fat. 181.14: ἐξουσίαν τῆς αἱρέσεώς τε καὶ πράξεως τῶν ἀντικειμένων; cf. Fat. 179.3.11; 189.10–11; Mant. 174.4; 175.24–5; 180.28–31.

      - 

41

doing and not doing something’—appears to be pertinent to the development of a concept of freedom of decision. Formulations of the account with exousia occur as standard in Alexander’s On Fate (thirty-three instances according to Thillet’s index) and in Mantissa ch. 23, and there can thus be little doubt that the use is philosophically motivated. Alexander seems to be the first—of whom we know— to use the term exousia in this kind of account of what depends on us.⁵⁵ It further occurs in Quaestio 3.13 of the Quaestiones ascribed to Alexander; in Nemesius (Nat. hom. 112.10; 115.25) in a passage I believe to be later than his reports from the Middle Platonist theory of fate; in Ammonius (Int. 148.14.23); and in slightly different wordings in Iamblichus (Stobaeus, Ecl. 2.173.21) and Simplicius (Ench. 20). How can we explain the appearance of exousia in the accounts? On this question, I can only offer conjecture. First, since the term exousia seems to have replaced the verb dunasthai in the account, and this verb was linked with Aristotle’s two-sided dunamis, exousia may have been meant to stand in for Aristotle’s two-sided, rational capacity from Int. 13 and Metaph. 9 (see section 6). Second, in Alexander exousia tēs haireseōs/ exousia tou haireisthai could take the place filled in other late second- and thirdcentury authors by the phrase prohairetikē dunamis.⁵⁶ This phrase in turn seems to be a descendant of prohairesis in the Epictetan sense and which Epictetus himself already used in place of prohairesis throughout in Diss. 2.23. (But it also experiences an Aristotelian interpretation, cf. e.g. Nemesius, Nat. hom. 119.11.) For Epictetus prohairesis does not refer to a person’s particular choice in a certain situation. First and foremost he uses the term to denote a disposition of the human mind which determines a person’s individual choices. The exertion of this disposition is the only thing that is never necessitated by external circumstances. What we choose thus depends on us.⁵⁷ If this is where exousia in the accounts comes from, it may refer to a specifically human disposition for making choices. Third, Alexander uses exousian (en hēmin) echein tou . . . as virtually synonymous with (hēmeis) kurioi tou . . . The latter formulation occurs about a dozen times in this context. This fact may provide another link to Aristotle’s Ethics (see Alex. Fat. 178.26–8; 180.9–12), but formulations with kurioi for eph’hēmin are standard in practically all schools. There could also be a link between Alexander’s use of echein exousian and Epictetus, who uses it to say whether we or some external influence have control over certain things. Fourth, the Middle Platonist Maximus of Tyre uses exousia twice in speech 41, in the context of explaining how vice entered the

⁵⁵ We do not find exousia in this context in Aspasius, the Anonymous on the Nicomachean Ethics, Alcinous, [Plutarch], On Fate, (nor in Justin and Tatian) nor in Mantissa ch. 22 or Alexander’s commentaries on the Topics and Prior Analytics. ⁵⁶ Clement, Strom. 6.135. 4 (500.20–1 Staehlin); cf. Nemesius, Nat. hom. 119.4–5.11. ⁵⁷ Epictetus’ prohairesis can perhaps be described as a precursor of a concept of the will. However, it is neither a separate faculty or part of the mind (i.e. one that can be in conflict with other faculties or parts of the mind), nor is it free in the sense of being causally independent in its choices.

42

, ,   

world: it is this power of the soul (exousia tēs psuchēs) which enables us to do bad things (Orat. 41.5a and g). More important than where exactly the use of the term exousia originates is the particular way in which the various influences are combined. It is the synonymy with kurios which best shows the significance of the replacement of dunasthai by exousia. ‘Having the exousia of acting (choosing) and not acting (choosing)’* can be understood in two different ways. Compare the sentences 1) the king has the power (authority, control) over living and dying (life and death) 2) the king has the power (ability, capacity) to live and to die Similarly, the above sentence can be understood as 1) we have the power (authority, control) over acting/choosing and not acting/choosing 2) we have the power (two-sided ability, capacity) to act/choose and not to act/choose. In the cases of type (1), with genitivus obiectivus, where someone has the power, authority, or control over certain things, we can separate the person who has the power from the things over which they have the power in a way that cannot be done in cases of type (2). The king’s power over living and dying can be concerned with other people’s lives. The king’s power to live and die is concerned with his own condition. In case (1), the agent becomes a ‘decision maker’, in case (2) this is not so. The synonymity of ‘having the exousia’ and ‘being kurios’ over something in Alexander suggests that we have case (1) in his accounts of what depends on us. Something depends on us if we are in control over doing/choosing and not doing/choosing it. This is noticeably different from the earlier formulation with dunasthai: there, clearly a two-sided capacity was at issue. But owing to the introduction of ‘doing/choosing something or its opposite’, kurios and exousia do not function in the same way anymore as they did in Aristotle and Epictetus: in the latter authors it was the fact that nothing hindered us from doing or choosing something that made us have control over them. In Alexander’s account, the terms are (at least at times)⁵⁸ understood differently: what makes us have control over things is the fact that we are causally undetermined in our decision and thus can freely decide between doing/choosing or not

* ἔχειν τὴν ἐξουσίαν τοῦ πράττειν (αἱρεῖσθαι) καὶ μὴ πράττειν (αἱρεῖσθαι). ⁵⁸ See the next section. Note that the phrase ἐξουσία τοῦ πράττειν καὶ μὴ itself can also be understood as ‘not being hindered either way by external or internal factors’; and also as ‘having a general two-sided capacity to act’. In neither case would freedom of decision need to be involved.

      - 

43

doing/choosing them. The element of free decision in Alexander’s account thus lies not in the addition of the phrase ‘choosing or not choosing’, but in the introduction of the term ‘exousia’. We can thus see that the change to exousia in the account may have been of great significance, since it provided a way to express that the agent is a causally undetermined decision maker. The introduction of the term into this context is then also one further step towards the concept of a free will, since such a concept requires an independent faculty of decision-making. It may also be worth considering in this context the relation between the expressions exousia and autexousion. A link between them, although in a deterministic setting, can be observed already in Epictetus.⁵⁹ Something is in my own power (autexousion) if I have power (exousia) over it in the sense that nothing can prevent me from doing it. Epictetus clearly contrasts autexousios with someone else’s exousia over oneself. Autexousios indicates that something is outside the sphere of influence of others, and because of that in the sphere of my power. I suggest that a similar relation was assumed by Alexander and other authors who favoured an indeterminist concept of eph’hēmin: for from the late second century onwards autexousion became more and more common as a philosophical term used instead of eph’hēmin, and at Nemesius’ time seems to have superseded it. Alexander uses it very rarely, but in one place he states that autexousion is what is actually meant by eph’hēmin and that his opponents miss this meaning of the term (Fat. 182.22–4; cf. 189.9–11). Thus something may have been considered as in someone’s own power (autexousion), and as truly eph’hēmin, precisely if that person has the exousia over doing/choosing it or its opposite. Autexousion may then have been understood by some as implying indeterminist freedom of the agent.

12. The Volatility of the Concept of Freedom to Do Otherwise Thus it seems that in Alexander’s accounts of what depends on us it was rather the expression exousia than haireisthai that served to express the element of freedom of decision. We saw at the beginning that Alexander had a concept of freedom to do otherwise. We have now seen how this concept developed, absorbing both Stoic and Aristotelian and, perhaps, Platonic elements on its way. However, we would be quite wrong to assume that at the turn of the second century a general awareness of the problem of causal determinism and freedom to do otherwise had arisen, and that it had become part of the philosophical standard repertory of the time. There are several points that suggest that at his time Alexander is almost an isolated case, and that concepts of freedom to do otherwise are a rather marginal phenomenon without a clear philosophical context.

⁵⁹ Cf. Diss. 1.25.2; 4.1.62; 4.1.68; 4.7.16; 4.12.8; see also Bobzien 1997, 82 with n. 48 and 86 n. 59.

44

, ,   

First, it is noteworthy that the one-sided, causative conception of what depends on us was by no means peculiar to the Stoic system, nor generally seen as a feeble attempt of the Stoics to nominally save moral responsibility—even if Alexander wants to make us believe this (Fat. ch. 13). On the contrary, it seems to have been regarded as a serious alternative or as a complement to the two-sided, potestative conception in second- and third-century Middle Platonist and Peripatetic writings. We find non-Stoic accounts of such concepts in [Plutarch], On Fate (‘That which depends on us is that part of the contingent which is already happening in accordance with our impulse’),⁶⁰ in the Mantissa, and in Nemesius. Had the general concern at the time been to preserve freedom to do otherwise as a prerequisite for moral accountability, this repeated approbation of a one-sided, causative eph’hēmin would be decidedly odd. On the other hand, if we assume that the two-sided, potestative concept was considered to express a two-sided general capacity which provides the vehicle through which rational or moral agents manifest themselves in their actions, this fact is far less startling. For in that case both the one-sided concepts and the two-sided one serve to ensure that the agent is causally—and hence morally—responsible for the action, if in slightly different ways. (Remember that the one-sided concepts seem to have assumed a two-sided general capacity through which the agent causes the action.) On this assumption it is also not surprising to find that the Neoplatonist conception of what depends on us is one-sided and causative (cf. Plot. Enn. 3.1.9–10; 3.2.10; 6.8.7), and that this fact seems not to have outraged anyone at the time. The philosophical origins of the non-Stoic accounts of a one-sided, causative eph’hēmin are nowhere explicitly stated. The accounts in Nemesius and in the Mantissa show a striking resemblance to the later Stoic account which defines that which depends on us as that which happens through us (di’ hēmōn, see above, sections 2 and 3). Thus Nemesius writes: We say that, generically, all things that are done voluntarily through us depend on us . . .⁶¹

and the Mantissa has: For those choices of which the cause is nature or upbringing or habit are said to depend on us in the sense that they happen through us.⁶²

⁶⁰ Τὸ δὲ ἐφ᾽ἡμῖν θάτερον μέρος τοῦ ἐνδεχομένου, τὸ κατὰ τὴν ἡμετέραν ὁρμὴν ἤδη γινόμενον ([Plut.] Fat. 571d–e). ⁶¹ Λέγομεν τοίνυν γενικῶς πάντα τὰ δι᾽ἡμῶν ἐκουσίως πραττόμενα ἐφ᾽ἡμῖν εἷναι . . . (Nemesius, Nat. hom. 114.15–16, cf. 102). ⁶² ὧν γὰρ προαιρέσεως ἡ φύσις ἢ ἀγωγαὶ καὶ ἔθη ἐστὶν αἴτια, αὗται οὕτως ἐφ᾽ἡμῖν λέγονται ὡς δι᾽ἡμῶν γιγνόμεναι (Mant. ch. 22, 172.7–9).

      - 

45

However, I surmise that these accounts—and the above-quoted account in [Plutarch]’s On Fate—originate from (incorrect) Aristotle exegesis, presumably from an interpretation of the Nicomachean Ethics Book 3, from passages like the following: We deliberate about the things that depend on us and that can be done; . . . All groups of human beings deliberate about the things that can be done through them . . . we deliberate about those things which come to be through us and not always in the same manner.⁶³ (my emphasis)

Thus, where Aristotle argues on the generic level of (possible) actions, the passages from [Plutarch] and Mantissa ch. 22 talk about individual happenings. (Nemesius is ambiguous here.) The unfamiliarity of second- or early third-century thinkers with an indeterminist concept of freedom to do otherwise is also beautifully illustrated by the awkward way in which it is handled in ch. 22 of the Mantissa, which seems to present a presumably Peripatetic alternative to Alexander’s position. Its charming solution to the problem of Stoic-style determinism lies in the introduction of ‘that which is not’ (to mē on) as an influence factor.⁶⁴ The argumentation runs roughly like this: If everything is caused, and the same causes have the same effects, and our choices are determined by a combination of our nature, habit, and upbringing, then our choices do not depend on us. But choice does depend on us, and accordingly not everything is caused. The reason for this is that a little bit of ‘not-being’ is mixed in with the earthly things. In particular it can be detected in the things responsible for that which happens for the lesser part. This ‘not-being’ weakens the things or causes in which it exists, and thus weakens the continuity of causes. In things external to us this fact leads to chance events. In the causes in us, i.e. in our nature and habit, it leads to that which depends on us in the proper sense (kuriōs). Whenever the ‘not-being’ in us is responsible for a choice of ours, then that choice depends on us in the proper sense.

It is hard to see how, in this theory, moral responsibility is to be attached to the choices that depend on us. Their causal undeterminedness appears to render them some sort of random motions. Perhaps wisely, moral responsibility is not mentioned in the whole chapter. What is more, since the choices that depend on us in ⁶³ Βουλευόμεθα δὲ περὶ τῶν ἐφ᾽ἡμῖν καὶ πρακτῶν˙ . . . τῶν δ᾽ἀνθρώπων ἕκαστοι βουλεύονται περὶ τῶν δι᾽ αὑτῶν πρακτῶν . . . ὅσα γίνεται δι᾽ἡμῶν, μὴ ὡσαύτως δ᾽ἀεί, περὶ τούτων βουλευόμεθα . . . Arist. EN 1112a30–b4; cf. also 1112b27; 1111b23–4, 26. ⁶⁴ This is reminiscent of Plato’s struggle with not-being, e.g. in the Sophist; cf. also Cic. Fat. 18: ut sine causa fiat aliquid, ex quo existet, ut de nihilo quippiam fiat . . . ; Augustine, Lib. arb. 201–5.

46

, ,   

the proper sense result from a weakness or lack of tension of our nature and habit, something’s depending on us in the proper sense can hardly have been judged as a positive thing.⁶⁵ A third point that shows that a concept of freedom to do otherwise was far from securely established is that not only is there no unambiguous evidence for it before Alexander, but also in Alexander’s On Fate and in the Mantissa there is a steady vacillation between various concepts of what depends on us, some advocating freedom to do otherwise, others implying only the absence of any predetermination by external and/or internal causal factors, and still others that are clearly compatible with determinism.⁶⁶ We have seen above that Alexander’s phrases of the kind ‘having the power to do/choose opposites’ are ambiguous between determinist, un-predeterminist, and indeterminist readings. In this context Alexander’s encounter with the secondcentury  orthodox Stoic theory of fate becomes crucial: it is only where the twosided, potestative eph’hēmin meets with the Stoic causal principle (that in the same circumstances the same causes necessarily bring about the same effect, see above, section 2) that the phrases are disambiguated, and that a concept of freedom to do otherwise is uncontroversially in play. Generally, wherever Alexander considers the possibility that the same person in the same circumstances acts or chooses otherwise than they do, phrases like ‘having the power to do/choose opposites’ seem to acquire an indeterminist meaning. In particular, indeterminist freedom is almost certainly at issue in the important passages in which Alexander depicts the fictitious situation of someone who acts against their character, or against what seems reasonable to them, in order to show that determinism is wrong (Fat. chs 6 and 29, Mant. 174.33–5). Equally, the passage in which Alexander argues that our regret shows that we have the power to choose opposites suggests a concept of freedom to do otherwise. He says: For it is on the grounds that it was possible for us also not to have chosen and not to have done this that we feel regret and blame ourselves for our neglect of deliberation. (Fat. 180.29–31, trans. Sharples)⁶⁷

Here the concept of freedom of decision appears—finally—to be connected with a conception of moral responsibility based on the agent’s ability to do otherwise (MR2). A third important argument is that the same circumstances do not necessarily lead the same agent to the same actions/choices, because there are several—incommensurable—ends looking towards which we decide and choose ⁶⁵ This is very different e.g. in Cic. Fat. 11 and Alex. Fat. ch. 6, where the agents are envisaged as overcoming their nature or habit. ⁶⁶ This point has been discussed in Sharples 1975, 37–63. ⁶⁷ Similarly, but not as clear, Fat. ch. 19 on pardon and blame, where Alexander plainly goes beyond Aristotle, EN 3.1, and Fat. ch. 16.

      - 

47

(Fat. ch. 15, Mant. 174.17–24). All these arguments strike one as thoroughly modern, and as easy to grasp within the framework of today’s discussions of the ‘free-will problem’. Contrasted with these are the many Alexander passages with arguments which, for someone who expects a defence of freedom to do otherwise, simply seem to beg the question. However, most arguments make perfect sense as soon as one understands them as concerned not with indeterminist freedom but with different philosophical questions. There are first those passages in which Alexander basically contents himself with paraphrasing Aristotle, for instance where he describes the agent as causally responsible, or as a beginning (archē) of action (Fat. chs 15, 20; Mant. 173.10–21); similarly where he opens up the vexed questions of character determination and of one’s responsibility for the formation of one’s character (Mant. 175.9–32, Fat. ch. 27). Here Alexander does not go beyond Aristotle, leaving it open whether, when we begin forming our character, we are ‘free’ or our dispositions predetermined. Moreover, the whole question of one’s responsibility for forming one’s character makes most sense on the assumption that (at least in some situations) what one does is fully determined by what character one has. Finally, determinist reasoning, quite similar to Chrysippus’ position (cf. Cic. Fat. 7–9, 41–3), can be found in Mantissa ch. 23 (174.35–9). It suggests that, if at different times the same person chooses similar things, the reason is not that the circumstances are similar (and function hence as external necessitating causes), but because the person’s dispositions are similar each time. These remarks may suffice as an illustration that Alexander is by no means clear and consistent about whether his phrases like ‘having the power to do/ choose opposites’ are to be understood as indeterminist.⁶⁸ This may be partly explained by the fact that Alexander does not have a fully fledged concept of a faculty of a will, and a fortiori not one of a will that is free in that it can operate independently of the agent’s beliefs and desires. This is so despite the fact that he has collected all the ingredients required for a notion of acting from free will: he has endowed human beings with a two-sided power (exousia) of decision-making, which: • • • •

is not necessitated by external or internal influence factors; is exercised as the result of a process of deliberation; is envisaged as separable from the agent’s character, disposition, or nature; is envisaged—it seems—as separable from the agent’s reason: we can decide against what appears to us as the most reasonable course of action;

⁶⁸ Note also the strange restriction on indeterminism when he writes: ‘we have this power of choosing the opposite and not everything that we choose has predetermining causes because of which it is not possible for us not to choose this’ (Fat. 180.26–8, my emphasis). Calling to mind the above-discussed argument from Mant. ch. 22, this suggests that it is sufficient if in some of our choices there are no predetermining causes.

48

, ,    • leads to decisions that are not causally predetermined by internal or external factors, so that it is possible that the same agent, with the same desires and beliefs, in the same circumstances, chooses differently.

But all these points do not add up to ‘choosing and acting from free will’, since for Alexander the human soul is not separable from the body and is in principle susceptible to causal impacts. It remains unclear what the independent decisionmaking faculty would be which has the power over choosing opposites: it can hardly be one of the non-rational parts of the soul. But if it is a, or the, rational part, the difficulty arises how it can—as Alexander suggests—decide against the course of action that appears as the most reasonable one to the agent: either this was not the most reasonable course of action after all, or it is not a superordinate rational part of the soul that decides. Thus even if a decision is not necessitated, predetermined, or externally determined, there seems to be no suitable place for an independent decision-making faculty in Alexander’s conception of the soul. A full concept of acting from free will and a full awareness of the free-will problem in the narrowest sense are not developed in the context of the Stoic– Peripatetic debate over the compatibility of universal causal determinism and freedom to do otherwise. Neither the Stoics nor the Peripatetics experience within their systems any of the free-will problems listed at the beginning. The Stoics did not require a concept of free will, since they did not connect moral responsibility with freedom to do otherwise. As a consequence, they had no reasons to concern themselves with any free-will problem. Theirs is the problem of the compatibility of autonomous agency and causal determinism. On the Peripatetic side, Alexander had no free-will problem either. It is true, at least at times he seems to regard a concept of freedom to do otherwise as a prerequisite for moral responsibility. But he secures such freedom by simply denying predetermination of human actions. Unlike Stoics and Platonists, he can do so because he does not believe in universal divine providence. A free-will problem (in the wider sense) thus arises only in the confrontation of the two philosophical systems, when later Stoic causal determinism meets late Peripatetic freedom to do otherwise—with such freedom understood as a necessary condition for moral responsibility. If we want to find philosophers who are troubled by a free-will problem within their system, we need to turn to Platonists and Christian thinkers. In their theory of hypothetical fate the Middle Platonists had severed the Stoic chain of causes at the point of human choices and actions (see above, section 6). This was made possible by the fact that they proposed an immaterial human soul which can initiate action in the material world.⁶⁹ In this way they had gained unpredeterminist freedom, thus guaranteeing the agent’s autonomy. However, as

⁶⁹ Cf. e.g. Alcinous, Didasc. 153.4–5, ἡ δὲ πρᾶξις ψυχῆς λογικῆς ἐνέργεια διὰ σώματος γινομένη.

      - 

49

the Middle Platonists also advocated the universal impact of divine providence, the severance of the chain of causes did not solve all their difficulties. For human actions and choices, even if not the result of the network of causes, are still in accordance with divine providence. The problem of determinism is thus no longer that of predetermination by a chain of corporeal causes, but of predetermination by god’s providence, even if this does not work through the network of material causes. In particular the problem became dominant of how to bring into agreement the evil choices and actions of human beings with god’s providence, given that god is by definition good. Early Christian thinkers struggled with a similar question, despite considerable differences in their ‘metaphysics’, and they, too, had the advantage of an immaterial soul which made it possible that human action became independent of the network of material causes. It is in this context that finally a faculty of the will is introduced (no doubt influenced again by Epictetus’ concept of prohairesis), to warrant the independence of human evil deeds from god’s providence or creation. In which way this will is considered as free varies and is often hard to determine: indeterminist freedom of decision, un-predeterminist freedom, and freedom from force or compulsion (voluntariness) seem to alternate in our sources. Since the problem is no longer the independence of preceding material causes (this has simply been postulated), formulations of determinism of the kind ‘same (corporeal) causes, same effects’ are no longer fitting. As a consequence, an unambiguous description of the freedom involved in the various theories, whether indeterminist, unpredeterminist, or neither, becomes hard to find. Accordingly, it is seldom clear what kind of problem of ‘freedom’ of the will the philosophers were dealing with. After Alexander, the problem of determinism and freedom to do otherwise is most clearly present in the commentaries on Aristotle’s De Interpretatione, where—if I am right—lay one vital element of its coming into existence.

13. Results The problem of the compatibility of causal determinism and freedom to do otherwise appears to have been formulated only in the second century . This seems to have been the result of a confrontation of a refined Stoic universal causal determinism, on the one hand, with a two-sided, potestative concept of what depends on us (is eph’hēmin), based on Aristotle’s ethics, on the other. Presumably in the early second century, and as a consequence of combining Aristotle’s theory of deliberate choice with his modal theory and with his theory regarding the truth values of future propositions, this concept was interpreted as implying freedom to do otherwise. Who exactly was responsible for this new indeterminist understanding of that which depends on us is unclear, but it seems to have been accepted thereafter both by some Peripatetics and by some Middle

50

, ,   

Platonists. Alexander’s accounts, and some later ones, of that which depends on us display two further developments of this indeterminist concept of freedom. First, the addition of choice (hairesis) to action in the accounts reflects a refinement of theory of action and moral responsibility which focuses more on intra-psychic events and, in particular, on the choice of good or bad and the culpability of that choice. Here Stoic and Platonist impacts become apparent. Second, the replacement in the account of ‘being capable of ’ by ‘having the power or authority (exousia) over’ introduces a decision-making faculty and thus leads to a concept of free decision—the result, probably, of a fusion of Epictetan and Aristotelian elements. Alexander stops short of a concept of free will, due, it seems in part, to the fact that he believes the human soul to be inseparable from the corporeal item of which it is the form. The need of a free will becomes pressing in Platonist and Christian philosophy in the context of the problems of how vice entered the world and how god’s providence and foreknowledge of the future is compatible with human responsibility. But this is no longer in the context of a physical theory of universal causal determinism, characterized by principles of the kind ‘like causes, like effects’. Rather, the determinism is now teleological only, and the context theological. From the third century onwards, physical causal determinism was, it seems, no longer considered an attractive or plausible theory, let alone a threat. Overall, the problem of causal determinism and freedom to do otherwise appears not to have been a very prominent topic in antiquity. There is a certain likelihood of an awareness of it for some second-century Stoics who were confronted with the Peripatetic or Middle Platonist indeterminist concept of what depends on us, and for those Peripatetics and perhaps Middle Platonists who in turn criticized Stoic determinism. There is good evidence of it in certain passages of Alexander’s On Fate and in the Mantissa. Origen may have been aware of it. And it lingers on in later commentaries on Aristotle’s De Interpretatione. It is then presumably only a slight overstatement when I conclude by saying: the problem of physical causal determinism and freedom of decision entered the scene in the second century  by way of a chance encounter of Stoic physics and the fruits of early Aristotle exegesis, with the contemporary focus on the culpability of mental events and the introduction of a power of decision-making as catalysts— and it was not part of the philosophical repertoire for long.

2 Choice and Moral Responsibility in Nicomachean Ethics 3.1–5 1. Introduction This chapter serves two purposes: (i) it can be used as an introduction to chapters 1–5 of Book 3 of the Nicomachean Ethics (EN 3.1–5); (ii) it suggests an answer to the unresolved question of what overall objective this section of the Nicomachean Ethics has.¹ The chapter focuses primarily on Aristotle’s theory of what it is that makes us responsible for our actions and character. After some preliminary observations about praise, blame, and responsibility (section 2), it sets out in detail how all the key notions of Nicomachean Ethics 3.1–5 are interrelated (sections 3–9). The setting out of these interconnections makes it possible to provide a comprehensive interpretation of the purpose of the passage. The primary purpose of Nicomachean Ethics 3.1–5 is to explain how agents are responsible for their actions not just insofar as they are actions of this kind or that, but also insofar as they are noble or base: agents are responsible for their actions qua noble or base, because, typically via choice, their character dispositions are a causal factor of those actions (section 10). The chapter illustrates the different ways in which agents can be causes of their actions by reference to Aristotle’s four basic types of agent (section 11). A secondary purpose of Nicomachean Ethics 3.1–5 is to explain how agents can be held responsible for consequences of their actions (section 12), and in particular for their character dispositions insofar as these are noble or base, that is, virtues or vices (section 13). These two purposes are not the only ones Aristotle pursues in the passage. But they are the ones Aristotle himself indicates in its first sentence and summarizes in its last paragraph, and the ones that give the passage a systematic unity.² The chapter also briefly considers the issues of freedom-to-do-otherwise, free choice,

¹ This essay first appeared in R. Polansky (ed.), The Cambridge Companion to Aristotle’s Nicomachean Ethics (Cambridge: Cambridge University Press, 2014), 81–109 and is reproduced here with permission. I would like to thank Susan Sauvé Meyer, Carlo Natali, and Ronald Polansky for their most helpful comments on a draft version of this essay. ² These two purposes also tie in with Aristotle’s claim at EN 2.4.1105a28–33 that for actions to be performed in a virtuous way, (i) the agent has to choose them insofar as they are good (just, temperate, etc.) and (ii) the choice has to be based in a firm and unchangeable character disposition (i.e. in the virtues of justice, temperance, etc.).

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0003

52

, ,   

and free will in the contexts in which they occur (i.e. in the final paragraphs of sections 6, 7, 12, 13). The passage Nicomachean Ethics 3.1–5 is located at the end of Aristotle’s account of the character virtues and their acquisition in general in Book 2 and before his main discussion of the particular virtues in Books 3.6–5.11. In the passage, Aristotle considers—using traditional translations—praise and blame, voluntariness, choice, deliberation, wanting, and the things in our power.³ Aristotle never explicitly states the purpose of Nicomachean Ethics 3.1–5. It is a matter of debate what his intention in this passage was. According to some of the major views of Aristotle scholars, Aristotle intends 1. to provide the notions required for the discussion of the individual virtues and vices; 2. to show that praise and blame are justified for our being noble or base in our actions or our virtuous or vicious dispositions, because these are voluntary; 3. to capture the causal conditions of praise and blame; 4. to discuss intentional action; 5. to discuss the efficient cause of action; 6. to discuss Socrates’ view that no one is willingly bad.⁴ De facto, Aristotle does all of 1–6. I prefer a version of 2, for reasons expounded below. There is no dispute that, if anywhere in the Nicomachean Ethics Aristotle considers what renders people responsible for their actions and character and justifies praise and blame, he does this in 3.1–5. These chapters are also central for Aristotle’s theory of action. In addition, many scholars have espoused the view that Nicomachean Ethics 3.1–5 is pivotal in Aristotle’s works for his treatment of freedom of choice, free will, and freedom-to-do-otherwise.⁵ For all these reasons, our text remains one of the most discussed passages of Aristotle’s oeuvre.

³ EN 3.1–5 is one of the parts of the EN that is not shared with Aristotle’s Eudemian Ethics (EE) and where there is a long parallel treatment of mostly the same issues (EE 2.6–11). Although one should always allow for the possibility that Aristotle changed his mind or refined his theory in the interval between writing (and revising) the one and the other, the close similarities make EE 2.6–11 an invaluable source for better understanding EN 3.1–5. Any reader with a serious desire to master EN 3.1–5 is advised to study EE 2.6–11. (See the commentary in Woods 1992 and Meyer 2006 for helpful notes on the EE passage.) It is also worth keeping in mind that EN 3.1–5 is bound into the whole treatise of the EN and involves many philosophical notions and distinctions that Aristotle introduces elsewhere in the EN (especially in Books 1, 2, 6, and 7), as well as in other writings, and that these passages also help in understanding EN 3.1–5. Furthermore, EN 5.8 also overlaps in content with EN 3.1. ⁴ See e.g Grant 1874, vol. 2, iii, Taylor 2006, 125 for 1; Irwin 1985, 315, 319 for 2; Meyer 2006, 137–8, 139, 144, 152, 154 for 3; Charles 1984 for 4; Burnet 1900 and Joachim 1951 for 5; Gauthier and Jolif 1959, 168–9 for 6. ⁵ This interpretative tradition starts already in antiquity and is still going strong today (Broadie 1991, Destrée 2011), but has little basis in the text; see Bobzien 1998, 2012; Frede 2007, 2011 ch. 2; also Grant 1866, 316–18.

      .–

53

First, some notes on terminology. For the central terms of Nicomachean Ethics 3.1–5, I generally follow standard translations and add alternatives in footnotes— both for the benefit of Greek-less readers. I use the expression ‘virtuous agent’ (‘virtuous individual’) for an agent who has, as character dispositions, the character virtues. Aristotle sometimes uses the term spoudaios for such individuals. I use the expression ‘vicious agent’ for an agent who has some character vices as dispositions. For such individuals, Aristotle sometimes uses phaulos. I use ‘virtuous disposition’ as shorthand for ‘character dispositions that are identical with the character virtues’; and ‘vicious disposition’ as shorthand for ‘character dispositions which are identical with the character vices’. I use the word ‘moral’ exclusively as a generic term which covers the evaluative element, aspect, or dimension of what Aristotle calls, on the side of positive value, noble (kalos), good (agathos), fine (epieikēs), or virtue (aretē); and, on the side of negative value, shameful (aischros), base (phaulos), wicked (mochthēros), bad (kakos), or vice (kakia) (when using these terms in the context of human agency and character respectively).⁶ Accordingly, I also use the expression ‘moral disposition’ as a generic term for virtuous and vicious disposition. My use of ‘moral’ should not be conflated with such modern notions as rule-following, altruistic or utilitarian norms, and religion-based morality, as we find these only in Post-Aristotelian thought.

2. From Praise and Blame to Moral Responsibility Aristotle considers it a fact that we praise and blame people with respect to their virtues and vices (e.g. EN 1106a1–2). In effect, for him virtues are the character dispositions that are praiseworthy (1103a8–10; cf. 1101b13–15, b30–1). He argues that what makes these dispositions virtues is that they are directed towards the intermediate in action (1106b36–1107a1), and that this is also what makes them praiseworthy (1109b24).⁷ At the beginning of Nicomachean Ethics 3.1, Aristotle draws the connection between virtue and vice, praise and blame, and voluntariness, if in a rather unspecific way: praise and blame are bestowed only on actions and emotions⁸ ⁶ I make no attempt at interpreting what it is that Aristotle considers to be of positive value in nobility, goodness, fineness, or virtue and what it is that Aristotle considers to be of negative value in shamefulness, baseness, wickedness, badness, or vice. I simply take it that his consistent use of the respective Greek terms is sufficient indication that he assumed there to be some relevant positive or negative value to these things. ⁷ Aristotle also uses the fact that we praise people who hit the intermediate in order to support his view that virtue is the mean; cf. e.g. 1106b25–8, 1108a14–16. But one must not confuse (i) Aristotle’s use of people’s actual praise and blame as evidence for his theory of the mean with (ii) Aristotle’s own theory of what makes the character dispositions praiseworthy or blameworthy. ⁸ ‘Emotions’ translates pathē. Alternative translations are ‘feelings’, ‘passions’, ‘undergoings’, and with them come different interpretations. Accordingly, some have questioned whether for Aristotle pathē are voluntary. I disregard these difficulties and focus solely on actions.

54

, ,   

that are voluntary (1109b30–1), and virtue is concerned with actions and emotions. In his summary at the end of Nicomachean Ethics 3.5, Aristotle states that he has shown that the virtues are voluntary and that virtues and actions are voluntary in different ways (1114b26–31). This suggests that Aristotle assumes that voluntariness is a necessary condition for justified praise and blame—a fact that is commonly used to support the assumption that Nicomachean Ethics 3.1 presents a discussion of moral responsibility. It is worth pointing out, though, that Aristotle does not have an expression for moral responsibility,⁹ and that in 3.1 he talks mostly about people’s actual practices of praising and blaming, not about praiseworthiness and blameworthiness. Thus in Nicomachean Ethics 3.1 Aristotle may—as he often does—simply start out from empirical considerations: people bestow praise and blame for certain things; they only do this when they believe that these things have certain characteristics; these characteristics are those commonly collected under the heading of voluntariness. Consider societies, like those in ancient Greece, which are simpler than ours in that they do not have the complex intellectual, religious, and ethical traditions instilled in us from childhood. Imagine typical cases of praising and blaming in such societies: a runner is praised for winning a sports contest (‘We are proud of you!’); an adolescent for their beauty (‘Your looks are those of a god(dess)’); a wife is blamed for poor cooking skills (‘You dimwit. Why can’t you cook like our neighbour?’); a slave for having broken a vessel (‘It’s your fault I have to buy a new vessel!’). The expression ‘moral responsibility’ does not come to mind readily to characterize such cases. It is far from clear whether morality, in any sense, is at issue. Praising and blaming are human activities which are reactive to certain kinds of—mostly—human behaviour and habits that please or annoy. The purpose of praising and blaming may simply be that (i) of expressing one’s appreciative or disapproving sentiments or (ii) of encouraging or discouraging repetition of the relevant behaviour. There are certain conditions that have to be met for such activities to be appropriate—that is, conditions not for whether the appropriate reaction is praise or blame, but for whether any such reaction is appropriate. (For example, it is not appropriate to blame anyone for the weather, or to praise a stone for its weight.) No notions like noble (kalos), shameful (aischros), fine (epieikēs), or base (phaulos) need to be involved in this context. We can first sort things according to whether they are the kind of thing that can be praised or blamed, and then correlate these things with acts of praising or blaming. To get from the situations of blaming and praising considered by Aristotle in Nicomachean Ethics 3.1 to moral responsibility (i.e. responsibility related in some ⁹ I use ‘moral responsibility’ as a generic term that covers the praiseworthiness and blameworthiness of someone for something with regard to the nobility, fineness, or goodness, on the one hand, or the shamefulness, wickedness, baseness, or badness, on the other, of that someone or that something.

      .–

55

sense yet-to-be-determined to what is noble or shameful) the things that are praised or blamed must in addition be classified as noble or shameful, and the praise or blame must attach to them because of their being noble or shameful.¹⁰ In the remainder of this chapter, I show that in Nicomachean Ethics 3.2–5 Aristotle provides a theory of the conditions of moral appraisal that does precisely this. Aristotle’s move from generally accepted conditions of praise and blame to moral responsibility is in parallel with his move from common-sense-based and common-opinion-based discussion of the voluntary as a requirement for praise or blame in Nicomachean Ethics 3.1 to his explanation of the underlying psychological apparatus in 3.2–5, an explanation that is based on his own view of the soul (as set out in 1.13 and 6.1). Aristotle introduces choice as a psychological feature that characterizes the appropriate subjects and objects of praise and blame and that provides the channel by means of which both rationality and the moral dispositions of character can manifest themselves in adult human action. Choice turns out to be characteristic for any kind of moral responsibility. I begin with the manifestation of moral responsibility in action (sections 3–11). I consider moral responsibility for character in section 13.

3. Action and Voluntariness According to Aristotle, action (praxis) is a specifically human activity.¹¹ An action is a change (kinēsis) which has its origin (archē) in the human being, who is also the efficient cause of the actions (EE 1222b28–31).¹² All action is goal-directed, that is, it aims at an end (telos), though its end may lie in itself (EN 1.1). Action requires the agent to have a reasoning capacity (EE 1224a29–30). This is why people do not say of toddlers or animals that they perform actions (EE 1224a28–30). Although Aristotle defines action as change, when he discusses voluntariness he is concerned equally with situations in which individuals perform an action and in

¹⁰ Aristotle is aware of the difference between praising someone (or something) (i) for their (or its) nobility or character-virtue related goodness and (ii) for some other trait that is not related to character-virtue: ‘(i) we praise the just or brave person, and generally both the good person (ton agathon) and virtue itself, because of the actions and pursuits involved, and (ii) we praise the strong person, the runner, and so on, because they are of a certain kind and are related somehow to something good and excellent’ (EN 1.12.1101b14–18). ¹¹ EE 1222b18–20, EN 1139a19–20. The classic work on Aristotle’s theory of action is Charles 1984. ¹² Aristotle says or implies that actions are changes (kinēseis) both in the EE (e.g. 1222b29) and in the EN (e.g. 1139a31–2). This is in line with his assumption that actions have an efficient cause (ibid.). Here ‘efficient cause’ is the standard translation for the kind of cause which Aristotle refers to as ‘that from which’ (ibid.) or as ‘the origin of change’ (Phys. 195a11), and that he defines as ‘that from which the alteration or absence of alteration has its origin’ (Phys. 194b29–30). In the EN Aristotle refers to actions both as changes (kinēseis) and as activities (energeiai) (e.g. EN 1.1; 3.5). I assume that, rather than contradicting himself, Aristotle does not use energeia in its narrow, technical sense when calling actions activities. This assumption is supported by the fact that Aristotle’s practical sciences tend not to make direct appeal to his theoretical positions.

56

, ,   

which they refrain from performing an action. (For something to be a refraining from an action, broadly, the action must be something the agent could have considered doing.) Praise and blame are bestowed on both action and refraining from action. Accordingly, Aristotle uses ‘action’ (praxis) sometimes in the wider sense that covers both action and refraining. This is in line with Aristotle’s theory of efficient causation, which allows for the case in which someone is an efficient cause of an absence of change, as opposed to an efficient cause of change (Phys. 194b29–30, 195a11–14). I shall also use ‘action’ in places to cover both actions and refrainings. Aristotle holds that an action is voluntary (hekousios)¹³ precisely if both the origin of the action is in the agent and the agent knows the relevant particular circumstances of the action. The action’s origin is in the agent when the agent is the action’s efficient cause and the action is not the result of external force. Aristotle is aware that there are cases in which it is doubtful whether the agent was externally forced, and that the boundary between the unforced and forced cases is difficult to draw. I will not dwell on the details.¹⁴ What matters for us are the following—undisputed—points. (i) For an action to be unforced, the agent has to contribute something (EN 1110b2–3), that is, the agent has to be at least a causal co-contributor. (ii) The problematic cases are those in which the force seems to take the form of compulsion.¹⁵ By and large, Aristotle maintains that compulsion, as long as it does not put the agent under duress that overstrains human nature (one may envision severe torture), does not render the act forced (EN 1110a24–6, EE 1225a21). (iii) Aristotle explicitly rejects the suggestion that an agent’s action is forced when it is the result of a desire (for pleasure or avoiding pain) in the agent that conflicts with the agent’s end. (Imagine someone following their desire for doughnuts despite having adopted healthy eating as an end.) The possibility of a conflict between non-rational desires (e.g. appetite and anger) and rational ones is characteristic of humans. Aristotle argues that the agents’ appetites and emotions are no less human than their reason, and originate no less from the agents than do their reasoned ends.¹⁶ The second criterion for voluntariness, absence of ignorance, covers only ignorance regarding specific circumstances of the action. Aristotle provides ¹³ Alternative translations of hekousios: ‘willing’, ‘intentional’, ‘witting’. For discussion of voluntariness in the EE, see EE 2.7–8. ¹⁴ For an excellent discussion, see Meyer 2006, 141–9. The classical treatment is Kenny 1979. ¹⁵ By compulsion I mean a force that does not cause or prevent movement of the individual’s body directly (like strong winds or shackles) but that instead makes the agent move (or refrain from moving) as a result of action-promoting (or action-refraining) activity in the individual’s soul. Thus, if an agent utters ‘he is in the basement’ in order to avoid further beatings, the compulsion works via the agent’s soul and is thus psychological. The force would be solely physical if it were to work directly on the agent’s vocal cords, making them produce the phonetic sounds ‘he is in the basement’. ¹⁶ EN 1111a24–b3, cf. EE 1224b7–14, EE 1224a23–5. Aristotle also rejects the possibility that pleasant things (the doughnut over there) can externally force an agent to act. Rather, insofar as they are envisaged as good, such external things function as the final cause or goal of the action (EN 1110b9–15).

      .–

57

many examples. Thus one may be unaware that one’s fencing sword is missing its button, and subsequently involuntarily injure a friend. Again, there may be doubtful cases: it may be unclear whether in a particular case the ignorance at issue qualifies for removing voluntariness (see section 12). Aristotle himself does not discuss such doubtful cases. But he explicitly excludes ignorance of general truths, including what the right end is (EN 1110b28–33) and ignorance that is itself the result of a voluntary action (see section 12).¹⁷ With each criterion (agent-internal origin, absence of ignorance), Aristotle seems to capture common-sense opinion. From a contemporary perspective, it is remarkable how closely Aristotle’s criteria tally with those of present-day criminal law. Even with today’s psychological and technological methods, the problems a judge or jury faces as to whether an agent can be held responsible are pretty much the same.

4. What Happens in the Agent’s Soul The common-sense account of voluntary action in Nicomachean Ethics 3.1 explains only why actions like throwing cargo overboard or eating doughnuts are praised or blamed. A—positive—connection between the agent’s praiseworthy and blameworthy dispositions (i.e. virtues and vices) and voluntary action is not drawn. Nor do we obtain any information about what happens in the agent’s soul in cases of voluntary action. Yet Aristotle thinks virtues and vices are dispositions of the human soul. Thus, to draw the relevant connection Aristotle needs to provide the psychological details of human agency. He does this in Nicomachean Ethics 3.2–5, where it becomes clear that, with the exception of spur-of-the-moment acts (for which see section 9), human agency essentially involves choice, and that via choice the connection between voluntary action and virtues and vices is drawn. Aristotle defines choice (prohairesis) as an agent’s deliberated desire for things that are in the agent’s power.¹⁸ Every choice has two aspects: it is for the sake of something and of something (EE 1227b36–7). In the EN, Aristotle expresses these two aspects by saying that choice is with regard to an end and concerns the means to that end (ta pros to telos, EN 1111b27, 1113b3–4). Choice occurs in the agent’s soul, more specifically in the ruling part in the agent’s soul (EN 3.3 1113a6), and ¹⁷ Aristotle never says or implies that responsibility is a matter of degree. Rather, either one is morally responsible for something (i.e. one deserves praise or blame) or one is not. Of course, if one is responsible, there is the independent question of what kind or amount of praise or blame, respectively, would be appropriate, and this will depend on the circumstances of the action and agent. Aristotle shows awareness of this second point when he mentions that the punishments for actions done while drunk are double (EN 1113a30–4). ¹⁸ Alternative translations: ‘decision’, ‘intention’, ‘purpose’, ‘resolution’. The reason for these alternatives will become clear in what follows.

58

, ,   

for each of the two aspects a different subpart of this part of the soul is responsible, one rational-in-itself, one desiring and capable-of-listening-to-and-obeying reason.¹⁹ The desiring part of the human soul is one of its two nonrational parts.²⁰ It comprises all human desires (orexeis), including appetite, anger, and wanting.²¹ A person’s character (ēthē) is manifested primarily in this desiring part of the soul. What type of character a person has is determined by what character dispositions (ēthikai hexeis) it comprises. For example, if the character is excellent, these character dispositions will be the character virtues (courage, generosity, etc.): the character virtues are the excellences of this part of the soul (EN 2 and 3.6–5.11). The agent’s character dispositions determine what ends the agent has—that is, what the agent’s goal-directed desires are directed towards. The specific desire a person has for their ends is called wanting (boulēsis).²² It is this wanting that is causally responsible for the ‘for-the-sake-of-which’, or the end, of the choice. Without character dispositions that determine what a person wants, there can thus be no choice (1139a33–4). (In short: the desiring part, via the agent’s character dispositions, including virtues or vices, determines the agent’s ends and provides the agent with a certain kind of desire, a wanting, for that end.) The relevant rational part of the soul is the one that deals with contingent or changing (as opposed to necessary, unchanging) things. This part is called calculative. Aristotle states that calculation and deliberation are the same (EN 1139a11–13). This explains why this part is responsible for deliberation. The excellence of this part of the soul is practical wisdom.²³ Via deliberation, the calculative part determines the means by which the wanted end can be realized. Thus, the desiring part of the soul provides that for-the-sake-of-which a choice is; the calculative part provides that of which that choice is. Together, the two parts make up the efficient cause and origin of the choice (EN 1139a31–4). Every choice necessarily needs both. Without a wanted end, there is nothing that can function as a means, and hence no action to choose. Without the means, there is no way to reach the end. I consider each aspect separately.

¹⁹ The main passages about the parts (or aspects) of the human soul are EN 1.13 and 6.2 and DA 3.9; cf. also DA 2.2. ²⁰ The other is the vegetative part, responsible for digestion, sleep, breathing, growth; see EN 2.13; DA 2.2. ²¹ The desiring element, though non-rational, shares in reason insofar as it ‘can listen to it and obey it’ and thus differs from the desiring part in non-rational animals. The human desiring part can wholly or partly be in conflict with the rational part or ‘speak on all matters with the same voice’ as it (EN 1102b13–31). Section 11 provides details. ²² Wanting is a desire that requires the listening and obeying ability of the desiring part of the soul to have been used. It presupposes, and depends on, the rational part of the agent’s soul. (For a different interpretation of how wanting is related to the soul, see Lorenz 2009.) ²³ Phronēsis; cf. EN 1103a3–6 and Natali 2014 for details.

      .–

59

5. Choice of, Deliberation, and the Things within the Agent’s Reach A choice is of the action-to-be-performed, where an action can also be a refraining (section 3). The calculative part of the soul is needed to figure out what action this is going to be. It does so by means of a certain activity: a process of deliberation (bouleusis).²⁴ For Aristotle, deliberation is practical enquiry (EN 1112b22–3) regarding the possible means that allow agents to reach their ends. In terms of Aristotelian syllogistic, deliberation is the search for the middle term (EN 1142b24) or terms of a practical (i.e. action-related) syllogism. For Aristotle, practical reasoning seems to be a form of syllogistic reasoning that involves at least one universal term which expresses something the agent has adopted as an end (e.g. healthy eating), and whose conclusion states the action-to-be-done as a case of the universal term.²⁵ My main interest regards the things which deliberation is about, that is, the objects of deliberation (bouleuta). Aristotle states that the objects of deliberation must belong to the class of things that are in the agent’s power.²⁶ There are two features of the things in the agent’s power that demarcate them as the proper objects of deliberation and choice: (i) that they are things that are within the agent’s reach; and (ii) that they provide alternative options. Deliberation is aimed at action that leads towards one of the agent’s ends. Hence deliberation is about things that are within our reach and that allow us to obtain our ends. Deliberating about any other things would be pointless. Aristotle argues this point repeatedly (EN 3.3.1112a21–1112b16, 6.2.1139b5–11; EE 2.10.1226a21–32), listing all the things that are not within our reach, in order to arrive by elimination at those that are. He rules out • Eternal (i.e. unchanging) things, like mathematical truths or astronomical facts. • Changing things that always happen the same way, like the rising of the sun. • Changing things that follow unpredictable patterns, like the weather. • Chance events, like finding a treasure. • Human affairs that are out of our reach, like the political affairs of remote states. • Particular facts, like whether this is bread. ²⁴ Also boulē, alternative translation: ‘reasoning’. Aristotle discusses deliberation also in EN 6.9 and in EE 2.10. ²⁵ For more information on deliberation, see Segvic 2009, §§1–3, and Natali 2014. ²⁶ epi with dative: Aristotle uses this mostly, but not exclusively, in the form ta eph’hēmin, ‘the things in our power’ and eph’hēmin plus infinitive of verb of action, ‘in our power to do’ something. Alternative translations are ‘up to us’, ‘depending on us’, ‘lie within us’. Aristotle takes this phrase from ordinary language. Only in Post-Aristotelian philosophy do we find to eph’hēmin used as a philosophical term; see Bobzien 1998.

60

, ,    • The past, like to have sacked Troy. • Our own ends, since we have them or set them, but do not deliberate about them. • Things that happen through us, but always in the same way.²⁷

By way of this elimination argument, Aristotle determines the possible objects of deliberation, that is, the things in our power. His narrowing of the scope is entirely common sense. As Aristotle himself says, he is concerned with what reasonable people would deliberate about; people who have a mental disability or are temporarily unstable may deliberate about things not in their control (EN 3.3.1112a19–21). Aristotle is correct that reasonable people do not deliberate about the types of things excluded. They deliberate only about things that they consider it to be, with a reasonable certainty,²⁸ within their control to achieve.²⁹ Positively, Aristotle determines the possible objects of deliberation as those things that can be brought about through us by action.³⁰ As sufficient condition for this, he mentions that the origin of the action is in us (1112b27–8). As we know, this entails that we are the efficient causes of those actions. Presumably preempting a possible objection, Aristotle notes that things we ask our friends to do for us count among the things brought about through us by action. Thus, the possible objects of deliberation turn out to be what we consider possible actions of ours—in a suitably loose sense.

6. The Necessity of Alternative Options Voluntariness is a necessary condition for praise and blame for actions. As such, it is a property of actions,³¹ that is, of things that actually occur. But restriction to things that actually occur, and their properties, is insufficient to capture a vital

²⁷ Aristotle does not explain these last ones. He may have in mind things like sleeping, breathing, blinking. We do not deliberate about whether we should sleep, breathe, or blink, although we may deliberate about whether we should sleep, breathe, or blink now. Alternatively, he intends unreflective habits, e.g. we may always put our left sock on before our right. There is nothing to deliberate about this (in ordinary circumstances); we may never have deliberated and chosen to do it this way round. We just do it. ²⁸ There is no guarantee ever: external circumstances can prevent the action. Aristotle indicates that the agent is aware of this when he states that agents deliberate about things that happen for the most part (hōs epi to polu, 1112b8–9). They do not need certainty that they will bring about the action, just that, regarding comparable situations, for the most part this is so. ²⁹ Even reasonable people may err, of course, due to ignorance of circumstances: you may think you can take the train, yet, unbeknownst to you, it just derailed. Whether such deliberation would be apparent only, as opposed to actual, Aristotle does not say. ³⁰ EN 1112a31–2, 1111a33–4. It seems that Aristotle takes it for granted that all agree that the things he needs to demarcate are the things in our power. His question is what classes of things satisfy this condition? ³¹ Or a relation between an action (or refraining) and the agent.

      .–

61

aspect of human agency—namely, that the agent must have alternative options. These come in at the stage of deliberation: agents will deliberate about a possible course of action only if they have more than one possible course of action. It does not suffice that the agent is the origin and efficient cause of some change. (For the agent is also, say, the origin and efficient cause of her digestion and her reflex actions, but does not deliberate about these.) Deliberation and choice both require that the agent has alternative options. It is an essential feature of the things we deliberate about, and that are in our power, that they come in pairs.³² Following ordinary Greek language use, Aristotle uses the phrase ‘in our power’ also to express this necessary feature of human agency. I articulate this by saying that he uses the expression for a two-sided notion of that which is in our power or up to us. That is, he assumes the following principle: if it is in the agent’s power to do something, then it is also in the agent’s power not to do it, and vice versa (see EN 1110a17–18, 1113b7–8 and section 9). Aristotle also sometimes uses ‘being master of ’ (kurios + genitive) and expressions of possibility (exēn) in this two-sided way to express the same requirement for human agency (e.g. EN 3.5.1113b32–3, 1114a2–3, 16–17, 19–20). For agents to act, it is not sufficient that they have alternative options. They also need to be aware that they have them. If agents do not believe that both doing and not doing something are in their power, they will not deliberate about whether to do it.³³ Possibly, when Aristotle says that it is unclear how the things we deliberate about will turn out (EN 1112b9), he wishes to draw attention to this fact.³⁴ In any event, in De Interpretatione 9, where Aristotle draws out the consequences of the hypothesis that everything happens by unconditional necessity,³⁵ he states ‘there would be no need to deliberate or busy oneself with anything, thinking that if we do this, this will happen, but if we do not, it will not’ (DI 9.18b31–2, tr. Ackrill, modified). This shows that Aristotle assumes (i) that agents generally believe that they have alternative options and consider them part of their deliberation, and, moreover, (ii) that they (or at least the reflective agents) believe their deliberation is a determining factor for which option is going to be realized. (ii) is confirmed when Aristotle, in the same context, appeals to the facts, writing ‘we see that what will be has an origin both in deliberation and in action’ (DI 9.19a7–9, tr. Ackrill). Thus, Aristotle takes it for granted not just that agents have alternative options but also that they are aware both that they have these options ³² The pairs are related to each other as F and not-F are. Of course, there may be many other options, G, H, I, etc. (with which there then also come not-G, not-H, and not-I, etc.). ³³ Of course, in the course of a deliberation, there may be, e.g., dead ends or discoveries that presumed options are actually not options. Aristotle is aware of this; cf. EN 1112b24–6. ³⁴ This is one of several possible interpretations of this sentence. Alternatively, Aristotle could refer to the fact that the type of actions deliberated about do not always reach their end (see Joachim 1951, 101; Taylor 2006, 150). But this would make either this or the ‘for the most part’ clause (see n. 28 above) redundant. ³⁵ Unconditional necessity: necessarily p, no matter what else is the case.

62

, ,   

(‘if I do this, this will happen; if not, it won’t’) and that their deliberation process and its result are necessary causal factors that (co-)determine what action they will perform and whether they reach the goal at which the action was directed. The required alternative options and the agents’ awareness of these entail neither (i) that the agents are causally undetermined in their action nor (ii) that the agents believe they are causally undetermined in their action. That is, Aristotle’s Nicomachean Ethics 3.1–5 and De Interpretatione 9 are compatible with (i) that, the agent and the circumstances being what they are, together they are sufficient causal factors to bring about the action, and (ii) that the agents do not hold beliefs that they are causally undetermined in their action. All that the text suggests is that (reflective) agents are aware that without their deliberation the action would not take place—that their deliberation is a necessary causal factor in the process that brings about the action (18b31–2, 19a7–9); and that Aristotle holds that through their deliberation (and choice) the agents themselves become a decisive causal factor of the action and its direct consequences (DI 18b31–2 and 19a7–9 together with EN 3.2, 5, e.g. 1113b30–3).

7. How Deliberation and Choice are Related Aristotle holds that the objects of choice (prohaireta) are taken from the same pool of things as the objects of deliberation: the things within our reach. Every object of choice was at some point an object of deliberation. However, being an object of choice is not the same as being an object of deliberation. An object of deliberation is something the agent deliberates about. An object of choice is something the agent has come to be choosing. There is a significant difference here. The object of deliberation has an element of indefiniteness to it. By contrast, the object of choice (i.e. the chosen course of action) is determinate (1112b9, 1113a3). An object of deliberation becomes an object of choice via the agent’s judging or deciding (krinein, 1113a4, 1113a12). At the moment when the agent judges to pursue one of the courses of action about which they deliberated, the agent starts to have a deliberated desire—that is, a desire in accordance with the judgement (krisis) resulting from the deliberation (1113a11–12)—for pursuing the respective course of action. This desire is the agent’s choice.³⁶ Hence, where there was an indeterminacy between at least two possible courses of action, there is only one course once the choice has come into being. The object of choice is thus determined and rationally desired, whereas the object of deliberation was not (yet). Being a desire, choice essentially has a

³⁶ In fact, Aristotle wavers about how to categorize choice. Since both elements of reason (via reasoning process) and of desire (via character disposition; see also section 8) are preserved in choice, Aristotle alternatively calls it reasoned desire and desiring reason (EN 1139b4–5, EE 1227a3–5, 1226b17).

      .–

63

duration, which in the standard case continues from the moment of judgement to the completion of the action.³⁷ The English translation ‘choice’ does not (clearly) capture this point, but it is important for a full understanding of Aristotle’s theory of agency. What is it that determines which course of action (of those under deliberation) becomes rationally desired or the object of choice? The answer is: the agents’ deliberation and ends. The agents assemble the premises and go through a— potentially rather complex—course of reasoning, and it is their drawing the conclusion that is the judgement that brings into being their choice. What conclusion an agent arrives at likely depends on a variety of factors. These include what ends the agents started with as general premises (section 8), how welldeveloped their reasoning ability is, how well their memory and perception function,³⁸ and probably the various external circumstances present during the deliberation process, such as possible distractions, time pressure, and which people the deliberator may consult for help (EN 1112b10–11, b27–8). Hence Aristotle’s choice (prohairesis), as he uses it in Nicomachean Ethics 3.1–5, is nothing like an act of deciding or an act of choice between alternatives.³⁹ Nor is it (or is it issued from) a faculty for causally undetermined choice or decision, or of free will, as is sometimes assumed (e.g. Grant 1874, vol. 2, 14). The judgement (krisis) that co-causes this desire is also not a faculty for undetermined decisionmaking, nor is there any decision-making faculty such as a will in the agent that determines which way the judgement will go.⁴⁰

8. Choice for the Sake of, Wanting, and the Transfer of the Moral Aspect to Choice Aristotle’s discussion of choice of in Nicomachean Ethics 3.2–3 elucidates the relation between the things in our power, deliberation, and choice, but links the voluntariness of actions neither with moral responsibility nor with virtue and vice. For this connection to become apparent, we need to examine for-the-sake-of-what

³⁷ This becomes more intelligible if we consider that for Aristotle the choice (prohairesis) of a particular action is the actuality (energeia) of a dispositional state of the soul and, as such, is complete at any moment and can be continued (until the action is completed). ³⁸ Cf. the mention of intellect (nous) in EN 1139a33 and Aristotle’s description of its role in deliberation in EN 6.11. ³⁹ When Aristotle says, at EN 1112a16–17, that the name prohairesis suggests that to prohaireton is something (to be) chosen before, or in preference to, other things, this does not imply that prohairesis is an act of choice between alternatives. Rather, one needs to read it in its context, where Aristotle emphasizes that reason and deliberation are necessary conditions for prohairesis. It is since prohairesis requires reason that prohairesis is of what is preferable to other things. ⁴⁰ In the virtuous agent, the dispositional state of the soul of which a choice is the actuality is the respective virtue (see sections 8 and 10). This is why Aristotle sometimes says that virtue is prohairesis (EN 1106a2–4), where he intends the dispositional state from which the desire issues.

64

, ,   

a choice is. Aristotle tackles this issue in 3.4 and draws the connection explicitly in 3.5.⁴¹ We saw that the part of the soul causally responsible for the for-the-sake-ofwhat of choice is the desiring part—that is, the part where a person’s desires, including their wants, as well as their character dispositions (e.g. their virtues and vices), are manifested (section 4). Without character dispositions that determine a person’s ends, there can be no choice (EN 1139a33–4, b3–4). Virtuous agents have as their end that which is truly good, since they have the correct view of what is good. In Nicomachean Ethics 2, Aristotle established this end to be the intermediate in action. In the Eudemian Ethics, Aristotle explicitly considers the function of the desiring part of the soul for the case of virtuous and vicious agents (EE 1227b34–1228a5): since virtuous individuals have (hitting) the intermediate as their end, their choice is for the sake of (hitting) the intermediate. Thus a character virtue is the cause of the intermediate qua being intermediate (EE 1227b36–8). It causes the end of the choice to be correct. That is, the virtuous agent, when judging, brings about a choice with a view to (hitting) the intermediate qua intermediate. (An agent may choose to tip a certain amount because it is the intermediate between being stingy and being wasteful, or because this happens to be the amount of cash in their pocket. In the latter case, tipping that amount is not chosen qua intermediate.)⁴² In Nicomachean Ethics 3.4, Aristotle explains by means of which element of the desiring part of the soul this causation works: it is via wanting.⁴³ Wanting is of the end (EN 1111b26–8, 1113a15, 1113b3), which in the case of action is (truly or apparently) good action and its (truly or apparently) good consequences. Virtuous agents both want what is truly good (1113a31–3, eupraxia, 1139a34) and have all their other desires aligned with their wanting. It is by this internally unchallenged want for the intermediate that a virtuous character disposition causes the intermediate. Vicious agents may not want what is truly good, but only what—since pleasant—appears good to them. Their vices make them want incorrect ends, and thus causes their not realizing the intermediate. Generally, the agents’ character dispositions (in the case of the virtuous and the vicious, their virtues and vices) cause the agents to have a certain type of goaldirected desire: wanting. This wanting provides the deliberative part of the soul with the starting point for their deliberation by providing a universal term (e.g. healthy eating) for the major premise, to which term is tied the agent’s desire for its realization. ⁴¹ Commentators on EN 3.1–5 are often somewhat unclear about what is the purpose of EN 3.4 and its location between EN 3.3 and EN 3.5. A reading of the passage like the one presented here that explains the purpose of EN 3.4 in its context thus has the edge over those. ⁴² For this distinction, see also Aristotle EN 5.8. ⁴³ Translators of the EN have also chosen ‘wish’ or ‘will’ (Rolfes 1921: ‘Wille’) or ‘willing’ instead of ‘wanting’ for boulēsis.

      .–

65

Since the virtuous and vicious agents’ character disposition has an aspect to it that is essentially moral (the disposition is virtue, or is vice), this moral aspect is, via their wanting, transferred to their choice. For example, if the agent’s end was not the intermediate, their choice would—mostly—not be for actions that are intermediate.⁴⁴

9. From Choice to Action Choice is caused by a cooperation of two causal factors, issued from two different parts of the agent’s soul. They both together make up the origin and efficient cause of choice (EN 1139a31–4). It is in choice that a person’s reason and their desire (which is expressive of their character disposition) are joined and preserved. This is a precondition for human agency—as opposed to animal behaviour. Choice, in turn, is the efficient cause and origin of action (EN 1139a31–2; see also Segvic 2009, §4). We saw that choice is a type of desire and, as such, has a duration. But not every choice ends in the action chosen. Choice continues from its beginning until either (i) the action has been completed, or (ii) something external interferes with the realization of the action, or (iii) the agent reconsiders their options and re-deliberates, with a different choice as result, or (iv) (if there is a time limit on its possible realization) the agent fails internally to realize the action—a case taken up in section 11. There are two further reasons why there is no one-to-one correlation between choice and adult human action. On the one hand, Aristotle acknowledges spurof-the-moment acts that are voluntary (EN 1111b9–10; EE 1224a2–4, 1226b3–4). So we may incur praise or blame for them. (We may shoo away a fly, thereby brushing a glass off the table, or throw ourselves in front of a car to save a child.) But, Aristotle holds, these spur-of-the-moment acts do not involve choice. No deliberation precedes them. They may be reflex actions or caused directly by one’s dispositions (EN 1117a20–2). On the other hand, one choice can be sufficient for repeated action. For example, you may resolve to floss your teeth daily, and then, caused by this reasoned desire for daily flossing, floss every day. It may be that Aristotle thought—correctly—that a large part of our actions are of this type.

10. Transference of the Moral Aspect from Choice to Action Via choice, the moral aspect—that is, the element of good or bad—is transferred from the agent’s character disposition to the agent’s action. It is this transference ⁴⁴ EN 1113a18; 1113b1–2; 1139a33–6; a39; EE 1128a4–5; cf. EN 1144a20: the virtue makes the choice right.

66

, ,   

of the moral aspect to action, and consequently the voluntariness of this aspect in the action, that Aristotle explains in the first part of Nicomachean Ethics 3.5, taking up what he laid out in 3.4:⁴⁵ (1) The end being the object of wanting (2) and the means to the end being the objects of deliberation and choice, (3) the actions with regard to these [means to the end ]* would be in accordance with choice and [that is] voluntary. (4) And the activities of the virtues are [activities] with regard to these [means to the end]. (EN 1113b3–6; emphasis added)

This is the first time since the beginning of Nicomachean Ethics 3.1 that Aristotle mentions virtue, action, and voluntariness together (EN 1109b30–1; cf. section 2). Naturally, here, we should expect him to provide information about the connection between them which he hinted at there. If read carefully, the passage does just this: (1) and (2) introduce the two causal strands that together cause choice. (1) Wanting, the goal-directed desire whose end is determined by the agent’s character dispositions, is responsible for what the end of an action is. (2) Deliberation, the process of reasoning determined by the calculative part of the agent’s soul, is responsible for what kind of action the action is. (3) Since these two together cause choice, and the choice causes the action (see section 9), the action is expected to be in accordance with choice and hence voluntary.⁴⁶ It is in (4) that Aristotle expressly draws the connection to virtue. I assume that by ‘activities of the virtues’ Aristotle denotes the actions of the virtuous agent.⁴⁷ Thus the character virtues are active via actions in which the intermediate is realized. In fact, the only way in which character virtues can be realized in action is via realizing the intermediate in action. As (1)–(3) make clear, not just any action of the agent will do. For the moral aspect of virtue (or vice) to be transferred to action, it has to be the result of choice. Thus we can make explicit an implicit conclusion of Aristotle’s argument: the activities of virtues are actions that are in accordance with choice. The only way in which virtues can be realized in action is via choice and realization of the intermediate in action. For example, the only way the end of healthy living can be reached is (via wanting of that end) by choices and actions that realize that end, e.g. choosing and eating a balanced meal rather than doughnuts galore.⁴⁸ ⁴⁵ Remember that the division into chapters is not Aristotle’s, so the beginning of EN 3.5 would have followed without pause upon the end of EN 3.4. * Square brackets in a translation indicate a phrase supplied by context. ⁴⁶ Anything that occurs in accordance with choice is voluntary; cf., e.g., EN 1112a14–15, EE 1226b34–5. ⁴⁷ Cf. EN 1094a4, a6, a16: actions are a kind of activity. ⁴⁸ This fact, that the moral aspect can only be expressed (or fail to be expressed) if there is choice, explains why toddlers and dogs, though they may be de facto praised and blamed, are not held morally responsible for what they do.

      .–

67

And since all action in accordance with choice is voluntary, we can make explicit a second implicit conclusion in Aristotle’s argument: the activities of the virtues are voluntary. This squares with Aristotle’s account of the voluntary. As the virtues (or vices, respectively) are one of the two causal factors of choice, and there is no external force, nor ignorance of the relevant kind, involved in their cocausing choice, what they contribute to the action (i.e. its moral aspect) should be expected to be voluntary. In the following lines, Aristotle provides an argument that has the function of explaining how it is that virtue and vice in action themselves are in our power (and thus voluntary). This argument is based on the point established in (1)–(4) that the activities of the virtues (and vices) concern the means to the end.⁴⁹ Here is, first, the passage: (5) And virtue, too, is in our power, and equally vice. (6) For where it is in our power to act, it is [also in our power] to not act, and where [it is in our power] to not [act], also to [act]. (7) Hence, if acting, being noble, is in our power, not acting, being shameful, will also be in our power, and if not acting, being noble, is in our power, acting, being shameful, [will] also [be] in our power. (8) But if doing noble things and doing shameful things are in our power, and equally, too, not doing [noble things and not doing shameful things], (9) and this was [as we said earlier] being good and being bad, (10) then being fine and being base will be in our power. (EN 1113b6–14)

Aristotle’s use of the singular terms ‘virtue’ and ‘vice’ suggests that he is talking about virtue and vice as characteristics manifested in an action, and about the agent qua agent of the action, rather than about the character dispositions.⁵⁰ (In (4) he makes the move from (i) virtue as character disposition to (ii) the manifestation in action of the character disposition that is virtue. In (5)–(10) he is talking about (ii).) I believe that Aristotle assumes an implicit conclusion (C), ‘the activities of the virtues are activities in accordance with choice and voluntary’; and parallel to (4), ‘the activities of the vices are [activities] with regard to these [means to the end]’, and to (C) ‘the activities of the vices are activities in accordance with choice and voluntary’. In other words, I believe that Aristotle takes it for granted that the ⁴⁹ Aristotle never states the purpose of these lines; moreover, they contain some textual issues. I here provide what I consider the most plausible reading and mention and assess alternative suggestions in footnotes. ⁵⁰ I agree with Meyer 2006, 129–31 that Aristotle appears to use the expressions for virtue and vice, and for our being good and bad, noble and base, in two quite different ways. When he uses the expressions in the plural, he talks about the moral dispositions, and our having virtue or vice, in this context, means that we have those character dispositions. When he uses the expressions in the singular, he often simply means the goodness and badness manifested in an agent’s action: ‘you were bad, you lied to me’; ‘you were good, you didn’t have dessert’. Here, when agents are called good or bad, etc., this is a ‘punctual’ and ‘reflective’ (from the action back to the agent) use: the agents are called good or bad insofar as their action was good or bad.

68

, ,   

same psychological structure-and-procedure by which virtues become active is in play when vices become active. The purpose of 1113b7–14 is then to set out how the moral aspects of the virtues and vices (that which makes them virtues or vices) are manifested in action in such a way that praise and blame can be attached to the agents for the resulting noble or base action (qua noble or base action). The argument starts with a statement of the thesis to be proved, that is, (5). This is standard procedure from Aristotle’s dialectic. Next, Aristotle introduces a twopart premise, (6), stating the two-sidedness of things being in our power. Its truth seems to be taken for granted. The points that human actions are goal-directed, and their end, which is an aspect of them, is determined by the agent’s character dispositions (including their virtues or vices), are taken from (1)–(4) as additional premises. Thus, a full description of the actions at issue would not be ‘(not) doing x’ but ‘(not) doing x, x being noble’ or ‘(not) doing x, x being shameful’ respectively. In (7), Aristotle combines (6) with (1)–(4), and draws an intermediate conclusion (‘hence’): since between good and bad actions the relations hold that if by doing x one is hitting the intermediate, then by not doing x one is not, and if by not doing x one is hitting the intermediate, then by doing x one is not, we get the expanded statement of the two-sidedness of things being in our power. From (7), there follows the antecedent of (8)–(10), that is (8), which rearranges the possible cases, putting the cases of acting before those of refraining. In (9) Aristotle restates as an additional premise an account of what it is to be good or bad in action that he used before. Finally, from (8) and (9) he concludes (10), assuming that if x is in our power and x = y, then y is in our power. The conclusion (10) is an alternative formulation of the thesis (5): being fine and being base are in our power. Aristotle’s argument (1113b3–14) is then in short: all action is goal-directed. The end is determined by the agent’s character dispositions and wanted by the agent because of this character. An action with which the agent aims at the intermediate is a good-and-noble action; an action with which the agent aims at an extreme is shameful-and-bad. The goodness or badness are transferred to the action from the character disposition virtue and vice, by means of a choice. (Character virtue is good; character vice is bad.) Actions that are the result of choice are in our power, and so are their opposites. But the actions at issue that result from a choice have a moral aspect, which is manifested in them because the moral disposition is a co-cause of the choice.⁵¹ Thus it is correct to say of these actions that they are in our power, being good, and that they are in our power, being bad, respectively, and hence that the good actions and the bad actions are in our power, respectively. And in the sense in which we are fine-and-good or base-and-bad, insofar as we performed a good or bad action, it is also correct to say that our being good, or our ⁵¹ Aristotle only considers the character dispositions of vice and virtue here, not any in-betweens, for which see section 11.

      .–

69

being bad, is in our power. Aristotle is here not talking about it being in our power that we are fine or base in the sense of having virtuous or vicious dispositions. Evidently, the argument proves no such thing.⁵² Rather, the goodness or badness of the agent that is in the agent’s power is derived from the good or bad actions. And this is all that is required to establish the moral responsibility of the agent for their actions. As their being good/bad (qua doing something good/bad) is in their power and thus voluntary, they can be praised/blamed as being fine/base (qua having done something good/bad).⁵³ At this point, Aristotle has completed showing how virtue concerns actions (EN 1109b30) in such a way that its manifestation in action is voluntary and in our power, so that praise or blame can be attached to the agent for having performed a noble or shameful action: we can say ‘you are noble insofar as you did this’, or ‘your doing this was bad’. The blame or praise can be attached to the agent, since it was something in the agent, namely the agent’s character disposition, which is causally responsible for the moral dimension of the action.⁵⁴

11. Choice and Aristotle’s Four Types of Agents It is instructive to consider the causal sequences from an agent’s dispositions and intellect, via their wanting, deliberating, and choice, to action, as they occur in

⁵² There are several alternative interpretations of EN 1113b3–14. These (i) either leave Aristotle arguing fallaciously, or (ii) needlessly change the text, or (iii) face some other interpretational difficulties. Ad (i): both the interpretations (a) that the passage is a self-contained proof that virtue and vice in action is in our power and (b) that the passage is a proof that virtuous and vicious dispositions are in our power make Aristotle reason incorrectly (cf. Meyer 1993, 129–31). Ad (ii): the interpretation that in this passage Aristotle shows that we have undetermined, free, choice (Broadie 1991, Destrée 2011) is based on an emendation, introducing a verb of saying into (6) and reading ‘no’ (ou) instead of ‘not’ (mē) (cf. Bobzien 2014b). Ad (iii): the asymmetry interpretation of this passage (e.g. Broadie and Rowe 2002, 317; Pakaluk 2005, 145–6) assumes either (a) that it is assumed ad hominem that virtue is in our power or (b) that this has been shown in lines 1113b3–6; and argues from there that vice, too, is in our power. One problem with this interpretation is that it sits badly with Aristotle summarizing 3.1–5 by saying that virtue is voluntary and in our power: (a) makes his summary equal an ad hominem assumption of his; (b) makes his summary regard only four lines of EN 3.5 and leads to the philosophical oddity that while the voluntariness of virtue was justified with its being caused by our virtuous disposition, via choice, the voluntariness of vice is justified in an entirely different and needlessly complex argument, although an equivalent argument to that for virtue would have suggested itself and is assumed, e.g., in EE 1228a4–5. ⁵³ Cf. also EN 3.5.1113b21–6, where private individuals and lawmakers are said to punish people who do wicked things and praise people who do noble things: it is only after the argument 1113b3–14 that the praise of noble actions is mentioned in EN 3.1–5. But Aristotle noted the link between virtue and noble action already at EN 1.12.1101b31–2: ‘as a result of [virtue] people tend to do noble things (ta kala)’. ⁵⁴ Perhaps this transference might be expressed using Aristotle’s model of efficient causation in Met Z: As the form of health in the doctor’s soul is the efficient cause of any particular coming-to-be-healthy produced (1032b21–4), so the form of good action (its intermediateness) in the agent’s soul is the efficient cause of the agent’s particular good (i.e. intermediate) action. Either way, it is the person’s desire (the doctor’s desire to heal, the agent’s prohairesis) via which the efficient cause realizes the effect.

70

, ,   

Aristotle’s four main types of agents: the virtuous, vicious, strong-willed, and weak-willed.⁵⁵ I disregard all possible external interferences in the process from forming the choice to completing the action. These would include cases in which, unbeknownst to the agent, external factors cause something usually in the agent’s power not to be so, and cases in which, unforeseeable for the agent, external factors thwart the realization of their choice. I consider only cases in which, in Aristotle’s view, the agent is the origin of the action. T  : The virtuous agent is one in whom all factors of voluntary action are realized in an excellent way. (i) The agent’s character dispositions are good; they are virtues. They cause the agent’s ends, which are the intermediate relative to each dimension of virtue. Since the character dispositions are states of the desiring part of the soul, the virtuous agent desires—more precisely, wants—the intermediate with respect to each dimension of virtue. (ii) The agent’s practical intellect works without flaws. The agent has practical wisdom. As a result, and given the agent’s wants, the agent reasons correctly about how to realize the intermediate in their action, using an accurate assessment regarding what is in their power. (iii) All other desires of the agent are fully aligned with what the agent wants. (Not only does the agent want health-promoting food, and deliberates that leafy greens are health-promoting, they also love leafy greens, that is, have an appetite from them, and have a heartfelt dislike for fast food.) The agent’s wanting-of-the-end and deliberation-of-the-means together then cause a choice.⁵⁶ (iv) Since the agent’s other desires are fully aligned with what the agent wants, at the appropriate time the agent’s choice causes the action. Thus the action is voluntary. It is also (internally and externally)⁵⁷ noble (kalos), and that aspect of the action is also voluntary. The action, qua being noble, is thus praiseworthy. One noteworthy aspect of the actions of completely virtuous agents is that, although nothing forces the agents to act, for psychological reasons (i.e. for reasons to do with their soul), the agents cannot choose or act otherwise than they do. In the same circumstances the same virtuous agent will always choose and follow up the same noble course of action. There is nothing in the agent that could interfere with this. For Aristotle, this is no reason to question the voluntariness of ⁵⁵ Of the remaining two named types, Aristotle mentions, at EN 7.1 and 5, that brutes may be beyond blame, godlike heroes beyond praise. ⁵⁶ What happens if there are two equally good actions towards reaching the end? In that case, we can assume, external circumstances or non-moral elements of the agent’s character (a preference for red over blue, say) will determine the agent’s choice. If there are no such circumstantial elements, rather than not acting at all (like Buridan’s Ass) the rational agent would do something comparable to throwing a coin. ⁵⁷ An internally good/bad action is one co-caused by the respective virtue/vice. And externally good/ bad action is one that is indistinguishable from an internally good/bad one, except for the fact that it is not co-caused by the respective virtue/vice. The internal/external terminology is ours. But the distinction is Aristotle’s: cf. EN 1105b5–9. (From the externally good actions one must distinguish the accidentally [kata sumbebēkos] good actions from EN 5.8 that were illustrated by our tip example in section 8. They do not have the intermediate as end.)

      .–

71

the action or the agent. On the contrary, completely virtuous agents act with the minimal possible amount of force, since in addition to the absence of external force, there is also a total absence of psychological, internal factors that could be considered as compelling them to act.⁵⁸ T - : (i) The strong-willed and the virtuous agents have and want the same ends: the intermediate in actions.⁵⁹ The ends wanted are thus good (as opposed to just appearing good to the agent). What differentiates the strong-willed from the virtuous is that not all their desires are aligned with their wanted ends. (ii) The agents may or may not reason correctly about the means to their ends, and they may or may not accurately assess what is in their power. For simplicity, we assume they do both right. (iii) The agent’s willing-of-the-end and deliberation-of-the-means together cause a choice for the action that promotes the intermediate. (iv) Despite the agent’s desires that conflict with their wanted end, at the appropriate time the agent’s choice causes the action (EN 1111b13–15). (The agent wants health-promoting food, chooses to eat, and eats, the leafy greens, although their appetite is for the greasy sugary doughnuts.) Thus the action is voluntary. It is also (externally) noble,⁶⁰ and the result of a strong-willed disposition. T - : In points (i) to (iii), the weak-willed agent does not differ from the strong-willed. (iv) The element of weak-willedness (akrasia) comes to the surface in the relation between choice and action. It is at the point when one would expect the choice to cause the action that the agent’s conflicting desires interfere (EN 1111b13–14). They cause a temporary state of ignorance about some property of some object involved in the action (e.g. that this thing is sugary; that doughnuts are sugary; that sugary food is unhealthy).⁶¹ As a result, the person’s behaviour is aligned with their conflicting desire, say their appetite for doughnuts. It is this desire that causes an alternative to the chosen action (not eating the leafy greens; eating the doughnuts instead). The action is thus neither externally nor internally noble. Actions caused by weak-willedness are nonetheless voluntary. Aristotle is unequivocal on this point. First, he argues that actions that might be described as actions in which the agent was compelled to act the way they did by irrational ⁵⁸ Cf. also EN 1111a29–31, where Aristotle mentions that it would be absurd to call involuntary the things that one ought to desire. The fact that for Aristotle the completely virtuous person cannot choose other than they do provides no support for the assumption that Aristotle was a determinist. (After all, few of us, if any, are completely virtuous, and thus barred from the exhilarating experience of conflicting desires.) Rather, what the fact shows is that being able to choose otherwise was for Aristotle in no way a condition for justified praise. (Cf. also Frede 2011 ch. 2.) ⁵⁹ Arguably, one could have an ignoble end and be strong-willed, or an ignoble end and be weakwilled, but I disregard these possibilities here, too. A full classification of actions based on the various factors involved would provide a much larger number of possible cases than four. ⁶⁰ If the deliberation process was flawed, there would be no guarantee that the action is noble. ⁶¹ Different interpretations of Aristotle’s theory of weak-willedness assume different objects of ignorance. For an in-depth discussion of the topic see Lorenz 2014.

72

, ,   

desires like appetite or anger do not thereby become involuntary. Rather, appetite and anger are no less manifestations of who the agent is than is his reason. Hence, actions that ‘come from anger or appetite, too, are the actions of the person’ (EN 1111b1–2). Such actions satisfy the account of voluntariness, since external force is absent and the action’s origin is in the agent. Moreover, the temporary forgetfulness or ignorance is also not of the k ind that exempts actions from voluntariness. For the cause of the forgetting is internal to the agent, not external. Virtuous and strong-willed individuals would not have forgotten. Thus, although for Aristotle choice is characteristic for actions that are not ‘spur-of-the-moment’, it does not follow that the choice is always the cause of the action. In weak-willed actions it is not. T  : (i) The vicious agent’s character dispositions are bad. They are character vices. They cause the agent’s ends, which are bad, although they appear good to the agent. The ends are extremes, rather than the intermediate, with respect to a dimension of virtue. The vicious agent wants those extremes. (ii) Vicious agents may or may not deliberate correctly or accurately assess what is in their power. For simplicity, we assume they do both right. (iii) None of the other desires of the agent seem to be in conflict with what the agent wants. (Not only does the agent want to eat huge amounts of fast food, they also love quadruple cheeseburgers and grease-dripping doughnuts.) The agent’s want for some extreme together with their deliberation causes a choice. (iv) Since the agent’s other desires do not conflict with their wants, their choice causes the action, which is thus voluntary. It is also (internally and externally) bad, and that aspect of it is also voluntary. The action, qua being bad, is thus blameworthy. (Qua being an action of burger-eating, it may be praiseworthy: no ketchup on the shirt, no pickles on the pants, no choking on the bun.) A      : Thus, the actions of all four agent types involve choice, but only in three cases does their choice cause the action. Yet the conditions of voluntariness are satisfied each time. Each case involves choice and thus the relevant rational psychological process. Each time, the origin of the action is in the agent, and there is neither external force nor ignorance of the kind that eradicates voluntariness. Of course, for Aristotle, adult humans do not start out with a fully developed character. Hence, I consider how action is brought about in agents who do not yet have a fully developed character. (i) Such agents do not yet have moral dispositions. Still, they have wanting (EN 1114a11–13), that is, goal-directed desire, in the desiring part of their soul. Depending on their natural dispositions, their ends may or may not coincide with the intermediate (EN 6.13). The agents also have other desires, which are likely in conflict with some of their wants. (ii) Such agents have an ability for practical reasoning, and for assessing what is within their reach. They may not be very accurate on either front. (iii) This notwithstanding, the agent’s wants and

      .–

73

deliberation together (and, thus, the agent) will cause a choice. (iv) Depending on the strength of the agent’s conflicting desires, the choice either causes the corresponding action, or the conflicting desires prevail. Either way, the action is voluntary. The action may be externally noble or bad. If it is bad, whether by choice or weak-willedness, the agents may be blamed for it. If it is good as a result of choice, they may be praised.

12. Voluntary Consequences of Adult Voluntary Action People are praised and blamed not just for things they do or do not do, but also for consequences of things they did or did not do. Since voluntariness is a prerequisite for praise and blame, Aristotle also discusses voluntariness as it attaches to consequences of things people did or did not do. I call this indirect voluntariness. The main case of indirect voluntariness that Aristotle explains in Nicomachean Ethics 3.5 is that of the moral dispositions. But he starts with a more straightforward kind of indirect voluntariness, using examples. In his first example, a drunk person does something which they would not have done had they had some relevant circumstantial information, and which is punishable by law (and thus assumed to be blameworthy). Since the person lacks some relevant circumstantial information, what they did would not count as directly voluntary. It is not voluntary action. How can blame be attached to it nonetheless? The answer is that it is a consequence of an action which was directly voluntary. The person desired to drink, deliberated, chose, and drank. (Whether the choice was to drink or the drinking resulted from weak-willedness is irrelevant.) Nothing external forced them to drink. Hence their drinking—at least of their first few drinks—was voluntary. It was a foreseeable consequence of their drinking that they became drunk. This consequence, being drunk, a temporary condition, is indirectly voluntary. It is a consequence of being drunk that the person did not have certain information, which, had they had it, would have prevented them from acting. Thus, their acting is also indirectly voluntary (twice removed from the action, as it were). Hence it can be blamed. In a second example, a person breaks a law that they did not know existed, but the existence of which would have been easy for them to discover, had they cared to find out. Such persons are punished by law (and Aristotle assumes they are blameworthy). Perhaps the person did not know that it is illegal to dump dioxin in the public garbage site. A quick online search would have revealed this to them. Common sense would have suggested they check. Here we have a refraining (i.e. from checking), which is directly voluntary. The refraining is an instance of not taking care. (Not taking care here is neither a temporary condition, nor a character disposition, of the agent.) The refraining caused the condition of ignorance, which

74

, ,   

is indirectly voluntary; it in turn led to an act of law-breaking, which is indirectly voluntary (twice removed from the refraining).⁶² What Aristotle’s examples have in common (what makes them blameworthy) is that they are foreseeable consequences of voluntary action.⁶³ In terms of causation, there is a foreseeable causal chain that has its origin in the agent (EN 1113b30, 33). To express this fact, Aristotle invokes the alternative options the agent had when performing the (directly voluntary) actions. Here he uses the expression ‘master of ’ (kurios + genitive, 1113b32) in a twosided way: at the time of their actions, the agents were masters both of doing and of not doing those actions. This kind of formulation has been used to argue that Aristotle was an indeterminist with respect to voluntary actions: if at a time t the agent was master of not doing what they did at t, then at t the agent must have been causally undetermined with respect to doing or not doing what they did. They had the—indeterminist—freedom to do otherwise. However, just as in the case of Aristotle’s use of ‘in our power’ (see section 6), this is not so. For Aristotle, it is the fact that the agents (including their ends, awareness of alternative options, deliberation, and choice) are causally responsible for what they do that makes their actions and their foreseeable consequences voluntary—not that their actions or choices were causally undetermined.

13. The Voluntariness of Virtuous and Vicious Dispositions So far, we have seen that, for Aristotle, praise and blame can be (and are) bestowed on people for their individual actions, for the moral aspects of those actions, and for the foreseeable consequences of individual actions, including temporary mental states of the agent. Aristotle recognizes that people are also praised and blamed for who they are, in the sense of what moral dispositions they have.⁶⁴ People may be blamed for being bad, that is, not for any particular bad action but for those character dispositions that make them perform bad actions with some regularity. Likewise, they may be praised for being virtuous. People may also be blamed for being weak-willed, in the sense of continuously falling short of their choices owing to conflicting desires. In Nicomachean Ethics 3.1, Aristotle introduced ⁶² The action of breaking the law is indirectly voluntary. The action of dumping dioxin in the garbage site is directly voluntary. ⁶³ At this point we see how the criterion of absence of ignorance for direct voluntariness can produce borderline cases. The person who injured their friend when fencing since they did not know the button was off was assumed not to have injured their friend—directly—voluntarily (see section 3). But if it was a case in which she simply didn’t bother to check, she injured her friend voluntarily nonetheless, just indirectly so. And there may be no natural cut-off points for situations in which people are expected to check from those in which they are not. (In ordinary circumstances, we would not be expected to do a chemical analysis of a drink we offer a friend, to check on potential poisoning.) ⁶⁴ EN 1101b14–15, 1101b31–2, 1103a8–10, 1114a28–31. For an excellent discussion of Aristotle’s theory of our responsibility for our character, see Meyer 1993.

      .–

75

voluntariness as a necessary condition for praise and blame. Accordingly, in 3.5, he provides a theory of the voluntariness (and being-in-our-power) of character dispositions (1114a3–1114b25). Whatever his primary purpose for the passage may be,⁶⁵ in it he explains that, how, and why virtuous and vicious dispositions are voluntary. At the end of the passage he believes he has shown that the virtues, qua character dispositions, are in our power and voluntary; and the text where he shows this is EN 3.5.1114a3–1114b25.⁶⁶ In outline, Aristotle’s theory of the voluntariness of moral dispositions is the following. Virtues and vices are character dispositions. They are produced by an agent by repeatedly performing the actions that the disposition, once developed, will be a disposition for. Any agent who is not entirely uneducated knows this. The disposition-producing actions are voluntary and usually the result of choice— which warrants their voluntariness. Thus, (i) the disposition can be regarded as a foreseeable consequence of accumulative voluntary action of certain types. Moreover, (ii) the origin of the disposition lies in the agent—since the agent is the originator of the repeated action. Factors (i) and (ii) together qualify the disposition itself as voluntary—even though the process of its coming into being is a long, complex one, and is more vulnerable to external interferences than individual actions. Agents can work towards having certain character dispositions (e.g. certain virtues) similar to the way they can work towards achieving certain consequences of individual actions. The main differences are, first, that the kind of action needs repetition and, second, that the location of the consequence is the agent’s own soul. Aristotle brings to the reader’s attention one fundamental difference between direct and indirect voluntariness. This is the fact that agents can end directly voluntary things (like actions) at any point, but not indirectly voluntary things, including character dispositions (EN 1114a13–21; 1114b30–1115a3). Take a voluntary action: in standard cases, it is possible for agents to ‘change their mind’, that is, to reverse their choice, at any point during their performing the action, and consequently stop the action. For example, on your way to the store you can turn around and go home, or you can terminate your daily flossing any day. In the case of voluntary character dispositions, the equivalent does not hold. Here a change of mind would amount to a reversal of wanting (EN 1114a13–14). But the reversal of wanting, say, from wanting extremes as the end to wanting the intermediate as the ⁶⁵ The purpose (in fact of all of 1113b14–1114b25) is debated. Some think it is to show that the character dispositions are voluntary. Others think that the passage presupposes that virtuous dispositions are voluntary and argues from this to the conclusion that vicious dispositions are voluntary, too (e.g. Meyer 2006). Still others think that the purpose of the passage is to show that we are morally responsible for our character since causally undetermined, free actions are a partial cause of our character formation (e.g. Destrée 2011). For a counter to this assumption of causally undetermined choice, see my subsequent discussion. ⁶⁶ EN 1114b20–5 uses ‘voluntary’ for virtues and vices, in the plural; 1114a28–31 uses ‘in our power’ for dispositions (of body and soul).

76

, ,   

end, or the change from wanting a vicious disposition to wanting a virtuous one, does not enable the person, as a result, to cease to have the disposition they have— or, in any case, not instantaneously. Thus it is not a prerequisite for something x to be voluntary that, at any time during which x is present, x’s originator can, upon reversing their desire for x, make x stop or cease to exist—even if at no point during that time x is externally forced or the originator is relevantly ignorant. This should not surprise. First, it is a common feature of many indirectly voluntary things that they are not, or not instantly, reversible upon the originator’s desire for reversal: consider being drunk or serious self-mutilation. Second, voluntariness helps single out cases for justified praise or blame, and it would be absurd if it were a requirement for justified praise or blame that its object could be reversed or undone by its originator. Persons could then only be blamed for the death of someone they deliberately killed if they could bring them back to life upon accordingly reversing their desire. It is sometimes claimed that Aristotle maintains that the reason why actions of a vicious agent are voluntary and hence blameworthy lies in the fact that at some point in the past the agent could have reversed their development towards a vicious disposition (and mutatis mutandis for virtuous agents). But Aristotle never suggests or implies anything like this. His view is that any action that satisfies the criterion for direct voluntariness is voluntary for that reason. The fact that at some past time it was in the agent’s power to reverse their path towards a vicious disposition is as irrelevant to the action’s blameworthiness or praiseworthiness as is the fact that for psychological reasons an agent may be unable to act otherwise than they do.

3 Aristotle’s Nicomachean Ethics 1113b7–8 and Free Choice 1. EN 1113b7–8 and Its Putative Role as Evidence for Indeterminist Free Choice One of the various arguments proposed in favour of the claim that Aristotle had an indeterminist notion of free choice is based on a sentence in the first paragraph of Book 3 chapter 5 of the Nicomachean Ethics (EN). Here is a much-quoted translation of this sentence: (A)

(1) For where we are free to act we are also free to refrain from acting, (2) and where we are able to say No we are also able to say Yes. (EN 1113b7–8, tr. Rackham 1926)

The phrases ‘we are free to’ and ‘we are able to’ both translate eph’hēmin [estin].¹ The fact that eph’hēmin [estin] underlies both clauses is made explicit in a more recent rendering: (B)

(1) For when acting depends on us, not acting does so too, and when saying no does so, saying yes does too. (tr. Rowe)

The sentence translated by (A) and (B) is a central sentence of a central passage in which Aristotle is concerned with certain things that are eph’hēmin. To see why this sentence is thought to support indeterminist free choice, we first need to see what indeterminist free choice is. Authors often do not provide a definition,

This essay first appeared in P. Destrée, R. Salles, and M. Zingano (eds), What Is Up to Us? Studies on Causality and Responsibility in Ancient Philosophy (Sankt Augustin: Akademia Verlag, 2014), and is reproduced here with permission. A first version of this essay was presented as part one of a talk at Princeton University in April 2011, and I wish to thank the members of the audience for their terrific discussion and helpful observations, Dorothea Frede for a set of incisive written comments, and Sara Protasi and Geoff Moseley for some editorial assistance. ¹ More literal common English translations of eph’hēmin are ‘it is up to us’, ‘it is in our power’, ‘it lies with us’, ‘it depends on us’. Context determines what translation is most suitable. For a detailed discussion of Aristotle’s use of eph’hēmin, see Meyer 2014.

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0004

78

, ,   

but the following account should be sufficiently general to cover the various relevant views: An agent is undetermined, and thus free, in their choice (prohairesis), if at the moment of making their choice, what choice they make is not fully or sufficiently determined by preceding, and/or simultaneously existing, causes. At the moment of the agent’s making the choice, no causal factors prevent them from not making that choice.²

The argument for undetermined choice is introduced or implied by several Aristotle scholars. Pierre Destrée is one author who has recently supported the claim that for Aristotle choice is causally undetermined with the sentence that corresponds to (A) and (B). He argues that it is Aristotle’s view that human agents, when they act, could have done otherwise ‘in the strong sense, implying the idea of freedom of choice’ (Destrée 2011, 289–96). He regards the sentence behind (A) as ‘the most crucial . . . passage’ for his case (Destrée 2011, 289).³ Here is what he says: The first thing to be noted is the force of the expressions ‘saying Yes’, and ‘saying No’: ‘For when acting depends on us, not acting does so too, and when saying No does so, saying Yes does too’.⁴ It is obvious that Aristotle is here relying on a common way of thinking about our actions: in any case where we can say Yes, we also have the very possibility of saying No. In other words, we only voluntarily do something when we have the possibility of saying ‘No, I will not do it’. (Destrée 2011, 292)

And a little further down: And since one always has the very possibility of saying either Yes, or No, it means that acting this or that way depends on the way one will answer in the given case. In sum, from the way Aristotle relies on the connection that we commonly presume between speaking (‘Yes’, ‘No’) and action (compliance or refusal), it seems to be beyond any reasonable doubt, as S. Broadie very aptly says, that Aristotle assumes that ‘the agent is at least implicitly aware of the options as options’ (1991, 153) and that, one might add, the agent has the real possibility, or the choice, to go for one option instead of the other. (Destrée 2011, 292; cf. also 294 n. 11)

² This formulation allows for the possibility that the agents are a sufficient cause of their action, but are themselves unpredetermined in what choice they make. Cf. e.g. Bobzien 1998, 133–4. ³ In fact, Destrée says this of his text T1, which includes, in addition to (A), EN 1113b8–14. But at 292 Destrée admits that the following two sentences are not strong evidence in favour of free choice, so this leaves (A), which is what he explicitly uses as evidence. ⁴ This is Destrée’s version of (A).

    –   

79

‘Real possibility’ is one of Destrée’s ways of expressing causal indeterminism. He takes the ‘yes’ and ‘no’ as practical, or action-related. He regards them as short for ‘yes, I will do it’, ‘no, I won’t do it’. He also maintains that Aristotle assumes a corresponding connection between speaking and action, where ‘yes’ leads to compliance, ‘no’ to refusal. Here is Christof Rapp, referring to the Greek underlying (A): eine Handlung um ihrer selbst willen wählen heisst gerade, dass wir diese Handlung und ihre immanenten Ziele selbst für zustimmungswürdig halten. Weil wir diese Zustimmung auch nicht geben könnten (III 7, 1113b7ff.) sind wir für Handlungen, die wir um ihrer selbst willen wählen, voll verantwortlich. (Rapp 1995, 131)

Another author who has connected the sentence with free choice is Sarah Broadie. Here is a quote from her Ethics with Aristotle on our sentence: At NE 1113b7–11, Aristotle speaks of ‘it depends on him’ in terms of ‘Yes’ and ‘No’, and says that if the ‘Yes’ depends on the agent, so does the ‘No’ and conversely.⁵ The saying of ‘Yes’ or ‘No’ is practical, in that when F depends on the agent, if F occurs it is because the agent said ‘Yes’ to F, or refused to say ‘No’ to it, or because he said ‘No’ or refused to say ‘Yes’ to not-F. However, this formulation assumes that it is indeed open to one to say ‘Yes’ or ‘No’. (Broadie 1991, 153; cf. 154, 156, 159, and n. 31)

As for Destrée, for Broadie an agent’s saying ‘Yes’ and ‘No’ corresponds to the affirmation or refusal of an action and is in this sense practical. Human action presupposes such an act of affirmation (saying ‘Yes’) for action, an act of negation (saying ‘No’) for refraining. It is open to, or undetermined for, agents to affirm or refuse action by saying ‘Yes’ and ‘No’. And these undetermined acts of affirmation or refusal cannot be prevented by force (cf. also Broadie 1991, 156). Broadie rightly assumes that Aristotle was—as far as his extant writings go—not aware of any debate that contrasts free choice with causal determinism. However, she believes that for Aristotle choice is causally undetermined and that this is supported by the Greek sentence underlying (A) (Broadie 1991, 158). She concludes that ‘Aristotle may . . . be described as a proto-indeterminist’ (Broadie 1991, 158).⁶ In the same vein, Francis Sparshott writes about the passage that contains (A): ‘This is Aristotle’s essential thesis: our vices and virtues are the outcome of our

⁵ There is no phrase corresponding to ‘and conversely’ in the Greek. ⁶ Broadie also interprets EE 1222b41ff. in terms of agents saying ‘yes’ or ‘no’, and concludes that for Aristotle the agent’s ‘yes’ or ‘no’ would be contingent (1991, 159). Cf. also Broadie’s discussion of the no and the yes in her introduction to Broadie and Rowe 2002, 40.

80

, ,   

choices, and whenever we choose to say yes we could choose to say no’ (Sparshott 1994, 130, about EN 1113b3–21), and that for Aristotle ‘it is still true that choices involve saying yes or no, that whoever says yes can say no and vice versa’ (Sparshott 1994, 134 n. 82; emphasis mine).⁷ The interpretation of the sentence behind (A) and (B) as evidence for free choice is not a phenomenon that has its beginnings in the 1990s. Here is a paraphrase of our passage by George Stock from 1897. Stock rewrites the Nichomachean Ethics as dialogue, and here he has Aristotle speak: But when I speak of the voluntariness of virtue or vice, you must understand me to mean that the virtuous or vicious man is a free agent, that there is no force acting upon him except what comes from his own nature, except, in fact, himself. If he knows the right and the wrong, it is as open to him to choose the one as the other. Where he can do, he can refrain from doing, and where he can say ‘no’, he can say ‘yes’.⁸ (Stock 1897, 179; emphasis mine)

This will suffice as evidence that (A) and similar translations of EN 1113b7–8 have been used by Aristotle scholars to support their thesis that Aristotle propounded free choice or indeterminist choice. Now, to get from (A) to indeterminist free choice one needs to supplement two interpretative steps. The first interpretative step is that the Yes and No are taken as practical: ‘saying no’ and ‘saying yes’ are understood as shorthand for ‘saying yes to doing something’ and ‘saying no to doing something’, where the saying either denotes something like an internal soliloquy that accompanies the act of choosing; or where ‘saying yes’ or ‘saying no’ is Aristotle’s metaphorical way of expressing ‘choosing to act’ and ‘choosing not to act’. This interpretation of ‘yes’ and ‘no’ as practical also implies that in (A)(2) two possible choices or choosings are at issue: (i) to choose to act and (ii) to choose not to act. Yet, the Greek of (A)(1) only has ‘to act’; ‘not to act’. So the expected parallel would be ‘to choose (to act)’; ‘not to choose (to act)’, rather than ‘to choose to act’; ‘to choose not to act’. Quite generally, there is no textual evidence that Aristotle ever discusses an agent’s alternative options (i) to choose to do x and (ii) to choose not to do x—as opposed to (i) to choose to do x and (ii) not to choose to do x. The interpretation of ‘yes’ and ‘no’ as practical is by no means the only option. For example, as Christopher Taylor notes, Aristotle could instead be talking about it being up to us (eph’hēmin)

⁷ The ‘vice versa’ is Sparshott’s addition. There is no corresponding phrase in the Greek. ⁸ Cf. also vol.1 of Hammerton 1936, from the summary of Aristotle’s Nicomachean Ethics 5.3: ‘Choice is not the same thing as a voluntary act; nor is it desire, or emotion, or exactly “wish”, since we may wish for, but cannot make choice of, the unattainable. Nor is it deliberation—rather, it is the act of decision following deliberation. If man has the power to say yes, he has equally the power to say no, and is master of his own action.’ This clearly connects the saying Yes and saying No with choice qua an act of decision.

    –   

81

what to say: ‘Saying Yes and No may be given as another example of opposites which are equally up to us, just as acting and not acting are’ (Taylor 2006, 164).⁹ The second interpretative step is from choice to indeterminist choice: One needs to interpret Aristotle’s—presumed—statement that it is up to us to choose to do or not to do something as implying that the choice is causally undetermined. Thus it must not be fully determined by the agent’s character and deliberation as joint sufficient causes. But as Stephen Everson and Michael Pakaluk have independently noted, the ‘saying yes’ and ‘saying no’ in their versions of (A), even if one interprets this as practical (as they both do), does not entail that the choice at issue is causally undetermined.¹⁰ And since (A) is used by several of the abovequoted scholars as support for their thesis that for Aristotle choice is causally undetermined, this second interpretative step to indeterminist choice cannot simply be assumed. It needs to be argued for.

2. How to Translate the Sentence EN 1113b7–8 from the Greek Having given a long spiel about these three points regarding (A), I now abandon them together with (A) until further notice. Instead, I turn to the Greek lines of which (A) is purported to be a translation. I start by looking at these lines, in isolation from their immediate textual context. (C) provides the Greek: (C)

(1) En hois gar eph’hēmin to prattein, kai to mē prattein, (2) kai en hois to mē, kai to nai.*

There are no variants for our sentence in the apparatus criticus. (D) and (E) are two word-for-word literal (and thus somewhat unsightly) translations: (D) (1) For, where to act is up to us, also to not act [is up to us], (2) and where to not [is up to us], also to yes [is up to us].¹¹ (E)

(1) For, where (the) acting is up to us to act, also (the) not acting is up to us, (2) and where (the) not [is up to us], also (the) yes [is up to us].

⁹ Then again, Aristotle may think of saying as a kind of action (see e.g. EE 1225a27–33). Either way, saying and believing would come apart. For Aristotle suggests both in De Anima (427b20) and in the Eudemian Ethics (EE 1226a1–3) that belief (doxa) is not up to us (eph’hēmin). It would of course be a perfectly reasonable view that the two come apart in this way. ¹⁰ See Pakaluk 2005, 144–5; Everson 1990, 90. Broadie is aware of this point, too; cf. Broadie 1991, 156. * (C)(1) ἐν οἷς γὰρ ἐφ’ ἡμῖν τὸ πράττειν, καὶ τὸ μὴ πράττειν, (2) καὶ ἐν οἷς τὸ μή, καὶ τὸ ναί· ¹¹ Here and below I stick to the awkward infinitive-splitting ‘to not act’ instead of ‘not to act’ in order to provide maximal structural agreement with the Greek.

82

, ,   

The insertions in square brackets in (D) and (E) seem not to be questioned by anyone.¹² So I will not argue for them. (D) and (E) differ as follows. (D) keeps the syntactical role of the Greek to (τὸ) the same in (1) and (2), i.e. as introducing an infinitival clause; in (2) these clauses would be abbreviated. (E) replicates the grammatical category of the Greek word to (as definite article) by adding the English ‘the’ in brackets. It does not imply a parallelism of infinitival clauses between (1) and (2). I do not see how one can retain both points in one translation, which is why I have opted for two. Now to the points that matter. (i) There is no word for ‘saying’, or for anything similar, in the Greek text. In any rendering that has a verb of saying, meaning, intending, etc., this verb is an addition by the translator. And by adding such a verb, translators imply that they read or interpret the passage in a particular way. This way is likely to be along the lines presented in the first section. (As we have seen, such a reading implies neither that Aristotle maintains causal indeterminism nor that he expresses that choices are up to us, as opposed to actions or utterances). (ii) There is no word for ‘no’ in the Greek text. The word that appears to have been translated by ‘no’ does not mean no, and in the present use it means ‘not’. For something like (A) to be justified as a translation of (C), we hence should want at least two of the following three points satisfied. 2.1. We would want some textual parallels in which the Greek for ‘to not . . . to yes’ has the meaning of ‘to say “no” . . . to say “yes” ’, with a verb of saying either explicitly given or indubitably understood; these passages should preferably be by Aristotle, or roughly from Aristotle’s time (plus/minus five hundred years, say). 2.2. We would want an explanation as to how exactly to read (C) to mean something like (A), i.e. involving a verb of saying and ‘yes’ and ‘no’. 2.3. We would want there to be no reasonable alternative interpretation that avoids inserting a verb of saying and rendering mē (μή) by ‘no’. I take these three points in turn.

2.1 Parallel Passages Considered Our passage of the Nicomachean Ethics seems unique among ancient Greek texts in having to mē (τὸ μή) and to nai (τὸ ναί) directly opposed to each other, functioning as part of an abbreviation for something (if unclear so far for what). A fortiori, there seem to be no parallels in which the Greek for ‘to not . . . to yes’ has the meaning of ‘to say “no” . . . to say “yes” ’, with a verb of saying either explicitly

¹² A more literal translation of en hois (ἐν οἷς) would be ‘in those in which’. As this point is irrelevant to my purpose, I have opted for the shorter ‘where’. In principle, en hois could also refer to people as third-person plural masculine. However, it is unclear what the relevance of acting ‘among people’ rather than ‘in cases/circumstances/situations’ could be.

    –   

83

given or indubitably understood. Or in any event nobody has put one forward yet. The closest passages in Aristotle’s oeuvre are presumably the following six: • A dialectical proposition must be such that one can answer it with yes or no. (Arist. Top. 158a15–17) • But if the question is clear and simple, he should answer either yes or no. (Arist. Top. 160a33–4) • The person questioned should answer either yes or no. (Arist. Soph.Elench. 175b9–10) • The answerer must say either yes or no. (Arist. Soph.Elench. 175b13–14) • It is possible . . . for it to be true to say either yes or no. (Arist. Soph.Elench. 176a10–11) . . . he should not say (the) yes or no in the case of homonyms. (Arist. Soph.Elench. 176a15–16)¹³ What makes these passages at least worth contemplating is that they each have a Greek verb of saying (‘to say’ and ‘to answer’) combined with the standard Greek words for ‘yes’ and ‘no’. However, the passages are not close enough to provide support for the reading of our Greek sentence (C) as (A). First, they are all in the context of Aristotle’s dialectic. Saying or answering ‘yes’ or ‘no’ was part of the dialectical game or method, as is well established (Whitaker 1996, 101). There are no comparable passages in Aristotle in contexts of human agency where the agent says ‘yes’ or ‘no’ as a way of making a choice (deciding, agreeing, telling themselves) to do or not to do something. The passages from Aristotle’s dialectic fail as parallels for two further reasons. None of them has a definite article (to) in front of the ‘yes’ and the ‘no’, as (C) has; and all have ou (οὔ), the Greek word used equally for ‘not’ and for ‘no’, not mē. But (C) has mē. And mē is a word used adverbially for ‘not’ in certain grammatical contexts—but not for ‘no’.¹⁴ So the six passages are not parallel passages. Nor are there any other parallels in the Corpus Aristotelicum (or in any of the ancient Greek texts in the Thesaurus Linguae Graecae) which have to mē and to nai opposed in a sentence or sequence of sentences. By contrast, there are quite a few passages that oppose to nai and to ou.

¹³ ἔστι γὰρ πρότασις διαλεκτικὴ πρὸς ἣν ἔστιν ἀποκρίνασθαι ναί ἢ οὔ (Top. 158a15–17). ἐὰν δὲ καὶ σαφὲς ᾖ καὶ ἁπλοῦν τὸ ἐρωτώμενον, ἢ ναί ἢ οὔ ἀποκριτέον (Top. 160a33–4). τὸ ἢ ναί ἢ οὔ ἀποκρίνεσθαι τὸν ἐρωτώμενον (Soph.Elench. 175b9–10). ἢ ναί ἢ οὔ ἀνάγκη λέγειν τὸν ἀποκρινόμενον (Soph.Elench. 175b13–14). ἐγχωρεῖ . . . ἢ ναί ἢ οὔ ἀληθὲς εἶναι λέγειν (Soph.Elench. 176a10–11). οὐδ’ ἐπὶ τῶν ὁμωνύμων τὸ ναί ἢ οὔ λεκτέον (Soph.Elench. 176a15–16—several manuscripts do not have the τὸ; some have ἢ before τὸ). ¹⁴ The grammatical contexts in which μή is used adverbially for ‘not’ are typically one of the following: with the imperative; with the subjunctive; with the optative; with the infinitive; and with participles when they have a conditional or general force; and in certain indirect questions.

84

, ,   

Christopher Taylor, one of the very few philosophers who attempt to explain the—presumed—‘saying’ in (A), suggests a possible parallel in Aristotle’s Ethics. He writes: It is . . . possible that Aristotle is thinking of acting as itself a way of giving an affirmative answer to the question ‘Should I Φ?’ and not acting as a way of giving a negative answer to that question. Cf. VI.2, 1139a21–2, ‘what assertion and denial are in thought, pursuit and avoidance are in desire’, which seems to mean that pursuing some end is itself a way of asserting that the thing is to be pursued (or that it is good) and avoiding something a way of denying that it is to be pursued (or that it is good). (Taylor 2006, 164)

This attempt is, however, unsuccessful. The passage Taylor quotes does not ‘seem to mean’ what he suggests. If a and b are in thought what c and d are in desire, it is neither implied nor indicated that c-ing and d-ing are ways of a-ing and b-ing. For illustration: if I say that truth is in thought what the good is in desire, it does not follow that the good is a kind of truth. Rather, the case suggests that there is something that a and c, and b and d, respectively, share. Putting this point aside, the Aristotle passage also does not support the view that choice is at issue in (A) and (C). Rather, if we take Taylor at his word, what Aristotle seems to intend to say is (F) (1) For, where to act is up to us, also to not act [is up to us], (2) and where to not act [is up to us], also to act [is up to us]. In (F) (2) action, not choice, is the subject under discussion. We will see in the next section that to obtain (F) no verb of saying needs to be invoked. The problematic translation (A) would thus be unnecessary.

2.2 ‘Not’ and ‘Yes’ as ‘Don’t’ and ‘Do’ Can one provide a plausible justification of the insertion of a verb of saying and the translation of mē as no? None of the scholars proposing (A) seem to have provided one. Here is one suggestion as to what they could have provided. First, there is the fact that the ancients had no quotation marks, and that the definite article to was frequently used in a way similar to quotation marks in English, to indicate that an expression or sentence is mentioned, not used. This fact could be exploited to justify the translation ‘the “not” ’ and ‘the “yes” ’. Second, perhaps Aristotle’s use of mē in (C) is not meant to be equivalent to ‘ou’, meaning ‘no’, as used in answers to questions. Rather, mē may be employed by Aristotle the way it is used in independent clauses that express a command, as in mē praxēs,

    –   

85

‘don’t act!’: a person deliberates whether to do something; it is up to them to choose not to do it; in this case, their choosing involves their telling themselves ‘don’t act!’. In support of this reading, reference could be made to Aristotle’s sporadic allusions to the use of syllogisms in practical reasoning in the Ethics. An example, of a universal premise, would be EN 1147a34, ‘it says to avoid this’. Another passage one might adduce is EN 1143a8–9, ‘for practical wisdom gives commands. For what one must do or not [do] is its end’. Moreover, it could be adduced that Plato takes thinking to be internal speech, and that, a century after Aristotle, the Stoic Chrysippus had a theory that, in rational beings, the impulse to act takes the form of the agent’s reason prescribing or commanding the agent to act (Plutarch, Stoic. Rep. 1037f.). One could imagine this to take the following form: Zoe to herself ‘Don’t eat the baklava!’ or Zoe to herself ‘(Do) eat the baklava!’ Thus we would get a reading of the kind: (G)

(1) En hois gar eph’hēmin to prattein, kai to mē prattein, (2) kai en hois to mē [praxēis] kai to [praxon],*

in translation: (H)

(1) For, where to act is up to us, also to not act [is up to us], (2) and where the ‘don’t [act!’ is up to us], also the ‘do [act!’ is up to us].

This seems to go some way towards explaining the negative expression to mē (as opposed to to ou) in conjunction with an implicit verb of saying.¹⁵ Still, such imperative understanding of to mē poses its own difficulties. • The understanding of the Greek definite article ‘to’ in lieu of quotation marks obliterates the—expected—parallel between the two uses of ‘to’ in (1) and the two uses in (2). • The imperatival ‘mē’ is no longer on a par with the ‘nai’. In particular, grammatically nai is not part of positive commands, and cannot be short for ‘do act!’. • Moreover, with this reading one would still expect the sequence positive– negative in both (1) and (2), now taking the form ‘do!–‘don’t!’ in (2). Instead we have ‘don’t!’–‘do!’.

* (G) (1) ἐν οἷς γὰρ ἐφ’ ἡμῖν τὸ πράττειν, καὶ τὸ μὴ πράττειν, (2) καὶ ἐν οἷς τὸ μὴ [πράξῃς] καὶ τὸ [πρᾶξον]. ¹⁵ I have not found this reading of (C) explicitly defended anywhere but would be surprised if I was the first to consider this option. Perhaps Taylor 2006, 164 (quoted above) had something similar in mind.

86

, ,    • And, last but not least, we would need some supporting evidence in Aristotle’s text for the assumption that to mē in the middle of (C) can be short for our telling ourselves ‘don’t act!’. Some hints at practical reasoning in other books of the Nicomachean Ethics are not sufficient,¹⁶ and nor are references to Plato’s internal speech and Stoic impulses. We would need evidence that in Aristotle to mē can be a reference to a self-addressed imperative, which I believe we do not have.

So (G) is not a viable option. (Nor, of course, would (A) be a translation of (C) understood as (G).) Thus, as far as I can see, there is no legitimate way of getting a translation like (A) from (C). One would need even better reasons to subsequently interpret (A), or similar, as meaning that it is up to an agent to choose not to act or to choose to act. The following two reasons would not be good reasons for such an interpretation. (i) Aristotle had a theory of free choice which involved some kind of internal monologue of choosing. This reason is not good, since (C) is used as part of the main evidence for just this claim. The reasoning would be circular. (ii) We would naturally or intuitively think that Aristotle wants to say what the translation provides (cf. e.g. Destrée 2011, 288). This reason is not good, since we cannot infer from what we—or some of us—naturally or intuitively think today that the ancients thought this as well. If this were a methodologically sound step, it would become impossible to show the absence in antiquity of certain thoughts that are common to present thinking.

2.3 Alternative Ways to Translate and Understand the Sentence as Abbreviation No doubt, (C) does involve some abbreviation, ellipsis, and contraction, and there is no straightforward easy reading. Still disregarding its immediate linguistic context, I next ask: how else could one complete the elliptical clause (C)(2)? First I look at to mē, then at to nai. There is a natural and simple way of supplementing ‘to mē’. This is by supplementing the infinitive ‘to act’ (prattein) and reading ‘to mē’ as short for ‘to not act is up to us’ (to mē prattein eph’hēmin). (I) (1) For, where to act is up to us, also to not act [is up to us], (2) and where to not [act is up to us], also . . .

¹⁶ Also, the above-quoted passage EN 1143a8–9 is about practical wisdom (phronēsis). But in Aristotle’s view, not everyone has practical wisdom, whereas (C), i.e. EN 1113b7–8, is about all adult human beings. So the relevance of EN 1143a8–9 is doubtful.

    –   

87

There are quite a few examples in Aristotle in which mē is short for mē prattein (or a similar verb of doing or happening); several are in the context of what is eph’hēmin. In all cases, just beforehand in the sentence we find prattein (or that other verb of doing or happening).¹⁷ (I)(2) unquestionably provides the most natural way of supplementing to mē in the context of the whole sentence (C). So it is a good starting point. What are we then to do with to nai? To nai does not square straightforwardly with to mē, no matter what. Any interpreter is saddled with this problem. Still, if (I) provides the correct supplementation after to mē, it is clear what to nai is intended to convey: the alternative ‘to act’. So, assuming (I) to be correct, we get: (J)

(1) For, where to act is up to us, also to not act [is up to us], (2) and where to not [act is up to us], also to [act is up to us].

In other words, ‘to yes’ is short for ‘to act is up to us’ (to prattein eph’hēmin), as contrasted with ‘to not act is up to us’; i.e. with emphasis on ‘act’. (J) is indeed the way countless translators and scholars across the ages have read (C). Sometimes, (C)(2) is simply rendered by ‘and vice versa’. This may be to evade the awkwardness of the ‘to nai’, or to reflect the extreme brevity of (C)(2). (Henceforward, I call all translations of type (J) vice-versa translations.)¹⁸ Can the expression ‘to nai’ be read as an abbreviation of ‘to act [is up to us]’—as contrasted with to not act? Grammatically, this amounts to having to show that the Greek word for ‘yes’ in a phrase ‘to yes’ can be an abbreviation for an emphatic ‘to Φ’, contrasted with ‘to not Φ’ (where Φ stands for a verb of action). Can this be shown? Not by ancient Greek parallels. But perhaps a case can be made by comparison with other languages, together with what we know about the use of nai. The English ‘yes’ can—informally—stand in for what is expressed by the emphatic ‘doch’ in German. Take this dialogue: ‘this is green’—‘(no,) it isn’t’— ‘(yes,) it is’—‘no’—‘yes’; or its variation: ‘this is green’—‘(no,) it isn’t’—‘(yes,) it is’—‘is not’—‘is too’ (or ‘is so’). Similarly, the Latin etiam does not only mean ‘yes’, but also ‘also’, ‘again’. I wonder whether something similar may not hold of the ancient Greek nai. Take the following German translation of (C) in the spirit of the vice-versa translations (double square brackets are used to indicate what would be understood, but need not be supplemented in the translation):

¹⁷ E.g. EN 1110a17–8; EN 1143a8–9; EE 1223a5–8; cf. also Arist. Metaph. 1042b7–8; Arist. Rhet. 1359a36; and Arist. Athen.Const. 43 section 5. ¹⁸ Vice-versa translations can be found e.g. in Gauthier and Jolif 1970, vol. 1, 68 (‘et réciproquement’); Wardman 1963, 359; Apostle and Gerson 1983, 459; and also in some scholarly books and articles, e.g. Hardie 1980, 178.

88 (K)

, ,    (1) Denn wenn es bei uns liegt zu handeln, [liegt es] auch [bei uns] nicht zu handeln, (2) und wenn [[es bei uns liegt]] nicht [[zu handeln]], dann auch doch [[zu handeln]].

In English, a crude (and not quite right) equivalent would be: (L)

(1) For, where acting is up to us, also not acting [is up to us], (2) and where not [[acting]], also [[acting]] too.

The German at least seems grammatically just fine. Until proven otherwise, I assume that the analogue in Greek for (K) (with nai for the German ‘doch’) is fine as well.¹⁹ And, accordingly, the English vice-versa translation (J). In that case, we have a reading of (C) which is superior to all others in that it makes full sense of the text as it stands.²⁰ Even if this reading was not fully grammatical, (J) would still be vastly preferable to (A). For none of the alternatives, including (A), makes (C) fully grammatical. In addition, unlike (A), (J) (i) does not require the supplementation of a verb of saying; (ii) reads mē correctly as ‘not’; and (iii) requires no complex not-quitefitting interpretation.²¹ Based on reading (J), we can also explain what Aristotle’s point is in stating (C). It is to make explicit an important element of the logical structure of the notion of something’s being up to someone (epi (ἐπὶ) + dative), an element which Aristotle indicates in at least ten other places: this is its two-sidedness (see Chapters 2 and 3).²² Aristotle never provides a philosophical account of what it is for something to be eph’hēmin (as he does of the voluntary, deliberation, choice, virtue, etc.). He uses eph’hēmin and other epi + dative personal pronoun constructions as expressions of ordinary language which are generally understood by speakers of the language. In reading (J), sentence (C) makes explicit something people who speak the language assume: that doing something is up to us if and only if not doing it is up to us as well. To express this biconditional, both (C)(1) and (C)(2) are required.

¹⁹ We find such use of ‘nai’ for example several times in the work of the ninth-century grammarian Georgius Choeroboscus. For these passages and further parallels see Chapter 4 section 4 including nn. 14 and 15. ²⁰ The reading (J)(K)(L) differs from the don’t/do reading (G)(H). In (J)(K)(L) the ναί stands in as an abbreviation for a phrase that occurred in the exact same form earlier in the same sentence (i.e. πράττειν). In the case of the don’t/do reading (G)(H) this is not so. ²¹ Moreover, (iv), (J) makes perfect sense of the order affirmative/negative–negative/affirmative in the sequence of the two conditionals (C)(1) and (C)(2). If instead (C)(2) also had the order affirmative/ negative, this would simply be a repeat of (C)(1). It is thus ruled out. The interpretations behind (A) do not explain the inverted order. Proponents of (A) could try and plead the rhetorical device of chiasmus, though it would be a rather unusual case. ²² EE 1123a6, 7–8, 1223a5–6, 1225a9–10, 1225b35–6, 1226a27–8, 1226b30–1; EN 3.1.1110a17–18; EN 3.5.1115a2–3, 1125a26; cf. Bobzien 1998, 143–5, also 139–40.

    –   

89

Why does Aristotle state this biconditional at the beginning of Nicomachean Ethics 3.5, though? The reason is this: he needs to make explicit the logical structure of the notion of eph’hēmin at this point, since he exploits it as part of the argument at EN 1113b6–14.²³ That is, the biconditional is needed for the context of (C).

3. The Linguistic Context of (C) Taken into Account Hence, next I consider how readings (A) and (J) of the Greek sentence (C) fare, when one takes the immediate linguistic context into account.²⁴ First note the following three points. (i) The sentence before (C) is (M)

Now, virtue is up to us, too, and equally also vice.²⁵ (EN 1113b6–7)

(ii) The argument ends with the conclusion (marked by the particle ara) (N) so it will be up to us to be virtuous people and to be vicious people.²⁶ (EN 1113b13–14) This final clause of the argument states almost the same thing as (M). (iii) Our sentence (C) immediately follows (M) and begins with ‘for’ (gar). It thus provides a reason for (M). Combining (i) to (iii) with what we know about Aristotle’s dialectic, we get the following set-up of an argument from (M) to (N), or from 1113b6 to b14: Aristotle provides the thesis he intends to prove in (M) at the beginning of his argument. This is standard in Aristotelian dialectic. He then argues for (M) up until and including the antecedent of the sentence 1113b11–14. The consequent of the sentence is (N). It also provides the conclusion of his argument. Thus from (C) to (N) (i.e. from 1113b7–8 to b13–14) Aristotle provides an argument for the thesis that (not just actions but also) virtue and vice are up to us. This is so regardless of what exactly the thesis (M) and the conclusion (N) amount too. (C) is a premise in this argument. ²³ At EN 1113b14, Aristotle moves on to consider a possible objection. ²⁴ Here I ignore the debate about whether (i) the whole paragraph 1113b6–14 is meant to show that vice is up to us, with Aristotle taking it to have been shown already that virtue is up to us (the asymmetry reading); or whether (ii) the whole paragraph is meant to show that both virtue and vice (acting virtuously and acting viciously) are up to us. For the question of whether EN 1113b7–8 supports undetermined choice in Aristotle, this debate is only of minor importance. (Cf. also the next footnote.) ²⁵ ἐφ’ ἡμῖν δὴ καὶ ἡ ἀρετή, ὁμοίως δὲ καὶ ἡ κακία. In all MSS that are considered in the Oxford edition, the sentence does not have the particle δή (‘hence’, ‘now’), but the particle δέ. Modern editions of the EN tend to put δὴ. The emendation may be unnecessary, since Aristotle uses the particle combination δέ καὶ . . . δέ . . . elsewhere in the EN in one sentence—cf. e.g. EN 1103a8–10. ²⁶ ἐφ’ ἡμῖν ἄρα τὸ ἐπιεικέσι καὶ φαύλοις εἶναι.

90

, ,   

With the sentence following (C), Aristotle argues towards an intermediate conclusion: it starts with hōst’ (ὥστ’), which in grammatical contexts such as ours translates most naturally as ‘hence’: (O) Hence, (1) if to act, being noble, is up to us, also to not act, being shameful, will be up to us, and (2) if to not act, being noble, is up to us, also to act, being shameful, [will be] up to us. (EN 1113b8–11) (C) and (O) are strikingly parallel. This can be made explicit by using columns. I provide first the Greek, then the translation. The first column presents EN 1113b7–8, divided into four phrases, one per line; the second does the same with EN 1113b8–11. Bold indicates complete textual agreement between the two sentences. Underlining indicates phrases that are explicit in (O) and for which, regarding (C), it is not debated that they are understood at the parallel places. Column I (EN 1113b7–8) For, where it is up to us to act also to not act [is up to us] and where to not [is up to us] also to yes [is up to us]

Column II (EN 1113b8–11) Hence if to act, being noble, is up to us also to not act, being shameful, is up to us, and if to not act, being noble, is up to us also to act, being shameful, is up to us

en hois gar eph’hēmin to prattein* kai to mē prattein [eph’hēmin] kai en hois to mē [eph’hēmin] kai to nai [eph’hēmin]

hōst’ei to prattein kalon on eph’hēmin esti, kai to mē prattein eph’hēmin estai aischron on, kai ei to mē prattein kalon on eph’hēmin, kai to prattein aischron on eph’hēmin.

This striking parallel provides a strong reason for reading (C)(2) as an abbreviation along the lines that I—in agreement with numerous other scholars—have suggested, that is, as short for (C)

(2) kai en hois to mē [prattein eph’hēmin esti] kai to [prattein eph’hēmin esti]**

(J) (2) and where to not [act is up to us], also to [act is up to us]. This reading provides four perfectly matching cases, in the right order. The only substantive difference is that in (M) each time an evaluative attribute is added.²⁷ The apparent lack of parallelism in (C), from positive/negative in (1) to negative/

* ἐν οἷς γὰρ ἐφ’ ἡμῖν τὸ πράττειν ὥστ’ εἰ τὸ πράττειν καλὸν ὂν ἐφ’ ἡμῖν ἐστί, καὶ τὸ μὴ πράττειν [ ἐφ’ ἡμῖν ] καὶ τὸ μὴ πράττειν ἐφ’ ἡμῖν ἔσται αἰσχρὸν ὄν, καὶ εἰ τὸ μὴ πράττειν καλὸν ὂν ἐφ’ ἡμῖν, καὶ ἐν οἷς τὸ μή [ ἐφ’ ἡμῖν ] καὶ τὸ πράττειν αἰσχρὸν ὂν ἐφ’ ἡμῖν. καὶ τὸ ναί [ ἐφ’ ἡμῖν ] ** (C)(2) καὶ ἐν οἷς τὸ μὴ [πράττειν ἐφ’ἡμῖν ἐστί] καὶ τὸ [πράττειν ἐφ’ ἡμῖν ἐστί]. ²⁷ For the purpose of this addition at this point in Book 5 of the EN, see Bobzien 2014b.

    –   

91

positive in (2), finds a full explanation in the move from noble to shameful in the two conditionals of (J). Aristotle’s intent is to cover all four possibilities and their interrelations (noble action is paired with shameful inaction; noble inaction with shameful action), and for each interrelation he starts with the noble case. The parallel between (C) and (O) also provides strong reasons for not adding a verb of saying as readings of type (A) do. First, with (A), the inferential ‘hence’ (hōst’) is very hard to explain. Second, no mention is made of saying yes or no, or the like, ever again in the argument at issue. But if (A) were correct, we would expect some such mention, given the parallel structure of (O) and (C). We can be more precise: if (A) were correct we would expect the second half of (O) to be something like (P)(2): (P)

(2) and if saying yes, being noble, is up to us, also saying no, being shameful, is up to us.

But we do not have this. Rather, if (A) were correct, (O)(2) would just hang in the air, so to speak. There is nothing in the previous sentence for it to latch onto. Even without the linguistic context taken into account, (A) turned out to be a doubtful translation of (C). Now, with the context considered in addition, it becomes clear that translations like (A) misrepresent the Aristotelian text.²⁸

4. Concluding Remarks Where does all this leave us regarding the question of free choice in Aristotle? We have seen that there is a flourishing tradition among Aristotle scholars of reading sentence (C)(2) as an abbreviation for ‘if saying No is up to us, saying Yes is also up to us’ (i.e. as (A)(2)); and thus of reading it as containing an implicit verb of saying, and of translating the mē in (C)(2) by ‘no’;²⁹ further of understanding the sentence, thus translated, as being about practical assent or denial; and as expressing free choice or causally undetermined choice. I noted that, even if the reading of (C) as (A) were correct, there would be no compelling reasons to understand the sentence as being about practical assent and denial, or to understand it as expressing free choice, and a fortiori not as expressing causally undetermined choice; moreover, that the proponents of (A) face the difficulty of explaining the lack of parallelism in (A)(1) and (A)(2). Next ²⁸ For a comprehensive study of the reception of (C) from antiquity to the present day and a historical explanation of the frequency of translations of type (A) cf. Bobzien 2013. ²⁹ In addition to the authors and passages mentioned in section 1 above (Rackham 1926, Broadie 1991, Sparshott 1994, Destrée 2011, Rapp 1995, Stock 1897, Taylor 2006, Everson 1990, Pakaluk 2005), saying-yes-saying-no renderings are also found in the translations of Stahr 1863, 86; Williams 1869, 64; Peters 1881; Oswald 1962, 65; Crisp 2000, 45; Rowe 2002, 130; Taylor 2006, 24.

92

, ,   

I argued that (A) is in fact not an acceptable translation of (C). To this end, I showed that there are no parallels in Aristotle, or in antiquity generally, that would support readings like (A); I noted moreover, that the Greek seems not to lend itself to such a translation; and, finally, that there is a simple, most reasonable, alternative translation (i.e. the vice-versa reading (J)), which is not fraught with any of the difficulties (A) faces, and for which there are partial parallels in Aristotle and other Greek texts. Finally, I demonstrated that the linguistic context of (C) strongly supports reading (J), and by contrast does not tally with (A). A question that remains to be addressed is whether vice-versa readings like (J) in any way support the proposition that Aristotle, in Nicomachean Ethics 3.5, advocates free or indeterminist choice. Here is (J) again: (J)

(1) For, where to act is up to us, also to not act [is up to us], (2) and where to not [act is up to us], also to [act is up to us].

First and without doubt, (J)(1) and (J)(2) are both about acting and refraining from acting. Choice is neither expressly mentioned nor in any way implied as the topic of the sentence. Second, the sentence states nothing that had not been stated earlier in Nicomachean Ethics 3.1–4. Its purpose in 3.5 is to provide a premise for an argument that concludes that being virtuous and being vicious are both up to us. Third, the fact that Aristotle states that both acting and refraining from acting are up to us does not imply that we are causally undetermined with regard to whether we act. The phrase eph’hēmin may simply be used to express that there is an absence of external force (in the form of something external either making us move or preventing us from moving); and that, given this condition is satisfied, it depends on us, in the sense of on who we are at the point of deliberating and choosing, whether we act or refrain in the particular circumstances of action at issue.³⁰ From all this it follows that (C), i.e. EN 1113b7–8, is not a text that supports an interpretation of Aristotle as a proponent of indeterminist free choice or free will. That is, the sentence that is sometimes adduced as the main piece of evidence for the claim that Aristotle was an indeterminist with respect to choosing (prohairesis) and actions (praxeis, prattein) is no evidence for this claim at all.

³⁰ Cf. e.g. Everson 1990, 90; Bobzien 1998, 143–4.

4 Found in Translation Aristotle’s Nicomachean Ethics 3.5, 1113b7–8 and Its Reception

This chapter is distinctly odd. It illustrates what happens when an analytical philosopher and historian of philosophy tries their hand at the presently trending topic of reception. As a novice to this genre, it seemed advisable to start small. Rather than researching the reception of an author or a book, chapter, section, or paragraph, the focus of this chapter is on one sentence: Aristotle’s Nicomachean Ethics (EN) 3.5, 1113b7–8. This sentence has markedly shaped scholarly and general opinion alike with regard to Aristotle’s theory of free will. In addition, it has taken on a curious life of its own. Part One of the chapter examines the text itself. Part Two explores its reception from antiquity to the present day, including in present-day popular culture and in later ancient, Byzantine, Arabic, Latin Medieval, Renaissance, Victorian, and contemporary scholarship. There are some surprises along the way.

 .      3.5, 1113b7–8 1. EN 1113b7–8: An English Translation and the Greek Text One of the most famous sentences from Aristotle’s Nicomachean Ethics comes from his discussion of the voluntary (τὸ ἑκούσιον), choice (προαίρεσις), and what is up to us or in our power (ἐφ’ ἡμῖν), towards the beginning of EN 3.5. Here is a much-quoted translation of the lines:

This essay first appeared in Oxford Studies in Ancient Philosophy 45(2) (2013), 103–48 and is reproduced here with permission. Thanks to Dimitri Gutas, Stephen Menn, and Geoffrey Moseley for assisting me with the Arabic text of the Nicomachean Ethics; to Katerina Ierodiakonou for sharing some of her expertise in Byzantine philosophy; and to the audience at the Princeton Colloquium in April 2011, where I presented parts of this essay in draft form. Additional thanks go to Charles Brittain for his insightful remarks on a later draft, to Dorothea Frede and Brad Inwood for their judicious suggestions for further improving the essay, and to Sara Protasi for help with the bibliography.

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0005

94 (A)

, ,    (1) For where we are free to act we are also free to refrain from acting, (2) and where we are able to say No we are also able to say Yes. (EN 1113b7–8, trans. Rackham)¹

The phrases ‘we are free to’ and ‘we are able to’ both (seem to) translate ἐφ’ ἡμῖν [ἐστιν]. This sentence from the Nicomachean Ethics is often considered to be crucial for the interpretation of Aristotle’s view on the question of whether humans have indeterminist free choice.² Here are the Greek lines of which (A) is purported to be a translation, from Bywater’s Oxford edition:³ (B) (1) ἐν οἷς γὰρ ἐφ’ ἡμῖν τὸ πράττειν, καὶ τὸ μὴ πράττειν, (2) καὶ ἐν οἷς τὸ μή, καὶ τὸ ναί· There are no variants for our sentence in the apparatus criticus. (C) and (D) are two word-for-word literal (and thus somewhat unsightly) translations: (C)

(1) For, where to act is up to us, also to not act [is up to us], and where to not [is up to us], also to yes [is up to us].

(D) (1) For, where (the) acting is up to us, also (the) not acting [is up to us], (2) and where (the) not [is up to us], also (the) yes [is up to us]. The insertions in square brackets in (C) and (D) seem not to be questioned by anyone. So I will not argue for them. (C) and (D) differ as follows. (C) keeps the syntactical role of the Greek τὸ the same in (1) and (2), i.e. as introducing an infinitival clause; in (2) these clauses would be abbreviated. (D) replicates the grammatical category of the Greek word τὸ (as definite article) by adding the English ‘the’ in brackets. It does not imply a parallelism of infinitival clauses between (1) and (2). I do not see how one can retain both points in one English translation, which is why I have opted for two. Now to the points that matter. (i) There is no word for ‘saying’, or for anything similar, in the Greek text. In any rendering that has a verb of saying, meaning, intending, etc., this verb is an addition by the translator. And in supplying such a verb, translators imply that they read or interpret the passage in a particular way. (ii) There is no word for ‘no’ in the Greek text. The word that appears to have been translated by ‘no’ does not mean no, and in the present use it means not. ¹ Rackham 1926. ² Destrée 2011, 285–318, §§1 and 2; Sparshott 1994, 130; Rapp 1995, 109–33, at 131; also many of the authors mentioned in section 7 below. The general idea is often that our free (un-predetermined, uncaused, or unforced) choice is manifested or expressed in our ability to say either ‘no’ or ‘yes’ to (to either reject or choose) the course of action we deliberate about. For a detailed discussion, see Chapter 3 (previously published as Bobzien 2014a). See also below, section 13. ³ Bywater 1894.

  .,   –   

95

For something like (A) to be justified as a translation of (B), we hence should want at least two of the following three points satisfied. 1. We would want some textual parallels in which the Greek for ‘to not . . . to yes’ has the meaning of ‘to say “no” . . . to say “yes” ’, with a verb of saying either explicitly given or indubitably understood; these passages should preferably be by Aristotle, or roughly from Aristotle’s time (plus/minus five hundred years, say). 2. We would want an explanation as to how exactly to read (B) to mean something like (A), i.e. involving a verb of saying and ‘yes’ and ‘no’. 3. We would want there to be no reasonable alternative interpretation that does not insert a verb of saying and does not render μή by ‘no’. I take these three points in turn.

2. Parallel Passages Considered Our passage from the Nicomachean Ethics seems unique in ancient Greek texts in having τὸ μή and τὸ ναί directly opposed to each other, functioning as part of an abbreviation for something (if unclear so far for what). A fortiori, there seem to be no parallels in which the Greek for ‘to not . . . to yes’ has the meaning of ‘to say “no” . . . to say “yes” ’, with a verb of saying either explicitly given or indubitably understood. Or in any event nobody has put one forward yet. The closest passages in Aristotle’s oeuvre are apparently the following six: •

A dialectical proposition must be such that one can answer it with yes or no. ἔστι γὰρ πρότασις διαλεκτικὴ πρὸς ἣν ἔστιν ἀποκρίνασθαι ναί ἢ οὔ· (Arist. Top. 158a15–17)

•

But if the question is clear and simple, he should answer either yes or no. ἐὰν δὲ καὶ σαφὲς ᾖ καὶ ἁπλοῦν τὸ ἐρωτώμενον, ἢ ναί ἢ οὔ ἀποκριτέον. (Top. 160a33–34)

•

The person questioned should answer either yes or no. τὸ ἢ ναί ἢ οὔ ἀποκρίνεσθαι τὸν ἐρωτώμενον. (SE 175b9–10)

•

The answerer must say either yes or no. ἢ ναί ἢ οὔ ἀνάγκη λέγειν τὸν ἀποκρινόμενον. (SE 175b13–14)

•

It is possible . . . for it to be true to say either yes or no. ἐγχωρεῖ . . . ἢ ναί ἢ οὔ ἀληθὲς εἶναι λέγειν· (SE 176a10–11)

•

. . . he should not say (the) yes or no in the case of homonyms. οὐδ’ ἐπὶ τῶν ὁμωνύμων τὸ⁴ ναί ἢ οὔ λεκτέον· (SE 176a15–16)

⁴ Several manuscripts do not have the ‘τό’. Some have ‘ἢ’ before ‘τό’.

96

, ,   

What makes these passages at least worth contemplating is that they each have a Greek verb of saying⁵ combined with the standard Greek words for ‘yes’ and ‘no’. However, the passages are not close enough to provide support for the reading of our Greek sentence (B) as (A). First, they are all in the context of Aristotle’s dialectic. Saying or answering ‘yes’ or ‘no’ was part of the dialectical game or method, as is well established.⁶ There are no comparable passages in Aristotle in contexts of human agency where the agent says ‘yes’ or ‘no’ as a way of making a choice (deciding, agreeing, telling themselves) to do or not to do something. The passages from Aristotle’s dialectic fail as parallels for two further reasons. None of them has a definite article (τὸ) in front of the ‘yes’ and the ‘no’, as (B) has, and all have οὔ, the Greek word used equally for ‘not’ and for ‘no’, not μή. But (B) has μή. Now, μή is a word used adverbially for ‘not’ in certain grammatical contexts—but not for ‘no’.⁷ So the six passages are not parallel passages. Nor are there any other parallels in the Corpus Aristotelicum (or in any of the ancient Greek texts in the Thesaurus Linguae Graecae) which have τὸ μή and τὸ ναί opposed in a sentence or sequence of sentences.⁸ Christopher Taylor, one of the very few philosophers who attempt to explain the—presumed—‘saying’ in (A),⁹ suggests a possible parallel in Aristotle’s Ethics. He writes: It is . . . possible that Aristotle is thinking of acting as itself a way of giving an affirmative answer to the question ‘Should I Φ?’ and not acting as a way of giving a negative answer to that question. Cf. VI.2, 1139a21–2, ‘what assertion and denial are in thought, pursuit and avoidance are in desire’, which seems to mean that pursuing some end is itself a way of asserting that the thing is to be pursued (or that it is good) and avoiding something a way of denying that it is to be pursued (or that it is good).¹⁰

This is a brave attempt, though unsuccessful. The passage Taylor quotes does not ‘seem to mean’ what he suggests. If a and b are in thought what c and d are in desire, it is neither implied nor indicated that c-ing and d-ing are ways of a-ing and b-ing. For illustration: if I say that truth is in thought what the good is in desire, it does not follow that the good is a kind of truth. Rather, the case suggests that there

⁵ These are ‘to say’ (λέγειν) and ‘to reply’ (ἀποκρίνεσθαι). ⁶ Cf. e.g. Whitaker 1996, 101. ⁷ The grammatical contexts in which ‘μή’ is used adverbially for ‘not’ are typically one of the following: with the imperative; with the subjunctive; with the optative; with the infinitive; and with participles when they have a conditional or general force and in certain indirect questions; cf. Smyth 1920, 604–6, 608–30. ⁸ See section 4 for some Byzantine Greek parallels. By contrast, there are quite a few passages that oppose ‘τὸ ναί’ and ‘τὸ οὔ’. ⁹ The reader is not told, though, that there is no expression for ‘saying’ in the Greek. ¹⁰ Taylor 2006, 164.

  .,   –   

97

is something that a and c, and b and d, respectively, share. What that something is would need to be separately identified. If we put this point aside and take Taylor at his word, what Aristotle seems to intend to say is (E)

(1) For, where to act is up to us, also to not act [is up to us], (2) and where to not act [is up to us], also to act [is up to us].

We will see in the next section that to obtain (E) no verb of saying needs to be invoked. The problematic translation (A) would thus be unnecessary.

3. ‘Not’ and ‘Yes’ as ‘Don’t’ and ‘Do’ Can one provide a plausible justification for the insertion of a verb of saying and the translation of μή as no? Not one of the scholars proposing (A) seems to have done so. Here is one suggestion as to what they could have provided. First, there is the fact that the ancients had no quotation marks, and that the definite article τὸ was frequently used in a way similar to quotation marks in English, to indicate that an expression or sentence is mentioned, not used. This fact could be exploited to justify the translation ‘the “not” ’ and ‘the “yes” ’. Second, perhaps Aristotle’s use of μή in (B) is not meant to be equivalent to ‘οὔ’, meaning ‘no’, as used in answers to questions. Rather, μή may be employed by Aristotle in the way it is used in independent clauses that express a command, as in μὴ πράξῃς, ‘don’t act!’:¹¹ a person deliberates whether to do something; it is up to them to choose not to do it; in this case, their choosing involves their telling themselves ‘don’t act!’. In support of this reading, reference could be made to Aristotle’s sporadic allusions to the use of syllogisms in practical reasoning in the Ethics. An example, of a universal premise, would be EN 1147a34, ‘it says to avoid this’ (ἣ μὲν οὖν λέγει φεύγειν τοῦτο). Another passage one might adduce is EN 1143a8–9, ‘for practical wisdom gives commands. For what one must do or not [do] is its end’ (ἡ μὲν γὰρ φρόνησις ἐπιτακτική ἐστιν· τί γὰρ δεῖ πράττειν ἢ μή, τὸ τέλος αὐτῆς ἐστίν). Moreover, it could be adduced that Plato takes thinking to be internal speech, and that a century after Aristotle, the Stoic Chrysippus had a theory that, in rational beings, the impulse to act takes the form of the agent’s reason prescribing or commanding the agent to act (Plutarch, Stoic. Rep. 1037f.). One could imagine this to take the following form: Zoe to herself ‘Don’t eat the baklava!’ or Zoe to herself ‘Eat the baklava!’ Thus we would get a reading of the kind:

¹¹ This use of ‘μή’ with the subjunctive aorist has a parallel in the Latin ‘ne’.

98

, ,   

(F)

(1) ἐν οἷς γὰρ ἐφ’ ἡμῖν τὸ πράττειν, καὶ τὸ μὴ πράττειν, (2) καὶ ἐν οἷς τὸ μὴ [πράξῃς] καὶ τὸ [πρᾶξον],

in translation: (G)

(1) For, where to act is up to us, also to not act [is up to us], (2) and where the ‘don’t [act!’ is up to us], also the ‘do [act!’ is up to us].

This seems to go some way towards explaining the τὸ μή (as opposed to τὸ οὔ) in conjunction with an implicit verb of saying.¹² Still, such an imperative understanding of τὸ μή poses its own difficulties. • The understanding of the definite article τὸ in lieu of quotation marks obliterates the—expected—parallel between the two uses of τὸ in (1) and its two uses in (2). • The μή is no longer on a par with the ναί. In particular, grammatically ναί is not part of positive commands, and cannot be short for ‘do act!’. • Moreover, with this reading one would still expect the sequence positive– negative in both (1) and (2), now taking the form ‘do!’—‘don’t!’ in (2). Instead the reading has ‘don’t!’—‘do!’. • And, last but not least, we would need some supporting evidence in Aristotle’s text for the assumption that τὸ μή in the middle of (B) can be short for our telling ourselves ‘don’t act!’. Some hints at practical reasoning in other books of the Nicomachean Ethics are not sufficient, and nor are references to Plato’s internal speech and Stoic impulses. We would need evidence that in Aristotle τὸ μή can be a reference to a self-addressed imperative, which I believe we do not have. So (F) is not a viable option. (Nor, of course, would (A) be a translation of (B) understood as (F).) Thus, as far as I can see, there is no legitimate way of getting a translation like (A) from (B).

4. Alternative Ways to Translate and Understand the Sentence as Abbreviation No doubt, (B) does involve some abbreviation, ellipsis, and contraction, and there is no straightforward easy reading. Still disregarding its linguistic context, I next

¹² I have not found this reading of (B) explicitly defended anywhere, but I doubt that I am the first to consider this option. Perhaps Taylor 2006, 164 (quoted above) had something similar in mind.

  .,   –   

99

ask: how else could one complete the elliptical clause (B2)? First I look at τὸ μή, then at τὸ ναί. There is a natural and simple way of supplementing ‘τὸ μή’. This is by supplementing the infinitive ‘to act’ (πράττειν) and reading ‘τὸ μή’ as short for ‘to not act is up to us’ (τὸ μὴ πράττειν ἐφ’ ἡμῖν). (H)

(1) For, where to act is up to us, also to not act [is up to us], (2) and where to not [act is up to us], also . . .

There are quite a few examples in Aristotle in which ‘μή’ is short for ‘μὴ πράττειν’ (or a similar verb of doing or happening); several are in the context of what is ἐφ’ ἡμῖν. In all cases, just beforehand in the sentence we find ‘πράττειν’ (or that other verb of doing or happening).¹³ (H2) unquestionably provides the most natural way of supplementing ‘τὸ μή’ in the context of the whole sentence (B). So it is a good starting point. What are we then to do with ‘τὸ ναί ’? ‘τὸ ναί’ seems not to square straightforwardly with ‘τὸ μή’, no matter how interpreted. Any interpreter is saddled with this issue. Still, if (H) provides the correct supplementation after ‘τὸ μή’, it is clear what ‘τὸ ναί’ is intended to convey: the alternative ‘to act’. So, assuming (H) to be correct, we get: (I)

(1) For, where to act is up to us, also to not act [is up to us], (2) and where to not [act is up to us], also to [act is up to us].

(1) ἐν οἷς γὰρ ἐφ’ ἡμῖν τὸ πράττειν, καὶ τὸ μὴ πράττειν, (2) καὶ ἐν οἷς τὸ μή [πράττειν ἐφ’ ἡμῖν], καὶ τὸ [πράττειν ἐφ’ ἡμῖν]· In other words, ‘τὸ ναί’ is short for ‘to act is up to us’ (τὸ πράττειν ἐφ’ ἡμῖν), as contrasted with ‘to not act is up to us’, i.e. with emphasis on ‘act’. (I) is indeed the way countless translators and scholars across the ages have read (B). Sometimes (B2) is simply rendered by ‘and vice versa’. This may be to evade the awkwardness of the ‘τὸ ναί’, or to reflect the extreme brevity of (B2). (Henceforward, I call all translations of type (I) vice-versa translations.) Can ‘τὸ ναί’ be read as an abbreviation of ‘to act [is up to us]’—as contrasted with ‘to not act’)? Grammatically, this amounts to having to show that in the phrase ‘τὸ ναί’ the Greek word usually translated ‘yes’ can be an abbreviation for an emphatic ‘to Φ’, contrasted with ‘to not Φ’ (where Φ stands for a verb in the infinitive). Can this be shown? Not by ancient Greek parallels. ¹³ EN 1110a17–18 ὧν δ’ ἐν αὐτῷ ἡ ἀρχή, ἐπ’ αὐτῷ καὶ τὸ πράττειν καὶ μή (i.e. πράττειν); EN 1143a8–9 τί γὰρ δεῖ πράττειν ἢ μή (i.e. πράττειν), τὸ τέλος αὐτῆς ἐστίν; EE 1223a5–8 φανερὸν ὅτι ἐνδέχεται καὶ γίνεσθαι καὶ μή (i.e. γίνεσθαι), καὶ ὅτι ἐφ’ αὑτῷ ταῦτ’ ἐστι γίνεσθαι καὶ μή (i.e. γίνεσθαι), ὧν γε κύριός ἐστι τοῦ εἶναι καὶ τοῦ μὴ εἶναι. ὅσα δ’ ἐφ’ αὑτῷ ἐστι ποιεῖν ἢ μὴ ποιεῖν. Cf. also Meta. 1042b7–8 τίς μὲν οὖν διαφορὰ τοῦ ἁπλῶς γίγνεσθαι καὶ μὴ (i.e. γίνεσθαι) ἁπλῶς, ἐν τοῖς φυσικοῖς εἴρηται; Rhet. 1359a36 τῶν ἐνδεχομένων καὶ γίγνεσθαι καὶ μή (i.e. γίνεσθαι); Athen.Const. 43.5 καὶ περὶ τῆς ὀστρακοφορίας ἐπιχειροτονίαν διδόασιν, εἰ δοκεῖ ποιεῖν ἢ μή (i.e. ποιεῖν).

100

, ,   

However, there are several Byzantine Greek parallels for just such a use of ‘ναί’ and even of the phrase ‘τὸ ναί ’. For example, we find such an abbreviatory use of ‘ναί ’ several times in the work of the ninth-century grammarian Georgius Choeroboscus, when he is talking about the conjugation of verbs. (His views have survived in notes taken by his pupils.) Here is one instance: . . . ἐπειδὴ οὐκ ἔστι τὸ α χαρακτηριστικὸν τῶν τρίτων προσώπων τῶν ἑνικῶν· τὸ δὲ ε ναί, οἷον ἔτυπτε ἔλεγε·¹⁴ (‘ . . . since the alpha is not a characteristic of the third-person singular; but the epsilon is, for example ἔτυπτε ἔλεγε.’ The emphatic ‘is’ translates ‘ναί ’.) In this clause, the ‘ναί’ is short for ‘ἔστι χαρακτηριστικὸν τῶν τρίτων προσώπων τῶν ἑνικῶν’ (‘is a characteristic of the third-person singular’), which is contrasted with the previous clause, which negates the same predicative phrase, just as the ‘ναί ’ in (B) does in the vice-versa reading (I). The earliest surviving parallel (of at least three) for the use of ‘τὸ ναί ’ as abbreviation of an infinitive or infinitival phrase occurs in the third of the Orationes of the bishop and philosopher Eustratius of Nicaea (c.1050/1060–c.1120), where he discusses the Filioque. He writes: Ὁ Λατῖνος. Καὶ δεῖξόν, φησιν, οὗ τῆς Γραφῆς εὕρηται τὸ μὴ ἐκ τοῦ Υἱοῦ ἐκπορεύεσθαι. Πρὸς τοῦτο ἐγώ. Οὐκ ἔγω γε ἀπαιτοῦμαι δεῖξαι τὸ μή, ἀλλ’ αὐτὸς τὸ ναί.¹⁵ (‘The Latin [i.e. the Roman Catholic]: and, he says, show me, where in the Scripture it is found that it (the Spirit, πνεῦμα from line 22) does not proceed from the Son. Against this I [say]: I am not required to show that it (the Spirit) does not [proceed from the Son]; but he [is required to show] that it does [proceed from the Son].’) Here the emphatic ‘that it does’ translates ‘τὸ ναί ’ and is short for ‘that it does proceed from the Son’. So there is good evidence that in the Greek language (B) would naturally have been understood as an abbreviation for the Greek version of the vice-versa reading (I). The abbreviatory use of ‘ναί ’ has at least partial parallels in contemporary languages. Thus the German emphatic ‘doch’ corresponds to ‘ναί ’ in this use and allows an almost literal translation of (B) in the spirit of the vice-versa translations:

¹⁴ Georgius Choeroboscus, Prolegomena et scholia in Theodosii Alexandrini canones isagogicos de flexione verborum, ed. A. Hilgard (GG 4.2; Leipzig, 1894), 86.29–31; cf. ibid. 85.17–18; 86.34–5; 336.25–6. ¹⁵ Eustratius of Nicaea, Orationes, ed. A. Demetrakopoulos (Ἐκκλησιαστικὴ Βιβλιοθήκη, 1; Leipzig, 1866; repr. Hildesheim, 1965), Oration 3, 86.23–6. The other two parallels are: Joannes XI Beccus [the Patriarch], Four Books to Constantine Meliteniotes, ed. J.-P. Migne (Patrologia Graeca, 141; Paris, 1865), 337–96 at 388.16–18: ἕτερον τῷ τὴν ἐξ Υἱοῦ διδόντι ἐκπόρευσιν φήσειέ τις εἶναι διὰ σπουδῆς, ἢ ἀπεναντίας τοῖς ἀντιλέγουσι, (15) μετεῖναι τῆς αἰτίας ταύτης τῷ Υἱῷ συνιστᾷν; Εἰ γὰρ ἡ ἄρνησις διὰ τὸ μὴ, ἡ ὁμολογία πάντως διὰ τὸ ναί; and Manuel II Palaeologus [the Emperor, orator and theologian (1350–1425)], Dialogi cum mahometano, dialogue 17, in Piissimi et sapientissimi imperatoris Manuelis Palaeologi opera omnia, theologica, polemica, panegyrica, paedagogica, ed. J.-P. Migne (Patrologia Graeca, 156; Paris, 1866), 221.29–31: ἀλλ’ εἰ τὸ μὴ διαφέρειν οὐκ εὔλογον, ὅπερ ἐχρῆν, ἀποδέδεικται· εἴη γὰρ ἂν εὔλογον τὸ διαφέρειν, ἀδύνατον γὰρ εἶναι καὶ τὸ μὴ καὶ τὸ ναὶ ἐν ἑνὶ τῷ ζητουμένῳ.

  .,   –    (J)

101

(1) Denn wenn es bei uns liegt zu handeln, [liegt es] auch [bei uns] nicht zu handeln, (2) und wenn [[es bei uns liegt]] nicht [[zu handeln]], dann auch doch [[zu handeln]].¹⁶

The English ‘yes’ or ‘too’ can—informally—stand in for what is expressed by the emphatic ‘doch’ in German. Take this dialogue: ‘this is green’—‘(no,) it isn’t’— ‘(yes,) it is’—‘no’—‘yes’; or its American English variation: ‘this is green’—‘(no,) it isn’t’—‘(yes,) it is’—‘is not’—‘is too’. So in English, a crude equivalent to (J) would be: (K)

(1) For, where acting is up to us, also not acting [is up to us], (2) and where not [[acting is up to us]], also [[acting]] too.

The German seems to be grammatically just fine and there is a high likelihood that the English would be understood by a native speaker. Thus we have sufficient reason to conclude that (B) is the Greek equivalent to (J) and (K), and moreover, given the grammatical restrictions of the English language, that English vice-versa translations like (I) that express the fleshed-out Greek version (I) of (B) represent the correct way of translating (B). Thus we have a reading of (B) which is superior to all others in that it makes full sense of the text as it stands.¹⁷ Overall, then, (I) is vastly preferable to (A). There are clear Greek parallels to (I) but not to (A). In addition, unlike (A), first, (I) does not require the supplementation of a verb of saying; second, (I) reads ‘μή’ correctly as ‘not’; and, finally, (I) requires no complex not-quite-fitting interpretation.¹⁸ Based on reading (I), we can also explain what Aristotle’s point is in stating (B). It is to make explicit an important element of the logical structure of the notion of something’s being up to someone (ἐπὶ + dative), an element which Aristotle indicates in at least ten other places: this is its two-sidedness (see Chapters 1 and 2).¹⁹ Aristotle never provides a philosophical account of what it is for something to be ἐφ’ ἡμῖν (as he does of the voluntary, deliberation, choice, virtue, etc.). He uses ἐφ’ ἡμῖν and other ἐπὶ + dative personal pronoun constructions as expressions of ordinary language which are generally understood by speakers of ¹⁶ Double square brackets are here used to indicate what would be understood, but need not be supplemented in the translation for the translation to make sense. ¹⁷ The reading (I)(J)(K) differs from the don’t/do reading (G)(H). In (I)(J)(K) the ‘ναί’ stands in as an abbreviation for a phrase that occurred in the exact same form earlier in the same sentence (i.e. ‘πράττειν’). In the don’t/do reading (G)(H) this is not so. ¹⁸ Moreover, (I) makes perfect sense of the order affirmative/negative—negative/affirmative in the sequence of the two conditionals (B1) and (B2). If instead (B2) also had the order affirmative/negative, this would simply be a repeat of (B1). It is thus ruled out. The interpretations behind (A) do not explain the inverted order. Proponents of (A) could plead the rhetorical device of chiasmus, but it would be a somewhat unusual and strained case. ¹⁹ EE 1223a5–9; 1225a9–10; 1225b35–6; 1226a27–8; 1226b30–1; EN 3.1.1110a17–18; 3.5.1115a2–3, 1125a26; cf. Bobzien 1998, 133–75 at 143–5, also 139–40.

102

, ,   

the language. In reading (I), sentence (B) makes explicit something that people who speak the language assume: that doing something is up to us if and only if not doing it is up to us as well. To express this biconditional, both (B1) and (B2) are required. Why does Aristotle state this biconditional at the beginning of Nicomachean Ethics 3.5, though? The reason is this: he needs to make explicit the logical structure of the notion of ἐφ’ ἡμῖν at this point, since he exploits it as part of the argument at EN 1113b6–14.²⁰ That is, the biconditional is needed for the context of (B).

5. The Linguistic Context of (B) Taken into Account Hence, next I consider how readings (A) and (I) of the Greek sentence (B) fare when one takes the immediate linguistic context into account.²¹ To start with, note the following three points. First, the sentence before (B) is (L) Now, virtue is up to us, too, and equally also vice. ἐφ’ ἡμῖν δὴ καὶ ἡ ἀρετή, ὁμοίως δὲ καὶ ἡ κακία.²² (EN 1113b6–7) Second, the argument ends with (M)

then it will be up to us to be virtuous people and to be vicious people. ἐφ’ ἡμῖν ἄρα τὸ ἐπιεικέσι καὶ φαύλοις εἶναι. (EN 1113b13–14)

This final clause of the argument states almost the same thing as (L). Third, our sentence (B) immediately follows (L) and begins with ‘for’ (‘γάρ’). It thus provides a reason for (L). Considering also what we know about Aristotle’s dialectic, we get the following set-up of an argument from (L) to (M), or from 1113b6 to 1113b14: Aristotle provides the thesis he intends to prove in (L) at the beginning of his argument. ²⁰ At EN 1113b14 Aristotle moves on to consider a possible objection. ²¹ Here I ignore the debate as to whether (i) the whole paragraph 1113b6–14 is meant to show that vice is up to us, with Aristotle taking it to have been shown already that virtue is up to us (the asymmetry reading); or whether (ii) the whole paragraph is meant to show that both virtue and vice (acting virtuously and acting viciously) are up to us. I believe (ii) is right, and that a good case can be made for this, which I hope to do elsewhere. For the question whether EN 1113b7–8 supports undetermined choice in Aristotle, this debate is only of minor importance. (Cf. also the next footnote.) ²² In all the manuscripts that are considered in the Oxford edition the sentence does not have the particle ‘δή’ (‘hence’, ‘now’), but the particle ‘δέ’. Modern editions of the Nicomachean Ethics tend to give ‘δή’. The difference is not relevant for present purposes. I just mention that Aristotle does use the particle combination ‘δὲ καὶ . . . δέ . . . ’ elsewhere in the Nicomachean Ethics in one sentence. Cf. e.g. ἐπαινοῦμεν δὲ καὶ τὸν σοφὸν κατὰ τὴν ἕξιν· τῶν ἕξεων δὲ τὰς ἐπαινετὰς ἀρετὰς λέγομεν (EN 1103a8–10).

  .,   –   

103

This is standard in Aristotelian dialectic. He then argues for (L) up until and including the antecedent of the sentence 1113b11–14. The consequent of the sentence is (M). It provides the conclusion of the argument. Thus from (B) to (M) (i.e. from 1113b7–8 to b13–14) Aristotle provides an argument for the thesis that (not just actions but also) virtue and vice are up to us. This is so regardless of what exactly the thesis (L) and the conclusion (M) amount to. (B) is a premise in this argument. With the sentence following (B), Aristotle argues towards an intermediate conclusion: it starts with ‘ὥστ’ ’, which, in grammatical contexts such as ours, is translated most naturally as ‘hence’: (N) Hence, (1) if to act, being noble, is up to us, also to not act, being shameful, will be up to us, and (2) if to not act, being noble, is up to us, also to act, being shameful, [will be] up to us. ὥστ’ (1) εἰ τὸ πράττειν καλὸν ὂν ἐφ’ ἡμῖν ἐστί, καὶ τὸ μὴ πράττειν ἐφ’ ἡμῖν ἔσται αἰσχρὸν ὄν, καὶ (2) εἰ τὸ μὴ πράττειν καλὸν ὂν ἐφ’ ἡμῖν, καὶ τὸ πράττειν αἰσχρὸν ὂν ἐφ’ ἡμῖν. (EN 1113b8–11) (I have used bold to indicate complete textual agreement with (B) and underlining to indicate parallels to understood additions in (B) that are generally accepted.) This striking parallel provides a strong reason for reading (B2) as an abbreviation along the lines which I—in agreement with numerous other scholars—have suggested, that is, as short for (B2full)

καὶ ἐν οἷς τὸ μὴ [πράττειν ἐφ’ ἡμῖν ἐστί] καὶ τὸ [πράττειν ἐφ’ ἡμῖν ἐστί]

(J2full)

and where to not [act is up to us], also to [act is up to us].

This reading provides four perfectly matching cases, all in the right order. The only substantive difference is that in (L) an evaluative attribute (καλόν, αἰσχρόν) is added each time.²³ The apparent lack of regularity in (B), from positive/negative in (1) to negative/positive in (2), finds a full explanation in the move from noble to shameful in the two conditionals of (I). Aristotle’s intent is to cover all four possibilities and their interrelations (noble action is paired with shameful inaction; noble inaction with shameful action), and for each interrelation he starts with the noble case. The parallel between (B) and (N) also provides strong reasons for not adding a verb of saying, as readings of type (A) do. First, with (A), the inferential ‘hence’ (‘ὥστ’ ’) is very hard to explain. Second, no mention is made of saying yes or no, or the like, ever again in the argument at issue. But if (A) were correct, we would

²³ For the purpose of this addition at this point in Book 3 of the Nicomachean Ethics, see Chapter 2 (Bobzien 2014b). For a different view, see Sorabji 1980b, 248 n. 14 and 230.

104

, ,   

expect some such mention, given the parallel structure of (N) and (B). We can be more precise: if (A) were correct we would expect the second half of (N) to be something like (O2): (O2full)

and if saying yes, being noble, is up to us, also saying no, being shameful, is up to us.

But we do not have this. Rather, if (A) were correct, (N2) would just hang in the air, so to speak. There is nothing in the previous sentence for it to latch onto. Given that, even without the linguistic context, (A) was somewhat grasping at straws, it appears that the linguistic context, i.e. (N), provides the bale of straw that breaks the camel’s back.

 .      3.5, 11137–8 Next, I turn to the startling phenomenon that, despite its utter implausibility, versions of the saying-no translation have made their way into the general consciousness of what Aristotle himself stated. That is, I now move to the topic of the reception of EN 3.5, 1113b7–8.

6. Reception in ‘Popular Culture’ Let us begin with the reception of translation (A) of our sentence in what may be called popular culture. (A) is equally popular on blogs, Twitter, self-help and alternative-healing websites, in coffee-table books, soap operas, and generally in collections of famous quotations. • In 1940, (A) provided the first lines of the opening narration of an episode of the soap opera ‘The Guiding Light’.²⁴ • It is listed second among the ‘immortal’ Proverbs, Sayings, and Curses at angelfire.com.²⁵ • It made it into the compilation of inspirational quotations for teens at oxygenfactory.com, categorized under ‘yes’.²⁶ ²⁴ [accessed 27 Apr. 2020]. I thank Paul Scade for this reference. ²⁵ [accessed 27 Jan. 2013]. ²⁶ [accessed 27 Jan. 2013]. (While some of the websites cited here have gone offline since the original publication of this article in 2013, our text continues to be used in similar ways.)

  .,   –   

105

• On Willow’s Dreamscapes it is one of the quotations in the category ‘daydreams’ and the subcategory ‘choose freedom’.²⁷ • At morequotations.com, in Life Quotes, Cute Quotes, Funny Quotes we find it as number one under the heading ‘will’.²⁸ • The keynote speech of the ‘Workshop on Clinical Teaching’ in 2009, delivered by the dean of the Faculty of Medicine, UiTM, includes the lines: ‘Sometimes we are overcome by doubts and self-inflicted inertia. Aristotle realized that and he opined, “Where we are free to act, we are also free to refrain from acting, and where we are able to say NO, we are also able to say YES.” ’²⁹ • Elsewhere, we find sentence (A) as ‘A Positive Thought For Saturday, March 15, 2008’.³⁰ • We find it as the quotation accompanying a photo of a charity celebration in Ninna Gay, Shifts: Beyond the Visible (Central Milton Keynes, 2010), 26. • In Ian McTavish’s A Prisoner’s Wisdom, (A) is incorporated to encourage choice that transcends the Ego.³¹ • It was the Tweet of the day on 3 October 2011 at Dance_with_life.³² • And finally, in the world of blogs, Edith Hall, in The Edithorial (Saturday 10 March 2012), uses (A) to gently criticize Queen Elizabeth II: ‘But Aristotle’s response to her record as ruler would immediately have been to point out that avoiding error is not enough to qualify a person as good. “Where we are free to act we are also free to refrain from acting, and where we are able to say ‘No’ we are also able to say ‘Yes’; if therefore we are responsible for doing a thing when to do it is right, we are also responsible for not doing it when not to do it is wrong.” (Nicomachean Ethics 1113b2 (sic)).’³³ In some of these occurrences, the alleged Aristotle quote is used (in an unAristotelian way) to remind us of our free will. In others, it is used (in an equally un-Aristotelian way) as a source for optimism: don’t wallow in your apathy and misery; say ‘yes’ to life. Well, Aristotle is dead, his copyright has run out, or, more accurately, never existed in the first place, and he is not the only philosopher who is used for alien and perplexing purposes in popular culture. The extra twist in our case is that Aristotle never wrote a Greek equivalent of (A) in the first place. But even this is in

²⁷ [accessed 27 Jan. 2013]. ²⁸ [accessed 27 Jan. 2013]. ²⁹ [accessed 10 Jan. 2013]. ³⁰ [accessed 27 Jan. 2013]. ³¹ McTavish 2012, 56. ³² [accessed 27 Jan. 2013]. ³³ [accessed 27 Jan. 2013].

106

, ,   

no way unique.³⁴ So, too much weight should not be put on this part of the reception of EN 3.5, 1113b7–8 from the point of view of veridicality. What we have, however, is evidence of how the same word-shell can be filled with different meaning in different millennia, catering to the varying and changing consumer desires for titbits of wisdom. Given the continuous stream of quotations of the saying-no translation, it may not be an exaggeration, though, to proclaim that the saying-no translation has become a meme.³⁵

7. Reception in Popular Philosophy and non-Ancient Philosophy Things get just a tad more serious when we move to popular philosophy and to professional philosophers who do not specialize in ancient philosophy. A good number of them who are not Aristotle scholars make use of (A) in their publications. I mention here just some examples: Teodros Kiros in Self-Construction and the Formation of Human Values, in the context of discussing human choice;³⁶ Jeremy Naydler in The Future of the Ancient World, attributing the birth of freedom of choice to Aristotle;³⁷ George Stack in ‘Aristotle and Kierkegaard’s Existential Ethics’;³⁸ John A. O’Brian in Truths Men Live By;³⁹ Max Hamburger in Morals and Law, ‘assuming that when we can say “no” we also can say “yes” and that it is therefore in our power to act in harmony with goodness as well as with badness’, with reference to 1113b7–11;⁴⁰ David Buchanan, in An Ethic for Health Promotion, uses (A) as evidence that Aristotle held that human choices are undetermined by prior causes;⁴¹ similarly, Bob Doyle, in his e-book on Free Will, quotes (A) in support of Aristotle being an agent-causal libertarian.⁴² The majority of the authors mentioned use (A) in order to attribute to Aristotle a theory of freedom of choice, uncaused choice, or agent causation, thus perpetuating the myth that Aristotle endorsed undetermined choice between alternatives. And if it was not for this (a false interpretation of Aristotle on a most important philosophical topic, based on a mistranslated sentence from the Nicomachean Ethics), the question of the reception of EN 1113b7–8 might be of little interest. As it is, it seems worthwhile to ask: however did we get to the general acceptance of the saying-no translation? Let us start at the beginning. ³⁴ For instance, many of the anecdotes in Diogenes Laertius’ Lives of the Philosophers are told, often almost verbatim, about different philosophers, sometimes in the same work, sometimes in other ancient authors. Cf. Diogenes Laertius, Vitae Philosophorum. ³⁵ The reader is encouraged to Google ‘where we are able to say no, we are also able to say yes’ (with the quotation marks) for verification. ³⁶ Kiros 1998, 84. ³⁷ Naydler 2009, 197–8. ³⁸ Stack 1974, 1–19. ³⁹ O’Brien 1946, 248. ⁴⁰ Hamburger 1951, 31. ⁴¹ Buchanan 2000, 55. ⁴² [accessed 27 Jan. 2013].

  .,   –   

107

8. Ancient and Byzantine Commentators and Commentaries There appear to be no saying-no translations of (B) in antiquity, or in Byzantine and Latin Medieval texts.⁴³ For antiquity, the only extant ancient commentary on Book 3 of the Nicomachean Ethics is by the early commentator Aspasius (fl. second century ). Aspasius considers and explains only the first clause of 1113b7–8, (B1).⁴⁴ He is silent on (B2). So is the Anonymous commentator of Book 5 (who wrote in the later second century ). I could find nothing of relevance in Alexander of Aphrodisias’ writings, or in any other ancient commentators. (We know that Porphyry wrote a commentary on the Nicomachean Ethics, but it is lost, though it may have been translated into Syriac and Arabic.)⁴⁵ The later ancient or, more likely, Byzantine Anonymous paraphrase (in the past wrongly attributed either to Andronicus of Rhodes or to Heliodorus of Prusa)⁴⁶ butchers EN 1113b7–8 in a strange and unhelpful way (50.8–16). (B1) has an unmotivated ‘τὸ ἀγαθὸν’ inserted, and the ‘γὰρ’ is moved from (B1) to (B2). The result is: (P)

εἰ δὲ τὸ πράττειν τὸ ἀγαθὸν ἐφ’ ἡμῖν, καὶ τὸ μὴ πράττειν ἐφ’ ἡμῖν ἐστιν· ἐν οἷς γὰρ τὸ μή, καὶ τὸ ναί. ([Heliodorus], EN 50.10–11)

This makes little sense, either taken on its own or in its context. Among (other) Byzantine commentators on the Nicomachean Ethics, neither Eustratius of Nicaea nor Michael of Ephesus comment on Book 3.⁴⁷ George Pachymeres’ very brief commentary or paraphrase does not touch on the passage.⁴⁸ The Byzantine Anonymous commentary on Books 2–5 has a short paragraph concerned with EN 1113b6–14 (Anonymi in Eth. Nic., 154.17–32) but there is nothing recognizable in it as paraphrase of, or comment on, 1113b7–8.⁴⁹ ⁴³ For detailed information about the ancient commentaries on the Nicomachean Ethics, see Mercken 1990, 407–10. For detailed information about the Byzantine commentaries on the Nicomachean Ethics see Barber and Jenkins 2009. ⁴⁴ Aspasius, In Ethica Nicomachea quae supersunt commentaria, ed. G. Heylbut (Commentaria in Aristotelem Graeca, 19.1; Berlin, 1889), 76.8–16. ⁴⁵ See Akasoy and Fidora 2005, 23–7 for the evidence. ⁴⁶ [Heliodorus of Prusa], In Ethica Nicomachea paraphrasis, ed. G. Heylbut (Commentaria in Aristotelem Graeca, 19.2; Berlin, 1889). The date of composition is unknown. The terminus ante quem is 1366, the date of the earliest manuscript. Michele Trizio argues that the anonymous author relied on Eustratius of Nicaea’s commentary on the Nicomachean Ethics (Trizio 2012, 199–224). ⁴⁷ Eustratius of Nicaea, Eustratii et Michaelis et anonyma in Ethica Nicomachea commentaria, ed. G. Heylbut (Commentaria in Aristotelem Graeca, 20; Berlin, 1892). Michael of Ephesus, Michaelis Ephesii in librum quantum Ethicorum Nicomacheorum commentarium, ed. M. Hayduck (Commentaria in Aristotelem Graeca, 22.3; Berlin, 1901). ⁴⁸ Georgios Pachymeres, Philosophia, 11. Ethica Nicomachea, ed. K. Oikonomakos (Corpus Philosophorum Medii Aevi, Commentaria in Aristotelem Byzantina, 3; Athens, 2005). ⁴⁹ [Anonymous], In Aristotelis ethica Nicomachea ii–v commentaria, in Eustratii et Michaelis et anonyma in Ethica Nicomachea commentaria, ed. G. Heylbut (Commentaria in Aristotelem Graeca, 20; Berlin, 1892), 122–255.

108

, ,   

9. Medieval Latin Translations and Commentaries (Twelfth and Thirteenth Centuries) In the twelfth century, Burgundio of Pisa (possibly the first translator of the Nicomachean Ethics from Greek into Latin) provided this translation:⁵⁰ (Q)

In quibus enim in nobis operari, et non operari; et in his, utique et non.

In the early thirteenth century, Robert Grosseteste (1175–1253), probably revising Burgundio, translates the sentence even more literally as:⁵¹ (R)

In quibus enim nobis operari, et non operari; et in quibus non, et eciam.

Such very nearly word-for-word translations were standard in Medieval Latin translations of Greek texts. Grosseteste’s translation is retained in the revision usually attributed to William of Moerbeke (c.1215–86), dating from later in the thirteenth century,⁵² and consequently becomes part of the standard Latin translation. (I return to this translation below in section 12.3.) The two earliest medieval commentaries of the Nicomachean Ethics are by Albertus Magnus and Thomas Aquinas. They were both influential for centuries. Albertus Magnus (c.1200–80) provides a clear vice-versa reading of 1113b7–8: (S)

In quibuscumque enim in nobis est operari illa, in illis eisdem in nobis est non operari eadem: et in quibus in nobis est non operari aliqua, in illis eisdem est etiam in nobis operari eadem. (Albertus Magnus, Liber III Ethicorum, tract. I, cap. XXI 28, p. 227 Borgnet)⁵³

Thomas Aquinas (c.1225–74), in his literal commentary, provides an extended exposition of the vice-versa reading of 1113b7–8: ⁵⁰ Burgundius Pisanus translator Aristotelis, Ethica Nicomachea: translatio antiquissima librorum II et III siue ‘Ethica uetus’, ed. R. A. Gauthier (Aristoteles Latinus, 26.1–3, 2nd fasc.; Turnhout, 1972), pp. 5–48, Clavis: 26.1.1 (M), liber: 3, cap.: 6, p. 32.22. Six commentaries on the Nicomachean Ethics were also translated by Burgundio of Pisa, but I have not been able to check these; two remain unpublished. ⁵¹ R. Grosseteste (tr.), Ethica Nicomachea: libri I—III; VIII.1–5 (6) (‘recensio pura’—Burgundii translationis recensio), ed. R. A. Gauthier (Aristoteles Latinus, 26.1–3, 3rd fasc.; Turnhout, 1972), pp. 141–201, 271–305.5, Clavis: 26.2.1 (M), liber: 3, cap.: 7, p. 187.23. This translation was known as the recensio pura. ⁵² William of Moerbeke (trans.), Aristotelis secundum exemplar Parisiacum—Ethica Nicomachea (‘recensio recognita’—Roberti Grosseteste translationis recensio), ed. R. A. Gauthier (Aristoteles Latinus, 26.1–3, 4th fasc.; Turnhout, 1972), pp. 375–588, Clavis: 26.3, liber: 3, cap. 7, p. 418.10. This translation was known as the recensio recognita. ⁵³ Albertus Magnus, Opera omnia, ed. A. Borgnet (Paris, 1891), vol. vii Ethica. ‘And in whatever cases it is in us to do them, in those cases it is also in us not to do them; and in the cases in which it is in us not to do some things, in those cases it is also in us to do them.’

  .,   –   

109

(T) Et dicit quod simili ratione etiam malitia est voluntaria et in nobis existens, quia operationes eius sunt tales. Et hoc sic probat: quia si operari est in potestate nostra, oportet etiam quod non operari sit in potestate nostra. Si enim non operari non esset in potestate nostra, impossibile esset nos non operari: ergo necesse esset nos operari: et sic operari non esset ex nobis, sed ex necessitate. Et similiter dicit quod in quibus rebus non operari est in potestate nostra, consequens est quod etiam operari sit in potestate nostra. Si enim operari non esset in potestate nostra, impossibile esset nos operari. Ergo necesse esset nos non operari: et sic non operari non esset ex nobis, sed ex necessitate. (Aquinas, Sent.Eth. lectio 11 (Sententia Ethic. lib. 3.l.11 n. 2); emphasis mine)⁵⁴ There are quite a few Latin commentaries on the Nicomachean Ethics written between the late thirteenth and the fifteenth century, many of them not available in modern editions, including those by Walter Burley (written 1334),⁵⁵ Albert of Saxony, Gerald of Odo, and John Buridan.⁵⁶

10. Arabic Translations and Commentaries The interest in Aristotle’s Nicomachean Ethics of early Arabic philosophers plays a key role in the reception of EN 1113b7–8, and I turn to it next.

10.1 Arabic Translations In the 1950s the Maghribī manuscript of an Arabic translation of Aristotle’s Nicomachean Ethics was discovered, in two parts, in Fez. The manuscript (‘Fez MS’) is dated to 1222 . Most certainly at least Books 1 to 4 go back to ⁵⁴ Thomas Aquinas, Sententia libri ethicorum, in Opera omnia, ed. R. A. Gauthier, vol. xlvii/1 (Rome, 1969). ‘And he says that for a similar reason vice is also voluntary and exists in us, because one’s actions are such [i.e. voluntary and exist in us]. And he proves this as follows: Since if acting is in our power, not acting ought to also be in our power. For if not acting was not in our power, then it would be impossible for us not to act; hence it would be necessary for us to act; and so acting would not be from us but from necessity. And similarly he says that in the cases in which not acting is in our power, it follows that acting is in our power. For if acting was not in our power, it would be impossible for us to act. Hence it would be necessary for us not to act; and so not acting would not be from us but from necessity.’ ⁵⁵ Walter Burley, Expositio librorum Ethicorum (Venice, 1481; 2nd edn 1500). Burley provides a vice-versa translation and there is no hint of a saying-no understanding in the commentary. Burley is concerned with the relation between what is ἐφ’ ἡμῖν and Aristotle’s notion of two-sided possibility or contingency. Sorabji 1980b, 234 and 228 n. 1, makes a similar point about (B). ⁵⁶ J. Buridan, Quaestiones super decem libros Ethicorum Aristotelis ad Nicomachum (Paris, 1513); repr. as Super decem libros Ethicorum (Frankfurt a. M, 1968). Buridan’s commentary is in questionand-answer format and I have found nothing in it that concerns (B) directly.

110

, ,   

a ninth- or early tenth-century translation made by Ishāq b. Hunain,⁵⁷ :
presumably via a Syriac intermediate.⁵⁸ This manuscript is the oldest surviving text which has a verb of saying added in the translation of (B) (i.e. EN 1113b7–8). A literal rendering from the Arabic of the corresponding sentence into English would be (U)

(1) For the things that are up to us to do, [then] it is up to us not to do; (2) and the things concerning which we say no(t), [then] concerning those we say yes.⁵⁹ (emphasis mine)

One can see that, apart from the two occurrences of ‘we say’, the Arabic is virtually a literal, word-for-word translation of the sentence—as was common for early translations into Syriac, Arabic, and Latin alike. We can illustrate the fact that we have—almost—a word-for-word translation by placing the Greek from (B) in parentheses after the English translation of each of the respective Arabic phrases, with phrases lacking a Greek ancestor put in italics. Thus we obtain in the parentheses a sort of back-translation from Arabic into Greek (with omission of the two verbs of saying): (V)

(1) For (γὰρ) the things that (ἐν οἷς) are up to us (ἐφ’ ἡμῖν) to do (τὸ πράττειν), [then] (καὶ) it is up to us ([ἐφ’ ἡμῖν]) not to do (τὸ μὴ πράττειν); (2) and (καὶ) the things concerning which (ἐν οἷς) we say no(t)⁶⁰ (τὸ μή), [then] (καὶ) concerning those ([ἐν οἷς]) we say yes (τὸ ναί).

In addition to the two occurrences of ‘we say’, in the Arabic translation the ἐφ’ ἡμῖν that is understood in the Greek in (B1) and the implicit reference to ἐν οἷς in the Greek in (B2) are supplied and thus made explicit in the Arabic. The translation is not, as it stands, a saying-no translation. For this, we would in addition need two modal expressions in (V2). These could be taken as understood in the Arabic. Dunlop, in his translation of the Arabic behind (U2), reads the text in this way and makes them explicit:⁶¹ (W)

(1) That is, the things which are in our power to do are in our power not to do, (2) and we may also say No in regard to the things in regard to which we may say Yes.

⁵⁷ Akasoy and Fidora 2005, 1–2, 27–8, 94–5. ⁵⁸ Ibid. 26, 62, 106. ⁵⁹ The Arabic is: ‫ﻭﺫﺍﻟﻚ ﺃﻥ ﺍﻷﺷﻴﺎﺀ ﺍﻟﺘﻲ ﺇﻟﻴﻨﺎ ﺃﻥ ﻧﻔﻌﻠﻬﺎ ﻓﺈﻟﻴﻨﺎ ﺃﻻ ﻧﻔﻌﻠﻬﺎ ﻭﺍﻷﺷﻴﺎﺀ ﺍﻟﺘﻲ ﻓﻴﻬﺎ ﻧﻘﻮﻝ ﻻ ﻓﻴﻬﺎ ﻧﻘﻮﻝ ﻧﻌﻢ‬ (Akasoy and Fidora 2005, 207.1–2). ⁶⁰ The bracketed ‘t’ is explained below. ⁶¹ Akasoy and Fidora 2005, 206. (It seems that all of (2) could be taken as subordinate to ‘it is up to us’ from (1).)

  .,   –   

111

In form at least, (W) qualifies as a version of the saying-no translation, if with a reversal of the order of the ‘No’ and the ‘Yes’ (which is not present in the Arabic). In any event, note the terminological mismatch between the two occurrences of ‘in our power’ and the two occurrences of ‘may’ in (W). The Greek (B) requires that two ἐφ’ ἡμῖν be understood in (B2)—even by those who chose a saying-no rendering. Whether the Arabic translator understood his text in this way, we cannot know. Significantly, the Fez MS also displays a textual difference with regard to the linguistic context of (B), more precisely, with regard to the sentence immediately following (N). Compared with the Greek manuscript tradition, there appears to be a lacuna in the text, so that instead of (N) Hence, (1) if to act, being noble, is up to us, also to not act, being shameful, will be up to us, and (2) if to not act, being noble, is up to us, also to act, being shameful, [will be] up to us (EN 1113b8–11) the Fez MS has (the Arabic equivalent of) (X)

And, if the doing of the noble is up to us, [assumed lacuna] then also the doing of the shameful is up to us.⁶²

The difference between the Fez MS and Bywater’s text is most easily explained by a combination of two factors. First, there is an omission in the Fez MS that is the result of one of the scribes in the history of its transmission inadvertently missing a line and somehow confounding the second clause of (N1) with the very similar second clause of (N2). Once this lacuna was part of the MS tradition, any reader and later scribe will have had difficulties in seeing the point of (B2), since (B) originally served as a preparation for the structurally parallel (N) (see section 5 above). This structural parallel would have been eliminated with the lacuna, and this elimination may have triggered or contributed to a reading with a verb of saying.⁶³ Second, the Arabic word used as the negation particle (lā,‫ )ﻻ‬has a different range of application than does the Greek ‘μή’. It can be used both to express ‘no’ and to express ‘not’. It would have been the word to be used for a literal translation of the ‘μή’ in (B2), if this was correctly understood as an abbreviated second half of a vice-versa translation. But it can also be understood as ‘no’ in parallel to

⁶² The Arabic is:

‫ﻭﺇﻥ ﻛﺎﻥ ﻓﻌﻞ ﺍﻟﺠﻤﻴﻞ ﺇﻟﻴﻨﺎ ﻓﻔﻌﻞ ﺍﻟﻘﺒﻴﺢ ﺃﻳﻀﺎً ﺇﻟﻴﻨﺎ‬

(Akasoy and Fidora 2005, 207.2–3). ⁶³ The lack of parallel from (N) in (X) may have led Dunlop in his translation (W) to reverse the order of ‘saying yes’ and ‘saying no’, thus introducing a parallel between acting and saying yes and not acting and saying no. The saying-no translations lack such a parallel, although we would expect it (see above, section 4).

112

, ,   

‘ναί’ understood as ‘yes’. This fact would have facilitated a reading as ‘no’ rather than ‘not’ in (U2). The editors of the 2005 edition of the Fez MS expressly caution readers that, despite its early age, it is most unlikely that the Fez MS represents what Aristotle wrote where the Arabic text differs from the surviving Greek manuscripts.⁶⁴ Generally by far the likeliest explanation of textual discrepancies is the repeated change of language and script: from Greek to Syriac, Syriac to Arabic, and finally from Oriental Arabic script into Maghribī Script.⁶⁵ The Arabic translation had a discernible direct impact only in the Arabicspeaking world. Ibn Sina (c.980–1037) probably knew it. Ibn Rushd (1126–98) and Moses Maimonides (1135–1204) were definitely familiar with it and used it for their works.⁶⁶ Its impact on the West seems have to been mostly indirect. This brings us to the Arabic commentaries on the Nicomachean Ethics.

10.2 Arabic Commentaries: The Summa Alexandrinorum and Ibn Rushd’s Middle Commentary of the Nicomachean Ethics Al-Fārābī (872–950) is likely to have produced a commentary on the Nicomachean Ethics, but it is lost.⁶⁷ Probably after 1177, Ibn Rushd wrote a commentary on the Nicomachean Ethics. It survives in Latin and Hebrew translations only.⁶⁸ Whether Ibn Rushd had direct access to Porphyry’s lost commentary is unclear. Another potentially relevant text surviving only via the Arabic is the summary or paraphrase of the Nicomachean Ethics known as the Summa Alexandrinorum. The Arabic presumably goes back to the late tenth-century Baghdad translator Ibn Zurʽah. The Greek original may have been composed in the early first century by the Peripatetic Nicolaus of Damascus.⁶⁹ Both Ibn Rushd’s middle commentary and the Summa Alexandrinorum were translated into Latin by Hermannus Alemannus in Toledo, the former in 1240, the latter completed in 1243/4. In the Summa the whole of EN 1113b6–14 is summarized in the following two sentences. (Y)

Et res quas agere in nobis est, non agere eas in nobis est. Si igitur agere actiones pulcras in nobis est, etiam res turpes agere in nobis est.⁷⁰

⁶⁴ Akasoy and Fidora 2005, 103–4. For details about the transmission process and the reliability of the manuscript, see also Schmidt and Ullmann 2012. ⁶⁵ Cf. ibid. 94–5. ⁶⁶ Ibid. 31–55. ⁶⁷ Ibid. 18, 41–2, 45. ⁶⁸ Aristotle/Averroes, Aristotelis opera cum Averrois commentariis [2nd Juntine], vol. III (Venice 1562; repr. Frankfurt a. M., 1963). Ibn Rushd, Averroes’ Middle Commentary on Aristotle’s ‘Nicomachean Ethics’ in the Hebrew Version of Samuel ben Judah, ed. L. Berman (Jerusalem, 1999). ⁶⁹ Akasoy and Fidora 2005, 62–79. Translation possibly via a Syriac intermediate (ibid.). ⁷⁰ Text from the 1904 Marchesi edition: C. Marchesi (ed.), L’etica nicomachea nella tradizione latina medievale, Documenti ed Appunti (Messina, 1904), lii–liii [accessed 27 Jan. 2013].

  .,   –   

113

There is no verb of saying nor any trace of the saying-no translation. This is in line with the aforementioned possibility that the Greek original goes back to the first century . However, Ibn Rushd’s commentary turns out to be important for us. It is a socalled middle commentary. In his middle commentaries Ibn Rushd intersperses sentences from the Aristotelian text with portions of paraphrastic commentary. (The beginning of a comment after a portion of text may be indicated by an intendo or the like, but there is otherwise no explicit distinction between the portions of text and the portions of commentary; in particular, there is rarely the ancient equivalent of quotation marks, i.e. a ‘he says’ before a quote from Aristotle’s text.) If we disregard Ibn Rushd’s comments, the text corresponding to EN 1113b6–14 that he quotes and comments upon is: (Z)

(1) Et hoc quoniam res, quas facere in nobis est: et non facere eas in nobis est: . . . (2) Et in quibus rebus dicimus sic, in his quoque possumus dicere non. (3) Et si fuerit in nobis facere res pulchras, ergo et facere res turpes in nobis est. (4) Cumque fuerit in nobis facere res pulchras, et res turpes, et fuerit in nobis facere eas et non facere.⁷¹

There can be no doubt that Ibn Rushd is here commenting on the Arabic text of the Nicomachean Ethics which we (also) have as the Fez MS.⁷² First, as in the Arabic (U2 above), we have a verb of saying before the Latin ‘sic’ and ‘non’ in (Z2)—which would correspond to the Greek ‘τὸ ναί’ and ‘τὸ μή’. Second, we have, as the subsequent sentence (Z3), almost exactly the sentence that results from the lacuna in the Fez MS (above, (X)). Note, though, two subtle changes. The Arabic manuscript has in (U2) ‘saying no’ before ‘saying yes’. This is in line with Aristotle’s original text (B2), which has ‘μή’ first and ‘ναί’ second. Moreover (U2) contains no explicit modal expression. Ibn Rushd has reversed the order of the positive and negative expressions and added a modal verb (‘possumus’) to the corresponding clause (Z2). This is but the next expected step for someone whose source for the Nicomachean Ethics has the lacuna. Since the parallel in the Greek between (B) and (N) (above, section 5) is lost, some other purpose would need to be found for the second clause of (U). The reversal of ‘non’ and ‘sic’ introduces a parallel between acting and saying yes and

⁷¹ Aristotle/Averroes, 2nd Juntine, 19.18–26. ‘(1) And those things which are in us to do are also in us not to do . . . (2) And in those things in which we say “yes”, in those we can also say “no”. (3) And if it has been/is in us to do good things, it is hence also in us to do bad things. (4) And when it has been in us to do good things and bad things, it has also been in us to do them and not to do [them].’ The ellipsis stands for Ibn Rushd’s comment ‘intendoque rerum, quarum in nobis est potentia faciendi eas, in nobis quoque est potentia non faciendi eas’. ⁷² This is in line with the general view that the copy of the Nicomachean Ethics that Ibn Rushd commented on was the Arabic version that also survived in the Fez MS.

114

, ,   

not acting and saying no.⁷³ This opens up the possibility of the readings of (B) discussed in Part One, sections 1–3. The introduction of the modal expression prevents an apparent inconsistency: ‘and where we say yes, we also say no’ without the ‘possumus’ could be understood as a confession of a penchant for paraconsistency. Unlike Dunlop in his translation (W), Ibn Rushd in his commentary adds only one modal expression. In similarity to (W), there is no direct parallel between ‘in nobis est’ (Z1) and ‘possumus’ (Z2), though the Greek (B) would suggest two understood occurrences of ‘in nobis est’ in (Z2). Hermannus’ translation of Ibn Rushd’s Nicomachean Ethics commentary is dated to 1240. Hermannus also translated the Nicomachean Ethics from Arabic into Latin. Some fragments of this translation have survived.⁷⁴ Given that the Arabic translation and Ibn Rushd’s commentary are our earliest (and before the fourteenth and fifteenth centuries, it seems, only) source of something akin to the saying-no reading of EN 1113b7–8, their transmission in the Western world is of major importance. We skip the details of the sparse evidence for transmission in the thirteenth to mid-fifteenth century and move directly to the Renaissance.

11. Modern Reception of EN 1113b7–8 The invention of the printing press in the mid-fifteenth century soon leads to the wide spread in Europe of both Aristotle’s works and Ibn Rushd’s commentaries. Between 1494 and 1498, the Aldine editio princeps of Aristotle’s works is printed in Venice. In 1497, Jacques Lefèvre d’Étaples (1450–1536) publishes his edition of the Nicomachean Ethics.⁷⁵

11.1 Latin Translations In the same year, 1497, a volume with three different translations of the Nicomachean Ethics is published:⁷⁶ one, the so-called antiqua traductio, which— very nearly—corresponds to Grosseteste’s;⁷⁷ one by Leornardo Bruni (c.1370–1444); and one by John Argyropoulos (1415–87). Bruni’s 1416–17 translation has

⁷³ As it did in Dunlop’s translation of the Fez MS (see above). ⁷⁴ Cf. Akasoy and Fidora 2002, 79–93. ⁷⁵ J. Lefèvre d’Etaples, Decem librorum Moralium Aristotelis tres conuersiones (Paris, 1497). ⁷⁶ Aristotle, Decem librorum Moralium Aristotelis tres conuersiones: Prima Argyropili Byzantij/secunda Leonardi Aretini/tertia vero Antiqua per Capita et numeros conciliate: communi/familiariq[ue] commentario [Jacobi Stapulensis] ad Argyropilum adiecto [Tres conuersiones] (Paris, 1497). [accessed 27 Jan. 2013]. ⁷⁷ ‘In quibus enim in nobis operari et non operari: et in quibus non et etiam’ (i.e. with ‘in nobis’ instead of ‘nobis’, like Burgundio): = Tres conuersiones, pdf p. 271.

  .,   –    (AA)

115

. . . nam in quibus utrum agamus uci non agamus in nobis est: et in nostra sunt potestate.⁷⁸

Here, instead of (B2) we have simply ‘they, too, are in our power’. Whether this is Bruni’s way of concealing a lack of understanding of the Greek is unclear. Argyropoulos’ 1450s rendering is closer to our Greek text. (BB)

Quibus enim in rebus nostra in potestate situm est agere in iis et non agere. Et in quibus non agere: in iis est et agere.⁷⁹

In agreement with Albertus Magnus’ and Aquinas’ commentaries, this is an explicit vice-versa translation and thus without verbs of saying. So, none of the three translations in the volume has a verb of saying. Bruni’s and Argyropoulos’ translations are the basis for many of the vast number of fifteenth- and sixteenth-century commentaries.⁸⁰ The sixteenth century sees a number of new Latin translations of the Nicomachean Ethics. Two influential sixteenth-century works also have vice-versa translations. In 1540/2, in Paris, Ioachim Perion translates: (CC)

Quas enim res in nobis situm est, ut agamus, eas etiam in nobis situm est, ut ne agamus. Quasque res in nobis situm est ut non agamus, easdem ut agamus in nostra est potestate.⁸¹

And in 1558, the French philologist Denys Lambin’s widely used Latin translation of the Nicomachean Ethics was published in Venice, and shortly after in Paris. Lambin, too, provides a fully fledged vice-versa translation of 1113b7–8: (DD) Quas enim res agere in nobis situm est, easdem non agere possumus: et quas non agere in nobis situm est, earundem quoque agendarum potestas nostra est.⁸² A comparison of Argyopoulos’, Perion’s, and Lambin’s translations with the medieval ones reveals that the word-for-word method has been superseded by exegetical translations. In 1566 the Swiss humanist scholar Theodor Zwinger

⁷⁸ = Tres conuersiones, pdf p. 365. The next sentence corresponds to (N). ⁷⁹ = Tres conuersiones, pdf p. 53. ⁸⁰ Cf. e.g. Bejczy 2008. Bejczy’s introduction gives a useful overview. ⁸¹ I. Perion (tr.), Aristoteles, De Moribvs qu[a]e Ethica nominantur, ad Nicomachu[m] filium, Libri decem (Paris, 1540; repr. Basel, 1542). ⁸² D. Lambin (tr.), Aristoteles, In libros De moribus ad Nicomachum annotiationes (Venice, 1558).

116

, ,   

chose Lambin’s translation to be printed aside his Greek edition.⁸³ In 1716 William Wilkinson of the Queen’s College, Oxford, does the same in his Aristotelis Ethicorum Nicomacheorum libri decem—except for some small modifications of the translation, but none in the sentence at issue.⁸⁴ Wilkinson’s book is reprinted in 1803, 1809, and 1818 and remains for over a hundred years the text with which the Nicomachean Ethics is taught at Oxford. In 1828 the work is replaced by Edward Cardwell’s Greek-only textbook.

11.2 Latin Commentaries and Paraphrases During the sixteenth and seventeenth centuries, thousands of commentaries on Aristotle’s works are written and published, including many on the Nicomachean Ethics.⁸⁵ In order to find an explanation for the origin of saying-no translations (of the Aldine and later the Bekker text, i.e. of the Western manuscript tradition) we need to follow the path of Ibn Rushd’s Nicomachean Ethics commentary in the West and in print. Ibn Rushd was held in high esteem in the Renaissance, and the influence of his works is multiply attested.⁸⁶ It appears that the first printed Latin version of Ibn Rushd’s commentary on the Nicomachean Ethics stems from a 1483 volume that also contains a Latin translation of the text that was printed before the commentary. The translation is identical to the Grosseteste/Moerbeke standard translation (version (R) above).⁸⁷ Since in Latin the word for ‘not’ (non) also means ‘no’ and the relevant word for ‘too/doch’ (etiam) also means ‘yes’, the Grosseteste/Moerbeke translation is theoretically compatible with a saying-no reading, with a verb of saying understood. Accordingly, in principle, the text of the Nicomachean Ethics and commentary on 1113b7–8 in this 1483 volume would be compatible. Things are different in a later printing of Ibn Rushd’s commentary, the Venice 1562 second Juntine edition of Aristotle, vol. 3.⁸⁸ The title page proclaims that the volume contains all of Aristotle’s ethical works together with Ibn Rushd’s

⁸³ D. Lambin (tr.), Aristoteles, Ethicorum Nicomachiorum libri decem, ed. T. Zwinger (Basel, 1566). Our sentence is at p. 144. ⁸⁴ D. Lambin (tr.), Aristoteles, Ethicorum Nicomacheorum libri decem, ed. W. Wilkinson (Oxford, 1716). Our sentence is at p. 99. ⁸⁵ See e.g. D. F. Lines, Aristotle’s Ethics in the Italian Renaissance (ca 1300–1650) (Leiden, 2002). ⁸⁶ See e.g. A. A. Akasoy and G. Giglioni (eds), Renaissance Averroism and Its Aftermath: Arabic Philosophy in Early Modern Europe (New York: Springer, 2013). ⁸⁷ N. Vernia (ed.), Aristoteles, Opera latina cum commentariis Averrois (Venice, 1483): ‘In quibus enim nobis operari et non operari et in quibus non et etiam’ (Book 3, ch. 8, third sentence). The text continues: ‘qua re si operari bonum existens in nobis est: et non operari in nobis erit malum existens. et si non operari bonum existens in nobis.’ This is the end of the page. The text lacks the consequent of (N2), probably a printer’s error. In any case it is not a translation of the Arabic version that Ibn Rushd used. ⁸⁸ Aristotle/Averroes, 2nd Juntine. The text of the relevant passages in the first edition from 1550–2 is identical.

  .,   –   

117

commentary on the Nicomachean Ethics and his paraphrase of Plato’s Republic. In this book, the commentary is not appended to the translation as a whole. Rather, the relevant bits are added after each chapter. The translation is not the antiqua, but a recent new translation by the humanist and translator Johannes Bernardus Felicianus. He renders 1113b7–8 as (EE)

In quibus enim in nostra potestate situm est agere, situm est et non agere. Et in quibus non agere, simili modo et agere.

This is an explicit vice-versa translation, as appears to be standard in the sixteenth century. Three pages later, following the end of Nichomachean Ethics 3.5, the Latin of Ibn Rushd’s commentary on the same passage, containing citations from the Arabic translation within his commentary, is added (text (Z) above). In this edition we thus have the—ill-fitting—combination of a vice-versa translation with a commentary which provides a reverse saying-no paraphrastic rendering. This is most significant: a critical reader who compares text and commentary is bound to be somewhat puzzled. A reader who wishes to retain the authority of both Aristotle’s text and Ibn Rushd’s commentary is bound to take some action to make the two match. It can only have been a matter of time until some such reader materialized. It is not my intent, nor within my expertise, to provide a comprehensive study of how EN 1113b7–8 was handled in the vast number of Latin translations, paraphrases, summaries of, and commentaries on the Nicomachean Ethics produced between the end of the fifteenth and the mid-seventeenth century. Rather, I have picked out some texts that appear to be paradigmatic of what happened, which is the following. The Greek text of the Western tradition becomes the standard text, via the Aldine edition. At the same time, Ibn Rushd is recognized as an authority on Aristotle and the Latin translation of his commentary on the Nicomachean Ethics (including fragments of text from the Arabic translation of the work that he used, and, in particular, the lacunose 1113b8–11) becomes a standard commentary. The Western Greek text and Ibn Rushd’s commentary are repeatedly printed together in one volume, first separately, later with parts of the latter inserted after chapters of the former. Latin summaries, explications, paraphrases, and commentaries rely on Ibn Rushd’s commentary (without their authors necessarily looking at the Greek text for comparison). By 1578, it seems, someone has compared Ibn Rushd’s commentary with the Aldine, and someone has attempted to make the two match. The evidence is an explicatio accuratissima published that year, which offers a hybrid paraphrasetranslation of 1113b7–8, combining the vice-versa translation of the Aldine with the Arabic-origin Latin translation in a new, specific way. The author of the summaryparaphrase with textual notes of the Nicomachean Ethics was the German classical scholar Joachim Camerarius (1500–74). The text was posthumously published in

118

, ,   

Frankfurt.⁸⁹ It gained a wide readership and was still used in the nineteenth century.⁹⁰ There are no textual notes on 1113b7–8, but the following paraphrase is offered: (FF)

Ubi enim penes nos est ut aliquid agamus, ibi est etiam ut non agamus. Et ubi affirmandi potestas est, ibi et negandi est.

This sentence was not proposed as a translation of the Greek (Aldine) text. Nor is it a saying-no translation. Rather, it manifests the parallel of ‘being able to do’ and ‘saying yes’ with ‘being able not to do’ and ‘saying no’ as we find it in Ibn Rushd (the reverse saying-no reading). What is noteworthy is that we have a structural parallel of (B1) and (B2) (‘ubi’, ‘ibi’) and particles that correspond to the Greek of the Aldine (‘enim’ in (B1), ‘et’ in (B2)), combined with Latin verbs for ‘saying yes’ and ‘saying no’ and an explicit mention of a ‘potestas’ in (B2), which parallels the ‘penes’ of (B1). One can see how it may only be a matter of time before, first, based on hybrid explications like these, a hybrid translation, based on two different texts, may see the light of day; and before, second, this hybrid translation is offered as an actual translation of the Greek Aldine (or, later, Bekker) edition. The saying-no translation is just such a hybrid translation. It seems that for actual instances of the saying-no translation of (B) we have to move into nineteenth-century Germany and Britain and translations into modern languages.

12. Commentaries and Modern-Language Translations in the Nineteenth and Early Twentieth Centuries The earliest translation of the Nicomachean Ethics into English, by John Gillies in 1789,⁹¹ conveniently leaves out the entire sentence 1113b7–8—or perhaps we should say paraphrases over it. The relevant passage reads (at p. 304): (GG)

Ends are then the objects of volition; and the means of attaining them are the objects of deliberation and preference; which, being conversant only about such things as are in our power, the virtues immediately proceeding from them must also be in our own power, and voluntary, as well as the contrary vices.

The Spectator bemoaned the lack of literalness of this translation (vol. 43 (1870), 179). The earliest German translation, by Christian Garve (1742–98) at the end of

⁸⁹ I. Camerarius, Ethicorum Aristotelis Nicomachiorum explicatio accuratissima (Frankfurt, 1578). ⁹⁰ Yale has the copy of the American nineteenth-century scholar Thomas Day Seymour. In Oxford, St John’s, Keble, and Trinity have copies. ⁹¹ J. Gillies (tr.), Aristotle’s ‘Ethics’ and ‘Politics’, Comprising his Practical Philosophy (London, 1789).

  .,   –   

119

the eighteenth century,⁹² is more literal and provides a traditional vice-versa translation of EN 1113b7–8: (HH)

Denn allenthalben, wo das Handeln in unsrer Gewalt ist, da ist auch das Nichthandeln in unsrer Gewalt: und wenn das Unterlassen von uns abhaengt, so haengt auch das Thun von uns ab.

And come the nineteenth century, EN 1113b7–8 is always translated. The nineteenth century also sees a further milestone in Aristotle scholarship: the publication of the Bekker edition (1831–70),⁹³ which soon becomes the standard edition used by scholars worldwide (including Oxford). The first Bekker edition of the Nicomachean Ethics appears in 1831. The passage 1113b7–8 is identical with (B). There are no comments on the sentence.⁹⁴

12.1 Vice-Versa Translations Between 1818 and 1925, I count at least eight English translations that provide variations of the vice-versa reading.⁹⁵ These start in 1818 with Thomas Taylor,⁹⁶ followed in 1828 by an Anonymous,⁹⁷ and then the frequently reprinted 1861 Chase translation.⁹⁸ Chase, it seems, is the first to actually use the phrase ‘vice versa’: (II)

Wherever it is in our power to do, it is also in our power to forbear doing, and vice versa.

Then in 1879 comes Walter Hatch, who pairs the following vice-versa translation

⁹² V. C. Garve (tr. and comm.), Die Ethik des Aristoteles, 2 vols (Breslau, 1798–1801) (partly posthumous edition). ⁹³ I. Bekker (ed.), Aristotelis Opera (Berlin, 1831). ⁹⁴ K. L. Michelet’s Ethicorum Nicomacheorum libri decem. Ad codicum manuscriptorum et veterum editionum fidem recensuit commentariis illustravit in usum scholarum suarum, 2 vols (Berlin, 1829; 2nd edn 1835), also has (B). In the preface (vol. i, p. vi) Michelet acknowledges that he used Ibn Rushd’s commentary. But it left no trace with respect to 1113b7–8 (ibid. 1.50; 2.156). ⁹⁵ There are also more Latin vice-versa translations, e.g. in Karl Zell’s Graeco-Latin edition of the Nicomachean Ethics: K. Zell (tr.), Aristotelis Ethica Nicomachea (Heidelberg, 1820). ⁹⁶ T. Taylor, The ‘Rhetoric’, ‘Poetic’, and ‘Nicomachean Ethics’ of Aristotle, 2 vols (London, 1818), 2.90: ‘for in those things in which to act is in our power, not to act is also in our power; and in those things in which we have the power not to act, we have likewise the power to act’. ⁹⁷ [‘Vincent’], A New Translation of the ‘Nicomachean Ethics’ of Aristotle, 2nd edn (Oxford, 1826), 82: ‘for whenever we have the power to do, we have also the power not to do; and where we have the power not to do, we have also the power to do’. ⁹⁸ D. P. Chase, The ‘Nicomachean Ethics’ of Aristotle: A New Translation, Mainly from the Text of Bekker (Oxford, 1861). There were at least twenty-one editions published between 1911 and 2010 in the Everyman’s Library (London and New York).

120

, ,   

(JJ)

Where the power of action depends upon our own selves, in such cases there is also the power of forbearing; and where there is a power of forbearing, there is also a power of acting.

with an explanatory saying-no comment on the same page: (KK) But if the doing of good be within our own power, the refraining from good will be within our power, since where there is ‘nay’ there is also ‘yea’.⁹⁹ This seems right in the tradition of the Latin 1562 Juntine edition (see above, section 11.2), a vestige of, or late witness to, the juxtaposition of a vice-versa translation, on the one hand, with a reference to ‘yes’ and ‘no’ as we have it in Ibn Rushd’s commentary, on the other. In fact, the commentary printed along with Hatch’s translation, of which (KK) is a part, is nothing but an English translation of the anonymous paraphrase of the Nicomachean Ethics (see above, section 8), here attributed to Andronicus of Rhodes. Curiously, the Greek relating to EN 1113b7–8 (above (P)) is mistranslated in (KK) in a manner that partially parallels the misrendering of (B) in (A)—notwithstanding the vice-versa translation (JJ). There follow with vice-versa translations R. W. Browne in 1889,¹⁰⁰ J. E. C. Welldon in 1892,¹⁰¹ Edward Moore in 1902,¹⁰² and, finally, in 1925, Ross with the translation which (in the revised Ackrill/Urmson version) is still one of the most used today:¹⁰³ (LL)

For where it is in our power to act it is also in our power not to act, and vice versa.

12.2 Nineteenth-Century Commentaries on the Nicomachean Ethics These translations are complemented by at least three of the standard nineteenthcentury Latin or English commentaries which judge (B2) to be unremarkable and ⁹⁹ W. M. Hatch, The Moral Philosophy of Aristotle (London, 1879), at 144. ¹⁰⁰ R. W. Browne, The ‘Nicomachean Ethics’ of Aristotle (London, 1889), 66: ‘for wherever we have the power to do, we have also the power not to do; and wherever we have the power not to do, we have also the power to do’. ¹⁰¹ J. E. C. Welldon, The ‘Nicomachean Ethics’ of Aristotle (London, 1892), 73: ‘for where it is in our power to act, it is also in our power to refrain from acting, and where it is in our power to refrain from acting, it is also in our power to act’. By 1912, there are five reprints. ¹⁰² E. Moore, An Introduction to Aristotle’s Ethics (New York, 1902) (chs 1–4, parts of ch. 10), 138: ‘If it is in our own power to act, it must also be in our own power not to act (else our action was not really in our power but was compulsory), and vice versa.’ ¹⁰³ D. Ross, Aristotle: The ‘Nicomachean Ethics’ (Oxford, 1925). Reprinted in Barnes 1984, vol. 2, 1758.

  .,   –   

121

do not comment on it at all: Ramsauer/Susemihl, Grant, and Stewart. On 1113b7–8 Ramsauer and Susemihl are silent; Grant notes ‘elsewhere (Met. VIII ii 2) Aristotle states in more philosophical form this first step in the doctrine of free-will, namely that every psychical δύναμι& is a capacity of contraries’; and Stewart writes: ‘Grant refers to Met.Θ 2 where αἱ μετὰ λόγου δυνάμεις are said to be τῶν ἐναντίων. Cf. Met.Θ 5 where ὄρεξις or προαίρεσις is said to be τὸ κύριον, and to determine which of the two possible ἐναντία shall be selected.’¹⁰⁴

12.3 Faux-Literal Translations In the nineteenth century there seems also to originate a different translation, and the earliest (apparently) which translates the ‘μή’—incorrectly—with a word like ‘no’ rather than ‘not’. The first perpetrators seem to be German. Rieckher, in 1856, translates (B2) as (MM)

Denn wo das Thun in unserer Gewalt ist, da ist es auch das Unterlassen, und wo das Nein, da ist es auch das Ja.¹⁰⁵

The influential 1911 translation by Rolfes is almost identical (!), but even shorter: (NN)

Denn wo das Tun in unserer Gewalt ist, da ist es auch das Unterlassen, und wo das Nein, da auch das Ja.¹⁰⁶

Both versions keep the parallel of the article ‘to’ for (B1) and (B2). I call this family of translations ‘faux-literal’, since they appear to be word-for-word but translate ‘μή’ with ‘no’. Perhaps it is no coincidence that faux-literal translations first appear in German works. Conceivably, it was the result of the above-mentioned ambiguity in the Latin (section 11.2). In the mid-nineteenth century a Latin version of the Nicomachean Ethics was likely used as a guide in translating the Greek. Now, as stated above, in Latin, the word for ‘not’ (‘non’) also means ‘no’; and in Latin, (as in Greek) the relevant word for ‘too/doch’ (‘etiam’) also means ‘yes’.¹⁰⁷ Thus the Grosseteste/ Moerbeke literal Latin for (B2), ‘et in quibus non, et eciam’ (above, (R)), can be understood in at least two ways, each of which treats the ‘non’ and the ‘etiam’ as ¹⁰⁴ G. Ramsauer and F. Susemihl, Aristotelis Ethica Nicomachea (Leipzig, 1878), 165 and 733; A. Grant, The ‘Ethics’ of Aristotle (London, 1866), vol. 2, 26; J. A. Stewart, Notes on the ‘Nicomachean Ethics’ of Aristotle (Oxford, 1892), 274. ¹⁰⁵ J. Rieckher (tr.), Nikomachische Ethik (Stuttgart, 1856), 77: ‘For where the doing is in our power, there the refraining is, too (i.e. in our power), and where the No, there the Yes is too (i.e. in our power).’ ¹⁰⁶ Rolfes and Bien 1985, 55: ‘For where the doing is in our power, there the refraining is, too (i.e. in our power), and where the No, there the Yes, too.’ ¹⁰⁷ See above, section 4, for the Greek and the German.

122

, ,   

grammatically parallel expressions. There is, on the one hand, the pair ‘nicht (zu handeln)’ and ‘doch (zu handeln)’ and, on the other, the pair ‘das Nein’ and ‘das Ja’. In German and English (unlike in Romance languages)¹⁰⁸ there is no straightforward way of covering both readings in one translation. The translator has to make a choice and may need some help in making this choice. At this point, commentaries and anything else that aids the understanding of a Greek passage—such as textual parallels—may become important. Linguistic context and argument structure were rarely among the aids used in early Victorian times. Thus, if a commentary based on Ibn Rushd’s was consulted on the lines, this may well have been decisive and have pushed a translator towards the second (faux-literal) option. Another observation that may have aided translators towards a faux-literal translation could have been a perceived similarity of EN 1113b7–8 to two passages in the New Testament, each of which contrasts ‘τὸ ναί’ and ‘τὸ οὔ’. These are Paul’s Second Epistle to the Corinthians 1:17 (ἵνα ᾖ παρ’ ἐμοὶ τὸ ναὶ ναὶ καὶ τὸ οὒ οὔ;) and Epistle of James 5:12 (ἤτω δὲ ὑμῶν τὸ ναὶ ναὶ καὶ τὸ οὒ οὔ, ἵνα μὴ ὑπὸ κρίσιν πέσητε).¹⁰⁹ Each time the context suggests reference to a spoken ‘yes’ and ‘no’, indicated by the neuter definite article ‘τό’. Christian scholars would have been familiar with these passages and would thus have had a precedent (if with ‘τὸ οὔ’ rather than ‘τὸ μὴ’), or at least a conceptual template, for the reading of ‘no’ and ‘yes’ in EN 1113b7/8, understood as something that is said.¹¹⁰ Be this as it may, quite generally, if someone is faced with a faux-literal translation, the easiest manoeuvre for them to fill it with sense is by assuming an implicit verb of saying. The saying-no translation simply makes this explicit, and Ibn Rushd’s commentary, or a derivative commentary, would have sanctioned it.

12.4 Saying-No Translations and Paraphrases (by Date) Finally, to the saying-no translations themselves. The first I found (and there may well be earlier ones) is a translation into German from 1863 by the German writer, literary historian, and philologist Adolf Stahr: (OO)

(1) Denn in allen Bereichen, wo das handeln bei uns steht, steht auch das nicht handeln bei uns, (2) und wo wir Nein sagen koennen, koennen wir auch Ja sagen.¹¹¹

¹⁰⁸ Some renderings into Romance language simply retain the ambiguity. ¹⁰⁹ In the New King James Version ‘that with me there should be yea yea, and nay nay?’ and ‘But let your yea be yea; and your nay nay; lest you fall into judgment.’ ¹¹⁰ The two biblical passages are referenced in Patristic and Byzantine Greek texts alone for over a hundred times (as a TLG search will confirm). ¹¹¹ ‘(1) For in all areas where the acting is with us, the not acting also is with us, (2) and where we can say No, we can also say Yes’ (A. Stahr (tr. and comm.), Aristoteles’ nikomachische Ethik (Stuttgart, 1863), 86). This German translation seems to be part of the first complete set of translations of Aristotle

  .,   –   

123

Note that the contrast between ‘bei uns steht’ in (1) and ‘koennen’ in (2) corresponds almost literally to Ibn Rushd’s ‘in nobis’ in the clause corresponding to (1) and ‘possumus’ in the clause corresponding to (2). This is probably not a coincidence. The next saying-no translation is from 1869, by Robert Williams:¹¹² (PP)

(1) For, where it is in our power to do a thing, it is equally in our power to abstain from doing it; (2) where refusal is in our power, assent is equally so.

This text was widely spread and had at least four editions. It was praised in The Spectator.¹¹³ In 1881 follows the much-lauded translation by Peters,¹¹⁴ who translates: (QQ)

(1) For where it lies with us to do, it lies with us not to do. (2) Where we can say no, we can say yes.

Again, we have the telling contrast between ‘lies with us’ in (1) and ‘can’ in (2) that evokes Ibn Rushd. This translation sees its fifth edition in 1893. In 1897, George Stock publishes a book in which he rewrites the Nicomachean Ethics as a dialogue, to make it palatable to the English readership. Here is a paraphrase of our passage (roughly 1113b6–14) where Stock has Aristotle speak: (RR)

But when I speak of the voluntariness of virtue or vice, you must understand me to mean that the virtuous or vicious man is a free agent, that there is no force acting upon him except what comes from his own nature, except, in fact, himself. If he knows the right and the wrong, it is as open to him to choose the one as the other. Where he can do, he can refrain from doing, and where he can say ‘no’, he can say ‘yes’.¹¹⁵

In 1900 Burnet, in his famous commentary, comments on our lines: ‘This is because our capacities are μετά λόγου, and every λόγος implies both “yes” and “no” ’.¹¹⁶ Thus Burnet reads ‘μή’ as ‘no’ rather than ‘not’. He adds ‘Cf. above into German. There are no comments on EN 1113b7–8, and, throughout, there are frequent references to the translation by Garve, who, as we saw above, provided a vice-versa translation, and very few references to other works. ¹¹² R. Williams (tr.), The ‘Nicomachean Ethics’ of Aristotle (London 1869, 1876², 1879³), at 62. There are no comments or references other than to Bekker 1861. ¹¹³ Review of Williams’ translation in The Spectator 43 (1870), 178–80, at 178–9. ¹¹⁴ F. H. Peters (tr.), The ‘Nicomachean Ethics’ of Aristotle (London, 1881), at 74. (There are no notes or comments on this sentence, only general acknowledgement of prior translations and commentators, without any names, as well as of use of Bekker, and in the fifth edition (1893) of Bywater.) ¹¹⁵ S. G. Stock, Lectures in the Lyceum; or: Aristotle’s ‘Ethics’ for English Readers (London, 1897), at 179. ¹¹⁶ J. Burnet (comm.), The ‘Ethics’ of Aristotle (London, 1900), 134.

124

, ,   

1103a20sqq’. But 1103a20ff does not provide any useful information in support of his statement that ‘every λόγος implies both “yes” and “no” ’. Nor does the Metaphysics passage 1046a36ff. (i.e. Arist. Met.Θ 2, which Grant and Stewart had referred to as evidence for Aristotle’s theory of free will), which Burnet adduces in support for his saying-no reading. In 1920, Samuel Sidney McClure summarizes the passage from Nicomachean Ethics 3.5 thus:¹¹⁷ (SS)

Choice is not the same thing as a voluntary act; nor is it desire, or emotion, or exactly ‘wish,’ since we may wish for, but cannot make choice of, the unattainable. Nor is it deliberation—rather, it is the act of decision following deliberation. If man has the power to say yes, he has equally the power to say no, and is master of his own action.

The ‘yes’ and ‘no’ in the last sentence can only be in reference to 1113b7–8. As in Ibn Rushd’s and Camerarius’ commentaries, though, the order of the negative and the positive are the reverse of that in 1113b8. Next in line seems to be Rackham’s widely used and much-quoted 1926 translation,¹¹⁸ which is nothing but our sentence (A), the sentence that has become the Aristotle-meme: (1) For where we are free to act we are also free to refrain from acting, (2) and where we are able to say No we are also able to say Yes.

In this translation, too, there is a contrast between being free in (1) and being able to in (2), which matches Ibn Rushd’s ‘in nobis’ and ‘possumus’. None of the four translations, nor any of the paraphrases, nor Burnet’s commentary mentions earlier translations or commentaries which adopt the saying-no reading. What are we to make of this situation in the reception of such a critical text? A couple of things come to mind. First and foremost, the hybrid translation which, above in section 11, we anticipated as a natural development of Camerarius’ hybrid explanation of the seeming incompatibility of the Greek text of the Western tradition with Ibn Rushd’s commentary has come into being in multiple forms. Second, it was standard in the nineteenth-century translations, summaries, paraphrases, analyses, etc. to provide Book 3 chapter 5 of the Nicomachean Ethics with titles such as ‘the freedom of man’, ‘freedom of the will’, and to assume that it contained Aristotle’s theory of free will. (Rackham still translates ‘ἐφ’ ἡμῖν’ with ‘free’.) And there is throughout in Victorian times a Christian background to ¹¹⁷ S. S. McClure, The World’s Greatest Books, xiii, Philosophy and Religion (New York, 1920). This is identical to vol. 1 of J. A. Hammerton (ed.), Outline of Great Books (New York, 1936). ¹¹⁸ H. Rackham (tr.), Aristotle: The Nicomachean Ethics (Cambridge, 1926).

  .,   –   

125

the translator’s comments. This Christian background assumes that human beings have free, undetermined choice: the ability to choose between good and bad. This idea goes back at least to the second century  in early Christian texts.¹¹⁹ And in the nineteenth century Aristotle was read with this conception of a free will and a choice between good and bad in mind. Thus nineteenth-century translators would have had little difficulty in slightly twisting the text to fit this idea—not in order to make a point of Aristotle exegesis, but simply because they believed anyhow that Aristotle advocated freedom of choice in 3.5. Aristotle’s argument would have been assumed to work from acting and not acting being up to us, via choosing not to act and choosing to act being up to us, to our acting or not acting nobly or shamefully being up to us; free choice would have provided the point where the goodness or badness of the person latches onto their actions. Third, it is also worth remembering that in Victorian times the strictures on translations were much more relaxed. The belief that the translator was to present what the author meant, if there are difficulties in translating the text as it stands, was widely accepted (although towards the end of the nineteenth century requirements became stricter). In addition, translators were not in the habit of considering the linguistic context and the logical structure of the argument as means for making sense of a piece of text. So they would have been unlikely to have considered sentence (N) for making sense of (B).¹²⁰ Thus, we can understand how the misrendering (A) made its way into translations of the Victorian era. It is the result of the confounding of the Greek text of the Western tradition with the Latin translation of Ibn Rushd’s middle commentary, which itself is, most probably, the result of a lacuna, and subsequent reinterpretation of (B2). This reinterpretation is in line with the late ancient and Christian tradition of reading Aristotle as advocating a free will and free choice between good and evil.

13. Contemporary Reception in Ancient Philosophy To conclude, a brief overview of contemporary reception, which I count as, somewhat arbitrarily, the period from 1950 onwards. I have looked at a good part of the countless contemporary English translations, commentaries, introductions, and companions of or to the Nicomachean Ethics. We find still several renowned commentaries that do not comment on 1113b7–8 at all—presumably because the commentators consider their reading of the text to be straightforward and unproblematic. Examples are Joachim in 1951, Dirlmeier in 1956, Gauthier and Jolif in 1970.¹²¹ There are vice-versa ¹¹⁹ Cf. Bobzien 1998, 172–4. ¹²⁰ See section 5 above for details. ¹²¹ Joachim and Rees 1951; Dirlmeier 1956 [2003]; Gauthier and Jolif 1970.

126

, ,   

translations in, for example, Gauthier and Jolif, Wardman in 1963, and in Apostle and Gerson in 1983,¹²² as well as in some scholarly books and articles (in all cases without comment on the choice of translation).¹²³ Regarding the faux-literal translations, we can observe an increase in these, including Dirlmeier, Gigon, Kenny, Irwin, and Sauve Meyer.¹²⁴ In recent English-language translations, saying-no renderings appear to outnumber their alternatives. We find them, for example, in the translations of Martin Oswald in 1962 (‘For where it is in our power to act, it is also in our power not to act, and where we can say “no”, we can also say “yes” ’); Roger Crisp in 2000 (‘Where it is in our power to act, it is also in our power not to act, and where saying “No” is in our power, so is saying “Yes” ’); Christopher Rowe in 2002 (‘For when acting depends on us, not acting does so too, and when saying no does so, saying yes does too’); and Christopher Taylor in 2006 (‘where acting is up to us, not acting is up to us, too, and where one can say No, one can also say Yes’).¹²⁵ They are also present in many scholarly commentaries, books, and articles, such as Everson (1990), Broadie (1991 and 2002), Sparshott (1994), Rapp (1995), Pakaluk (2005), Taylor (2006), and Destrée (2011).¹²⁶ None of these scholars and philosophers mentions that the Greek does not include a verb of saying, or that the Greek includes no equivalent to ‘no’. In fact, none of the translations and commentaries I have been able to consult, regardless of the rendering of EN 1113b7–8 offered or used, mentions that the Greek could be translated differently.

¹²² Gauthier and Jolif 1970, vol. 1, 68: ‘en effet, là où il est en notre pouvoir d’agir, il est aussi en notre pouvoir de ne pas agir (et réciproquement, là où le non est en notre pouvoir, le oui l’est aussi)’ (in French non + adj. translates ‘not’; cf. also si non for ‘if not’; hence ‘le non’ for ‘the not’ is possible). Wardman 1963, 359: ‘The point is that where we can act, we can also refrain, and vice versa.’ Apostle and Gerson 1983, 459: ‘For where it is in our power to act, it is also in our power not to act, and where it is in our power not to act, it is also in our power to act.’ ¹²³ Such as Hardie 1968 [1980], 178: ‘where it is in our power to act it is also in our power not to act, and vice versa’ (1113b7–8; cf. 1115a2–3)’. ¹²⁴ Dirlmeier 2003, 66: ‘denn überall wo es in unserer Macht steht zu handeln, da seht es auch in unserer Macht, nicht zu handeln, und wo das Nein, da auch das Ja’; Gigon 1967; Kenny 1979, 7–8: ‘we are told that where it is in our power to do something, it also in our power not to do it, and when the “no” is in our power, the “yes” is also (1113b7–8)’; Irwin 1985, 66: ‘for when acting is up to us, so is not acting, and when No is up to us, so is Yes.’; Meyer 1993, 130 and 2006, 152: ‘For in those cases in which it is up to us to do something it is also up to us not to do it, and in cases in which no is up to us, so is yes.’ ¹²⁵ Ostwald 1962, 65; Crisp 2000, 45; Broadie 2002, 130; Taylor 2006, 24. ¹²⁶ Everson 1990, 81–103 at 90 (repr. Gerson 1999, 252–76); Broadie 1991, 153–4, 156, 159 with n. 31, and Broadie 2002, 9–91 at 40; Sparshott 1994, 130, 134 with n. 82; Rapp 1995, 131; Pakaluk 2005, 145: ‘In cases in which it’s up to us to carry out an action, it’s also up to us to refrain from carrying it out; and in cases in which saying “no” is up to us, saying “yes” is also up to us’; Taylor 2006, 164; Destrée 2011, 289. Cf. also the advertising and blurb for Bernard William’s short (posthumously published) book A History of Freedom (2010): ‘One of the things that distinguishes human beings from animals is the sense of ourselves as free, autonomous individuals. In the words of the ancient Greek philosopher Aristotle: “Where we are free to act, we are also free not to act, and where we are able to say ‘No’, we are also able to say ‘Yes’.” ’

  .,   –   

127

14. Conclusion In Part One, I provided a textual analysis of EN 1113b7–8. The result was that there is only one type of correct translation of the sentence: as an abbreviated version of what I have called vice-versa translations; moreover, that saying-no translations are not accurate renderings of the text we have. In Part Two, sections 6 and 7, I offered a glimpse of the meme-like proliferation of saying-no translations of EN 1113b7–8 in present-day popular culture and popular philosophy. In sections 8 to 12 I offered an explanation of how, in the nineteenth century, sayingno translations came about as the result, if you will, of a historical accident: the confluence of the Western textual tradition of the Nicomachean Ethics with a defective Arabic translation, via the Latin translation of an Arabic commentary; and combined with this, a ubiquitous belief in the authority of previous generations of scholars, centuries-old persistent misinterpretation of Aristotle’s theory of freedom (often based on Christian teachings of free will), lack of consideration for linguistic context and the course of Aristotle’s argument, and the prevalence of interpretative over literal translation in Victorian times. As to contemporary reception, it is still developing.

5 Moral Responsibility and Moral Development in Epicurus’ Philosophy In this chapter I argue that Epicurus had a notion of moral responsibility that is based on the agent’s causal responsibility, as opposed to the agent’s ability to act or choose otherwise; and that Epicurus considered it a necessary condition for praising or blaming an agent for an action that it was the agent and not something else that brought the action about. Thus, the central question regarding moral responsibility was whether the agent was the, or a, cause of the action, or whether the agent was forced to act by something else. Actions could be attributed to agents because it is in their actions that the agents, qua moral beings, manifest themselves (section 1). As a result, the question of moral development becomes all-important. I discuss the extant evidence for Epicurus’ views on moral development, i.e. (i) on how humans become moral beings and (ii) on how humans can become morally better. It becomes clear that Epicurus envisaged a complex web of hereditary and environmental factors as shaping the moral aspect of a human being (section 2). It emerges that, in line with Epicurus’ theory of moral responsibility and moral development, Epicurean ethics does not have the function of developing or justifying a moral system that allows for the effective allocation of praise and blame. Rather, for Epicurus the function of ethics—and in fact of the whole of philosophy—is to give everyone a chance to morally improve (section 3).

1. Moral Responsibility For the purposes of this chapter, I assume that if a person is morally responsible for an action, this is a necessary and sufficient condition for moral appraisal of that person for that action. For instance, if the action is morally wrong, moral blame is in order. Other morally relevant responses that are sometimes connected This essay first appeared in B. Reis and S. Haffmans (eds), The Virtuous Life in Greek Ethics (Cambridge: Cambridge University Press, 2006), 206–29, and is reproduced here with permission. Early versions of this essay have been presented at the Institute for Classical Studies in London, at Cornell University, and at a meeting of the Cambridge Philological Society, and I am grateful to my audiences for their stimulating discussions. Special thanks go to David Blank, Charles Brittain, and David Sedley for their most helpful comments. It is my pleasure to dedicate this essay to Dorothea Frede, from whose wisdom, kindness, and generosity I have benefited many times over the past two decades.

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0006

     

129

with moral responsibility are praise, pardon, shame, pride, reward, punishment, remorse. I now introduce two quite different concepts of moral responsibility: one is grounded on the causal responsibility of the agent for an action, the other on the ability of the agent to do otherwise. The concept based on the agent’s causal responsibility considers it a necessary condition for praising or blaming an agent for an action that it was the agent and not something else that brought about the action. The question of moral responsibility becomes one of whether the agent was the, or a, cause of the action, or whether the agent was forced to act by something else. On this view, actions or choices can be attributed to agents because it is in their actions and choices that the agents, qua moral beings, manifest themselves. The second idea of moral responsibility considers it a prerequisite for blaming or praising an agent for an action that the agent could have done otherwise. This idea is often connected with the agent’s sentiments or beliefs that they could have done otherwise, as well as the agent’s feelings of guilt or regret, or pride, for what they have done. Some philosophers consider the causal indeterminedness of the agent’s decision to act as necessary to warrant that the agent could have done otherwise. Depending on what conception of moral responsibility philosophers have, they will have to produce different reasons if they want to show that or how moral responsibility is preserved or integrated in their philosophical system. With a concept of moral responsibility based on causal responsibility, philosophers have to show that agents themselves (and not something else) are the causes of their actions; and they have to determine what characteristics an agent needs to have in order to be the sort of cause of an action to which responsibility can be attributed. With a concept of moral responsibility based on the agent’s ability to do otherwise, philosophers will have to show that in their systems agents are in fact capable of acting or of deciding and acting otherwise than they do. In this first section I set out to show that our sources univocally suggest that Epicurus had a concept of moral responsibility based not on the agent’s ability to do otherwise, but on the agent’s causal responsibility. There are only a few texts that provide information about Epicurus’ concept of moral responsibility. The most important information comes from Epicurus’ On Nature Book 25. In this book Epicurus considers three different causal factors that are involved in human behaviour: 1. Our original constitution (hē ex archēs sustasis; sometimes Epicurus simply uses archē) or nature (phusis), i.e. the package of soul atoms we come with, and which in part differs from person to person.

130

, ,   

2. The environment (ta periechonta); most commonly, the environment influences our behaviour via our perception of it, e.g. when I perceive that it is starting to rain heavily, I will open up my umbrella. 3. Us ourselves, or, as Epicurus also says, ‘the cause from ourselves’ (hē ex hēmōn aitia)¹ or ‘that through ourselves’ (to di hēmōn autōn), etc.² When our initial constitution and our environment together (i.e. nature and nurture, as it were) fully determine what we do, then our actions are the result of necessity.³ When we ourselves are causally involved in the actions, then they are not the result of necessity. Epicurus takes the fact that we blame each other, and try to reform each other, as an indication that the cause of our actions lies in ourselves, or that the actions happen through ourselves: (2) [And we can invoke against the argument that our behaviour must be caused by our initial constitution or by environmental factors] by which we never cease to be affected, [the fact that] we rebuke, oppose, and reform each other as if we have the cause also in ourselves, and not only in our initial constitution and in the mechanical necessity of that which surrounds and penetrates us.⁴,⁵ (Laursen 1997, 35; LS 20C(2). The text is Laursen’s. LS numbers are added for ease of reference. Greek and English translation of all of (1)–(12) are given in Chapter 6, section 3.)

The concept of blame presupposes that the beings that are blamed were themselves causally responsible for their actions. It makes no sense to blame individuals for certain events if those events came about through necessity and the individuals were forced in bringing them about.⁶ Moreover, Epicurus thinks, we have a preconception that we are the causes of our actions: (8) . . . using the word ‘necessity’ of that which we call ‘ . . . by ourselves’, he is merely changing a name; but he [i.e. Epicurus’ opponent] must⁷† prove that we

¹ Cf. also Laursen 1997, 47. ² Τὴν ἐξ ἑαυτοῦ αἰτίαν; τὸ δι᾽ἡμῶν αὐτῶν; ἔχοντας ἐν ἑαυτοῖς τὴν αἰτίαν; τὸ ἐξ ἡμῶν αὐτῶν πραττόμενον; ἡ ἐξ ἡμῶν γεννομένην (Laursen 1997, 45; 697.4). With ‘us’ or ‘we’ Epicurus refers to human beings. ³ Long and Sedley 1987 (= LS) 20C(2), Laursen 1997, 28. ⁴ Namely by means of perception, cf. Epic. Nat. 25, Laursen 1997, 33, LS 20C(1), quoted below. ⁵ (2) ἐστήκει, ὧν οὐ . . . ἀπολειπείτα πάθη τοῦ γίνεσθαι . . . νουθετεῖν τε ἀλλήλους καὶ μάχεσθαι καὶ μεταρυθμίζειν ὡς ἔχοντας καὶ ἐν ἑαυτοῖς τὴν αἰτίαν καὶ οὐχὶ ἐν τῇ ἐξ ἀρχῆς μόνον συστάσει καὶ ἐν τῇ τοῦ περιέχοντος καὶ ἐπεισιόντος κατὰ τὸ αὐτόματον ἀνάγκῃ. ⁶ Cf. LS 20C(3). ⁷ I follow Laursen’s reading (Laursen 1997); ‘†’ indicates differences from the text in LS.

     

131

have a preconception of a kind which has faulty delineations when we call that which [comes] through ourselves causally responsible.⁸ (Epic. Nat. 25; Arr. 34.28, Laursen 1997, 37; LS 20C(8)) (4) . . . when he blames or praises. But if he were to act in this way, he would be leaving intact the very same behaviour [i.e. praising and blaming] which we think† of as concerning ourselves, in accordance with† our preconception of the cause;⁹,¹⁰ (Epic. Nat. 25; Arr. 34.28, Laursen 1997, 36; LS 20C(4))

A preconception is some kind of veridical general conception or true opinion that we have acquired empirically by having repeatedly the same sort of perceptual experience.¹¹ Preconceptions are self-evident.¹² Epicurus, it seems, holds that we acquire the preconception of us as causes precisely through our observations that human beings, including ourselves, are praised and blamed for their actions. And given that Epicurus thinks we have such a preconception, and that all preconceptions are self-evident, we can surmise that he thought it to be self-evident that we are the causes of our actions. Two other passages from the same book, On Nature 25, suggest that what makes an action blame- or praiseworthy, or exempted from blame and praise, is not the fact that it is a physical event of a certain kind but that it has causal factors of a certain kind: And if, on account of the cause which is now already from itself, it actually goes to what is similar to the original constitution, and this is a bad one, then we censure it at times even more, but rather in an admonitory way . . .¹³ (Epic. Nat. 25; Arr. 34.25, Laursen 1997, 31)

and: (1) But many naturally capable of achieving such and such results fail to achieve them, because of themselves, not because of one and the same causal responsibility of the atoms and of themselves. (2) And with these [people] we especially do battle, and rebuke them . . . because they behave in accordance with their

⁸ (8) . . . φ ἡμῶν αὐτῶν καλούμενον τῶι ἀνάγκης ὀνόματι προσαγορεύειν ὄνομα μόνομ μετατίθεται† δεῖ δ᾽ἑπιδῖξαι ὅτι τοιοῦτο τι ω μοχθηρ[οί εἰσι τύ]ποι προειληφότες τὸ δι᾽ἡμῶν αὐτῶν αἴτιον καλοῦμεν, οὐτιδ . . . ⁹ I follow Laursen’s reading (cf. next footnote). However, I do not quite understand what it means. I hope it still means the same as what Sedley 1983 suggested, viz. that our observation of blaming and praising produces our preconception of us as causes of our actions. ¹⁰ (4) μεμφόμενος ἢ ἐπαινῶν ἀλλ᾽εἰ μὲν τοῦτο πράττοι, τὸ μὲν† ἔργον ἂν εἴη καταλείπον ὃ ἐφ᾽ἡμῶν αὐτῶν κατὰ† τὴν τῆς αἰτίας πρόληψιν, ἐννοοῦμεν†. ¹¹ Cf. DL 10.33. ¹² Cf. DL 10.33, ἐναργές. ¹³ ἂν δὲ καὶ βαδίζη διὰ τὴν ἑαυτοῦ ἤδη αἰτίαν εἰς τὸ ὅμοιον τῃ ἐξ ἀρχῆς συστάσει φαύλη οὒση, ἔτι μᾶλλον ἐνίοτε κακίζομεν, ἐν νουθετητικῷ μέντοι μᾶλλον τρόπῳ . . .

132

, ,   

disordered original nature, as we do with the whole range of animals. (3) For the nature of their atoms has contributed nothing to some of their behaviour and degrees of behaviour and dispositions, but it is their developments which themselves possess all or most of the causal responsibility for certain things.¹⁴ (Epic. Nat. 25; Arr. 34.21, Laursen 1997, 19; LS 20B(1–3); tr. Long and Sedley, modified)

Thus, if an action of a certain kind is caused by the initial constitution of a person, in response to the environment, then the person is—it seems—not to be reproached. Thus if, say, toddler Tina throws a tantrum for not getting a toy, she is presumably not to be morally blamed. However, if an action of the same kind is caused by Tina herself, and thus not by the initial constitution, then she is to be blamed for it. For example, if Tina keeps having tantrums about trivia as an adult, presumably she is to be blamed for them. One last passage concerned with moral responsibility comes from the Letter to Menoeceus: [he says that some things happen by necessity,] others by chance, and others again because of us, since necessity is not accountable to anyone, and chance is an unstable thing to watch, whereas that [which happens] because of us is without master, and culpability and its opposite are naturally attached to it. (Epic. Ep. Men. 133–4)¹⁵

Here we learn that Epicurus distinguished between three types of events: those that happen by necessity, those that happen by chance, and those that happen because of us (ginesthai par’hēmas). He adds two pieces of information about the things that come to be because of us. First, they are said to be without master. By this I take Epicurus to mean that it is we who bring the events about and not something else; in particular not fate, which he called master (or rather mistress) just the sentence before. Second, the things that happen because of us are said to have praiseworthiness and culpability naturally attached to them. This suggests that they are precisely those things for which we can be held morally accountable. Taking the two points together, it—again—appears that I can be held morally

¹⁴ Πολλὰ δὲ καὶ τῶνδε καὶ τῶν[δε φ]ύσιν ἔχοντα ἀπεργαστικὰ [γί]νεσθαι δι᾽ἑαυτὰ οὐ γίνεται ἀ [πε]ργαστικὰ (οὐ διὰ τὴν αὐτὴν αἰτία[ν] τῶν τε ἀτόμων καὶ ἑαυτῶν), οἷς δὴ καὶ μάλιστα μαχόμεθα καὶ ἐπιτιμῶμεν. [.] . οῦντες κατὰ τὴν ἐξ ἀρχῆς [τ]αραχώ[δ]η φύσιν ἔχοντα, κα[θά]περ ἐπί τῶν πάντων ζῷων. [*οὐ]θὲν γὰρ αὐτοῖς συνήργηκεν εἰς ἔνια ἔργα τε καὶ μεγέθη ἔργων καὶ διαθέσεων ἡ τῶν ἀτόμων φύσις ἀλλ᾽αὐτὰ τὰ ἀπογεγεννημένα τὴν πᾶσαν ἢ τὴν πλείστην κέτηται αἰτίαν τῶνδέ τινων, ἐκ δ᾽ἐκείνης ἔνιαι τῶν ἀτόμων κινήσεις ταραχώδεις κινοῦνται, οὐχὶ διὰ τατ[ . . . πά]ντως [διὰ δὲ τῶ]ν [παρεμ]π[ι]πτόν[των . . . ἐκ τοῦ περι]έχοντος. ¹⁵ [ἃ μὲν κατ᾽ἀνάγνκην γίνεσθαι λέγει], ἃ δὲ ἀπὸ τύχης, ἃ δὲ παρ᾽ἡμᾶς, διὰ τὸ τὴν μὲν ἀνάγκην ἀνυπεύθυνον εἶναι, τὴν δὲ τύχην ἄστατον ὁρᾶν, τὸ δὲ παρ᾽ἡμᾶς ἀδέσποτον ᾧ καὶ τὸ μεμπτὸν καὶ τὸ ἐναντίον παρακολουθεῖν πέφυκεν.

     

133

responsible for something when I am somehow causally responsible for its occurrence. And I assume that those things that come to be because of us (ginesthai par’hēmas) are those of which we ourselves are the causes. There is, then, one element all these passages on moral responsibility have in common: they connect the concept of moral responsibility with us as causal factors of the things for which we are considered morally accountable. We thus need to see what it is that makes us the causes of our behaviour—as opposed to the mere combination of our initial constitution (or atoms) and our environment. In the surviving evidence, Epicurus never directly addresses this question, but his On Nature 25 gives some hints as to how he would have answered it. If we are the cause of an action, this involves, first, that we are not forced in bringing it about (LS 20C(10); Ep. Men. 133); second, that we have an impulse (hormēma) or desire (prothumia) to perform that action; and third, that we act in accordance with that impulse or desire (LS 20C(9–11)). Fourth, the most important element seems to come from Epicurus’ gloss on what it is that something comes to be because of us (ginesthai par’hēmas): Hence at some point† it is unqualifiedly because of (para) us that the development comes to be now of this kind or that kind; i.e. the things which on account of the pores flow in of necessity from what surrounds us at some point come to be [of this kind or that kind] because of (para) us, that is, because of (para) our beliefs [that come] from ourselves.¹⁶ (Epic. Nat. 25; Arr. 34.26, Laursen 1997, 32–3; Long and Sedley, vol. 2, 20C(1))

It seems (if, as I suggest we should, we understand the ‘or rather’ (kai)¹⁷ in the last sentence as epexegetic) that if something comes to be because of us, then it does so because of our own beliefs (doxai).¹⁸ If this is so, we can infer several things about Epicurus’ concept of moral responsibility.

¹⁶ ὥστε παρ᾽ἡμᾶς π[οθ]† ἁπλῶς τὸ ἀπογεγεννημένον ἤδη γείνεσθαι, [τ]οῖα ἢ τοῖα, καὶ τὰ ἐκ τοῦ περιέχοντος κ[α]τ᾽ἀνάγκην διὰ τοὺς πό[ρους] εἰσρέο[ν]τα παρ᾽ἡμᾶς π[ο]τε γε[ίνε]σθαι καὶ παρὰ τ[ὰς] ἡμε[τέρα]ς [ἐ]ξ ἡμῶν αὐτ[ῶν] δόξ[ας. †This is Laursen’s reading. ¹⁷ Laursen’s tentative reading in papyrus 1056 is ἤ, which also supports an epexegetical interpretation. ¹⁸ David Sedley has suggested to me that ‘happens (or comes to be) depending on us’ captures γίνεσθαι παρ’ ημᾶς better than ‘happens (or comes to be) because of us’. I disagree. I take the παρὰ in γίνεσθαι παρ’ ημᾶς in the general sense illustrated in LSJ C III 7, with possible alternative translations ‘happens on account of us’, ‘happens through us’, or ‘happens with us being the cause’. Phrases like ‘this happened depending on me’ seem to strain the English, and generally I believe that the use of παρὰ + acc. together with γίνεσθαι makes ‘depending on us’ a less desirable translation. In any case, I believe the causal reading I give to παρὰ + acc. is justified by the fact that this is the best reading for the grammatically parallel παρὰ τ[ὰς] ἡμε[τέρα]ς [ἐ]ξ ἡμῶν αὐτ[ῶν] δόξ[ας in the very same clause, in which, I believe, Epicurus must have used παρὰ in the same sense as in the immediately preceding παρ᾽ἡμᾶς π[ο]τε γε[ίνε]σθαι καὶ.

134

, ,   

First, this suggests that if we are morally responsible for something (actions, characteristics), then we must have certain beliefs, and hence be capable of having beliefs. These requirements preclude very small children and some animals from being responsible for their behaviour, since they do not have beliefs. This squares with Epicurus’ view, as elsewhere attested, that wild animals are not to be held responsible for their behaviour.¹⁹ Second, our beliefs must somehow be causally involved in bringing about the actions and dispositions that ‘come to be because of us’. Now we know that, for Epicurus, beliefs are indeed causally involved in most of our behaviour.²⁰ By beliefs (doxai) Epicurus does not intend—or at least not primarily—that, for example, when I sit down, I probably thought ‘I should sit down’ just beforehand. Beliefs (doxai) are not (just) volatile, momentary, thoughts of this kind, but are lasting and firmly held by the individual; they make up a fundamental part of our mental dispositions; they underlie or are part of our dispositions to experience emotions; and they determine what kinds of desires we will have, and hence what we do. For example, if I believe that thunder is a form of divine punishment, then this will invoke certain fears in me, which in turn will make me tend to react in certain ways to external factors like thunderstorms. We can thus understand why Epicurus thought that our beliefs play an essential role for our moral responsibility. Switching to the level of atoms, we can describe this as follows: the atomic structure of our mind is in part determined by the beliefs we hold. For our beliefs are certain structures of the atoms in the mind which make it possible that certain external influences can enter our mind (and be thought, and reacted upon), whereas others cannot. For the latter there will be no pathways or channels of the right shape, that is, of the shape that would be needed for it to be possible for them to enter the mind. I picture this process analogously to that of perception, as described by Epicurus. Third, it seems that the fact that Epicurus emphasizes that the beliefs are the agents’ own beliefs (τ[ὰς] ἡμε[τέρα]ς [ἐ]ξ ἡμῶν αὐτ[ῶν] δὁξ[ας), implies at the very least that the agents must have thought the beliefs themselves, and that they must have accepted them in some way. However, presumably what Epicurus has in mind is something stronger. For instance, he may have envisaged a distinction between, on the one hand, beliefs that a person simply took over from others, without thinking them through, and, on the other, beliefs that the person thought through, and then adopted as their own, on the basis of some rational grounds for the belief.²¹

¹⁹ Laursen 1997, 31. ²⁰ E.g. Ep. Men. 132; DL 10.149 = Sent. 30; see also Furley 1967, 202 and Mitsis 1988, 141. ²¹ Cf. the distinction between false assumptions (ὑπολήψεις ψευδεῖς) and προλήψεις about the gods, Ep. Men. 124; see also Mitsis 1988, 141 n. 27; cf. further ‘if we do not grasp what the canon is, i.e. that which judges all things that come to be through beliefs, but irrationally follow the tendencies of the

     

135

Not all the details of Epicurus’ conception of moral responsibility are then clear. But all passages connect the concept of moral responsibility with our being causally responsible for the things for which we are morally responsible. There is no trace of a concept of moral responsibility which takes it to be a necessary condition that we (the same persons, in the same circumstances) are capable of deciding or acting otherwise than we do. On the contrary, it seems that Epicurus held that, when we are the causes of our actions, then how we react to external stimuli at any given time will depend fully on us as we are at the time when we set out to act,²² i.e. on our overall mental disposition, including our beliefs, at that time.²³ That is, if in two situations my overall mental disposition was the same, I would in my actions necessarily react in the same way to the same external stimuli. There is—in this sense—no possibility for us to act ‘out of character’. *** Let us now turn to the topic of moral development and its connection with moral responsibility.²⁴ I call an individual a ‘moral being’ if they can in principle be held morally responsible for at least some of their behaviour. We can then distinguish two different issues concerning moral development: 1. The development of someone from not being a moral being (i.e. not to be held morally responsible for anything) to their being a moral being (being held morally responsible—in principle—for some things). There is thus a conceptual connection between moral responsibility and moral development. 2. The development of a moral being from their present state to being morally better or morally worse (in some particular respect or overall); for Epicurus, the important case is that of becoming morally better, i.e. of moral improvement or moral progress.

2. Moral Development I: How Do We Become Moral Beings? Our first question is ‘How do we become moral beings?’ or ‘What does our transition from the non-moral or pre-moral stage of our lives to the moral stage consist in?’ Following what has been said above, we can also rephrase this question

many, everything with respect to which we investigate will be lost’ (Laursen 1997, 43–4; context and Greek in section 3 below). ²² At least nothing in the sources is incompatible with the assumption. ²³ See also Philodemus. Ir. 32 and cf. Annas 1992, 180–1 and Mitsis 1988, 141. ²⁴ One can talk about the moral development of individuals as well as of groups of people, such as societies or communities. In this chapter, I am interested exclusively in the moral development of individuals.

136

, ,   

as ‘How do we become beings that are morally responsible for their behaviour?’ For philosophers who think that we are morally responsible for an action if we could have done otherwise, this question boils down to: ‘When and how do we become capable of doing otherwise than we do?’ For instance, such philosophers presumably do not consider toddlers as capable of doing otherwise (in the relevant sense), but consider eighteen-year-olds as having that capacity. (The question, however, is rarely asked in this fashion.) By contrast, for philosophers like Epicurus, who think that we are morally responsible for our actions when we ourselves are the causes of our actions, the question amounts to: ‘When and how do we ourselves become causally responsible for our actions?’* Roughly, Epicurus’ answer to this question appears to be this: we become causally responsible for our actions when our mind has developed to a point at which we are capable of consciously adopting as our own beliefs beliefs that do not square with our initial constitution, and can internalize these beliefs so that they in turn causally influence our actions. But let us proceed more slowly. I start with a reminder of the very basics of Epicurus’ ethics. In a nutshell, and ruthlessly simplified, it amounts to this: • The end (telos) which all human beings do and should aim at is a life of happiness (eudaimōnia) above all in the form of tranquillity (ataraxia),²⁵ that is freedom from mental disturbance (tarachē). For Epicurus such happiness consists in pleasure²⁶ and the absence of pain. • It is the task of ethics to aid us in achieving happiness. • The core of Epicurean ethics is a complex theory of desires, pains, and pleasures that enables us to grasp with our intellect which of all the many pleasures and pains to choose, and which to shun, so that we can attain tranquillity.²⁷ • It turns out that as a matter of fact we can reach such tranquillity only if we are virtuous.²⁸ However, virtue is understood strictly as a means to the end— pleasure.²⁹ So far Epicurus’ ethics in a nutshell. Back to moral development. First, what do we know about Epicurus’ view of the mental constitution of human beings very early on, say, at birth?³⁰ We have at birth two different kinds of

* There is to my knowledge no evidence that Epicurus ever considered a question like ‘When or how do we become able to do otherwise?’ On the other hand, there is some evidence that he asked when and how we become causes of our actions. ²⁵ Ep. Men. 128, 131. ²⁶ E.g. Ep. Men. 129. ²⁷ E.g. Ep. Men. 127–32. ²⁸ Ep. Men. 132; DL 10.140. ²⁹ DL 10.138. As we take medicine for the sake of health, so we choose the virtues on account of pleasure and not for their own sake. ³⁰ Lucr. 3.344–7: anima and body are combined already in the mother’s womb.

     

137

atomic substructures in our mind: those which we all share and those in which we differ. Our minds are all equal, for example, in so far as from birth we all aim at obtaining pleasure and avoiding pain.³¹ We do so instinctively, not as the result of deliberation, since we cannot deliberate yet.³² Our minds differ from each other, for instance, in that we have different emotional tendencies, different dispositions for experiencing certain emotions and for behaving accordingly. Thus, in terms of atoms, if we have more fire-like atoms in the nature of our mind, we get angry more easily; if we have more air-like atoms in the nature of our mind, we are more easily afraid; and if there is an abundance of breath-like atoms, we are naturally calm.³³ These tendencies are morally relevant, since they later co-determine how successful we will be when trying to achieve tranquillity. Next, how does the mind develop from birth to adulthood? Lucretius writes: Furthermore, we perceive that the mind is born jointly with the body, grows up jointly with it, and ages jointly with it. For just as infants walk unsteadily with a frail and tender body, so too their accompanying power of mental judgement is tenuous. Then when they have matured to an age of robust strength, their judgement is greater and their mental strength increased.³⁴ (Lucr. 3.445–50)

This passage suggests a gradual development of the mind, starting in early childhood, leading to a more fully developed intellect and capacity for reasoning at the onset of adulthood. There is some evidence that Epicurus thought some of our mental development to be necessitated by the initial constitution of our minds—both concerning developments based on elements shared by all humans and developments based on elements that differ across individuals. Thus, in On Nature 25 Epicurus writes: But if in the mind the first constitution expels something from the development, this sort of thing not being developed from necessity all the way to these

³¹ Cf. DL 10.137, the so-called ‘cradle argument’. ³² Cf. DL 10.137, ‘naturally and without reason’ (φυσικῶς καὶ χωρ`ς λόγου); also Cic. Fin. 1.30. ³³ ‘(4) The mind also has that kind of heat which it takes on when it boils with anger and the eyes shine with a fiercer flame; it has plenty of cold wind, the companion of fear, which excites fright in the limbs and rouses the frame; and it has that state of the still air which is found in a tranquil chest and in a calm face. But there is more heat in those with fierce hearts and angry minds which easily boil over with anger. A prime example is the lion, which regularly bursts its chest with roaring and groaning and cannot contain the billows of rage in its chest. But the cold mind of stags is more windy, and quicker to rouse through their flesh those chilly gusts which set the limbs in trembling motion. The nature of cattle, on the other hand, is characterized more by calm air. Neither does ignition by the smouldering brand of anger ever over-excite it and cloud it with blind darkness, nor is it transfixed and numbed by the icy shafts of fear. It lies midway between stags and fierce lions. (5) Likewise the human race’ (Lucr. 3.289–307, LS 14D(4–5)). ³⁴ Praeterea gigni pariter cum corpore et una j crescere sentimus pariterque senescere mentem. j Nam velut infirmo pueri teneroque vagantur j corpore, sic animi sequitur sententia tenuis. j Inde ubi robustis adolevit viribus aetas, j consilium quoque maius et auctior est animi vis.

138

, ,   

particular things; but on the one hand to the point where it has come to be a soul, or rather a soul which has a disposition and motion of such a size, this sort of thing being developed from necessity from these sorts of things; and on the other hand to the point . . . (Arr. 34.24, Laursen 1997, 28, my emphasis; text continued below)³⁵

We may perhaps think of these necessitated developments as a kind of genetically directed ‘maturing’ of the mind. (Thus in this way it is determined that an individual develops a soul, and that that soul has a disposition and motion of a particular size.) We can imagine that with age our minds unfold to greater and greater complexity (in the combination of the atoms), and this means that we acquire more and more capacities and dispositions. For instance, we all learn how to speak (a complicated business);³⁶ we all develop some ‘preconceptions’ (e.g. of horses and cows);³⁷ we all learn to reason.³⁸ Despite genetic pre-programming, these ongoing differentiations of our minds still mostly require certain environmental conditions in order to be realized: if we perceived nothing, we would not obtain any preconceptions of things; and if no one around us spoke a language, nor would we.³⁹ The dispositions we develop will at least in part vary from person to person, owing to the differences in our initial constitutions—and this will be so even when the environments are similar. Moreover, all these developments take a certain time—thus we won’t learn how to speak before the approximate age of one; and we won’t learn how to reason properly before adolescence.⁴⁰ In this context, Epicurus considers time and age as causes,⁴¹ that is—I take it—as factors we can use rightly in explanations in answer to questions like ‘Why can’t Tina speak yet?’—‘She’s not yet one.’ In addition, differences in the environment will influence us in such a way that different individuals turn out differently, even where their initial natures do not vary significantly: thus, only when we perceive kangaroos will we obtain a preconception of kangaroos, and only when people around us speak Greek will the language we learn as a child be Greek.⁴² Moreover, Epicurus is known to have recommended that one should watch over the young so that they do not develop maddening desires.⁴³ Thus he thought that different external influences on children make them develop different desires. Hence, before reaching the age of

³⁵ Κἂν κατὰ διάνοιαν δέ τι ἐκβιάζηται ἡ πρώτη σύστασις τοῦ ἀπογεγεννημένου, μὴ ἐξ ἀνάγκης μέχρι τωνδί τινων ?ἐξ ἀνάγκης ?τοιοῦνδε ἀπογεννωμένου ἀλλὰ μέχρι μὲν τοῦ ψυχήν γενέσθαι ἢ καὶ τοσαυτηνὶ διάθεσιν καὶ κίνησιν ἔχουσαν ψυχὴν ἐξ ἀνάγκης ?τοιοῦδε ἀπογεννωμένου ἐκ τῶν τοιουτωνί, μέχρι δὲ . . . (text continued in n. 48 below). ³⁶ Inferred from Lucr. 5.1028ff. on the origin of language. ³⁷ Inferred from DL 10.33. ³⁸ Lucr. 3.445–50, quoted above, LS 19B. ³⁹ Again inferred from Lucr. 5.1028ff., LS 19B. ⁴⁰ e.g. Lucr. 3.445–50, quoted above. ⁴¹ Cf. Laursen 1997, 28–9 and 42–5. ⁴² Cf. Mitsis 1988, 147. ⁴³ Sent. 80.

     

139

morality, education and environment generally are factors that will influence our chances to become moral. Epicurus’ recommendation to watch over the young also exemplifies that there is one very important way in which the environment may have an influence even on our pre-moral mental development. Once we can speak and start having thoughts, we will be able to, and will indeed, take in the views or beliefs of other people around us, in particular of parents, teachers, and peers. Thus we are likely to take in the views that death is bad and to be feared, that the soul is immortal, that there is punishment in the afterlife, and more such nonsense. We will initially take these views in without questioning them, and will unreflectively internalize them; they will determine our desires and fears, and how we react to new, incoming, influences, just as our preconceptions will. Thus if I have been told and now believe that thunder and lightning are a form of divine revenge, and I have a predominantly air-like mind, this will invoke fear of thunderstorms in me and I will behave accordingly. So our mental development is a rather complicated affair. But however complex our mind becomes, until we are ourselves the causes of our actions, for Epicurus, everything we become and do is a function of hereditary and environmental factors. That is, we are fully determined or necessitated by a combination of those two factors.⁴⁴ And as long as this is so, we cannot be held morally responsible for our behaviour.⁴⁵ How should we envisage the transition from this pre-moral stage to the stage when we are moral beings (i.e. when we are causal factors of our actions)? We can assume that Epicurus believed that we all have the innate potential for becoming moral beings. For first he thinks that we can all morally improve (a point I will get to later). And second, he seems to think that one of those features of mental development that we all share is that of becoming a cause of our actions.⁴⁶ For this, as I said earlier, we need to have ‘our own beliefs’, and these have to be causally relevant for what we do.⁴⁷ Epicurus implies also that we need to reach a certain age or stage of maturity, a stage at which our mind has developed a sufficient complexity, in order for it to be possible that we are causes ourselves (this is the continuation of the text from On Nature 25 quoted above): . . . on the other hand, to the point where [it has come to be] this kind of soul or that, this sort of thing being developed not from necessity, or rather this sort of

⁴⁴ LS 20C(2); also Laursen 1997, 28. Our development is necessary as long as we are not causes ourselves. ⁴⁵ LS 20C; and inferred from Laursen 1997, 28. ⁴⁶ Laursen 1997, 43–5; also 28. ⁴⁷ LS 20C(1); also Laursen 1997, 43–5; LS 20B(5–7) (Laursen 1997, 22).

140

, ,   

thing being developed not by necessity whenever there is advancement in age, but from itself being able—or [from] the cause which comes from itself—[to develop? to become? to bring about?] something else as well . . . (Arr. 34.24, Laursen 1997, 28)⁴⁸

In this passage internal necessitation and lack thereof do not concern individual actions or volitions, but what a person’s soul comes to be like. The emphasis is on the non-necessity of mental development and, in particular, on the fact that we ourselves (or the cause from ourselves) are causally responsible for the changes in our soul, and that these changes are not necessary.⁴⁹ One fragmentary passage from On Nature 25 seems to suggest that we become causes ourselves at the moment when there is in us a certain development that differs qualitatively from (what is there in) the atomic structure of our mind, and we make this new development part of our ‘nature’ or mind, as it were. (5) In this way whenever something is developed which takes on some distinctness from the atoms in a way that pertains to judgement⁵⁰—not in the way as from a different distance—he receives the causal responsibility which is from himself; (6) and then he immediately imparts this to his first natures and somehow makes the whole of it into one. (7) This is why those who cannot correctly make such distinctions confuse themselves about the adjudication of causal responsibility. (Epic. Nat. 25, Arr. 34.22; Long and Sedley 20B(5–7), Laursen 1997, 22; tr. Long and Sedley, modified)⁵¹

What exactly this means, I do not know. But it seems to me that Bob Sharples is right when he says about this passage: the obvious, indeed inevitable way of interpreting this in the atomic context is to say that we, by thought and effort, can modify our character, and hence also the atomic structure of our minds . . . the downwards causation in the passage . . . may thus relate to the process by which we modify our characters, and not to the explanation of free choice by volition causing atomic swerves.⁵²

⁴⁸ . . . μέχρι δὲ τοῦ τοιανδὶ ψυχὴν ἢ τοιανδὶ οὐκ ἐξ ἀνάγκης τοιοῦδε ἀπογεννωμένου ἢ οὐκ ἐπειδὰν προβῇ γε τῇ ἡλικίᾳ τοιοῦδε ἀπογεννωμένου κατ᾽ἀνάγκην ἀλλ᾽ἐξ ἑαυτοῦ δυναμένου καὶ τῆς ἐξ ἑαυτοῦ αἰτίας καὶ ἄλλο . . . (text continued from n. 35 above). ⁴⁹ Cf. Bobzien 2000, chapter 6 in this volume. ⁵⁰ Or ‘in a discriminating way’. ⁵¹ (5) οὕτως ἐπειδὰν ἀπογεννηθῆ τι λαμβάνον τινὰ ἑτερότητα τῶν ἀτόμων κατὰ τινα τρόπον διαληπτικόν, οὐ τὸν ὡς ἀφ᾽ἑτέρου διαστήματος, ἰσχάνει τὴν ἐξ ἑαυτοῦ αἰτίαν, (6) εἶτα ἀναδίδωσιν εὐθὺς μέχρι τῶν πρώτων φύσεων καὶ μίαν πως ἅπασαν αὐτὴν ποιεῖ†. (7) ὅθεν δὴ καὶ οἱ μὴ δυνάμενοι κατὰ τρόπον τὰ τοιαῦτα διαιρεῖν χειμάζουσιν αὑτοὺς περὶ τὴν τῶν αἰτιῶν ἀπόφασιν. †Laursen 1997, 22. ⁵² Sharples 1991–3: 186.

     

141

Regarding the details, I believe Epicurus’ idea could have been something like this: we ourselves become causes at the moment at which we—consciously—identify with an incoming idea or thought which is not in keeping with the beliefs we have so far taken in from our environment ‘unthinkingly’, as it were, and in accordance with the original nature of our mind. More precisely, when we identify with this new thought, we incorporate it into our mind, and thus change our mental dispositions. As a result, from then on our actions can be caused by behavioural dispositions that are at least partially the result of our identifying with something that was not part of our original constitution.⁵³ (Earlier in On Nature 25, Epicurus mentions that at some point in our life we become able to ‘think ourselves by means of ourselves’.⁵⁴ So it is tempting to assume that only when we are able to ‘think ourselves by means of ourselves’ will we be able to have beliefs that are our own—not just absorbed ‘unthinkingly’, from others.)⁵⁵ At that point, thus, the disposition is no longer—fully—the result of internal and external necessity, but in part the result of conscious, rational, influencing. When, then, someone acts from such a disposition, they are the cause of the action, and no longer ‘the atoms’, i.e. those of their initial constitution. Of course, nothing guarantees that the incoming thoughts we identify with are correct—they, too, may be based on cultural prejudice. Equally, if we rethink some of our culturally induced beliefs, nothing prevents us from ending up retaining false beliefs, or replacing false beliefs by false beliefs. In another passage in On Nature 25—which is rather badly preserved— Epicurus may refer to the same kind of development: And the same thing was both generated as permanent and was a kind of seed, as I say, leading from the origin[al nature] to something else, and when this is present, we think or form beliefs . . . and there is much that [happens] with [our] nature helping, and much [that happens] when [our] nature is not helping, and there is something that [happens] when our nature is rearranged by us. (Arr. 34.31, Laursen 1997, 44–5, context and Greek in section 3 below)

⁵³ This is then a quasi-empirical proof that it is not our original nature. Another passage (LS 20B (1–4), Laursen 1997, 19–20, in fact just before the one quoted) deals with the point that, once we are causes ourselves, even if the behaviour is the same as that of our original nature would have been, we are responsible, since we (not our original nature) are the cause. Generally, thus, it would presumably suffice if we were capable of incorporating ‘alien’ elements into our mind, whether or not we actually do so. ⁵⁴ Laursen 1995, 46–7: ἐν] [ἑ]αυτῷ κατὰ τὸ ὅμοιον καὶ ἀδιάφορον ἑαυτὸν [ῥ]ηθήσεται διανοεῖσθαι. ⁵⁵ Laursen himself suggests in Laursen 1995: when we ‘think ourselves by means of ourselves’ we have the mental faculty of ‘subsequent reasoning’ (45–6); ‘this makes it possible to realize one’s real goal (τέλος)’ (46); ‘reason is acquired by time . . . at a comparatively late stage, we acquire the capacity for a reasoned consideration of our state as a whole’ (47), but then, strangely, continues the last sentence thus, ‘that is, we become moral philosophers’. In my view, it is much more likely that at that point we have become adult, rational beings.

142

, ,   

This passage—at least as I have reconstructed it following Laursen—suggests that all human beings, at some point, reach a developmental stage at which they start thinking and forming opinions.⁵⁶ Epicurus seems to assume then that—in the normal course of events—at some point during the process of their growing up every person reaches this stage of development at which they start ‘thinking for themselves’. Yet, it is unclear whether the quoted passages describe a unique event in a person’s life or a gradual process in which a person changes or confirms (part of) their beliefs one by one upon reflection over a longer period of time. If this is a unique event, we may consider our becoming moral beings as instantaneous. If it is a gradual process, Epicurus could have thought that we become moral beings when we ourselves become causes of something for the first time (since we then have the capacity for becoming causes ourselves); or else, our becoming morally responsible could itself be a gradual process. This last possibility seems to me the most exciting, as it seems both a very modern idea and to come closest to the truth.⁵⁷

3. Moral Development II: How Can We Become Morally Better? Let us assume that we are now moral beings (as opposed to pre-moral beings). How can we then develop morally according to Epicurus? In particular, how can we become morally better? Theories of moral responsibility based on the agent’s ability to do otherwise tend to spend little time on the question of moral development. When they deal with it, they may connect measures for moral improvement with an agent’s freedom of decision in two ways. (i) Agents need to be given a maximum of relevant information for mental storage that they can make use of at times when they have to make a moral decision. Relevant information here covers anything that enables them to find out what the morally right thing to do is. (ii) Agents need to strengthen their will—or whatever it is that decides freely—so that in a situation of choice, in which they know what the right thing to do is, they will actually decide to do the right thing instead of satisfying some adverse more immediate desire. By contrast, theories of moral responsibility that are based on the agents qua rational agents as the cause of their behaviour tend to display substantial interest in the question of moral development. There is one simple reason for this: if the ⁵⁶ The end of the passage suggests the possibility that we ourselves rearrange the atoms in our minds, which plausibly could be our changing our dispositions as a result of our adopting new beliefs— see also section 3 on this point. ⁵⁷ Remember that Lucretius, in the passage quoted above, describes the development of our rationality as if it were gradual.

     

143

assumption is that how a person acts or reacts in a given situation depends fully on that person’s mental dispositions at the time, then moral improvement becomes a question of how one can alter one’s mental dispositions in such a way that one will react in the morally right way to external stimuli. This is the only possible way of getting oneself to act differently or better than one tends to do at the present. For there is no way of ‘deciding or acting out of character’ at the very moment one has to decide what to do: one’s decision is a function of the overall state of one’s mind. In line with his concept of moral responsibility, this is the approach to moral development Epicurus took. He believed that, in order to make moral progress, one has to change one’s mental dispositions: . . . Likewise the human race. Even though education may produce individuals equally well turned out, it still leaves those original traces of each mind’s nature. And we must not suppose that faults can be completely eradicated, so that one person will not plunge too hastily into bitter anger, another not be assailed too readily by fear, or the third type not be over-indulgent in tolerating certain things. There are many other respects in which the various natures and consequently the behaviours of human beings must differ, but I cannot now set out their hidden causes, nor can I find enough names for all the shapes of primary particles from which this variety springs. But there is one thing which I see I can state in this matter: so slight are the traces of our natures which reason cannot expel from us, that nothing stands in the way of our leading a life worthy of the gods.⁵⁸ (Lucr. 3.307–22, tr. Long and Sedley, modified, my emphasis)

The relevant points in this passage are these: the initial nature of a human mind includes certain morally relevant dispositions, which are present in different people in various strengths. Through education peoples’ minds can develop in such a way that these differences are by and large evened out. The reason is that by the use of our intellect we can modify our mental dispositions to a large extent. This passage corroborates the assumption that the Epicureans worked with a model of disposition-dependent agency on several counts. First, it makes it clear that Lucretius took a person’s mind to include that person’s character dispositions (3.309). Second, it implies that Lucretius thinks that one’s nature determines one’s behaviour; and third, that in order to change one’s behaviour, one has to change one’s nature, that is, the nature of one’s mind, by the use of one’s intellect. ⁵⁸ Sic hominum genus est. Quamvis doctrina politos j constituat pariter quosdam, tamen illa relinquit j naturae cuiusque animi vestigia prima. j Nec radicitus evelli mala posse putandumst, j quin proclivius hic iras decurrat ad acris, j ille metu citius paulo temptetur, at ille j tertius accipiat quaedam clementius aequo. j Inque aliis rebus multis differre necessest j naturas hominum varias moresque sequaces; j quorum ego nunc nequeo caecas exponere causas j nec reperire figurarum tot nomina quot sunt j principiis, unde haec oritur variantia rerum. j Illud in his rebus video firmare potesse, j usque adeo naturarum vestigia linqui j parvula quae nequeat ratio depellere nobis, j ut nil impediat dignam dis degere vitam (my emphasis).

144

, ,   

The last sentence of the Lucretius passage suggests that our original nature cannot be expelled completely, but that most of it can. This expelling can only consist in some change of the dispositions (or atomic structure) of our mind. In support of this assumption, Diogenes Laertius’ report of Epicurus’ view that ‘someone who has once become wise never again takes hold of the opposite disposition’⁵⁹ implies that in order to become wise, and that is morally good, one has to change one’s mental disposition. Accordingly, the major part of Epicurus’ ethics is geared to the development of the individual’s behavioural dispositions. Since the end to which all human behaviour should be above all directed is tranquillity (ataraxia), it follows that, in order to improve one’s behaviour, one has to try to develop one’s mind in such a way that one’s emotions, desires, and actions will become conducive to attaining tranquillity. With regard to the opportunities for moral improvement, there is variation across individuals. Epicurus starts out with the optimistic belief that everyone can morally improve⁶⁰ and get to a life of true pleasure and tranquillity,⁶¹ and that it is never too late to start trying to morally improve.⁶² However, moral improvement will not be equally easy for everyone, nor will the development be of the same kind. The starting situations of individuals can vary quite significantly, as is clear from the Lucretius passage on fire-, air-, and breath-like atoms.⁶³ At the point when people have become moral beings, some of them will be closer to a life of tranquillity, others further away. This will depend (a) on the education and external influences they have been exposed to up to then, and (b) on the fact that their natures are different in morally relevant ways. Thus it will be more difficult for someone with a very irascible nature to get to the right degree of irascibility than for someone of a calmer nature. But both can succeed.⁶⁴ A passage in a letter by Seneca confirms that Epicurus considered the question of whether humans and their natures are susceptible to moral improvement, and shows that he distinguished three types of human beings with respect to their moral progress:⁶⁵

⁵⁹ τὸν ἅπαξ γενόμενον σοφὸν μηκέτι τὴν ἐναντίαν λαμβάνειν διάθεσιν μηδὲ πλάττειν ἑκόντα (DL 10.117). Cf. also the passages in Nat. 25 about changing or developing one’s disposition (διάθεσις). ⁶⁰ Ep. Men. 122. ⁶¹ Lucr. 3.320–2. ⁶² ‘Let no one be slow to seek wisdom when he is young or weary in the search of wisdom when he is old; for no age is too early or too late for the health of the soul’. Μήτε νέος τις ὢν μελλέτω φιλοσοφεῖν, μήτε γέρων ὑπάρχων κοπιάτω φιλοσοφῶν˙ οὔτε γὰρ ἄωρος οὐδείς ἐστιν οὔτε πάρωρος πρὸς τὸ κατὰ ψυχὴν ὑγιαῖον, etc. (Ep. Men. 122). ⁶³ Lucr. 3.289–307, quoted in section 2. ⁶⁴ Lucr. 3.307–22. ⁶⁵ Sen. Ep. 52.3–4: (3) Quosdam ait Epicurus ad veritatem sine ullius adiutorio exisse, fecisse sibi ipsos viam. Hos maxime laudat, quibus ex se impetus fuit, qui se ipsi protulerunt. Quosdam indigere ope aliena, non ituros, si nemo praecesserit, sed bene secuturos. Ex his Metrodorum ait esse; egregium hoc quoque, sed secundae sortis ingenium . . . Ne hunc quidem contempseris hominem, qui alieno beneficio esse salvus potest; et hoc multum est, velle servari. (4) Praeter haec adhuc invenies genus aliud hominum ne ipsum quidem fastidiendum eorum, qui cogi ad rectum conpellique possunt, quibus non duce tantum opus sit, sed adiutore, et, ut ita dicam, coactore. Si quaeris huius quoque exemplar, Hermarchum ait

     

145

1. Some people’s nature is such that they can acquire tranquillity by their own impulse and efforts. (Apparently, Epicurus was thought to fall in this class.) 2. Others (want to do the right thing—velle servari—but) need some moral role model, as it were, ‘to show them the way’—which way, once they have found it, they will follow faithfully. (Metrodorus was placed here.) 3. Others again need someone to actively encourage them and perhaps even to force them along. (Hermachus is an example for this case.) Epicurus does not doubt that individuals of all three types can morally improve.⁶⁶ How then should we imagine the mental process of moral development in detail? We know that the mind of an adult human being (or of a moral being) encompasses their individual ‘initial constitution’ or nature,⁶⁷ preconceptions, and also a set of beliefs that are their own in the sense that by means of them the individual can become causally responsible for their behaviour. The mind further includes some conception of the end (telos), namely that it consists in pleasure.⁶⁸ The person’s set of beliefs will determine their emotions and desires,⁶⁹ and it will ordinarily include true and false beliefs, and among the false ones what Epicurus calls empty beliefs, i.e. beliefs that are counterproductive to reaching tranquillity— such as beliefs about vengeful gods.⁷⁰ Moral improvement will then consist in the main in restructuring⁷¹ a person’s system of concepts and beliefs in the light of the end (telos), and in strengthening the new, true, beliefs, and thus aligning the accompanying habits.⁷² False and empty beliefs will have to be first identified, and then measures will have to be taken so that the person gives them up and replaces them by true beliefs.⁷³ Epicurus was aware of the fact that my simply pointing out to myself (or having it pointed out to me) that one of my beliefs is incorrect will not make me abandon it. Just saying to myself: ‘The belief that death is an evil is false’ is unlikely to suffice

Epicurus talem fuisse. Itaque alteri magis gratulatur, alterum magis suspicit; quamvis enim ad eundem finem uterque pervenerit, tamen maior est laus idem effecisse in difficiliore materia. ⁶⁶ Did Epicurus distinguish more than these three types? Possibly. Could they all morally improve at least to some degree? Most probably yes; see Ep. Men. 122 and below. ⁶⁷ Cf. also Lucr. 3.289–322. ⁶⁸ Cic. Fin. 1.55, LS 21U(1); Cic. Fin. 1.29–39, LS 21A(4). Some Epicureans hold that this preconception of pleasure is understood by mind and reasoning. ⁶⁹ Ep. Men. 132. ⁷⁰ DL 10.144 = Sent. 15; DL 10.149 = Sent. 29–30; Porph. Marc. 31.34 8P = 239 in Arrighetti 1973, 567. ⁷¹ κατακοσμουμένης, Laursen 1997, 45; μετακοσμήσει, LS 20C(10), Laursen 1997, 39. ⁷² συνεθίζειν (accustom oneself to, make it a habit), Ep. Men. 124, 131; μελετᾶν (train, exercise oneself, practise), Ep. Men. 122; ‘Accustom yourself to believe that death is nothing to us’, Συνέθιζε δὲ ἐν τῷ νομίζειν μηδὲν πρὸς ἡμᾶς εἶναι τὸν θάνατον, followed by reasons for why we should believe this, Ep. Men. 124. ⁷³ See e.g. Ep. Men. 124.

146

, ,   

to change that belief and the desires and emotions tied to it. One obvious reason for this should be the fact that we do not have beliefs in isolation, but our beliefs are interconnected in complicated ways. Changing one’s beliefs about something will thus take some time, and it will usually involve a plurality of causal factors. Epicurus seems to have given some thought to the question of what different causal factors can be involved in moral improvement, as (yet another) fragmentary passage from On Nature 25 suggests (the information most relevant here is found after the first lacuna):⁷⁴ But often both [the considerations of the end itself and the origin(al natures)?] actually are equally causally responsible, even without the ones having been attracted by the others, or without [the ones] being attracted [by the others] and forcing many such things to happen because of time and age and other causes. Hence both the consideration of the end itself and the origin[al nature] were causally responsible, but we were, too. The [causal factor] from us was the perception⁷⁵ of the fact that if we do not grasp what the canon is, i.e. that which judges all things that come to be through beliefs, but we irrationally follow the tendencies of the many, everything with respect to which we investigate will be lost and excess . . . (lacuna of c. three words⁷⁶). And the same thing was both generated as permanent and was a kind of seed, as I say,⁷⁷ leading from the origin[al nature] to something else, and when this is present, we think or form beliefs . . . and there is much that [happens] with [our] nature helping, and much [that happens] when [our] nature is not helping, and there is something that [happens] when our nature is rearranged by us; but there is also something that [happens] when [our nature] itself leads the way (lacuna of c. three words) [not only] matured, but also because the things which flow in from the environment take the lead to improvement, and do not merely follow . . . ⁷⁸ (Laursen 1997, 43–5) ⁷⁴ I provide the context, since I refer to individual phrases from that context below. ⁷⁵ For this perception to become the causal factor from us, we must somehow have adopted it, retained it, made it our own (see above, section 2). Perhaps Epicurus’ use of ἐπαίσθησις instead of the simple αἴσθησις hints at that fact? ⁷⁶ Some phrase expressing the general idea that excess will follow or will be the result or will reign would make sense here. ⁷⁷ ὥσπερ ληρῶ would translate roughly as ‘as I keep ranting about’. However, I just don’t believe this is what Epicurus intended to say at this point. Maybe he wrote ὥσπερ λέγω, ‘as I say’, referring to ‘seed’, which is Epicurean terminology, in which the word is used metaphorically; or, perhaps, in Hellenistic times the meaning of ληρῶ had been watered down, so that it sometimes simply meant ‘I say’. ⁷⁸ . . . νον πολλάκις δὲ καὶ τὴν αὐτὴν ἀμφότερα κέκτηται μὲν αἰτίαν καὶ μὴ συνεπεσπασμένα τὰ ἕτερα ὑπὸ τῶν ἑτέρων μηδὲ συνεπισπώμενα καὶ βιαζόμεν παρὰ τε χρόνους πολλὰ τῶν τοιούτων συνπίπτειν καὶ ἡλικίας καὶ ἄλλας αἰτίας. ὅθεν καὶ τὸ τοῦ τέλους αὐτοῦ ἐπιλόγισμα εἶχε μὲν καὶ ἡ ἀρχή τὴν αἰτίαν εἴχομεν δὲ καὶ ἡμεῖς. ἦν δὲ τὸ ἐξ ἡμῶν ἐπαίσθησις τοῦ ‘εἰ μὴ ληψόμεθα τίς ὁ κανὼν καὶ τὸ ἐπικρεῖνον πάντα τὰ διὰ τῶν δοξῶν περαινόμενα ἀλλ᾽ἀκολουθήσομεν ἀλόγως ταῖς τῶν πολλῶν φοραῖς, οἰχήσεται πάντα καθ᾽ἃ διερευνώμεθά τι καὶ ὑπεροχή . . . ’ [lacuna of c. three words] τὸ δ᾽αὐτὸ καὶ ἀΐδιον ἐγεννήθη καὶ σπέρμα ἦν τι ὥσπερ ληρῶ, ἐξ ἀρχῆς πρὸς ἕτερον ἀγωγόν, παρόντος δὲ τούτο νοοῦμεν ἢ δοξάζομεν . . . πολὺ δὲ αὐτό ἐστιν μὲν ὃ συνεργούσης τῆς

     

147

First, Epicurus seems to think that in some cases of moral improvement one’s original nature (or constitution) functions as a causal factor while in others it does not.⁷⁹ One way this nature can be influential may simply be that the nature of a person’s mind ‘matures’, or develops further.⁸⁰ Then, presumably, our natural instincts to pursue pleasure may be helpful for adopting beliefs that we should pursue pleasure.⁸¹ By contrast, if someone who is by nature very irascible wants to reduce their irascibility then that aspect of their nature will not be of much help. Epicurus equally seems to acknowledge that we ourselves may restructure our nature (and thus, I assume, some of the atoms in our mind) up to a certain point.⁸² This should mean that certain of our beliefs which we have adopted as our own, and which make up ‘us’ as causes,⁸³ are used in order to get rid of certain other beliefs—beliefs which we took in either unreflectively from others, in accordance with our initial nature,⁸⁴ or as the result of inaccurate reasoning. For example, if I have many air-like atoms in my soul, I will have readily embraced the views that the gods use lightning to punish sinful mortals and hence I fear thunderstorms; but now my newly acquired beliefs about the true nature of the gods allow me to get rid of some of those views. For this method to work, presumably I will first have to come to realize the true nature of the gods, such as that they have better things to do than being concerned with earthly events. Then, I have to realize that, if this is so, the gods will not waste their time hurling lightning. However, in order actually to lose my fear of thunderstorms, I will have to thoroughly convince myself that there are no grounds for having it.⁸⁵ For this I may have to rehearse the arguments against the existence of vengeful gods repeatedly, and as many such arguments as possible, and especially so when the clouds get darker. I will also have to cultivate a replacement set of true beliefs, which, if firmly held, will provide me with the dispositions and desires needed to reach tranquillity. Finally, Epicurus also acknowledges environmental impacts on our moral development; as we have seen earlier, not many people will be able to restructure their belief system all by themselves.⁸⁶ Most people need some help from outside. They may realize that the end is a life of pleasure, which takes the form of φύσεως, ἔστι δ᾽ὃ οὐ συνεργούσης, ἔστι δ᾽ὃ κατακοσμουμένης ὑφ᾽ἡμῶν, ἔστι δὲ καὶ αὐτῆς προηγουμένης τι [lacuna of c. three words] ἐπαυξόμενον ἀλλὰ καὶ διὰ τὰ ἐκ τοῦ περιέχοντος ἐπεισιόντα τὰς καθηγεμονίας εἰς τὸ βέλτιον† , οὐ μόνον τὰς συνακολουθήσεις λαμβάνοντα† . The reading of εἰς τὸ βέλτιον (‘to the better’) is uncertain. ⁷⁹ Laursen 1997, 45, ‘there is much that [happens] with [our] nature helping, and much [that happens] when [our] nature is not helping’. ⁸⁰ Laursen 1997, at the end of 45, ‘matured’, ‘grown’ (ἐπαυξόμενον). ἐπαυξάνω also occurs at [5] VIII 8; [34.27] 8; [34.31] 31 Arr. Cf. also Epicurus calling times and ages ‘causes’ just before, Laursen 1997, 43. ⁸¹ Cf. ‘the consideration of the end itself ’, Laursen 1997, 43. ⁸² Laursen 1997, 45, ‘and there is something that [happens] when our nature is rearranged by us’. ⁸³ See above, section 1. ⁸⁴ Cf. Laursen 1997, 43, ‘we irrationally follow the tendencies of the many’. ⁸⁵ βέβαιος (‘firm’), Epic. Ep. Hdt. 63; Ep. Pyth. 85; Sent. 40; Laursen 1997, 46–7, κατανοείν, see below. ⁸⁶ Sen. Ep. 52.3–4.

148

, ,   

tranquillity,⁸⁷ and that they are not in its possession, and somehow do not succeed in getting there. In order to succeed, they have to change their environment. For example, they could join the Epicurean school, and receive the required guidance which will help them to shed their empty beliefs, and replace them by a set of true beliefs. In this way they may establish new dispositions and come closer to a life of tranquillity. Note that education of this kind constitutes a special way in which the environment has an influence on us. As we have seen earlier,⁸⁸ for Epicurus, ordinarily our reactions to external influences will be determined either by our nature or by ourselves qua causes,⁸⁹ and our overall mental dispositions will not change as a result of our reactions to our environment. For instance, if I have many air-like atoms, and hear a thunderstorm coming, then I will get frightened and hide under a blanket, say. But my dispositions won’t change. Next time there’s a thunderstorm I’ll be under the blanket again. The special case, on the other hand, is one in which environmental influences can also ‘take the lead to the better’, as Epicurus says.⁹⁰ In particular, we should assume from what else we know about Epicurus’ ethics that if we are externally influenced by teaching, we can get rid of ingrained false beliefs, and can, by thus changing our dispositions, also change our behaviour. Thus if someone provides me with a convincing scientific explanation of thunder, I may eventually stop being frightened and consequently not go into hiding anymore. Moral development, whether self-caused or triggered by others, thus consists primarily in the change of one’s mental dispositions, in particular one’s beliefs. There is a passage in On Nature 25 which suggests that Epicurus may have made an attempt at explaining such a process on the atomic level of the mind. (This passage is rather lacunose in character and its first two sentences are fragmented, but the text is still full enough to offer some interesting information.) . . . of [speech] sounds and thinking and thoughts and representations of the everlasting or non-everlasting disturbance or happiness in the soul [being/is/ are?] the cause for hunting down, little by little, the principle, or canon or criterion. For these things lead to the consideration of the criterion, and from the criterion it[self] . . . perception . . . considera[tion] . . . to the investigation little by little of the things I mentioned earlier. For these things furnished each other with their cause and use⁹¹ and each thought coming in immediately pulled along ⁸⁷ Perhaps even there help is required in some cases? Remember the third category of people in the passage from Seneca’s letter. ⁸⁸ Above, in section 1. ⁸⁹ LS 20C(1). ⁹⁰ If the—uncertain—reading ‘to the better’ or ‘to improvement’ (εἰς τὸ βέλτιον) is correct. Laursen 1997, 45, ‘but there is also something that [happens] when [our nature] itself leads the way (lacuna of c. three words) [not only] matured, but also because the things which flow in from the environment take the lead to improvement’ (context and Greek above). ⁹¹ I.e. one makes use (takes advantage of) the other.

     

149

in turn the other thought, at first coming in little by little and flowing out again quickly, then being understood more and more, in part because of the natural cause of the growth and of loss of fluidity,⁹² in part because of that [cause] which comes to be from ourselves.⁹³ (Laursen 1997, 46–7)

I very tentatively suggest that one idea underlying this gap-ridden passage is as follows: when other people say something to us (i.e. something that can be thought), very fine atomic structures enter our mind from outside and leave it again,⁹⁴ being thought by us only briefly while they were there, and at first they may leave hardly a trace in the mind. But this may happen repeatedly with the same sort of thought, and thus the thought can be understood better each time. It will be connected with other thoughts, which ‘pull it along’ as Epicurus puts it, so that when we think one thing, we think another also (i.e. some sort of association theory of learning). In part owing to ourselves as causes (perhaps as focusing on it or connecting it with other thoughts that are our own already), the new thought is, it seems, eventually anchored in the mind and becomes part of it (by leaving a durable impression, and thus having changed the atomic structure of the mind permanently).⁹⁵ The thought has thus become a belief of ours, and at the same time our behavioural dispositions have changed.⁹⁶ It is in this light, I suggest, that one has to understand the method of philosophizing and teaching philosophy in the Epicurean school. The practice of memorizing the canon of Epicurean philosophy by repeating it again and again to oneself and others is,⁹⁷ on this interpretation, in no sense a ‘mindless’ enterprise. The repetition is meant to increase one’s understanding of new beliefs (especially those that are incompatible with the beliefs one has so far held) and thus to increase the firmness with which one holds those beliefs (they have to make a ⁹² Or loss of flaccidity: the idea may be that when the soul gains in firmness and structure, because more and more (hopefully correct) thoughts are being adopted and integrated as beliefs, the more easily a new thought is interconnected with these and retained and understood. ⁹³ . . . ψόφων τε καὶ νοήσεων καὶ ἐπινοημάτων καὶ φαντασμάτων καὶ τὴς αἰωνίας κατὰ ψυχὴν ὀχλήσεως ἢ εὐδαιμονίας ἢ μὴ αἰωνίας τὴν αἰτίαν τοῦ θηρεύειν τὴν ἀρχὴν καὶ κανόνα καὶ κριτήριον καὶ κατὰ μικρόν. ταῦτά τε γὰρ εἰς τὸν ἐπιλογισμὸν τοῦ κριτηρίου ἦγεν καὶ ἐκ τοῦ κριτηρίου αὐτ.. ἐπαισθα[5/6 ἐ]πιλογι . . . εἰς τὴν κατὰ μικρὸν ὥνπερ ἔνπροσθεν εἶπα διερεύνησιν. ἀλλήλοις γὰρ ταῦτα τὴν αἰτίαν καὶ χρείαν παρείχετο καὶ ἐναλλὰξ ἑκάτερον παρενπίπτον ἐπεσπάσατο εὐθὺς τὸ ἕτερον ἐπινόημα κατὰ μικρὸν πρῶτον ἐγγεινόμενον καὶ ταχέως ἐκρέον, εἶτα μᾶλλον μᾶλλον κατανοούμενον, τὰ μὲν διὰ τὴν φυσικὴν αἰτίαν τῆς ἀπαυξήσεως καὶ ἀπαλλάξεως πλαδαρότητος, τὰ δὲ διὰ τῆν ἐξ ἡμῶν γεινομένην καὶ . . . ⁹⁴ Atoms are not mentioned in so many words in the passage. However, Epicurus’ talk of thoughts ‘coming in’ (i.e. into the soul, which consists of atoms) and ‘flowing out’, and of ‘growth’ and ‘loss of fluidity’, suggests that atomic structures are at issue (cf. also Lucr. 3.510–16). This is in any case what we would expect in line with Epicurus’ materialist view of the soul—except perhaps adherents of Sedley’s emergentist interpretation of Epicurus’ psychology (e.g. Long and Sedley 1987, vol. 1, 109–11), an interpretation to which I have provided an alternative above in section 2 and, in more detail, in Bobzien 2000, chapter 6 in this volume. ⁹⁵ Cf. also Ep. Pyth. 85, ‘firm belief ’ (πίστιν βέβαιον). ⁹⁶ Cf. also Lucr. 3.510–16 for the possibility of change of our mind: such change involves either adding or taking away atoms, or changing the structure of the present atoms. ⁹⁷ Cf. e.g. Epic. Ep. Hdt. 35–6.

150

, ,   

‘lasting impression’ in the mind, quite literally).⁹⁸ Similarly, the Epicurean practice of producing a number of different arguments to prove the same point becomes comprehensible in this way. (Recall the twenty-nine or so proofs for the mortality of the soul in Lucretius, Book 3.) For the new beliefs have to be integrated and harmonized with one’s other beliefs—and here different arguments for them will lead to connections with different beliefs. All this is needed since only if one firmly holds the new beliefs will they be able to result in a change of one’s desires and emotions, and thus lead to a change in what actions one tends to perform.⁹⁹

4. Concluding Remarks To conclude, I return briefly to the subject of moral responsibility. From what we have seen about Epicurus’ ideas of moral development and moral progress, it is clear that at any time the state of an adult human’s moral dispositions is only in part the result of their own critical restructuring of their mind. We are not causally responsible for the initial constitution of our mind, but this constitution will determine, for instance, whether our moral progress will be faster or slower, effortless or arduous. Similarly, we are not causally responsible for the surroundings we find ourselves in initially, nor are we ever completely causally responsible for our environment; but this environment, too, will be a causal factor in our moral development. Consequently, both our nature and our past and present environment will—indirectly—co-determine what sort of actions we perform, and hence also whether we should be blamed or praised for our actions. Thus not only will the calm-natured offspring of a family of Epicureans growing up within a circle of Epicurean friends have a much smoother and shorter path to a life of tranquillity than the fire-natured, irascible youth who grew up in a society that indulges in luxury and fervently teaches religious superstition, but the latter will also be likely to deserve blame for a much larger proportion of their actions. This is a fact about which we find no sign of worry in Epicurean writings, nor, as far as I am aware, in any other philosophical writings from antiquity.¹⁰⁰ Epicurus simply seems to accept that blame is attached to those actions of moral beings of which they are themselves the main causal factor, regardless of how far the overall disposition of their mind is a result of their rational reflection and belief, and how far the result of necessitating factors.¹⁰¹ The agents can be ⁹⁸ The many repetitions in Lucretius’ On the Nature of Things may be intended to serve the same purpose. ⁹⁹ For the importance of reason in the process of moral improvement cf. e.g. Lucr. 3.321, ‘reason’ (ratio); Epic. Ep. Men. 124–5, ‘correct understanding’ (γνῶσις ὀρθή); 132, ‘reasoning’ (λογισμός) and DL 10.117 and 120. ¹⁰⁰ Cf. Bobzien 1998, section 6.3.6. ¹⁰¹ He does, though, make the interesting distinction between ‘respect’ one should have for people who approach tranquillity with difficult starting positions, and congratulations to those who had it easy; cf. Sen. Ep. 52.4.

     

151

blamed for their actions because they were not in any sense forced to bring them about and because the actions are not the outcome of a ‘mindless’ co-operation of nature and nurture. Rather, they result from the agents’ own beliefs, which in turn determine their desires and emotions. What is important here is the fact that, for Epicurus, ethics does not have the function of developing or justifying a moral system that allows for the effective allocation of praise and blame. The function of ethics—and in fact of the whole of philosophy¹⁰²—is rather to give everyone a chance to morally improve, that is, a chance to understand that in order to reach true happiness one has to learn to distinguish between pleasures conducive to that end and pleasures distracting from it; and in the course of this, to give up prejudicial and irrational beliefs that one has unthinkingly absorbed from the social surroundings one lives in. Epicurean ethics is, thus, exclusively forward-looking. It takes praise and blame for actions as in principle justified, based on the rationality of the agent. But praise and blame are not themselves a topic of ethics.¹⁰³ Human failure is taken into account only as a starting point for moral progress towards a life of happiness and tranquillity.

¹⁰² SE M 11.169; Epic. Ep. Pyth. 85; Sent. 11–13. ¹⁰³ The role of the swerve is not to justify moral praise and blame for individual actions, but to make it possible that our mental constitution, and in particular our beliefs, can change in response to environmental influences; see Bobzien 2000, chapter 6 in this volume.

6 Did Epicurus Discover the Free-Will Problem? 1. Introduction In 1967 Epicurus was credited with the discovery of the problem of free will and determinism.¹ Among the contestants were Aristotle and the early Stoics. Epicurus emerged victorious, because—so the argument went—Aristotle did not yet have the problem, and the Stoics inherited it from Epicurus. In the same year 1967, David Furley published his essay ‘Aristotle and Epicurus on Voluntary Action’, in which he argued that Epicurus’ problem was not the free-will problem.² In the thirty-odd years since then, a lot has been published about Epicurus on freedom and determinism.³ But it has only rarely been questioned again whether Epicurus, in one way or another, found himself face to face with some version of the free-will problem.⁴ In this essay I intend to take up the case for those who have questioned the point, combining a fresh perspective on the debate with a selection of new arguments and a detailed textual analysis of the relevant passages. I begin with a brief sketch of the problem of freedom and determinism with which Epicurus is widely taken to have been concerned. The determinism Epicurus defends himself against is usually understood as causal determinism: every event is fully determined in all its details by preceding causes. These causes are commonly pictured as forming an uninterrupted chain or

This essay first appeared in Oxford Studies in Ancient Philosophy 19 (2000), 287–337, and is reproduced here with permission. Versions of this essay have been presented at Cambridge, Bern, Yale, and Cornell universities. I am grateful to members of my audiences for stimulating discussion; and I thank David Sedley, Tad Brennan, Gail Fine, Terry Irwin, and especially Charles Brittain for helpful written comments on earlier drafts of this essay. ¹ By Huby 1967, 353–62. ² Furley 1967. ³ For example, Annas 1992, ch. 8; Asmis 1970; 1990, 275–91; Avotins 1979, 95–100; 1980, 75–9; Conway 1981, 81–9; Englert 1987; Fowler 1983, 329–52; Gulley 1990, 37–52; Huby 1969, 17–19; Kleve 1980, 27–31; Laursen 1988, 7–18; 1990, 3–22 (21–2); 1992, 143–54; 1995, 5–109; 1998, 5–82; Long 1974, 56–61; Long and Sedley 1987, vol. 1, 102–12; vol. 2, 104–13 [Long and Sedley] and [LS]; Mitsis 1988, ch. 4; Purinton 1999, 253–99; Saunders 1984, 37–59; Sedley 1983, 11–51; 1988, 295–327; Sharples 1991, 172–7, 195; 1991–3, 174–90. ⁴ Besides Furley there is Conway 1981. Authors have however become more careful on this point, see e.g. Mitsis 1988 and Laursen’s publications (see n. 3).

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0007

    -  ? 153 network, reaching back infinitely into the past, and as governed by an allembracing set of laws of nature, or as manifestations of such a set of laws of nature. On the side of freedom, Epicurus is generally understood to have been concerned with freedom of decision (the freedom to decide whether or not to do some action) or freedom of choice (the freedom to choose between doing and not doing some action)⁵ or freedom of the will (where the freedom to will to do something entails the freedom to will not to do it, and vice versa; I call this twosided freedom of the will). Epicurus is taken to have introduced an indeterminist conception of free decision or free choice or two-sided free will: agents are free in this sense only if they are causally undetermined (or not fully causally determined) in their decision whether or not to act or in their choice between alternative courses of action; undetermined, that is, by external and internal causal factors alike. There is assumed to be a gap in the causal chain immediately before, or simultaneously with, the decision or choice, a gap which allows the coming into being of a spontaneous motion. In this way every human decision or choice is directly linked with causal indeterminism. The assumption of such indeterminist free decision, free choice, or two-sided free will does not presuppose that one specifies an independent mental faculty, like, for example, a will, and indeed it is not usually assumed that Epicurus’ theory involved such a faculty. The ‘free-will problem’ that Epicurus is assumed to have faced is then roughly as follows: if determinism is true, every decision or choice of an agent between alternative courses of actions is fully determined by preceding causes and forms part of an uninterrupted causal chain. On the other hand, if an agent has (twosided) freedom of the will, it seems that the agent’s decision or choice must not be fully determined by preceding causes. Hence, it appears, determinism and freedom of the will (freedom of decision, freedom of choice) are incompatible.⁶ I do not believe that Epicurus ever considered a problem along the lines of the one just described. In particular, I am sceptical about the assumption that he shared a conception of free decision or free choice akin to the one I have sketched. (I also have my doubts that he ever conceived of a determinism characterized by a comprehensive set of laws of nature; this is a point I only mention in passing.) To avoid misunderstandings, I should stress that I do believe that Epicurus was an indeterminist of sorts—only that he did not advocate indeterminist free decision or indeterminist free choice. Why do I surmise that Epicurus did not deal with a concept of free decision or choice? I have first a circumstantial reason, as it were: elsewhere I argue that neither Aristotle nor the early Stoics nor any contemporaries of Epicurus were

⁵ Or alternatively to choose between doing one action or another. ⁶ This is the kind of free-will problem Giussani 1896, 125–69 and Bailey 1928, 318–23, 433–7 connected with Epicurus, and which the majority of more recent scholars take him to have dealt with, so e.g. Asmis 1990; Gulley 1990; Huby 1967; Purinton 1999; Sedley 1983; Sharples 1991–3.

154

, ,   

concerned with such a concept.⁷ However, in this essay I want to focus exclusively on Epicurean philosophy. I argue—negatively—that nothing in Epicurus’ extant writings and in our other sources for Epicurus’ philosophy provides any unambiguous evidence that Epicurus ever discussed a free-will problem as just described;⁸ and—positively—I suggest that Epicurus was caught in a related, but different, set of problems, for which we do have some direct evidence. One can best understand why Epicurus was concerned with a different problem if one realizes that he worked with different models of agency and moral responsibility than many modern scholars who see him as a proponent of indeterminist freedom of decision or choice take him to. The concept of indeterminist freedom that has been attributed to Epicurus presupposes that agents are free only if they are undetermined (or not fully determined) in their decision or choice by external and internal factors alike. The stipulation of the absence of these determining factors is typically linked with a model of agency that is based on the following distinction: on the one hand, there are the possible causal influence factors, some external to me (the agent), like the environment, some internal to me, such as my character dispositions. On the other hand, there is myself, the one who decides, causally detached not only from external impacts but also from my past experiences, from my present character and dispositions, from my desires and inclinations, perhaps even from my memories and factual beliefs. I, the one who decides, am thus identified with some sort of a decision-making faculty which is detached from the rest of ‘me’, where this rest includes my character and my present mental dispositions. This decision-making faculty has control over all the other factors and can in principle decide against any one of them. Thus the entity that makes decisions is not identified with the whole person (body, soul, and mind), nor even the whole mind, but is causally independent of large parts of that person or of that mind. I call this the independent-decision-faculty model of agency. It is often, and sometimes explicitly, taken to underlie Epicurus’ philosophy.⁹ This model and the concept of free decision or choice on which it is based are commonly linked with the following conception of moral responsibility: I can be held morally responsible for an action of mine only if I could have decided or chosen to act otherwise, or if I was free to decide or choose otherwise. This

⁷ For the Stoics see my 1998a, ch. 6. For Aristotle, the Peripatetics, the Middle Platonists, and the general development of the problem of free decision and causal determinism, see Chapter 1 in this volume [Bobzien 1998b, 133–75], Chapters 2–4, and Chapter 7. ⁸ Whether Epicurus discussed free will depends on what one means by ‘free will’. For example, if one intends ‘free will’ to render Lucretius’ ‘libera voluntas’ (see below, section 4), and to mean whatever element of Epicurus’ doctrine Lucretius meant to capture by this phrase, then Epicurus evidently was concerned with free will. My concern is only to show that he did not discuss a problem of free will that involves a conception of freedom of decision or choice as adumbrated in the main text. ⁹ Explicitly e.g. in Asmis 1990, 283; Sharples 1991–3, 178, 187–8; it is implicit in e.g. Bailey 1928, 435 and in Sedley 1983 at 49 and before.

    -  ? 155 conception of responsibility is typically backed up with the fact that people are generally convinced that they could decide and act otherwise than they do, and with suggestions that moral emotions like regret or guilt indicate that we are free to decide and act otherwise. Epicurus, I suggest, operated with a different conception of moral responsibility and with a different model of agency, a model that does not causally sever a person’s present character and mental dispositions from that person qua decision maker. Rather, Epicurus identified the agent with the person’s mind (or large parts thereof), including the person’s system of beliefs, memories, character dispositions, desires, and emotions. In this model, what volition a person forms, and what action they perform, fully depends on the overall disposition of the person’s mind at the time of forming the volition. It is considered essential for the causal attribution of an action to me as a person that it is the person as I am when I form the volition who is (at least in the main)¹⁰ causally responsible for what action is performed. I call this the whole-person model of agency.¹¹ In this model, moral responsibility for an action is typically not based on the causal undeterminedness of the agent’s decision but, on the contrary, on the fact that the action is dependent on the person’s present overall disposition of their mind. What makes an agent morally responsible, and an action voluntary, is what for convenience I shall call the autonomy of the agent. By ‘autonomy’ I intend to indicate nothing but the fact that a rational agent, and not something else, causes the action (or is at least its main cause), and that the agent is not compelled or forced to act. This second concept of moral responsibility is thus not grounded on a concept of freedom to decide or choose otherwise, but on a notion of agent autonomy.¹² Characteristically, in the whole-person model of agency, agents can influence their ways of acting and behaving only indirectly, by changing the basis of their actions, that is, their mental dispositions. This is so since their volitions and actions are a function of these dispositions. There is in this model no space for indeterminist freedom of decision. For the fact that agents act in accordance with their overall mental disposition is considered a necessary condition for attributing the action to them, whereas indeterminist freedom of decision takes the detachment of the decision-making faculty from (most of) the rest of the person’s mind as a necessary condition. The concept of an internally undetermined decision made by the agent would be difficult to incorporate in this conceptual framework, and so would be the free-will problem in the form sketched above.

¹⁰ Epicurus allowed for a multiplicity of causes for human behaviour; see section 2 below. ¹¹ Two Lucretius passages could perhaps be adduced as providing some support for the assumption that Epicurus worked with the whole-person model of agency: ‘that is still nothing to us, who are constituted by the conjunction of body and spirit’ (Lucr. 3.845–6, tr. Long and Sedley) and ‘the mind and the spirit are firmly interlinked and constitute a single nature’ (Lucr. 3.136–7, tr. Long and Sedley). ¹² Evidence that Epicurus had this concept of moral responsibility is provided in section 3 below. For a more detailed discussion of Epicurus’ notion of moral responsibility see Chapter 5.

156

, ,   

In the whole-person model of agency the main problem that arises from a deterministic assumption is how, if everything, including my actions, were necessitated by something other than me, I, the agent, could still rightfully be considered as being myself causally and hence morally responsible for my actions. I call this is the problem of necessitation and agent autonomy. In the following I aim to show that all the surviving texts which have been adduced as evidence that Epicurus dealt with a free-will problem as set out above are in fact compatible with the assumption that Epicurus did not work with an independent-decision-faculty model of agency or a concept of moral responsibility based on freedom to decide otherwise; and that there is a reasonable amount of evidence that Epicurus worked with a whole-person model of agency and a concept of moral responsibility based on the idea of agent autonomy as just described; and accordingly, that his problem was not the one commonly assumed, but a problem of necessitation and agent autonomy. Why, then, is it such a popular view among philosophers that Epicurus dealt with a free-will problem as described above, or at least with some of its main aspects? Apart from less relevant factors—such as that it would be nice if Epicurus had anticipated the modern discussion of causal determinism and free decisionmaking, and even nicer if he had anticipated, with his swerve, the ‘solution’ some contemporary libertarians (e.g. David Wiggins) squeezed out of the theory of quantum mechanics—there is one important reason: many of the key expressions used in those Epicurean passages that are commonly held to be about indeterminist choice appear to be ambiguous in such a way that—although they never state the problem of free decision and causal determinism directly—they can be read as alluding to it. These expressions include in particular phrases and terms like ‘free volition’, ‘freedom’, ‘being without master’, ‘beginning of motion’, ‘moving oneself ’, and ‘that which depends on us’ (to par’hēmas). When, in the following pages, I analyse those passages that have been adduced as evidence that Epicurus dealt with indeterminist freedom of decision or choice, I shall, on the way, also spell out the ambiguities and vagueness in the key expressions involved.

2. To par’hēmas There is, in the extant Epicurean texts, no Greek term for freedom of decision, freedom to do otherwise, or free will.¹³ The Greek expression in Epicurean sources which is most often regarded as an indication that some such freedom is at issue is par’hēmas. This expression occurs in Epicurus’ Letter to Menoeceus 133–4, and we find it twice in one sentence in Epicurus’ On Nature 25. (It also appears in some ¹³ ἐλευθερία, in Sent. 77 (τῆς αὐταρκείας καρπὸς μέγιστος ἐλευθερία) does denote neither freedom of decision nor—two-sided—freedom of the will.

    -  ? 157 later Epicurean texts. The related expression eph’hēmin does not occur in surviving passages by Epicurus.) Both passages from Epicurus have been understood as dealing with freedom of choice. However, I believe that this interpretation is based on an inadequate understanding of the expression par’hēmas. The expression could be taken either as what I call ‘causative one-sided’ or as what I call ‘potestative two-sided’ (see Chapter 1), and only the former does the passages justice whereas only the latter could provide support for the view that Epicurus was concerned with free choice. Let me explain the difference between these two ways of understanding the phrase par’hēmas. The interpretation of par’hēmas as two-sided potestative commonly assumes a verb of being (rather than of happening or becoming) to go with the phrase: something is par’hēmas. When I call par’hēmas ‘two-sided’, I mean that if some action, for example walking, is said to be para se, this can be read as short for ‘it is para se whether or not you walk’. On this understanding of the expression, the class of things that are par’hēmas includes precisely 50 per cent unrealized possibilities. Thus, when at a certain time it is para se whether you walk, then it is para se whether you don’t walk, as well; but you will be able only either to walk or not to walk at that time—not both. Hence either one or the other will remain unrealized. By ‘potestative’ I mean that if some action or occurrence is understood as par’hēmas, we are taken to have some sort of power to bring it about or not bring it about. A good translation of par’hēmas in its two-sided, potestative reading would be ‘up to us’. This two-sided, potestative understanding of par’hēmas entails neither determinism nor indeterminism.¹⁴ But it is often understood as indeterminist, in the following way: it is assumed that walking is up to me at a certain time if at that time whether or not I will walk is causally undetermined (not fully determined) and depends on my free decision. Epicurus is commonly interpreted as using the expression par’hēmas both as two-sided potestative and as indeterminist.¹⁵ When the expression is understood in this manner, the ‘we’ or ‘us’ (hēmas) in it is given the status of an active decision maker. We can decide freely whether or not we walk (or choose freely between walking and not walking). This is quite different in the case of what I call the one-sided, causative par’hēmas. This interpretation of par’hēmas usually assumes a verb of happening or becoming (rather than of being) to go with the phrase: something happens or comes to be par’hēmas. By ‘causative’ I mean to indicate the fact that the

¹⁴ For a determinist reading compare what I say about the related expression ἐφ᾽ἡμῖν in Chapter 1 [Bobzien 1998b, 143–4] or in Bobzien 1998a, 281–2. ¹⁵ For a two-sided, potestative reading, see Long and Sedley 1987, vol. 1, 20C(1) (tr.) ‘that which we develop is up to us’ (γίνεσθαι παρ᾽ἡμᾶς), interpreted as ‘it is up to us whether or not we develop a particular characteristic c’ or ‘whether or not we develop c or not c’; further Annas 1992, 129 n. 18; Englert 1987, 129, who assumes that τὸ ἐφ᾽ἡμῖν and τὸ παρ᾽ἡμᾶς are equivalent; similarly Purinton 1999, 261–2.

158

, ,   

prepositional phrase ‘para y’ in ‘x happens para y’ refers to something which is a (main) cause or reason of x. Thus here translations such as ‘because of us’ or ‘due to us’ are preferable. When I call the phrase ‘one-sided’, I mean that if an action happens (comes to be) par’hēmas then its opposite does not happen (come to be) par’hēmas. If at a certain time my walking happens because of me, then it is not the case that my not walking happens because of me, too, for in that case my not walking does not happen at all. Here the natural understanding of the phrase is that it expresses who has the causal responsibility for the action in question. ‘The walking happens para se (because of you)’ can be paraphrased as ‘you are the cause of your walking’. When understood as one-sided causative, ‘whether-or-not’ paraphrases of par’hēmas are impossible. The general idea is that, when sifting through the things that actually happen, one distinguishes: well, this came to be by necessity, and that happened because of you, and so forth. The one-sided, causative par’hēmas, too, can be made use of in an indeterminist as well as a determinist theory. However, whereas the two-sided par’hēmas can be used to express an element of undeterminedness, by implying that we, qua decision makers, can freely decide between alternative options, the one-sided par’hēmas cannot be so used. Its function is rather to indicate who bears the causal responsibility for an event. It does not imply the possibility of free choice. (The one-sided, causative par’hēmas is however compatible with the assumptions that we can be the cause of our actions only if, say, we are not forced in our action, or if we have a general two-sided capacity for acting and not acting, or if it is in some sense possible for us—or up to us, eph’hēmin—both to act and not to act. I return to this point at the end of section 3.) Now, it is my view that in Epicurus, and quite generally in Hellenistic philosophy in the context of determinism and moral responsibility, the phrase par’hēmas was commonly understood as one-sided and causative. There are several good reasons for assuming this. First, in all the Epicurean (and related) texts in which par’hēmas is connected with a verb, this verb is always ‘gignesthai’, never ‘einai’, and thus fits the causative understanding better. It is true, gignesthai can occasionally mean ‘to be’. However, ‘to come to be’ or ‘to happen’ usually fit the context better text passages, and the consistency in the use of gignesthai rather than einai also suggests this latter meaning. (It is also, I believe, a more common use of para with personal accusative.) Second, par’hēmas occurs repeatedly as one of a triad of expressions for kinds of causal factors. Take first the above-mentioned passage from the Letter to Menoeceus: Whom do you believe to be better than the one . . . who would denounce , which some introduce as mistress of all things, others by chance, and others again because of us, since necessity is not accountable to anyone, and chance is an unstable thing to watch, whereas that because of us is without master, and culpability and its opposite are naturally

    -  ? 159 attached to it; for it would be better to follow the myths of the gods than to be enslaved by the fate of the natural philosophers. For the one indicates some hope for pardon from the gods if we honour them, but the other comes with inexorable necessity.¹⁶ (Epic. Men. 133–4)

In this passage par’hēmas is contrasted with the expressions ‘by necessity’ (kat’anagkēn) and ‘by chance’ (apo tuchēs). These latter two expressions are both used to refer to causes (i.e. to that because of which some things happen), and as one-sided: if something happens by necessity, then its opposite does not happen; and again, if something happens by chance, then its opposite does not happen. Since par’hēmas is co-ordinated and treated as on a par with these two expressions, it is natural to infer that it has the same function of establishing the cause of an actual event. And this implies that it, too, is used as causative and one-sided.¹⁷ This finds support in the fact that in Stobaeus’ report of Epicurus’ tripartition, par’hēmas has been substituted by the Peripatetic expression kata prohairesin and the three expressions are taken to denote so many kinds of being caused: . . . many causes: choice, chance, necessity. Epicurus: in accordance with necessity, in accordance with choice (prohairesis), in accordance with chance.* (Stob. Ecl. 1.4–5, Diels, Doxogr. graec. 326.1–4)

Moreover, in On Nature 25 we find Epicurus talking about tēn kath’hēm[as] . . . ai [tian] (roughly ‘the through-us cause’).¹⁸ Similarly, the later Epicurean Diogenianus speaks about tēn par’hēmas aitian (roughly ‘the because-of-us cause’),¹⁹ and he uses the phrase par’hēmas as one-sided causative throughout.²⁰ ¹⁶ ἐπεὶ τίνα νομίζεις εἶναι κρείττονα τοῦ . . . τὴν . . . ὑπό τινων δεσπότιν εἰσαγομένην πάντων αγγέλλοντος ἃ δὲ ἀπὸ τύχης, ἃ δὲ παρ᾽ἡμᾶς, διὰ τὸ τὴν μὲν ἀνάγκην ἀνυπεύθυνον εἶναι, τὴν δὲ τύχην ἄστατον ὁρᾶν, τὸ δὲ παρ᾽ ἡμᾶς ἀδέσποτον ᾧ καὶ τὸ μεμπτὸν καὶ τὸ ἐναντίον παρακολουθεῖν πέφυκεν; ἐπεὶ κρεῖττον ἦν τῷ περὶ θεῶν μύθῳ κατακολουθεῖν ἢ τῇ τῶν φυσικῶν εἱμαρμένῃ δουλεύειν˙ ὁ μὲν γὰρ ἐλπίδα παραιτήσεως ὑπογράφει θεῶν διὰ τιμῆς, ἡ δὲ ἀπαραίτητον ἔχει τὴν ἀνάγκην. αγγέλλοντος: cf. Herodian 5.2.2: δοῦλοι ὅσοι δεσπότας κατήγγελον; ἂν γελῶντος, Long and Sedley; διαγελῶντος, Usener. : Usener (cf. also VS 40, where γίγνεσθαι κατ᾽ ἀνάγκην occurs three times); , Long and Sedley. ¹⁷ The tripartition is also found in SE M 5.46: τῶν γινομένων τὰ μὲν κατ᾽ἀνάγκην γίνεται τὰ δὲ κατὰ τύχην, τὰ δὲ παρ᾽ ἡμᾶς . . . The passage M 5.46–8 may well be an argument put forward by later Epicureans. * . . . αἰτιῶν ποικίλων προαιρέσεως, τύχης καὶ ἀνάγκης. Ἐπίκουρος κατ᾽ ἀνάγκην, κατὰ προαίρεσιν, κατὰ τύχην. ¹⁸ Laursen 1995, 99. 1056 corn. 1 pz. 3 z. 1 col. 2 = Arr. 34.7. This is Laursen’s reading of the papyrus. ¹⁹ Praep. Ev. 6.8.34. Cf. also [Plut.] Epit. 27.3 (Diels, Doxogr. graec. 322.5–8) for Plato: τὴν παρ᾽ ἡμᾶς αἰτίαν. ²⁰ Praep. Ev. 6.8.2, 6, 23, 30 (with γίνεσθαι); 6.8.32 (with συμβαίνειν). A passage in [Plutarch] ([Plut.] Epit. 27.4; Diels, Doxogr. graec. 322.9–14) suggests that the Stoics used παρ᾽ ἡμᾶς as onesided causative. The formulation in Philodemus, Sign. 36 is neutral with respect to a one-sided or two-sided reading.

160

, ,   

A strong reason for the assumption that Epicurus used par’hēmas as one-sided causative is finally provided by the way it occurs in Epicurus’ On Nature 25: Consequently that which we develop (characteristics of this or that kind) comes to be at some point absolutely because of (para) us; and the things which of necessity flow in through our passages from that which surrounds us, at one point come to be because of (para) us and because of (para) the beliefs of ours which are from us ourselves.²¹ (Epic. Nat. 25; Arr. 34.26, Laursen 1997, 33)

In its last part this sentence contains the following parallel construction: . . . the things which of necessity flow in . . . from our surroundings come to be because of (para) us and because of (para) the beliefs of ours which are from us ourselves.

I take it that para (with accusative) has the same meaning in both phrases, since they are syntactically co-ordinated. But in the phrase ‘para our beliefs’ we cannot construe para as two-sided potestative. We cannot understand our beliefs as a decision maker in the way we can see ourselves; we cannot paraphrase ‘our beliefs make a decision as to whether or not we develop characteristic c’. Our beliefs cannot actively decide anything. They are specific beliefs which we actually have, and if something depends on them, that can here only mean that they determine that we develop one way and not the other—not whether we develop one way or the other. But if para is one-sided causative in the case of our beliefs, it should also be one-sided causative in the case of par’hēmas.²² The interpretation of par’hēmas as one-sided causative obtains additional confirmation from the context in which the sentence in On Nature 25 belongs. It is the last sentence before a gap in the papyrus. The question discussed by Epicurus that leads up to the sentence (Arr. 34.24–5, Laursen 1997, 29–31) is whether we are morally responsible for our bad actions if there is a correlating badness in our original constitution (i.e. the one we are born with). Epicurus’ answer is that if we act through our initial disposition we cannot be held responsible, but if, when pursuing the bad actions, we ourselves as we developed (presumably once we reached adulthood) are the cause, then we can be morally criticized, even if what we do is in line with our initial constitution. I assume that in the above-quoted sentence that contains the expression ‘par’hēmas’ Epicurus is still in some way concerned with this same question. Thus the surrounding ²¹ ὥστε παρ᾽ἡμᾶς ποτε ἁπλῶς τὸ ἀπογεγεννημένον ἤδη γείνεσθαι, τοῖα ἢ τοῖα, καῖ τὰ ἐκ τοῦ περιέχοντος κατ᾽ἀνάγκην διὰ τοὺς πόρους εἰσρέοντα παρ᾽ἡμᾶς π[ο]τε γε[ίνε]σθαι καὶ παρὰ τὰς ἡμετέρας ἐξ ἡμῶν αὐτῶν δόξ[ας . . . (Epic. Nat. 25; Laursen 1997, 33). ²² Contrast Long and Sedley, who translate παρ᾽ἡμᾶς as ‘up to us’ and παρὰ τ[ὰς] ἡμε[τέρα]ς [ἐ]ξ ἡμῶν αὐτ[ῶν] δόξ[ας as ‘dependent on beliefs of our own making’ (Long and Sedley 1987, vol. 1, 20C(1)).

    -  ? 161 context of our sentence does not deal with alternative choices, or with a decisionmaking power of any sort (nor do any other passages from On Nature 25, as far as I can see). Rather, the context is human mental and, especially, moral development, and the general question is what or who is causally responsible for our dispositions and actions (see Chapter 5). The occurrence of the phrase ‘par’hēmas’ in Epicurean writings is then no evidence for the claim that Epicurus discussed the free-will question as set out in section 1, since the phrase is used by him not as two-sided and potestative but as one-sided and causative.

3. The Digression in On Nature 25 Next I consider the so-called digression in Epicurus’ On Nature 25 (Arr. 34.27–30, Laursen 1997, 35–42; LS 20C(2–15); I follow Laursen’s edition of the text). In it Epicurus digresses from the book’s main topic of human psychological development in order to refute the necessitarian views of an opponent. This passage has sometimes been taken to be concerned with the topic of the compatibility of determinism with free choice. It contains a series of arguments against a philosophical view that is incompatible with Epicurus’ own. The structure and philosophical strength of these arguments have been analysed in detail by David Sedley.²³ Here I primarily intend to follow up the question of what philosophical problems Epicurus is concerned with in the passage and what information the passage provides on his concepts of action and moral responsibility. The main thesis of Epicurus’ opponent was that everything is caused by necessity, or that everything is necessitated (LS 20C(5 and 13); Laursen 1997, 36, 41). Thus his is a position of universal necessitation. The repeated use of phrases such as hē kata to automaton anagkē (LS 20C(2, 3); Laursen 1997, 35, cf. 41) makes it clear that the mechanical necessitation of the atomists is at issue, and not teleological predetermination. Necessity is not some mysterious divine power; it is broken down into various causal factors. In the case of human behaviour, the two factors that are explicitly mentioned are our initial congenital constitution (hē ex archēs sustasis), as necessitating us internally, and the mechanical external necessity of that which happens to surround us and which we perceive (LS 20C(2–3); Laursen 1997, 33, 35). Epicurus’ opponent believes that our actions are a function of necessitating hereditary and environmental factors. By contrast, Epicurus concedes that some occurrences are necessary,²⁴ but maintains that we are the cause of our actions and that they are not necessary. He takes it for granted that the same action cannot be caused both by necessity ²³ Sedley 1983. ²⁴ See also Epic. Men 133–4 (quoted above) and Epic. Nat 25, Arr. 34.24 (discussed below).

162

, ,   

and by ourselves. The reason for this seems to be that he construes necessity in terms of compulsion or force (e.g. LS 20C(10)), and considers it a necessary condition for us to be the cause of our actions that we are not compelled. Epicurus believes that his opponent faces the problem that in his theory he cannot guarantee the causal responsibility of the agents for their actions. In all the arguments of the digression the underlying question is who or what is the cause of, or causally responsible for, human action. Epicurus considers two candidates: necessity on the one hand, us ourselves on the other (LS 20C(2, 5), Laursen 1997, 35, 37). Where Epicurus uses the terms aitia, aitios, and aitiasthai, they have sometimes been taken as denoting moral responsibility. However, the surviving parts of Epicurus’ On Nature 25 (Laursen 1995 and 1997) suggest that Epicurus used the expressions aitia, aitios, and aitiasthai indistinguishably in order to express the causal responsibility of something, or its being a cause. For example, they are all also used of things such as atoms, the environment, our nature, etc. This suggests strongly that their meaning is not that of having moral responsibility.²⁵ Epicurus talks of moral responsibility in terms of praise and blame and similar evaluative expressions (see below, this section). The opponent’s problem, as Epicurus presents it, is then not that he is unable to accommodate in his theory that human actions and decisions must be causally undetermined, or that we can choose freely between alternative courses of actions. Rather, Epicurus and his opponent seem to share the assumption that my actions are caused and in that sense determined. The opponent’s problem is that on his theory it is difficult to see how I myself can be causally responsible for my behaviour, since he claims that something else, i.e. necessity in the form of hereditary and environmental factors, is fully causally responsible for it. I quote the text in chunks, with some comments interspersed, in order to make it apparent that it does not deal with free choice or free decision, but with the question of whether we, the agents, or necessity causes our actions. (Suspension points indicate lacunae in the text; all emphases are mine; the Greek text is Laursen’s but I have kept Long and Sedley’s numbering for convenience.) (2) ²⁶ by which we never cease to be affected, [the fact that] we rebuke, oppose, and reform each other as if the cause lay also in ourselves, and not just in our initial make-up and in the ²⁵ E.g. τὴν ἀνάγκην . . . πάντα αἰτιᾶσθαι, Laursen 1997, 41; LS 20C(13); καὶ τὴν αὐτὴν ἀμφότερα κέκτηται μὲν αἰτίαν . . . ἡ ἀρχὴ εἶχε τὴν αἰτίαν, εἴχομεν δὲ καὶ ἡμεῖς (Laursen 1997, 43); διὰ τὴν ἐξ ἡμῶν γεινομένην αἰτίαν . . . διὰ τὴν φυσικὴν αἰτίαν (Laursen 1997, 46–7). ²⁶ Long and Sedley add ‘.’ But nothing in the text suggests that choice is at issue.

    -  ? 163 accidental necessity of that which surrounds and penetrates us.²⁷ (3) For if someone were to attribute to the very processes of rebuking and being rebuked the mechanical necessity and always . . . understand . . . (4) . . . when he blames or praises. But if he were to act in this way, he would be leaving intact the very same behaviour [i.e. praising and blaming] which we think of as concerning ourselves, in accordance with our preconception of the cause;²⁸ and he would have changed the name [only]. (5) . . . so great an error. For this sort of account is self-refuting, and can never prove that everything is of the kind which we call ‘by necessity’; but he debates this very question on the assumption that his opponent talks nonsense on account of himself. (6) And even if he goes on to infinity saying that, again, he does this action of his by necessity, always appealing to arguments, he is not reasoning it empirically so long as he goes on imputing to himself the cause for having reasoned correctly and to his opponent that for having reasoned incorrectly. (7) But unless he were to stop attributing his actions to himself, and to pin it on necessity instead, he would not even . . .²⁹ (Epic. Nat. 25; Arr. 34.27–8, Laursen 1997, 35–7)³⁰

In this section, the opponent faces the charge that he pragmatically refutes himself when he argues his position of universal necessitation. Epicurus provides him with the alternatives of either attributing causal responsibility to himself and his interlocutor and not to necessity—because the opponent in fact attributes evaluative criticism to his and his interlocutor’s verbal acts—or giving up his argument. Epicurus’ main argumentative step is this: when someone evaluates a person’s acts morally or veridically, they implicitly attribute causal responsibility to that person for that act. Freedom of decision or free choice are not involved in the argument; nor is freedom to do otherwise explicitly mentioned anywhere. This passage also provides important insight into Epicurus’ concept of moral responsibility. Epicurus takes the fact that we blame each other, and try to reform ²⁷ Namely by means of perception, cf. Epic. Nat. 25, Laursen 1997, 33; LS 20C(1), quoted above. ²⁸ I follow Laursen’s reading (cf. next footnote, sentence (4)). However, I do not quite understand what it means. I hope it still means the same as what David Sedley (1983) suggested, viz. that our observation of blaming and praising produces our preconception of us as causes of our actions. ²⁹ (2) ἐστήκει, ὧν οὐ . . . ἀπολείπει τὰ πάθη τοῦ γίνεσθαι . . . νουθετεῖν τε ἀλλήλους καὶ μάχεσθαι καὶ μεταρυθμίζειν ὡς ἔχοντας καὶ ἐν ἑαυτοῖς τὴν αἰτίαν καὶ οὐχὶ ἐν τῇ ἐξ ἀρχῆς μόνον συστάσει καὶ ἐν τῇ τοῦ περιέχοντος καὶ ἐπεισιόντος κατὰ τὸ αὐτόματον ἀνάγκῃ (3) εἰ γὰρ τις καὶ τῷ νουθετεῖν καὶ τῷ νουθετεῖσθαι τὴν κατὰ τὸ αὐτόματον ἀνάγκην προστιθείη καὶ ἀεὶ τοῦ ποθ᾽ἑαυτῷ ὑπάρχοντος . . . συνιέναι . . . (4) μεμφόμενος ἢ ἐπαινῶν˙ ἀλλ᾽εἰ μὲν τοῦτο πράττοι, τὸ μὲν ἔργον ἂν εἴη καταλεῖπον ὃ ἐφ᾽ἡμῶν αὐτῶν κατὰ τὴν τῆς αἰτίας πρόληψιν ἐννοοῦμεν, τὸ δ᾽ὄνομα μετατεθειμένος . . . (5) τοσαύτης πλάνης. περικάτω γὰρ ὁ τοιοῦτος λόγος τρέπεται, καὶ οὐδέποτε δύναται βεβαιῶσαι ὡς ἒστιν τοιαῦτα πάντα οἷα τὰ κατ᾽ἀνάγκην καλοῦμεν ἀλλὰ μάχεταί τινι περὶ αὐτοῦ τούτου ὡς δι᾽ἑαυτὸν ἀβελτερευομένῳ. (6) κἂν εἰς ἄπειρον φῇ πάλιν κατ᾽ἀνάγκην τοῦτο πράττειν ἀπὸ λόγων ἀεί, οὐκ ἐπιλογίζεται ἐν τῷ εἰς ἑαυτὸν τὴν αἰτίαν ἀνάπτειν τοῦ κατὰ τρόπον λελογίσθαι εἰς δὲ τὸν ἀμφισβητοῦντα τοῦ μὴ κατὰ τρόπον. (7) εἰ δὲ μὴ ἃ ποιεῖ ἀπολήγοι εἰς ἑαυτόν ἀλλ᾽εἰς τὴν ἀνάγκην τιθείη, οὐδ᾽ἃν . . . ³⁰ In my rendering of the digression I have made use freely of Long and Sedley’s excellent translation, but have modified it in line with Laursen’s new readings of the text, and in some other places, in order to bring out my understanding of the text more clearly.

164

, ,   

each other, as an indication that the cause of our actions lies in ourselves, or that the actions happen through ourselves (LS 20C(2), cf. C(8)). The concept of blame presupposes that the beings that are blamed were themselves causally responsible for their behaviour. It makes no sense to blame individuals for certain events if those events came about through necessity (LS 20C(3)). There are several other passages that confirm that Epicurus based his concept of moral responsibility for an action on that of our causal responsibility for it: (i) Epicurus’ On Nature 25, Arr. 34.25 (Laursen 1997, 29), implies that if an action of a certain kind is caused by the initial constitution of a person in response to the environment, then the person is not to be blamed. However if an action of the same kind is caused by the person themself, and thus not (exclusively) by the initial constitution, then the person is to be blamed for it. (ii) In Epicurus’ Letter to Menoeceus 133, the things that happen because of us are said ‘to have praise and blame naturally attached to them’. With the above interpretation of that which happens because of us (par’hēmas), this suggests that when I am causally responsible for something happening, then I can be held morally responsible for it. (iii) Similarly, Epicurus, On Nature 25, Arr. 34.21 (Laursen 1997, 19–20, LS 20B(1–4)), suggests that if a person (or their ‘developments’) is causally responsible for something, then they can be held morally responsible for it.* Back to Epicurus ‘digression’: (8) . . . using the word ‘necessity’ of that which we call ‘ . . . by ourselves’, he is merely changing a name; but he must prove that we have a preconception of a kind which has faulty delineations when we call that which through ourselves causally responsible . . . ³¹ (Epic. Nat. 25; Arr. 34.28, Laursen 1997, 37)

Epicurus’ point in (8) is that we have a preconception that we are causally responsible for our actions by means of that which comes through ourselves, i.e. presumably our own beliefs (doxai, LS 20C(1)) and impulses (hormēmata, hormai) or endeavours (prothumiai, LS 20C(9–11)) (cf. also Chapter 5), and that his opponent is unable to show that we are mistaken about having this preconception. This argumentation is not a version of the modern one that we have the intuition, or know by introspection, that we could have acted or decided otherwise. The preconception is not that we have an ability to act or to decide otherwise, but that we (by means of our beliefs, impulses, and endeavours), and not something else, are causally responsible for what we do.

* These passages are discussed in more detail in Chapter 5. ³¹ (8) . . . ]φ ἡμῶν αὐτῶγ καλούμενον τῶι τῆς ἀνάγκης ὀνόματι προσαγορεύειν ὄνομα μόνομ μετατίθεται. δεῖ δ᾽ἐπιδίξ[α]ι ὅτι τοιοῦτό τι ᾧ μοχθηρ[οί εἰσι τύ]ποι προειληφότες τὸ δι᾽ἡμῶν αὐτῶν αἴτιογ καλοῦμεν, οὐτιδ[ . . .

    -  ? 165 (9) . . . but to call necessity as a result of your claim. If someone won’t show this, and has no auxiliary element or impulse in us which he might dissuade from those actions which we perform calling the cause of them ‘through ourselves’, but is giving the name of necessity to all the things that we endeavour to do in accordance with our position, calling the cause of them ‘through ourselves’, he will be merely changing a name. (10) He will not be modifying any of our actions in the way in which in some cases the one who sees what sort of things are necessitated usually dissuades those who endeavour to do something in the face of compulsion. (11) And the mind will be inquisitive to learn what sort of action it should then consider that one to be which is performed in some way out of ourselves through our endeavour to act. For he has nothing else to do but to say . . . ³² (Epic. Nat. 25; Arr. 34.29, Laursen 1997, 39)

The point of the argument (9)–(11) is in short this: if saying ‘our actions happen by necessity’ is to be more than just another way of saying ‘our actions happen through ourselves’, then our recognition of their necessity would have to change our behaviour, as our behaviour is generally changed when it is pointed out to us that something must necessarily occur. For then we will not endeavour to do anything to prevent the necessitated thing from occurring, because there is no point in doing so. Similarly, if our actions were necessitated, there would be no point in doing anything about them; in particular there would be no point in making an effort to bring them about. Thus Epicurus seems to envisage necessity as some kind of compulsion, and to presuppose that, if our actions are necessitated, they will happen even if we do not endeavour to bring them about. By contrast, if we are the causes of our actions, our endeavouring to act will be causally connected with the action itself.³³ Freedom of decision, free choice between alternatives, or freedom to do otherwise are not part of the argument. If any surviving passage from Epicurus deals with what Lucretius renders as ‘voluntas’ (volitional act), I suppose it is this one. In any case, we can extract from this passage what Epicurus regards as essential characteristics of human action. (i) Human action is not necessary, that is, it does not happen by force (bia). (ii) The agent has an impulse (hormēma) towards the action and endeavours

³² (9) ἀλλὰ κενὸν καὶ τὸ δι᾽ ἀνάγκην καλεῖν πρὸς ὧν φάτε. ἂν δὲ μή τις τοῦτο ἀποδείξει, μηδ᾽ ἔχει ἡμῶν τι συνεργὸν μηδ᾽ὅρμημα ἀποτρέπειν ὧν καλοῦντες δι᾽ ἡμῶν αὐτῶν τὴν αἰτίαν συντελοῦμεν, ἀλλὰ πάντα ὅσα νῦν δι᾽ ἡμῶν αὐτῶν ὀνομάζοντες τὴν αἰτίαν πως προθυμούμεθα πράττειν κατὰ χώραν ἀνάγκην προσαγορεύων, ὄνομα μόνον ἀμείψει; (10) ἔργον δ᾽οὐδὲν ἡμῶν μετακοσμήσει, ὥσπερ ἐπ᾽ ἐνίων ὁ συνορῶν τὰ ποῖα κατ᾽ἀνάγκην ἐστὶν ἀποτρέπειν εἴωθε τοὺς προθυμουμένους παρὰ βίαν τι πράττειν. (11) ζητήσει δ᾽ ἡ διάνοια εὑρεῖν τὸ ποῖον οὖν τι δεῖ νομίζειν τὸ ἐξ ἡμῶν αὐτῶν πως πραττόμενον τῇ προθυμίαι τοῦ πραττειν. Οὐ γὰρ ἔχει ἀλλ᾽ οὐθὲν πράττεν ἢ φάναι . . . ³³ This argument is reminiscent of the so-called Idle Argument (ἀργὸς λόγος, for which see chapter 5 of Bobzien 1998a); cf. the presentation and criticism of the Stoic refutation of the Idle Argument by the Epicurean Diogenianus for the emphasis on προθυμία and σπουδή (Eus. Praep. Ev. 6.8.25, 29, 30) and on us as causes (ibid. 6.8.34, 38).

166

, ,   

(prothumeisthai, prothumia) to perform it. Impulse (hormēma) or endeavour (prothumia) are the two candidates for Lucretius’ voluntas. I assume that physically they are motions in the agent’s mind that are directed at the action. They are our contribution (sunergon) to the action. They are not portrayed as choices between alternatives or decisions whether or not to do something; they are volitions, impulses, endeavours to do something. (iii) Epicurus repeatedly says that we call the cause of an action ‘through ourselves’ (di’hēmōn autōn). I assume that the cause through ourselves are precisely our impulse and endeavour. (And they are, I take it, called ‘through ourselves’ because they are the result of our own beliefs and desires, which, when externally triggered, produce the impulse and endeavour,³⁴ and they make us the cause of our actions.)³⁵ We can then also see what Epicurus meant when in the Letter to Menoeceus he calls that which happens because of us (par’hēmas) ‘without master’ (adespoton, see passage quoted above). He refers to the fact that the things which happen because of us are not forced. More precisely, that we are not forced when we bring them about. There is nothing that compels (us to do) them. In particular, our actions are not subordinated to fate (or necessity), which a few lines before was characterized as master (or rather mistress, despotis).³⁶ The contrast in this passage is thus the same as the one in the digression in Epicurus’ On Nature. (12) . . . supremely unthinkable. But unless someone perversely maintains this, or makes it clear what fact he is rebutting or introducing, it is merely a word that is being changed, as I keep repeating. (13) The first [thinkers] to give a satisfactory account of causes, [thinkers] not only much greater than their predecessors but also, many times over, than their successors, contradicted themselves unawares— although in many matters they had alleviated great ills—in this respect that they held necessity and . . . causally responsible for everything. (14) Indeed, the actual account promoting this view came to grief when it left the great man blind to the fact that in his actions he was clashing with his doctrine; and that if it were not that a certain blindness to the doctrine took hold of him while acting he would be constantly perplexing himself; and that wherever the doctrine prevailed he would be falling into desperate calamities, while wherever it did not he would be filled with conflict because of the contradiction between his actions and his doctrine.³⁷ (Epic. Nat. 25; Arr. 34.30, Laursen 1997, 40–2) ³⁴ Epic. Nat. 25, Arr. 34.26, Laursen 1997, 33, LS 20C(1); and perhaps Arr. 34.31, Laursen 1997, 44–5, suggest that in order to be causes ourselves, we must have beliefs, and that these beliefs must be our own beliefs. ³⁵ Cf. Epic. Nat. 25, Arr. 34.22, Laursen 1997, 32–3, LS 20B(1), τὴν . . . αἰτίαν . . . ἑαυτῶν, and LS 20B (5), τὴν ἐξ ἑαυτοῦ αἰτίαν, etc. See also Chapter 5 of the present volume. ³⁶ Cf. also δουλεύειν a few lines later. (In natural-language use, locutions such as ‘our actions are free’ are generally not understood as ‘independent of us’, but as ‘free from interference by things other than us’.) ³⁷ (12) . . . μάλιστα ἀδιανοήτων. ἂν δέ τις τοῦτο μὴ παραβιάζηται, μηδ᾽αὖ ὃ ἐξελέγχει γε ἢ ὃ εἰσφέρει πρᾶγμα ἐκτιθεῖ, φωνὴ μόνον ἀμείβεται, καθάπερ πάλαι θρυλῶ. (13) οἱ δ᾽αἰτιολογήσαντες ἐξ ἀρχῆς ἱκανῶς, καὶ οὐ μόνον τῶν πρὸ αὑτῶν πολὺ διενέγκαντες ἀλλὰ καὶ τῶν ὕστερον πολλαπλασίως ἔλαθον ἑαυτοῖς—

    -  ? 167 This last section of the digression presents again a pragmatic argument: philosophers who hold that every event is caused by necessity ‘contradict’ themselves when they act. The point seems to be that one embarks on an action only if, while one acts, one thinks of oneself as not necessitated and as causally responsible for one’s actions. Freedom of decision or choice are not at issue. We can conclude that in the digression Epicurus’ concern is to refute the view that our actions are necessitated, in the sense of being caused by something other than us. He does not discuss the question of whether we are undetermined in our decisions. As a matter of fact, all his arguments in this digression could be consistently proposed by a compatibilist determinist. This does not rule out that Epicurus was an indeterminist—and I believe he was. All I suggest is that in the arguments of the digression what is at issue is not indeterminism or free decision or choice, but agent autonomy. A remark on freedom to do otherwise (as different from freedom of decision and freedom of choice). I have described the contrast between necessity and that which happens because of us as that between compulsion and absence of compulsion. This is a common Greek way of understanding necessity. For Epicurus, a person’s behaviour happens by necessity if the person is compelled to behave that way. If a person’s behaviour (action) results from their having an impulse to act that is based on their present beliefs and desires in response to some external stimulus, then the person is not compelled in their behaviour (action). Freedom to do otherwise is not explicitly involved. It has been objected, however, that if one’s action is not necessary, then this entails, or even means, that it was possible for one not to act, and consequently that one was free (in some sense) to do otherwise than one did; hence that Epicurus must have held that we are free (in some sense) to do otherwise than we do. I am unsure about how to respond to this objection. First of all, I am inclined to think that Epicurus believed that it is usually up to us (eph’hēmin) whether or not we act, either in the sense that if we had different beliefs or desires we would act differently, or in the sense that we have some general two-sided capacity for certain things, such as walking and not walking (see Chapter 1). And if someone wants to call either of these ‘freedom to do otherwise’, so be it—as long as they are aware that such kinds of freedom are in principle compatible with determinism. Second, however, I am uncertain whether Epicurus ever de facto drew the connection between such freedom and non-necessity,³⁸ although, again, it is likely that he thought it a precondition for an action to καίπερ ἐν πολλοῖς μεγάλα κουφίσαντες—ἐν το τὴν ἀνάγκην καὶ ταυτομεγ [ . . . ]ν πάντα αἰτιᾶσθαι. (14) ὁ δὴ λόγος αὐτὸς ὁ τοῦτο διδάσκων κατεάγνυτο καὶ ἐλάνθανεν τὸν ἄνδρα τοῖς ἔργοις πρὸς τὴν δόξαν συνκρούοντα˙ καὶ εἰ μὴ λήθη τις ἐπὶ τῶν ἔργων τῆς δόξης ἐνεγείνετο, συνεχῶς ἂν ἑαυτὸν ταράττοντα˙ ᾗ δ᾽ἐκράτει τὸ τῆς δόξης κἂν τοῖς ἐσχάτοις περιπείπτοντα˙ ᾗ δὲ μὴ ἐκράτει στάσεως ἐμπιπλάμενον διὰ τὴν ὑπεναντιότητα τῶν ἔργων καὶ τῆς δόξης. add. Laursen 1997, 42. ³⁸ At Epic. Nat. 25, Arr. 34.27, Laursen 1997, 28: ‘but out of itself or out of the cause out of itself being able to [develop] also something else’ is contrasted with necessity of development. For the context of the passage, see below, section 6.

168

, ,   

happen because of us (par’hēmas) that it was up to us. Third, regardless of whether he expressly drew this connection, my point is that we have no evidence for the assumption that he ever regarded the compatibility of such kinds of freedom with atomistic mechanical necessity or causal determinism as problematic. We do not know whether he did. In the surviving passages he appears to discuss different—if related—problems.

4. Lucretius, De Rerum Natura 2.251–93 and 4.877–91 The passage 2.251–93 of Lucretius’ De Rerum Natura has traditionally been adduced as the main evidence for the claim that Epicurus was concerned with a free-will problem as set out in section 1. It is, however, generally agreed that Lucretius’ immediate concern in the passage is a different one: he provides a— second—argument for the existence of the swerve. This argument very roughly follows modus tollens: (A) If the swerve does not exist, neither does volition (2.251–60) (B) But volition exists (2.261–83) (C) Therefore the swerve exists (2.284–93) Thus—on the assumption that Lucretius is sufficiently faithful as a witness of Epicurean doctrine—we can infer that Epicurus regarded the swerve as a necessary condition for the existence of volition. However, Lucretius does not tell us anywhere in the passage in what way the swerve is required for volition. As a consequence, scholars have, with much inventiveness and subtlety, produced a host of interpretations, each designed to answer this question. Those who argue that the Lucretius passage shows that Epicurus discussed free choice or free decision generally agree on the following point: the swerve is meant to help solve the free-will problem; its function is to provide the element of indeterminedness that Epicurus thought is needed for individual decisions or choices to be free. Most commonly the assumption is that one or more swerves are involved in the formation of every volition.³⁹ Since I do not believe that Epicurus was faced with a free-will problem as set out in section 1, I also do not believe that it is the role of the swerve to preserve free choice or decision. Nor do I believe that it has to feature in every act of volition. Before I say anything more about the swerve, I want to present what I hope to be a consistent alternative interpretation of the Lucretius passage. (It will be easier to follow the suggested interpretation if the reader assumes at least hypothetically ³⁹ Asmis 1990; so for Bailey 1928; Gulley 1990; Huby 1967; Purinton 1999; Sedley 1983; Sharples 1991–3.

    -  ? 169 that Lucretius has at the back of his mind the disposition-dependency model of agency and the problem of autonomous agency, and not the independentdecision-faculty model and a free-will problem as set out in section 1.) At the same time, I shall point out a number of those ambiguities in the text that I have mentioned above and which may have furthered the view that Lucretius was concerned with freedom of decision or choice. One possible ambiguity should be mentioned at the beginning, since it stretches through the whole passage: this is the fact that ‘voluntas’ can be used equally for an act of volition and for a capacity or power of volition. It seems to me that ‘voluntas’ is always used for volitional acts, and that when Lucretius refers to a power of volition he uses different phrases. But this is not very important. What matters— and will be argued in this section—is that volitions are not acts of choosingbetween or deciding-whether, but our willing (or impulse or endeavouring) to perform an action; and that accordingly our power of volition is not a power of choice-between, or a decision-making faculty, but our ability to form in response to external stimuli volitions in accordance with our own beliefs and desires. Let me begin then with section (A). In this section Lucretius does more than just state, in a somewhat passionate way, that the swerve is a prerequisite for volition. He also provides us with information about the determinism he attacks, and with an implicit account of volition (voluntas): (A)(1) Moreover, if all motion is always linked, and new motion arises out of old in a fixed order, and atoms do not by swerving make some beginning of motion to break the decrees of fate, so that cause should not follow cause from infinity, (2) from where does this free volition exist for animals throughout the world? (3) From where, I ask, comes this volition wrested away from the fates, through which we proceed wherever each of us is guided by their pleasure and likewise swerve off our motions at no fixed time or fixed region of space, but wherever the mind itself carries us.⁴⁰ (Lucr. 2.251–60)

(1) describes the predetermination of all events which the swerve is said to prevent. The theory is one of causal determinism. There is a sequence of causes which reaches back infinitely into the past; there is a fixed order of all motion; this order is in accordance with the ‘decrees of fate’.⁴¹ We can infer that this fixed order and those ‘decrees of fate’ also go back infinitely into the past, and that all motions

⁴⁰ (1) denique si semper motus conectitur omnis | et vetere exoritur novus ordine certo | nec declinando faciunt primordia motus | principium quoddam quod fati foedera rumpat, | ex infinito ne causam causa sequatur, | (2) libera per terras unde haec animantibus exstat, | (3) unde est haec, inquam, fatis avulsa voluntas | per quam progredimur quo ducit quemque voluptas, | declinamus item motus nec tempore certo | nec regione loci certa, sed ubi ipsa tulit mens? ⁴¹ ‘Decrees of fate’ in quotes, since (as in Ep. Men. 133–4) the reference should be to the ‘fate of the natural philosophers’, i.e. to mechanistic necessity, not to a theory of teleological determinism.

170

, ,   

are understood as being in this way eternally predetermined. (3) contains the implicit account of volition as that ‘through which we proceed wherever each of us is guided by their pleasure and likewise swerve off our motions at no fixed time or fixed region of space, but wherever the mind itself carries us’. I take this account in two parts. First, ‘this volition . . . through which we proceed wherever each of us is guided by their pleasure’ (Lucr. 2.257–8). This phrase suggests that volition is the vehicle by means of which we pursue or realize the satisfaction of our desires. If I find smoking cigarettes pleasant, then an act of volition directed at my smoking will be a necessary step to get me to smoke. Volition is here described as the consequence of our pleasure directing us somewhere: if—in a situation of possible smoking—I find smoking pleasant, then, it seems, by means of a volition a motion towards getting a cigarette will be started. For the second part of the sentence (2.259–60) it is important to take it in its entirety and not to cut it off before the ‘but’. We are presented with a contrast: we swerve off our motions through volition not at a fixed time or space but wherever the mind itself carries us. Two things are unclear here. First, what does ‘mind’ (mens) mean in this sentence? Proponents of the view that Lucretius discusses free choice or decision have repeatedly suggested that when Lucretius says ‘mind’ here and later in the passage, what he actually means is volition, and that he is only speaking loosely.⁴² I prefer to think that when Lucretius says ‘mind’ in our passage, what he means is actually mind. That is, he means the central part of the soul, which is located in the heart and which elsewhere he calls animus or mens.⁴³ Second, the phrase ‘neither at a fixed time, nor in a fixed space’ can mean two things. It can mean ‘(spatio-temporally) at random’. The opposite is ‘not at random’, implying ‘with some order’. Or else, it can mean ‘not at a pre-determined time or space’. Here the opposite is ‘at a pre-determined time or space’. These two options differ: logically, the second does not require that the motion is random. I believe that Lucretius intended this second reading, i.e. that time and space of the motions are not fixed in advance, from eternity (cf. 2.255 ex infinito).⁴⁴ For the contrast in our sentence is between ‘swerving off a motion at a fixed time or space’ and ‘swerving off a motion wherever the mind carries us’. But if, as I take it, ‘mind’ means mind, and we ‘swerve off ’ our motions⁴⁵ where our mind carries us, this cannot properly be described as a random motion. For our mind carries us ⁴² E.g. Sedley 1983, 47 n. 65; Long and Sedley 1987, vol. 2, 111–12; Gulley 1990, 42. ⁴³ E.g. 2.270, animi . . . voluntate, and 3.139, consilium quod nos animum mentemque vocamus. The Greek would be διάνοια, or something similar. ⁴⁴ There is the parallel, with the prefixed order (ordine certo), from line 2.252, and with the whole predeterministic scenario (ex infinito, fatis, etc.) from the first part of the sentence (2.251–7). All this evokes the familiar idea of predetermination from infinity. Cf. Cic. Fat. 21 for certo with the sense of ‘predetermined’ and thus implying necessity. Similarly, certus in Cic. Nat. D. 1.69. ⁴⁵ I agree with Sharples 1991–3, 182, that Lucretius here has our movements in mind, not the swerving atomic movements in us.

    -  ? 171 wherever it carries us in accordance with our pleasure—as is implied by the first part of the sentence, and this is not at random: without pleasure the mind would not carry us there.⁴⁶ Lucretius’ contrast is this: either our mind is the cause of our motion (by means of a volition) or something else (fate, necessity, our initial constitution, etc.) predetermines the time and location at which our motion occurs (and thus is its true cause). This is a variation of a point we encountered earlier, in the digression of Epicurus’ On Nature 25. To sum up, the implicit account of volition in (A) can be read as saying that it is an essential characteristic of a volition that it is initiated directly by the mind, in accordance with our desire, and that it is not predetermined by something else. Section (A) also contains the one phrase which offers perhaps the main reason for the persistent assumption that Lucretius is discussing free will in the passage under consideration: libera voluntas. The meaning of the phrase is however not ‘freedom of choice’ or ‘freedom of decision’, nor does it denote or imply a faculty of the will capable of making un-predetemined choices between alternative courses of actions. Rather, an act of volition is free (libera) since it is not forced by fate, necessity, our initial constitution, etc. It is an unforced volitional act. Importantly, libera voluntas is pleonastic here: if the volition were not free or unforced, it would not be a volition. Section (B) (2.261–83) has the function of establishing the existence of volition. For this step in the argument reference to the swerve is not required and we should not normally expect it. What Lucretius produces in this passage is an empirical argument that backs up the existence of volition by contrasting the phenomena of volitional movement of living beings with those of externally induced movement. Movement that is initiated by something external to the moving body is described as non-volitional (invitus, 2.275, 2.278) and as forced (coactu, 2.273; cogat, 2.278). Volitional motion, on the other hand, is unforced.⁴⁷ It is characterized as involving (at least) two kinds of motion in our passage (B), and as involving three kinds of motion in the passage on volition in Book 4 which is often employed to illuminate our text (4.877–91). First, there is the volition or desire of the mind (mens avet 2.265; studium mentis 2.268; voluntas animi 2.270). It is itself a motion of the mind, and is initiated by the mind (4.886, 2.269–70). The mind, thus moved, then sets in motion the soul, which extends through the entire body (4.887–8; hinted at at 2.271). The soul, thus moved, in turn moves the body, and in this way sets in ⁴⁶ If in 2.258 one reads voluntas instead of the commonly accepted emendation voluptas, a similar argument can be produced by taking voluntas to signify individual instances of volition (‘this volition . . . through which we proceed wherever each of us is guided by it’). For instance, if I have a volition to smoke (based on some belief of mine that it is pleasant, hence good for me), then this volition will guide me towards smoking a cigarette. ⁴⁷ This is similar to the point that that which happens because of us (τὸ παρ᾽ἡμᾶς) is without master (ἀδέσποτον). I assume that that which happens because of us and our volitional movement are coextensional.

172

, ,   

motion the whole living being (2.266–8, 4.890–1).⁴⁸ This enables Lucretius to say that the body follows the mind’s desire or endeavour (studium mentis, 2.268). The phenomenological difference between volitional (and thus unforced) motion and forced motion of living beings is that in the former case there is an observable time delay between the formation of the will to act and the eventual movement of the entire body. And this—Lucretius maintains—can only be explained by the assumption of the existence of volition, which needs some time to internally mobilize the body via the soul. So far this step of the argument.⁴⁹ I have indicated above that, as regards the logical function of section (B) in the argument, we should be surprised if we found in it any discussion of the swerve. This fact has not prevented scholars from searching for traces of the swerve in virtually every line of it. Together with the parallel passage in Book 4 (4.877–91), this passage has repeatedly been exploited to back up the views that Epicurus considered the swerve to be a necessary condition for free decision or choice, and that the swerve features in every instance of volition. The main reason for this, I suspect, is another characteristic ambiguity, namely that which we find in phrases like ‘the beginning of motion’ and ‘setting oneself in motion’. For when something is said to bring about a beginning of motion, or to move itself, this can be understood in two rather different ways. First, it can refer to some absolute beginning of motion: something produces a motion, or causes itself to move, but is itself not fully causally determined to do so by any prior motion. There is a gap in the ‘causal chain’. The motion is spontaneous.⁵⁰ Second, talk about moving oneself and producing a beginning of motion can refer to a relative beginning of motion: here a thing is said to move itself, or to produce a beginning of motion, if, given certain external or internal impacts, the thing will start to move because it is the sort of thing it is. For instance, two people are each offered a cigarette: one takes it

⁴⁸ Cf. also 3.159–60, facile ut quivis hinc noscere possit | esse animam cum animo coniunctam, quae cum animi vi | percussast, exim corpus propellit et icit. ⁴⁹ Passage (B) has been used to argue that Lucretius was concerned with freedom of action rather than freedom of the will (Conway 1981). Freedom of action is the freedom I have when there is nothing that prevents me from acting as I desire or choose to act. Freedom of action is lacking when I desire or have chosen to do something, but the realization of the action is prevented by physical hindrances, and my desire is thus frustrated. In this case things happen against my desire or will. Now it is true that (B) includes an adequate description of what we call free action. (Cf. also Lucr. 4.877–8, ‘how it comes that we can step forward when we want to’, where free action seems to be the topic.) However, I think that in (B) Lucretius is not primarily interested in free action. At least, his contrast is not between free action and prevention of free action, i.e. frustrated desire or choice. Rather it is between motion that has an internal beginning by means of the agent’s desire and volition, and motion that has an external beginning, without the agent’s desire or volition being involved. In this latter case, something happens without my volition (2.275, 278) but not necessarily against my desire. The contrast is between volitional and non-volitional motion, not between free action and thwarted intention. Lucretius makes use of the phenomenon of free action in order to back up his thesis that volition exists. This is how (B) starts (‘for no doubt it is volition that gives these things their beginning for each of us, and it is from volition that motions are spread through the limbs’) and this is how (B) ends (‘So do you now see that . . . there is something in our chest capable of . . . ’, namely the power of volition). ⁵⁰ So e.g. Bailey 1928, 435; Purinton 1999, 255, 276.

    -  ? 173 and smokes it, the other does not. One is the sort of person who considers smoking pleasurable, the other is not. Here—given the external stimulus—it depends entirely on the person whether movements of taking the cigarette will be initiated. If we look at the three relevant passages in Lucretius, we can see that none of them compels us to interpret them as postulating an absolute beginning of motion. Hence, none of them implies direct involvement of the swerve. First 2.261–2: his rebus sua cuique voluntas principium dat. In this phrase ‘his rebus’ refers to ‘motus’ (our movements, i.e. our volitional actions) from 2.259. We can then translate: ‘volition gives our movements the beginning for each of us’. This phrase need not mean anything more than ‘the motion of volition makes us move’ (or ‘makes our body move’), as contrasted with the case in which something external makes us (or our body) move. Nothing is said about what brings about the volition. Hence there is in this case certainly no reason to assume an absolute beginning of motion.⁵¹ Second 2.269: initum motus a corde creari . . . , which can be translated as ‘a beginning of motion is brought about by the mind . . . ’.⁵² This case differs from the previous one in that, as far as the wording is concerned, the beginning in question could be absolute or relative. It could be understood as ‘the mind, qua decisionmaking faculty, may, without being caused in any way to do so, produce a motion of volition’. This would mark an absolute beginning. Or it could be understood that in a certain situation the mind, because it is the way it is, may produce a motion of volition. For example, in a situation of a certain kind, a smoker may start smoking a cigarette (as a result of their beliefs and desires), whereas a nonsmoker may do nothing (as a result of their beliefs and desires). This would be a relative beginning. The third passage, 4.886, is similar to the second: ergo animus cum sese ita commovet ut velit ire . . . ; in English: ‘thus, when the mind sets itself in motion so that it wants to go forwards . . . ’. This, too, can easily be read as being about a relative beginning of motion, for example in the following way: because of the mind’s individual constitution, images of walking strike the mind, and presumably appear to it as pleasant,⁵³ when it ‘previews’ them (4.884–5). (Another mind, in the same situation, may not be struck by such images of walking as pleasant.) As a result, the mind sets itself in motion in accordance with the images. More precisely, it sets in motion its faculty of volition. But once the mind is set in ⁵¹ If voluntas denoted a power of volition, this could be understood as ‘the capacity of volition, by producing volitions, gives animal movements their beginning of motion’. This too would be contrasted with an external beginning of motion, and need not express an absolute beginning, since it is not ruled out that there are causes that make the power of volition produce volitions. ⁵² Or alternatively ‘in the heart’. (Englert, Furley: ‘in the heart’; Sedley: ‘from’; Smith: ‘by’. Lucretius: ‘cor’: 1.737, 923 (introductions), 4.53 (5.882) 5.1107 = LS 22L, tr. ‘mind’; 6.5 = LS 21X, tr. ‘insight’.) ⁵³ Remember 2.257–8: voluntas per quam progredimur quo ducit quemque voluptas.

174

, ,   

motion, that means that it velit ire, wants to go forward, which I understand as ‘it now has the volition to go forward’.⁵⁴ Thus all three passages in Lucretius harmonize well with the assumption that in them Epicurus had only a relative beginning of motion in mind. Accordingly, none of them provides compelling evidence for the view that swerves are involved directly in each act of volition or for the view that they provide the element of undeterminedness required in an act of free choice or decision. This leaves us with the last section, (C), of the Lucretius passage on the swerve (2.284–93). In it the conclusion that the swerve exists is drawn from the premises set out in (A) and (B), but we have again more than that. An explanation is added of why the conditional premise is true. And here we finally obtain two valuable bits of information about the relation between volition and swerve. The first is in the first sentence: Therefore in the atoms too one has to admit another cause of motions besides impacts and weights from which this power is born in us, since we see that nothing can come to be from nothing.⁵⁵ (2.284–7)

The term ‘power’ (potestas) in this sentence is another chameleon expression. Potestas can denote a disposition or capacity, i.e. something which is possessed continuously, both when it is actualized and when it is not. Potestas can also denote something like energy or force (or power as in ‘power station’), i.e. the force released in an instance of volition. Such a power is something we do not have continuously. It exists only as long as the volition lasts. The phrase ‘from which this power is born in us’⁵⁶ (nobis innata) in the sentence suggests that Lucretius is talking about a capacity. To what capacity does Lucretius then refer? As he calls it ‘this power’, it must be a power he has talked about shortly before. No power is explicitly mentioned in the whole passage, but there is an implicit reference: with potestas Lucretius can only really refer to lines 2.279–83:

⁵⁴ It is perhaps not without interest that Epicurus seems to have considered this kind of relative beginning of motion in the context of the formation of volitional action in On Nature 25 (Arr. 35.10, Laursen 1995, 44 and 91 on PHerc 1420.2.2). There he seems to hold that an external influence may affect different people differently. Simon Laursen considers this to be a parallel to the Lucretius passage in Book 4 (4.877–91) which I have just discussed. ⁵⁵ quare in seminibus quoque idem fateare necessest, esse aliam praeter plagas et pondera causam motibus, unde haec est nobis innata potestas, de nilo quoniam fieri nil posse videmus. ⁵⁶ Or ‘from which we have this inborn power’; it makes little difference whether one takes innata as attributive adjective of potestas or as predicative adjective belonging to est. (Cf. ‘from which comes this innate power in us’ (Englert), ‘from which we have this inborn power’ (Furley)—cf. also 3.270, mobilis illa vis.)

    -  ? 175 There is something in our chest that is capable of (possit) fighting and resisting at whose bidding the mass of matter is also forced at times to be turned throughout the limbs and frame . . . ⁵⁷ (2.279–82)

This something in our chest is (that aspect of our mind which is) the power of having volitions, i.e. the power which can make the body move. Again, this should be a capacity rather than some kind of energy. Thus the swerve is a necessary condition for our having this power of volition. It is then possible to understand the phrase ‘from which this power is born in us’ in the following way: the power of volition is a capacity which we acquire at some stage of our life; and the swerves are somehow responsible for the coming to be (and perhaps for the sustaining) of this capacity in us. But let us look at the next sentences: For weight prevents that all things come about by impact, by, as it were, external force . . . ⁵⁸ (2.288–9)

The mind’s weight (presumably including its atomic structure, see below) is sufficient to warrant that the mind’s movements are not completely externally forced. However, as the next sentence in Lucretius makes clear, the mind’s weight is not sufficient to prevent its movements from being forced by internal factors: . . . but that the mind should not itself possess an internal necessity in all its behaviour, and be overcome and as it were forced to suffer and be acted upon, that is brought about by a tiny swerve of atoms at no fixed region of space or fixed time.⁵⁹ (2.289–93)

Similarly to the ‘digression’ in Epicurus’ On Nature 25, in these lines necessity is connected with force or coercion.⁶⁰ The crucial distinction in 2.288–93 is that between external and internal compulsion. The swerve is said to prevent internal necessity of the mind,⁶¹ and thus the mind’s being ‘overcome and as it were forced ⁵⁷ . . . esse in pectore nostro | quiddam quod contra pugnare obstareque possit? | cuius ad arbitrium quoque copia materiai | cogitur interdum flecti per membra per artus . . . ⁵⁸ Pondus enim prohibet ne plagis omnia fiant | externa quasi vi. ⁵⁹ sed ne mens ipsa necessum | intestinum habeat cunctis in rebus agendis | et devicta quasi cogatur ferre patique, | id facit exiguum clinamen principiorum | nec regione loci certa nec tempore certo. mens Lambinus: res OQ. ⁶⁰ In On Nature 25 (quoted in section 3 above), forced and necessitated actions were contrasted with actions performed with impulse (ὅρμημα) and eagerness or desire to act (προθυμία). This seems to be the closest parallel in a text by Epicurus to Lucretius’ voluntas. (If studium is a translation of προθυμία, voluntas may be a translation of ὁρμή, ὅρμημα, or a similar term.) The power of volition would then have been a δύναμις of ὁρμή or ὅρμημα, and a volition (voluntas) a particular impulse to act or intention a person has, e.g. the impulse or intention to smoke a cigarette. ⁶¹ I adopt the generally accepted emendation of res to mens. But I think (pace Avotins 1979) that if the original res were to be kept, this would not make much of a difference for my interpretation.

176

, ,   

to suffer’. One’s interpretation of this internal necessity or coercion will differ depending on what one takes Epicurus’ model of agency to be. (Cf. Chapter 5, section 3 for some elements of Epicurus’ theory of action.) Proponents of the independent-decision-faculty model will be prone to the following reading: they will understand the internal necessity of the mind as necessitation, or coercion, of one part of the mind (the power of volition)⁶² by other parts of the mind, in particular by the person’s present character dispositions.⁶³ These would—if the mind’s motions were internally necessitated—in the case of each action force the power of volition to initiate movement in accordance with them. Thus, effectively, there would be no power of volition. The swerve’s role in the case of each action would naturally be somehow to sever the decision-making power from the agent’s present dispositions. The whole-person model of agency suggests a different interpretation of the internal necessity. In it, it is presupposed that a person’s volitions to act (in response to environmental stimuli) are always fully determined by the person’s present overall mental disposition as it is while the person forms the volition. The decision-making is understood as a ‘function’ of the mind as the mind is when it decides and of external circumstances. Since in this model there is no independent decision-making faculty, ‘mens’ (2.289) refers to the mind qua conglomerate of atoms in which are manifested a set of dispositions. The mind’s internal force is distinguished temporally from a person’s present overall mental disposition: internal coercion is coercion of someone’s present overall mental state by temporally prior mental states or dispositions, which in turn were necessitated by temporally prior mental states, etc. back to a time at which the individual is thought not yet to be responsible for their actions. The necessitation thus concerns the development of the mind, not its decision-making. The difference between internal necessity and its absence concerns the point whether the person was internally forced to become, or develop into, the person they are when they set out to act. In the case of force, the action cannot be attributed to the agent, since the agent is not truly causally responsible, but some other factors which predate adult agenthood and which necessitate the agents in their action by necessitating their mental dispositions. For I take it that omnia fiant in 2.288 need not refer to semina from 2.284, but that it can just as well refer generally to the things that happen on the everyday level. This might also help explain the use of quasi, by which Lucretius qualifies the external force. The point of lines 2.288–9 would then be that not all things react in the same way when externally pushed, for the reason that different things have different weights, which make them react in different ways. Sed ne res ipsa . . . in line 2.289 could then be translated as ‘but that a thing itself should not . . . ’, where by ‘thing’ Lucretius refers in an indeterminate way to all the things on the everyday level that may not be internally necessitated, the most important of which would be human beings, or their minds. Other such things might be all those things that develop or change in a random way. ⁶² Or perhaps rather that which would be the power of volition if it were not necessitated in its activity. ⁶³ E.g. Asmis 1990, 283.

    -  ? 177 This raises the question of what, in this interpretation, the internal factors would be that necessitate the agents in their action, if there were no swerves, or, in other words, the question of what the internal necessity is. The answer, I believe, can be gauged from a passage from Epicurus’ On Nature 25 which I have already quoted above in section 3. In this passage, Epicurus critically assesses the case for holding that all our actions are determined by a combination of internal and external necessity: by which we never cease to be affected, we rebuke, oppose, and reform each other as if the cause lay also⁶⁴ in ourselves, and not just in our congenital makeup and in the accidental necessity of that which surrounds and penetrates us. (Arr. 34.27, Laursen 35; LS 20C(2); Greek text above, n. 29)

A person’s behavioural response to the environment thus seems to be internally necessitated, inasmuch as the cause of the behavioural response to the environment lies in the person’s initial constitution instead of in the person themself.⁶⁵ This is consistent with the assumption of the whole-person model that a person’s behavioural responses to the environment are determined by the overall disposition of the person’s mind when the person sets out to act. It can be interpreted as suggesting that if the person themself is not the cause of the action, this means that the person’s overall disposition at that time is internally necessitated by the person’s initial constitution (and thus by something that is temporally prior to the person’s formation of a volition—or any volition, for that matter). Returning to Lucretius 2.288–93, it would then be the initial constitution that internally necessitates the mind in all its behaviour, and it would be the swerve’s function to prevent this necessitation. Internal necessity corresponds to the weight of the atoms (see above). The internal necessity of the mind should thus correspond to the weight—and composition—of the mind atoms. And this is exactly what we can imagine the initial constitution to be: it is a collection of atoms, the precise composition and structure of which varies from person to person. The atomic composition and structure make up a person’s mental dispositions, and are as such relatively firm; and they determine which of the impinging images the mind takes in and acts upon, and which not.⁶⁶ The mere penetration of the mind by individual atoms or groups of atoms (e.g. eidōla) coming from outside ⁶⁴ I.e. we are a co-cause, presumably the main causes but not necessarily the only cause. There are other passages to this effect in On Nature 25, e.g. Arr. 34.31, Laursen 1997, 43: ὅθεν καὶ τὸ τοῦ τέλους αὐτοῦ ἐπιλόγισμα εἶχε μὲν καὶ ἡ ἀρχὴ τὴν αἰτίαν εἴχομεν δὲ καὶ ἡμεῖς. ⁶⁵ Note also Epic. Nat. 25, Arr. 34.33, Laursen 1997, 48, the contrast between δι᾽ἡμᾶς, διὰ τὴν φύσιν, καὶ τὸ περιέχον. ⁶⁶ Cf. Epic. Nat. 25, Laursen 1997, 33, LS 20C(1), quoted above in section 2.

178

, ,   

will not usually lead to a change of the mental structure. However, swerves may lead to a new structure of these atoms, perhaps to the integration of incoming atoms into this structure, and it will be the partial change of atomic structure (or the development of structure) which prevents internal necessitation of our behaviour by our initial constitution. The swerve leads to changes in the structure of the mind. Hence the mind is not necessitated or compelled in its movements by its original structure. And our decisions and actions with which we respond to our environment depend (at least to a major extent) on the new structural elements of the mind, and are hence not internally necessitated. I conclude then that the Lucretius passage on swerve and volition can be read consistently as based on the whole-person model of agency, and as treating internal necessity as necessitation of the development of the mind, as opposed to necessitation of its individual decisions. Individual decisions are generally fully determined by the person’s mind as it is when the decision is made. If a swerve comes in at this point, it will usually have no effect on what decision is made, in the same sense in which a swerve that occurs in the mind in the process of perception will usually have no effect on the reliability of perception. In both cases the mental structure can be conceived of as so firm that a single swerve does not easily make a difference to our decision or perception. If it ever did, the result would be chance events; this possibility could have been used by Epicureans in order to explain the occasional apparent malfunction of our volitional or perceptual apparatus. Proponents of the view that Lucretius discussed a free-will problem as set out in section 1, however, may not yet be satisfied. They have repeatedly objected that mind- or character-formation interpretations cannot satisfactorily explain Lucretius’ phrase ‘we swerve off our motions’ (2.259), which they take as strong evidence that there is a swerve involved in every act of volition. I shall attempt to show that Lucretius’ analogy between the three types of atomic motions and the three types of motions on the everyday level is not as close as this. Rather, for all three types of motion the primary point of the analogy is that the atomic motion is a necessary condition for the corresponding type on the everyday level. First, collision: if atoms did not collide, everyday objects could not collide. However, the collision of objects is qualitatively different from atomic collision; it can involve penetration of a complex of atoms by one or more atoms, and the destruction of a complex or aggregate of atoms. Second, weight: if atoms had no weight internally directing their downwards movements (their default movements), then objects would have no weight (and atomic structure) internally directing their movements and changes (their default movements). What is analogous to the necessary downward movement of an atom is the necessary movement or change of an object in accordance with its internal physical properties, which are determined by certain stable combinations of atoms. Again, the analogous movements are clearly much more complex than the atomic ones. They are also qualitatively different, as they need not be ‘downward’ movements at all.

    -  ? 179 Third, the swerve: if atoms did not swerve, there would be no volitions (and no random movements). There are a number of corresponding elements. If there were no atomic swerves, all atomic movement would be a function of weight and collision of atoms. If there were no volitions, all movements of the mind would be a function of the mind’s initial atomic composition and its collision with external things (e.g. ‘images’). With atomic swerves, the mind can develop in such a way that its movements are no longer a function of the mind’s weight and collision with external things. Like the swerves, and because of them, the volitional movements of the mind are not eternally predetermined. As it is the nature of the atoms to swerve, so it is our (mind’s) nature to have volitions; and as the atoms swerve from their downwards path, so do we, with our volitions, swerve from our path of hereditary development. (Since Lucretius’ work is a poem which amply uses analogies for literary purposes, his use of ‘swerve’ in ‘swerve off ’ is nothing but another use of this literary device.)

5. Cicero, On Fate 23 and Diogenes of Oenoanda 32.1.14–3.14 A sentence from Cicero, On Fate 23 has been plausibly suggested as a parallel to the lines 2.288–93 from Lucretius:⁶⁷ Epicurus introduced this theory because he was afraid that, if the atom was always carried along by its weight in a natural and necessary way, there would be nothing free for us, since our mind would be changed in such a way as the movement of the atoms would compel it.⁶⁸

The parallels to Lucretius are obvious. In addition, there are two elements in this sentence which we do not find in the De Rerum Natura. Cicero implies that we have some sort of freedom, which is contrasted with our mind being compelled; and he mentions explicitly what it is that would force our mind if there were no swerve, viz. the movement of the atoms. This sentence has been adduced as a proof that Epicurus used the swerve to save freedom of choice or decision, and not for the ‘freedom’ of the development of dispositions. However, just like the verses in Lucretius, this sentence is compatible with the view that the swerve secures the non-necessity of the development of our mental dispositions on the basis of the whole-person model of agency. This model here leads to the following interpretation: Cicero’s emphasis is on the contrast between internal compulsion by the

⁶⁷ Long and Sedley 1987, vol. 2, 112. ⁶⁸ hanc Epicurus rationem induxit ob eam rem, quod veritus est ne, si semper atomus gravitate ferretur naturali ac necessaria, nihil liberum nobis esset, cum ita moveretur animus ut atomorum motu cogeretur.

180

, ,   

atoms and freedom from such compulsion. The argument works from the atomic level to the everyday level. If there were no swerve, all atomic movement would be necessary. Whether movements at the everyday level are necessary depends on whether all movements at the atomic level are necessary. In particular, if all the atomic movements that ‘make up’ an everyday-level movement were necessary, so would be the corresponding everyday-level movement. (In this way, atomic movement, since necessary, would transmit its necessity to the everyday-level movement and thus in a sense ‘compel’ it.) In particular, ‘our mind would be moved in such a way as the movement of the atoms would compel it’. That is, combinations of collision and weight would fully determine the way the structure of the mind changes or develops and accordingly how the mind reacts to external influences. At any time our mental dispositions would be a function of our initial constitution and external influences, and so would, accordingly, our behaviour. On the other hand, as there are swerves, we have freedom from compulsion and it is not the case that ‘our mind would be moved in such a way as the movement of the atoms would compel it’. The reason is that, as a result of the swerving movements, not all atomic motions are necessary, and hence they no longer convey necessity to all change at the everyday level. In particular—as swerves occur in our mind—our mental dispositions are not the result of compulsion by the atoms, and nor, accordingly, will be the volitions and actions that flow from it. Rather, they are free, that is uncompelled. (The phrase ‘cum ita moveretur animus’ can refer either to what would otherwise be our volitions or to the change of our mental dispositions. In the latter case the translation ‘be changed’ is preferable to ‘be moved’. For my point it is immaterial which way one reads the text. For when the swerve prevents necessitation of our mental dispositions, neither the mind’s development nor its volitional movements are compelled or necessary.) Hence the passage is consistent with my proposed interpretation of Lucretius. Almost immediately after the quoted passage, still in On Fate 23, a couple of sentences lend further support to the interpretation that Epicurus connected the swerves with the development of a person’s mental dispositions rather than directly with every act of volition. After a remark that Democritus’ position is superior to Epicurus’, because he can do without the swerve, Cicero continues: More astutely, Carneades taught that the Epicureans could have maintained their position without this fictitious swerve. For, seeing that they taught that there could be some volitional movement of the mind, it would have been better to defend that than to introduce the swerve . . . ⁶⁹ (Cic. Fat. 23) ⁶⁹ acutius Carneades, qui docebat posse Epicureos suam causam sine hac commenticia declinatione defendere. nam cum docerent esse posse quendam animi motum voluntarium, id fuit defendi melius quam introducere declinationem . . .

    -  ? 181 Here the introduction of the swerve is contrasted with the existence of volitions of the mind. But if the swerve really was needed in the formation or execution of every volition, we would expect not a contrast but, rather, the claim that a volition can exist without involving a swerve. For the sentence implies that the Epicurean believed that the volitional movements of the mind were at least not directly dependent on their introduction of the swerve. On the other hand, if, as I suggest, the swerve is not concerned with the formation or execution of volition directly, the text as it stands poses no problems. It has also been claimed that a passage in Diogenes of Oenoanda (32.1.14–3.14) is evidence that Epicurus introduced the swerve in order to preserve free will and freedom to do otherwise than we do.⁷⁰ Again, I disagree. Here is the passage: (1) Once prophecy is eliminated, how can there be any other evidence for fate? (2) For if someone uses Democritus’ account, saying that because of their collisions with each other atoms have no free movement, and that as a result it appears that all motions are necessitated, we will reply to him: (3) ‘Don’t you know, whoever you are, that there is also a free movement in atoms, which Democritus failed to discover but Epicurus brought to light, a swerving movement, as he demonstrates from evident facts?’ (4) But the chief point is this: if fate is believed in, that is the end of all censure and admonition, and even the wicked ⁷¹ (Diogenes of Oenoanda 32.1.14–3.14, tr. Long and Sedley)

In this passage Diogenes gives two independent reasons for why not everything is compelled by necessity or fate. The first, in (3), implies that Epicurus introduced the swerve in order to prevent universal necessitation—nothing more—and I assume it to go back to the same arguments of Epicurus’ which Lucretius reports. The second reason, in (4), is that universal necessitation is incompatible with praise and blame—nothing more—and I assume it to go back to Epicurean arguments such as those in the digression of On Nature 25.⁷² Thus the passage corroborates neither the thesis that Epicurus introduced the swerve in order to ⁷⁰ Purinton 1999, 265–6, 299. ⁷¹ [πῶς ἀνῃρημέ]νης οὖν | μαντικῆς σημεῖ | ον εἱμαρμένης ἔστιν | ἄλλο; ἂν γὰ[ρ] τῷ Δημο | κρίτου τι χ[ρ]ήσηται | λόγῳ, μηδεμίαν μὲν ἐλευθέραν [φ]άσκον | ταῖς ἀτόμο[ι]ς κείνη | σιν εἶναι δι[ὰ] τὴν πρὸς | ἀλλήλας σ[ύν]κρουσιν | αὐτῶν, ἔν[ε]ν δὲ φαί | νεσθαι κατ[η]νανκας | μένως π[άντ]α κεινεῖσ | θαι, φή[σομε]ν πρὸς | αὐτόν˙ ‘[οὔκουν] οἶδας, ὅσ | τις ποτὲ εἶ, καὶ ἐλευθέ | ραν τινὰ ἐν ταῖς ἀτό | μοις κείνησιν εἶναι, ἣ[ν] Δημόκριτος μὲν οὐ | χ εὗρεν, Ἐπίκουρος δὲ | εἰς φῶ[ς] ἤγαγεν, παρεν | κλιτικὴν ὑπάρχουσαν, ὡς ἐκ τῶν φαινομέ | νων δείκνυσιν;’ τὸ δὲ | μέγιστον˙ πιστευθεί | σης γὰρ εἱμαρμένης | αἴρεται πᾶσα νουθεσ[ί]α καὶ ἐπιτείμησις καὶ | οὐσὲ τοὺς πονηροὺς [ The adjective ἐλεύθερος seems to be used in the context of physical determinism not before the first century  (see Chapter 1 of this volume [Bobzien 1998b]). ⁷² Purinton 1999 does not translate the δέ (‘but’) in (4), and runs the two reasons together: ‘According to Diogenes . . . the main reason that Epicurus posited the swerve was to preserve the phenomena of “admonition and rebuke” ’. I do not think that the text bears this out.

182

, ,   

preserve freedom to do otherwise, nor the thesis that swerves were involved directly in every volitional act.⁷³

6. Epicurus on Internal Necessity and Character Development Before I offer a conjecture about how the swerve was thought to prevent the internal necessity of the mind, I want to present some further evidence in support of the suggestion that both Epicurus and Lucretius are concerned with the internal necessitation of the development of the mind, and generally with the question of the autonomy of the agent, and not with the free-will problem as set out in section 1. We have a second passage in Epicurus’ On Nature 25 (Arr. 34.24, Laursen 1997, 28) that is concerned with internal necessitation of the mind. It occurs a little before the ones discussed in section 3 (Arr. 34.27): And if the first constitution of the development exerts some kind of compulsion in the mind,⁷⁴ and such a thing is not developed out of necessity to the point of (developing) these specific things, but on the one hand, such a thing is developed from such conditions out of necessity to the point where there comes to be a soul or rather a soul with a disposition and movement of this particular size, on the other hand, such a thing is not developed out of necessity to the point of (developing) a soul of this or that kind, or at least such a thing is not developed with necessity once it proceeds in age, but out of itself or out of the cause out of itself being able to [develop] also something else. (tr. Laursen 1997, 51–2, modified)⁷⁵

This passage is not easy to make sense of, in particular since we lack the immediate context. Nonetheless, it provides some hints as to which things Epicurus thought to be internally necessary and which he did not. That an individual develops a soul and that that soul has a disposition and motion of a particular size are necessary. But the specific qualities of that soul and its specific developments when it (or the person whose soul it is) advances in age are not internally necessitated. Rather, ⁷³ Plut. Stoic. Rep. 1050b–c, too, gives no hint about whether the swerve was meant to come in in character formation or directly in the formation of volitions. ⁷⁴ David Sedley points out to me that it would fit the context much better if one rendered this phrase as ‘and if by the power of thought the first constitution is forcibly separated from the development’, although, as he adds, this has the drawback of taking ἐκβιάζεσθαι plus genitive in an uncommon way. ⁷⁵ Κἂν κατὰ διάνοιαν δέ τι ἐκβιάζηται ἡ πρώτη σύστασις τοῦ ἀπογεγεννημένου, μὴ ἐξ ἀνάγκης μέχρι τωνδί τινων ? ἐξ ἀνάγκης ? τοιοῦνδε ἀπογεννωμένου ἀλλὰ μέχρι μὲν τοῦ ψυχήν γενέσθαι ἢ καὶ τοσαυτηνὶ διάθεσιν καὶ κίνησιν ἔχουσαν ψυχὴν ἐξ ἀνάγκης ? τοιοῦδε ἀπογεννωμένου ἐκ τῶν τοιουτωνί, μέχρι δὲ τοῦ τοιανδὶ ψυχὴν ἢ τοιανδὶ οὐκ ἐξ ἀνάγκης τοιοῦδε ἀπογεννωμένου ἢ οὐκ ἐπειδὰν προβῇ γε τῇ ἡλικίᾳ τοιο ῦδε ἀπογεννωμένου κατ᾽ἀνάγκην ἀλλ᾽ἐξ ἑαυτοῦ δυναμένου καὶ τῆς ἐξ ἑαυτοῦ αἰτίας καὶ ἄλλο . . .

    -  ? 183 when the soul (or the person whose soul it is) advances in age, it will be able to develop from itself, or from the cause from itself. Here, again, internal necessitation and lack thereof are concerned not with individual actions or volitions, but with what a person’s soul comes to be like. The emphasis is on the non-necessity of the development of the mind, and in particular on the fact that we ourselves (or the cause from ourselves) are causally responsible for the changes in our soul, and that these changes are not necessary. The possibility of influencing the development of one’s soul dispositions is allimportant, if what a person’s soul is like at a certain time determines how the person sets out to act at that time. For the only way of getting oneself to act differently from the ways one tends to act is by changing one’s dispositions to act. This holds in particular for moral development.⁷⁶ We become the causes of the changes of our mental dispositions if we receive the right moral education and use our intellect to assimilate this education. A passage in Lucretius confirms that Epicurus had this conception of moral development; it deals with the development of the human mind and is based on the whole-person model of agency: . . . Likewise the human race. Even though education may produce individuals equally well turned out, it still leaves those original traces of each mind’s nature. And we must not suppose that faults can be completely eradicated, so that one person will not plunge too hastily into bitter anger, another not be assailed too readily by fear, or the third type not be over-indulgent in tolerating certain things. There are many other respects in which the various natures and consequently the behaviours of human beings must differ, but I cannot now set out their hidden causes, nor can I find enough names for all the shapes of primary particles from which this variety springs. But there is one thing which I see I can state in this matter: so slight are the traces of our natures which reason cannot expel from us, that nothing stands in the way of our leading a life worthy of the gods.⁷⁷ (3.307–22, tr. Long and Sedley, modified, my emphases)

The relevant points in the passage are these. The initial nature of a human mind includes certain moral dispositions, which are present in different people in various strengths.⁷⁸ Through education people’s minds can develop in such a ⁷⁶ Both before and after the quoted passage of On Nature 25 morality is at issue: Laursen 1997, 23, 26, 29. ⁷⁷ Sic hominum genus est. quamvis doctrina politos | constituat pariter quosdam, tamen illa relinquit | naturae cuiusque animi vestigia prima. | nec radicitus evelli mala posse putandumst, | quin proclivius hic iras decurrat ad acris, | ille metu citius paulo temptetur, at ille | tertius accipiat quaedam clementius aequo. | inque aliis rebus multis differre necessest | naturas hominum varias moresque sequaces; | quorum ego nunc nequeo caecas exponere causas | nec reperire figurarum tot nomina quot sunt | principiis, unde haec oritur variantia rerum. | illud in his rebus video firmare potesse, | usque adeo naturarum vestigia linqui | parvula quae nequeat ratio depellere nobis, | ut nil impediat dignam dis degere vitam. ⁷⁸ This point has a close parallel in the talk of seeds (σπέρματα) in Epicurus, On Nature 25 (Arr. 34.26).

184

, ,   

way that these differences are by and large evened out. The reason is that by the use of our intellect we can modify our mental dispositions to a large extent. This passage corroborates the assumption that the Epicureans worked with a wholeperson model of agency on two counts. First, it makes it clear that Lucretius took a person’s mind to include that person’s character dispositions. Second, it implies that Lucretius thinks that one’s nature determines one’s behaviour, and third, that in order to change one’s behaviour, one has to change one’s nature, that is, the nature of one’s mind, by the use of one’s intellect. The explanation on the atomic level of what happens when we modify our character (and how we become the causes of our dispositions, and consequently actions) seems to be provided by another passage from On Nature 25. (This passage and the next are discussed in more detail in Chapter 5.) In this way whenever something is developed which takes on some distinctness from the atoms in a discriminating way⁷⁹—not in the way as from a different distance—he receives the causal responsibility which is from himself; and then he immediately imparts this to his first natures and somehow makes the whole of it into one.⁸⁰ (Epic. Nat. 25; Arr. 34.22, Laursen 1997, 22)

It seems to me that Bob Sharples is right when he says about this passage: ‘the obvious, indeed inevitable way of interpreting this in the atomic context is to say that we, by thought and effort, can modify our character, and hence also the atomic structure of our minds . . . the downwards causation in the passage . . . may thus relate to the process by which we modify our characters, and not to the explanation of free choice’ (Sharples 1991–3, 186). This passage seems to be concerned with absolutely essential occurrences in people’s mental developments: namely how they themselves become causes first of their dispositions⁸¹ and consequently of their actions. I understand it in the following way. A person may encounter beliefs, including value beliefs, which differ from those they have adopted and developed in line with their initial constitution. These different beliefs are then transmitted to the initial ‘disposition of the soul’ (the ‘first natures’) and made part of it, and as a result the overall disposition is—slightly—changed.⁸² (Thus new beliefs are adopted, and beliefs ⁷⁹ Perhaps: ‘in a way that pertains to judgement’? ⁸⁰ Οὕτως ἐπειδὰν ἀπογεννηθῇ τι λαμβάνον τινὰ ἑτερότητα τῶν ἀτόμων κατά τινα τρόπον διαληπτικόν, οὐ τὸν ὡς ἀφ᾽ἑτέρου διαστήματος, ἰσχάνει τὴν ἐξ ἑαυτοῦ αἰτίαν, εἶτα ἀναδίδωσιν εὐθὺς μέχρι τῶν πρώτων φύσεων καὶ μίαν πῶς ἅπασαν αὐτὴν ποιεῖ. ⁸¹ Whether this is (i) a unique event in a person’s life, or (ii) a gradual process in which a person changes or confirms their beliefs one by one upon reflection, or whether (iii) such events happen all through one’s adult life, is unclear. ⁸² For Epicurus our dispositions to act, and our emotions, are grounded on, and perhaps partially identical with, our beliefs. Remember also that at Nat. 25, Arr. 34.26, the text suggests that that which happens because of us (παρ᾽ἡμᾶς) happens because of beliefs of ours which are from us ourselves (παρὰ τὰς ἡμέτερας ἐξ ἡμῶν αὐτῶν δόξας), see above, section 2.

    -  ? 185 inconsistent with a new belief may be discarded.) At that point, the disposition is no longer (fully) the result of internal and external necessity, but in part the result of conscious, rational, influencing. When, then, someone acts from such a disposition, they are the cause of the action, and no longer ‘the atoms’, i.e. those of their initial constitution. Again, the way things work seems to be: presumably externally induced changes in beliefs lead to changes in dispositions (which in turn lead to changes in behaviour).⁸³ Thus, not only do the Lucretius and Cicero passages on the swerve allow a consistent interpretation of the swerve as having its main function in the formation of one’s mental disposition but there are also several other texts that support this interpretation. In fact, Epicurus’ ethics as a whole is geared to the development of one’s character or mental dispositions, as opposed to a canon of right and wrong actions to choose from.⁸⁴

7. The Atomic Swerve, Mental Development, and Moral Responsibility It remains for me to provide a plausible story about the function of the swerve within the proposed interpretation. In this context I consider the following points. (A) First, I address the two main objections that have been voiced against interpretations that see the swerve as being involved in the formation of character, rather than in each volition. (B) Second, if, as I argue, Epicurus was not concerned with a free-will problem as set out in section 1, and in particular not when he introduced the swerve, I need to show that he had some other real problem, so that the introduction of the swerve does not seem gratuitous.⁸⁵ (C) Third, if such a problem can be identified, it remains to be demonstrated how the swerve could in ⁸³ In Nat. 25 (Arr. 34.31, Laursen 1997, 44–5) Epicurus may refer to the same kind of development; but this passage is rather badly preserved. ⁸⁴ For completeness, I should mention that Cicero presents in a number of short passages in different works some information about Epicurus’ treatment of the Principle of Excluded Middle for future propositions (Cic. Fat. 21, 28, 37; Acad. 2.97; Nat. D. 1.70). These passages show that Epicurus feared that arguments of the family of the ‘Mower’ prove that if the Principle of Excluded Middle held for future propositions, then all future events would be certain or predetermined, and hence necessary. In order to escape this consequence, Epicurus apparently took the step of denying universal application to the principle. (For details see Bobzien 1998a, 75–86). Neither the arguments nor Epicurus’ reply mention action, volition, choice, or freedom of any sort; free decision or free choice are not under discussion, but the non-necessity of future events more generally. This is confirmed by the example for non-necessity that Epicurus chose: ‘either Hermachus will be alive tomorrow, or he will not’. What is at issue is not whether Hermachus will or won’t do something tomorrow, nor whether he will or won’t decide to do something tomorrow, but (assuming that he was not suicidal) whether or not something will happen to him tomorrow—human death being one of those occurrences that were paradigms of fated events which would happen in the form of ‘accidents’, such as drowning at sea, being struck by lightning, dying of diseases, etc. ⁸⁵ As Asmis 1990, esp. at 288, has convincingly shown, Englert’s interpretation suffers from this defect.

186

, ,   

principle be thought of as contributing to its solution—even if we do not know how Epicurus actually envisaged this to work. (A) Here are the two main objections that have been levelled against the view that the swerve, as described in Lucretius, plays a role in the formation of character.⁸⁶ First, it has been repeatedly objected that if one’s character can be randomly altered by the swerve, Epicurus would have problems explaining why people’s characters remain relatively stable.⁸⁷ The point has been memorably illustrated thus: ‘a man of good Epicurean character will live in fear of an unpredictable event which may change him into a Stoic or something worse’.⁸⁸ It is however important to see that this criticism is merely a special case of a more general objection which, in slightly different forms, arises equally for interpretations that consider the swerve as necessary for forming or exerting a volition. The general problem is this: once the swerves have been furnished with a specific function, how can it be ruled out that additional swerves occur that undermine, obstruct, or undo the workings of the swerve as one has determined them? For swerves are by definition random motions and thus can in principle happen at any place at any time. For instance the problem would manifest itself as follows in an interpretation that favours the swerve in the formation of volition in order to preserve freedom of choice: if a swerve is correlated with my forming a volition for performing some action, what if, at the same time, or immediately afterwards, a second swerve counteracts the first and, as a result, I have a volition for its opposite, or at least no longer have a volition for that action? Alternatively, if I utilize one swerve for a volition towards some action, what if, whoops, another one undoes this by rechannelling the atom to produce a volition for its opposite, or to result in no volition either way? Thus any interpretation of the function of the swerve needs to tackle this problem in the particular guise in which it occurs each time. Solutions depend—as far as I can see—mainly on inventiveness. In the case of characterformation interpretations, the criticism can easily be countered as follows. Generally, single swerves go unnoticed, since they are so tiny that they do not interrupt the course of events at the everyday level. In particular, the atomic structure of someone’s mental dispositions is relatively fixed and stable, so that one swerve usually has little chance of doing much damage. It is only in certain developmental situations—which will have to be described in the respective interpretations—that single swerves can contribute to the initiation of a new development in a different direction (see below).⁸⁹ ⁸⁶ I hope to have dispelled the frequently made objection that character-formation interpretations do not square with the Lucretius and Cicero passages in sections 4 and 5 above. ⁸⁷ Long 1974, 61, Englert 1987, 3; similarly Sharples 1991–3, 187 n. 56; Purinton 1999, 275–6. ⁸⁸ Long 1974, 61, quoted by Englert and Purinton. ⁸⁹ Ceteris paribus, proponents of the involvement of the swerve in the formation or execution of volition seem actually to have a harder nut to crack. For in such interpretations a swerve typically

    -  ? 187 The second main objection is this: how can a random movement like the swerve that happens in the course of the development of one’s character introduce and guarantee moral responsibility of the agents in their actions? This, again, is a specific version of a more general problem with which any interpretation of the swerve is confronted. A random motion or a number of random motions cannot by themselves warrant moral responsibility, regardless of whether they occur in the process of developing one’s character, or of forming or exerting a volition, and regardless of whether the concept of moral responsibility is based on freedom of decision or on autonomous agency. (We do not know whether Epicurus was aware of this problem in one form or other, but I would like to think that he was.) This objection can be invalidated by the observation that Epicurus—like most ancient philosophers—generally thought that moral responsibility is based on the fact that agents are beings of a certain kind; namely rational beings who have the capacity to base their actions on their rationality, i.e. their own beliefs. Epicurus’ swerve thus is a necessary but not a sufficient condition for our having volitions and autonomous agency. I sketch below how this could work in the case of character-formation interpretations. (B) My second point was: what was Epicurus’ problem that led him to introduce the swerve? I assume that this was not a ‘timeless’ philosophical problem such as the free-will problem as set out in section 1, but a problem specific to the ancient defence of atomistic philosophies such as those of Democritus and Epicurus. (The usefulness of the swerve is after all restricted to such theories.) Atomists have to explain the entire universe, and everything that happens in it, in terms of atoms and their movements, and the void. The problem is then this: on the one hand, atomists thus have the enormous task of accounting for the order and regularity in the universe;⁹⁰ on the other, they have to explain both the existence of chance events, i.e. disorderly events, and how human beings can be causes or can have volitions. The function of the swerve is thus to provide an explanation of the possibility of chance events⁹¹ and volitions without undermining the atomistic explanation of the order in the universe. The nature of this problem becomes clearer when one follows the ancients in their depiction of Epicurean atomism in two stages: first atomism without the comes in at a point where whether or not a certain decision is made (or action performed) depends fully on whether a swerve occurs. Thus, if one swerve can determine whether or not an action is performed, one further swerve may suffice to counteract the first and lead to the opposite result. ⁹⁰ Cf. e.g. Plot. Enn. 3.1.3. Given a philosophical climate in which teleological views are the norm, this is one of the major challenges Epicurus has to meet. We can gather this also from the fact that Lucretius gives the point a lot of space. The readiness of many modern philosophers to believe in physicalism, mechanism, and reductionism is quite a different scene, in particular, since no empirical sciences such as modern chemistry and molecular physics were available to the ancients. ⁹¹ This is what Philodemus, Sign. 36.11–17, and Plut. Soll. An. 964e say—and (pace Bailey 1928, 327 and Long 1977, 63–88) I cannot see why this should not be what Epicurus said.

188

, ,   

swerve; then atomism with the swerve. Whether the first stage was thought to be historically real, or fictitious and merely an explanatory device, is immaterial in this context. Atomism without the swerve is designed to meet the challenge to give a nonteleological, mechanical, explanation of the order in the universe. In atomism without the swerve every movement would be necessary, both at the atomic level and at the everyday level. There are two kinds of necessitating factors, internal and external ones. At the atomic level these are internally the weight of an atom, and externally other atoms that collide with this atom. At the everyday level, concerning the movements of everyday objects, they seem to be the weight and atomic composition and structure of the object as internally necessitating factors, and atoms and things or clusters of atoms external to the object, which ‘collide’ with it (and might enter it), as externally necessitating factors. The order and regularity in the world are explained as the result of the cooperation of these two factors. Such explanation proves most difficult in the case of objects of complex structure like plants and animals, where phenomena such as reproduction and self-motion need to be accounted for. Plants and animals display stable properties, stable patterns of behaviour, and patterns of development that follow a fixed temporal order; for example, all animals have a soul, birds generally build nests, all boys sprout a beard when approaching manhood.⁹² The properties are explained by the types of atoms involved and their structural combinations and patterns of movements. The patterns of behaviour are explained as reactions to external stimuli that are determined by the atomic structure and movement of the object. The developmental patterns (for the development of both properties and dispositions) require some inbuilt ‘time release’ in the atomic base of things, which will be activated by suitable external stimuli. An atomistic theory can, thus, in principle explain the order and complex structure of the world. In this swerveless atomic theory the following difficulty arises. Being entirely the result of internal and external necessitating factors, all motions—both at the atomic level and at the everyday level—are necessary. But it was generally accepted among Hellenistic philosophers that it is an essential characteristic of voluntary behaviour (and of chance events) that they are non-necessary. Hence in the— swerveless—atomic system voluntary behaviour appears to be precluded. (C) The swerve is introduced to solve this problem by satisfying two conditions: positively, it needs to make voluntary action possible by preventing internal necessitation of certain states and movements; and, negatively, it must not undermine the explanation of the order and regularity in the world in terms solely of atoms and void. If one assumes that the swerve’s function—as described in

⁹² Lucr. 5.849–54.

    -  ? 189 Lucretius and Cicero—is to remove internal necessity from the agent’s mental dispositions, we can imagine this to work as follows. I start with a note on the frequency of the swerves—a point on which our sources are silent.⁹³ Since the swerve is a third basic motion, and none of the texts about the swerve mentions a particular scarcity of swerves, we may assume that swerves happen quite frequently. However, the frequency is limited by the fact that the everyday-level world appears generally ordered (although arguably less so to ancient philosophers than to Newton, say). Thus we can state, as it were, an upper and a lower limit for the frequency of the swerves: on the one hand, the swerves must occur sufficiently often to guarantee the possibility of a frequency of chance events and non-necessitated character developments (especially character changes) that corresponds to the frequency of such kinds of occurrences as we can ‘observe’ them happening; on the other hand, relative to the size of the deviation of a swerve (the elachiston) and the time it takes, the force of a single swerve and the number of swerves must be restricted in such a manner that they do not interfere with the order of things at the everyday level. For instance, the reliability of sense perception has to be preserved, and those kinds of developments that show great regularity must not be interfered with. (The theory of quantum leaps may give us an idea of how this could in principle be possible.) Next, the removal of necessity. What the introduction of the swerve does to the modalities of the atomic movements is uncertain. We can assume the following points. (i) Swerving movements are not necessary. (I assume, however, that they are caused by the atom, whose nature it is to swerve randomly every now and then—within the bounds of frequency determined above, although when these swerves happen is not fixed in advance. Atoms may thus be conceived of as having a built-in random generator as part of their nature.) (ii) Any movement that is neither itself a swerve nor in any way the close or remote effect of a swerve is necessary (but given the eternity of the universe, it is uncertain whether there are such movements). (iii) Certain kinds of movements are impossible, for example swerves that exceed one minimum and ‘upward’ movements if there is no collision involved. (iv) This leaves the majority of atomic movements, those which are not swerves but have a swerve somewhere in their ‘causal history’. I am inclined to think that at least all such movement as would not have occurred if one of the preceding swerves had not occurred would be non-necessary, assuming some sort of transitivity of non-necessity. What the modalities of movements at the everyday level are is equally underdetermined by our sources. I assume that they roughly correspond to those of the atomic movements. (i) All those events or changes of an object in which no swerve has been causally effective are necessary. These will include facts such as that ⁹³ Except if one accepts Kleve’s (Id facit exiguum clinamen, 28) argument that Lucretius’ etiam atque etiam (2.243) means ‘again and again’ and refers to the frequency of the swerves.

190

, ,   

people will die by a certain age (since the clusters of atoms that make up humans just ‘give in’ after a certain time, and no swervings can prevent this from happening); and that, given certain circumstances, human beings desire pleasure and shun pain. (ii) All those events or changes that concern an object in which one or more swerves have been causally effective somewhere in their ‘causal history’ are non-necessary. Thus, whenever in a person the mental dispositions have been changed as a result of a swerving, all subsequent behaviour of that person that is (internally) co-caused by the part of the overall disposition thus changed is nonnecessary. (Here it does not matter whether it would also have happened without the change; what matters is what was causally effective in its coming about.⁹⁴) We can then see why on this interpretation the swerves that happen in the mind do not lead to chaos and why Epicureans have little reason to fear that they may suddenly metamorphose into a Stoic or something worse. We have to imagine the mind as a relatively stable atomic structure but with a large number of developmental ‘potentials’ (spermata, Epic. Nat. 25; Arr. 34.26) with built-in time release, which ensures that certain developments do not happen before a certain age (certain potentials are not actualized before a certain age), or before other developments have happened. We can then imagine that at a time when new developments are due to start, swervings can be effective much more easily, determining the direction the development takes (out of a number of possible directions or ‘trajectories’), whereas once a new property or characteristic is fully developed, single swerves are not strong enough to make a difference. Thus most swerves that occur in the mind do not affect the mind’s structure and thus a person’s mental dispositions. If we postulate a sufficient number of swerves, and a sufficient number of developmentally sensitive periods for individual potentials (we could even assume that a swerve is a necessary condition for triggering a new development in such sensitive periods), most specific mental dispositions, and hence most human behaviour, will become non-necessary, it being in part the result of a causally effective swerve. We can also now explain how such random modifications of the initial state of our mind could make us any more causally or morally responsible for our movements than if they flowed directly from our initial make-up. As hinted above, the answer is roughly that the swerve is no more than one of many necessary conditions for moral responsibility. All the other necessary conditions have to be found in the nature of those beings who are to be held responsible. (This point can also be inferred from the fact that swerves serve to explain chance events as well,⁹⁵ since there must then be factors other than the swerve that account for the

⁹⁴ See Epic. Nat. 25, Arr. 34.24–5, Laursen 1997, 29–30. ⁹⁵ Philodemus, Sign. 36.11–17, and Plut. Soll. An. 964e. Purinton’s reading of the two passages as referring only to the chance existence of the cosmos (1999, 261–2) is unconvincing.

    -  ? 191 difference between chance events and those events for which we are causally and hence morally responsible.) Swerves can lead to the development of the power of volition and to moral responsibility only in the right kind of things. My point is not that because of the fineness of the mind-atoms swerves initiate noticeable changes only in the mind, and that this explains why humans are morally responsible but plants, for instance, are not. For even if the swerves can have effects in the mind more easily, this does not rule out that, at the right point in the development of a plant, or if occurring in a sufficiently large number, they could also, for example, change the constitution of a plant. Imagine that a ‘cluster of swerves’ in some part of a plant brings about a change in the plant’s constitution. Imagine that the plant now grows an extra leaf at the wrong place. This would hardly make the plant morally responsible for the changes. Nor would one hold the whole plant itself causally responsible. Rather, for Epicurus such a change would be by chance (apo tuchēs).⁹⁶ For an action that happens because of us (par’hēmas), i.e. of which we are the cause, a lot more than this is required: only those beings can become morally responsible that come with a primary constitution of a certain complexity. Human beings do—and arguably some tame animals. The initial make-up of human beings is categorically different from that of plants. We are—or so Epicurus thinks—from birth led by pleasure. Plants are not. From birth, too, we have a large number of potentials for developments. Plants do not. It lies in our atomic structure that we will develop a mind, preconceptions, memory, the ability to speak a language, and many more things which plants do not develop. And presumably we have to develop quite a bit before the swerves can exert their catalytic function in the process that leads to the development of a capacity for fully fledged volitions by means of which we can become the cause of our actions. I surmise that Epicurus held that we ourselves become causes and morally responsible at the point when we start changing our disposition by way of developing our own thoughts, value judgements, and so on, that is, when we become capable of reconsidering the judgements and desires we have adopted or developed solely as a result of hereditary and environmental impacts. Such reconsideration influences what we regard as pleasant, and thus in what circumstances we have a volition for something and thereby initiate an action. I further suspect that this is what Epicurus is concerned with in much of On Nature 25.⁹⁷ There are several different ways in which one can conceive of the relation between, on the one hand, the swerves’ prevention of internal necessitation of our mental dispositions, and on the other, the way we become (and remain) ourselves the causes of our actions, so that the actions happen because of us and we are ⁹⁶ See e.g. Lucr. 4.1223–6 for random variation in heredity. For a similar use of the expression ἀπὸ τύχης see also Aristotle and Alexander of Aphrodisias’ remarks on humans with six digits, etc. ⁹⁷ See in particular Epic. Nat. 25, Arr. 34.22, quoted above in section 6, with my comments.

192

, ,   

morally responsible for them. I give the two possibilities which I find most plausible; there are others. I do not maintain that either of the two was Epicurus’. I merely intend to show that such explanations, which are based on a character-formation interpretation, are possible and that they are no more fantastic than the ones suggested by those who think that the swerve was meant to solve the free-will problem as set out in section 1. The first possibility is what one may call the minimalist approach: Epicurus’ change of his conception of atomic movement to the effect that in addition to their natural downwards movements atoms have swerving movements introduces an element of non-necessity into the world. This solves the problem of autonomous agency as follows: adult human beings can be called the cause of their actions because (a) they influence the development of their character by way of their rationality or reason; (b) what their character is at any time is not a function of hereditary internal and circumstantial external influences, since swerves occur in the mind in the process of its development and have at crucial developmental moments effects on the changes of the atomic structure of the mind; this development, the resulting mental dispositions, and the actions flowing from those dispositions, are thus not necessary. (c) It therefore makes no sense to call necessity, or the atoms (at preceding times), or the initial constitution of the mind, the cause of the action. The person can be called the cause of their actions, and can be held morally responsible, because of (a) and (b), which however are not necessarily connected.⁹⁸ The second possibility is some kind of correlation approach: there is a direct correspondence between (some of the) swerving movements in a person’s mind and their conscious effort to change their mental dispositions. For instance, I study Epicurean ethics and, aiming at tranquillity (ataraxia), I try to adopt and follow the theory, try to change my old ingrained beliefs and prejudices, etc. In this context, making a conscious effort to believe firmly that P, and to replace non-P by P, happens—at least occasionally—simultaneously with swerving motions of mind atoms; and when it does it leads to a change of my character dispositions (the main constituents of which are beliefs) by integrating P into my system of firmly held beliefs. Exactly how the randomness of the atomic level can be the foundation of change is not explained in our texts. Perhaps Epicurus could not answer this. Agents are the cause of their actions because their mental dispositions, which determine their volitions and actions, are the result of the agents’ own changing of (or confirmation of) their beliefs.⁹⁹

⁹⁸ This approach is close to Furley’s position in his 1967; it also utilizes some of Conway’s suggestions in his 1981. ⁹⁹ This is a modification, in the light of the character-formation interpretation and the whole-person model of agency, of interpretations such as those by Sharples 1991–3; Mitsis 1998, 164–6; Gulley 1990, 45–6; Asmis 1990, 291.

    -  ? 193

8. Conclusion My overall goal in this chapter was this: to show that there is no compelling textual evidence for the assumption that Epicurus was concerned with freedom of decision or choice or with a problem of free will as set out in section 1. There is no evidence that he discussed, or even had a conception of, freedom of decision or freedom of choice. There is no evidence that he had a concept of moral responsibility that is grounded on freedom of choice, or on freedom of decision. There is not even any direct evidence that he thought that freedom to do otherwise was jeopardized by atomistic determinism. There is further no compelling evidence that the swerve played a role in the formation of volitional acts or decision processes. I hence suggest that the whole idea that Epicurus was concerned with the free-will problem as set out in section 1 is anachronistic, and that—at least as long as no positive evidence comes to light—the view that Epicurus thought there was such a problem, and that he endeavoured to solve it, should be dropped. I have attempted to draw an alternative picture, based on the evidence we have. This picture suggests that Epicurus—in line with philosophers before and after him (Aristotle and the Stoics)—had a different concept of human agency and of moral responsibility: human actions are fully determined by the mental disposition of the agents when they set out to act. Moral responsibility presupposes not free decision or free choice, but both the absence of coercion and autonomous agency, i.e. that the person, and not something else, is causally responsible for the actions for which they are to be held morally responsible. Autonomous human agency requires the ability of the agents to influence causally, on the basis of their own beliefs, the development of their behavioural dispositions. In the context of his mechanistic atomism, Epicurus faced the problem that he had to explain the non-necessity of human agency (and chance events) without undermining the atomistic explanation of the order in the universe. The swerve— which is of use only within atomism—was meant to solve this problem by making the mental dispositions of adult human beings non-necessary (perhaps by allowing a person’s rational attempts at altering their dispositions to gain a foothold). This is possible without great interruptions and ‘out of character’ developments if one assumes a certain frequency of the swerves and a generally stable atomic structure of the mind, which is susceptible to influence by single swerves only when new developments or realizations of potentials are about to start. How exactly this was meant to work in detail we do not know.

7 Stoic Conceptions of Freedom and Their Relation to Ethics Early Stoics, Epictetus, Late Stoics

1. Introduction In modern discussions of freedom in Stoic philosophy we often encounter the following assumptions: (i) the Stoics discussed the problem of free will and determinism;¹ (ii) since in Stoic philosophy freedom of the will is in the end nothing but an illusion, the Stoics understood the freedom of the sage as a substitute for it, and as the only true freedom;² (iii) in the five hundred years or so of live Stoic philosophical debates, the Stoics were by and large concerned with the same philosophical problems of freedom. In the following pages I argue that (i) can be upheld only in a rather restricted way, that (ii) is altogether untenable, and regarding (iii) that although there may have occurred little change in the Stoic philosophical position on freedom and related concepts over the centuries, we can detect more than one transformation of the philosophical problems that were at the forefront of the discussion. All the conceptions and problems of freedom were linked in one way or other to Stoic ethics, and the differences between them become particularly transparent when one considers their various roles in this context. I shall be concerned with three different stages in Stoic philosophy, in each case looking at how the Stoics of that period deal with freedom and affiliated concepts, and how they relate to ethics. These stages are, first, the early Stoics, in particular Chrysippus; then Epictetus; and finally, some later Stoics, presumably of the second century. I shall not discuss the question of the development from one

This essay first appeared in R. Sorabji (ed.), Aristotle and After (London: Institute of Classical Studies, 1997), 71–89, and is reproduced here with permission. An early version of this essay was presented to the Internationales Kolloquium zur praktischen Philosophie der Stoiker in Bern, 1992, and I have benefited from the discussion of the paper and from the comments Michael Frede offered in his reply. I am indebted to Jonathan Barnes for his suggestions on some of the issues I have dealt with, to Christopher Gill, who drew my attention to an imprecision in my presentation of Epictetus, and to the 1997 volume editor, Richard Sorabji, for several sheets of written comments. ¹ Cf. e.g. Gercke 1885, 698; Pohlenz 1959, vol. 1, 104–5 and vol. 2, 60; Sambursky 1959, 57ff.; Huby 1967, 358–9; Forschner 1981, 98; Mackendrick 1989, 200–2; Hankinson 1995, 101. ² See n. 54 for proponents of this view.

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0008

   :    

195

stage to the next. I shall simply show that the stages differ noticeably from one another, and how they differ. Each stage faced a philosophical problem of freedom (or even several), and these problems—although partly related—were not the same. In the first stage, for instance, the topic of the freedom of the sage and that of the kind of freedom which is a necessary condition for moral responsibility were totally unrelated. And—contrary to general opinion—the ‘modern problem of free will (or freedom of decision) and determinism’ can be detected in ancient philosophy with some certainty only in the last of the three stages. One reason why these points have often been overlooked is easily explained: the sources are usually examined from a contemporary point of view, with a contemporary conceptual apparatus and contemporary philosophical problems in mind, instead of an approach from within Greek terminology and the conceptual system of the Stoics themselves. If one wants to discuss Stoic freedom, the first problem that one encounters is one of terminology. The English word ‘freedom’ is used in very different philosophical contexts. We talk about freedom of choice, freedom of decision, freedom of action, political freedom (liberty), freedom of indifference, physical and metaphysical freedom; about free will, freedom of the will, being free, and having freedom.³ What kind(s) of freedom did the Stoics talk about then, and what words, terms, and expressions were available to them? The only Greek noun that is standardly translated as ‘freedom’ is eleutheria, together with its cognate adjective eleutheros, ‘free’. These expressions no doubt denote freedom in some sense, and they play an important role in Stoic philosophy. However, in our sources for Hellenistic philosophy eleutheria⁴ never occurs in the context of the debate over freedom as a necessary condition for moral responsibility and purposeful action. What words were used in the Hellenistic discussion of this problem? As a matter of fact, there appears to have been no noun for ‘freedom’ in this context. Instead, all we have are a couple of propositional phrases: ‘eph’hēmin’ and ‘par’hēmas’: I translate them both as ‘depend(ing) on us’.⁵ The first expression, ‘eph’hēmin’, is familiar from Aristotle’s Nicomachean Ethics (EN).⁶ Later it gained

³ The Italian libertá, the French liberté, and the German Freiheit have comparable ranges of use. ⁴ I transliterate eleutheria and eleutheros, rather than translating them, since one purpose of this chapter is to keep clear the various concepts of freedom of the ancients, and not to evoke incongruous connotations, as the word ‘freedom’ easily does. ⁵ The option of transliteration is not really open in this case, since we do not know which expression(s) the early Stoics used, cf. n. 10. The translation ‘depend(ing) on us’ is not ideal, but for my purposes it is preferable to the alternatives ‘up to us’, ‘in our power’, and ‘attributable to us’, since it shares an ambiguity present in the Greek eph’hēmin. Eph’hēmin can convey the sense both of something’s being ‘up to someone’ and of something’s being ‘dependent (e.g. causally) on someone’. There is a tradition of rendering or understanding to eph’hēmin and even to par’hēmas as ‘free will’ or ‘freedom’ (so for instance Sambursky 1959, 61; Theiler 1946, 85–6 with n. 164; Talanga 1986, 150; ‘Chrysippos’ in Der kleine Pauly 1170; Dobbin 1991, 124; and Dillon 1993, 35). We shall see that at least in the case of the Stoics this is rather misleading. See also chapter 1, this volume, for antiquity in general. ⁶ Arist. EN 3.3 and 3.5 (see also chapters 1 and 2, this volume).

196

, ,   

general acceptance as the standard way of referring to the kind of human self-determination at issue in the discussions of the compatibility of moral responsibility with determinism.⁷ The second phrase, ‘par’hēmas’, is more common in Epicurean contexts.⁸ We do not know for certain whether Chrysippus used either (or both) of these expressions. I consider it likely that he used the phrase ‘eph’hēmin’.⁹ If we retranslate into Greek from Cicero’s report from Chrysippus, it appears that ‘eph’hēmin’ was used as a prepositional phrase in predicative position: eph’hēmin esti, ‘(something) depends on us’—just as we find it earlier in Aristotle (EN 3.5). Only much later do we encounter a substantivation of the phrase, obtained by pre-positing a neuter article: ‘to eph’hēmin’, ‘the depending on us’ or ‘that which depends on us’ (see section 4). The lack of such a noun phrase in the context of moral responsibility and determinism in early Stoic philosophy may be an indication that this concept of that which depends on us was not itself a subject of the philosophical discussion. Two further differences in terminology should be noted when one deals with the Stoics on ‘free will’ and moral responsibility. First, as is well known, the early Stoics had no noun expression for ‘will’ (as in ‘free will’), presumably since they had no concept of a separate faculty of the will.¹⁰ So whatever the Stoic problem of freedom and responsibility was, it was not centred around some faculty of the will. Secondly, they had no term for moral responsibility. Instead of using an abstract phrase, moral responsibility was talked about in terms of attributability of praise and blame, punishment and reward. This is a shared feature in Aristotle, Epicurus, and the Stoics.¹¹ Epictetus occasionally uses the term ‘accountable for’ (hupeuthunos) in a way that suggests moral accountability might be at issue (Diss. 1.12.32–5). Thus we can see that the linguistic tools available to the Stoics were quite different from ours. Accordingly, we should expect that their phrasing of the problems, their thoughts and argumentation, and also their concepts may sometimes have no straightforward parallels in modern philosophical theories of freedom and responsibility. This is of some relevance to what follows. ⁷ Cf. e.g. Alex. Fat.; Epict. Diss.; Porphyry in Stobaeus, Ecl. 2.163–73 (Wachsmuth); Nemesius, Nat. hom. 104–20 (Morani). ⁸ E.g. DL 10.133; Epic. Nat. 25, Arr. 34.26.6, 12; Philodemus, Sign. 36.14. In the first passage Epicurus uses ‘without master’ (adespoton) to describe that which depends on us; cf. also n. 30 below. ⁹ This is the standard term in later Stoic philosophy (cf. e.g. Epict. Diss. 1.1; Marcus Aurelius 6.32; Alex. Fat.); moreover, the Latin phrases ‘in nostra potestate’ and ‘(sita) in nobis’, which we find in discussions of Chrysippus’, and in general of Stoic, compatibilism—e.g. in Cicero (Fat. 40, 41, 43), Gellius (NA 7.2.15), and Augustine (De civ. D. 5.10)—are most probably translations of the phrase eph’hēmin. On the other hand, in Diogenianus’ presentation and criticism of Chrysippus’ books on fate, we find par’hēmas (Diogenianus in Eusebius, Praep. Ev., e.g. 321.1, 323.1, 326.22, 327.9). This suggests that either Chrysippus used par’hēmas (in addition to or instead of eph’hēmin) or Diogenianus made use of Epicurean terminology in these passages. ¹⁰ Cf. Dihle 1982; Voelke 1973; Kahn 1988, 234–59. ¹¹ E.g. EN 1109b30–35. DL 10.133–4, Gellius, AN 7.2.4–13. In the same passage in Diogenes, Epicurus calls necessity ‘not accountable’ (anupeuthunon), but I doubt that moral accountability is intended here.

   :    

197

2. Early Stoics: Zeno, Cleanthes, Chrysippus 2.1 The Early Stoic Concept of That Which Depends on Us In the time of the early Stoics—in fact it seems from Aristotle to Carneades—it was generally assumed that a condition for holding a person responsible for an action (or omission) is that this action depended on that person.¹² There is no evidence in the period of a discussion or dispute over which kinds of things depend on someone. Apparently there was agreement that human actions (and omissions) in general depended on the agents. Accordingly, we do not find a philosophical (or other) definition of that which depends on us.¹³ There is no reason to believe that ‘eph’hēmin’ and ‘par’hēmas’ were technical terms, and it seems that the meaning of the phrases in these contexts was taken to be unproblematic and generally understood. The reason why Chrysippus nonetheless dealt with the concept of that which depends on us is that the Stoics had run into the accusation that their theory of fate would conflict with the fact that there are things that depend on us. The Stoic doctrine of fate (heimarmenē) is basically a theory of universal causal determinism (combined with a teleological element).¹⁴ Crudely simplified, it states that everything that happens does so in accordance with fate (DL 7.149), or that everything is fated, with fate defined as the network of causes in the world (Nemesius, Nat. hom. 108.15–17). The claim that everything is fated is hence understood as meaning that every event has causes that precede it and because of which it occurs (Cic. Fat. 21.41–2, [Plut.] Fat. 574e). Moreover, the Stoics maintained that that which will happen in the future is already determined in the present and, what is more, was always determined in the past (Cic. Div. 1.125–6, Fat. 27; Stobaeus, Ecl. 1.79). The conflict between this Stoic ‘causal fate-determinism’¹⁵ and the claim that there are things that depend on us is then as follows: when something happens in accordance with fate and by virtue of preceding causes, then it happens no longer by virtue of ourselves. Hence it is not in our power, does not depend on us, but

¹² The Stoic theory of causal determinism and the arguments of the Stoics and their opponents concerning the compatibility of this theory and moral responsibility, which I present in sections 2.1 and 4, are discussed in more detail in Bobzien 1998a (and cross refs). ¹³ wExcept if one reads the characterization of that which is par’hēmas in Diogenianus in Eusebius, Praep. Ev. 326.21–3 as a definition and as Chrysippean: εἴ γε καλεῖν προείληφεν [i.e. Chrysippus] . . . παρ᾽ ἡμᾶς δὲ ὅσα ἔκ τοῦ σπουδάζειν ἡμᾶς καὶ ἐνεργεῖν ἐπὶ τέλος ἔρχεται ἢ παρὰ τὸ ἀμελεῖν καὶ ῥᾳϑυμεῖν οὐκ ἐπιτελεῖται. Aristotle, although not presenting a definition, states more than once that in situations in which it is eph’hēmin to act, it is also eph’hēmin not to act; cf. e.g. EN 1113b7–8, i.e. he had what I will call a two-sided concept of that which is eph’hēmin. As we shall see, this feature is not implied by the Stoic concept of that which depends on us. ¹⁴ Cf. Stobaeus, Ecl. 1.79.1ff.; Plut. Stoic. Rep. 1056c, Comm. Not. 1076e. ¹⁵ I include the reference to fate in this expression as a reminder of the teleological element that is part of Stoic determinism.

198

, ,   

depends on fate. And hence it becomes impossible to attribute praise and blame to us. A typical argument that frames this objection can be found in Gellius: If Chrysippus . . . believes that all things are set in motion and ruled by fate, and that it is not possible that the ways and coils of fate are bent or transcended, then the sins and misdeeds of human beings too should not . . . be attributed to them and their volitions, but to a certain necessity and perseverance, which arises from fate . . . ¹⁶

In response to such arguments, Chrysippus set out to demonstrate that although everything is fated or causally predetermined, nonetheless some things depend on us.¹⁷ Of Chrysippus’ various counterarguments and defences of the Stoic position, the following two are of special relevance to our understanding of his concept of that which depends on us, and of his compatibilist position. First, he argues that the fact that everything is fated does not entail that every event is necessary or necessitated, nor that what does and does not happen depends completely on the impact of external forces and hindrances. This he does by making use of his modal theory. Secondly, Chrysippus shows that and how actions (and certain mental events), i.e. the things generally considered to depend on us, depend on us, despite their being fated. This part of his argumentation belongs to the area of psychology. M. First, then, a few words about Chrysippus’ use of his modal logic. Chrysippus endeavoured to show that the fact that something is fated to happen does not make it necessary, and that the fact that something is fated not to happen does not make it impossible (Cic. Fat. 13). In other words, he argued that his theory of fate is not deterministic in the sense that everything that happens is necessary and everything that does not happen is impossible. His concepts of possibility, impossibility, necessity, and non-necessity are defined in such a way that those things that are commonly considered to depend on us come out as neither impossible nor necessary; that is they come out as contingent—despite their being fated. What is more, the modal concepts bring out one necessary condition for something’s depending on us. Chrysippus defines the concept of possibility and non-necessity by giving two conditions for each. Following Diogenes Laertius 7.75,¹⁸ that which is possible

¹⁶ Gellius, NA 7.2.5; see Cic. Fat. 40 for a similar argument against the Stoics. ¹⁷ This he did in his second book on fate. Cf. Diogenianus in Eusebius, Praep. Ev. 325.26–8. ¹⁸ We can be confident that Diogenes reports Chrysippus’ definitions, for we know that Chrysippus developed his own modal concepts, which were meant to improve on Diodorus’ (Cic. Fat. 12–14), and in Plutarch (Stoic. Rep. 1055d–f) we find remnants of Diogenes’ accounts—with identical formulations—ascribed to Chrysippus. For a detailed discussion of Stoic modalities, see Frede 1974, 107–17; Bobzien 1993, 63–84.

   :    

199

(i) must be capable of being true (of happening)¹⁹ and (ii) must not be prevented by external things from being true (from happening); and accordingly, that which is non-necessary (iii) must be capable of being false (of not happening) and (iv) must not be prevented by external things from being false (from not happening); i.e. must not be forced to be true (to happen). There are then four conditions for contingency: two concerning general ‘capability’, two concerning ‘external things or circumstances’. Something is contingent precisely when it is capable of happening and not happening, and when it is neither prevented by external things from happening nor forced to happen by such things. That which is contingent includes most human actions, i.e. the things usually considered to depend on us. For instance, if it now depends on me to walk, then it is now possible that I walk; and it is not necessary that I walk, and hence it is also possible that I do not walk. It is important to see that this ‘two-sided’ possibility (to walk and not to walk; see Chapters 1 and 2) is not the same as the claim that when I then walk there was no cause for my walking. (It also does not entail that if it depended on me to walk, it also depended on me not to walk. Rather, Chrysippus’ concept of that which depends on us appears to have been ‘onesided’ only.)²⁰ All that is maintained is that I am capable of both walking and not walking, and that in the particular situation in which I am, I am neither externally forced to walk nor externally hindered from walking, for example I am not chained to a wall or I have not passed out, etc. Nonetheless, according to Chrysippus there are causes for my walking, for instance that I had the sense impression of a Renaissance church in the distance, and—because of my interest in Renaissance architecture—I thereupon decided to go and have a closer look at it. This fits in with Chrysippus’ causal fate-determinism as follows: if every movement or change of an object were completely causally determined by causes ¹⁹ In Diogenes Laertius, the modal accounts are given in terms of truth and falsehood, and the modalities are understood as properties of Stoic propositions (axiōmata). However, in the context of the discussion of fate-determinism and freedom, modalities were also used of events or that which happens (ginomena) and of that which does not happen (mē ginomena) and of causes (e.g. Cic. Fat. 13, 42, 44; Alex. Fat. 176.14–18; Augustine, De civ. D. 5.10). In the following I disregard various subtleties of Stoic ontology and simply assume e.g. that the proposition ‘Diotima walks’ is possible precisely when it is possible that Diotima walks, that ‘Diotima walks’ is capable of being true precisely if Diotima is capable of walking, and that ‘Diotima walks’ is prevented by external things from being true precisely when Diotima is prevented by external things from walking. ²⁰ This is at least suggested by his claim that that which depends on us (that which is par’hēmas) is encompassed by fate. Cf. Diogenianus in Eusebius, Praep. Ev. 327.4–5. [Plut.] Epit. 1.27 (Diels, Doxogr. graec. 322.9–13). See also below, section 4.

200

, ,   

external to that object, and the object were hence in some sense externally forced to change, then every event would not only be fated but also necessary. However, there are changes or movements that are not completely causally determined by outside causes and which are the kind of change that, as far as the nature of the object is concerned, in general could or could not take place. For Chrysippus, these changes are, although caused, not necessary. That is, for Chrysippus ‘being fated’ is not synonymous with ‘being externally necessitated’ or ‘being externally forced or hindered’. There is a realm of contingent events, of events which are neither forced to happen nor hindered from happening, and this is where the things that depend on us belong. P. Chrysippus’ second argument that furthers our understanding of the early Stoic concept of that which depends on us is based on his explication of how what is fated can, nonetheless, depend on us. Chrysippus set out to show how it can be that, even though every event is causally determined in every detail, the things that were traditionally regarded as depending on us are still in some sufficient sense ours, so that we cannot pass the buck to fate. (Recall the Gellius passage quoted above.) What Chrysippus provides is in fact a causal analysis of the psychological process that takes place in the case of human actions. Again strongly simplified, Chrysippus’ analysis of action is as follows: an ‘impulsory’ or impulse-prompting impression (phantasia hormētikē), i.e. an impression of something as desirable, occurs in a human being’s soul.²¹ It is, at least ultimately, triggered by something external to the soul. This impression sets in motion the faculty of assent in that it suggests that the human being should assent to it. But—unlike non-rational animals—human beings have in principle the ability to withhold assent to incoming impressions. Whether or not a person yields to the impression by assenting to it depends on that person’s state of the soul. In the case that assent is given, the impression is a cause of it insofar as it activated the faculty of assent. Such an assent to an impulse-prompting impression automatically brings about—or according to one source, is nothing but—a human impulse or intention (hormē), and this intention is then translated into an action (praxis).²² To take a simple example, the impression of a bar of chocolate as something desirable (‘I should eat this chocolate’) activates someone’s faculty of assent, so that, in response, it either yields to this impression (response: ‘yes’) or does not yield to it. In the case of a positive response, this automatically sets in motion—or is nothing but—the intention to eat the chocolate. This intention in turn will make the person start and carry through the action of eating the chocolate in due course. Thus, although there is a cause of the action that ²¹ For the Stoics all impressions human beings have are rational impressions, i.e. they have propositional structure. An impression that presents something as desirable may be represented along the lines of ‘I should pursue . . . ’. ²² Cf. Cic. Fat. 40–3; Plut. Stoic. Rep. 1037f and 1057a–c; Stobaeus, Ecl. 2.86–8. For a thorough study of Stoic theory of action, see Inwood 1985.

   :    

201

ultimately comes from outside, the action is not completely externally determined. For it is the faculty of assent in the human mind that determines whether the person gives in to the impulse-prompting impression.²³ Now the faculties of assent, and in general the souls of different individuals, are in different states or conditions. It is for this reason that, when receiving comparable external impressions, one person may give, another withhold assent, depending on the state of their souls: for instance, one may be very greedy, the other more restrained.²⁴ Thus, although the externally induced impression may be the only active cause, the result does not exclusively depend on this cause, but rather on the persons themselves. The person is the originator of the resulting action. The action hence depends on that person, who can then be held responsible for it.²⁵ However, at the same time it is true that the person’s soul, the externally induced cause, and presumably the external circumstances,²⁶ all taken together determine the outcome completely. They all are part of the causal network that is fate.²⁷ Thus Chrysippus has shown that the Stoic concept of causal fate-determinism is compatible with the fact that there are things that depend on us. The conflict between fate and things that depend on us has turned out to be only apparent: we are part of the causal network that is fate, and the actions we do are not determined by external factors, or force, or necessity, but by the people we are. For the attribution of moral responsibility it is thought sufficient to trace an action back to the person that acted, or more precisely to that person’s soul or character.²⁸ The concept of freedom involved in this early debate on fate and responsibility is that of not being externally determined to act, or of not being in any way forced to act. But note that this concept of freedom is not the same as the concept of that which depends on us. An action depends on me if (having given assent to the corresponding impulse-prompting impression) I bear the chief causal responsibility for it and am in this sense its originator. It is in order for this to be possible that I must not be compelled to act or externally prevented from acting, i.e. that I must be free from necessitating and hindering influences. Such a concept of freedom is implied both by Chrysippus’ modal theory and by his defence of that which depends on us, but it is never discussed directly in these contexts.

²³ Cf. Gellius, NA 7.2.7–9. ²⁴ I take it that in Cic. Fat. 42–3 the two examples of the cylinder and the spinning top are chosen as analogues of two human souls or minds that are in different states, and hence would move differently, despite having comparable impressions. ²⁵ Cf. Gellius, NA 7.2.11; Cic. Fat. 42–3. ²⁶ The external circumstances have to be ‘favourable’, as it were. As Chrysippus’ modal theory makes clear, external activating cause and assent alone are not sufficient for an action’s depending on us. In addition, possible hindering and forcing circumstances have to be absent. ²⁷ Cf. Gellius, NA 7.2.7–10. ²⁸ The person’s state of soul and character are of course in the end externally determined. But that is a different story.

202

, ,   

There are no signs that Chrysippus or any other early Stoic or any of their Hellenistic opponents considered the problem of free will or of freedom of decision, i.e. the question of whether in a particular situation where a decision is made, given the same or relevantly similar circumstances, and given the same character or dispositions of the soul, the agent could have decided and acted otherwise.²⁹ It is regarded as a necessary condition for the attributability of praise and blame (and hence for the possibility of ethics) that we are the originators of our actions. But there are no signs that freedom of decision (as introduced in section 1) was considered a necessary condition for justified praise or blame.

2.2 Early Stoics on eleutheria The concept of eleutheria was part of Stoic philosophy from its very beginning, i.e. from Zeno onwards, but there are few testimonies about eleutheria that can be assigned with certainty to the early Stoa. In Diogenes Laertius we find a passage from Zeno’s Republic where he declares that ‘only the wise or virtuous are true citizens or friends or kindred or free men (eleutherous), and that all those who are not wise are enemies, slaves, and aliens to one another’ (DL 7.32–3, with parallels in Cic. Mur. 61). Further, we have from Zeno a remark in Philo of Alexandria that no one could ‘force any wise man to do unwillingly anything that he does not want to do’ (Prob. 97), and a modification of a Sophoclean verse on freeman and slave (Plut. Quomodo adul. 33d). Cleanthes wrote a book on eleutheria (DL 7.175), but apart from its title nothing has survived.³⁰ We have nothing at all on eleutheria that is attributed directly to Chrysippus. Of the passages listed in von Arnim that

²⁹ It is true, Epicurus may have postulated uncaused movements. However, as far as I can see, there is no decisive evidence that he connected this idea with the questions of whether our decisions are undetermined by preceding causes, whether under the same circumstances we could act differently or choose otherwise, or even generally whether alternative courses of actions are open to us. Rather, just as in the case of the early Stoics, the problem is how we can be causally responsible for our actions, beliefs, etc., rather than being necessitated by something else (this is the reason for the occurrence of adespoton in DL 10.133); and the requirements for moral responsibility are freedom from compulsion and hindrances and self-origination of our actions. There is one difference though. Due to Epicurus’ atomism he faces the additional threat of internal necessitation by the atoms of which a living being, or its soul, is constituted (cf. Nat. 25, Arr. 34.21ff.; Lucr. 2.264–93). Arguably, this brings Epicurus closer to the modern discussion of neurophysiological determinism and free will than the Stoics seem to have been. I discuss these questions in more detail in Chapters 5 and 6 of the present volume. As to Aristotle and the Peripatetics, it has been argued repeatedly that we have no reasons to assume that Aristotle resorted to uncaused events to preserve the concepts of voluntary action and moral responsibility (e.g. Sorabji 1980b, 228–33; Irwin 1980, 117–55; Englert 1987, ch. 5). The passage which most resembles modern discussions of the free-will problem, including the question of our responsibility for our character, is EN 3.5. Meyer 1993 has recently suggested a plausible interpretation of this passage, according to which Aristotle was not concerned with the problem of free will and freedom of decision in the modern sense. ³⁰ Perhaps Philo’s book On the fact that only the wise are free and the fifth of Cicero’s Stoic paradoxes on the same topic may give us some general idea of the issues that Cleanthes dealt with.

   :    

203

deal with eleutheria, four more might be early Stoic.³¹ They are Stobaeus, Eclogues 2.101.15–20; Cicero, De Finibus 3.75 and Academica 2.136; and Diogenes Laertius 7.121. This last passage gives a definition of eleutheria as exousia autopragias, which could be early Stoic.³² Now, we know that Chrysippus used ‘autopragia’ in the sense of ‘doing what is one’s own’ (ta autou prattein, Plut. Stoic. Rep. 1043b). Accordingly, eleutheria would then be the power of managing one’s own things. That is, you are eleutheros if you have the power of doing what you want or what you should do. All testimonies on Stoic eleutheria, without exception, belong to ethics or politics.³³ Eleutheria and eleutheros are typically contrasted with douleia and doulos, and the philosophical use of the concepts in ethics seems to have taken its origin from the analogy with politics and public life, as in Zeno’s Republic (see above). Most sources make the point that only the sage is (truly) eleutheros whereas common mortals are all (truly) slaves.³⁴ Being eleutheros is further typically cited as one of a number of positive attributes all of which belong only to wise persons;³⁵ in early sources often with political attributes, later with any attributes of virtue. ‘True’ eleutheria depends on the disposition (diathesis) of the wise person’s soul, which is stable and in a state of ideal tension. As regards eleutheria, this state of soul has a twofold effect on the person’s behaviour: internally, the eleutheros is master of his emotions (nec obediens cupiditati, Cic. Fin. 3.75); externally, the eleutheros cannot be bribed or blackmailed into actions which he does not want to perform (nec dominationi cuiusquam parens, ibid.). In none of the testimonies on early Stoics is there any connection drawn between eleutheria and that which depends on us.³⁶ Eleutheria, as I have said, belongs exclusively to ethics (including politics); that which depends on us belongs to Stoic physics, and in particular to Stoic psychology. The function and purpose of the two concepts in early Stoic philosophy are quite different. Only the wise are eleutheroi, but wise and non-wise people alike are responsible for their actions; their actions depend on them in exactly the same way. Whether one is a slave of one’s passions or bribed into actions by other people, one is accountable

³¹ As to the remaining passages quoted by von Arnim (SVF 3.359–64 from Philo’s Quod omnis probus liber, 3.603 from Philo’s De sobrietate, and 3.356 and 365 from Dio Chrysostom), they do not mention Stoics at all, and although clearly Stoicizing, must contain material later than Chrysippus. ³² Most of Diogenes’ report is early Stoic. Similarly, Origen reports as an opinion of ‘the Greeks’, καὶ μόνον καὶ πάντα τὸν σοφὸν εἶναι ἐλεύϑερον, ἐξουσίαν αὐτοπραγίας ἀπὸ τοῦ ϑείου νόμου εἰληφότα (Evang. Ioannis 21.10). ³³ This holds also of the later passages in Philo’s Quod omnis probus liber and of Dio Chrysostom, and of Cicero’s fifth Stoic paradox. ³⁴ E.g. DL 7.32–3; Stobaeus, Ecl. 2.101.15–20; Cic. Acad. 2.136. ³⁵ E.g. DL 7.32–3; Cic. Acad. 2.136. ³⁶ Note also that Cicero deals with both topics, with eleutheria in the Stoic Paradoxes, with the Stoic position on that which depends on us in On Fate, and that he draws no connection at all between the two.

204

, ,   

for one’s deeds and omissions in exactly the same way as a wise person is.³⁷ (There is also no mention of the wise having more choices or possibilities in their actions; they differ from common people rather in that they follow up the right choices and are not tempted by the wrong alternatives. This is compatible with the fact that by performing certain actions one widens one’s future choices, by performing others one narrows them down, and that it cannot be ruled out that on balance a wise person may end up with more choices.)

3. Epictetus 3.1 Epictetus and That Which Depends on Us Both the concept of eph’hēmin and the concept of eleutheria are central to the part of Epictetus’ philosophy which is extant, recorded by his pupil Arrian.³⁸ I shall deal with these two Epictetan concepts, and their relation to each other. This is of course not the place to give a detailed analysis of Epictetus’ philosophy of freedom through self-restriction. I shall confine myself to pointing out some changes in emphasis, interest, topics, and use of terminology in Epictetus, in order to be able to outline the main differences between his and Chrysippus’ conceptions of freedom. Regarding the concept of that which depends on us, Chrysippus’ concern was, first, with securing certain necessary conditions for that which depends on us, second, with the proof that the things commonly seen as depending on us depend in fact on us, and that Stoic causal fate-determinism does not conflict with this. Chrysippus did not ask what the things were that depend on us. It appears that he took it for granted that there was general agreement about that question. For Epictetus, on the other hand, the main concern is which particular kinds of things depend on us. An agreement about that is no longer taken for granted. The two different topics were also terminologically differentiated: Chrysippus was concerned with what was later called to eph’hēmin; Epictetus is primarily interested in ta eph’hēmin, i.e. the kinds of things that depend on us.³⁹ With this difference in interest there goes hand in hand a difference in how the expression ‘depending on

³⁷ The difference between the early Stoic concept of that which depends on us and the early Stoic concept of eleutheria has been aptly pointed out by Stough 1978, 224. ³⁸ Epictetus (c.55–c.135 ) was a pupil of the Stoic Musonius (c.30–c.100 ), and for most of the following points we find some related ideas in the few extant fragments of Musonius (in Stobaeus). This makes it likely that Epictetus took over many of his ideas from his teacher. ³⁹ For Chrysippus neither of the substantivized forms are known. To eph’hēmin occurs in Plutarch, e.g. Stoic. Rep. 1056d. I have not found ta eph’hēmin as a (semi) technical term in any Stoic text before Musonius. (The explanatory clause about things that depend on us in Epiphanius, Adv. Haer. 3.36 (Diels, Doxogr. graec. 692.25–6), which is ascribed to Zeno, I take to be later and not originally by Zeno.)

   :    

205

us’ is used. According to Chrysippus, an individual state or change depends on me if I, qua rational being, am causally responsible for it; that is, if it is caused primarily by an act of assent of mine. Hence (i) the element of causation and, in this sense, of self-origination is predominant, and (ii) in order to determine whether something depends on me, the individual case has to be examined. In Epictetus (i) the element of causation is not discussed, and the factor of self-origination is seldom emphasized, although the occasional use of phrases like kurios and autexousios suggests that it was presupposed. Moreover (ii), in order to establish whether an individual change or state depends on us, it seems that for Epictetus what matters is whether it belongs to a class of things that cannot be externally hindered or forced. This means that whereas according to Chrysippus, if I take a walk and nothing hinders me from walking, my walking presumably depends on me (I caused it by assenting to an impulse-prompting impression of the kind ‘you should take a walk’), for Epictetus it does not depend on me, since in principle something could prevent me from walking, even if in this case nothing does (cf. Diss. 4.1.68–73). Thus for Epictetus the modality of the expression ‘depending on us’ has changed and he singles out as things that depend on us only those which are in our power in all possible circumstances and with absolute certainty. He is concerned with the things (activities, behaviour) that in the course of one’s life can in no circumstances be prevented or spoiled by external circumstances, including by other people. It is the question of someone who plans their future life and actions and wants to know which factors of this future life can be relied on with certainty. The motivation for such questions—as it is usually presented by Epictetus—is how one can avoid failures and disappointments and how one can keep or attain an undisturbed and well-poised emotional state. That is, the motivation is chiefly prudential and pragmatic: if one is aware of the limits of one’s power, one’s future plans will be more realistic (they will probably include more ifs and whens)⁴⁰ and consequently one’s future disappointments are minimized. The field thus is practical philosophy and ethics inasmuch as ethics helps one to attain happiness (eudaimonia). For, and not surprisingly, he holds that the only way to minimize frustration and to subsequently acquire true happiness is by consistently pursuing the morally right end (telos), which is, in one common Stoic description, to live in accordance with nature. The perspective has thus become a perspective towards the future, concerned with guidance of actions and behaviour; before, in Chrysippus and the old Stoa, it was primarily a backwards perspective or was time independent, concerned with the attribution of responsibility and with the moral assessment of actions.⁴¹ ⁴⁰ Cf. Inwood 1985, 119–26 and 165–75 for ‘reservation’ (hupexairesis) in Seneca and Epictetus. ⁴¹ It ties in with this change of perspective that whereas Chrysippus’ concept of that which depends on us may have been one-sided (i.e. the opposite of that which depends on us does not depend on us,

206

, ,   

Accordingly, for Epictetus, the range of things that satisfy the condition of depending on us is much more limited than it was ‘traditionally’. He puts in the class of the things that depend on us (ta eph’hēmin) primarily—or perhaps even exclusively—the ‘use of impressions’ (hē chrēsis tōn phantasiōn), which for him means: the mental events of assenting and intending, and certain refrainings from action.⁴² In the class of things that do not depend on us (ta ouk eph’hēmin) he puts everything else that has some influence on our life, i.e. our (positive) actions, the results of our activities, the things called ‘indifferents’ in Stoic ethics—like health, wealth, life, death—and, as Chrysippus and the tradition had done, the impressions themselves.⁴³ One can easily see how this restriction was brought about: if we use the Epictetan criterion for future actions, the result is meagre. For nearly every activity that involves intentional bodily movements, however small, we can imagine some external obstacles that will prevent it from being carried out. With (positive) actions such as walking, eating, or escaping one’s enemy, we can never be sure that they will be in our power. Things look a bit better in cases in which assent is given in favour of not being active, i.e. in favour of refraining from action. In those cases one can argue (as Epictetus in effect does repeatedly) that external influences cannot prevent one from ‘inactivity’.⁴⁴ But the safest bets are assent and intention. External hindrances are not conceivable there. Neither natural nor human force can prevent them from occurring, as Epictetus never tires of repeating. For Epictetus, assent, intention, and refraining from action depend on us, because we have the general ability to perform them and no one and nothing external to us has the power of interfering and keeping us from performing them. Philosophically there is perhaps less of a difference between Epictetus and the early Stoa than this account makes one think. The underlying general idea of which things depend on us seems to be the same as for Chrysippus: in order to depend on us something (i) has to be originated by us, hence we must not be forced, and (ii) it must not be prevented by external circumstances. But Epictetus is not concerned with establishing that there are things that depend on us. Rather he employs the concept of that which depends on us within ethics, and although such an application is not recorded for early Stoic ethics, we can easily see how

see section 2.1), some passages in Epictetus suggest that his concept is two-sided, for instance, if it depends on me to assent to an impression, it also depends on me not to assent to it (see e.g. Diss. 1.1.17; 1.1.23; 2.13.10–11; 2.19.32–4; 3.24.23). In this respect, the things that depend on us would double, compared to Chrysippus’ approach. This does not entail that for Epictetus that which depends on us is uncaused or undetermined. ⁴² These also depend on us according to Chrysippus, cf. Cic. Fat. 40–3; Gellius, NA 7.2.11. ⁴³ For Epictetus, typical passages are: Ench. 1, 2, and 5; Diss. 1.1.7; 1.1.12; 1.1.22–3; 1.6.40; 1.12.34; 1.22.10; 2.5.8; 2.19.32; 3.24.3; 3.24.69; 3.24.108; 4.1.68–75. For Musonius, see Stobaeus, Ecl. 2.159.25–160.11 (fr. 38 Hense). ⁴⁴ E.g. Diss. 1.1.22–4.

   :    

207

Epictetus’ concept of that which depends on us fits in with it. Antipater had described the end (telos) of human life as ‘to do everything in one’s power (kath’hauton) continuously and undeviatingly with a view to obtaining the predominating things which accord with nature’ (Stobaeus, Ecl. 2.76.13–15, tr. Long and Sedley). What Antipater had in mind becomes clearer from the analogy with the archer, in which the Stoics distinguished between aiming straight at the target and actually striking it (Cic. Fin. 3.22). The end is doing everything to attain one’s object; from this has to be distinguished the actual achieving of the object. Of these only the former is in our power. Here Epictetus comes in with the expression ‘that which depends on us’: he uses it to denote the realm within which we can do everything to attain our object, and he relegates the actual achieving of the object (or not, depending on the circumstances) to the sphere of that which does not depend on us. This, it seems to me, is all there is to Epictetus’ concept of that which depends on us. There is no evidence that this concept had anything to do with the choice of alternatives, let alone free (undetermined) choice or agent causality. (Behind it there lies the ‘I-cannot-be-bribed-or-blackmailed-into-doing-certain-things’ idea of freedom.) As in the case of Chrysippus, there is no hint that Epictetus considered whether someone could have done otherwise in the sense that in identical circumstances the very same person could end up once doing something, once not doing it. The interpretation that according to Epictetus we are free to choose whether we give assent and whether we want something (in the sense that in the narrow boundaries of that which depends on us we are not causally predetermined in our choice) is simply wrong.⁴⁵ Epictetus does not reflect upon the ability to choose freely between performing or not performing an activity. Rather, in some sense, this ability seems always implicitly presupposed.⁴⁶ Suppose, for example, that I am externally forced to go to prison. If I then happen to choose not to go, my choice will be frustrated, inasmuch as the object of my willing is externally prevented from realization. If, complying with god’s plan, I happen to choose to go to prison, my choice will not be frustrated. In the same way my choice will never be frustrated in cases in which I ‘choose’ to give assent or to have an intention to do something. But, as far as the act of choice itself and its determinedness are concerned, there is no difference in all these cases—whether

⁴⁵ A more recent advocate of this view is Dobbin 1991, e.g. 121, 133. Generally, this interpretation assumes that for Epictetus we have no choice in the case of things that do not depend on us, e.g. whether we go to prison. Independently of what we decide or want, if we have to go, we will have to go. But we have undetermined free choice in the case of things that depend on us: we cannot choose whether or not we will go to prison, but it is a matter of undetermined choice whether or not we accept (give assent in favour of) going to prison, and whether we go to prison willingly or unwillingly. ⁴⁶ Within the realm of that which depends on us we are repeatedly advised to go for those things that are in accordance with nature and to avoid those that are contrary to nature; cf. e.g. Ench. 2.48.3; Diss. 3.10.10–11, 3.11.15, 3.24.101–2. Equally, prohairesis, often rendered as ‘moral choice’, is never characterized as undetermined choice by Epictetus, nor does it imply indeterminedness.

208

, ,   

or not my choice is frustrated.⁴⁷ Epictetus’ occasional use of the term autexousios in the context of the discussion of things that depend on us does not indicate any concern with the topic of undetermined choice either. In later texts to autexousion becomes synonymous with to eph’hēmin (see section 4), and sometimes involves the ability to choose freely or to start a motion without preceding cause, but in Epictetus it does not have these connotations. Epictetus clearly contrasts autexousios with someone else’s power or authority (exousia) over oneself. Autexousios indicates that something is outside the sphere of influence of others, and because of that in the sphere of my power. That is, we have the same contrast between the possibility of external hindrances and the absence of that possibility, as in the case of the distinction between ‘things that depend on us’ and ‘things that do not depend on us’.⁴⁸ The absence of concern with the topic of choice is familiar from Chrysippus and the early Stoics. Similarly, despite a difference in the use of expressions, there seems to be little difference between Epictetus’ position on moral responsibility and that of the early Stoics. In the extant works, Epictetus never considers the problem of whether it is consistent with Stoic philosophy in general to maintain that people are morally responsible for some things. But he deals with the questions of praise and blame and moral accountability within ethics. Praise and blame occur in two different contexts. On the one hand, praise and blame, reward and punishment are rather some amongst those external things from the influence and importance of which one should try to free oneself, exactly because they are not in our power (e.g. Diss. 2.13.2; 2.16.6–11). In line with this we learn that we should not praise or blame others (Ench. 48.2). On the other hand, there is the occasional mention of the idea that the moral goodness and badness and the praiseworthiness and blameworthiness of an action does not depend on external factors, but is always derivative of the morally right or wrong beliefs (dogmata) from which the agent acts (Diss. 4.4.44; 4.8.1–4). We may praise or blame an agent for an action only insofar as it was brought about as the result of the right or wrong use made of the impressions. In line with this we learn that we are morally accountable (hupeuthunos) only for those things that depend on us (Diss. 1.12.32–5). Thus Chrysippus and Epictetus agree fully on the point that actions are praiseworthy and blameworthy only insofar as they are the result of an act of assent to a practical impression—only that Epictetus restricts the terms ⁴⁷ The passages in which Epictetus says that even god cannot prevent or hinder us in the case of things that depend on us (e.g. Diss. 1.6.40) cannot be invoked to back up the claim that Epictetus deals with free choice. God’s ‘inability’ to interfere does not mean that, say, I can give assent to some impression although god does not want me to but is unfortunately unable to prevent me. The point is rather (as usual) that assent and intention, on the ground of their very nature, cannot be subjected to coercion or force, which includes possible coercion or force exercised by god. ⁴⁸ Cf. the use of ἐξουσίαν ἔχειν in Diss. 4.7.16, εἰς ἐμὲ οὐδεὶς ἐξουσίαν ἔχει. ἠλευϑέρωμαι ὑπὸ τοῦ ϑεοῦ, (see also 1.25.2, 4.12.8) with the use of αὐτεχούσιον in 4.1.62, τί οὖν ἐστὶ τὸ ποιοῦν ἀκώλυτον τὸν ἄνϑρωπον καὶ αὐτεξούσιον . . . , and 4.1.68, πότερον οὖν οὖδὲν ἔχεις αὐτεξούσιον, ὃ ἐπὶ μόνῳ ἐστὶ σοί . . .

   :    

209

‘depending on us’ and ‘accountable for’ to the use we make of our impressions, whereas Chrysippus seems to have allowed at least the first one to be applied to actions as well, if only derivatively. (Since he appears to hold that actions take place in the mind, the discrepancy is even smaller than it may seem.) As for the individual cases in which someone is held responsible, i.e. for individual assent– action combinations as it were, there should be no principal disagreement between the two philosophers. Again, the difference is rather one of perspective. The change of viewpoint in Epictetus’ treatment of that which depends on us has its reason in a shift of topic, interest, and emphasis from the question of the possibility of moral accountability to the guidance of action. There is no indication that it is the result of any discontent on Epictetus’ part with early Stoic or Chrysippean compatibilism—as appears to be the case in the Middle Platonist theory of fate, in Posidonius, and others. Rather, it seems that the theoretical, abstract question of the compatibility of fate and that which depends on us has completely dropped out of sight in Epictetus’ extant works. His concept of that which depends on us is basically independent of whether or not everything is fated. It is likely that both Epictetus’ teacher Musonius and Epictetus himself believed that, in the end, everything is predetermined. Musonius, it seems, may have stuck at least to the spirit of universal fatedeterminism. For him the claim is reported that whatever happens cannot happen otherwise.⁴⁹ He presumably did not exempt the things that depend on us from this claim. Epictetus appears to have avoided talking about fate (heimarmenē) qua network of causes and the Chrysippean theory of causal fate-determinism. He repeatedly employed Cleanthes’ verses on destiny (peprōmenē) for practical purposes (Ench. 53.1, Diss. 2.23.42, 3.22.95, 4.1.131, 4.4.34), and the concept of destiny used is that of personal fate, i.e. the common belief that the gods allot to every person their destiny. The one time when Epictetus speaks in propria persona about ‘destiny’ (Diss. 1.12.25), we are presented with the idea of personal fate, too, and the terminology seems to be borrowed from Cleanthes’ verses. There can be little doubt that Epictetus was convinced that god was responsible for what happens in the world, and that he preordained everything (Diss. 1.12.15–17). But his remarks are throughout so vague that from them one cannot make out whether god’s predetermination concerns generics only, or particulars as well, and no connection is drawn to the causal interrelation of things. To sum up, the role of that which depends on us in Epictetus’ philosophy (as far as it has come down to us) differs noticeably from the function it has for the early Stoa. In early Stoic philosophy the notion of that which depends on us has its place in physics—more precisely, in Stoic psychology. It has the function of preserving the possibility of moral responsibility. It is thus a basic precondition for ethics, but

⁴⁹ Stobaeus, Ecl. 4.44, 60 (fr. 42 Hense).

210

, ,   

not itself part of ethics. For Epictetus, the notion of that which depends on us serves—on the basis of an already established theory of morals—a predominantly practical purpose. It has its role within (first-order) ethics. It is intended to provide a means that helps people to plan and lead a good and undisturbed life. Its primary function is guidance of life and actions. Despite the different primary function assigned to the concept of that which depends on us, there is little difference between Chrysippus and Epictetus in the relevant parts of Stoic philosophy— except that in his extant works Epictetus shows no signs of any critical awareness of the problem of compatibilism. For Epictetus, as for Chrysippus, the concept of that which depends on us is not a concept concerned with the freedom of decision or of undetermined choice between alternative courses of actions.

3.2 Epictetus on eleutheria The concept of eleutheria is central to Epictetus’ philosophy as it has survived. We have an extensive essay especially on that topic (Diss. 4.1) and eleutheria plays a prevalent role in numerous other essays. There seems to be little difference between Epictetus’ concept of eleutheria and the early Stoic one. But again, there is a shift in emphasis. According to Epictetus, the eleutheros is ‘someone who lives as he wills, who is neither necessitated nor hindered nor forced, whose impulses are unhampered, whose desires reach their end’ (Diss. 4.1.1); and again, ‘someone for whom all things happen in accordance with choice and whom none can constrain’ (Diss. 1.12.9) and someone ‘who is rid of pain, fear, and trouble’ (Diss. 2.1.24). As in the early Stoa, so in Epictetus eleutheria is a moral quality which only wise people possess, and it is frequently opposed to slavery (e.g. Diss. 2.2.13). It is not connected primarily with actions but is linked to persons and their characters or states of mind. But in Epictetus eleutheria occurs typically within a group of positive terms that are connected with tranquillity of mind, like ataraxia, apatheia, akōlutos.⁵⁰ What mainly concerns us here is how Epictetus sees the relation between eleutheria and that which depends on us. Unlike the early Stoics, Epictetus expressly made the connection between the two concepts, and it is essential for an understanding of his philosophy to comprehend their relation. The relation between his notions of that which depends on us and eleutheria is then as follows. Eleutheria is a virtuous state of mind, desirable and to be aimed at. In order to achieve this virtue you must (i) know exactly what things depend on you and (ii) align your desires, life plan, etc. in such a way that you only ever want what depends on you and expect only what is within the boundaries of what depends on ⁵⁰ E.g. Diss. 2.1.21, 3.5.7, 3.15.12, 4.1.27–8; Ench. 29. Akōlutos also often occurs together with eph’hēmin.

   :    

211

you. In short, you have eleutheria if, knowing what depends on you, you do not ever desire or deplore anything that does not depend on you.⁵¹ Such a relation between eleutheria and that which depends on us is conceivable only with Epictetus’ modified understanding of what depends on us. Chrysippus’ conception of that which depends on us does not admit of such a relation. For whether a particular thing, an action, say, depends on us in the Chrysippean sense is contingent upon whether there are hindrances at the time of the attempted performance, and frequently agents will learn about them only when they actually try to act. Nevertheless, for Epictetus, as for the early Stoics, in general there are no more things that depend on the free (eleutheros) sage than there are that depend on common people. The only difference is that the sage knows about the restrictions of that which depends on us. As to moral responsibility, common persons are no less accountable for their actions than the free sage, and Epictetus’ restriction of that which depends on us does not diminish our moral responsibility. Thus, as in the case of the early Stoics, there is no direct connection between the concept of eleutheria and that of moral responsibility.⁵²

3.3 Some Remarks on a Contemporary Interpretation of Stoic Freedom It can now be seen that a prevalent interpretation of Stoic theory of freedom is quite unfounded. This is the interpretation that for the Stoics eleutheria was to fulfil the role of a substitute for ‘the real freedom of the will’ (that which depends on us, interpreted as free will or freedom of decision), which, alas, was destroyed by Stoic determinism.⁵³ This assumption is often made in the context of the idea that free will was a mere illusion (the concept of free will a mere epistemic concept) in Stoicism and was regarded as such by the Stoics. It is then assumed that the Stoics intended to make up for this lack of freedom with their concept of eleutheria, the freedom reserved for the wise only, which is then referred to as ‘the only freedom left for a man’, and as ‘the only true freedom’.⁵⁴ Thus the two concepts of that which depends on us and eleutheria are taken to have played roles ⁵¹ Cf. the essay on eleutheria; consider also ‘And because of this, if he holds that that which is good for him and beneficial is only in those things that are free from hindrance and depend on him, he will be eleutheros, serene, happy . . . ’ (Diss. 4.7.9). Cf. also Sen. Vit. Beat. 15.7. ⁵² This does of course not preclude that our beliefs about what we are morally responsible for can be pertinent to our progress towards eleutheria in various ways. For instance, as long as I believe that I bear moral responsibility for the disastrous consequences of an earthquake, since I did not succeed in preventing it from happening, the accompanying emotional distress I feel about my ‘moral failure’ will, presumably, be detrimental to my progress towards eleutheria. ⁵³ So Long 1971, 175, 189ff.; Long and Sedley 1987, vol. 1, 394; Inwood 1985, 109–11. Similarly Pohlenz 1959, vol. 1, 106; Forschner 1981, 110–11. ⁵⁴ Inwood 1985, 110; Long and Sedley 1987, vol. 1, 394. But note that when Stoic sources talk about ‘true freedom’ in the context of eleutheria of the wise (e.g. Stobaeus, Ecl. 2.101.15–20), the contrast is

212

, ,   

in the discussion of a single philosophical problem, and moreover, they are often not properly distinguished.⁵⁵ This treatment of Stoic eleutheria and what depends on us ignores some essential points. First, there is no evidence that the topics of eleutheria and of that which depends on us were connected in any way before Epictetus. Second, there is no reason to believe that the early Stoics thought their concept of that which depends on us to be in any way deficient, or to be founded on an illusion. Rather, it seems that the early Stoics (and the other Hellenistic philosophers) did not have a concept of freedom of the will or of freedom of decision. Third, there is no parallel in Hellenistic times to the English ambiguity of ‘freedom’, as freedom of decision and freedom from hindrances and political freedom; there was no one word in early Stoicism that covered these various meanings. Fourth, and, most importantly, the debate over fatalism, determinism, and that which depends on us assigned the concept of that which depends on us a special function: namely to guarantee the possibility of moral responsibility, of purposeful action, of a legal system, etc. The concept of that which depends on us was considered the most important necessary condition for these. The early Stoic concept of eleutheria, with its restriction to wise people, and being in the first instance ‘freedom from’ and not connected with the origination of intention and action, obviously could not and never was meant to fulfil this function.⁵⁶ Hence, not only is there no evidence that any Stoic thought of eleutheria as an ersatz for the ‘lost’ freedom of that which depends on us in the debate about determinism, but we should find it surprising if any Stoic had suggested this. For they would have suggested using a completely different concept, fulfilling a different function, in a different part of Stoic philosophy, and referred to by a different term, in order to make up for the lack of things depending on us. Obviously, such a suggestion would have been quite without point or philosophical motivation.⁵⁷

4. Later Antiquity The story of Stoic freedom does not end with Epictetus. In Alexander of Aphrodisias’ treatise On Fate, another Stoic compatibilist theory is discussed, a

with the ‘illusionary freedom’ of the non-wise freeman who thinks of himself as free but is in fact a slave. The contrast here is not with ‘illusionary freedom of the will’, i.e. ‘that which depends on us’. ⁵⁵ Long 1971, 189–90; Inwood 1985, 109–10. This mix up leads Inwood to deliberate whether the non-wise are not in fact more free (eleutheros or eph’hēmin) than the wise (ibid. 109), and prompts Long to wonder why ‘this [Stoic] concept of freedom [i.e. eleutheria] was not attacked by critics of Stoic determinism’ (ibid. 175). ⁵⁶ One may remember that it was in particular bad actions and errors for which responsibility had to be secured; see Gellius, NA 7.2.4–14. ⁵⁷ This is not to deny that modern philosophers have tried, in various ways, to substitute concepts of freedom similar to Stoic eleutheria for concepts of ‘undetermined freedom of decision’.

   :    

213

theory that most probably originates from the first half of the second century , and which seems to have been defended by a contemporary of Alexander.⁵⁸ Let me begin with a remark about a change in terminology in Greek texts of which we have the first traces shortly after Epictetus. In Alexander, and generally from the end of the second century onwards, we repeatedly find the locutions to eph’hēmin and eleutheron closely connected: in particular the adjectival expression eleutheros is predicated of to eph’hēmin.⁵⁹ But in these passages eleutheros no longer has anything to do with the early Stoic or the Epictetan special, moral sense of the word (see sections 2.2 and 3.2). This means that despite this later near synonymy of the expressions eleutheros and eph’hēmin, there is still no evidence that the Stoic concept of eleutheria played a role in the debate about the compatibility of fate and the kind of freedom that is required for moral responsibility. The Stoic compatibilist position that we find in Alexander’s treatise On Fate undoubtedly stands in the Chrysippean tradition. Like Chrysippus these Stoics are concerned with the compatibility of causal fate-determinism with moral responsibility. However, although the Stoic position seems to have remained by and large the same, the focus of discussion has once more shifted. It appears that now, for the first time, the discussion involves a concept of freedom of decision and an awareness of the problem of the compatibility of such freedom and causal determinism. Chrysippus had defended the compatibility of causal determinism and that which depends on us by maintaining that different people react to comparable impressions in different ways, and that because of this acts of assent and actions depend on the people themselves. At the same time they are causally determined by fate, which simply acts through these people. In Alexander, the Stoic position on causal determinism has been refined: it is now set out explicitly in terms of the idea of counterfactual situations. We encounter a new physical principle that has been introduced in order to back up the principle that every event has preceding causes (Fat. 192.22–5). The new principle states that under the same circumstances the same cause will necessarily bring about exactly the same effect.⁶⁰ For the special case of human assent and action this means: in the ⁵⁸ Cf. e.g. Frede 1982, 276–7. ⁵⁹ In Alex. Fat. 188.21 we read in the context of that which depends on us τὸ ἐλεύθερόν τε καὶ αὐτεξούσιον, and in 189.10 εἶναι τὸ ἐφ’ ἡμῖν ἐλεύθερόν τε καὶ αὐτεξούσιον; in Nemesius, Nat. hom. 36.26–37.1 ἐλεύθερον γάρ τι καὶ αὐτεξούσιον τὸ λογικόν, in 105.24 ἐλεύθερον γὰρ εἶναι δεῖ τὸ ἐφ’ ἡμῖν. ἐλεύθερος in this meaning occurs only in Justin, e.g. Apol. 1.43; in Tatian, Ad. Gr. 7; in Plot. Enn. 6.8.4–6; in Bardesanes in Eusebius, Praep. Ev. 336, etc. The term αὐτεξούσιον, which we encounter already a few times in Epictetus (see section 3.1), comes to be another near-synonym of τὸ ἐφ’ ἡμῖν. It becomes quite common in the third century , and in Nemesius’ time it seems to have taken over from τὸ ἐφ’ ἡμῖν, cf. e.g. Nemesius, Nat. hom. 112.7. Περὶ τοῦ ἐφ’ ἡμῖν, ὅ ἐστι περὶ τοῦ αὐτεξουσίου. αὐτεξούσιον occurs several times in Alexander (Fat. 188.21, 189.10, 196.10), in Bardesanes (in Eusebius, Praep. Ev. 337, 340, 342), in Origen (Princ. 3.1.5), and in Justin (Apol. 1.43) and Tatian (Ad. Gr. 7); Eusebius (Praep. Ev. 334.13–14) and Tertullian (An. 21.6) treat it as a technical term. ⁶⁰ ἀδύνατου εἶναι, τῶν αὐτῶν ἁπάντων περιεστηκότων περί τε τὸ αἴτιον καὶ ᾦ ἐστιν αἴτιον, ὁτὲ μὲν δὴ μὴ οὑτωσί πως συμβαίνειν, ὁτὲ δὲ οὕτως. ἔσεσϑαι γὰρ, εἰ οὕτως γίνοιτο, ἀναίτιόν τινα κίνησιν (Fat. 192.22–5). Cf. Alex. Fat. 185.7–9 and Nemesius, Nat. hom. 105.18–21. For some details, see Chapter 8 of this volume, originally printed as Bobzien 2005.

214

, ,   

same situation the same incoming impression in the same person with the same state of soul will elicit the same reaction. If—in a certain situation—a person gives assent to an impression, then that person would also give assent to that impression in any otherwise identical situation. We now also find a Stoic account or definition of that which depends on us: it is that which happens through us, i.e. via our assent to impressions and the subsequent intention to act (Fat. 182.12–13;16). This definition is in keeping with Stoic causal fate-determinism. The position of the—Peripatetic—critics of the Stoics in Alexander differs from that of Chrysippus’ opponents. These Peripatetics are proponents of indeterminist freedom. (One could think of them as the ancient variety of libertarians.) They provide accounts of that which depends on us which differ from the Stoic one and which are similar to that of the modern concepts of metaphysical freedom and freedom of decision. They state that ‘depending on us’ is predicated of the things of which we have in us the ability to choose also the opposite* (Fat. 181.5–6; cf. 181.12–14)

and that those things depend on us of which we seem to be masters of their being done and of their not being done.** (Fat. 169.13–15; cf. 181.12–14, 199.8–9, 211.31–3)

Thus, first, in some formulations of the account the stress is now on choosing to do something, not simply on doing it. Second, the power of doing or choosing alternative courses of actions seems essential to this account.⁶¹ And third, these Peripatetic ‘libertarians’ require expressly that the choice or action has to be independent of preceding causes: we take it that we have this power in our actions of choosing the opposite and not that everything we choose has predetermining causes because of which it is not possible for us not to choose this*** (Fat. 180.26–8, cf. Mant. 174.3–12)

* τὸ ἐφ’ ἡμῖν ἐπὶ τούτων κατηγορεῖται, ὧν ἐν ἡμῖν ἡ ἐξουσία τοῦ ἑλέσθαι καὶ τὰ ἀντικείμενα. ** ἐφ’ ἡμῖν δὲ ταῦτα, ὧν καὶ τοῦ πραχθῆναι καὶ τοῦ μὴ πραχθῆναι ἡμεῖς εἶναι δοκοῦμεν κύριοι. For details, see Chapter 1. ⁶¹ Note again the following difference between the Stoic and Peripatetic accounts. For the Peripatetics, as for Aristotle, if an action is eph’hēmin, so is its opposite. For the Stoics this does not hold: only the actions which we actually perform are eph’hēmin; only they happened through us. However, for Stoics and Peripatetics alike, if something is eph’hēmin, then it is possible and its opposite is possible as well (see section 2.1). However, we have no reason to believe that Aristotle maintained, as these Peripatetics did, that there were actions that are independent of preceding causes. *** ταύτην ἔχειν τὴν ἐξουσίαν ἐν τοῖς πρακτοῖς προειλήφαμεν, ὡς δύνασθαι διαιρεῖσθαι τὸ ἀντικείμενον, καὶ μὴ πᾶν ὃ αἱρούμεθα ἔχειν προκαταβεβλημένας αἰτίας, δι’ ἃς οὐχ οἷόν τε ἡμᾶς μὴ τοῦτο αἱρεῖσθαι.

   :    

215

These preceding causes include not only the external circumstances but also the disposition of the agent’s soul or the agent’s character (Fat. 199.27–200.7): Alexander stresses that there are situations in which people can act or choose against their dispositions or character;⁶² this implies that he distinguished between a person who chooses and that person’s character or set of dispositions. Thus the debate has moved to a related but different problem of compatibility. The new conflict is this: the determinists claim that every change in the world is causally predetermined. The same cause under the same circumstances will necessarily bring about the same effect. On the other hand, for the ‘libertarians’, the claim that there are things that depend on us now has come to mean: some changes in the world are not causally predetermined; these include human choices or decisions. In the very same situation, we, the very same causes, could decide one time one way, another time another way, undetermined by preceding external and internal factors. The question then is: how can I myself be said to make a decision if there are causes prior to my decision by which it is predetermined which way I decide (Fat. 180.26)? This is the problem of the compatibility of— indeterminist—freedom of decision and causal determinism. There is no solution to this conflict: causal determinism and partial causal indeterminism are mutually exclusive. For the libertarian, decisions are either uncaused or—more commonly—caused by the person, soul, or will who decides, and who in turn is causally independent from previous states of the world. (This is sometimes called agent causality.) This concept of freedom and the related concept of that which depends on us are very different from those the early Stoics and their opponents dealt with.⁶³ Note that with the development of this new problem of compatibilism, the relation between freedom and morality has also changed. The opponents of the Stoics now consider their freedom of decision as a necessary condition for moral responsibility, and correspondingly now accuse the Stoics of not preserving freedom of decision and of thus destroying moral responsibility (e.g. Fat. 190.1–19). The Stoics do not give up at this point. However their strategy is not to attempt to show that their causal fate-determinism is compatible with the Peripatetic concept of that which depends on us, i.e. the concept of free decision, based on metaphysical freedom. Instead, they take the offensive and set out to show that the Peripatetic concept of that which depends on us is unacceptable in that rather than preserving moral responsibility, in certain cases it actually makes it impossible to attribute praise and blame to an agent. And—as is implied—they hold that ⁶² One may of course doubt the consistency of his position in this respect. ⁶³ It combines two features: the first is that of indeterminism, i.e. the freedom from previous states of the world, including the agent; the second feature is that of active causation of the action or decision by the agent. The element of indeterminism alone also covers chance events. The element of self-causation is unique to beings that have some faculty of self-causation, e.g. a will, a faculty of assent or of choice, etc.

216

, ,   

ethics rather requires a concept of that which depends on us such as is promulgated by themselves. That is, the debate now focuses on the question of which is the right concept of that which depends on us. (This question was, as far as I can see, not discussed in Chrysippus’ time.) In outline the Stoic argument against their opponents’ concept of that which depends on us runs as follows (cf. Alex. Fat. 196.24–197.3).⁶⁴ Someone who is truly virtuous does not⁶⁵ have the ability to be vicious, which is the opposite of being virtuous (and hence presumably also lacks the ability to act viciously). So someone who is virtuous does not have the ability to be the opposite of what they are (and to do the opposite of what they do). But virtuous people are praised for being virtuous. And accordingly, their being virtuous (and presumably their acting virtuously) is assumed to depend on them.⁶⁶ Hence the account of that which depends on us as ‘that of which we can do the opposite’ must be wrong. This argumentation implies that for the attribution of moral responsibility it is not essential to have the ability to act differently (or choose otherwise) in the very same circumstances. What is essential—according to the Stoics—is that it is oneself who acts, that one’s action is performed as the result of an act of assent (Fat. 182.16–18, which for them presupposes that the agent is a rational being), and that this in turn is based on the state or disposition of one’s soul. Thus it is a concept of that which depends on us like the Stoic one that is needed for the attribution of moral responsibility. And this concept fits in with Stoic causal determinism—as Chrysippus had already shown. In conclusion, these later Stoics no longer simply endeavour to show (as Chrysippus did) the compatibility of causal fate-determinism with some undefined, presupposed, concept of that which depends on us. Rather, they attempted to show that the concept of freedom of decision as introduced by contemporary Middle Platonists or Peripatetics is not suitable as a basis for moral responsibility, whereas their—that is the Stoic—concept is suitable as a basis for moral responsibility, and hence as a basis for a system of ethics.

⁶⁴ Sharples 1983, 159 entertains the idea that this argument is not Stoic, but produced by Alexander himself. However, this seems unlikely to me, since (i) the argument is introduced as one of the arguments about which the opponents (of the Peripatetics) are most confident, and since (ii) the definition of that which depends on us used in the argument is not one of those Alexander uses otherwise in the De Fato, but a presumably earlier one, which we know also from Nemesius, Nat. hom. 114.21–4. ⁶⁵ The Greek has μηκέτ’, ‘no longer (able to be vicious / virtuous)’ (Fat. 196.27, 29). Perhaps Alexander slightly misrepresented the Stoic argument, which may have claimed only ‘inasmuch as they are virtuous / vicious they are not able to act viciously / virtuously’, cf. Stough 1978, 208–13. Alternatively, the argument may be ad homines, against the Peripatetics, cf. Sharples 1983, 160. ⁶⁶ This is, at least, the Peripatetic view, cf. Aristotle, EN 3.5.

8 Early Stoic Determinism The Merging of Teleology and Universal Causation

The distinctive feature of early Stoic determinism is its unique combination of teleological and mechanical elements. In this essay, I set out the details of how this combination works. (Section 1) Although from the second century  to the third century  the problems of determinism were discussed almost exclusively under the heading of fate, early Stoic determinism—as introduced by Zeno, the founder of the Stoa, and elaborated by Chrysippus—was developed largely in Stoic writings on physics, independently of any specific ‘theory of fate’. Stoic determinism was firmly grounded in Stoic cosmology, and some basic concepts from Stoic physics and ontology are indispensable for a full understanding of the theory. (Section 2) Stoic determinism was originally not presented as causal determinism, but with a strong teleological element and in the context of a theory of natural motions and states. (Section 3) However, Chrysippus also employed his conception of causality in order to explicate his determinism.¹ (Section 4) The teleological and mechanical elements of early Stoic determinism were brought together in Chrysippus’ conception of fate, which places elements of rationality in every cause and thus combines Nature or Cause with the individual causes within the world. The early Stoics used several analogies to explain how we are to think of this combination. Examination of these analogies in the context of Stoic philosophy illuminates how the Stoics provide for the coexistence of science and theology, as well as of the compatibility of science and moral responsibility and of theological predetermination and human moral responsibility.

This essay first appeared in Revue de métaphysique et de morale 48(4) (2005), 489–516, and is reproduced here with permission. ¹ A more detailed discussion of early Stoic determinism, its relation to freedom and moral responsibility, and later Stoic determinism can be found in Bobzien 1998a. For an in-depth study of Chrysippus’ theory of causes, see Bobzien 1999 (Chapter 9 in this volume).

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0009

218

, ,   

1. Physical and Ontological Foundations I start with some basic elements of Stoic physics and ontology that are required for an understanding of Stoic determinism. These are, in order, the active principle, causation, and changes, qualitative states, and events. For the Stoics, the world (kosmos) is a unitary and continuous body without any gaps; it is located in the void, and it contains no smallest parts.² Stoic physics is thus a continuum theory.³ The world is constituted of two principles, the active and the passive.⁴ The passive principle is called ‘matter’ and ‘substance’; it is amorphous and unqualified; it possesses neither power of cohesion nor power of movement.⁵ The active principle is a power that is eternal and self‐moved; it is responsible for all form, quality, individuation, differentiation, cohesion, and change in the world.⁶ Both principles are material.⁷ In physical terms, for Chrysippus, the active principle is pneuma or breath, which is a special combination of air and fire; the passive principle is a combination of earth and water.⁸ (For Zeno, in physical terms the active principle is creative fire.) The two principles form a complete blending (krasis) both in the world as a whole and in any object in the world.⁹ That is, they are completely co‐extended but they and their respective qualities are fully preserved in that mixture, and they are, in principle, separable again. The function of the active principle can be contemplated from two viewpoints: from a global perspective, which considers the whole cosmos as one unified entity; and from the inner-worldly perspective, which looks at particular objects and their interrelations. The active principle is responsible both for the cohesion, form, and change of the cosmos as a whole, and for the individuation, cohesion, form, change, and duration of the objects in the world. The individual objects are each held together (as the objects they are) by the active principle, which gives them a certain tension or tenor or cohesion (hexis). Different objects have different complexity, owing to the complexity of their tenor. With increasing complexity, inanimate objects have tenor (in the specific sense); plants have nature; non‐rational animals have soul; and rational beings have reason as their highest organizing principle. They each also have all the lower kinds of tenor. Physically, these kinds of tenor are pneuma of increasing purity or fineness. The finest pneuma, reason, is situated in beings of the highest order, i.e. rational beings, in the ruling part of their soul. The world as a whole is also a being

² DL 7.140, 143, 150; SE M 9.332. ³ For details on this point and for Stoic physics in general cf. Sambursky 1959, Bloos 1973, and Lapidge 1978. ⁴ DL 7.139. ⁵ DL 7.134, 139, 150; SE M 9.75; Sen. Ep. 65.2. ⁶ DL 7.139; SE M 9.75–6; Sen. Ep. 65.2; Plut. Stoic. Rep. 1054a. ⁷ E.g. Aristocles in Eusebius, Praep. Ev. 15.14.1. ⁸ Nemesius, Nat. hom. 52.18–9; Plut. Comm. Not. 1085c–d. ⁹ Alex. Mixt. 224–5.

  

219

of the highest order, i.e. a rational being.¹⁰ Like human beings, in addition to tenor, nature, and soul, it has reason. This rational organizing principle, the reason of the cosmos, is called ‘god’ or ‘god’s soul’, ‘nature’, or ‘ruling principle’.¹¹ Sometimes it is placed in the aether, as the accumulation of the finest pneuma in the cosmos. Sometimes all of the active principle, sometimes only its finest part, is called the ‘reason of god’. Stoic determinism is causal determinism in one of its main aspects. However, we must beware of rash comparisons with modern theories of causal determinism. Modern theories may consider causes as events, facts, things, or properties, but mostly there is the assumption that cause and effect belong to the same ontological category, and that what is effect in one instance of causation can be cause in a subsequent instance. On this point, the Stoics differ: for them causes and effects belong to two different ontological categories. This has various consequences for their account of determinism.¹² Here are two Stoic accounts of ‘cause’: Chrysippus states that a cause is that because of which; and that the cause is an existent thing and a body, and that the cause is ‘because’, and that of which it is a cause is ‘why?’.¹³ . . . the Stoics say that every cause is a body which becomes a cause, to a body, of something incorporeal; as for instance the scalpel, which is a body, becomes a cause, to the flesh, which is a body, of the incorporeal predicate ‘being cut’.*

Every instance of causation involves at least three main factors, two corporeal, one incorporeal. (For reasons of convenience, I individuate instances of causation by assuming one such instance per effect.)¹⁴ One corporeal is the cause, the other the object to which it is the cause, and at which the effect obtains. The effect, or that which is caused, is incorporeal and obtains at the second corporeal, and it is a predicate.¹⁵ In cases in which different causes work together in one instance of causation, the factors are multiplied. Moreover, cause and effects are relative (pros ti), that is, they are relative to each other, and inseparable: a cause is not a particular thing but that thing insofar as it produces its effect.¹⁶

¹⁰ DL 7.142–3. ¹¹ Plut. Stoic. Rep. 1052c; Alex. Mixt. 225.1–2; Plut. Stoic. Rep. 1053b. ¹² For Stoic theory of causation in general, see Frede 1980 and Bobzien 1998a. ¹³ Χρύσιππος αἴτιον εἶναι λέγει δι’ ὅ· καὶ τὸ μὲν αἴτιον ὂν καὶ σῶμα, * καὶ αἴτιον μὲν ὅτι, οὗ δὲ αἴτιον διά τι (Stobaeus, Ecl. 1.138.23–139.2). Some emendation is required. My suggestion (in angled brackets) is based on the close parallels for Zeno (Stobaeus, Ecl. 1.138.5–16) and Posidonius (Stobaeus, Ecl. 1.139.7–8). * . . . Στωικοὶ μὲν πᾶν αἴτιον σῶμά φασι σώματι ἀσωμάτου τινὸς αἴτιον γίνεσθαι, οἷον σῶμα μὲν τὸ σμιλίον, σώματι δὲ τῇ σαρκί, ἀσωμάτου δὲ τοῦ τέμνεσθαι κατηγορήματος (SE M 9.211). ¹⁴ The individuation of effects is discussed below. ¹⁵ Stobaeus, Ecl. 1.138.15–16; 139.7–8; SE M 9.211; Clement, Strom. 8.9 96.23–97.1. ¹⁶ Cf. e.g. SE PH 3.25, SE M 9.207.

220

, ,   

For the Stoics, thus, all interaction is interaction between bodies, with one body acting upon another. Incorporeals can neither act nor be acted upon.¹⁷ Hence effects, being incorporeal, can neither bring something about, nor can they be acted upon. When the interaction between bodies is described as causation, it is described as the relation between one or more bodies as causes and the incorporeal effect as obtaining at another body. The most basic Stoic distinction between interactions of bodies is that between (i) sustenance (sustaining-something/being-sustained-by-something) and (ii) change (changing-something/being-changed-by-something).¹⁸ Correspondingly, the most basic distinction of causes is that between causes of qualitative states (scheseis) and causes of changes (kinēseis): on the level of causal explanation, the Stoics not only ask for the cause why something has changed, but also for the cause why something keeps on being what it is and the cause of what state it is in. The requirement of causes of (the sustenance) of qualitative states¹⁹ is a consequence of the Stoic theory of the active and passive principle. The active principle is not only the sole source of change, it is also required for any qualitative state of an object to continue, and for any object to continue to be that object. The active principle is regarded as the cause of both changes (it brings them about) and qualitative states (it keeps them up).²⁰ The active principle is also the main reason why for the Stoics all causes are active causes, and Aristotelian formal, material, and final causes, as well as mere necessary conditions, do not qualify, but would merely count as circumstances.²¹ However, the active causes combine some of the functions of Aristotelian formal, final, and effective causes, and are involved equally in the causation of change and of qualitative states. A main characteristic of causes of qualitative states is that they are simultaneous with their effects. The most prominent causation of qualitative states is that involving a cohesive cause (sunektikon aition): the cohesive cause is that portion of pneuma in an object that is—and insofar as it is—responsible for the object being the object it is. This seems to be the same pneuma as the tenor in each object. Thus for the ongoing existence of an object, the presence of a cohesive cause is a necessary condition. The main characteristic of causation of change is that a cooperation of two causal factors is required, of which one has to be antecedent to its effect. Since for the Stoics all causes are corporeal, some elucidation is needed for what it means

¹⁷ Nemesius, Nat. hom. 21.6–7 (Morani); SE M 8.263; Plut. Comm. Not. 1085b. ¹⁸ All sustenance and change is two‐faced in this way: it depends on one’s chosen perspective whether one describes one body as affecting another, or that other body as being affected by the first; cf. Simplicius, Cat. 306.14–15. ¹⁹ I render σχέσις by ‘qualitative state’ in order to remind the reader that this is a kind of effect, i.e. something that requires a cause for its sustenance. ²⁰ Cf. e.g. Plut. Stoic. Rep. 1054a. ²¹ Cf. Sen. Ep. 65.4, SE PH 3.14.

  

221

for a body to be such an antecedent cause. For in most cases of causation this body will exist before, during, and after the effect obtains. ‘Antecedent’ can thus not refer simply to the time at which the body exists. One needs to take into account that qua cause the body actively contributes to its effect, and that causes are relative. Thus the body is the antecedent cause only insofar as it actively contributes to the effect. I thus suggest that, as a minimal condition, c is an antecedent cause of an effect e, if the period of time at which c is active in contributing to e precedes, at least in part, the period of time at which the effect obtains. The role of causation of qualitative states in Stoic theory of determinism is often neglected. This is so partly because it has no equivalent in later physical theories, and partly because it is the causal predetermination of changes that was the primary matter of contention, not the causal simultaneous determination of states. But the pair of nouns ‘change / qualitative state’ (kinesis/schesis) together with the corresponding pair of verbs ‘to change / to be in a qualitative state’ (kineisthai/ ischesthai) appear frequently in Stoic philosophy, and they denote a standard Stoic distinction.²² The individual changes and qualitative states are of special interest since they are precisely those entities that are effects of instances of causation. What is their ontological status? We can be confident that they are incorporeals, since they are effects, and effects are incorporeal. If they are incorporeals, they must be void, time, place, or ‘sayables’ (lekta)²³. Of these, they can reasonably only be sayables. This tallies with the account of effects as belonging to the ‘things that can be predicated’ or ‘predicates’, which are a subclass of the sayables. But matters are more complex. First, consider the definitions of changes and qualitative states: Chrysippus had, it seems, a narrower and a wider concept of change. He defined change as ‘alteration concerning space, either to the whole or to a part of it’.²⁴ Here change is locomotion. Another definition is ‘alteration concerning space or shape ’.²⁵ In this account, what is at issue includes any change or process. For, ‘shape’ (schēma) stands for all qualitative features of an object.²⁶ This wider definition is the one of primary relevance to the theory of determinism. Chrysippus also produced two definitions of ‘continuity’ (monē): ‘absence of change of a body’, and ‘qualitative state (schesis) of a body, concerning the same features and in the same way, now and before’.²⁷ The qualitative states, here, being a type of continuity, must be incorporeal. This

²² Cf. Plut. Stoic. Rep. 1050c–d, 1056c; Comm. Not. 1076e; Stobaeus, Ecl. 2.73.1, 82.11–17 and 95.6–8, DL 7.104, Cic. Fin 3.33; Origen, Orat. 2.368; Plot. Enn. 3.1.7; Stobaeus, Ecl. 1.166.24–167.14; Simplicius, Cat. 212–13. ²³ SE M 10.218. ²⁴ Stobaeus, Ecl. 1.165.15–17, cf. SE M 10.52. ²⁵ κίνησιν μεταλλαγὴν κατὰ τόπον ἢ σχῆμα , Stobaeus, Ecl. 1.165.17–18. The supplement is taken from Chrysippus’ previously quoted narrower definitions of κίνησις; it finds justification in its occurrence in Apollodorus’ parallel definition, Stobaeus, Ecl. 1.166.24–6. ²⁶ Cf. also SE Μ 9.75, Plut. Comm. Not. 1054a. ²⁷ Stobaeus, Ecl. 1.165.19–21.

222

, ,   

is confirmed by Apollodorus’²⁸ definitions of ‘qualitative state’ (schesis) as (i) ‘continuity/coherence (sunochē) concerning space or form’ and (ii) ‘the being in such a state’ (to ischesthai toiouto).²⁹ The second definiens defines an incorporeal. And it is qualitative state as equivalent to being-in-a-qualitativestate, and change as equivalent to changing, which are incorporeals, predicates, lekta, and the effects of causation.* Qualitative states, as is clear from the definition, have duration. So, too, do changes: in line with Stoic continuum theory and the indefinite divisibility of time, Apollodorus states that for every change there is a change that is part of it, and for every qualitative state there is such a state that is part of it.³⁰ Recalling the definitions of qualitative states, we can conclude that qualitative states are homogeneous: every state that is a temporal part of a state is of exactly the same kind as the state of which it is a part. Changes, though, are not all homogeneous: the fact that for every change there is a change that is part of it does not entail that the change that is part of it is of the same kind as the change of which it is a part. Rather, some Stoic changes are homogeneous, some are not. For further understanding of the Stoic concepts of change and qualitative state we have to follow the trail of the Stoic claim that they are predicates (since they are effects and effects are predicates). The Stoics define a predicate as ‘what is asserted of something, or a thing (pragma, here, as often, meaning “sayable”) attachable to some thing or things’.³¹ Examples for predicates are ‘to walk’ and ‘to be alive’. Predicates can be actualized, i.e. subsist at something or hold of something, or can be not actualized. When actualized, they are called ‘actualized predicates’ or ‘attributes’ (sumbebēkota).³² The being actualized of a predicate is best explained by example: the predicate ‘to walk’ holds of me, and is thus actualized at me, precisely when I am walking.³³ Whenever a predicate is actualized, it is so for a period of time tm to tn. The same predicate, ‘to walk’, can be actualized repeatedly, e.g. (at Thea) from 9 to 10 a.m. and again (at Thea) from 3 to 5 p.m. today; it would then presumably also be actualized (at Thea) from 9.30 to 9.45 a.m., etc.³⁴ We can now draw the connection between predicates and changes and qualitative states as follows: whenever a predicate is actualized, there exists and can be discerned an actualization of that predicate. So, in place of talking about a ²⁸ The Stoic Apollodorus was presumably a pupil of Antipater (Philodemus, Ind. Stoic. col. 53.7). The excerpt on σχέσεις and κινήσεις from his Handbook of Physics in Stobaeus, Ecl. 1.166–7 overlaps in large parts with Chrysippus’ accounts (Stobaeus, Ecl. 1.165–6). As there are no inconsistencies with earlier Stoic theory, I assume that the excerpt generally squares with Chrysippus’ view. ²⁹ Stobaeus, Ecl. 1.166.26–7. * For the connection between lekta and effects of causation, see now also Bronowski 2019. ³⁰ Stobaeus, Ecl. 1.167.9–14. Cf. also, for Chrysippus, Stobaeus, Ecl. 1.106.8–9. ³¹ DL 7.64. ³² Stobaeus, Ecl. 1.106.20–3. ³³ Stobaeus, ibid: . . . κατηγορήματα ὑπάρχειν λέγεται μόνα τὰ συμβεβηκότα, οἷον τὸ περιπατεῖν ὑπάρχει μοι ὅτε περιπατῶ, ὅτε δὲ κατακέκλιμαι ἢ κάθημαι οὐχ ὑπάρχει. ³⁴ If the time span becomes very small, perhaps no longer the predicate ‘to walk’ but some other predicate of change would be actualized at those times (at Thea), e.g. ‘is moving her right leg forward’.

  

223

predicate’s being actualized many times, one can talk about as many actualizations of the one predicate. Thus the various cases of walking are individuated, and what one arrives at are individual changes and qualitative states. For instance, in our example we have an actualization of the predicate ‘to walk’ from 9 to 10 a.m., from 3 to 5 p.m., and from 9.30 to 9.45 a.m., i.e. three different actualizations. Each has duration. The same holds for predicates that are concerned with qualitative states, e.g. ‘to be alive’, actualized at Dio; there were in fact indefinitely many actualizations of that predicate at Dio during Dio’s lifetime. Changes and qualitative states can thus be understood as a subclass of all actualizations of predicates: namely the subclass of those things that are effects. For instance, in the above example of ‘to walk’, actualized at Thea, we specified three motions. And in the case of ‘to be alive’, actualized at Dio, we had indefinitely many qualitative states. But is the actualization of a predicate a predicate? Asked in this way, the answer seems to be ‘no’. However, the phrase ‘actualization of a predicate’ was something I introduced for reasons of explication only. In fact, we can just as well say: the motion of Thea’s walking from 4 to 5 p.m. is the predicate ‘to walk’ actualized (at Thea) from 4 to 5 p.m. Thus individual changes and qualitative states can be understood as something like phases of predicates, more precisely, phases of predicates-while-they-are-actualized. If we look at changes and qualitative states qua effects, we obtain similar results. Effects are predicates. And effects, as dealt with by the Stoics in the context of determinism, are particulars. Thus, effects also must be predicates actualized at some thing from tm to tn. And indeed it is testified, at least for Zeno, that effects are sumbebēkota, that is, actualized predicates.³⁵ Thus changes and qualitative states, qua effects, are tokens, not types; they obtain in the world during some time tm to tn.³⁶ Besides ‘change’ and ‘qualitative state’, there is another expression prevalent in the debate over Stoic determinism: this is ‘that which happens’ or ‘events’ (gignomena), derived from the verb ‘to happen/to be’ (gignesthai). In Stoic sources ‘event’ and ‘change’ are sometimes used interchangeably.³⁷ However, ‘event’ is equally used to cover both changes and qualitative states.³⁸ In this second sense, events make up that important subclass of all actualizations of predicates that are effects. The Stoic concepts of events, changes, and qualitative states share elements both of contemporary concepts of events and of contemporary concepts of facts. The above exposition is, however, not meant to represent a Stoic theory of events or facts, as they themselves consciously developed it. Rather, I have tried to make apparent several features of their concepts of change, qualitative state, and event ³⁵ Stobaeus, Ecl. 1.138.15. ³⁶ This is not to deny that the Stoics ever talked about motions, qualitative states, and effects on a general level as well. On the contrary—it is likely that when the Stoics talked about general motions, qualitative states, or effects, they precisely took them to be predicates (κατηγορήματα). ³⁷ E.g. Cic. Fat. 20; Alex. Fat. ch. 13. ³⁸ E.g. Plut. Stoic. Rep. 1049f–1050d.

224

, ,   

which follow from theories we know they had about other things: causation, effects, sayables, predicates, and actualized predicates. As many modern philosophers build their theories without a fully worked out theory of meanings, facts, and events, and indeed sometimes mix up their scopes of application, so too did the Stoics.

2. Teleological Determinism Defined The only surviving passages in which Chrysippus systematically sets out and justifies his determinism are a number of quotes from his first book On Nature, preserved by Plutarch. This sequence of short passages shows how Chrysippus’ determinism grows out of the basic assumptions of Stoic cosmology and is thus anchored in Stoic physics: (1)

(2)

(3a)

(4)

(5)

. . . in his first book On Nature he said this: Since the organization of the universe proceeds in this way, it is necessarily in accordance with this organization that we are in whatever qualitative state we may be, whether contrary to our individual nature we are ill or maimed or have become grammarians or musicians . . . We shall on this principle state similar things both about our virtue and about our vice and generally about skills and the lack of skills as I have said . . . . . . that nothing at all, not even the smallest, is in a qualitative state or changes otherwise than in accordance with the reason of Zeus, which is the same as fate. For since the common nature stretches into all things, it must be the case that everything that happens in any way whatsoever in the universe and in any of its parts will have happened in accordance with that nature and its reason in unimpeded sequence; for neither is there anything to obstruct the organization from outside nor can any of its parts change or be in any qualitative state except in accordance with the common nature. . . . everywhere, but mainly in his physical books he [i.e. Chrysippus] has written: that there happen many obstacles and impediments for particular natures and (their) changes but none for the nature of the universe.³⁹

³⁹ Plut. Stoic. Rep.: (1) 1049f–1050a; (2) 1050a; (3) 1056c (cf. Stoic. Rep. 1050a and Comm. Not. 1076e); (4) 1050c; (5) 1056d: (1) . . . ἐν τῷ πρώτῳ περὶ Φύσεως . . . ταῦτ’ εἴρηκεν· Οὕτω δὲ τῆς τῶν ὅλων οἰκονομίας πραγούσης, ἀναγκαῖον κατὰ ταύτην, ὡς ἄν ποτ’ ἔχωμεν, ἔχειν ἡμᾶς, εἴτε παρὰ φύσιν τὴν ἰδίαν νοωοῦντες, εἴτε πεπηρωμένοι, εἴτε γραμματικοὶ γεγονότες ἢ μουσικοί. (2) . . . Κατὰ τοῦτον δὲ τὸν λόγον τὰ παραπλήσια ἐφοῦμεν καὶ περὶ τῆς ἀρετῆς ἡμῶν καὶ περὶ τῆς κακίας, καὶ τὸ ὅλον τῶν τεχνῶν καὶ τῶν ἀτεχνιῶν, ὡς ἔφην.

  

225

Passages (1) to (5) appear all to be part of one argumentative context: the thesis Chrysippus wants to back up is that every change and every qualitative state of any object is in accordance with the common nature of the world.⁴⁰ The above-mentioned distinction between global and inner-worldly perspectives is essential for Chrysippus’ reasoning, and is manifested in his distinction between common and individual nature. The common nature is the same as the active principle or god (3a, 3b).⁴¹ This universal nature extends into all things (4).⁴² It is responsible for the organization of the universe as a whole (4, 1, 3b), and it is unhindered (4). On the intra‐cosmic level naturally things are more complex. All objects in the world have their individual natures (1), and these go together with qualitative states and changes, which are in accordance with the object’s individual nature; e.g. for human beings, to be in a state of health and to move in accordance with reason are qualitative states and changes in accordance with nature.⁴³ But there are also counter-natural changes and qualitative states, i.e. those contrary to an object’s individual nature: being ill or maimed are examples for states (1), bodily movements that go beyond a person’s impulse (hormē) for movement;⁴⁴ and we can deduce from (1) that there are changes of an object that are externally forced, like being maimed. Counter‐natural states can be the result of counter‐natural changes. But they can also be frustrated natural changes: a tree’s growing can be prevented by a roof that is ‘in the way’. That is, on the inner-worldly level there can be impediments that interfere with the natural (and also with counter‐natural) changes of individual objects. An impediment is external to the object (or part of the object) whose changing is prevented. Changes and qualitative states that are not natural have the reason of their occurrence in the fact that there is a multitude of objects in the world and that they can get in each other’s way when performing their natural changes. Both

(3) . . . μηδὲν ἴσχεσθαι μηδὲ κινεῖσθαι μηδὲ τοὐλάχιστον ἄλλως ἢ κατὰ τὸν τοῦ Διὸς λόγον, ὃν τῇ εἱμαρμενῃ τὸν αὐτὸν εἶναι. (4) . . . Τῆς γὰρ κοινῆς φύσεως εἰς πάντα διατεινούσης, δεήσει πᾶν τὰ ὁπωσοῦν γινόμενον ἐν τῷ ὅλῳ καὶ τῶν μορίων ὁτωοῦν κατ’ ἐκείνην γενέσθαι καὶ τὸν ἐκείνης λόγον κατὰ τὰ ἑξῆς ἀκωλύτως· διὰ τὰ μήτ’ ἔξωθεν εἶναι τὸ ἐνστησόμενον τῇ οἰκονομίᾳ, μήτε τῶν μερῶν μηδὲν ἔχειν ὅπως κινηθήσεται ἢ σχήσει ἄλλως * κατὰ τὴν κοινὴν φύσιν. (5) . . . πανταχοῦ, μᾶλλον δ’ ἐν πᾶσι τοῖς Φυσικοῖς γέγραφε ταῖς μὲν κατὰ μέρος φύσεσι καὶ κινήσεσιν ἐνστήματα πολλὰ γίνεσθαι καὶ κωλύματα, τῇ δὲ τῶν ὅλων μηδέν. * The insertion of ἢ is justified by its occurrence in the parallels in (3), Stoic. Rep. 1050a, and Comm. Not. 1076e. ⁴⁰ Cf. Philodemus, Piet. col. 11.32–col. 12.1; Cic. Nat. D. 1.39. ⁴¹ Cf. Philodemus, Piet. col. 11.32–col. 12.1 and Cic. Nat. D. 1.39. ⁴² The same verb, ‘to extend’ (διατείνειν), is used to describe how the pneuma of the soul stretches through an animal’s body, e.g. Stobaeus, Ecl. 1.368. ‘Aetius’ 4.21 (Diels, Doxogr. graec.410–11); cf. also DL 7.138. ⁴³ Galen, PHP 4.2.12, quoting from Chrysippus. ⁴⁴ Galen, PHP 4.2.16–18, quoting from Chrysippus.

226

, ,   

counter‐natural changes (i.e. changes by force) and counter‐natural states (prevention of an object’s natural changes or results of counter‐natural changes of an object) are the result of such clashes. The interplay of the objects leads to the result that one object—in moving naturally—hinders another from moving or from being in a certain state; and one object can by force move another object, or change the state it is in, or even destroy it. Chrysippus’ reasoning in his first book On Nature is then based on the relation between cosmic and intra‐cosmic perspectives. His main thesis is that ‘every qualitative state and every change in the world is in accordance with the universal nature’. Our passages suggest that Chrysippus put forward the following two reasons to back up this thesis. First, nothing can obstruct or destroy the organization of the universe from outside, for the reason that there is nothing (no thing) external to the universe— only the void. The natural movement of the universe as a whole can thus not be thwarted.⁴⁵ (Why can the impediments at the level of individual objects not be regarded as obstacles to the nature of the universe? This is Plutarch’s criticism.⁴⁶ We should expect Chrysippus’ reply to be (a) that the common nature extends into all these obstacles, too; (b) that all obstruction can in the end be reduced to natural changes of some things (that are part of the world); and (c) a teleological reason, viz. that these obstructions lead to the realization of the best possible world.) Second, nothing in the world can be in a qualitative state or change otherwise than in accordance with the universal nature, since the universal nature is ‘all‐ embracing’: it includes the individual natures, in that it extends into the individual objects, making up their individual nature (4). Although it would make sense, say, that my individual nature goes against yours, it makes no sense to say that my individual nature goes against the universal nature, for my individual nature is part of the one universal nature.⁴⁷ Combining the two reasons, we can see that the thesis that every state and every change is in accordance with the one universal nature (and its organization) is a consequence of the Stoic conception of the world as a unity that is held together by the one active principle. The teleological aspect of the active principle is taken into account in our passage by the references to the organization of the world (1), (4), and to the reason of the universal nature (3), (4). There can be little doubt that Chrysippus proclaims some sort of universal determinism in these passages from On Nature. His formulations in (1), (3), and (4) suggest that the universal nature leaves no room for alternative developments ⁴⁵ Cf. Cic. Nat. D. 2.37–8. ⁴⁶ Stoic. Rep. 1056d. ⁴⁷ What about the passive principle, matter (ὕλη)? In the same first book On Nature, Chrysippus applies the same distinction between global and inner-worldly perspectives to it as well. Whereas the matter of the individual things both increases and decreases as a result of the workings of the active principle, the matter of the whole world remains constant (DL 7.150; Stobaeus, Ecl. 1.133.6–11).

  

227

of the world. There is exactly one course of events that is in accordance with the rational universal nature, and that is the course of events which is the actual one. But what sort of determinism is this? 1. As to its scope, it is universal: first, changes and qualitative states are equally determined (by the rational universal nature). Second, every qualitative state and every change is included. Given that Stoic physics is a continuum theory, changes and qualitative states can be indefinitely small; but however small, they are included in the universal nature. There is no room for quantum leaps, nor is Nature or god concerned with the ‘weighty’ events only. Later Stoics illustrated this point with the raising of one’s eyebrow, stretching out of one’s finger, or turning one’s head as examples for events that are trifling but predetermined nonetheless.⁴⁸ 2. The world, as it is determined, is not chaotic, or like something produced by a random generator. Rather, it is an ordered and organized whole (1), (4). 3. This organized whole has not evolved randomly (‘order from chaos’). Rather, its development follows a rational principle of organization: the reason of nature or Zeus (3), (4). 4. This rational organizing principle works completely immanently in the world, and also from the inside of each individual object. 5. There is no universal inner-worldly harmony of the kind that every object can realize its individual nature unhindered and unforced. Rather, the rational organizing principle works like this: the individual nature in each object provides the object with its characteristic qualities and activities, which in favourable circumstances it will display and perform. However, not all objects realize their individual nature in all respects. Rather, some objects prevent others from performing their natural changes, and some objects force others to go through certain counter‐natural changes. The world is such that the objects are, as it were, left to battle the conflicts out between themselves. Yet—from the cosmic perspective—the way this happens does not include any element of chance, for it is in accordance with the reason of the world, which works from the inside of these objects. 6. Chrysippus expressly mentions our virtues and vices as among the qualitative states that cannot be in any other way than they are (1), (2), and there is no hint that he considered this problematic. Virtues and vices, being dispositions (diatheseis), are part of the portion of active principle in us. This picture of universal determinism is far removed from modern theories of mechanical, causal determinism, and is so on two counts in particular. First, the

⁴⁸ Alex. Fat. 175.6–13.

228

, ,   

overall picture is strictly teleological: there is a rational organization of the world, and every state and change conforms to the rational organizing principle. Second, causation is not mentioned in any of the passages, nor is there any notion of empirically detectable regularity or individual laws of nature that govern all events. The changes in the world—like the qualitative states—are defined in terms of the natural changes and states of the objects in the world, and of the counter-natural changes and obstructions that are the consequence of many different objects trying to realize their natural changes. What makes an object change ‘naturally’ is primarily the active principle in that object, although such changes always presuppose some external stimulus. One could express the Stoic theory in terms of ‘laws of nature’, but these would be laws concerning the individual natures of individual objects. They would have the form: an object with the individual nature n will perform changes of type m as long as the circumstances are favourable, i.e. as long as it is—appropriately prompted and—not prevented by external circumstances from doing so. Such ‘laws of individual natures’ connect objects that have certain qualities with certain changes, in line with Stoic theory of causation.⁴⁹ They are, however, not suitable to define determinism—not so much because they do not connect events with events, but, rather, (a) since they do not take into account the circumstances in which the objects find themselves, and (b) since they only account for natural changes. Precise prediction of the future on the basis of such laws, even if one knew them all, is not possible. But it would be a mistake to infer from this that Chrysippus’ determinism was not a causal determinism in which efficient causes play a central role. In order to see how Stoic determinism and Stoic theory of causation fit together, we have to turn to another passage in Plutarch.

3. Causal Determinism Defined In antiquity, Stoic physics stands out not so much because it is a deterministic system but because it contains a worked out theory of universal causal determinism. No early Stoic exposition of this theory has survived, but a passage in Plutarch gives us the basic information we need. The passage is a summary from a Chrysippean work; it consists of an argument by some opponents of the Stoics (the ‘arguments from indistinguishables’) and Chrysippus’ reply. I start with these. ⁴⁹ But it is unlikely that the Stoics themselves thought of laws of such a kind as governing the natural movements, or that they would take recourse to such laws for explaining such movements. Rather, the rational pneuma in every natural object will have been thought to contain the relevant ‘information’, embodied in the specific state of tension of their pneuma, and this will make the object perform its natural movements, if prompted. The regularity results from the fact that natural objects of a certain kind all have the same kind of ‘information’ stored in their pneuma.

  

229

Here is Plutarch’s presentation of the opponents’ argument from indistinguishables: (1) Some philosophers, believing that they bring about release for the impulses from being forced by external causes, construct in the ruling faculty of the soul some adventitious motion (2) which becomes evident best in the case of indistinguishables. (3) For when it is necessary to take one of two things, (4) when the two are of equal power and are in the same state, (5) [and] when no cause leads to either of them, (6) since it in no way differs from the other, (7) [then] the power of spontaneity itself in the soul, by taking from itself an inclination, (8) cuts through the puzzle.⁵⁰ (Stoic. Rep. 1045b–c)

The philosophers in this passage try to justify the existence of spontaneous motions by singling out a subclass of the situations in which such changes are assumed to occur and in which their introduction appears most plausible. These are situations in which an agent encounters some indistinguishable alternatives between which to choose (2). The philosophers’ argument for their point is crammed into (3)–(8):⁵¹ they make the assumption that if two alternatives are indistinguishable, then no (external) cause will lead the faculty of impulse to either (see (6), (5)). They also presuppose that at least sometimes an action ensues. From these two assumptions they seem to infer that in the cases in question impulse and action occur without an external cause. The power of spontaneous motion in the mind is thus introduced to explain how the agent’s predicament was solved (8). This power takes from itself (i.e. not from outside) an inclination for one of the two alternatives. What sort of power it is we are not told. The argument neither states nor entails motion without a cause (with the Hellenistic concept of corporeal causes). Rather, the passage propagates some kind of ‘agent causality’ based on a self‐moving force in the agent’s soul (7). The spontaneous motion is introduced not as causally undetermined, but as ‘un-predetermined’. (See Chapter 1 for the distinction between ‘undetermined’ and ‘un-predetermined’.) Plutarch summarizes Chrysippus’ reply thus: (9) Chrysippus, speaking against these, as people who violate nature through the uncaused, (10) adduces in many places the die and the scales and many of the things that cannot take now one fall or inclination, now another (11) without some cause and there being a difference either concerning just the ⁵⁰ (1) Τοῦ κατηναγκάσθαι δοκοῦντες ὑπὸ τῶν ἔξωθεν αἰτιῶν ταῖς ὁρμαῖς ἀπόλυσιν πορίζειν ἔνιοι τῶν φιλοσόφων ἐπελευστικἠν τινα κίνησιν ἐν τῷ ἡγεμονικῷ κατασκευάζουσιν, (2) ἐπί τῶν ἀπαραλλάκτων μάλιστα γιγνομένην ἔκδηλον· (4) ὅταν γὰρ δυοῖν ἴσον δυναμένων καὶ ὁμοίως ἐχόντων (3) θάτερον ᾖ λαβεῖν ἀνάγκη, (5) μηδεμιᾶς αἰτίας ἐπὶ θάτερον ἀγούσης (6) τῷ μηδὲν τοῦ ἑτέρου διαφέρειν, (7) ἡ ἐπελευστικὴ δύναμις αὕτη τῆς ψυχῆς ἐπίκλισιν ἐξ αὑτῆς λαβοῦσα (8) διέκοψε τὴν ἀπορίαν. ⁵¹ I pass over various ambiguities.

230

, ,   

things themselves or concerning the external circumstances; (12) for, he thinks, the uncaused is completely non‐existent, and so is self‐motion, (13) and in the case of these things which some people invent and name ‘spontaneous’, (14) concealed causes sneak in and, (15) without our noticing it, they lead the impulse to one of the two alternatives.* (Stoic. Rep. 1045c)

Chrysippus proceeds as follows. First he states that the introduction of the power of spontaneous movement amounts to the introduction of uncaused motion (9). Then he gives an alternative explanation of what happens in the situations of indistinguishables (10)–(15). Since the intra-psychical events are not open to direct investigation, instead of a direct account, he presents an analogy from ‘everyday physics’ to the situations of indistinguishables.⁵² The die and scales analogy is given in a very compressed form. But we can secure from the text a few general points about the situations at issue in the analogy: • The analogue (the analogous case of the analogy) concerns the same object (a die, a pair of scales), or two identical-looking objects in two spatio‐ temporally distinct situations. • In these two situations the object, or the two objects, react in noticeably different ways. • The reactions are characteristic movements of the objects at issue: falling on one of its sides (die); inclining to one side (scales). • From the fact that the situations are meant to be in some way analogous to those of the indistinguishables, we can infer that in each case the two starting situations must at least prima facie look indistinguishable in all relevant respects. (For example, a die, thrown twice, under seemingly identical circumstances, comes up once with one side, then with another.) Chrysippus then explains what is really going on in these cases: the seemingly indistinguishable starting situations are in fact not completely the same; there is a difference either in the object (die, scales) or in the surroundings which is responsible for the difference in the outcome. And—given the assumption that the situations appear alike in all relevant respects—this difference must be hidden to the observer. Thus Chrysippus clearly acknowledged the fact that tiny * (9) πρὸς τούτους ὁ Χρύσιππος ἀντιλέγων, ὡς βιαζουμένους τῷ ἀναιτίῳ τὴν φύσιν, (10) ἐν πολλοῖς παρατίθησι τὸν ἀστράγαλον καὶ τὸν ζυγὸν καὶ πολλὰ τῶν μὴ δυναμένων ἄλλοτ’ ἄλλας λαμβάνειν πτώσεις καὶ ῥοπὰς (11) ἄνευ τινὸς αἰτίας καὶ διαφορᾶς ἢ περὶ αὐτὰ πάντως ἢ περὶ τὰ ἔξωθεν γιγνομένης· (12) τὸ γὰρ ἀναίτιον ὅλως ἀνύπαρκτον εἶναι καὶ τὸ αὐτόματον, (13) ἐν δὲ ταῖς πλαττομέναις ὑπ’ ἐνίων καὶ λεγομέναις ταύταις ἐπελεύσεσιν (14) αἰτίας ἀδήλους ὑποτρέχειν (15) καὶ λανθάνειν ἡμᾶς ἐπὶ θάτερα τὴν ὁρμὴν ἀγούσας. ⁵² Chrysippus and Zeno used this method repeatedly, cf. Gellius, NA 7.2.11, Cic. Fat. 42–3, Acad. 2.145, Galen, PHP 4.2.14–18.

  

231

variations in the antecedent situation can lead to—comparatively—large changes in the outcome. Since Chrysippus intends to argue by analogy, we can assume that he thought that his audience accepts that there are such unnoticeable differences present in the case of the die and scales. For only then can the analogue be used to elucidate Chrysippus’ explanation of what happens in the human mind in the situations of indistinguishables. Chrysippus draws the analogy in (13) to (15): as in the case of the die and scales, so in the case of indistinguishables the reactions are neither uncaused nor spontaneous; rather, there are factors concealed from us which are responsible for the agent’s impulse to go for one and not the other alternative. In parallel with the case of the die and scales, these hidden factors should be either in the person’s mind or in the surroundings. We can assume Chrysippus to have concluded that hence the so‐called indistinguishable alternatives do not prove the existence of spontaneous—and thus uncaused—self‐motion. So far Chrysippus’ reply. A comparison between the two arguments shows that the controversy between Chrysippus and his opponents is at base metaphysical. The opponents, starting from the premises that there is no difference in the antecedent data and that one option is actually chosen, conclude that there is spontaneous ‘self‐movement’. Chrysippus, starting from the premises that there is no uncaused motion (and that spontaneous self‐motion implies uncaused motion) and that one option is chosen, concludes that the opponents’ assumption of ontological indistinguishability in the antecedent data is false: there are differences, but they are hidden; the indistinguishability is merely epistemic. Chrysippus thus rightly rejects the validity of his opponents’ step from epistemic to ontological indistinguishability. But he in turn can conclude that there is always a hidden difference only by virtue of his assumption that there are no uncaused motions. And this assumption is as metaphysical as the one that there are ontologically indistinguishable situations. How is Chrysippus’ argument related to causation? We know from elsewhere that the causal factors responsible for the movements in the case of the die would be the person throwing the die and the nature (i.e. shape, etc.) of the die.⁵³ But Chrysippus, in his explanation of the analogue, says there is a difference either in the object (the die) or in the surroundings (which should include more than just the person who throws the die). That means that there is not necessarily a difference in the causes in the Stoic strict sense. There is a difference somewhere in the overall situation—perhaps in the causal factors, perhaps in the circumstances. And this difference explains the difference in outcome. This is why we have to read (11) as saying that dice cannot fall now in one way, now in another ‘without some cause and [without] there being a difference either concerning just

⁵³ Cf. Gellius, NA 7.2.11, Cic. Fat. 42–3.

232

, ,   

the things themselves or the external circumstances’. That is: there are causes for every fall of a die; and if there are different outcomes in comparable situations, this is due to a difference either in those causes or in the surroundings. When Chrysippus draws the analogy, he emphasizes the presence of causes: ‘and in the case of these things which some people invent and name “spontaneous”, concealed causes sneak in’ (above (13), (14) in response to (5)). This is so because the situation of indistinguishables is assumed to leave the agent in a stalemate, i.e. inactive, if there is neither a difference in overall situation nor a spontaneous motion. Thus, in order for something to happen in such situations, there must be an active causal factor involved—whichever way the agent decides. (The opponents assumed that in the cases of indistinguishables there are no known such causes.) What does Chrysippus’ argument tell us about Stoic determinism? The Plutarch passage allows us to extract two types of principles of universal causality which were part of Chrysippus’ theory. First, Chrysippus assumes a principle, ‘the uncaused and the self‐moved are non‐existent’. Variations of this principle are recorded for him elsewhere. Thus Cicero reports that Chrysippus maintained that there is ‘no change without cause’ and that ‘nothing can happen without a cause’.⁵⁴ This principle is, in the first instance, about changes or events; it links corporeal objects (the causes) with motions. (The formulation in Plutarch does not mention motion, but the whole argument is about the motions of the soul.) I call it the ‘General Causal Principle’: (GCP)

Nothing happens without a cause.

Elsewhere, Chrysippus backed up this principle with the Principle of Bivalence. Here, he seems to justify it by the idea that uncaused motion would violate nature (9). As long as the General Causal Principle is not further specified, it does not, within Stoic philosophy, entail determinism, since causes are not events, but bodies, considered in relation to the event they bring about. Thus the principle does not preclude that a change is not fully determined by its causes. Nor does it rule out that in relevantly similar situations an object is once the cause of one kind of change, once of another. That is, it is not ruled out that there is an element of spontaneity in the (thing that is the) cause. The General Causal Principle does not guarantee universal regularity or uniformity between cause and effect. But such uniformity is the hallmark of most modern theories of causal determinism.

⁵⁴ Fat. 20 and Div. 2.61. And Galen writes: ‘For in this way they would concede some uncaused change, something they commend us to beware of ’ (PHP 4.4.35–6).

  

233

This lack of specification cannot be made up for by the fact that we have determinism of some (teleological) kind secured for Chrysippus elsewhere (cf. section 2). For it does not follow that, if we combine this teleological determinism with the General Causal Principle, we obtain universal causal determinism. In order to show that Chrysippus’ determinism was universal causal determinism, we need to show that he holds a different kind of physical principle. We find such a principle underlying our Plutarch passage. The kind of principle I have in mind belongs to the family described by such catchphrases as ‘same causes—same effects’ or ‘like causes—like effects’. I dub such principles ‘specified causal principles’. We are nowadays all familiar with such principles, and indeed many consider them trivial. However, this ‘triviality’ is based on the fact that in everyday life we have absorbed a certain physical world view, thinking of causes and effects as events, and as governed by ‘laws of nature’ which somehow connect types of causes and types of effects. And we seem to hang on to this kind of idea of the world fairly undisturbed by the fact that it has little to do with modern physics—or philosophy of causation for that matter. However, this world view has not always been the prevalent one. Someone had to come up with the idea of such a specified causal principle first, and, evidently, with a concept of cause as corporeal, any version of such a principle will have to look different than do the modern ones. Let us return to Chrysippus’ reply to the proponents of spontaneous motions. It is plain that he maintains the General Causal Principle in his reply (12). But how is a specified causal principle involved in his reasoning? One way of understanding specified causal principles of the kind ‘like causes—like effects’ is that they imply the existence of a plurality of particular empirical causal laws which state some universal regularities, and which allow us to say something like ‘effect e happened because cause c preceded’, based on reasoning of the kind ‘ceteris paribus, whenever a cause of type C, then an effect of type E; hence in this case e (which is of type E) followed, since c (which is of type C) preceded, and cetera pariba’. But this is not what we find for Chrysippus. In our passage, what Chrysippus intends to show is that in the case of any one motion (including those in situations of indistinguishables) causes and circumstances jointly determine the motion in every detail. To show this, he argues by analogy from two (or more) similar motions. The analogy makes use of the following type of situation: we have two sections of the spatio‐temporal universe, each consisting of a starting situation s which includes an object o, and a subsequent change of o, which ends in a resultant state r. The two starting situations, as far as they are considered pertinent to the changes, are epistemically indistinguishable but temporally and/or spatially distinct. The subsequent changes, however, differ from each other in that they lead to noticeably dissimilar resultant states. Chrysippus inferred from this—and takes his audience to agree with him—that the two epistemically indistinguishable starting situations are in fact ontologically distinct.

234

, ,   

So it looks as if Chrysippus has some kind of specified causal principle on the side of the analogue, roughly ‘difference in effects—difference in overall starting situations’; or more precisely: (SCP‐analogue)

When from two seemingly indistinguishable starting situations two different effects ensue, then these situations were ontologically distinct: either in the object or in the surroundings there was a factor that differed.

This looks like the converse of the principle ‘like causes—like effects’. More accurately, it is contraposed to ‘no difference in starting situations—no difference in effects’. Why has Chrysippus the converse of what we usually find? The answer could be that when formulating the principle in contraposition, starting with the effects in the antecedent, as is implied by our text, Chrysippus conveniently bypasses the difficulty of having to determine what is relevant to the starting position. For in the form of such a principle that begins with the starting situation, a notorious difficulty is to determine what counts as part of the starting situation. First, the factors can be indefinitely many, owing to the Stoic continuum theory. Then, given the Stoic theory of sympathy, i.e. of the physical influence of everything on everything, the relevant starting situation may well include the whole universe! Does Chrysippus base the principle (SCP‐analogue) on the assumption of the existence of particular empirical causal laws? First, here we have to distinguish between what we would consider empirical causal laws and what Chrysippus would. For the modern idea of uniformity between certain kinds of causes and certain kinds of effects, where both causes and effects are events, is very different from Chrysippus’ ideas of active causation and corporeal causes. And our passage certainly does not entitle us to assume that Chrysippus had empirical laws built on his own concept of cause in mind, since in SCP‐analogue the relevant differentiating factor in the comparable situations is either cause or surrounding. And if he had had something of the modern idea in mind, it would not have been qua causal laws. Second, neither from SCP‐analogue nor from the Plutarch passage itself can we infer that Chrysippus thought of the world as exhaustively determined by a plurality of particular empirical laws, short of laws that govern the entire situation of the world at a time as starting situation (and which I would not call ‘particular’). However, these points are in any case irrelevant since principle SCP‐analogue is not the one Chrysippus is after. For the empirical situation of scales and dice is only the analogue in his analogy. He argues from this analogue, i.e. from two empirical spatio‐temporally different situations to one situation together with counterfactual reasoning about this one situation. First, the text implies that Chrysippus elaborates from the empirical analogue of his Specified Causal Principle SCP‐analogue as follows: in starting situation s₁ there is some factor f₁ that is responsible for the fact that e₁ and not e₂ (e₃, . . . ) happens,

  

235

and in starting situation s₂ there is some factor f₂ that is responsible for the fact that e₂ and not e₁ (e₃, . . . ) happens. f₁, f₂, . . . need not be causal factors; for instance, f₁ can be the presence of some causal or some hindering element, in the object or the surroundings, and f₂ the absence of that element. Chrysippus then draws the analogy to one starting situation in which there appear to be two equally likely outcomes, i.e. to the situations of indistinguishables. These are situations of human choice, but for the present this fact is immaterial. Chrysippus’ point, as relevant here, is this: (SCP‐indist.)

If in one starting situation s it looks as if there are two different but equally likely outcomes e, e*, and e occurs, then there is a factor f in s such that because of f e (and not e*) occurred; that is, had instead of f f* (≠ f) been present, then e* (and not e) would have occurred.

The empirical case of two numerically different starting situations was only an analogue. Here, now, we can see that Chrysippus’ Specified Causal Principle is based on counterfactual reasoning about one and the same situation. Particular empirical causal laws play no role in it. The point is made, as before, negatively, by introducing one factor that differs from the overall starting situation (this time this factor is counterfactual: ‘If the effect were different, there would have been a difference in the starting situation’). Accordingly, again there is no need to describe a ‘causally relevant section’ of the world.⁵⁵ By generalizing from the situations of indistinguishables to all movements, we can formulate Chrysippus’ (unrestricted) Specified Causal Principle. If we include Chrysippus’ notion of cause and the General Causal Principle (as (i)), we obtain: (SCP)

For every change e (i) there are causal factors c₁ . . . cn which are actively responsible for e, and (ii), for any possible alternative change e* to e there is a factor f in the starting situation s of e, such that because of f e (and not e*) occurred; i.e. for e* (and not e) to occur there would have to have been some f* (≠ f) instead of f in s.

Add to this the fact that for the Stoics motions are indefinitely divisible, and you can see that here we have a specified causal principle that defines a fully deterministic system. This principle does not involve particular empirical causal laws, nor does it rely on comparable empirical situations with comparable effects— except as analogy. Rather, it is concerned with the total actual state of the world or

⁵⁵ We have no evidence of converse formulations, starting with similarity of antecedent situations, rather than difference in resulting states, before the second century ; cf. Alex. Fat. 192.22–4 and Nemesius, Nat. hom. 105.18–21.

236

, ,   

an unspecified part thereof. Backing of the principle would not be expected from empirical quarters. (Accordingly, no element of predictability comes in.) Instead, we would expect the justification of the principle to come from cosmological, theological, or teleological theory. Nonetheless, with this principle SCP, we have (in (ii)) a full formulation of causal determinism in the modern sense. For the principle entails that every change is fully determined by the antecedent situation. If the outcome were any different, in however minute a detail, then the antecedent situation would also have been different. However, remember two points. First, Chrysippus’ determinism is ‘stricter’: in addition to uniformity (of whole-world-state starting situations and effects), there is active causation—in the Stoic sense—involved in (i) in SCP, which moreover is regarded as the main determining factor. Second, what we may consider to be ‘causality’ in this account of determinism was not regarded as causality by Chrysippus, and if he had considered his determinism as causal, then not because of (ii) in SCP. These two points exemplify the substantial difference between the Stoic and any prevalent contemporary views, and illustrate at the same time how a different conceptual approach can still manifest universal causal determinism.

4. Teleological and Mechanical Aspects Combined: Fate The question that remains is: how does Chrysippus’ seemingly mechanical causal determinism as tentatively captured by SCP combine with the teleological determinism proclaimed in the passages discussed in section 2? How do scientific and theological elements of his theory come together? For an answer, I turn to the early Stoic concept of fate. For Chrysippus, the causal and the teleological aspect of determinism are neither alternative nor mutually exclusive explanations of the world. Rather they complement each other in one comprehensive theory. The teleological part of the theory leaves it undetermined in which way exactly the rational principle makes the world develop in accordance with it. This is where causation comes in. The early Stoic theory of fate gives us some indication of how this works. Chrysippus stated that fate was the same as the active principle, as god or Zeus, as providence, as the nature of the universe, as the reason and the will of Zeus, and as the reason and the cause of the world.⁵⁶ Most of these identifications are known already for Zeno, who adorned the creative Reason or fire with the names of ‘fate’,

⁵⁶ Plut. Stoic. Rep. 1049f; Philodemus, Piet. col. 11.31–col. 12.2; Plut. Stoic. Rep. 1056c, Comm. Not. 1076e, cf. Diogenianus in Eusebius, Praep. Ev. 323.10–11; Stobaeus, Ecl. 1.79.5–7, Plut. Stoic. Rep. 1055e.

  

237

‘god’, ‘soul of Zeus’, and the ‘necessity of all things’, and maintained that fate is the same as providence and Nature.⁵⁷ Chrysippus is thus taking over traditional Stoic doctrine. All these identity statements suggest that at least the extensions of the terms involved are the same. This is confirmed by the fact that physically all these entities are considered to be pneuma by Chrysippus.⁵⁸ But the meanings of the terms were not all the same. For instance, although Chrysippus and Zeno identified fate with god or god’s reason, some properties that belong to god are never associated with fate: e.g. god is said to be perfect in happiness, blessed, benevolent, caring, beneficent, and not admitting any evil.⁵⁹ On the other hand, all properties of fate seem to have parallels in those of god, except that for god the detailed description of the network of causes (see below) is not recorded. Thus, although co‐extensive with god, fate seems to denote a particular group of aspects of god, i.e. god qua being the active principle that structures and moves the world. From Chrysippus’ accounts of fate we can extrapolate four aspects of the active principle that were predominantly connected with the term ‘fate’. Some had been traditionally associated with fate, others are specifically Stoic. Gellius has preserved a Chrysippean definition of fate in which all the aspects are assembled. Fate is a certain natural arrangement of the universe, with things following upon other things and being involved with other things from eternity, such a weaving‐ together being inexorable.⁶⁰ (Gellius, NA 7.2.3)

Stobaeus presents a collection of accounts of fate, taken from Chrysippus’ works: Chrysippus the substance of fate is a power of breath, administering order of the all. This he does in his second book On the World. But in his second book On . . . in his books On Fate, and occasionally in others he puts forward various views, stating that ‘Fate is the Reason of the universe’ or ‘the Reason of the things in the universe administered by providence’, or ‘the Reason in accordance with which past events have happened, present events happen, and future events will happen’; and instead of ‘Reason’ he uses ‘Truth’, ‘Cause’,

⁵⁷ Tertullian, Apol. 21, Lactantius, SVF 1.160, DL 7.135, cf. Stobaeus, Ecl. 1.133.3–5, 1.78.18–20. ⁵⁸ Stobaeus, Ecl. 1.79.1–2; cf. SE PH 3.218; [Plut.] Epit. 1.7 (Diels, Doxogr. graec. 306.1–6); Alex. Mixt. 224–5; Clement, Strom. 5.14.89.2. ⁵⁹ DL 7.147; Plut. Comm. Not. 1075e. ⁶⁰ φυσικήν τινα σύνταξιν τῶν ὅλων, ἐξ ἀϊδίου τῶν ἑτέρων τοῖς ἑτέροις ἐπακολουθούντων καὶ μεταπολουμένων*, ἀπαραβάτου οὔσης τῆς τοιαύτης ἐπιπλοκῆς (Gellius, NA 7.2.3). * The text is corrupt; I follow Sharples’ suggestion, Sharples 1991, 96 and 197.

238

, ,   

‘Nature’, and ‘Necessity’ and adds other terms which apply to the same substance from different perspectives.⁶¹ (Stobaeus, Ecl. 1.79.1–12)

Both passages confirm that ‘fate’ was used to refer to a whole set of aspects of the active principle. First, there is an element of teleology, rationality, organization, and order. This is an aspect that was prevalent in Chrysippus’ exposition of the nature of the universe (see section 2), which—as the passages just quoted show—is essential also to the meaning of ‘fate’: fate is a natural arrangement or organization, the reason of the world and the order which administers everything.⁶² What has been said about the common nature of the world in section 2 holds generally for this aspect of fate. Second, like the active principle and god, fate is described as eternal.⁶³ That is, (a) fate itself does not have a beginning, but has always been there as the organizing principle of the world; (b) the organization is such that—in some way—whatever occurs had been organized, hence settled, before it occurred. And this advance organization or fixing is also eternal: what occurs was always organized and settled to occur. Thus fate does not ever determine future events in any finite time before they happen.⁶⁴ In particular, a person’s destiny is not determined at the time of their birth or conception, as some popular views had it.⁶⁵ The third aspect—necessity, inevitability, and immutability—is traditionally connected with fate. It is also recorded for Zeno⁶⁶ and is repeatedly emphasized in Chrysippus’ accounts. Fate is called ‘greatest Necessity’,⁶⁷ and there is a colourful selection of adjectives used to express this aspect of fate: e.g. ‘invincible’, ‘unpreventable’, ‘immutable’, ‘unchangeable’, and ‘inexorable’ (aparabatos), which becomes the standard attribute of fate in later texts. We have seen above the main reason for the inevitability of fate: nothing external to the universal Nature (i.e. fate) can interfere with what occurs, since there is nothing external to it. In particular, human beings cannot interfere, since their natures are themselves part of the common Nature or fate. Connected with the eternal predetermination

⁶¹ Χρύσιππος δύναμιν πνευματικὴν τὴν οὐσίαν τῆς εἱμαρμένης, τάξιν* τοῦ παντὸς διοικητικήν. Τοῦτο μὲν οὖν ἐν τῷ δευτέρῷ Περὶ κόσμου, ἐν δὲ τῷ δευτέρῳ Περὶ . . . καὶ ἐν τοῖς Περὶ τῆς εἱμαρμένης καὶ ἐν ἄλλοις σποράδην πολυτρόπως ἀποφαίνεται λέγων· εἱμαρμένη ἐστὶν ὁ τοῦ κόσμου λόγος, ἢ λόγος τῶν ἐν τῷ κόσμῳ προνοίᾳ διοικουμένων, ἢ λόγος καθ’ ὃν τὰ μὲν γεγονότα γέγονε, τὰ δὲ γινόμενα γίνεται, τὰ δὲ γενησόμενα γενήσεται. Μεταλαμβάνει δ’ ἀντὶ τοῦ λόγου τὴν ἀλήθειαν, τὴν αἰτίαν, τὴν φύσιν, τὴν ἀνάγκην, προστιθεὶς καὶ ἑτέρας ὀνομασίας, ὡς ἐπὶ τῆς αὐτῆς οὐσίας τασσομένας καθ’ ἑτέρας καὶ ἑτέρας πιβολάς (Stobaeus, Ecl. 1.79.1–12). * I read τάξιν for τάξει (with G). But nothing much hinges on this. ⁶² Cf. DL 7.149, Stobaeus, Ecl. 1.79.17–18. ⁶³ Gellius, NA 7.2.3; Cic. Fat. 20, 27, 28; Diogenianus in Eusebius, Praep. Ev. 324.3–5; cf. also Stobaeus, Ecl. 1.79.15–16. ⁶⁴ This is so despite the fact that for the Stoics the course of the world is cyclical. ⁶⁵ E.g. Tacitus, Ann. 6.22. ⁶⁶ Tertullian, Apol 21. ⁶⁷ Plut. Stoic. Rep. 1055e and Stobaeus Ecl. 1.79.1–12, quoted above in this section.

  

239

of all events is the point that it is futile to attempt to influence one’s fate, since at any time all future events have been settled already, and are thus unchangeable. This contrasts with theories which allow for the—occasional—later change of what has been predetermined as a result of prayers.⁶⁸ The fourth factor concerns fate insofar as it links individual objects. Several Chrysippean accounts speak of fate as linking the things in the world in some orderly manner. Chrysippus named the way the things are connected an ‘interweaving’ (epiplokē).⁶⁹ This interweaving is explained by him as ‘things following upon other things and being involved with other things (from eternity)’.⁷⁰ The idea of interconnection is found also in his etymological exegesis of ‘fate’ as ‘connecting cause of the things’.⁷¹ What are the things that are linked? Plotinus, reporting Stoic doctrine, mentions the ‘interconnection of causes with each other’.⁷² Then there is a later Stoic etymological explanation of ‘fate’ as ‘chain (or series) of causes’ (heirmos aitiōn). Thus Chrysippus was probably concerned with the concatenation of all causes or all things. There is not much difference here, since the causes are all things— considered in a certain respect, namely insofar as they have effects (see above, section 1.2): fate, qua pneuma in all things, links these things through space and time; through time by way of antecedent causes, through space by way of sustaining causes and sympathy.⁷³ However, we need to exclude an interpretation which is sometimes read into the accounts from a modern perspective, viz. that we have a chain of causes and effects, in which the effect of one instance of causation is the cause of the next. This cannot be the early Stoic position, since cause and effect are ontologically distinct, one being a material thing, the other an actualized predicate. How are we to interpret the metaphors of interweaving and chain? Stoic causes are corporeal and relative (pros ti); that is, they are bodies while and insofar as they actively produce (or contribute to producing) an effect in a body. Both causes and effects thus have duration (see above, section 1.2). We can then imagine the simplified case of a chain of causes thus: a body b₁ is the cause c₁ from t₁ to t₃, producing an effect e₁ at a body b₂, from t₂ to t₃. As a result, at t₃ (perhaps a little earlier), body b₂ starts being the cause c₂ of another effect e₂ at a body b₃; c₂ may last from t₃ to t₅, e₂ from t₄ to t₅. At t₅ (or a bit earlier), b₃ starts being a cause of a further effect e₃ at a body b₄, etc. A case of a section of a network, which includes a combination of causes for one effect, would be this: a body b₁ is the cause c₁ from

⁶⁸ Nemesius, Nat. hom. 106.15–20. Mostly, however, the ancients assumed that if something is fated, it is immutable. ⁶⁹ Gellius, NA 7.2.3. Stobaeus, Ecl. 1.78.4–6; cf. also Aristocles in Eusebius, Praep. Ev. 15.14.2. ⁷⁰ Gellius, NA 7.2.3. ⁷¹ Cf. Diogenianus in Eusebius, Praep. Ev. 323.11–12; DL 7.149. ⁷² Plot. Enn. 2.1.2 236.30–1 (Henry/Schwyzer); [Plut.] Epit. 1.27.4 (Diels, Doxogr. graec. 322.11–12). ⁷³ These two dimensions of the connection of fate, time, and space are suggested by [Plut.] Fat. 574e.

240

, ,   

t₁ to t₃, of an effect e₁ from t₂ to t₃ at a body b₂. b₂ as cause c₂ from t₃ to t₅ and b₃ as cause c₃ from t₄ to t₆ are co-causes of effect e₂ at a body b₄ from t₄ to t₆. As a result, at t₆ (or a little earlier) b₄ may start being a cause of an effect e₃ at a body b₅ from t₇ to t₈, and perhaps also being a cause of an effect e₄, perhaps at b₁, etc. That is, we have a temporal concatenation of bodies insofar as they are causes. This is in line with the formulations of interweaving of things or chains of causes, i.e. of corporeal entities, as opposed to of the incorporeal effects, motions, states, etc. In one respect, Chrysippus’ statement of the ‘interweaving’ of things seems superior and better in keeping with his theory than the metaphor of a chain of causes. A chain leading through time has one link at a time, and conjures up the picture of isolated parallel ‘strings’ of causation.⁷⁴ The picture of an interweaving or concatenation of things on the other hand allows for indefinite complexity: a network rather than isolated strings, in which many instances of causation can occur at the same time, or be temporally staggered and overlap, etc. The account of fate as ‘concatenation of causes (in the plural)’, which arises from the inner‐cosmic perspective, is not to be confused with the identification of fate with Cause (in the singular) from the global perspective, where both ‘fate’ and ‘Cause’ refer to the one active principle. The difference is marked by a terminological distinction. Chrysippus always referred to the one cause which is identical with fate and the active principle by the feminine noun Aitia.⁷⁵ Aitia is identified with fate, god, and the active principle (pneuma) of the universe, which are all one.⁷⁶ This one Cause is the same as Reason. So Stobaeus: Fate is the Reason of the universe . . . and instead of Reason he uses ‘Truth’, ‘Aitia’, ‘Nature’, and ‘Necessity’ . . . (Ecl. 1.79.5–10)

And Seneca: As you know, our Stoics state that there are two in the nature of things from which everything occurs: Cause and matter . . . Cause, i.e. Reason, moulds matter and turns it wherever it wants.⁷⁷ (Ep. 65.2)

⁷⁴ Except if one assumes that at any time the whole world state counts as a link, which may be suggested by the singular noun ‘chain’, but for which there is otherwise no evidential support. ⁷⁵ Cf. Plut. Stoic. Rep. 1055e, 1056b, 1056c; Stobaeus, Ecl. 1.79.5–10. Cf. also Plot. Enn. 3.1.2.17–22. I use, as is sometimes done, ‘Aitia’ with a capital ‘A’ when it is clear from the context that the Stoic one active principle is at issue, otherwise no capitals. ⁷⁶ Cf. Marcus Aurelius 5.8; 8.27; 9.29; Sen. Ben. 4.7; SE M 9.75; Syrianus, Met. (SVF 2.308); Stobaeus, Ecl. 1.31.13–14. ⁷⁷ Dicunt, ut scis, Stoici nostri duo esse in rerum natura ex quibus omnia fiant, causam et materiam. Materia . . . Causa autem, id est ratio, materiam format et quocumque vult versat. Here ‘Stoici nostri’ refers to the early Stoics. For similar statements see also Stobaeus, Ecl. 1.132.27–133.5; DL 7.135.

  

241

‘Reason’ here is the pneumatic world reason, which pervades the universe as a whole, including all things.⁷⁸ What is the relation between the one Cause (Aitia) and the many causes (aitiai or aitia), in the context of fate? Seneca uses the phrase ‘the cause of the causes’.⁷⁹ The above evidence about the identification of Cause (Aitia), Reason (logos), and the active principle suggests that the relationship is as follows: the one Cause (Aitia) is the pneumatic world reason, which penetrates all material objects, and is responsible for their shape and changes. The Cause of any individual cause (aition) is the portion of (rational) pneuma which permeates that cause.⁸⁰ For example, in an instance of causation of bread-cutting, the pneuma in the cause (aition) knife which cuts the bread is the Cause of that knife, i.e. that portion of the world reason in the knife that makes it cut the bread.⁸¹ Individual material objects are thus causes insofar as they are pervaded by a portion of the world reason. This causal function of the world reason is the ground why it is also called ‘Cause’ and ‘fate’. This relation between the active principle, qua Reason or Cause, and the individual causes is crucial for understanding early Stoic determinism. It brings us closer to an answer to the question of how the teleological and causal (‘mechanistic’) elements in the theory combine; how the predetermination of every movement by the situation of the world prior to it, and including antecedent causes of it, cohere with the eternal, rational, divine world order. We know that the picture cannot be that of a transcendental deity who devises a plan and then realizes it in the world. There is no space for either a god or a god’s plan outside the world. Both god and god’s will or reason are part of the one material world. God, qua fate or Cause, is what makes everything a cause. It is the pneuma in every thing, by which all things are linked and through which the world progressively develops. God, or god’s will, thus works—in part—from the inside of all things. Hence we should imagine every individual cause as containing a piece of information about where it is heading. Every cause carries with it, and in itself, the relevant bit of god’s will or plan. The Stoics offer several analogies to explain how this teleological element is combined with their causal determinism. One analogy compares Nature (god, the active principle, fate) to a human being: like human beings, Nature has a soul and that soul has a ruling part. The ruling part of Nature’s soul is placed by some Stoics

⁷⁸ Cf. Origen, Cels. 6.71.5–7 (Borret); DL 7.134. ⁷⁹ QNat. 2.45. ⁸⁰ This explains the passage [Plut.] Epit. 1.12 (Diels, Doxogr. graec. 310.6–7): ‘the Stoics hold that all causes are corporeal; for they are portions of pneuma (πνεύματα)’. ⁸¹ These considerations suggest a plausible way of understanding the difficult closing sentence of Stobaeus’ report from Chrysippus’ theory of causes: ‘Αἴτια is the Reason in the cause (αἴτιον), or the Reason in respect of the cause qua cause’ (Ecl. 1.139.3–4, Αἰτίαν δ᾽εἶναι λόγον αἰτίου, ἢ λόγον τὸν περὶ τοῦ αἰτίου ὡς αἰτίου). Making use of the above example, we can say that that aspect or part of the pneuma in the knife that makes it cut the bread is the reason in respect of the knife qua cause of the cutting. For although the whole knife is the cause (αἴτιον) of the cutting, strictly speaking, it is the pneuma in the knife that is responsible for the effect.

242

, ,   

in the aether, the fiery sphere that encompasses the earth and the layer of air that surrounds it. The aether is pure, condensed rationality. From this ruling part, Nature literally stretches into everything and connects all things, just as the human soul stretches from the human ruling part through the entire human body: Chrysippus says in the first book On Providence that the world is a living being and rational and ensouled and with intellect.⁸² (DL 7.142)

And: As Chrysippus says in the fifth book of On Providence and Posidonius in the third book of On Gods, the world is administered in accordance with intellect and providence, and the intellect (nous) pervades through every part of it, as in us the soul, and more [of it] through some parts and less [of it] through others. For through some parts [of the world] it passes as cohesion (hexis), as through our bones and sinews, but through other parts [it passes] as intellect, as through our ruling part of the soul. Thus the entire world is a living being and ensouled and rational, and has as its ruling part the aether, as Antipater of Tyre says in the eighth book of On the world, and Chrysippus in the first book of On providence . . . Chrysippus again somewhat differently says it is the purer part of the aether in the world.⁸³ (DL 7.138–9)

Thus (the purer part of) the aether is to the world what the intellect in the ruling part of the human soul is to the human being. This analogy helps us understand how the world can be pervaded by divine rationality. First, the analogy clearly presupposes the Stoic continuum theory, and thereby the interconnection of all things that later Stoics called ‘sympatheia’. (We do not know whether the early Stoics used this term.) It appears then that, for the Stoics, just as I can take a step with my right foot by sending impulses from my mind, via my nervous system, to my foot, so Nature can make something happen by sending impulses to the place where it is meant to happen. And just as my foot shares in my rationality insofar as it is an extension of my mind and, once I have formed the intention to take a step forward, my right foot will move forward as directed by the impulse in it, so ⁸² Ὅτι δὲ καὶ ζῷον ὁ κόσμος καὶ λογικὸν καὶ ἔμψυχον καὶ νοερὸν καὶ Χρύσιππός φησιν ἐν πρώτῳ Περὶ προνοίας (DL 7.142). ⁸³ Τὸν δὴ κόσμον διοικεῖσθαι κατὰ νοῦν καὶ πρόνοιαν, καθά φησι Χρύσιππός τ᾽ ἐν τῷ πέμπτῳ Περὶ προνοίας καὶ Ποσειδώνιος ἐν τῷ τρίτῳ Περὶ θεῶν, εἰς ἅπαν αὐτοῦ μέρος διήκοντος τοῦ νοῦ, καθάπερ ἐφ᾽ ἡμῶν τῆς ψυχῆς: ἀλλ᾽ ἤδη δι᾽ ὧν μὲν μᾶλλον, δι᾽ ὧν δὲ ἧττον. [139] δι᾽ ὧν μὲν γὰρ ὡς ἕξις κεχώρηκεν, ὡς διὰ τῶν ὀστῶν καὶ τῶν νεύρων: δι᾽ ὧν δὲ ὡς νοῦς, ὡς διὰ τοῦ ἡγεμονικοῦ. οὕτω δὴ καὶ τὸν ὅλον κόσμον ζῷον ὄντα καὶ ἔμψυχον καὶ λογικόν, ἔχειν ἡγεμονικὸν μὲν τὸν αἰθέρα, καθά φησιν Ἀντίπατρος ὁ Τύριος ἐν τῷ ὀγδόῳ Περὶ κόσμου. Χρύσιππος δ᾽ ἐν τῷ πρώτῳ Περὶ προνοίας καὶ Ποσειδώνιος ἐν τῷ Περὶ θεῶν τὸν οὐρανόν φασι τὸ ἡγεμονικὸν τοῦ κόσμου, Κλεάνθης δὲ τὸν ἥλιον. ὁ μέντοι Χρύσιππος διαφορώτερον πάλιν τὸ καθαρώτερον τοῦ αἰθέρος ἐν ταὐτῷ (DL 7.138–9, cf. also DL 7.156, Cic. Acad. Pr. 2.126).

  

243

everything shares in Nature’s rational pneuma and is directed by it and moves and changes as directed. This reading of the analogy with the human soul finds further evidence in the Stoic view that Nature is a rational agent. First, Zeno clearly stated that the Nature of the world is not just similar to an agent, but is actually an agent: Of the nature of the world itself, which encloses and contains all things in its embrace, [Zeno] says that it is not just like an artisan but actually is an artisan, as it advises and provides for everything that is useful and advantageous . . . ⁸⁴ (Cic. Nat. D. 2.57–8)

From the fact that Zeno calls Nature an actual artisan (and also from the fact that he has just before said that artistic work is creative and productive) we can infer that he regards nature to be a creative agent,* and as such someone who plans the future and plans ahead for the future. The same Cicero text provides more detail just two sentences later: Of the nature of the world itself, which encloses and contains all things in its embrace, [Zeno] says . . . has the impulserelated movements, the desires and intentions, which the Greeks call hormai, and it employs the appropriate actions just as we do ourselves, who are moved by our minds and our senses.⁸⁵ (Cic. Nat. D. 2.57–8)

Nature is endowed with all the elements that Stoic theory of action considers part of what happens in each action: a desire—more specifically an impression of something as desirable (hormetikē phantasia)—intention or impulse (hormē), presumably assent, and action (cf. Bobzien 1998a, chapter 6.1.2).⁸⁶ Here the individual natures and the Nature of the world are juxtaposed, and features of the individual natures that are required for human action, and human actions

⁸⁴ ipsius vero mundi, qui omnia conplexu suo coercet et continet, natura non artificiosa solum sed plane artifex ab eodem Zenone dicitur, consultrix et provida utilitatum oportunitatumque omnium. * Paul Scade reminds me that further support for this might be found in the wider context of Cicero’s presentation of Stoic views in On the Nature of the Gods, which draws heavily on the Timaeus and aligns Stoic Nature with Plato’s Demiurge (as well as the world soul). For example, Nat. D. 2.142ff., on the artistry of the design of the human sensory organs, appears to be taken in parts from the Timaeus. ⁸⁵ ipsius vero mundi, qui omnia conplexu suo coercet et continet, natura . . . † omnes†† motus habet voluntarios conatusque et adpetitiones quas ὁρμάς Graeci vocant, et his consentaneas actiones sic adhibet ut nosmet ipsi qui animis movemur et sensibus. † There seems to be a lacuna here, for which see below, note 88. The expression in angled brackets is a placeholder rather than a serious suggestion for emendation. †† For this reading see below note 88. ⁸⁶ There is explicit evidence that Nature or god has some kind of perception or impression/ representation (phantasia). Thus Olympiodorus (Phd. 2.14) reports that the Stoics assumed that god was a body that acts (energein) in accordance with phantasia. Cf. also DL 7.156, which states that for the Stoics souls have perception.

244

, ,   

themselves, are by analogy attributed to Nature: desire, impulse, and action. As Nature is rational (section 1, DL 7.142–3), its action here has to be analogous to human action. And our texts are quite clear that the analogy is not between agency and something similar to agency, but rather between human agency and Nature’s agency. The following example illustrates the analogy. Just as—having some impulseprompting (hormetic) impression, say, of something I want and which is a step away—I have a desire to take a step forward with my right foot, then assent to that desire and create an impulse that leads to my action of stepping forward, so Nature has a desire, or impulse-prompting (hormetic) impression and impulse, and acts accordingly. Here Nature’s actions include every event that occurs in the world. In Nature’s case, impulse moves from the celestial aether to whichever part of Nature is going to move or change. As, according to the Stoic assumption, the world is the best possible one, Nature will only desire and have impulse(s) towards those changes that make the world (continue to be) the best possible one. (Thus the world is wise, sophos.) The Stoic comparison of Nature’s agency with human agency, as I have so far developed it, is helpful for understanding some aspects of how teleology and apparently mechanical causation come together in the explanation of a single event. However, the account so far lacks—among other things—the crucial element of the predetermination from eternity of Nature’s action. Here a second Stoic analogy is more successful. It is taken from biology. As early as Zeno, Stoics called Nature, or god, a seminal or sperm-like reason (logos spermatikos). In line with this there survives a Stoic analogy of the development of the world with the development of living things from seeds. In the same Cicero passage from On the Nature of the Gods, nested in between the previous two quotations, we read: And just as the other [i.e. individual] natures are generated, and grow and are held together by their seeds, so the Nature of the world. (Cic. Nat. D. 2.58)⁸⁷,⁸⁸ ⁸⁷ Atque ut ceterae naturae suis seminibus quaeque gignuntur augescunt continentur, sic natura mundi. ⁸⁸ Cicero appears to combine two distinct Stoic analogies into one: the world-soul-based analogy with the human soul and human agency, and the seed-like-reason-based analogy with seed. The text as it stands does not make sense. If we take it as it is, we have an analogy between, on the one hand, individual natures that are ‘generated, and grow and . . . held together by their seeds’, and, on the other hand, Nature as having ‘the impulse-related movements, the desires and intentions . . . and . . . actions just as we do ourselves’. So Nature is compared, in what appears to be one analogy, to two distinct things at the inner-worldly level: seeds and human agents. Either we have one analogon too many or one analogy too few. Since we have independent evidence for both the seed analogy and the soul analogy elsewhere, I infer it must be the second. For the remainder of this footnote, I am largely indebted to Charles Brittain. Mayor in his commentary seems to have thought the same (Mayor 1883, see later in this footnote). Charles Brittain, in agreement with my suggestion, offers that we expect those two analogies since, at the beginning of the sentence, Cicero has Balbus (the Stoic who speaks mostly in Book 2) say that the Nature of the world is not only made with craft, or according with the rules of

  

245

This second analogy, too, is between the Nature of the world as a whole and individual natures within the world, this time it seems of living beings generally. There are (at least) four points of comparison: generation, nourishment, growth, and coherence. I will draw out this analogy with the help of what we know about Stoic cosmology. Some elements are (again) inevitably conjectural. G  . At the level of natures, the individual seeds are active matter. The seeds generate or give birth to the individual natures. In a seed is assembled all the information that is required to generate a plant or animal, and it will do so provided that there is nourishment available. (A mammalian sperm and the embryo are nourished by the pregnant body, see Hierocles 1.5ff, quoted below.) The position is analogous in the case of the world seed. The active principle or aether at the end of a conflagration (ekpurōsis, a periodic complete destruction of the world through fire), and separated from passive matter, correlates to the initial stage of the seed without matter (cf. Philo, De Aeternitate Mundi 19, p. 505M for Chrysippus). The world seed is (the finest part of) this aether. In it is assembled all the information that is needed to generate the world. For the (re-)generation (palingenesis), matter (the passive principle) is available and functions as ‘nourishment’. Cicero’s text is phrased in terms of natures rather than objects. This is understandable if one assumes that ‘nature’ is a relative term, and that in some sense a piece of aether, or the whole of it, becomes a nature (of a thing, of the world) with the generation of that thing or the world, and does not exist beforehand independent of the thing/world as that nature, although it exists as the seed for that nature. G  . With the help of further nourishment, and in accordance with the information the seed carries, the (nature of the) thing grows and develops into the (nature of the) individual living thing the form of which was already fully determined in the seed. For this growth and development to be successful, in addition to nourishment being present, external impediments need

crafts (artificiosa), and as such has all its generation, growth, and sustainment through its seed(s), but also is a craftsperson (artifex) who, as such, has cognition and impulse. These two expressions thus correlate with the seed analogy and the human agent analogy respectively. In their commentaries, Pease and Reid (Pease 1958, 686, quoting Reid as quoted in Mayor 1883, 161) try to make sense of the text as is: ‘art is shown by arrangement and purpose; all parts of nature show this and are therefore artificiosa. The art is as it were stored in the seed; the universe has no seed, but . . . has its movements and feelings belonging to itself ’. This move is unsuccessful, though, since the world does develop from seed, for which see the main text and also Zeno’s expression ‘seminal fire’ (above). So I assume that the suggestion that we have two analogies holds up. What are we to do with the text? The construction atque ut . . . sic . . . sic . . . ut already intimates the presence of two analogies. Mayor (Mayor 1883, 161) suggests a full stop after ‘sic natura mundi’ and continuation with ‘omnes’. This is not plausible, since it requires us to understand ‘natura mundi’ as implicit subject expression of the second sentence, as the subject of ‘omnes habet motus’. More likely is the presence of a lacuna. The assumption of haplography of ‘natura mundi’ would be a simple remedy, but leaves us with a clumsy, entirely unCiceronian, sentence construction (so also Charles Brittain). So the lacuna would have to have had a few more words in it to give the text the required elegance. I leave this to the Latinists.

246

, ,   

to be absent (no delphinium if the slugs eat the seedling; a disfigured rose blossom if inhabited by a beetle). Analogously in the case of the world seed: Nature (the world pneuma, creative fire, aether) has as one of its functions feeding on the matter and growing and developing into the structured world (cosmos) the form of which was already fully determined in the seed of the world. However, unlike in the case of seedlings, there can be no external obstacles to the development of the world (section 2 above, where Plut. Stoic. Rep. 1056d is cited), no super-slugs or super-beetles. (Matter is entirely passive and cannot obstruct the working of the active principle.) Thus we can assume that the development of the world is determined in its entirety by the seed (aether), which initiates the beginning of the development of the world in each world cycle. C. The function of the seeds continues beyond the developmental stage. Once the development is complete, the seeds operate as cohesive causes, causes that hold the object together, that give it its properties, and—if the relevant external circumstances and triggers are present—causes of the internal development and change of an individual (see Chapter 9 for Stoic kinds of causes). In the same way, we have to understand the world seed (aether) as a coherent cause that continues to hold the world together once it has taken in all the matter and is fully grown and developed. From the Cicero passage we can gather that, at the individual level, an important point in this analogy is that the seeds of the individual things do not vanish once the individual thing is in the process of developing or is fully developed. Rather, the seeds are considered to be retained and present within the thing during its development and during its entire lifetime. This is a notion of seed that comes much closer to contemporary DNA than to the way in which an acorn is the seed of an oak, say. The Stoic seed is in the embryo or seedling from its very beginning, but it also remains still fully present during the lifetime of the individual thing. The parallel would be, rather, with the information contained within the acorn that then remains within the oak. The information is in the acorn (say), and the same information is still in the oak. Analogously, we can assume that the Nature of the world is a rational seed that carries all the information for the development of the whole world in it, and that this information is retained as a whole in the world both while it develops and when it is fully developed. Drawing on the first analogy and DL 7.138–9 (quoted above), we can conjecture that this information is retained in the celestial aether, with the relevant bits of it stretching into the individual things on earth. There is another important point. We can assume that, at the individual level, a tree seed includes the information that leads to the production of new seeds which carry the same information. It is also possible that a tree seed was considered to include some information that eventually leads to the tree’s demise. Analogously, the Nature of the world, we can assume, includes the information responsible for the demise of the world (during the conflagration) and for the production or

  

247

generation of a new world, and so on ad infinitum, all while the basic information, the ‘DNA of Nature’, remains intact throughout. If this is right, we can say that as the seed determines the life cycle of the individual natures, so Nature determines the course of the world from conflagration to conflagration. Did the Stoics really have a notion of information that is ‘stored’ in the world seed and in all the individual seeds? Is this not a case of introducing into ancient philosophy a notion that is entirely alien to them? I have used the terms ‘information’ and ‘DNA’ in order to provide the reader with an easy way into the Stoic theory of the all-pervading Nature or, physically speaking, the all-pervading active principle and chemically speaking (the purest part of) the aether. In Stoic terms, we would instead speak of divine knowledge. God, i.e. Nature, has knowledge of the past, present, and future. This systematic knowledge (epistēmē) or Truth (alētheia) is a physical correlate to the true propositions (axiômata) about the world (e.g. SE M 7.40–1). Truth is corporeal and is located in the ruling part of rational beings. Where I talk about stored information and compare it to DNA, I mean precisely the Truth that is part of God’s, i.e. Nature’s, knowledge. Just as there are certain physical imprints in the human soul that correspond to human true belief, so there is a system of physical imprints in God/Nature that is Truth or the physical manifestations of the true propositions about the world, past, present, and future. Continuum theory allows the Stoics to pack indefinitely finely structured elements into the active principle, wherever it is present. The understanding of seeds as DNA-like rather than acorn-like and of the similarities between individual seeds and the aether as seed of the world is confirmed by the late Stoic Hierocles (second century ). In particular, Hierocles’ description of the development of animals from sperm is instructive. The sperm absorbs matter from the pregnant body and develops into the embryo ‘in accordance with certain inexorable patterns’.⁸⁹ This terminology of a pattern or order (taxis) that is inexorable (aparabatos) is the same as is used in the description of fate. Related to the absence of external impediments, there is a further significant difference between the development of individual natures and the development of Nature. In the Stoic view, changes within the world need an external cause to trigger them (see Chapter 9); an agent needs a desire, or impulse-prompting (hormetic) impression, to trigger action. This is not the case for Nature and the world as a whole. There is no cause external to Nature (section 2). Appropriately, for Nature’s actions and development there is also no need for a cause that is external to Nature. The entire (physical encapsulation of the) information about the world is the seed that is Nature. According to the Stoics, this information has

⁸⁹ κατὰ [τι]νας ἀπαρα[β]άτους τάξεις (Hierocles, 1.5ff.).

248

, ,   

been there from eternity. Moreover, it is likely that Nature’s intention or intentions (hormai) have also been present from eternity. It is even possible that there is only one—very complex—action by Nature per world cycle, which would be a rational action, and thus one for which Nature carries responsibility. If this is not so, Nature’s intentions will have causes internal to the world that have causes internal to the world etc., going back to eternity. Either way, unlike human action, Nature’s action does not require an auxiliary antecedent cause (see Chapter 9). Nature is itself the principal cause of all future development and change of the world as a whole. The fact that this cause goes back eternally does not interfere with the temporal linearity of the world’s development. Related to this and to the seed analogy is a Stoic metaphor that describes eternal predetermination in terms of the ‘unrolling’ or ‘unfolding’ of fate through time.⁹⁰ This metaphor is meant to illustrate the fact that everything has been predetermined from eternity to happen in a certain sequence, and that nothing is newly or spontaneously created when it happens. There is no room in Stoic philosophy for chance events like the Epicurean swerves (for which see Chapter 6). For the Stoics, chance events are entirely epistemic phenomena, events for which we lack knowledge of their causes. We can assume that, just as the development of a tree or mammal is orderly, involving certain sequences, such as the progression from leaves to blooms (or the other way round, depending on species), so the seed of the world has information that bestows on the world an orderly sequence of changes and continuities. The Stoic analogies and metaphors leave unanswered some vexing theological questions. One is this: how is the divine rationality that is Nature translated into the perfection and orderliness of the world that displays the rational design? For an answer, we need to combine the information given in the analogies: to the seed analogy we have to add Nature’s intention from the human agency analogy, and hence an awareness of the world that is perhaps the world’s self‐awareness. Such intention-accompanying awareness will have to be thought of as analogous to that of animals.⁹¹ Moreover, as the intention-accompanying awareness is that of a rational being, it will include not only consciousness of everything that happens at the present moment, but also of everything that has happened in the past and that will happen in the future. Of particular interest here is how the Stoics imagined Nature’s intention-accompanying awareness to effect that which is still unrealized in the world. Cicero reports, though not necessarily reflecting the views of the early Stoics, that:

⁹⁰ See Diogenianus in Eusebius, Praep. Ev. 6.8.9 for Chrysippus on the etymology of the names of the three Fates. Cf. also Gellius, NA 7.2.3, quoted above. ⁹¹ We have some brief passages on self-awareness from later Stoics: Sen. Ep. 121.6–15; Hierocles 1.34–9, 51–7, 2.1–9, 4.38–53.

  

249

Someone who grasps the causes of future things necessarily grasps what the future things will be. But . . . nobody can do this except god.⁹² (Cic. Div. 1.127)

There is no textual evidence that helps with the question of how god’s awareness is related to such future things. One option is that the aether-DNA is theoretical information, and is ‘translated’ at the relevant times into divine impulseprompting (hormetic) impressions, leading to divine (assent and) intention (hormē), which is then realized in divine action. This would require a teleological mechanism that determines when individual intentions for each and everything that happens are to be triggered. Nothing suggests that the Stoics were committed to such a complex teleological picture. An alternative option is that a complex intention or network of intentions is in existence from eternity (with Nature’s assent also being given from eternity), and that whenever a particular change has occurred the relevant part of the intention ceases to exists, just as is the case with human action in Stoic theory. Divine awareness is then a necessary concomitant to divine intention, but there is no divine activity of any kind that initiates divine action during the course of the world, since all that is needed to generate an action in accordance with the divine plan is also already part of what has been within the divine aether from eternity. This picture fits decidedly better with the Stoic theory of fate, which does not allow bargaining with fate or a change of fate or exceptions to fate, and in particular with the analogy of the unfolding or unrolling of what occurs in the world over time. This picture would also be compatible with the possibility that the Stoics may have invoked their theory of the eternal recurrence of the world in cycles, so that this awareness is god’s memory of the last world cycle, as is hinted at in a later source.⁹³ We are equally in the dark about a second question, which is where divine awareness and divine perception (assuming the two to be related) reside in the world.⁹⁴ Is it that the celestial aether has sense-organ equivalents with which it perceives and has awareness of the individual objects and their development ‘from above’, as it were? Does it perhaps additionally perceive through the sense organs of living beings? Such a view may find support in Chrysippus holding that when someone takes a step, both the main chunk of the ruling part of the soul in the heart and the part of the ruling part of the soul that stretches into his foot act (cf. Sen. Ep. 113.18, Aetius, Plac. 4.21 (SVF 1.525)). Or does the aether, alternatively, perceive the world solely through the sense organs of living beings? In the latter

⁹² qui enim teneat causas rerum futurarum, idem necesse est omnia teneat, quae futura sint. quod cum nemo facere nisi deus possit. ⁹³ Cf. Nemesius, Nat. hom. 111.25–112.3. ⁹⁴ Cf. Diogenes Laertius’ report that ‘[the Stoics] called [the purest aether] also the primary god who perceiving as it were pervades through those in the air and all animals and plants, but through the earth by coherence’ (ὃ καὶ πρῶτον θεὸν λέγουσιν αἰσθητικῶς ὥσπερ κεχωρηκέναι διὰ τῶν ἐν ἀέρι καὶ διὰ τῶν ζῴων ἁπάντων καὶ φυτῶν: διὰ δὲ τῆς γῆς αὐτῆς καθ᾽ ἕξιν), DL 7.139.

250

, ,   

case, would each thing in the world be imbued just with the relevant share of that awareness and intention or with all of Nature’s awareness? I do not know. No doubt there are also other interpretations compatible with the surviving evidence of the Stoic theory of fate, which are for others to explore. The suggestions advanced here in relation to both these questions—that of the temporal origin of Nature’s intention (and assent) and that of the location of Nature’s awareness and perception—are entirely conjectural. I offer them primarily as an indicator of both the difficulties that the Stoic version of a living world soul brings with it and the many ways in which surviving elements of Stoic philosophy from other areas can help guide our view of how the Stoics might have thought about these things, if they did in fact delve into them in so much detail. * *

*

Turning back to firmer ground, it should be clear by now that early Stoic philosophy combines what we would think of as science and theology into one overarching theory. Science and theology are but two aspects of the same theory. Science is not seen in contrast to, and certainly not in a battle for dominance with, theology. Rather, science has the divine literally incorporated within its subject matter. Conversely, theology is not relegated to a mythical realm—although the Stoics have no shortage of mythical allegories. Theology is a rational, scientific, theory that has as its subject the rational order and organization of the world, the cosmos. Science fills in the details and could just as well be described as the theology of the specifics. The relationships (i) between morally responsible agents and theology and (ii) between morally responsible agents and science are also both treated in a satisfactory manner in Stoic philosophy. A Stoic scientific explanation of a human action is an explanation in terms of the interaction of individual causes that bring about the action (for details, see Chapter 9). Such an explanation identifies the human being as the main cause of their action. External antecedent auxiliary causes ‘trigger’ a chain of changes via our impulse-prompting (hormetic) impressions, assent, and intention (hormē). Since in the same circumstances a different person—that is a different main cause—might intend and act differently, it is we, qua being that main cause, who are morally responsible for our actions, and not any external antecedent causes of the action (see Chapter 7). Science is thus compatible with action and moral responsibility. More specifically, the scientific explanation does in fact provide a reason for our responsibility. It is by means of this scientific explanation that it is possible to understand (ii), that is, how the fact that Nature, or divine reason, has predetermined every event, even the smallest, is compatible with us being morally responsible for our actions. To see how this works, it is convenient to turn the Stoic analogy of rational agency upside down. We do what Nature has planned for us to do and what fate does through us, since our individual natures replicate Nature or god’s rationality. Only

  

251

insofar as we are ourselves rational creatures with perception and agency, whose actions are the result of our forming intentions by agreeing with and assenting to our desires (or by withholding such agreement), are we executing what Nature has planned for us to do (Bobzien 1998a, chapter 5). We manifest and complete Nature’s intentions or rational impulses in our actions by means of the (finest rational) aether in the ruling part of our soul. It is this aether that makes our behaviour our action (praxis), as opposed to simple activity, and that makes us rational. And this rational aether is at the same time (the finest part of) Nature. This constellation of facts impacts our relationship with Nature. First, it makes it impossible for us to transfer our responsibility for what we do to fate. Blaming fate, saying ‘Fate did it, not I’, is not a valid excuse. Second, it is equally impossible for us to be inactive and let fate act instead of us, e.g. by saying ‘Okay, there is no point in doing anything at all if fate does everything’ (the ‘Lazy Argument’, Bobzien 1998a, chapter 5). Third, accordingly, at the level of philosophical discourse it is impossible for philosophical opponents of the Stoics to argue successfully that the Stoic theory removes moral responsibility from human beings, or that Stoic theory provides an incentive for us to be lazy or to just follow our unexamined desires by default. All this is impossible since it is insofar as we are rational that we act or refrain from acting. Rationality is, thus, embodied in our assenting or refraining from assenting, in our subsequent impulse and action (or inaction). Rationality is a blessing and a burden. Once we have acquired practical rationality as a mental disposition, there is no escape from it: when we perform an action (or refrain from doing so), the buck stops with us. Human beings can do good or bad. It is the fact that our being able to do good or bad results from our rationality that makes us responsible for our actions. Unlike humans, Nature can do good but no bad. Nonetheless, in the Stoic view it is possible that Nature does good through us doing bad (just as the sage can). In that case, our doing bad is a part of the overall good, but it is not good on its own qua our action. Nature is causally responsible for our action only insofar as our souls are literally part of the divine soul, of Aitia, Cause, Providence, as part of the overall good. Since we, like Nature, are rational beings, moral responsibility for our action is connected with us alone. In sum, the Stoic theory of Nature as world soul and as a rational agent that is contrasted with the individual souls and agents within Nature helps to resolve three philosophically fundamental but puzzling relations. First, science and theology, in particular the issue of divine predetermination, can consistently be combined within one theory of the world. There is no need to choose between theological explanation and scientific explanation of the world. These are two complementary explanations of the same thing. Second, science and human responsibility can also be combined consistently in one theory of the world. A scientific explanation of our action in terms of causation (individual causes, Chapter 9) can be given. Such an explanation identifies us as the main cause of our

252

, ,   

actions. This fact is sufficient to make us morally responsible for them. Third, human agency and divine predetermination can be combined consistently in one theory of the world as well. In the Stoic view, human agency is essentially rational, and for any rational action (or refraining from action) the buck stops with the rational agent who is its main cause. This is so quite independently of the Stoic tenet that everything that happens is also a manifestation of Nature or divine agency.

9 Chrysippus’ Theory of Causes We know very little about Chrysippus’ theory of causation. Our textual evidence which names Chrysippus directly and can straightforwardly be considered as belonging to a theory of causes is this: a passage in Stobaeus (Ecl. 1.138.23–139.4) that presents Chrysippus’ basic account of causation; a distinction of causes in Cicero’s On Fate (Fat. 41–5); and an indirect reference to a distinction of causes in Plutarch’s On Stoic Self-Contradictions (Stoic. Rep. ch. 47). There is far better evidence for later theories: in Cicero’s Topics, various works by Galen, Sextus’ Outlines of Pyrrhonism Book 3 and Against the Mathematicians Book 9, and Clement’s Miscellanies Book 8.9 we find excerpts and summaries of causal theories of medical, later Stoic, and Peripatetic origin. (A trace of a later Stoic theory of causes can be found in Alexander of Aphrodisias’ On Fate 192.18–19.) The reports are mainly eclectic in character, often uncritically juxtaposing and mixing together various theories. But they have one thing in common: they treat the theories they report as finished taxonomies of causes. They present sets of technical terms that are used as names for mutually exclusive classes of causes, so that it is possible to assign any cause to precisely one class (and naturally there are no empty classes). In most cases, causes of more than one type are described as cooperating in one instance of causation. Based on Cicero’s report of a Chrysippean distinction of ‘genera’ of causes (Fat. 41) it has been the prevalent view that Chrysippus aimed at developing a complete classification of causes, like those later ones, introducing names that express the class membership of a cause.¹ The assumption is that some of the later theories were derived from Chrysippus’ theory, and efforts are then made to match Chrysippus’ ‘types of causes’ with these later theories and to extract them from those later texts. In contrast, my view is that Chrysippus neither developed a finished taxonomy of causes, nor intended to do so, and that he did not have a set of technical terms for mutually exclusive classes of causes, so that each cause can be assigned its class. This essay first appeared in K. Ierodiakonou (ed.), Topics in Stoic Philosophy (Oxford: Oxford University Press 1999), 196–242, and is reproduced here with permission. I am very grateful to Catherine Atherton and Jonathan Barnes, who each provided me with valuable comments on an earlier version of this essay. Thanks are due also to Charles Brittain, who detected several mistakes in what I had taken to be the final version of the essay, and to Katerina Ierodiakonou for drawing my attention to several ambiguities in my formulations. ¹ E.g. Long and Sedley 1987, vol. 1, 342; Duhot 1989, 172.

Determinism, Freedom, and Moral Responsibility: Essays in Ancient Philosophy. Susanne Bobzien, Oxford University Press (2021). © Susanne Bobzien. DOI: 10.1093/oso/9780198866732.003.0010

254

, ,   

Rather, Chrysippus made some conceptual distinctions. The various adjectives which he used for causes had the function of describing or explaining particular features of certain causes. Clarification, explanation, specification of particular philosophical points are his purposes—not the assignment of all causes to an already existent, exhaustive, classification, nor the construction of one; nor yet the empirical enterprise of dividing up into groups causes as they are found in the world. (Accordingly, the adjectives he uses are not mutually exclusive in their application to causes.) Chrysippus explained special features and functions of causes in ordinary language, wherever and to whatever degree this was needed in the relevant philosophical context—for example, in an argument for compatibilism. This evidence-based approach to Chrysippus’ causal theory leads to an interpretation of our main sources that differs from the received view. The following reconstruction of Chrysippus’ theory is grounded on the Stoic tenets that causes are bodies, that they are ‘relative’ (pros ti), and that all causation can ultimately be traced back to the one ‘active principle’ (archē poioun) which pervades all things, and of which every case of causation is a manifestation. The reconstruction relies primarily on the reports of Chrysippus’ own theory, secondarily on texts that uncontroversially present early Stoic doctrine. Later texts are generally adduced only to confirm certain suggestions further, not to establish points independently.

1. Definitions and Accounts Chrysippus states that a cause is that because of which; and that the cause is an existent thing and a body, and that the cause is ‘because’, and that of which it is a cause is ‘why?’. (Stobaeus, Ecl. 1.138.23–139.2)² the Stoics state that every cause is a body that becomes a cause, to a body, of something incorporeal; as the scalpel, which is a body, becomes a cause to the flesh, which is a body, of the incorporeal predicate ‘being cut’. (SE M 9.211; cf. Clement, Strom. 8.9 96.18–97.1)³ ² Χρύσιππος αἴτιον εἶναι λέγει δι᾽ὅ. καὶ τὸ μὲν αἴτιον ὂν καὶ σῶμα, * καὶ αἴτιον μὲν ὅτι, οὗ δὲ αἴτιον διά τι. * Some emendation is required. My suggestion is based on the parallel for Zeno (Stobaeus, Ecl. 1.138.5–16) and Posidonius (Stobaues, Ecl. 1.139.7–8): both contrast αἴτιον as σῶμα with οὗ δὲ αἴτιον as κατηγόρημα and both provide positive characterizations of the οὗ δὲ αἴτιον. These parallels are not born out by Wachsmuth’s emendation , which has been taken over by von Arnim, Stoicorum Veterum Fragmenta (SVF), 2.336, and Long and Sedley, vol. 2, 55A. ³ . . . Στωικοὶ μὲν πᾶν αἴτιον σῶμά ϕασι σώματι ἀσωμάτου τινὸς αἴτιον γίνεσθαι, οἷον σῶμα μὲν τὸ σμιλίον, σώματι δὲ τῇ σαρκί, ἀσωμάτου δὲ τοῦ τέμνεσθαι κατηγορήματος.

’   

255

Every instance of causation involves at least three main factors, two corporeal, one incorporeal. (For reasons of convenience, I individuate instances of causation by assuming one such instance per effect.) One corporeal is the cause, the other the object to which the effect happens. The effect—i.e. that which is caused—is immaterial and happens to the second corporeal. It is standardly determined as being a predicate (Stobaeus, Ecl. 1.138–9). (Note that the early Stoics, as far as we know, had no special term for ‘effect’. Apotelesma, which is the standard expression later, does not occur.) In cases in which different causes work together in one instance of causation the factors are multiplied. The account of the cause as ‘that because of which’ (di’ho) as well as the description of the pair of cause and effect as ‘because’/‘why?’ (if the text is not corrupt here), and as ‘cause’/’that of which it is a cause’, emphasize that cause and effect are relative to each other, and inseparable: a cause is not a particular thing, but that thing in so far as it produces its effect. In fact, for the Stoics, cause is relative in two respects: ‘They say that cause is a relative [pros ti]. For it is cause of something and to something, as the scalpel is the cause of something, viz. the cutting, and to something, viz. the flesh’ (SE M 9.207).⁴ Stobaeus’ report on Chrysippus’ concept of causation closes with the sentence ‘αἰτίαν δ᾽ εἶναι λόγον αἰτίου, ἢ λόγον τὸν περὶ τοῦ αἰτίου ὡς αἰτίου’ (Stobaeus, Ecl. 1.139.3–4). This has been variously translated as ‘But an aitia . . . is an account of the aition, or the account of the aition as aition’ and as ‘an explanation [aitia] is the statement of a cause [aition], or statement concerning the cause qua cause’.⁵ The aitia has accordingly been interpreted as being a ‘propositional item of a certain kind’,⁶ or as explanation.⁷ It is taken to be incorporeal, and thus categorically different from the aition. I believe that it is mistaken to place the Stoic aitia among the incorporeals, or making it a proposition, or translating it by ‘explanation’.⁸ The only evidence that has been given for this interpretation of the Stobaeus passage is the use of aitia, as opposed to aition, synonymously with ‘the account about the aition’ by the fourth-century  doctor Diocles of Carystus, as reported by Galen, Alim. fac. 6.455–456 K (fr. 112 Wellmann).⁹ This appears to me insufficient to establish that ⁴ Tὸ αἴτιον τοίνυν, ϕασί, τῶν πρός τι ἐστίν˙ τινὸς γάρ ἐστιν αἴτιον καὶ τινί, οἷον τὸ σμιλίον τινὸς μέν ἐστιν αἴτιον καθάπερ τῆς τομῆς, τινὶ δὲ καθάπερ τῇ σαρκί (SE M 9.207; cf. also SE PH 3.25, M 9.239 and see below in this section and section 5). It is evident from the very close parallel in SE M 9.211, quoted above, that this is Stoic. The same example is used in largely the same formulation. Whether it goes back to Chrysippus we do not know. ⁵ Frede 1980, 222, and Long and Sedley 1987, vol. 1, 333, respectively. Similarly, Duhot 1989, 146: ‘Αἰτία est le concept de cause, ou, si on prefère, le concept de cause en tant que cause.’ ⁶ Frede 1980, 222. ⁷ Long and Sedley 1987, vol. 1, 333. I find the translation of λόγος as ‘statement (of a cause)’ unconvincing in itself. Λόγος meaning ‘statement’ is exceptionally rare in Stoic philosophy. ⁸ As Frede (1980, 224–5) points out, the early Stoics are not interested in explanation; their concern is the attribution of responsibility. ⁹ Frede 1980, 222.

256

, ,   

for Chrysippus aitia was a propositional item or a kind of causal explanation. (We may also wonder why we hear nowhere else of this Stoic concept of explanation (aitia). Should it not have its place somewhere in Stoic logic or epistemology? Would we not even expect it to be central there? But there is not a trace of such a concept of explanation/aitia anywhere in our sources on Stoic epistemology.) An alternative is, instead, to look at Chrysippus’ own use of the word aitia, and how the Stoics used abstract feminine nouns in contrast with substantivized neuter singular adjectives. For, if Chrysippus was sensitive to the grammatical distinction aitia/aition, we should expect him to stick to his own philosophical interpretation of such distinctions. His use of aitia can be found in another passage in Stobaeus, Ecl. 1.79.5–12: Fate is the reason of the universe, or the reason of the things in the universe governed by providence . . . And instead of ‘reason’ he uses ‘truth’, ‘aitia’, ‘nature’, and ‘necessity’, and adds other terms, which apply to the same substance from different perspectives.¹⁰ (my emphasis)

Here Chrysippus uses aitia as coextensive with ‘reason’ (logos). And, for the Stoics in general, Seneca reports: As you know, our Stoics state that there are two [principles] in the nature of things from which everything occurs: cause and matter. Matter . . . But cause, i.e. reason, moulds matter and turns it wherever it wants. (Ep. 65.2, my emphasis)¹¹

Aitia, and equally alētheia and anagkē, are hence in this context identified with fate, reason, and the active principle (archē poioun) of the universe, which are all one for Chrysippus, and which physically are pneuma (by which the Stoics mean a breath-like corporeal substance that permeates all material objects) (e.g. Stobaeus, Ecl. 1.79.1–2). In particular, aitia is used synonymously for logos in Stobaeus, and causa for ratio in Seneca, just as in the passage under discussion (Stobaeus, Ecl. ¹⁰ Εἱμαρμένη ἐστὶν ὁ τοῦ κόσμου λόγος ἢ λόγος τῶν ἐν τῷ κόσμῳ προνοίᾳ διοικουμένων . . . Μεταλαμβάνει δ᾽ἀντὶ τοῦ λόγου τὴν ἀλήθειαν, τὴν αἰτίαν, τὴν ϕύσιν, τὴν ἀνάγκην, προστιθεὶς καὶ ἑτέρας ὀνομασίας, ὡς ἐπὶ τῆς αὐτῆς οὐσίας τασσομένας καθ᾽ ἑτέρας καὶ ἑτέρας ἐπιβολάς. Compare in Plutarch’s On Stoic Self-Contradictions, as Chrysippean names for fate: ‘the greatest cause’ (ἡ μεγίστη αἰτία (Stoic. Rep. 1055e)), ‘the cause of all things’ (πάντων αἰτία (Stoic. Rep. 1056b)), ‘invincible and unpreventable and unchangeable cause’ (αἰτίαν ἀνίκητον καὶ ἀκώλυτον καὶ ἄτρεπτον (Stoic. Rep. 1056c)). ¹¹ Dicunt, ut scis, Stoici nostri duo esse in rerum natura, ex quibus omnia fiant, causam et materiam. Materia . . . Causa autem, id est ratio, materiam format et quocumque vult versat. Cf. also Marcus Aurelius 8.27 ‘divine cause’ (θεία αἰτία); 9.29 ‘the cause of the all’ (ἡ τῶν ὅλων αἰτία); 5.8 . . . οὕτως ἐκ πάντων τῶν αἰτίων ἡ εἱμαρμένη τοιαύτη αἰτία συμπληροῦται; Sen. Nat. quaest. 2.45 causa causarum. See also Plot. Enn. 3.1.2.17–22, κυριωτάτη αἰτία for fate as all-pervading, greatest cause of all things, reporting Stoic doctrine; and Stobaeus, Ecl. 1.31.13–14, for Chrysippus: Δία δὲ αὐτὸν (i.e. god) λέγουσιν, ὅτι πάντων ἐστὶν αἴτιος καὶ δι᾽ αὐτὸν πάντα.

’   

257

1.139.3–4). However, here reason is corporeal and consists of pneuma, being the world reason, which pervades the universe as a whole, including all things in it.¹² Now in our initial quote (Stobaeus, Ecl. 1.139.3–4) aitia and logos were not the world cause and the world reason, but they were aitia and logos of an individual cause (aition)—i.e. an individual material object. But the relation between world cause and individual aitia is simple. Every object contains a portion of the world pneuma or active principle. And these individual portions of pneuma, too, according to their various functions, were referred to with abstract feminine nouns, in correspondence with the names of the world pneuma. For example, the feminine noun alētheia (‘truth’) is used also of an individual portion of pneuma in the human mind (e.g. SE PH 2.81). Equally, I believe, an individual portion of the world pneuma in an object, in so far as it is responsible for an effect, is referred to by the feminine noun aitia.¹³ This understanding of aitia also helps to make sense of the doxographical report that ‘the Stoics hold that all causes are corporeal; for they are pneumata’.¹⁴ This sentence refers to the individual portions of pneuma in those objects that function as cause. The relation between individual aitiai and the world cause may also be expressed in Marcus Aurelius, when he says ‘and just as the cosmos is made up into such a body from all bodies, so fate is made up into such a cause from all causes’ (Marc. Aur. 5.8.1).¹⁵ All this suggests that Chrysippus’ above-quoted statement about aitia (Stobaeus, Ecl. 1.139.3–4) should be understood as follows: the logos mentioned in it is the ‘pneumatic’ world reason. This—corporeal—logos penetrates all matter and is responsible for the qualities and movements of all material objects. The aitia of an aition is thus a portion of (rational) pneuma which permeates the aition. For instance, the pneuma in the aition knife which causes this bread being cut is the aitia of that knife (qua aition of this bread being cut). More precisely, it is a part or aspect of the pneuma which permeates the matter (hylē) of the aition, and together with the matter makes up the aition. Thus we can understand the sentence in Stobaeus as ‘aitia is the reason in the cause, or the reason in respect of the cause as cause’ (‘cause as cause’ serves as a reminder that causes are relative). In the example, that aspect or part of the pneuma in the knife that makes it cut this bread (as opposed to that aspect of the knife’s pneuma which makes its handle green) is the reason in respect of the knife qua cause of this bread being cut. Although the whole ¹² Cf. e.g., for Zeno, DL 7.134, for the Stoics in general, Origen, Cels. 6.71.5–7 (Borret), Stobaeus, Ecl. 1.79.1–2, together with 5–12. ¹³ In contrast, the Stoics used the substantivized neuter singular τὸ ἀληθές and τὸ ἀναγκαῖον to refer to incorporeal, propositional items, SE PH 2.81, DL 7.79. And, although for Chrysippus τὸ αἴτιον is clearly not incorporeal, or a propositional item, the fact that the feminine nouns ἀλήθεια and ἀνάγκη were reserved for corporeal entities points strongly in the direction that αἰτία was not an incorporeal, propositional item either, but, like ἀλήθεια and ἀνάγκη, corporeal, just as on the cosmic level. ¹⁴ [Plut.] Epit. 1.12; (Diels, Doxogr. graec., 310.6–7), Οἱ Στωικοὶ πάντα τὰ αἴτια σωματικά˙ πνεύματα γάρ. As in most later sources, the terminological distinction αἴτιον/αἰτία seems lost in this passage. ¹⁵ . . . καὶ ὥσπερ ἐκ πάντων τῶν σωμάτων ὁ κόσμος τοιοῦτον σῶμα συμπληροῦται, οὕτως ἐκ πάντων τῶν αἰτίων ἡ εἱμαρμένη τοιαύτη αἰτία συμπληροῦται.

258

, ,   

knife is the cause of the cutting, strictly speaking, it is the pneuma in the knife that is responsible for the effect: strictly speaking, the pneuma in the knife is the only part that is active—the rest of the knife is active only through it. The relation between active principle (aitia, logos) and cause (aition) is basic for Chrysippus’ causal theory. Nothing is uncaused for Chrysippus.¹⁶ Every effect, and that is every change or motion (kinēsis) as well as every qualitative state (schesis) of a thing, requires a cause and can be traced back to some pneuma, and is caused by the active principle.¹⁷ Every cause, of whatever kind, contains this active power. This is the main reason why for the Stoics all causes are active causes (poioun, energoun), and Aristotelian or Peripatetic formal, material, and final causes, as well as mere necessary conditions, do not qualify. For example Seneca writes, ‘The Stoics hold that there is one cause, viz. that which does something’ (Ep. 65.4),¹⁸ and Sextus reports of all dogmatists, including the Stoics, that a cause is that ‘because of which, it being active, the effect comes about’ (SE PH 3.14).¹⁹ In an instance of causation, only those factors that actively contribute to the effect are causes proper. There is thus for the Stoics a difference between the cause or causes of an effect, and all the other factors involved, including those things that are necessary conditions for the effect, but do not actively contribute to it. How does this square with Chrysippus’ above-quoted account of cause as that ‘because of which’ (di’ho)? It has been suggested that, when Chrysippus characterizes causes in this way, he allows for a very general notion of cause, which includes non-active causes, and which differs from his narrower concept, which is restricted to active causes.²⁰ I consider this unlikely, for the following reasons: the phrase di’ho itself can be understood in a wider sense (anything that helps explain why something happened or is the case) and in a narrower one (asking for that which is in some stronger sense responsible for or contributing to the effect). Now, as Stobaeus presents the di’ho as a definition for both Zeno and Chrysippus, and the Stoic concept of cause is that of an active cause, and furthermore in the same passage Zeno gives a very narrow interpretation of di’ho indeed,²¹ I assume that by di’ho Chrysippus, too, meant to cover his philosophical, ‘narrow’, notion of cause, according to which a cause actively contributes to the effect. But, of course, we do not know for sure what Chrysippus intended by the phrase di’ho in his account of cause. In any event, in the following I assume that, ¹⁶ See e.g. Plut. Stoic. Rep. 1045c τὸ γὰρ ἀναίτιον ὅλως ἀνύπαρκτον εἶναι. ¹⁷ Plut. Stoic. Rep. 1050c–d, 1056c, Comm. Not. 1076e. For the Stoic κίνησις/σχέσις distinction, see further Stobaeus, Ecl. 1.166.24–167.14, 2.73.1, 2.82.11–17, 2.95.6–8; DL 7.104; Origen, Orat. 2.368; Plot. Enn. 3.1.7; Cic. Fin. 3.33; Simplicius, Cat. 212–13. Cf. on the topic Bobzien 1998a, sections 1.1.2 and 1.1.3. ¹⁸ Stoici placet unam causam esse, id quod facit. ¹⁹ Δι᾽ ὅ ἐνεργοῦν γίνεται τὸ ἀποτέλεσμα. See also Frede 1980, 225–6, on the Stoic conception of causes as active. ²⁰ Frede 1980, 220. ²¹ Zeno held that ‘being alive’ occurs because of the soul, since it is impossible for someone who possesses a soul not to be alive (cf. Stobaeus, Ecl. 1.138.5–16). The soul is pneuma and hence active. The same holds for Zeno’s other examples.

’   

259

when Chrysippus deals with causation in philosophical contexts, he has active causes alone in mind, and that, in particular, mere necessary conditions for an effect do not count as causes. In which way, then, are Chrysippus’ causes active? One important point here is that ‘being active’ does not mean ‘actively bringing about some change’. For Chrysippus, the activity of causes is equally required for changes (kinēseis) and qualitative states (scheseis). The qualitative states of an object depend on the particular tension in its pneuma. Causes of a change actively contribute to something’s changing. Causes of qualitative states actively contribute to something’s keeping up or sustaining certain qualities. To understand this better, we may consider simple everyday observations such as: fire is needed to heat water up as well as to keep it hot, and muscles are needed to lift up a heavy object as well as to keep it lifted up. In that sense, very roughly, one can imagine that Chrysippus thought that the sustenance of qualitative states required some steady activity (to keep up a certain tension in the object); though in his view that which is active is not external to the object that is in the qualitative state, but is the active principle in the object. Keeping something up in a certain state is thus doing something or being active just as making something change is. The distinction between causes of qualitative states and causes of change is fundamental to Chrysippus’ physics. Unlike qualitative states, every change or motion requires an antecedent cause (Cic. Fat. 21, 40, 41, 43). There is no need to assume that Chrysippus introduced antecedent causes as a specific Stoic type of cause, or that he had a technical term for them. Presumably, one of the words he used was proēgoumenon.²² Nonetheless, since for the Stoics all causes are corporeal, some elucidation is needed for what it means for a body to be an antecedent cause. For in most cases of causation this body will exist before, during, and after the effect obtains; ‘antecedent’ thus cannot refer simply to the time at which the body exists. We need to take into account that, as cause, the body actively contributes to its effect, and that cause is a relative (pros ti). The body is the antecedent cause only in so far as it actively contributes to the effect. I thus suggest that, as a minimal condition, c is an antecedent cause of a change e, if the period of time at which c is active in contributing to e precedes, at least in part, the period of time at which the effect obtains.

2. Cicero, De Fato 41–5 Our main evidence for Chrysippus’ distinction of causes deals with causes of change only. The most discussed text on Stoic causation is Cicero, De Fato 41–5. ²² This is confirmed by [Plut.] Fat. 574e, which seems to be ‘orthodox’ Stoic, and is close to Chrysippus. The passage gives the principle μηδὲν ἀναιτίως γίγνεσθαι ἀλλὰ κατὰ προηγουμένας αἰτίας as backing up the Stoic principle that everything is fated. So too Hankinson 1987, 90.

260

, ,   

In it a distinction between two types of causes leads up to Chrysippus’ defence of his compatibilism by way of the notorious cylinder analogy. This passage is fraught with difficulties, which have led to numerous distinct interpretations. It will be seen that most of the problems disappear once one realizes that the relation between Chrysippus’ distinction of causes and the cylinder analogy is different from that which has traditionally been assumed. In De Fato 41 Cicero reports that Chrysippus distinguished between proximate and auxiliary causes on the one hand, and perfect and principal causes on the other. In De Fato 42–3 Cicero reports that Chrysippus maintained that, in instances of causation of certain types, two determining factors cooperate in bringing about the effect. One of these is an antecedent cause. The other is characterized as the main determining factor and is internal to the object to which the effect happens. The standard interpretation of the relation between these two passages is (i) that the antecedent determining factor is a proximate and auxiliary cause and the internal main determining factor is a perfect and principal cause; (ii) that Chrysippus introduced the distinction between proximate and auxiliary causes and perfect and principal causes in order to make this point; and (iii), inferred from (i), that proximate and auxiliary causes and perfect and principal causes cooperate in one instance of causation, and that in De Fato 42–3 only proximate and auxiliary causes are considered as candidates for antecedent causes.²³ It seems to me that the relation between the two Cicero passages is rather as follows: in any one instance of causation either a proximate and auxiliary cause or a perfect and principal cause is involved, but not both—that is, they are not cooperating causes but alternative possibilities for any instance of causation. The difference between the two types of causes is such that it allows for both to be in principle conceived of as antecedent causes. For Chrysippus’ compatibilist argument and the cylinder example, this means that the antecedent determining factor is a proximate and auxiliary cause (just as the standard interpretation assumes), but the internal main determining factor is not a perfect and principal cause.²⁴ The implications of this interpretation for our understanding of Chrysippus’ compatibilism are not world-shattering. Rather, its advantages are that it helps to dissolve some of the difficulties the standard interpretation encounters and fits better with our remaining evidence for Stoic causation. Generally, it sheds new

²³ So e.g. Botros 1985, 283–5; Dihle 1984, 164–5; 1982, 103; Dillon 1977, 86; Dobbin 1991, 119; Donini 1974/5; 1988, 31; Duhot 1989, 174–5, 179–80; Forschner 1981, 96–7; Frede 1980, 234–6; Hankinson 1987, 85; Inwood 1985, 46; Kleywegt 1973, 342–3; Long 1968, 340; 1970, 261–2; 1971, 182; 1976, 84; Long and Sedley 1987, vol. 1, 393; Meyer 1994, 76; Pohlenz 1959, 105; Reesor 1965, 285–97; Sambursky 1959, 62; Sedley 1993, 322–3; Sharples 1981, 81–97; 1986, 272–3; 1991, 199–200; Sorabji 1980a, 273; Steinmetz 1994, 611; v. Straaten 1977, 510–12; Talanga 1986, 132–7; Theiler 1946, 62. ²⁴ This view is shared by Görler 1987, 254–74; Schröder 1990, 5–26; Ioppolo 1994, 4492–545.

’   

261

light on some aspects of Chrysippus’ theory of causation and makes it possible to re-evaluate the relation between early Stoic and later ancient theories of causes. I shall now, in a somewhat painstaking way, work my way through Cicero, De Fato 41–5, recording what the sections say about Chrysippus’ distinction between types of causes, and showing that this passage does not contain any compelling evidence for the standard interpretation. (Impatient readers may find summaries of my results on pages 269–71.) Cicero’s De Fato 39–45 form a unit, with the tail end missing. The subject at issue is the question whether the motions of the soul (in particular impulse and assent) are necessitated by external factors, and hence not in our power.²⁵ De Fato 39–40.1 and 44–5 form a kind of exegetical framework in which Cicero speaks, comparing Chrysippus’ position with other philosophical views on that topic. De Fato 41–3 is a report from Chrysippus. Cicero partly quotes (in Latin translation) from a Chrysippean text, presumably his second book On Fate,²⁶ and partly summarizes and paraphrases Chrysippus’ argument. De Fato 41–42.1 presents Chrysippus’ reply to an argument which is presented in 40 and which challenges his theory of fate. This argument is designed to prove that the Stoic principle that everything is fated is untenable, by showing that it implies that our impulse and assent and consequently our actions are not in our power and thus that praise and blame for our actions are not just. The conclusion drawn is that, since the last statement is false, the initial assumption, that everything happens by fate, must be given up.²⁷ The main point of the argument of De Fato 40 is that moral accountability is destroyed by fate. Accordingly, Chrysippus’ reply combines a formal refutation of the argument (Fat. 41–42.1) with an exposition of some points from Stoic psychology which serves to determine where exactly in the complex process that leads to an action human responsibility for the action takes its origin (Fat. 42.2–43). The passage begins like this: (1) But Chrysippus, since he both rejected necessity and insisted that nothing happens without preceding causes, distinguished kinds of causes, in order that he both escape necessity and retain fate. (Fat. 41)²⁸

Chrysippus made a distinction between types of causes. The reason for this distinction is that he both rejects necessity and wants to maintain his claim that ²⁵ The general topic is the compatibility of Stoic determinism and theory of fate with moral responsibility. In this chapter I disregard the problems this Cicero passage poses for the interpretation of Chrysippus’ compatibilism. I deal with the question in considerable detail in Bobzien 1998a, ch. 6. ²⁶ Cf. Bobzien 1998a, section 6.1.1.1 for details. ²⁷ Cf. Ioppolo 1988, 397–424 and Bobzien 1998a, section 6.2.2 for two reconstructions of this argument. ²⁸ (1) Chrysippus autem, cum et necessitatem inprobaret et nihil vellet sine praepositis causis evenire, causarum genera distinguit, ut et necessitatem effugiat et retineat fatum.

262

, ,   

(i) nothing happens without an antecedent cause.²⁹ (i) is a corollary of Chrysippus’ principle that everything happens in accordance with fate. This is why Cicero says that Chrysippus insists on (i) in part in order to ‘retain fate’. The phrases ‘to reject necessity’ and ‘to escape necessity’ in passage (1) can be understood in two ways: Chrysippus claimed either that nothing that happens is necessary, or that (ii) not everything that happens is necessary. The context (Fat. 40) makes it clear that Chrysippus intended to deny that, if something is fated, it is necessary; in other words, he purports to show that fate does not entail necessity. In addition, we know that Chrysippus admitted that some things are necessary.³⁰ Hence I assume that in passage (1) Cicero refers to principle (ii). The reason why Chrysippus needs this principle is that he wants to retain the claim that assent, impulse, and action are in our power; and only non-necessary things can be in our power—as was universally agreed in antiquity. Thus Chrysippus’ distinction of causes is meant to ensure or clarify the compatibility of the principles (i) ‘nothing happens without an antecedent cause’ and (ii) ‘not everything that happens is necessary’. The next sentence introduces the distinction: (2) For, he says, some causes are perfect and principal, others auxiliary and proximate. (Fat. 41)³¹

Each type of cause is referred to by two adjectives. The wording of (2) alone does not permit us to decide whether the disjunction of types of causes is exhaustive. But it is clearly exclusive: one and the same thing cannot be both a perfect and principal and an auxiliary and proximate cause of the same effect. The text continues: (3) Because of this, when we say that everything happens through fate by way of antecedent causes, we do not want this understood as ‘by perfect and principal causes’, but ‘by auxiliary and proximate causes’. (Fat. 41)³²

Here Chrysippus applies his distinction to antecedent causes. The proponents of the argument in De Fato 40, it is implied, treated all antecedent causes as if they were perfect and principal causes—i.e. including the antecedent causes of impulse and assent. Chrysippus points out that in the phrase ‘everything happens through fate by way of antecedent causes’ the Stoics (at least in the context of impulse and

²⁹ This principle, more often occurring in the form ‘every event has an antecedent cause’, was used by Chrysippus in the argument given in Cic. Fat. 21 and criticized by Carneades in Cic. Fat. 23, 24, 31, 33. Cf. [Plut.] Fat. 574e (see n. 22) for a Greek formulation of the principle in the same context. ³⁰ Cf. Bobzien 1993, 63–84; Cic. Fat. 45; cf. also Augustine, De civ. D. 5.10. ³¹ (2) Causarum enim, inquit, aliae sunt perfectae et principales, aliae adiuvantes et proximae. ³² (3) Quam ob rem, cum dicimus omnia fato fieri causis antecedentibus, non hoc intellegi volumus: causis perfectis et principalibus, sed: causis adiuvantibus et proximis.

’   

263

assent) understand antecedent causes as auxiliary and proximate causes only. This clarification of the Stoic position is followed by Chrysippus’ formal reply to the argument of the opponents. (4) (a) Thus he retorts in the following way to the argument which I presented a little while ago: (b) if everything happens through fate, it follows indeed that everything happens by preceding causes, but not by perfect and principal [preceding] causes, but by auxiliary and proximate [preceding] causes. (c) If these are not in our power, it does not follow that impulse also is not in our power. (d) But this would follow, if we said that everything happens by perfect and principal [preceding] causes, so that, since these causes are not in our power, neither would impulse be in our power. (Fat. 41)³³

Again, the whole passage discusses antecedent causes only. (My addition of ‘preceding’ in the translation only makes explicit what is clearly implied.) The opponents’ mistake is that they understand the Stoic statement ‘everything happens by antecedent causes’ as ‘everything happens by perfect and principal antecedent causes’. And a perfect and principal cause would not only itself not be in our power, but its effect, as being completely determined and necessitated by it, would not be in our power either. The next sentence summarizes what Chrysippus thinks he has achieved by his counter-argument, linking up the distinction between causes with the concept of necessity that figures in the introductory sentence of De Fato 41, and referring back to the opponents’ argument in De Fato 40: (5) (a) Therefore, against those who introduce fate in such a way that they add necessity, the above argument will be valid; (b) but against those who will not claim that the antecedent causes are perfect and principal, it will not be valid. (Fat. 42.1)³⁴

³³ (4) (a) Itaque illi rationi, quam paulo ante conclusi, sic occurrit: (b) si omnia fato fiant, sequi illud quidem, ut omnia causis fiant antepositis, verum non principalibus causis et perfectis, sed adiuvantibus et proximis. (c) Quae si ipsae non sunt in nostra potestate, non sequitur, ut ne adpetitus quidem sit in nostra potestate. (d) At hoc sequeretur, si omnia perfectis et principalibus causis fieri diceremus, ut cum eae causae non essent in nostra potestate, ne ille quidem esset in nostra potestate. (4) (b) repeats the first premise of the opponents’ argument from Fat. 40 (si omnia . . . antepositis)— which they in turn seem to have taken over from the Stoics—and explicates it by specifying that the type of antecedent causes in the consequent are proximate causes (verum . . . proximis). (4) (c) conveys that it is a consequence of this understanding of the antecedent causes that the third premise of the opponents’ argument, si causa adpetitus non est sita in nobis, ne ipse quidem adpeditus est in nostra potestate, is false. ³⁴ (5) (a) Quam ob rem, qui ita fatum introducunt ut necessitatem adiungant, in eos valebit illa conclusio; (b) qui autem causas antecedentes non dicent perfectas neque principales, in eos nihil valebit.

264

, ,   

Again, it is antecedent causes that are considered as possible candidates for being perfect and principal, and this time explicitly so. We can infer from passage (5)— and equally from (3)—that, if antecedent causes were perfect and principal, they would necessitate their effect, but when they are auxiliary and proximate only, they do not. And, since Chrysippus takes them to be auxiliary and proximate only (at least in the relevant cases of impulse and assent), the argument from De Fato 40 either does not apply to his theory—namely, if it is assumed that all antecedent causes are perfect and principal, as the opponents wrongly assume—or the argument is not sound—namely, if it is assumed that there are antecedent causes that are auxiliary and proximate only, as Chrysippus actually assumes. Passages (3), (4) (d), and (5) thus allow us to establish a further important point about the perfect and principal causes: these passages all imply that Chrysippus takes his opponents to assume that all antecedent causes are perfect and principal. Moreover, the text suggests that the opponents themselves neither made—or made use of—any distinction between causes (e.g. a distinction they picked up from the Stoic position they criticize), nor called the antecedent causes ‘perfect and principal’, nor reflected expressly on their being perfect and principal. Rather, Chrysippus makes explicit what he regards as an implicit assumption of his critics, and then, in making his distinction, contrasts this conception of antecedent causes with that of antecedent causes that are auxiliary and proximate only. If this is right, it follows that the concept of the causes which Cicero labels ‘perfect and principal’ was not a subtle, highly technical concept from Stoic physics that makes sense only in the context of Stoic philosophy. Rather, whatever expression(s) Chrysippus used for such causes, it must have been possible to use them for causes which non-Stoic philosophers like Chrysippus’ opponents³⁵ believed to be involved in ordinary, standard cases of causation. To recapitulate briefly, from De Fato 41–42.1 we obtain the following information about Chrysippus’ distinction of causes: • the distinction is applied exclusively to antecedent causes in this context; • perfect and principal antecedent causes necessitate their effect, auxiliary and proximate antecedent causes do not; • the concept of perfect and principal causes is such that it cannot be a highly specialized Stoic technical concept, but must be such that non-Stoics such as Chrysippus’ addressees would readily employ it for ordinary cases of causation. After the refutation of the opponents’ argument in De Fato 41–42.1, Chrysippus sets out to make us understand how he thinks it is in the agent’s power to assent to

³⁵ Or, if these are fictitious, at least those non-Stoics whom Chrysippus addresses.

’   

265

impressions—that is, how it is possible that acts of assent (and impulses) happen by antecedent causes without being necessitated by them.³⁶ For this purpose Chrysippus first employs his distinction of causes to the case of assent: (6) (a) For as to the fact that assents are said to happen by means of preceding causes, Chrysippus believes that he can easily explain how this works. (b) For, even though an assent cannot occur unless set in motion by an impression, nonetheless, since the assent has this impression as proximate cause and not as principal cause, it has the explanation, as Chrysippus holds, which I stated earlier: (c) it is not the case that assent can happen without being prompted by some force from outside—for it is necessary that an assent be set in motion by an impression . . . (Fat. 42.2)³⁷

In (6) (c) we are not given the full reason or explanation (ratio) Cicero talks about in (6) (b). The full reason includes, of course, the fact that the antecedent causes need not necessitate their effect—as had indeed been stated earlier, in De Fato 41–42.1. This point is only made by way of analogy in De Fato 43 (passage (7) below). Still, in (6) we can see how Chrysippus made use of his distinction of causes in the case of assent. For my purposes it is of paramount importance to see that the distinction is here used in exactly the same way it was used in De Fato 41–42.1: it is applied to antecedent or preceding causes only. The impression is identified as antecedent and proximate cause of the assenting, and is introduced as a necessary, but not sufficient, condition of it. Thus assent is one of the things that, although they have an antecedent cause, are not necessitated by it. There is one new element in Chrysippus’ explanation: assent has to be prompted by some force from outside.³⁸ The rest of Chrysippus’ exposition is given in the form of an analogy. Cicero writes: ³⁶ Fat. 40 announces the topic to be assent, Fat. 41 mentions impulse only, Fat. 42–3 talks exclusively about assent. This does not mean that these passages discuss different topics. Rather, human impulse is sometimes treated by the Stoics as a kind of assent—namely, assent to an impulse-prompting (ὁρμητική) impression. Cf. Chapter 7, pp. 200–02 (previously published as Bobzien 1997, 76–7); Frede 1993, 58–60. ³⁷ (6) (a) Quod enim dicantur adsensiones fieri causis antepositis, id quale sit, facile a se explicari putat. (b) Nam quamquam adsensio non possit fieri nisi commota viso, tamen, cum id visum proximam causam habeat, non principalem, hanc habet rationem, ut Chrysippus vult, quam dudum diximus, (c) non ut illa quidem fieri possit ulla vi extrinsecus excitata—necesse est enim adsensionem viso commoveri . . . ³⁸ There is another difficulty here: according to Stoic psychology, the impression itself is not external to the mind (ἡγεμονικόν), which is the place where assent occurs. Assent and impression are not spatially distinct, but are produced by different capacities of the mind. It is true, though, that the impressions are externally induced (most straightforwardly in the case of perceptual impressions) in that they require external objects in order to come into existence. Later sources such as Cicero and Plutarch often carelessly confound the internal impression (ϕαντασία) and the external object (ὑποκείμενον κινοῦν ἡμᾶς; ϕανταστὸν; τὸ ποιοῦν τὴν ϕαντασίαν) (cf. [Plut.] Epit. 4.12.1–3 (Diels, Doxogr. graec. 401–2); SE M 7.241; τὸ ϕαινόμενον, Plut. Stoic. Rep. 1057b).

266

, ,   

(7) (a) . . . but Chrysippus turns to his cylinder and cone, which cannot start moving without being pushed. However, when this has happened, he believes that from then on the cylinder rolls and the cone spins by their own nature. (Fat. 42.3)³⁹ (b) As thus, he states, the person who shoved the cylinder gave it the beginning of its motion, but did not give it its roll-ability, so likewise, an impression, when encountered, will imprint and so to speak stamp its form on the mind, but assent [to it] will be in our power; and, just as was said in the case of the cylinder, being pushed from outside, for the rest it will move by its own power and nature. (Fat. 43.1)⁴⁰

The example is used as an explanatory analogy: in (6) it had been announced that it is meant to explain the function of the proximate antecedent cause in the case of acts of assent. A succession of physical events involving movements of perceptible everyday objects is employed in order to make comprehensible the empirically inaccessible, non-observable, mental processes that take place in the mind. The instances of causation on the level of perceptible objects have the rolling of the cylinder and the spinning of the cone as their effects. At the level of the mind, the effect is the assenting of the person. At either level, the instances of causation involve two cooperating determining factors. The first factors in each case—the person who pushes the cylinder and cone, and the impression—are external or externally induced, and can safely be assumed to be proximate causes; for the impression had been identified as proximate cause of the assent in De Fato 42.2. The second factors are described each time as the nature of the object that actually moves—i.e. that in which the effect happens: the cylinder and the person or, more precisely, the person’s mind. (For the Stoics there is a difference between a body and its nature. Both are corporeal, but the first consists of matter and form or active principle, the latter is restricted to that form or active principle. This nature of an object is pneuma and hence corporeal itself. Thus, the second determining factor is partially identical with the body in which the effect takes place, i.e. the cylinder and the person.) Is the second determining factor a cause? That is, is the volubilitas a cause of the rolling, and the nature of the person a cause of the assenting? It is never called a cause in Cicero.⁴¹ But it satisfies the conditions for something’s being a cause as set ³⁹ (7) (a) . . . sed revertitur ad cylindrum et ad turbinem suum, quae moveri incipere nisi pulsa non possunt. Id autem cum accidit, suapte natura, quod superest, et cylindrum volvi et versari turbinem putat. ⁴⁰ (7) (b) Ut igitur, inquit, qui protrusit cylindrum, dedit ei principium motionis, volubilitatem autem non dedit, sic visum obiectum inprimet illud quidem et quasi signabit in animo suam speciem, sed adsensio nostra erit in potestate, eaque, quemadmodum in cylindro dictum est, extrinsecus pulsa, quod reliquum est, suapte vi et natura movebitur. ⁴¹ Nor is it called a cause in the parallel passage in Gellius, NA 7.2.11.

’   

267

out in Chrysippus’ account in Stobaeus (see above, section 1): it is a body (pneuma), and it is a ‘that because of which’, as the Latin ablatives (suapte natura (Fat. 42), suapte vi et natura (Fat. 43)) validate. Furthermore, it is the decisive factor, which is responsible for whether a particular effect ensues, and, in the case of assent, it is the reason why the person can be held morally accountable. Hence it must be a cause. (I shall consider the issue in which way it is active in section 5.) The question that needs to be addressed next is how the perfect and principal causes are related to the analogy. The standard interpretation (see section 2) takes it for granted that the second determining factor is a perfect and principal cause of the effect and that it produces the effect in cooperation with the external auxiliary and proximate cause, which is a necessary condition for the effect. The perfect and principal cause is assumed to exist simultaneously with the effect, and to be internal to the object to which the effect happens.⁴² So the distinction between perfect and principal causes and auxiliary and proximate causes is regarded as corresponding to the distinction internal/external. Now, it is no doubt correct that in the cases at issue there is a cooperation of an external and an internal causal factor which are somehow together responsible for the effect. But it is not stated anywhere that the internal factor is a perfect and principal cause. As said above, in the whole of passage (7) perfect and principal causes are not mentioned once. A fortiori, they are not identified with the internal factor.⁴³ This identification of the internal factor with the perfect and principal cause can—if at all—be obtained by inference only. But there is certainly nothing in the passage that compels one to infer this—or even suggests this in any way. I shall come back to the analogy. But let me first present the rest of the Cicero passage. The last part of Cicero’s report from Chrysippus comes rather abruptly, and why Chrysippus says what he says in this place is far from clear. (The section makes perhaps most sense if read as a very condensed final summary of Chrysippus’ whole argumentation in De Fato 41–3: since the opponents had attacked the principle that everything is fated (Fat. 40), its validity has ultimately to be defended, and this connection is drawn here.) The passage reads: (8) If, then, anything were brought about without an antecedent cause, it would be false that everything happens through fate; if, however, it is plausible that everything that happens has a cause preceding it, what can one put forward for

⁴² Cf. n. 23 for proponents of this view. ⁴³ The same holds for the parallel passage Gellius, NA 7.2.11, where we also find the contrast of external and internal determining factors. Although the distinction of causes is central to the argument reported by Cicero, Gellius does not report it. If the distinction was in that context primarily between two types of antecedent causes, its absence from Gellius poses no major problem. In that case Gellius— or Chrysippus—just disregarded the distinction, since the analogy made it clear on its own that the antecedent causes were not necessitating.

268

, ,   

not conceding that everything happens through fate? One only has to understand the distinction and difference amongst causes. (Fat. 43.2)⁴⁴

It is not the details of this passage that concern us here, but rather how it relates to Chrysippus’ distinction of causes. Note that the only kinds of causes mentioned explicitly are again antecedent causes, and it is their relation to the principle that everything is fated that is at issue. The distinction and difference amongst causes must be that introduced by Chrysippus in De Fato 41—i.e. that between perfect and principal causes and auxiliary and proximate causes. The sentence is neutral as to whether these causes are cooperating or alternatives. (The focus on antecedent causes, and the stress on difference rather than on cooperation, more naturally suggests a distinction between causes that are alternatives.) With De Fato 43 ends Cicero’s report from Chrysippus. In De Fato 44–5 Cicero reverts to his framing story, which he left off in De Fato 39. These surrounding passages do not quote Chrysippus. In De Fato 44–5 Cicero argues that Chrysippus and his libertarian opponents have factually the same view, and differ only in terminology. This passage is of relevance since it still talks about Chrysippus’ theory of causation. Both De Fato 44 and De Fato 45 discuss two types of antecedent causes; I quote the relevant portions: (9) As this is how these things are expounded by Chrysippus, if those who deny that assents happen through fate . . . admit that impressions precede, and nonetheless hold that assents do not happen through fate, since that proximate and cohesive cause does not bring about the assent, see whether they do not say the same thing. For while Chrysippus concedes that the proximate and cohesive cause of the assent is placed in the impression, he neither concedes that this cause is necessitating for the assenting, nor will he concede that, if everything happens through fate, everything must happen by antecedent and necessitating causes. (Fat. 44)⁴⁵

In this passage two types of causes are contrasted. First, there are proximate and cohesive (continens) causes. This type of cause is identical with the proximate and auxiliary cause: it is referred to as ‘that’ (illa) proximate and cohesive cause— which is a back reference to the proximate and auxiliary cause of De Fato 41–2. Moreover, Chrysippus is said to hold that the impression is the proximate and ⁴⁴ (8) Quod si aliqua res efficeretur sine causa antecedente, falsum esset omnia fato fieri; sin omnibus, quaecumque fiunt, verisimile est causam antecedere, quid adferri poterit, cur non omnia fato fieri fatendum sit? modo intellegatur, quae sit causarum distinctio ac dissimilitudo. ⁴⁵ (9) Haec cum ita sint a Chrysippo explicata, si illi, qui negant adsensiones fato fieri . . . concedunt anteire visa, nec tamen fato fieri adsensiones quod proxima illa et continens causa non moveat adsensionem, vide, ne idem dicant. Neque enim Chrysippus, concedens adsensionis proximam et continens causam esse in viso positam neque eam causam esse ad adsentiendum necessarium, concedet ut, si omnia fato fiant, omnia causis fiant antecedentibus et necessariis.

’   

269

cohesive cause of assent. Second, and contrasted with the first kind, are the necessitating causes.⁴⁶ The distinction between proximate and cohesive causes and necessitating causes is used of antecedent causes, and thus the causes are treated as alternative, not as cooperating. Chrysippus’ claim that the impression is the proximate and cohesive, but not a necessitating, cause of assent exactly mirrors his claim in De Fato 42.2 (passage (6)) that the impression is the proximate and auxiliary, but not the perfect and principal, cause of assent. This general parallelism, in tandem with the identity of auxiliary and proximate causes and proximate and cohesive causes, and the fact that De Fato 41–2 implied that perfect and principal causes, if antecedent, necessitated their effect (see above), suggest very strongly that the necessitating causes in De Fato 44 are the same as the perfect and principal causes from De Fato 41–2 (and that the latter are hence not cooperating). De Fato 45 presents a distinction of antecedent causes in terms of what is in our power: (10) Generally, then, there is the following distinction: in some things one can truly say that, since these causes preceded, it is not in our power to prevent from occurring those things of which they were the causes; but in other things, although causes preceded, it is still in our power that that thing should occur differently; this distinction is approved of by both parties . . . (Fat. 45)⁴⁷

There are antecedent causes whose effects (once the causes are in play) are not in our power, since the causes necessitate their effect; and there are antecedent causes where it is in our power that things turn out differently. These latter must correspond to the proximate and auxiliary or cohesive causes from De Fato 41–2 and 44 respectively. Here is a brief summary of the results of the analysis of Cicero, De Fato 41–5: • Fat. 41–42.1 introduces the distinction between causes and applies it to antecedent causes, in order to refute the argument of Fat. 40. • Fat. 42.2 applies the distinction of the two types of antecedent causes to the case of assent.

⁴⁶ See Cic. Top. 60–2 for a related view of necessary and non-necessary causes. In Top. 61 we have an example of a necessitating cause that is external: At cum in Aiacis navim crispisulcans igneum fulmen iniectum est, inflammatur navis necessario, following causa necessaria in Top. 60. Cicero, for one, does not thus consider the necessitating causes as cooperating with, but as an alternative to, nonnecessitating ones. Cf. Sharples 1995, 247–71, for a detailed discussion of Cicero’s conception of cause. ⁴⁷ (10) Omninoque, cum haec sit distinctio, ut quibusdam in rebus vere dici possit, cum hae causae antegressae sint, non esse in nostra potestate, quin illa eveniant, quorum causae fuerint; quibusdam autem in rebus causis antegressis in nostra tamen esse potestate, ut illud aliter eveniat: hanc distinctionem utrique adprobant . . .

270

, ,   

• Fat. 42.2–43.1 introduces the cylinder analogy, and two cooperating causal factors, one internal, one external. The first factor can be identified with the auxiliary and proximate cause. The second causal factor is the nature of the thing in which the effect takes place. Perfect and principal causes are not mentioned. • Fat. 43.2 talks about fate and antecedent causes, and mentions the distinction of causes. • Fat. 44 contains a distinction between proximate and cohesive antecedent causes and necessitating antecedent causes which corresponds to that between auxiliary and proximate antecedent causes and perfect and principal antecedent causes from Fat. 41–2. • Fat. 45 presents a distinction between non-necessitating antecedent causes and necessitating antecedent causes. The first seem to be auxiliary and proximate causes. Thus De Fato 41–42.2, 44, 45 discuss alternative causes, De Fato 41–42.2, 43.2, 44, 45 talk expressly only about antecedent causes. No section talks about a cooperation of causes in one instance of causation, nor of a distinction between causes in which one is an antecedent cause, the other is not. Given that this is the situation of the sources, why is it so consistently assumed (and generally not even argued for) that the perfect and principal causes are the second determining factor in the cylinder and cone analogy? A reply I have encountered more than once is: ‘this is obvious’. Before I move on to our evidence in other texts, I shall say why this is not obvious. In De Fato 41.2–42.1 Chrysippus aims at refuting an argument designed to challenge the Stoic principle that every event is fated. His central point is that an antecedent cause need not necessitate its effect—having in mind in particular the case of impulse and assent. In order to show this, he introduces a distinction between causes of which one type, if antecedent, would necessitate its effect (the perfect and principal cause), the other does not (the auxiliary and proximate cause). In De Fato 42.2–43 Chrysippus aims at explaining how those things which have a non-necessitating antecedent cause (in particular, assents) are brought about, and why we can be held responsible for them. For this purpose he makes use of a second distinction between determining factors in causation: cooperating external and internal factors. The external factor is an auxiliary and proximate antecedent cause. The internal factor is not an antecedent cause. Now it is certainly not obvious that the sort of cause which the antecedent cause of assent is not, even though Chrysippus’ opponents assumed it to be such (i.e. the perfect and principal cause), is the same sort of cause which forms the second determining factor in the instance of causation where the (auxiliary and proximate) antecedent non-necessary cause is the first determining factor. Chrysippus makes two distinct points. For each point he draws on a distinction between

’   

271

determining factors. In the first, they are alternatives; in the second, they cooperate. Why should the alternative factor of the first point, which does not apply to assent, be identical with the cooperating factor in the second point?

3. Plutarch, On Stoic Self-Contradictions 1055f–1056d The second text in which we find a Chrysippean distinction of causes is Plutarch’s On Stoic Self-Contradictions. In chapter 47, two types of causes are distinguished: the self-sufficient (autotelēs) cause and the procatarctic (prokatarktikos) cause. In this chapter Plutarch’s goal is to demonstrate that Chrysippus’ theory of fate is incompatible with his theory of impression and assent. His argumentation is somewhat convoluted, but the main structure of the relevant section (Stoic. Rep. 1055f–1056d) can by and large be sifted out.⁴⁸ Plutarch’s general procedure is this: as the basis of the ‘Stoic self-contradiction’ in question he uses a Chrysippean argument in which Chrysippus intends to prove that the impressions are not self-sufficient causes of human assent, and that the main responsibility lies with the assenting person, because of their moral character.⁴⁹ (Chrysippus uses as an example the false impressions which, in certain special circumstances, are caused in ordinary people by the sages.) Drawing on this argument, Plutarch develops a dilemma for Chrysippus. This dilemma is grounded in the assumption that for Chrysippus fate, being a cause—namely, the cause of all things—is either a self-sufficient or a procatarctic cause of human assenting. It runs like this: If, on the one hand, fate is a self-sufficient cause, it will not be the cause of all things: for, according to Chrysippus’ argument, human assents do not have a self-sufficient antecedent cause, and hence fate would not be the cause of assent (Stoic. Rep. 1056a–b, p. 55.8–18 (Teubner)). Or, supposing it were the selfsufficient cause of all things, then it would follow that nothing is in our power (Stoic. Rep. 1056c–d, p. 56.7–9; 12–15). If, on the other hand, fate is a procatarctic cause, neither does it determine everything (Stoic. Rep. 1056b–c, p. 55.18–56.2), nor is it invincible and unpreventable (Stoic. Rep. 1056c; 1056d p. 56.2–7; 10–14; 15–16). Either way, Chrysippus contradicts himself: elsewhere he holds that fate is the cause of all things, that there are things in our power, and that fate determines everything and is invincible and unpreventable. We have good reason to assume that Plutarch has taken over the twofold distinction of causes from his Chrysippean source. First, the expression ‘self-

⁴⁸ For a detailed analysis of the structure of this passage and its relation to Chrysippus’ theory of fate, see Bobzien 1998a, sections 6.1.1.3, 6.3.4, and 6.4.2. ⁴⁹ Cf. the reference to the person’s badness and weakness in the parallel passage in Plut. Stoic. Rep. 1057a–b.

272

, ,   

sufficient cause’ (autoteles aition) occurs in the Chrysippean argument reported by Plutarch and should be definitely Chrysippus’. Then, in the passage under discussion, the term autotelēs occurs six times and the term prokatarktikos four times. They are the only terms used for causes and are treated as a pair. This suggests both that Plutarch understood them as at least semi-technical terms and that he took them from his source—especially as it is Plutarch’s distinctive habit in the On Stoic Self-Contradictions constantly to weave in Stoic terms from his source. Moreover, Plutarch’s entire ‘proof ’ of Chrysippus’ inconsistency in Stoic. Rep. 1055f–1056d is built on this distinction between causes: the validity of his argument depends on the assumption that fate, when functioning as antecedent cause, is either self-sufficient or procatarctic. In particular, Plutarch’s argument has weight against Chrysippus only if Chrysippus agreed that the two types of causes form an exhaustive disjunction—at least in the context, as antecedent causes. For this Chrysippus must have had, if not the terms, at least the concepts of both kinds of causes. What information does the text provide about the two types of causes? The expression autoteles, which I translated as ‘self-sufficient’, means ‘self-sufficient’ or ‘self-complete’. Attributed to a cause, this must mean: sufficient in itself to bring about the effect. And, given that Chrysippus chose the term, its ordinary meaning should still be at least partly preserved in its more technical use.⁵⁰ Hence, we can assume that, according to Chrysippus, a self-sufficient cause is a cause that is by itself sufficient to produce its effect. This is confirmed by the way in which Clement gives an account of the term ‘self-sufficient’ used of a cause: ‘The cohesive cause is synonymously also called “self-sufficient”, since it is self-sufficiently by itself productive of its effect’ (Clement, Strom. 8.9.95.31–96.2, cf. 101.19–20).⁵¹ Equally, in Sextus’ discussion of causation in Against the Mathematicians Book 9 the adverb autotelōs is used with respect to causes, with the meaning ‘bringing about the effect self-sufficiently’: Furthermore, if there is such a thing as a cause, it is the cause of something either self-sufficiently and making use only of its own power, or it needs in addition the help of the affected matter, so that the effect is thought of as in accordance with a combination of both. (SE M 9.236–7, cf. 242)⁵² ⁵⁰ This is perhaps confirmed by the fact that instead of the adjective αὐτοτελής, at one point, he employs the adverb αὐτοτελῶς. This may indicate that the terms had not yet become rigid, and that the common meaning of the word was still partly retained. ⁵¹ συνεκτικὰ δὲ ἅπερ συνονύμως καὶ αὐτοτελῆ καλεῖται, ἐπειδήπερ αὐτάρκως δι᾽αὑτῶν ποιητικά ἐστι τοῦ ἀποτελέσματος. For Clement’s identification of αὐτοτελής with συνεκτικός, see section 6. ⁵² Kαὶ μὴν εἰ ἔστι τι αἴτιον, ἤτοι αὐτοτελῶς καὶ ἰδίᾳ μόνον προσχρώμενον δυνάμει τινός ἐστιν αἴτιον, ἢ συνεργοῦ πρὸς τοῦτο δεῖται τῆς πασχούσης ὕλης, ὥστε τὸ ἀποτέλεσμα κατὰ κοινὴν ἀμϕοτέρων νοεῖσθαι σύνοδον.

’   

273

The contrast here is clearly that between a cause bringing about the effect wholly by itself and the effect requiring a combination of causal or determining factors. Thus the very name of the self-sufficient cause, as well as the accounts of it in Clement and Sextus, rule out the possibility that, in order for it to bring about the effect, the cooperation of another cause is required. It is more difficult to establish the meaning of prokatarktikos. It may have been coined by Chrysippus as a technical term. If we take it that usually the prefix prokata- means something like ‘beforehand’, the whole term might translate (i) as ‘that which commences [intr.] beforehand’, i.e. before the effect occurs or before a second causal factor comes in;⁵³ or (ii) as ‘that which initiates [tr.] beforehand’, e.g. initiates the effect before a second causal factor comes in.⁵⁴ Alternatively, (iii) if we assume that ‘procatarctic’ is understood as ‘before the catarctic’, the meaning would be something like ‘that which precedes that which is the origin of the effect’.⁵⁵ All three interpretations make sense in the present context, and I do not see how one can decide between them. We cannot resort to the later medical and philosophical uses of the term for a decision between them, because it is quite clear that the later senses differ from the one Chrysippus has in mind.⁵⁶ But we find more information in the Plutarch passage itself: in Stoic. Rep. 1056c Plutarch provides an account of the procatarctic cause. (It is unlikely that Plutarch developed this account himself, since his argumentation is based on it, and for Chrysippus to be affected by the argument, he must have accepted this account. I hence assume that Plutarch took the account from his Chrysippean source.) He writes: ‘the procatarctic cause is weaker than the self-sufficient cause, and it falls short when it is controlled by other [causes], which stand in the way’ (Stoic. Rep. 1056c).⁵⁷ This account proceeds by delimitation, distinguishing procatarctic causes from self-sufficient ones, which further confirms that the two kinds of causes formed part of one and the same theory. The characterization of the procatarctic cause as weaker than the other implies that no cause can be both a self-sufficient and procatarctic cause of the same effect. The ‘and’ (kai) in the account may be epexegetic, so that the second clause serves to explicate the weakness of the procatarctic cause: it is weaker in that the effect does not come about if the procatarctic cause is dominated by other things, presumably other causes, that obstruct it. (In Chrysippus’ case of an impression as procatarctic

⁵³ Clement, Strom. 8.9.101.17 states that the procatarctic cause ceases before the effect does. ⁵⁴ Clement, Strom. 8.9.95.28–9 gives an account of ‘procatarctic’ as ‘that which first contributes a starting point for something to happen’ (τὰ πρώτως ἀϕορμὴν παρεχόμενα εἰς τὸ γίγνεσθαί τι). ⁵⁵ This was suggested by Frede 1980, 243 n. 6; similarly, Rieth 1933, 147. ⁵⁶ Often the procatarctic and the antecedent (προηγούμενα) causes are considered as different types that cooperate in one instance of causation. Cf. [Galen], Def. med. 19.392, Galen, Caus. cont. 2 (LS 55F), Caus. puls. 9.2–3; see also Hankinson 1987, 86–9, Frede 1980, 241. ⁵⁷ . . . τὸ μὲν προκαταρκτικὸν αἴτιον ἀσθενέστερόν ἐστι τοῦ αὐτοτελοῦς καὶ οὐκ ἐξικνεῖται κρατούμενο ὑπ’ ἂλλων ἐνισταμένων, (or ἐξανισταμένων) . . .

274

, ,   

cause, the things that might obstruct it include the nature and moral disposition of the person who has the impression.)⁵⁸ We can infer from the quoted account and its context that the characterization given does not hold for the self-sufficient cause. Hence, a self-sufficient cause never falls short, which should mean it always brings about its particular effect, and no other forces can prevent it from bringing about its effect. This harmonizes well with the information concerning its self-sufficiency that was discussed above. It has become clear that the two Chrysippean types of causes in Plutarch are alternatives for, and not cooperative in, one instance of causation. The structure of Plutarch’s argument implies this: fate is supposed first to be a self-sufficient cause, then to be a procatarctic cause, and both times it would be an antecedent cause. Cooperation is not considered—there is no hint of it in the entire passage. On the contrary, both the name and the characterization of the self-sufficient cause preclude this. Being self-sufficient to produce the effect, it does not admit of another causal factor that works together with it, and a fortiori not of one that is a necessary condition for the effect.

4. Cicero versus Plutarch Next I want to argue that the Chrysippean conceptual distinction between causes which we find in Plutarch is the same as that announced in Cicero, De Fato 41, and in particular that the sort of cause Cicero labels ‘perfect and principal’ is the same as Plutarch’s self-sufficient cause. This has been accepted by many,⁵⁹ but is not self-evident. Consider first that the general context in which the distinction of causes is made is the same in both cases: in Cicero, De Fato 41, we learn that Chrysippus introduced that distinction in the context of the debate about fate and moral responsibility in order to show that impressions do not necessitate human assents. The topic of chapter 47 of On Stoic Self-Contradictions is fate, assent, and that which is in our power, and Plutarch presents Chrysippus’ argument with the introductory sentence that ‘he [Chrysippus] wants to prove that the impression is not a self-sufficient cause of assent’ (Stoic. Rep. 1055f).* This is exactly the thesis which Chrysippus wants to demonstrate in Cicero, and which is illustrated and explained by the cylinder analogy. Moreover, the types of causes are assigned roughly the same function in both texts. One is the type of cause that externally induced impressions (or the external objects) are in the case of assents; it cooperates with a second causal factor, which is in the examples the assenting person’s ⁵⁸ Cf. again the parallel passage in Stoic. Rep. 1057a–b. ⁵⁹ So among others Rieth 1933; Theiler 1946; Pohlenz 1959; Frede 1980; Duhot 1989. * τὴν γὰρ ϕαντασίαν βουλόμενος οὐκ οὖσαν αὐτοτελῆ τῆς συγκαταθήσεως αἰτίαν ἀποδεικνύειν.

’   

275

individual moral nature.⁶⁰ The other is a kind of cause which impressions are not in the case of assents, and which, if they were of this kind, would prevent moral responsibility for the assents lying with the person who assents. We should, therefore, expect the distinction of causes to be the same in both authors. It would be surprising indeed if in the same context Chrysippus introduced two different distinctions, both for the same purpose, and in both of which the causes have the same function. A very strong reason for the identification of Plutarch’s self-sufficient cause with Cicero’s perfect cause is that ‘perfectus’ would be a natural translation of autoteles. The identification finds additional confirmation in a passage in Origen in which the expression autotelēs aitia is used in a context very much like that in Cicero and Gellius (Orig. Princ. 3.1.4). The passage draws heavily on Stoic philosophy, and deals with the issue of moral responsibility, emphasizing the point that assenting to evildoing is not fully determined by external factors. This is illustrated with the example of adultery. The claim is that the woman is not the autoteles cause of the man’s indiscretion; rather, it is implied, the man’s assenting makes him himself responsible. It is rejected that the woman is a self-sufficient cause, just as the impression or the external object was rejected as a perfect and principal cause in Cicero (and a self-sufficient cause in Plutarch). It may also be worth mentioning that Clement, after having introduced the procatarctic cause as that which first produces the starting point for something to happen (see n. 54), adds an example of the same family: ‘to the licentious, beauty is the procatarctic cause of erotic love. In them it produces amorous inclinations, but it does not do so by necessity’ (Clement, Strom. 8.9.95.29–31). The procatarctic cause thus pairs well with the self-sufficient one—as alternatives, not cooperating—in examples like those of Cicero and Gellius. The denial of necessitation by the first causal factor in the Clement passage corresponds to Cicero, De Fato 41–2 and 44. Taking these various pieces of evidence together, I conclude that the distinctions of causes in Cicero and Plutarch are the same, and in particular that the Chrysippean concept of an autoteles cause as it occurs in Plutarch is the same as the one dealt with in Cicero under the name ‘perfect and principal’. But autoteles means ‘self-sufficient’, and we established from the Plutarch passage that Chrysippus understood it in this way. In line with this, the Plutarch passage implied that a self-sufficient cause brings about its effect always and by itself—a characteristic that was confirmed by accounts in Clement and Sextus. It follows that Cicero’s perfect and principal cause, too, is self-sufficient and produces its effect without the cooperation of any other cause. But that means that the standard view—i.e. the view that perfect and principal causes cooperate with ⁶⁰ It is more obvious in Gellius, NA 7.2.7–14 than in Cicero that moral nature is at issue in the cylinder analogy; in Cicero this can be established from the context, in particular Fat. 40.

276

, ,   

auxiliary and proximate causes in one instance of causation—is mistaken. It assumes that the perfect and principal causes of events are not self-sufficient, but require an auxiliary and proximate cause to be effective. And this is exactly what is precluded, if—as has been shown—a perfect and principal cause is selfsufficient. Hence, in addition to the lack of any positive evidence for the standard interpretation, here we have a strong argument against it. This fact has not gone unnoticed. It has led to a motley variety of ad hoc explanations, designed to explain away the unwelcome incongruity—none of them in my view successful.⁶¹

5. The Relations between the Various Properties of Causes There are, however, various factors which either have led scholars to adopt the standard interpretation or have been understood as confirming it, and which in any event complicate matters of interpretation. These centre on several of the terms for causes in Cicero and Plutarch, and on the relation between cohesive (sunektikos) and self-sufficient causes, following their identification in Clement. In any case, both issues need sorting out if we are to paint a comprehensive picture of Chrysippus’ causal theory. First, we encounter the puzzling fact that Plutarch uses one term for each type of cause, whereas Cicero gives two, and, moreover, that, with the exception of the above-discussed pair autoteles/perfectus, they do not correspond very well. This raises the questions: what were the Greek terms which Cicero, or his source, translates, and are we to infer that Plutarch left two terms out, or rather that Cicero added two? Much has been written about this, and in fact any of the following situations could lie behind the discrepancy: • Chrysippus distinguished two types of causes, and refers throughout to each of them with a pair of names. • One of the Latin words is a translation of Chrysippus’ term for the cause, the other translates a Greek term used by Chrysippus to explain his concepts. • The Latin words each translate Greek terms Chrysippus used to explain his concepts. • Cicero tried to translate by using two words for one. • One of Cicero’s terms is a translation, the other added by him (or a source) in order to explicate Chrysippus’ terminology to contemporary readers. • Cicero did not attempt to translate the Greek words strictly, preserving their general meaning, but chose expressions that he thought would properly describe the function of the causes in Chrysippus’ theory. ⁶¹ E.g. Theiler 1946, 74 [62] n. 122; Long 1971, 196 n. 32; Frede 1980, 236, 239; Sorabji 1980a, 260–1; Long and Sedley 1987, vol. 1, 341; Wolff 1988, 538; Duhot 1989, 214. See Sharples 1991, 200, for an overview of some of them.

’   

277

The last two possibilities gain their plausibility from the fact that Cicero—like philosophers of his time generally—has at his disposal a large number of different concepts, classifications, and names of causes and would naturally seek to make Chrysippus’ terminology comprehensible to those (like himself) who are acquainted with contemporary causal theories.⁶² The range of possible explanations of Cicero’s terms makes it clear that, in the absence of additional evidence, we have to rely on conjecture for their Greek origin.⁶³ I shall neither attempt to decide between the possibilities nor make a definite suggestion for the underlying Greek words. Instead I shall show that Cicero’s choice of terms, whatever its origin, in conjunction with our additional evidence on Stoic causal theory, does not support the standard interpretation, and that instead it points to a straightforward, simple, and consistent understanding of Chrysippus’ theory of causes which has several advantages over the various versions of the standard interpretation. There is, first, Cicero’s term ‘principalis’, which is paired with ‘perfectus’. It also occurs in De Fato 9, a passage in which Cicero does not report Stoic doctrine but speaks in propria persona. There the word is used in a way very similar to its use in De Fato 41–2: influencing factors such as the climate, which are admitted to be antecedent causes of some of our character traits, are ruled out as principal causes of our individual actions. Thus, as in De Fato 41–2, ‘principalis’ is used negatively only. Nothing suggests that a principal cause is the second factor in a cooperation of causes. On the contrary, De Fato 7–9 suggests that the climate is seen by Cicero as a principal and antecedent cause of some of our general character traits. And this may reflect Chrysippus’ view.⁶⁴ The phrase ‘principalis causa’ occurs also in the context of Stoic fate and modalities in Boethius: Opposite to this view [i.e. the Peripatetic one that there are things that both can and cannot happen] is the one that states that everything happens through fate, which is taught by the Stoics. For what happens through fate occurs by way of principal causes; but if this is so, that which does not happen cannot be changed. (Boethius, Int. II 197 Meiser)⁶⁵

This is unlikely to be an adequate presentation of early Stoic theory. But it makes clear that principal causes were seen either to necessitate their effects, or at least fully to determine them. This squares well with what we learn about perfect and

⁶² Cf. e.g. Top. 59; see also later authors such as Galen, Sextus, Clement (texts cited above). ⁶³ Sharples 1991, 199–201 gives a fairly comprehensive overview of suggestions that have been made in this context. ⁶⁴ This has also been pointed out by Ioppolo 1994, 4515; I disagree, however, with her assumption that Fat. 9 presents Chrysippus’ argumentation. ⁶⁵ cui sententiae contraria est illa quae dicit fato omnia fieri cuius Stoici auctores sunt. quod enim fato fit ex principalibus causis evenit, sed si ita est, hoc quod non fiat non potest permutari.

278

, ,   

principal causes in Cicero (e.g. their being described as ‘causae necessariae’ in De Fato 44) and about self-sufficient causes elsewhere. ‘Principalis’ has been repeatedly suggested to be a translation of kurios, although there are no Stoic kuria aitia or kuriōs aitia which we know of. I am inclined to think that ‘principal cause’ was not a technical term for Chrysippus, nor presumably for Cicero, and that it rather simply conveyed the idea that, in an instance of causation, something is the cause to which the main responsibility for the effect is to be attached. But this is, of course, speculative. Still, in this case, there could be principal causes that are not self-sufficient, and hence although a selfsufficient (or perfect) and principal cause brings about the effect all by itself, and does not admit of another cause as a necessary condition, a principal cause that is not self-sufficient would be one that cooperates with another cause.⁶⁶ On this assumption, the second determining factor in the cylinder analogy could be a principal cause, though not a self-sufficient cause.⁶⁷ And it is even conceivable that the first factor in a cooperation is a principal cause, and the nature of the object is only helping to bring about the effect.⁶⁸ ‘Principalis’ has also been suggested to be a translation of sunektikos (cohesive),⁶⁹ mainly on the grounds that in Clement’s exposition of causal theory in his Stromata we learn in two places that the cohesive causes were also synonymously called selfsufficient (Strom. 8.9. 95.31–96.1 and 101.19–20). I do not believe that Cicero could have rendered the Greek word sunektikos with ‘principalis’. There are any number of better Latin words to express sunexis (cohesion), and the assumption that Cicero picked out ‘principalis’ is too far-fetched. More serious is the frequently propounded view that by ‘principalis’ Cicero refers to Chrysippus’, or generally Stoic, cohesive causes.⁷⁰ This opens up the general question of the relation of the Stoic cohesive causes to the perfect and principal causes and the cylinder analogy. The main philosophical reasons why it has been assumed that Cicero’s perfect and principal cause is the Stoic cohesive cause, are: (i) that it has been taken for granted that the perfect and principal cause forms the second determining factor in the causal transaction which is initiated by the proximate and auxiliary cause; and (ii) that the second determining factor has been held to be a cohesive cause. This view has been backed up with the textual point mentioned above (iii) that Clement reports that the cohesive cause is ⁶⁶ In Top. 59 Cicero makes a similar distinction between efficient causes (causae efficientes): ‘for there are some causes which plainly produce the effect without anything helping them, and others which need to be helped’ (sunt enim aliae causae quae plane efficiant nulla re adiuvante, aliae quae adiuvari velint). Here we have efficient causes which only work with, and also those which work without, help. The former could include those of the type of the second determining factor in the cylinder example; the latter could correspond to the self-sufficient ones. ⁶⁷ So also Ioppolo 1994, 4514–15. ⁶⁸ Clement, Strom. 8.9.96.3–5 calls the pupil’s nature a helping cause of the pupil’s learning, though this is most certainly not Stoic. ⁶⁹ So e.g. Duhot 1989, 170–1. ⁷⁰ E.g. Frede 1980, 242–4, Long and Sedley 1987, vol. 1, 341, Wolff 1988, 537, Duhot 1989, 170–1.

’   

279

synonymously called self-sufficient, and that in one case this is followed by an example of cooperation of procatarctic, cohesive, and helping (sunergon) cause (Strom. 8.9.96.2–5). However, none of these points stands up to scrutiny. Point (i) has been dealt with above. Points (ii) and (iii) require some discussion of the Stoic concept of the cohesive cause and of its development. Originally, in early Stoicism, something would be described as a ‘cohesive cause’ if it caused a thing’s being in a certain state, having certain essential properties, literally holding that thing (qua being that thing) together.⁷¹ This illustrates the fact that the Stoics required efficient causes not only of changes or movements (kinēseis) but also of states (scheseis).⁷² A cohesive cause is the cause of a thing’s qualities (poiotēs), as opposed to its being merely ‘being qualified’ (poion). For the Stoics, such states involve the presence in the object of a particular tension of the pneuma, and this kind of tension is made up by a special sort of movement (tonikē kinēsis). But this tensional movement does not count as a change proper (kinēsis) of the object and thus does not require an antecedent cause. When calling the pneuma in an object ‘cohesive’ the Stoics thus express a particular function of that pneuma, namely the function of holding the object, qua being that object, together. The term sunektikon aition, which is standard in later taxonomies of causes, is not attested for Chrysippus, or for the early Stoics in general. Rather, the cohesive function of pneuma is talked about in various forms of the verb sunechein.⁷³ On the assumption that Chrysippus had no worked-out classification of causes with technical terms for the different kinds, this is not surprising. Galen provides us with evidence that later non-Stoics used the expression ‘cohesive cause’ in a different way from the early Stoics: However, it is above all necessary to remember how we said we were speaking of the cohesive cause—not in its strict sense, but using the appellative loosely. For no one before the Stoics either spoke of or admitted the existence of the cohesive cause in the strict sense. And what have before our time been spoken of as cohesive have been causes of something’s coming about, not of existence. (Galen, Syn. puls. 9.458.8–14, trans. Long and Sedley, modified)⁷⁴

⁷¹ E.g. Galen, Caus. cont. 1.1–5 (7 K); Plen. 3 (7.535 K); Alex. Mixt. 223–4; Plut. Stoic. Rep. 1053f. ⁷² See section 1; for the Stoic κίνησις/σχέσις distinction, see in particular n. 17. ⁷³ Cf. the only passage (I have found) attributed to Chrysippus, Plut. Stoic. Rep. 1053f: ὑπὸ τούτων [i.e. ἕξεις] γὰρ συνέχεται τὰ σώματα; καὶ τοῦ ποιὸν ἕκαστον εἶναι τῶν ἕξει συνεχομένων αἴτιος ὁ συνέχων ἀήρ ἐστιν . . . [or ὁ συνέχων αἴτιος ἀήρ ἐστιν]. Similarly, Alex. Mixt. 233–4, reporting Stoic doctrine: συνέχεσθαι, συνοχή, συνεχῆ, συνέχεια, συνέχον. ⁷⁴ Μεμνῆσθαι μέντοι χρὴ πρὸ πάντων ὅπως ἕϕαμεν ὀνομάζειν ἐνίοτε συνεκτικὸν αἴτιον, ὅτι μὴ κυρίως, ἀλλὰ καταχρώμενοι τῇ προσηγορίᾳ. Τὸ μὲν γὰρ κυρίως λεγόμενον αἴτιον συνεκτικὸν οὔτ᾽ ὠνόμασέ τις ἄλλος πρὸ τῶν Στωικῶν οὔτ᾽ εἶναι συνεχώρησε˙ τὰ δὲ καὶ πρὸ ἡμῶν οἷον συνεκτικὰ λεγόμενα γενέσεώς τινος, οὐχ ὑπάρξεως αἴτια.

280

, ,   

For the early Stoics cohesive causes are causes of something’s existing, not of something’s coming about. Galen here seems to contrast two kinds of effects: coming into being and existence.⁷⁵ Similarly, in Caus. cont. 1.5 Galen implies that the Stoics take the cohesive cause to be the cause of existing things.⁷⁶ I take this, in both passages, as short for ‘being the cause of that object existing as that object’. On the other hand, the things that come about are things like diseases, in the sense that a person contracts a disease; that is, Galen talks about changes of bodies rather than about their coming into existence (Caus. cont. 2.1–4). Galen’s distinction is thus roughly equivalent to the Stoic one between scheseis and kinēseis. The quoted Galen passage then further suggests that later, and by non-Stoics, ‘cohesive cause’ was used in the loose sense of what causes something’s coming about, and not in the original Stoic sense of something’s existing as the object it is.⁷⁷ Clement (Strom. 8.9.96.2–5), when presenting the example of the cooperation of a cohesive cause with other causes, uses some such later, loose sense of ‘cohesive cause’. The teacher is said to be the cohesive cause of the pupil’s learning, and learning is a change, not a state. In any event, Clement’s example does not match the cylinder analogy, even in the standard interpretation: for in the standard interpretation we would expect the pupil to be the cohesive cause of his learning. In the strict, early Stoic, understanding of the cohesive cause, it is indeed selfsufficient: it is the sufficient cause of an object being that object, or being in the state of being that object. However, this does not mean that ‘cohesive’ and ‘selfsufficient’ were understood by Chrysippus as two alternative ways of expressing membership of a cause in a particular class. Rather, originally, the terms ‘cohesive’ and ‘self-sufficient’ were employed to describe different, conceptually independent, relational properties of causes. And, if one cause was both cohesive and selfsufficient, this was so contingently. (When, on the other hand, it is understood in the later, non-Stoic, loose way, as a cause of change, the cohesive cause is no longer self-sufficient to produce the effect.) We can now see how the identification of Cicero’s perfect and principal cause with the cohesive cause comes about and where it goes wrong. First, as I said above, Cicero’s perfect and principal causes are wrongly identified with the second causal factor. Then, the second causal factor is wrongly identified with the cohesive cause in the following way: prompted by later sources which present the cohesive cause as a cause of change, the term ‘cohesive cause’ is understood not as a relational term that expresses a particular function (holding together), but

⁷⁵ I do not take ὕπαρξις as ὕπαρξις in the Stoic sense (the obtaining of predicates of some object), since the Stoics hold that all causes are causes of an ὕπαρξις in this sense, which fits the context in Galen badly. ⁷⁶ Vocantur autem a Stoycis non hec coniuncte cause entium, sed subtili partis substantia materialis . Cf. Caus. cont. 1.3, about the pneuma: eius opus esse continere alia corpora physica et ea que animalium. ⁷⁷ See also Hankinson 1987, 82–3; Ioppolo 1994, 4541–2.

’   

281

as a term that refers to a particular part of an object (the object’s nature), one of whose functions happens to be to hold the object together. The reasoning that would lead to the identification of the second, internal, causal factor with the cohesive cause can then be presented as follows: ‘that which holds a thing together, i.e. its cohesive cause, is the pneuma in the thing. And the nature of the thing is also this very pneuma. But the nature of the thing is the second causal factor in the instance of causation in the cylinder analogy. Therefore, the cohesive cause is this second causal factor.’ This argumentation may prima facie appear valid. But it is not. What can be derived is at most that the thing that functions as a cohesive cause is the same thing as that which functions as the second determining factor in the kinds of causation at issue. For ‘cohesive cause’ is a functional term. It describes, elliptically, a relation between a cause and its effect: the pneuma is the cause; the state of that thing being that thing is the effect. From this it does not follow that the second causal factor of a change is a cohesive cause of this change. The need to understand sunektikon, or suneches, or sunechon aition as describing a function becomes clearer when one examines the two main lines of interpretation which, taking the cohesive cause to be part of the object, identify it with the second causal factor. First, it has been argued that the external cause will activate the pneuma in the object so that its tensional states change from being non-active to being ‘active states’. By taking in energy from outside, the cohesive cause, when changed into an ‘active state’, becomes the cause of motion.⁷⁸ Alternatively, the assumption is that no such activation takes place, and that the second determining factor—again identified with the cohesive cause—is simply the shape or disposition of the object.⁷⁹ Thus, whereas in the first kind of interpretation the second causal factor is active in the sense that it is in a special ‘active state’, energy-laden from outside, and is thus an active force, in the second it can be called active only in so far as the form or disposition is the manifestation of the active principle (i.e. the pneuma) in the object. In both interpretations the identification of the cohesive cause with the second causal factor confounds two distinct functions of the pneuma in a changing object. Take the case of the rolling cylinder. On the one hand, the pneuma is responsible for the cylinder remaining a cylinder while it is rolling: this is its nature qua cohesive cause of the cylinder being a cylinder—e.g. of its having cylindrical shape. And this is important: the Stoics need to be able to say: this same cylinder was first at rest, and then, having been pushed, it rolled. But it was a cylinder all along, and the very same one. On the other hand, this same pneuma is also responsible for a certain kind of reaction of the whole body to a certain kind of external stimulus or antecedent ⁷⁸ The main exponent of this view is Frede 1980, 242. ⁷⁹ This view has been propounded e.g. by Wolff in his 1988, 533–9.

282

, ,   

cause. The effect here is a movement, not a state. In this case the function of the pneuma is not holding the object together, and hence it is not the pneuma qua cohesive cause, but it is the pneuma qua second causal factor of the rolling (i.e. depending on which interpretation one favours, either qua being activated to roll, or qua disposition to react in a certain way, if appropriately stimulated). There is one way of interpreting the cylinder analogy in which the above mistake is avoided, and the second causal factor is nevertheless understood as both self-sufficient and cohesive of the rolling: ‘the external push initiates the moving of the cylinder; but at the moment at which the cylinder has started to roll, its nature has changed from that of a cylinder at rest to that of a rolling cylinder. And the nature of the rolling cylinder is a self-sufficient cause of the cylinder’s rolling.’ But this kind of argumentation is unsatisfactory, for several reasons. First, for the Stoics there is no such thing as the nature of the rolling cylinder, or analogously of the assenting person. There may be a particular ‘being qualified’ (poion) of rolling, but the nature of a thing is what it always has, whether or not it is moving.⁸⁰ And Chrysippus maintains that the nature (and not any odd temporary ‘being qualified’) is the second causal factor. Even disregarding this point, it is the pneuma of the cylinder as it is both when at rest and when rolling (and not as it is only when it is rolling) which is responsible for its rolling: its shape (Gellius, NA 7.2.11) or perhaps its volubilitas (Gellius, NA 7.2.11; Cic. Fat. 43). Furthermore, Chrysippus insists that every movement has a preceding cause, which is a cause of that movement; accordingly, the rolling of the cylinder requires an antecedent cause, and hence the internal causal factor cannot be self-sufficient in the strict sense.⁸¹ The argumentation is even less convincing when one considers the analogous case of the assenting person, where it would run like this: ‘when the impression initiates the act of assent the person’s nature changes to that of an assenting person (if such is the person’s nature), and the nature of the assenting person (person-while-assenting) is the self-sufficient cause of the person’s assenting’. This suggestion does not work, since, first, again, there is no such thing as the nature of the assenting person, as opposed to that of the person when not assenting; and, second, it is necessary for the attribution of moral responsibility that the nature of the person, as it exists independently of whether the person is in the course of assenting, be the second causal factor of the assent. Moral responsibility is to be attached to the person, not to the person-while-assenting. In the cylinder analogy it is thus the nature of the object (its pneuma as it exists independently of whether

⁸⁰ Cf. Simplicius, Cat. 212–13; Alex. Fat. 181.13–182.20. ⁸¹ Perhaps the being qualified (ποιόν) as rolling of the rolling cylinder is a self-sufficient cause of the cylinder’s (being in the state of) being a rolling cylinder. But that seems very different from saying that the nature (ϕύσις) of the cylinder is a causal factor of the cylinder’s (movement of) rolling—which is what our sources state.

’   

283

the object is moving) which is the second causal factor. It is hence neither a selfsufficient cause nor a cohesive cause of the effect of assenting. To sum up, there is no direct evidence that the second causal factor in the cylinder analogy is a cohesive cause nor is there any compelling indirect evidence for this. The interpretation that the second causal factor is the cohesive cause has the disadvantages that it squares neither with the meaning of the word ‘cohesive’ (sunektikos, sunechēs, or sunechōn), nor with the evidence we have that for the Stoics cohesive causes are causes of states, not changes, and, further, it neglects the fact that causes are relative (pros ti) and consequently confounds the distinct functions of the pneuma in an object of (i) holding the thing, qua that thing, together, and of (ii) being a co-cause of its movement. I conclude that the second causal factor is not a cohesive cause. For the moment this may suffice with reference to the terms ‘perfect’, ‘principal’, ‘self-sufficient’, and ‘cohesive’, and the second causal factor. We are left with Cicero’s terms for the antecedent cause which is necessary but not sufficient for bringing about the effect: ‘proximus’ and ‘adiuvans’. On ‘proximus’ I have nothing new to say. Prosechēs and prokatarktikos have been suggested as Greek equivalents.⁸² Prokatarktikos fits well with Plutarch, but can hardly be justified as a translation.⁸³ But then, Cicero may not be translating the Greek word in its general meaning, but—influenced by contemporary theories—he may be trying to convey the special meaning he thinks the word has in this context.⁸⁴ In any event, ‘proximate’ is not obviously helpful in describing the kind of cause the impression is in the case of assent—except, perhaps, as picking out from the many antecedent factors that are necessary conditions of the effect the one that, besides being active, is temporally closest to the time at which the effect obtains. The second term, ‘adiuvans’, is easier to make sense of. From the way the auxiliary and proximate causes are described in Cicero, and the procatarctic cause in Plutarch, ‘helping’ or ‘auxiliary’ would be adequate characterizations of one aspect of such causes: namely, that, being neither the main cause nor sufficient in themselves to bring about the effect, they nonetheless assist in bringing it about. It has been noted that ‘adiuvans causa’ would be a natural translation of the Greek sunergon aition (helping cause) that was used in later antiquity. However, the helping cause as described, for example, in Clement (Strom. 8.9.101.13–102.12) is clearly a different type of cause from Chrysippus’ auxiliary and proximate (or procatarctic) cause. In Clement it is contrasted with the

⁸² E.g. Frede 1980, 241; Duhot 1989, 172. ⁸³ Cf. Sharples 1991, 200. ⁸⁴ As Donini has observed (Donini 1974–5, 23), Cicero’s ‘some causes provide a preparation for bringing something about’ (praecursionem quandam adhibent ad efficiendum) in Top. 59 could be an attempt to render προκαταρκτικός. See below for the context of this passage.

284

, ,   

procatarctic cause, and it is not necessary for the effect, but only intensifies it (ibid.). I assume that, if Chrysippus used a word like sunergos, this was not yet in any technical sense and that he did not introduce a type of cause with the name sunergon aition. A clearly non-technical use very similar to that in On Fate can also be found in the passage on causation in Cicero’s Topics: In this group of causes, without which something is not brought about . . . some causes provide a preparation for bringing something about, and contribute things that are themselves helping, although they are not necessitating. (Cic. Top. 59)⁸⁵

These causes (like the ones described in Cic. Fat. 41–5 and Plut. Stoic. Rep. ch. 47) are necessary but not sufficient conditions for the result; ‘adiuvans’ is used to describe a general function of these causes, and is not part of the nomenclature of a taxonomy of causes. In relation to the term ‘adiuvans’ we encounter a peculiarity in Cicero, De Fato 41–5, which many have considered problematic: in De Fato 41–2, Cicero consistently calls the non-sufficient antecedent cause an auxiliary and proximate cause; but in De Fato 44 the same kind of cause is twice referred to as proximate and continens (i.e. cohesive or contiguous) cause.⁸⁶ This has been found puzzling on two counts: first, why should the same type of cause, in the same context, be referred to in two different ways; second, on the assumption that ‘continens’ translates a Greek term such as sunektikos or sunechēs, with the meaning ‘cohesive’, it has been claimed that ‘cohesive’ is an inadequate name for or description of auxiliary and proximate, or procatarctic, causes—for, the reasoning runs, it is only the pneuma or nature in a thing which could be rightfully called ‘cohesive’ by the Stoics. But we are not confronted with a real problem here. Since any solution to the alleged difficulties will remain conjectural, I shall sketch two alternatives, leaving the ultimate choice to the reader. My own favoured explanation of this discrepancy between De Fato 41–2 and De Fato 44 is that there was an underlying Greek word ambiguous between ‘helping’ and ‘holding together’, and translated in different ways in the two passages. Sunergon is such a Greek word. This can either come from the adjective sunergos, cognate to the verb sunergeō, ‘to help’, and meaning ‘helping’, ‘auxiliary’. But it

⁸⁵ huius generis causarum, sine quo non efficitur . . . alia autem praecursionem quandam adhibent ad efficiendum et quaedam afferunt per se adiuvantia, etsi non necessaria. Ioppolo 1994, 4530–1 argues that Cic. Top. 58–9 is not Chrysippean. I agree on this point, but believe that Cicero is putting together various theories, perhaps from notes or from memory, and thus we may well find bits from Chrysippus’ theory in the text, since he wrote about Chrysippus’ distinction between causes only a month or so before he composed the Topics. ⁸⁶ Cf. section 2.

’   

285

can also be the neuter singular present active participle from the verb sunergō (suneirgō), ‘to hold together’, accordingly meaning ‘holding together’. This ambiguity holds only for accusative and nominative singular of the adjective sunergos (from sunergeō) and the participle sunergōn (from sune(i)rgō). Appropriately, in De Fato 44 the first occurrence of ‘continens’ is in the nominative singular, the second in the accusative singular. Hence both could be translations of sunergon (aition) when read as forms of the participle sune(i)rgōn. Perhaps some may find this explanation too whimsical. If one insists that Cicero or his source translated an expression such as sunechēs or sunektikos which meant ‘holding together’ for the Stoics, there need still be no inconsistency. It is helpful to remember that these terms originally described the function of a cause. Thus even in Stoic physics antecedent causes can meaningfully be called ‘cohesive’. All one has to do is apply the familiar Stoic distinction between the level of everyday experience and the cosmic level. The antecedent causes of individual motions can be looked at in two ways: as procatarctic causes they contribute to the motions of individual objects; but, if one considers their function in the universe as a whole, they serve to hold the universe qua universe together. This is, in fact, the reason why the Stoics do not permit the existence of events that have no antecedent causes: that would (as it were) explode the universe.⁸⁷ In this regard it is worth recalling that the whole context of Cicero, De Fato 39–45 is that of fate and causal determinism, and that the absence of any isolated events is a chief point of the Stoic theory of fate. Hence Cicero’s use of continens, meaning ‘cohesive’, for auxiliary and proximate causes, would be in no way inconsistent, and not even un-Stoic.

6. Does Chrysippus’ Causal Theory Admit of Self-Sufficient Causes of Change? For a full picture of Chrysippus’ theory of causes one final question needs to be addressed: does the theory admit of any self-sufficient causes of change? We know that for Chrysippus all instances of change involve antecedent causes. But do all instances of causation of change involve two determining factors, one antecedent, the other internal to the object in which the effect takes place? Or are there cases in which the antecedent cause self-sufficiently produces the effect? Before I take

⁸⁷ Cf. Alex, Fat. 192.11. Evidence for the use of συνέχειν, etc., in this second sense, on the macro level, in Stoic philosophy is: Alex. Mixt. 223–4, in particular, 223.26–7 . . . τὸ πᾶν ἡνῶσθαί τε καὶ συνέχεσθαι, πνεύματός τινος διὰ παντὸς διήκοντος αὐτοῦ . . . 224.7–8 . . . αυμμένειν τὰ σώματα αἴτιον τὸ συνέχον αὐτὰ πνεῦμα; cf. Mant. 131.5–10 and Clement, Strom. 5.8; Galen, Plen. 3.7.525.10–14, 526–7 K (SVF 2.439 and 440) for the use of συνεκτικὴ αἰτία in a related context; see also the heavily Stoicizing [Aristotle], De Mundo 6, about god, περὶ τῆς τῶν ὅλων συνεκτικῆς αἰτίας (397b9); cf. 399bff. συνέχεται.

286

, ,   

another look at Cicero, I shall sketch three possible answers to this question which may seem plausible in the context of Stoic philosophy. The first is based on the Stoic distinction between natural motions and other (non-natural or counter-natural) motions.⁸⁸ One may expect that in the case of every natural motion two determining factors are involved: first, an antecedent cause; second, the nature of the object to which the motion happens—as in the cases of the rolling cylinder and the assenting person. On the other hand, in the case of counter-natural motion, the antecedent, external cause would be selfsufficient in bringing about the effect, and the nature of the object is not a causal factor. For example, when you throw the cylinder in the air, its upward movement is not a natural movement, and—one may think—you would be the sufficient cause of its moving upward. (A variant of this suggestion could be formulated in terms of a distinction between forced and unforced motion.) Second, one could adduce the Stoic distinction of motions into actions and affections. One may think that the nature of the object, as second causal factor, contributes to the effect only if the object does something, whereas, when the object is only affected by the external, antecedent cause, this latter is the selfsufficient cause of the change. For instance, when I give assent, my nature is causally involved; but if, say, someone stabs me, the change (suffering being wounded) is something I am, it is true, affected by, but I am not doing anything, and hence I am not considered a causal factor of the effect.⁸⁹ Third, there is the possibility that in any kind of causation of change the nature of the object in which the change occurs is a causal factor, since it always depends on the nature of the object whether and what change occurs. Even when you throw a pebble into the air, although the upward movement is not natural to the pebble, it will depend on the nature of the pebble that it moves upwards—a generously sized rock would not, given the same antecedent effort of your making an attempt to throw it. Similarly, even when someone stabs me, the reason that I suffer being wounded is that I am a living being—not, for example, a piece of Camembert. So, although I am only affected and do not do anything, my nature still contributes to the effect—namely, my being wounded. Thus, even if the object does not do anything, or if the change is counter-natural, the object’s nature (the active principle in the object) is still involved in producing the effect. These, then, are the three possibilities. Which would most likely be Chrysippus’ choice? Cicero, De Fato 39–45 provides some information which points to the third. There are three sentences which imply that Chrysippus maintained that there are no perfect and principal antecedent causes:

⁸⁸ Cf. Alex. Fat. ch. 13, and Nem. Nat. hom. 105–6 for a Stoic theory of natural motions. See also Clement, Strom. 8.9.101.14 καὶ τὰ μὲν (i.e. τῶν αἰτίων) τοῦ κατὰ ϕύσιν, τὰ δὲ τοῦ παρὰ ϕύσιν. ⁸⁹ A similar suggestion has been made by Frede 1980, 236–7.

’   

287

(3) Because of this, when we say that everything happens through fate by way of antecedent causes we do not want this understood as ‘by perfect and principal causes’, but as ‘by auxiliary and proximate causes’. (Fat. 41, Latin in n. 32) (4) (b) if everything happens through fate, it follows indeed that everything happens by preceding causes, but not by perfect and principal [preceding] causes, but by auxiliary and proximate [preceding] causes. (Fat. 41, Latin in n. 33) (5) (a) Therefore, against those who introduce fate in such a way that they add necessity, the above argument will be valid; (b) but against those who will not claim that the antecedent causes are perfect and principal [i.e. Chrysippus and like-minded Stoics], it will not be valid. (Fat. 42.1, Latin in n. 34)

But if there are no perfect and principal antecedent (or preceding) causes, then (since there is no change without an antecedent cause) all causation of change requires the cooperation of an antecedent cause and an internal cause. There may be a way of arguing that the content of De Fato 41–2 is contextdependent, that Chrysippus is in fact talking only about the motions of the soul impulse and assent, and that in those cases, when he says ‘antecedent causes’ he means ‘auxiliary and proximate antecedent causes’. This would permit interpretations along the lines of the first two suggestions. However, given the absence of any further evidence on this point, and given the fact that the formulations in Cicero are quite straightforward, I propose—tentatively—that my third suggestion, that there are no self-sufficient causes of change, is what Chrysippus had in mind. (Plut. Stoic. Rep. ch. 47, is of no help: in Chrysippus’ argument (1055f–1056a) no antecedent self-sufficient causes are postulated, and the dilemmatic assumption that fate is either a self-sufficient or a procatarctic cause is presumably Plutarch’s own, but certainly not Chrysippus’.)⁹⁰ This interpretation may appear extreme. For instance, reusing my above example, one may think that—in parallel to the cylinder example—it entails that Chrysippus maintained that being stabbed is in my power and that I am morally responsible for it. However, this does not follow. There is no dispute over the fact that Chrysippus regarded the cylinder’s nature as a second determining factor for its rolling, and still did not consider the rolling as in the cylinder’s power, nor the cylinder morally responsible for its rolling. Hence we have no reason to believe that he thought that the fact that the nature of an object is involved as a determining factor makes the object morally responsible for the effect or renders the effect to be in its power. The absence of a self-sufficient antecedent cause and

⁹⁰ Cf. Bobzien 1998a, section 6.4.2.

288

, ,   

the presence of an internal second causal factor are necessary conditions for moral responsibility—nothing more.⁹¹ Furthermore, in this interpretation, the existence of a second, internal, causal factor does not make this factor automatically the one that bears the main causal responsibility. There is room for further distinctions: the internal second determining factor may be a main factor in cases of natural movements, or in those cases in which the object in question does something, but the external cause may be the main factor in other cases. For example, returning to Cicero, De Fato 7–9, we saw that the passage suggests that the climate is the principal cause for the formation of some character traits. Here the nature of the person may be a mere auxiliary cause (assuming the climatic influence to be postnatal). The involvement of the nature of the object as internal causal factor in all cases of change is less outlandish in a theory such as the Stoic one according to which causes are corporeal and the nature of things is active pneuma: given that the Stoics defined causal relations as involving two bodies and one predicate (see section 1), they may have recognized that (in the case of change) the effect is always a function of two factors, in that it is always dependent on the cooperation of the external body that initiates the change and the nature of the body in which the change takes place. The nature of the second body is always part of that ‘because of which’ (di’ho) the effect obtains. It is more than just a necessary condition of the effect. The nature of an object (which is pneuma and part of the active principle) includes its characteristic dispositional properties, and these concern equally the object’s ‘active’ reactions (such as giving assent) and ‘passive’ reactions (such as suffering being wounded) to external triggers. The object’s nature thus actively contributes to the effect in all cases of change: either in so far as the internal factor’s active contribution consists in just this fact that it is a manifestation of the active principle, or pneuma;⁹² or in so far as—even in the cases of ‘mere’ affections—there is a ‘transfer of energy’ such that the external trigger makes the pneuma in the object change itself and consequently the object.⁹³ The interpretation suggested in the present section has the additional advantage of making sense of Clement’s statement that the cohesive (sunektikon) cause is synonymously called ‘self-sufficient’ (autoteles). The standard interpretation (that the cohesive and self-sufficient cause cooperates with the procatarctic cause in

⁹¹ For the Stoics, any attribution of moral responsibility presupposes assent. One main point of the cylinder and cone analogy is to prove the agent’s responsibility by means of the fact that different people react differently to comparable externally induced stimuli (i.e. they give or withhold assent to different impressions). See Chapter 7, pp. 200–01 (Bobzien 1997, 76–8); and, in more detail, Bobzien 1998a, sections 6.3.3 and 6.3.5. ⁹² If one takes Wolff ’s 1988 position as a basis; cf. section 5. ⁹³ If one takes Frede’s 1980 view as a basis; cf. section 5.

’   

289

cases of change) faces the difficulty that both terms, ‘cohesive’ and ‘self-sufficient’, would be used in a way alien to their ordinary meaning. An alternative has recently been suggested⁹⁴ which considers the identification of cohesive and self-sufficient cause not to be Stoic. The explanation given of the identification is that, because the Stoics had no examples of a self-sufficient cause (since there were no such causes), later authors identified it with the cohesive cause as defined by medical writers. I do not find this explanation very plausible. The account of ‘self-sufficient’ in Clement does not fit the later medical concept of the cohesive cause, so that one wonders how someone could have got the idea of adding it as a second name to this cause. It seems to me that, on the contrary, the best explanation of the claimed synonymy is the fact that at some point (namely, for the early Stoics) the two attributes ‘cohesive’ and ‘self-sufficient’, when used of causes, actually had the same extension. For according to the suggested interpretation, for Chrysippus the cohesive causes, being causes of states, are indeed the only causes that are selfsufficient in bringing about their effect. As I said above, ‘self-sufficient’ originally described only a certain feature of causes, and I suggest that later the term became an alternative class name for the class of cohesive causes, since in the Stoic system for causes it had the same extension as ‘cohesive’.⁹⁵ It may be of interest in this context that Clement and Sextus both report from causal theories which (i) stress that the effect depends on the suitability or fitness (epitēdeiotēs) of the body at which the effect takes place, and hence the same thing becomes a cause of different effects at different objects;⁹⁶ (ii) maintain that every cause is that cause relative to the thing at which the effect takes place;⁹⁷ and (iii) state that suitability is a necessary condition and thus a sine qua non type cause of the effect.⁹⁸ I doubt that these theories are early Stoic. However, they still show that it was part of the debate over causation that the suitability of the object (which is part of the object’s nature) is a necessary condition and a cause of the change even in those cases where the change is a mere affection of the object. The difference between early Stoics and this later position seems to be this: whereas, for the orthodox Stoa, the internal causal factor is corporeal pneuma, and thus actively involved in bringing about the effect, for causal theories which do not postulate a corporeal active principle, the internal causal factor is demoted to a mere necessary condition—at least in those cases in which the object at issue is affected only, and not doing something. ⁹⁴ By Schröder 1989, 237, followed by Ioppolo 1994, 4542. ⁹⁵ That a self-sufficiently productive cause was later understood in a way that fits the Stoic cohesive cause well is shown in SE M 9.238 and 242, where it is implied that, if it is the nature of a cause to bring about an effect self-sufficiently and by using its own power, then it brings about its effect all the time. This is exactly what the early Stoic cohesive cause does. ⁹⁶ Clement, Strom. 8.9.100.20–101.3; cf. SE M 9.250–1. ⁹⁷ Clement, Strom. 8.9.98.25–30; cf. SE M 9.239 and 243. ⁹⁸ Clement, Strom. 8.9.98.7–12; cf. SE M 9.243.

290

, ,   

7. Results We end up with the following picture of Chrysippus’ theory. In any instance of causation any cause is that cause only relative to the relevant effect. If there is any basic Stoic distinction between causes, it is that between causes of states and causes of change. Causes of states are required in the Stoic system in particular in order to explain the existence and continuation of individual objects. These causes are cohesive (sunechē, sunektika, . . . ) causes. It follows from their nature (i.e. active pneuma) and function of holding together objects qua being these objects that they are self-sufficient (autotelēs, perfectus) in bringing about their effect. We can assume that they were thought to necessitate their effects. By contrast, any instance of causation of change requires the cooperation of at least two causal factors. Hence no cause of change is self-sufficient. Every change requires at least one antecedent (proēgoumenon, antecedens, . . . ) cause to get the change started. Changes always take place in a corporeal object, and the nature or pneuma of this object is always a second causal factor of the effect: whether and what change occurs depends in part on the constitution of the object that changes or is changed. The second causal factor is not a cohesive cause (sunektikon aition) of the effect. Cohesion is not its function. In the cases of change that Chrysippus discusses in Cicero De Fato 41–5, the main responsibility for the effect is attached to the second causal factor. Accordingly, this second causal factor might have been referred to as principal cause (principalis, kurios/kuriōs). However, we have no evidence for such a usage of the term and in any event this second factor is not part of the Chrysippean distinction Cicero reports. When, as in Chrysippus’ examples, the first causal factor is not the main cause, it is not a necessitating (necessarius) cause of the effect. It is only auxiliary (adiuvans, sunergos) in bringing about the effect. This is so in the cases of human assent and the rolling of the cylinder, in which the externally induced impression and the pushing, respectively, are not necessary causes. It is possible that Chrysippus allowed for cases in which the first causal factor—that is, the antecedent cause—is the principal cause, while the nature or pneuma of the object at which the change occurs only helps to bring about the effect. We do not know whether such antecedent principal causes would have been thought to necessitate their effect. As regards the (formerly) standard view that the Chrysippean distinction of causes that Cicero reports is one of causes that cooperate in one and the same instance of causation, it has become clear that this is not so. This Chrysippean distinction is between two possible ways of thinking of antecedent causes of change, as auxiliary or as main cause of the effect, and he argues that, in the case of assent, the antecedent cause is only an auxiliary cause and not the main cause.

Bibliography Editions, Translations, Etc. Alcinous Dillon, J. Alcinous: The Handbook of Platonism (Oxford, 1993).

Alexander of Aphrodisias Sharples, R. W. Alexander of Aphrodisias on Fate: Text, Translation, and Commentary (London, 1983). Sharples, R. W. Alexander of Aphrodisias: Ethical Problems (London, 1990).

Anonymus [Anonymus], In Aristotelis Ethica Nicomachea ii–v commentaria, in Eustratii et Michaelis et anonyma in Ethica Nicomachea commentaria, ed. G. Heylbut, Commentaria in Aristotelem Graeca 20 (Berlin, 1892), 122–255.

Aquinas Aquinas, T., Sententia libri ethicorum, in R. A. Gauthier (ed.), Opera omnia, Vol. 47.1 (Rome, 1969).

Aristotle Albertus Magnus (comm.), Liber III Ethicorum, in A. Borgnet (ed.), Opera omnia (Paris, 1891), Vol. 7, Ethica. Aristotle/Averroes, Aristotelis opera cum Averrois commentariis, Vol. 3 [2nd Juntine] (Venice, 1562, repr. Frankfurt a. M., 1963). Aristotle, Decem librorum Moralium Aristotelis tres conuersiones: Prima Argyropili Byzantij/secunda Leonardi Aretini/tertia vero Antiqua per Capita et numeros conciliate: communi/familiariq[ue] commentario [Jacobi Stapulensis] ad Argyropilum adiecto, [tres conuersiones] (Paris, 1497) . Barnes, J. (ed.), Aristotle: The Complete Works of Aristotle: The Revised Oxford Translation (Oxford, 1984). Bekker, I. (ed.), Aristotelis Opera (Berlin, 1831). Browne, R. W., The Nicomachean Ethics of Aristotle (London, 1889). Burgundius Pisanus (trans.), Aristotelis: Ethica Nicomachea: translatio antiquissima librorum II et III siue ‘Ethica uetus’, ed. R. A. Gauthier, Aristoteles Latinus 26.1–3, 2nd fasc. (Turnhout, 1972), 5–48. Buridan, J., Quaestiones super decem libros Ethicorum Aristotelis ad Nicomachum (Paris, 1513). Repr. as Super decem libros Ethicorum (Frankfurt a. M., 1968). Burley, Walter, Expositio librorum Ethicorum (Venice, 1481, 1500²). Burnet, J. (comm.), The Ethics of Aristotle (London, 1900). Bywater, J. Aristotle’s Ethica Nicomachea (Oxford, 1894).

292



Camerarius, I., Ethicorum Aristotelis Nicomachiorum explicatio accuratissima (Frankfurt, 1578). Chase, D. P., The Nicomachean Ethics of Aristotle: A New Translation, Mainly from the Text of Bekker (Oxford, 1861). Crisp, R. 2000, Aristotle: Nicomachean Ethics, Cambridge Texts in the History of Philosophy (Cambridge, 2002). Dirlmeier, F., Aristoteles: die Nikomachische Ethik (Berlin, 1956, repr. Stuttgart, 2003). Garve, V. C. (trans. and comm.), Die Ethik des Aristoteles, 2 vols (Breslau, 1798–1801). Gauthier, R. A. and J. Y. Jolif, Aristote: L’Ethique à Nicomaque. 3 vols (Louvain, 1958–9, 2nd edition, with a new introduction, 1970). Gigon, O., Die Nikomachische Ethik (Zürich, 1967). Gillies, J. (trans.), Aristotle’s Ethics and Politics, Comprising his Practical Philosophy (London, 1789). Grant, Arthur, The Ethics of Aristotle, Vol. 1 (London, 1866). Grant, Arthur, The Ethics of Aristotle, Vol. 2 (London, 1874). Grosseteste, R. (trans.), Ethica Nicomachea: libri I—III; VIII.1–5 (6) (‘recensio pura’— Burgundii translationis recensio), ed. R. A. Gauthier, Aristoteles Latinus 26.1–3, 3rd fasc. (Turnhout, 1972), 141–201, 271–305. Hardie, W. F. R., Aristotle’s Ethical Theory (Oxford, 1968, 1980²). Ibn Rushd, Averroes’ Middle Commentary on Aristotle’s Nicomachean Ethics in the Hebrew Version of Samuel ben Judah, ed. L. Berman (Jerusalem, 1999). Joachim, H. H., Aristotle: The Nicomachean Ethics: A Commentary (Oxford, 1851). Joachim, H. H. and D. A. Rees, Aristotle: The Nicomachean Ethics: A Commentary (Oxford, 1951). Lambin, D. (trans.), Aristoteles: In libros De moribus ad Nicomachum annotiationes (Venice, 1558). Lambin, D. (trans.), Aristoteles: Ethicorum Nicomachiorum libri decem, ed. T. Zwinger (Basel, 1566). Lambin, D. (trans.), Aristoteles: Ethicorum Nicomacheorum libri decem, ed. W. Wilkinson (Oxford, 1716). Lefèvre d’Etaples, J., Decem librorum Moralium Aristotelis tres conuersiones (Paris, 1497). Marchesi, C. (ed.), L’etica nicomachea nella tradizione latina medievale, Documenti ed Appunti (Messina, 1904). Michael of Ephesus, Michaelis Ephesii in librum quartum Ethicorum Nicomacheorum commentarium, ed. M. Hayduck, Commentaria in Aristotelem Graeca 22.3 (Berlin, 1901). Michelet, K. L., Ethicorum Nicomacheorum libri decem: Ad codicum manuscriptorum et veterum editionum fidem recensuit commentariis illustravit in usum scholarum suarum, 2 vols (Berlin 1829, 1835²). Moerbeke, William of (trans.), Aristotelis secundum exemplar Parisiacum: Ethica Nicomachea (‘recensio recognita’—Roberti Grosseteste translationis recensio), ed. R. A. Gauthier, Aristoteles Latinus 26.1–3, 4th fasc. (Turnhout, 1974), 375–588. Ostwald, M., Aristotle: Nicomachean Ethics, translated with an introduction and notes (Indianapolis, 1962). Pachymeres, Georgios, Philosophia, 11: Ethica Nicomachea, ed. K. Oikonomakos, Corpus Philosophorum Medii Aevi, Commentaria in Aristotelem Byzantina 3 (Athens, 2005). Perion, I., (trans.), Aristoteles: De Moribvs qu[a]e Ethica nominantur, ad Nicomachu[m] filium, Libri decem (Paris, 1540; Basel 1542). Peters, F. H. (trans.), The Nicomachean Ethics of Aristotle (London, 1881).



293

Rackham, H. (trans.), Aristotle: The Nicomachean Ethics (Cambridge, MA, 1926). Rieckher, J. (trans.), Nikomachische Ethik (Stuttgart, 1856). Rolfes, E. (trans.), Aristoteles, Nicomachische Ethik, ed. G. Bien (Hamburg, 1985). Rolfes, Eugen, Aristoteles: Nikomachische Ethik (Leipzig, 1921). Ross, D., Aristotle: The Nicomachean Ethics (Oxford, 1925). Stahr, A. (trans. and comm.), Aristoteles’ nikomachische Ethik (Stuttgart, 1863). Taylor, C. C. W., Aristotle: Nicomachean Ethics: Books II—IV, translated with introduction and commentary, Clarendon Aristotle Series (Oxford, 2006). Taylor, T., The Rhetoric, Poetic, and Nicomachean Ethics of Aristotle, 2 vols (London, 1818). Vernia, N. (ed.), Aristoteles: Opera latina cum commentariis Averrois (Venice, 1483). Wardman, A. E. and R. Bambrough, The Philosophy of Aristotle (New York/London, 1963). Welldon, J. E. C., The Nicomachean Ethics of Aristotle (London, 1982). Williams, R. (trans.), The Nicomachean Ethics of Aristotle (London, 1869, 1876², 1879³). Woods, John, Aristotle’s Eudemian Ethics: Books I, II, and VIII, 2nd edition (Oxford, 1992). Zell, K. (trans.), Aristotelis Ethica Nicomachea (Heidelberg, 1820).

Aspasius Aspasius, In Ethica Nicomachea quae supersunt commentaria, ed. G. Heylbut, Commentaria in Aristotelem Graeca 19.1 (Berlin, 1889). Aspasius: Earliest Extant Commentary on Aristotle’s Ethics, ed. Antonina Alberti and Robert W. Sharples, Peripatoi. Philologisch-historische Studien zum Aristotelismus (Berlin, 1998).

Boethius Boethius, Commentaries on De Interpretatione, ed. C. Meiser (Leipzig, 1887, 1880). Sharples, R. W., Cicero: On Fate and Boethius: The Consolations of Philosophy IV.5–7, V (Warminster, 1991).

Choeroboscus Choeroboscus, Georgius, Prolegomena et scholia in Theodosii Alexandrini canones isagogicos de flexione verborum, ed. A. Hilgard, GG 4.2 (Leipzig, 1894).

Chrysippus Gercke, A. ‘Chrysippea’, Jahrb. für klass. Phil. Supplbd 14 (1885).

Cicero Cicero, De Natura Deorum Libri Tres, with introduction and commentary by J. Mayor, Vol. 2 (Cambridge, 1883, repr. 2010). Pease, Arthur Stanley, M. Tulli Ciceronis De Natura Deorum libri secundus et tertius (Cambridge, MA/London, 1958). Sharples, R. W., Cicero: On Fate and Boethius: The Consolations of Philosophy IV.5–7, V (Warminster, 1991). Whittaker, J. Alcinoos: Enseignement des doctrines de Platon (Paris, 1990).

Diogenes Laertius Marcovich, M. (ed.), Diogenes Laertius: Vitae Philosophorum, Vol. 1 (Berlin, 1999).

Epicurus Epicurus—Graziano Arrighetti, Epicuro: Opere. Introduzione, testo critico, traduzione e note, Classici della Filosofia 4 (Turin, 1960).

294



Giussani, C., Lucreti Cari De Rerum Natura Libri Sex, Vol. 1 (Turin, 1896). Laursen, S., ‘Epicurus On Nature Book XXV’, Cronache Ercolanesi 18 (1988): 7–18. Laursen, S., ‘The Summary of Epicurus’ On Nature Book 25’, Papiri Letterari Greci e Latini (1992): 143–54. Laursen, S., ‘The Early Parts of Epicurus, On Nature, 25th Book’, Cronache Ercolanesi (1995): 5–109. Laursen, S., ‘The Later Parts of Epicurus, On Nature, 25th Book’, Cronache Ercolanesi (1997): 5–82.

Eustratius Eustratius of Nicaea, Eustratii et Michaelis et anonyma in Ethica Nicomachea commentaria, ed. G. Heylbut, Commentaria in Aristotelem Graeca 20 (Berlin, 1892). Eustratius of Nicaea, Orationes, ed. A. Demetrakopoulos, ̓Εκκλησιαστικὴ Βιβλιοθήκη 1 (Leipzig, 1866, repr. Hildesheim, 1965).

[Heliodorus of Prusa] [Heliodorus of Prusa], In Ethica Nicomachea paraphrasis, ed. G. Heylbut, Commentaria in Aristotelem Graeca 19.2 (Berlin, 1889).

Ioannes Stobaeus Ioannes Stobaeus, Anthologii libri duo priores qui inscribi solent Eclogae physicae et ethicae, ed. K. Wachsmuth, Vols 1–2 (Berlin, 1884).

John XI Beccus Joannes XI Beccus, Four Books to Constantine Meliteniotes, ed. J.-P. Migne, Patrologia Graeca 141 (Paris, 1865), 337–96.

Manuel II Palaeologus Manuel II Palaeologus, Dialogi cum mahometano, in Piissimi et sapientissimi imperatoris Manuelis Palaeologi opera omnia, theologica, polemica, panegyrica, paedagogica, ed. J.-P. Migne, Patrologia Graeca 156 (Paris, 1866).

Nemesius Nemesius, De natura hominis, ed. M. Morani (Leipzig, 1987).

Philo of Alexandria Philo of Alexandria, Philonis Alexandrini Opera quae supersunt, ed. L. Cohn and P. Wendland, 6 vols (Berlin, 1886–1915, Vol. 6 repr. 1962).

Collections of Testimonies Arnim, H., Stoicorum veterum fragmenta, Vols 1–3 (Leipzig, 1903–5); M. Adler, Stoicorum veterum fragmenta, Vol. 4, indices (Leipzig, 1924) (SVF). Arrighetti, G., Epicuro: Opere (Turin, 1960, 1973²). Diels, H., Doxographi Graeci (Berlin, 1879, 1958³). Usener, H., Epicurea (Leipzig, 1887).



295

Studies Akasoy, A. A. and A. Fidora (2002) ‘Hermannus Alemannus und die alia translatio der Nikomachischen Ethik’, Bulletin de philosophie medieval 44: 79–93. Akasoy, A. A. and A. Fidora (eds) (2005) The Arabic Version of the Nicomachean Ethics with intro. and trans. by D. M. Dunlop (Leiden). Akasoy, A. A. and G. Giglioni (eds) (2013) Renaissance Averroism and Its Aftermath: Arabic Philosophy in Early Modern Europe (New York). Algra, K., J. Barnes, J. Mansfeld, and M. Schofield (eds) (1998) The Cambridge History of Hellenistic Philosophy (Cambridge). Annas, J. (1992) Hellenistic Philosophy of Mind (Berkeley, Los Angeles, London). Apostle, H. G. and L. P. Gerson (ed. and trans.) (1983) Aristotle: Selected Works (Grinnell). Asmis, E. (1970) The Epicurean Theory of Free Will and Its Origins in Aristotle, PhD dissertation, Yale University. Asmis, E. (1990) ‘Free Action and the Swerve: Review of Walter G. Englert, Epicurus on the Swerve and Voluntary Action’, OSAP 8: 275–91. Avotins, I. (1980) ‘Notes on Lucretius 2.251–293’, Harvard Studies in Classical Philology 84: 75–9. Avotins, I. (1979) ‘The Question of mens in Lucretius 2.289’, Classical Quarterly 29: 95–100. Bailey, C. (1928) The Greek Atomists and Epicurus (New York). Barber, C. and D. Jenkins (2009) Medieval Greek Commentaries on the Nicomachean Ethics (Leiden). Barnes, J., M. Burnyeat, and M. Schofield (eds) (1980) Doubt and Dogmatism: Studies in Hellenistic Epistemology (Oxford). Barney, R. (2016) ‘Gorgias’s Encomium of Helen’, in E. Schliesser (ed.), Ten Neglected Classics of Philosophy (Oxford), 1–25. Bejczy, I. P. (ed.) (2008) Virtue Ethics in the Middle Ages: Commentaries on Aristotle’s Nicomachean Ethics 1200–1500 (Leiden). Bloos, L. (1973) Probleme der stoischen Physik (Hamburg). Bobzien, S. (1993) ‘Chrysippus’ Modal logic and Its Relation to Philo and Diodorus’, in K. Döring and Th. Ebert (eds), Dialektiker und Stoiker (Stuttgart), 63–84. Bobzien, S. (1997) ‘Stoic Conceptions of Freedom and Their Relation to Ethics’, in R. Sorabji (ed.), Aristotle and After (London), 71–89. Bobzien, S. (1998a) Determinism and Freedom in Stoic Philosophy (Oxford). Bobzien, S. (1998b) ‘The Inadvertent Conception and Late Birth of the Free-Will Problem’, Phronesis 43: 133–75. [Reprinted, with minor corrections, in R. Sorabji (ed.), Aristotle Re-Interpreted, London, 2016, 125–60.] Bobzien, S. (1999) ‘Chrysippus’ Theory of Causes’, in K. Ierodiakonou (ed.), Topics in Stoic Philosophy (Oxford), 196–242. Bobzien, S. (2000) ‘Did Epicurus Discover the Free-Will Problem?’, Oxford Studies in Ancient Philosophy 19: 287–337. Bobzien, S. (2005) ‘Early Stoic Determinism’, Revue de métaphysique et de morale 48(4): 489–516. Bobzien, S. (2006) ‘Moral Responsibility and Moral Development in Epicurus’ Philosophy’, in B. Reis and S. Haffmans (eds), The Virtuous Life in Greek Ethics (Cambridge), 206–29. Bobzien, S. (2012) Review of Michael Frede, A Theory of the Will, Berkeley 2011, Journal of the History of Philosophy 50(2): 292–3.

296



Bobzien, S. (2013) ‘Found in Translation: Aristotle’s Nicomachean Ethics III.5 1113b7–8 and Its Reception’, Oxford Studies in Ancient Philosophy 45: 103–48. Bobzien, S. (2014a) ‘Choice and Moral Responsibility in Aristotle’s Nicomachean Ethics iii 1–5’, in R. Polansky (ed.), The Cambridge Companion to Aristotle’s Nicomachean Ethics (Cambridge), 81–109. Bobzien, S. (2014b) ‘Free Choice in Aristotle’s Nicomachean Ethics III.5?’, in P. Destrée, R. Salles, and M. Zingano (eds), What Is up to Us? Studies on Causality and Responsibility in Ancient Philosophy (Bonn). Botros, S. (1985) ‘Freedom, Causality, Fatalism and Early Stoic Philosophy’, Phronesis 30: 274–304. Bown, Alexander. (2016) ‘Epicurus on Bivalence and the Excluded Middle’, Archiv für Geschichte der Philosophie 98(3): 239–71. Broadie, S. (1991) Ethics with Aristotle (Oxford). Broadie, S. (2002) ‘Philosophical Introduction’, in S. Broadie and C. Rowe (eds), Aristotle: Nicomachean Ethics (Oxford), 9–1. Broadie, Sarah and Christopher Rowe (2002) Aristotle: Nicomachean Ethics: Translation, Introduction, and Commentary (Oxford). Bronowski, A. (2019) The Stoics on Lekta: All There is to Say (Oxford). Buchanan, D. R. (2000) An Ethic for Health Promotion: Rethinking the Sources of Human Well-Being (Oxford). Charles, David (1984) Aristotle’s Philosophy of Action (London). Conway, P. (1981) ‘Epicurus’ Theory of Freedom of Action’, Prudentia 13: 81–9. Coope, Ursula (2020) Freedom and Responsibility in Neoplatonist Thought (Oxford). Den Boeft, J. (1970) Calcidius on Fate: His Doctrines and Sources (Leiden). Destrée, P. (2011) ‘Aristotle on Responsibility for One’s Character’, in M. Pakaluk and G. Pearson (eds), Moral Psychology and Human Action in Aristotle (Oxford), 285–318. Destrée, Pierre, R. Salles, and Marco Antonio De Zingano (eds) (2014) What Is up to Us? Studies on Causality and Responsibility in Ancient Philosophy (Sankt Augustin). Dihle, A. (1982) The Theory of Will in Classical Antiquity (Berkeley/Los Angeles/London). Dihle, A. (1984) ‘Zur Schicksalslehre des Bardesanes’, in Antike and Orient (Heidelberg), 161–73. Dillon, J. M. (1977) The Middle Platonists (London). Dobbin, R. (1991) ‘Προαίρεσις in Epictetus’, Ancient Philosophy 11: 111–35. Donini, P. L. (1974–5) ‘Fato e volontà umana in Crisippo’, Atti dell’Accademia delle Scienze di Torino 109: 1–44. Donini, P. L. (1988) ‘Plutarco e il determinismo di Crisippo’, in I. Gallo (ed.), Aspetti dello stoicismo e dell’epicureismo in Plutarco (Ferrara), 21–32. Doyle, B., ‘Free Will’, . Duhot, J. J. (1989) La conception stoicienne de la causalité (Paris). Englert, W. G. (1987) Epicurus on the Swerve and Voluntary Action (Atlanta, GA). Everson, S. (1990) ‘Aristotle’s Compatibilism in the Nicomachean Ethics’, Ancient Philosophy 10(1): 81–103 (repr. in L. P. Gerson (ed.), Aristotle: Psychology and Ethics, 1999, 252–76). Forschner, M. (1981) Die stoische Ethik (Stuttgart). Fowler, D. (1983) ‘Lucretius on the Clinamen and “Free Will” (II 251–93)’, in Συζήτηεσις: Studi sull’Epicureismo Greco e Romano offerti a Marcello Gigante (Naples), 329–52. Frede, D. (1982) ‘The Dramatisation of Determinism: Alexander of Aphrodisias’ “De fato” ’, Phronesis 27: 276–98.



297

Frede, D. (1985) ‘The Sea-Battle Reconsidered’, Oxford Studies in Ancient Philosophy 3: 31–87. Frede, M. (1974) Die stoische Logik (Göttingen), 107–17. Frede, M. (1980) ‘The Original Notion of Cause’, in J. Barnes, M. Burnyeat, and M. Schofield (eds), Doubt and Dogmatism (Oxford), 217–49. Frede, M. (1993) ‘The Stoic Conception of Reason’, in K. J. Boudouris (ed.), Hellenistic Philosophy, Vol. 2 (Athens), 50–63. Frede, Michael (2007) ‘The ἐφ’ ἡμῖν in Ancient Philosophy’, ΦΙΛΟΣΟΦΙΑ 37: 110–23. Frede, Michael (2011) A Free Will: Origins of the Notion in Ancient Thought, Sather Classical Lectures (Berkeley, CA). Furley, D. (1967) ‘Aristotle and Epicurus on Voluntary Action’, the second study in his Two Studies in the Greek Atomists (Princeton, NJ). Göransson, T. (1995) Albinus, Alcinous, Arius Didymus (Göteborg). Görler, W. (1987) ‘Hauptursachen bei Chrysipp und Cicero? Philologische Marginalien zu einem vieldiskutierten Gleichnis (De fato 41–44)’, Rheinisches Museum fuer Philologie 130: 254–74. Gulley, N. (1990) ‘Lucretius on Free Will’, Symbolae Osloenses 65: 37–52. Hamburger, M. (1951) Morals and Law: The Growth of Aristotle’s Legal Theory (New Haven, CT). Hammerton, J. A. (ed.) (1936) Outline of Great Books, Vol. 1 (New York). Hankinson, R. J. (1987) ‘Evidence, Externality and Antecedence: Inquiries into Later Greek Causal Concepts’, Phronesis 32: 80–100. Hankinson, R. J. (1995) The Sceptics (London/New York). Hatch, W. M. (1879) The Moral Philosophy of Aristotle (London). Hecht, J. (2014) ‘Freedom of the Will in Plato and Augustine’, British Journal for the History of Philosophy 22(2): 196–215. Huby, P. (1967) ‘The First Discovery of the Freewill Problem’, Philosophy 42: 353–62. Huby, P. (1969) ‘The Epicureans, Animals, and Freewill’, Apeiron 3: 17–19. Ierodiakonou, K. (ed.) (1998) Topics in Stoic Philosophy (Oxford). Inwood, B. (1985) Ethics and Human Action in Early Stoicism (Oxford). Ioppolo, A. M. (1988) ‘Le cause antecedenti in Cic. De Fato 40’, in J. Barnes and M. Mignucci (eds), Matter and Metaphysics (Naples), 397–424. Ioppolo, A. M. (1994) ‘Il concetto di causa nella filosofia ellenistica e romana’, in W. Haase and H. Temporini (eds), Rise and Decline of the Roman World, pt II, 36.7 (Berlin), 4492–545. Irwin, T. (1980) ‘Reason and Responsibility in Aristotle’, in A. O. Rorty (ed.), Essays in Aristotle’s Ethics (Berkeley, CA), 117–55. Irwin, Terence (1985) Nicomachean Ethics: With Introduction, Notes, and Glossary (Indianapolis, IN). Kahn, C. H. (1988) ‘Discovering the Will: From Aristotle to Augustine’, in J. M. Dillon and A. A. Long (eds), The Question of Eclecticism (Berkeley/Los Angeles/London), 234–59. Kenny, Anthony (1979) The Aristotelian Ethics (Oxford). Kiros, T. (1998) Self-Construction and the Formation of Human Values (Westport, CT). Kleve, K. (1980) ‘Id facit exiguum clinamen’, Symbolae Osloenses 55: 27–31. Kleywegt, A. J. (1973) ‘Fate, Free Will and the Text of Cicero’, Mnemosyne 26: 342–49. Kraut, Richard (ed.) (2005) The Blackwell Guide to Aristotle’s Ethics (Oxford). Lapidge, M. (1978) ‘Stoic Cosmology’, in J. M. Rist (ed.), The Stoics (Berkeley/Los Angeles). Laursen, S. (1990) ‘Against Democritus: Towards the End’, in M. Capasso et al., Miscellanea Papyrologica (Florence), 3–22.

298



Lines, D. F. (2002) Aristotle’s Ethics in the Italian Renaissance (ca 1300–1650) (Leiden). Long, A. A. (1968) ‘The Stoic Concept of Evil’, Philosophical Quarterly 18: 329–43. Long, A. A. (1970) ‘Stoic Determinism and Alexander of Aphrodisias’ De Fato i–xiv’, Archiv für Geschichte der Philosophie 52: 247–68. Long, A. A. (1971) ‘Freedom and Determinism in Stoic Theory of Human Action’, in A. A. Long (ed.), Problems in Stoicism (London), 173–99. Long, A. A. (1974) Hellenistic Philosophy (New York), 56–61. Long, A. A. (1976) ‘Early Stoic Concept of Moral Choice’, in Images of Man in Ancient and Medieval Thought, Festschrift for G. Verbeke (Louvain), 77–92. Long, A. A. (1977) ‘Chance and Natural Law in Epicureanism’, Phronesis 22: 63–88. Long, A. A. and D. N. Sedley (1987) The Hellenistic Philosophers, 2 vols (Cambridge). Lorenz, Hendrik (2009) ‘Virtue of Character in Aristotle’s Nicomachean Ethics’, Oxford Studies in Ancient Philosophy 37: 177–212. Lorenz, Hendrik (2014) ‘Aristotle’s Analysis of Akratic Action’, in R. Polansky (ed.), The Cambridge Companion to Aristotle’s Nicomachean Ethics (Cambridge), 242–62. Mackendrick, P. (1989) The Philosophical Books of Cicero (London). Mansfeld, J. (1999) ‘The Sources’, in K. Algra, J. Barnes, J. Mansfeld, and M. Schofield (eds), The Cambridge History of Hellenistic Philosophy (Cambridge), 3–30. Masi, F. G. and S. Maso (eds) (2013) Fate, Change, and Fortune in Ancient Thought (Amsterdam). McClure, S. S. (1920) The World’s Greatest Books, Vol. 13: Philosophy and Religion (New York). McTavish, I. (2012) A Prisoner’s Wisdom: Transcending the Ego (Bloomington, IN). Mercken, H. P. F. (1990) ‘The Greek Commentators on Aristotle’s Ethics’, in R. Sorabji (ed.), Aristotle Transformed (London), 407–10. Meyer, S. S. (1993) Aristotle on Moral Responsibility (Oxford). Meyer, S. S. (1994) ‘Self-Movement and External Causation’, in M. M. L. Gill and L. G. Lennox (eds), Self-Motion (Princeton, NJ), 65–80. Meyer, S. (2006) ‘Aristotle on the Voluntary’, in R. Kraut (ed.), The Blackwell Guide to Aristotle’s Nicomachean Ethics (Oxford), 137–57. Meyer, S. S. (2014) ‘Aristotle on What Is up to Us and What Is Contingent’, in P. Destrée, R. Salles, and M. Zingano (eds), What Is up to Us? Studies on Agency and Responsibility in Ancient Philosophy (Bonn), 75–89. Mitsis, P. (1988) Epicurus’ Ethical Theory (Ithaca, NY). Moore, E. (1902) An Introduction to Aristotle’s Ethics (New York). Moraux, P. (1984) Der Aristotelismus bei den Griechen von Andronikos bis Alexander von Aphrodisias II: Der Aristotelismus im I und II Jh. V. Chr. (Berlin). Natali, Carlo (2014) ‘The Book on Wisdom’, in R. Polansky (ed.), The Cambridge Companion to Aristotle’s Nicomachean Ethics (Cambridge), 180–202. Naydler, J. (2009) The Future of the Ancient World: Essays on the History of Consciousness (Rochester, NY). O’Brien, J. A. (1946) Truths Men Live By (New York). Pakaluk, M. (2005) Aristotle’s Nicomachean Ethics: An Introduction (Cambridge). Pakaluk, Michael and Giles Pearson (eds) (2011) Moral Psychology and Human Action in Aristotle (New York). Pohlenz, M. (1959) Die Stoa, 2nd edition (Göttingen). Purinton, J. S. (1999) ‘Epicurus on “Free Volition” and the Atomic Swerve’, Phronesis 44: 253–99.



299

Rapp, C. (1995) ‘Freiwilligkeit, Entscheidung und Verantwortlichkeit (III 1–7)’, in O. Höffe (ed.), Aristoteles: Die Nikomachische Ethik (Berlin), 109–33. Reesor, M. E. (1965) ‘Fate and Possibility in Early Stoic Philosophy’, Phoenix 19: 285–97. Rieth, O. (1933) Grundbegriffe der stoischen Ethik (Berlin). Rist, J. M. (ed.) (1978) The Stoics (Berkeley/Los Angeles). Sambursky, S. (1959) Physics of the Stoics (London). Saunders, T. (1984) ‘Free Will and the Atomic Swerve in Lucretius’, Symbolae Osloenses 59: 37–59. Schmidt, E. A. and M. Ullmann (2012) Aristoteles in Fes: Zum Wert der arabischen Überlieferung der ‘Nikomachischen Ethik’ für die Kritik des griechischen Textes (Heidelberg). Schröder, S. (1989) ‘Philosophische und medizinische Ursachensystematik und der stoische Determinismus’, Prometheus 15: 209–39. Schröder, S. (1990) ‘Philosophische und medizinische Ursachensystematik und der stoische Determinismus’, Prometheus 16: 5–26. Sedley, D. N. (1983) ‘Epicurus’ Refutation of Determinism’, in ΣΥΖΗΤΗΣΙΣ: Studi sull’Epicureismo Greco e Romano offerti a Marcello Gigante (Naples), 11–51. Sedley, D. N. (1993) ‘Chrysippus on Psychophysical Causality’, in J. Brunschwig and M. C. Nussbaum (eds), Passions and Perceptions (Cambridge), 313–31. Segvic, Heda (2009) ‘Deliberation and Choice in Aristotle’, in her From Protagoras to Aristotle: Essays in Ancient Moral Philosophy (Princeton, NJ), 144–71. Reprinted in M. Pakaluk and G. Pearson (eds), Moral Psychology and Human Action in Aristotle (New York), 159–86. Sharples, R. W. (1975) ‘Responsibility, Chance, and Not-Being in Alexander of Aphrodisias Mantissa 169–172’, Bulletin of the Institute of Classical Studies 22: 37–63. Sharples, R. W. (1981) ‘Necessity in the Stoic Doctrine of Fate’, Symbolae Osloensis 56: 81–97. Sharples, R. W. (1986) ‘Soft Determinism and Freedom in Early Stoicism’, a reply to Botros, Phronesis 31: 266–79. Sharples, R. W. (1991–3) ‘Epicurus, Carneades and the Atomic Swerve’, Bulletin of the Institute of Classical Studies: 174–90. Sharples, R. W. (1995) ‘Causes and Necessary Conditions in the Topica and De Fato’, in J. G. F. Powell (ed.), Cicero the Philosopher (Oxford), 247–71. Sharples, R. W. (2001) ‘Determinism, Responsibility and Chance’, in P. Moraux (ed.), Der Aristotelismus bei den Griechen von Andronikos bis Alexander von Aphrodisias III: Alexander von Aphrodisias (Berlin), 513–92. Smyth, H. W. (1920) Greek Grammar (Cambridge, MA). Sorabji, R. (1980a) ‘Causation, Laws and Necessity’, in J. Barnes, M. Burnyeat, and M. Schofield (eds), Doubt and Dogmatism (Oxford), 250–82. Sorabji, R. (1980b) Necessity, Cause and Blame: Perspectives on Aristotle’s Theory (Ithaca, NY). Sparshott, F. (1994) Taking Life Seriously: A Study of the Argument of the Nicomachean Ethics (Toronto). The Spectator 43 (1870), 178–80. Stack, G. J. (1974) ‘Aristotle and Kierkegaard’s Existential Ethics’, Journal of the History of Philosophy 12: 1–19. Steinmetz, P. (1994) ‘Die Stoa’, in H. Flashar (ed.), Die Philosophie der Antike, Vol. 4: Die hellenistische Philosophie (Basel), 491–716. Stewart, J. A. (1982) Notes on the Nicomachean Ethics of Aristotle (Oxford). Stock, S. G. (1897) Lectures in the Lyceum, or Aristotle’s Ethics for English Readers (London).

300



Stough, C. (1978) ‘Stoic Determinism and Moral Responsibility’, in J. M. Rist (ed.), The Stoics (Berkeley/Los Angeles), 203–31. Talanga, J. (1986) Zukunftsurteile und Fatum (Bonn). Theiler, W. (1946) ‘Tacitus und die antike Schicksalslehre’, in Phyllobolia für P. von der Muehll (Basel), 35–90. Trizio, M. (2012) ‘On the Byzantine Fortune of Eustratios of Nicaea’s Commentary on Books I and VI of the Nicomachean Ethics’, in K. Ierodiakonou and B. Bydén (eds), The Many Faces of Byzantine Philosophy (Athens), 199–224. v. Straaten, M. (1977) ‘Menschliche Freiheit in der stoischen Philosophie’, Gymnasium 84: 501–18. [‘Vincent’] (1826) A New Translation of the Nicomachean Ethics of Aristotle, 2nd edition (Oxford). Voelke, A. J. (1973) L’idée de la volonté dans le stoicisme, Bibliothèque de philosophie contemporaine (Paris). Whitaker, C. W. A. (1996) Aristotle’s De Interpretatione: Contradiction and Dialectic (Oxford). Williams, B. (2010) A History of Freedom (London). Wildberger, J. (2006) Seneca und die Stoa: Der Platz des Menschen in der Welt (Berlin/New York). Wolff, M. (1988) ‘Hipparchus and the Stoic Theory of Motion’, in J. Barnes and M. Mignucci (eds), Matter and Metaphysics (Naples), 471–545. Ziegler, K. (ed.) (1979) Der kleine Pauly: Lexikon der Antike, Vol. 1 (Munich).

Index Locorum Aetius, De Placita Philosophorum 4.21 (Diels, Doxogr. graec. 410–11) 225 n42, 249 Albertus Magnus, [Second] Commentary on the Nicomachean Ethics Bk 3, Tract. I, Ch. XXI 28 108 n53 Alcinous, Didascalicos Ch. 26 25 n14, 32 153.4–5 48 n69 179.8–13 38 179.10–11 32 179.20–23 32 179.31–33 32 Alexander of Aphrodisias, Commentary on Aristotle’s Prior Analytics 162.31–32 34 163.21–29 29 n28 169.6–9 34 270.23–25 34 Alexander of Aphrodisias, Commentary on Aristotle’s Topics 177.19–27 34 Alexander of Aphrodisias, On Fate Chs 2–6 11 Ch. 6 24, 46, 46 n65 Ch. 11 39 n53 Ch. 12 37 Ch. 13 18 n4, 22, 44, 223 n37, 286 n88 Ch. 15 46–47 Ch. 16 46 n67, 214 Ch. 20 47 Ch. 26 35 Ch. 27 47 Ch. 29 46 Ch. 39 23 n9 169.12 39 169.13–15 214 171.11–17 18 174.30–175.4 29 n28 175.6–13 227 n48 176.14–18 199 n19 176.21–22 17 n3 177.7–14 33 n36 178.26–28 41

179.3.11 40 n54 180 25, 39 180.2 36 n43 180.9–12 41 180.26 215 180.26–28 18, 37 n44, 47 n68, 214 180.29–31 46 181.5 18 181.5–6 29–30, 36, 37 n44, 214 181.12–14 214 181.13–182.20 282 n80 181.13–14 37 n44 181.14 40 n54 181.18–21 22 181.21–25 17 n3 182.11–16 22 182.12–13 214 182.16–18 216 182.22–24 43 184.18–19 37 n44 185.7–9 17 n3, 213 n60 188.21 213 n59 189.9–11 43 189.10–11 40 n54 189.10 213 n59 190.1–19 215 191.30–192.14 17 192.11 285 n87 192.18–19 253 192.22 17 192.22–25 213, 213 n60 192.22–24 17, 235 n55 194–95 39 196.10 213 n59 196.24ff 35 196.24–197.3 216 196.27 216 n65 196.29 216 n65 199.8–9 214 199.27–200.7 18, 215 199.29–200.7 36 206.16–18 40 211.1–4 26 n17 211.31–33 214 212 39 224–25 237 n58

302

 

Alexander of Aphrodisias, Mantissa Ch. 22 23 n9, 34, 37, 45, 47 n68 Ch. 23 23 n9, 41 Ch. 25 26 n17 131.5–10 285 n87 169.38 25 170.1 36 n43 171.22–24 37 n45 171.24 36 n43 172.7–9 44 172.10–12 37 n45 173.10–21 47 174.2–7 17 n3 174.4 40 n54 174.9–12 37 n45 174.17–24 46–47 174.32 36 n43 174.33–35 46 174.35–39 47 175.9–32 47 175.23–25 37 n45 175.24–25 40 n54 180.28–31 40 n54 183 25 n14 186 24 Alexander of Aphrodisias, On Mixture 223–24 279 n71 and n73, 285 n87 224–25 218 n9 225.1–2 219 n11 Alexander of Aphrodisias, Quaestiones 3.13 41 Ammonius, Commentary on Aristotle’s de Interpretatione 130.23–33 33 130.30–32 37, 40 130.32–33 40 131.25–32 27 n20 143.1–7 29 n28 148.14.23 41 242.19–20 28 242.24–27 28 242.24–25 25 242.27–28 28 Anonymous, Commentary on Aristotle’s Nicomachean Ethics 122–255 107 n49 149.34–35 24 150.1–4 24 n11 154.17–32 107 n49 Apuleius, De Platone et dogmate eius 1.12 25 n14 Aquinas, Commentary on the Ten Books of [Nicomachean] Ethics

Lectio 11 (Sententia Ethic. lib. 3.l.11 n. 2) 108–09 n54 Aristotle, Prior Analytics 32b8 34 Aristotle, Constitution of the Athenians 43, section 5 87 n17, 99 n13 Aristotle, De Anima Bk 2 Ch. 2 58 n19, 58 n20 Bk 3 Ch. 9 58 n19 427b20 81 n9 Aristotle, De Interpretatione Ch. 13 28 18b9 29 18b31–32 32, 61, 62 18b31 29 18a39–b9 29 19a7–9 61, 62 19a9 29 19a18 29 22b36ff 27 22b36–23a6 27 23a3–4 28, 29, 30, 31 n31 23a15–16 27 Aristotle, Eudemian Ethics Bk 2 Ch. 6 23 Bk 2 Chs 6–11 52 n3 Bk 2 Chs 7–8 56 n13 Bk 2 Ch. 10 23, 59 n24 1222b18–20 55 n11 1222b28–31 55 1222b29 55 n11 1222b41 79 n6 1223a1–9 23 1223a5–9 101 n19 1223a5–8 87 n17, 99 n13 1223a5–6 88 n22 1123a6 88 n22 1223a7–8 30, 88 n22 1224a2–4 65 1224a23–25 56 n16 1224a28–30 55 1224a29–30 55 1224b7–14 56 n16 1225a9–10 88 n22, 101 n19 1225a21 56 1225a27–33 81 n9 1225b35–36 88 n22, 101 n19 1226a1–3 81 n9 1226a21–32 59 1226a27–28 23, 88 n22, 101 n19 1226b3–4 65 1226b17 62 n36

  1226b30–31 30, 32, 88 n22, 101 n19 1226b34–35 66 n46 1227a3–5 62 n36 1227b34–1228a5 64 1227b36–38 64 1227b36–37 57 1128a4–5 65 n44, 69 n52 Aristotle, Nicomachean Ethics Bk 1 Ch. 1 55 Bk 1 Ch. 13 58 n19 Bk 2 Ch. 13 58 n20 Bk 3 Chs 1–5 3 Bk 3 Ch. 1 67 Bk 3 Ch. 2 23, 25, 28 Bk 3 Ch. 3 195 n6, 196 Bk 3 Ch. 5 3–4, 17 n2, 23, 195 n6, 196, 202 n29, 216 n66 Bk 5 Ch. 3 80 n8 Bk 5 Ch. 8 52 n3, 64 n42, 70 n57 Bk 6 Ch. 2 58 n19 Bk 6 Ch. 9 59 n24 Bk 6 Ch. 11 63 n38 Bk 6 Ch. 13 72 Bk 7 Ch. 1 70 n55 Bk 7 Ch. 5 70 n55 1094a4 66 n47 1094a6 66 n47 1094a16 66 n47 1101b13–15 53 1101b14–18 55 n10 1101b14–15 74 n64 1101b30–31 53 1101b31–32 69 n53, 74 n64 1102b13–31 58 n21 1103a3–6 58 n23 1103a8–10 53, 74 n64, 89 n25, 102 n22 1103a20ff 123–124 1105a28–33 51 n2 1105b5–9 70 n57 1106a1–2 53 1106a2–4 63 n40 1106b25–28 53 n7 1106b36–1107a1 53 1108a14–16 53 n7 1109b24 53 1109b30–31 53–54, 66 1109b30 69 1110a17–18 87 n17, 88 n22, 99 n13, 101 n19 1110a24–26 56 1110b2–3 56 1110b9–15 56 n16 1110b28–33 57 1111a24–b3 56 n16

1111a33–34 60 n30 1111b1–2 72 1111b9–10 65 1111b13–15 71 1111b13–14 71 1111b23–24 45 n63 1111b26–28 64 1111b26 45 n63 1111b27 57–58 1112a14–15 66 n46 1112a16–17 63 n39 1112a19–21 60 1112a21–1112b16 59 1112a21–33 23 1112a30–b4 45 1112a31–33 26 1112a31–32 60 n30 1112a31 23 1112a34 23 1112b8–9 60 n28 1112b9 61, 62 1112b10–11 63 1112b22–23 59 1112b24–26 61 n33 1112b27–28 60, 63 1112b27 45 n63 1113a3 62 1113a4 62 1113a6 58 1113a10–11 23 1113a11–12 62 1113a12 62 1113a15 64 1113a18 65 n44 1113a30–34 57 n17 1113a31–33 64 1113b1–2 65 n44 1113b3–1115a3 23 1113b3–21 80 1113b3–14 68, 69 n52, 69 n53 1113b3–6 66 1113b3–4 57–58 1113b3 64 1113b6–1113b14 102 1113b6–14 67, 89, 89 n24, 102, 113, 123 1113b6–7 89, 102 1113b7–14 68 1113b7–11 106 1113b7–8 4, 23, 61, 102 n24, 197 n13 1113b7–8 to b13–14 89–90, 103 1113b8–14 78 n3 1113b8–11 90, 103, 111, 117

303

304

 

Aristotle, Nicomachean Ethics (cont.) 1113b11–14 89, 103 1113b13–14 89, 102 1113b14–1114b25 75 n65 1113b14 89 n23, 102 n20 1113b21–26 69 n53 1113b30–33 62 1113b30 74 1113b32–33 61 1113b32 74 1113b33 74 1114a2–3 61 1114a3–1114b25 17 n2, 75 1114a11–13 72 1114a13–21 75 1114a13–14 75 1114a16–17 61 1114a19–20 61 1114a28–31 74 n64, 75 n66 1114b20–25 75 n66 1114b26–31 54 1114b30–1115a3 75 1115a2–3 88 n22, 101 n19, 126 n123 1117a20–22 65 1125a26 88 n22, 101 n19 1139a19–20 55 n11 1139a11–13 58 1139a21–22 84, 96 1139a31–34 58, 65 1139a31–32 55 n12, 65 1139a33–36 65 n44 1139a33–34 58, 64 1139a33 63 n38 1139a34 64 1139a39 65 n44 1139b3–4 64 1139b4–5 62 n36 1139b5–11 59 1142b24 59 1143a8–9 85, 86 n16, 87 n17, 97, 99 n13 1144a20 65 n44 1147a34 85, 97 Aristotle, Metaphysics 1032b21–4 69 n54 1042b7–8 87 n17, 99 n13 1046a36ff 124 1046b1–2 27 1046b4–7 27 1048a2–3 27 1048a8–9 27 1048a10–15 28 n23 1050b30–34 27

Aristotle Physics 194b29–30 55 n12, 56 195a11–14 56 195a11 55 n12 Aristotle, Rhetoric 1359a36 87 n17, 99 n13 Aristotle, Sophistical Refutations 175b9–10 83, 83 n13, 95 175b13–14 83, 83 n13, 95 176a10–11 83, 83 n13, 95 176a15–16 83, 83 n13, 95 Aristotle, Topics 112b1 34 158a15–17 83, 83 n13, 95 160a33–4 83, 83 n13, 95 177.19–27 34 [Aristotle], De Mundo 6 397b9 285 n87 399bff 285 n87 Aspasius, Commentary on Aristotle’s Nicomachean Ethics 71.25–27 24 74.10–14 24 n11 74.10–13 24 76.11–14 24 76.8–16 107 n44 Augustine of Hippo, De civitate Dei 5.10 196 n9, 199 n19, 262 n30 Augustine of Hippo, De libero arbitrio 201–05 45 n64 Boethius, Commentary on Aristotle’s de Interpretatione 2.197 277 2.203 37 2.208.1–3 33 n36 Burgundio of Pisa, Translation of Aristotle’s Nicomachean Ethics (Ethica uetus) Bk 3, Ch. 6, p. 32.22 108 n50 Calcidius, Commentary on Plato’s Timaeus 142–87 25 n14 151 37, 39 154 38 155–56 26 n16 156 28 n25 160–01 27 n21 Choeroboscus, Prolegomena et scholia in Theodosii Alexandrini canones isagogicos de flexione verborum 85.17–18 100 n14 86.29–31 100 n14

  86.34–35 100 n14 336.25–6 100 n14 Cicero, Academica 2.97 185 n83 2.136 203, 203 n34, 203 n35 2.145 230 n52 Cicero, On Divination 1.125–26 197 1.127 248–49 2.61 232 n54 Cicero, On Fate 7–9 47 11 46 n65 12–14 198, n 18 13 27 n22, 198, 199 n19 18 45 n63 20 223 n37, 232 n54, 238 n63 21 170 n44, 185 n84, 259, 262 n29 21.41–42 197 23 179–81, 262 n29 24 262 n29 27 197, 238 n63 28 185 n84, 238 n63 31 262 n29 33 262 n29 37 185 n84 39–45 261, 285, 286 39–40.1 261 39 268 40–43 200 n22 40 196 n9, 198 n16, 259, 261, 262, 263, 264, 265 n36, 269, 275 n60 41–45 253, 259, 259–71, 261, 284, 290 41–43 47, 261, 267 41–42.1 261, 264, 265, 269 41–42 268, 269, 270, 275, 277, 284, 287 41 196 n9, 259, 260, 261, 261 n28, 262, 262 n31, 262 n32, 263, 263 n33, 265 n36, 268, 274, 287 41.2–42.1 270 42–43 201 n24, 201 n25, 230 n52, 231 n53, 260, 265 n36 42 199 n19, 267 42.1 263, 263 n34, 287 42.2–43 261 42.2–43.1 270 42.2–43 270 42.2 265 n37, 266, 269 42.3 266, 266 n39 43 196 n9, 259, 267, 268, 282 43.1 266, 266 n40 43.2 268, 268 n44, 270 44–45 261, 268

305

44 199 n19, 268 n45, 269, 270, 275, 278, 284, 285 45 262 n30, 269 n47, 270 Cicero, On Ends 1.29–39 145 n68 1.30 137 n32 1.55 145 n68 3.22 207 3.33 221 n22, 258 n17 3.75 203 Cicero, Pro Murena 61 202 Cicero, On the Nature of the Gods 1.39 225 n40, 225 n41 1.69 170 n44 1.70 185 n84 2.37–38 226 n45 2.57–58 243, 243 n84, 243 n85 2.58 244 2.142ff 243 n84 Cicero, Topics 58–59 284 n85 59 277 n62, 278 n66, 283 n84, 284, 284 n85 60–62 269 n46 Clement of Alexandria, Stromata 5.8 285 5.14.89.2 237 n58 6.135.4 41 n56 8.9.95.28–29 273 n54 8.9.95.29–31 275 8.9.95.31–96.2 272, 278 8.9.96.2–5 279, 280 8.9.96.3–5 278 n68 8.9.96.18–97.1 254 n3 8.9.96.23–97.1 219 n15 8.9.98.7–12 289 n98 8.9.98.25–30 289 n97 8.9.100.20–101.3 289 n96 8.9.101.13–102.12 283 8.9.101.14 286 n88 8.9.101.17 273 n53 8.9.101.19–20 272, 272 n52, 278 Dio Chrysostom SVF 3.356 203 n31 SVF 3.365 203 n31 Diogenes of Oenoanda 32.1.14–3.14 181–82 Diogenes Laertius, Lives of Eminent Philosophers 7.25 27 n20 7.32–33 202, 203 n34, 203 n35 7.64 222 n31

306

 

Diogenes Laertius, Lives of Eminent Philosophers (cont.) 7.75 198 7.79 257 n13 7.104 221 n22, 258 n17 7.121 203 7.134 218 n5, 241 n78, 257 n12 7.135 237 n57, 240 n77 7.138–39 242, 242 n83, 246 7.138 225 n42 7.139 218 n5, 218 n6, 249 n94 7.140 218 n2 7.142–43 219 n10, 244 7.142 242, 242 n82 7.143 218 n2 7.147 237 n59 7.149 197, 238 n62, 239 n70 7.150 218 n2, 218 n5, 226 n47 7.156 242 n83, 243 n86 7.175 202 10.33 131 n11, 131 n12, 138 n37 10.117 144 n59, 150 n99 10.120 150 n99 10.133–34 196 n11 10.133 196 n8, 202 n29 10.137 137 n31, 137 n32 10.138 136 n29 10.140 136 n28 10.144 145 n70 10.149 134 n20, 145 n70 Epicurus, Letter to Herodotus 35–36 149 n97 63 147 n85 Epicurus, Letter to Menoeceus 122 144 n60, 144 n62, 145 n66, 145 n72 124–25 150 n99 124 134 n21, 145 n72, 145 n73 127–32 136 n27 128 136 n25 129 136 n26 131 136 n25, 145 n72 132 134, 134 n20, 136 n28, 145 n69, 150 n99 133–34 132, 156, 158–59, 161 n24, 169 n41 133 133, 164 Epicurus, Letter to Pythocles 85 147 n85, 149 n95, 151 n102 Epicurus, On Nature Arr. 34.7, Laursen 1995, 99 159 n18 Arr. 34.10, Laursen 1995, 46–47 141 n54, 147 n85, 148–49, 162 n25 Arr. 35.10, Laursen 1995, 44 174 n54 Arr. 35.10, Laursen 1995, 91 174 n54

Arr. 34.21, Laursen 1997, 19–20, LS 20B (1–4) 141 n53, 164 Arr. 34.21, Laursen 1997, 19, LS 20B (1–3) 131–32 Arr. 34.21ff. 202 n29 Arr. 34.22 191 n97 Arr. 34.22, Laursen 1997, 22, LS 20B (5–7) 139 n47, 140 n51 Arr. 34.22, Laursen 1997, 22 184 Arr. 34.23, Laursen 1997, 23 183 n76 Arr. 34.23, Laursen 1997, 26 183 n76 Arr. 34.23, Laursen 1997, 28 137–38, 139–40, 139 n44, n45, and n46, 140, 167 n38, 182 Arr. 34.24 161 n24 Arr. 34.24, Laursen 1997, 28, LS 20C 130 n3, 139 n44, 139 n45 Arr. 34.24–25, Laursen 1997, 28–29 138 n41 Arr. 34.24–25, Laursen 1997, 29–31 160 Arr. 34.24–25, Laursen 1997, 29–30 190 n94 Arr. 34.25, Laursen 1997, 29 164, 183 n76 Arr. 34.25, Laursen 1997, 31 131, 134 n19 Arr. 34.26 183 n78, 184 n82, 190, 196 n8 Arr. 34.26, Laursen 1997, 32–33, LS 20C(1) 133 Arr. 34.26, Laursen 1997, 32–33 Arr. 34.26, Laursen 1997, 33, LS 20C(1) 130 n4, 163 n27, 166 n34, 177 n66 Arr. 34.26, Laursen 1997, 33 LS 20C(2) 161 Arr. 34.26, Laursen 1997, 33 160 n21 Arr. 34.27–30, Laursen 1997, 35–42, LS 20C (2–15) 161 Arr. 34.27–28, Laursen 1997, 35–37 163 Arr. 34.27, Laursen 1997, 35, LS 20C(2) 130, 162, 177 Arr. 34.27, Laursen 1997, 35, LS 20C(3) 161 Arr. 34.28, Laursen 1997, 36, LS 20C(4) 131 Arr. 34.28, Laursen 1997, 36, LS 20C(5) 161 Arr. 34.28, Laursen 1997, 37, LS 20C(5) 162 Arr. 34.28, Laursen 1997, 37, LS 20C(8) 130–31 Arr. 34.28, Laursen 1997, 37 164 Arr. 34.29, Laursen 1997, 39, LS 20C(10) 145 n71 Arr. 34.29, Laursen 1997, 39 165 Arr. 34.30, Laursen 1997, 40–42 166 Arr. 34.30, Laursen 1997, 41, LS 20C(13) 161, 162 n25 Arr. 34.30, Laursen 1997, 42 167 n37 Arr. 34.30–1, Laursen 1997, 42–45 138 n41 Arr. 34.31, Laursen 1997, 43–45 139 n46, and n47, 146 Arr. 34.31, Laursen 1997, 43–44 134–35 n21

  Arr. 34.31, Laursen 1997, 43 147 n80, n81, and n84, 162 n25, 177 n64 Arr. 34.31, Laursen 1997, 44–45 141, 166 n34, 185 n83 Arr. 34.31, Laursen 1997, 45 130 n2, 145 n71, 147 n79, n80, and n82, 148 n90 Arr. 34.32, Laursen 1997, 47 130 n1 Arr. 34.33, Laursen 1997, 48 177 n65 Arr. 34.27, Laursen 1997, 51–52 182 LS 20B(5) 166 n35 LS 20C(1) 139 n47, 148 n89 LS 20C(2) 164 LS 20C(3) 164 LS 20C(3) 130 n6 LS 20C(8) 164 LS 20C(10) 162 Epicurus, Vatican Sayings 11–13 151 n102 15 145 n70 29–30 145 n70 30 134 n20 40 147 n85 77 156 n13 80 138 n43 Epictetus, Discourses 1.1 196 n9 1.1.7 206 n43 1.1.12 206 n43 1.1.17 206 n41 1.1.22–24 206 n44 1.1.23 206 n41 1.1.22–23 206 n43 1.6.40 206 n43, 208 n47 1.12.9 210 1.12.15–17 209 1.12.25 209 1.12.32–35 196, 208 1.12.34 38, 206 n43 1.22.10 206 n43 1.25.2 43 n59, 208 n48 2.1.21 210 n50 2.1.24 210 2.2.13 210 2.5.8 206 n43 2.13.2 208 2.13.10–11 206 n41 2.16.6–11 208 2.19 27 n20 2.19.32–34 206 n41 2.19.32 206 n43 2.23 41 2.23.42 209 3.5.7 210 n50

307

3.10.10–11 207 n46 3.11.15 207 n46 3.15.12 210 n50 3.22.95 209 3.24.3 206 n43 3.24.23 206 n41 3.24.69 206 n43 3.24.101–02 207 n46 3.24.108 206 n43 4.1 210 4.1.1 210 4.1.27–28 210 n50 4.1.62 43 n59, 208 n48 4.1.68–75 206 n43 4.1.68–73 205 4.1.68 43 n59, 208 n48 4.1.131 209 4.4.34 209 4.4.44 208 4.7.9 211 n51 4.7.16 43 n59, 208 n48 4.8.1–4 208 4.12.8 43 n59, 208 n48 Epictetus, Handbook 1 206 n43 2 206 n43 2.48.3 207 n46 5 206 n43 29 210 n50 48.2 208 53.1 209 Epiphanius, Adversus Haereses 3.36 (Diels, Doxogr. graec. 692.25–6) 204 n39 Eusebius, Praeparatio Envangelica 6.8.2 159 n20 6.8.6 159 n20 6.8.23 159 n20 6.8.29 166 n33 6.8.30 159 n20, 166 n33 6.8.32 159 n20 6.8.34 159 n19, 166 n33 6.8.38 166 n33 15.14.1 218 n7 15.14.2 239 n69 321.1 196 n9 323.1 196 n9 323.10–11 236 n56 323.11–12 239 n70 324.3–5 238 n63 325.26–28 198 n17 326.21–23 197 n13 326.22 196 n9 327.4–5 199 n20 327.9 196 n9

308

 

Eusebius, Praeparatio Envangelica (cont.) 334.13–14 213 n59 336 213 n59 337 213 n59 340 213 n59 342 213 n59 Eustratius, Orationes Oration 3, 86.23–6 100 n15

Grosseteste, Robert, Translation of Aristotle’s Nicomachean Ethics Bk 3, Ch. 7, p. 187.23 108 n51 [Heliodorus of Prusa], In Ethica Nicomachea paraphrasis 50.10–11 107

The Fez Manuscript 207.1–2 110 n59 206 110–11 n61 207.2–3 111–12 n62

Hierocles 1.5ff 245, 247 1.34–39 248 n91 1.51–57 248 n91 2.1–9 248 n91 4.38–53 248 n91

Galen, De alimentorum facultatibus 6.455–456 K (fr. 112 Wellmann) 255

Herodian 5.2.2 159 n16

Galen, On the Doctrines of Hippocrates and Plato 4.2.12 225 n43 4.2.14–18 230 n52 4.2.16–18 225 n44 4.4.35–6 232 n54 Galen, De Causis Contentivus 2 273 n56 2.1–4 280 1.1–5 279 n71 1.3 280 n76 1.5 280 Galen, Synopsis Librorum Suorum, Sexdecim, de Pulsibus 9.458.8–14 279, 279 n74 Galen, De Causis Pulsuum 9.2–3 273 n56 Galen, De Plenitudine 3.7.535K 279 n71 3.7.525.10–14K 285 n87 3.7.526–27K 285 n87 [Galen], Medical Definitions 19.392 273 n56 Gellius, Aulus, Attic Nights 7.2.3 237, 237 n60, 238 n63, 239 n69, and n70, 249 n90 7.2.4–14 212 n56 7.2.4–13 196 n11 7.2.5 198, 198 n16 7.2.7–14 275 n60 7.2.7–10 201 n27 7.2.7–9 201 n23 7.2.11 201 n25, 230 n52, 266 n41, 267 n43, 282 7.2.15 196 n9 14.1.23 39 n52

Hippolytus of Rome, Refutation of all heresies 19.19 (Doxogr. Graec. 569.19–22) 38 n49 Ibn Rushd, 2nd Juntine 19.18–26 113–114 n71 Joannes XI Beccus, Four Books to Constantine Meliteniotes 388.16–18 100 n15 Josephus, Antiquities of the Jews 13.172 24 n11 Justin, Apology 1.43 213 n59 1.44 38 n49 Lactantius SVF 1.160 237 n57 Lucretius, On the nature of things 1.737 173 n52 1.923 173 n52 2.243 189 n93 2.251–93 168ff 2.251–93 168 2.251–60 168 2.261–823 168 2.284–93 168 2.251–60 169 2.251–57 170 n44 2.252 170 n44 2.255 170 2.257–58 170, 173 n53 2.258 171 n46 2.259–60 170 2.259 173, 178 2.261–83 171 2.261–62 173 2.264–93 202 n29 2.265 171 2.266–68 172 2.268 171, 172

  2.269–70 171 2.269 173 2.270 170 n43, 171 2.271 171 2.273 171 2.275 171, 172 n49 2.278 171, 172 n49 2.279–83 174 2.279–82 174–75 2.284–93 174 2.284–87 174 2.284 175 n61 2.288–89 176 n61 2.288–2.89 175 2.288–93 175, 177, 179 2.288 175 n61 2.289 176 3.136–37 155 n11 3.139 170 n43 3.159–60 172 n48 3.289–322 145 n67 3.289–307 137 n33, 144 n63 3.307–22 143, 144 n64, 183 3.309 143 3.320–22 144 n61 3.321 150 n99 3.344–47 136 n30 3.445–50 137, 138 n38, 138 n40 3.510–16 149 n94, 149 n96 3.845–46 155 n11 4.53 173 n52 4.877–91 168ff, 171, 172, 174 n54 4.877–78 172 n49 4.884–85 173 4.886 171, 173 4.887–88 171 4.890–91 172 4.1223–26 191 n96 5.849–54 188 n92 5.882 173 n52 5.1028ff 138 n36 5.1028ff 138 n39 5.1107 173 n52 Manuel II Palaeologus, Dialogi cum mahometano, dialogue 17, in Piissimi et sapientissimi imperatoris Manuelis Palaeologi 221.29–31 100 n15 Marcus Aurelius, Meditation 5.8 240 n76, 256 n11 5.8.1 257 6.32 196 n9

309

8.27 240 n76, 256 n11 9.29 240 n76, 256 n11 Maximus of Tyre 41.5a 38 n49, 41–42 41.5g 41–42 Nemesius, On the nature of man 21.6–7 220 n17 36.26–37.1 213 n59 52.18–19 218 n8 102 44 103–04 26 n16, 34 103.20–21 27 n18 104–20 196 n7 104.1–2 28 n25 104.2–4 28 n26 104.4–5 28 n27 104.6–7 29 104.18–21 39 n52 105–06 18 n4, 22, 286 n88 105.18–21 17 n3, 213 n60, 235 n55 105.24 213 n59 106.15–20 239 n68 108.15–17 197 109 38–39 110.5–9 38–39 110.7–9 38 n49 110.125–26 25 n14 111.25–112.3 249 n93 112.7 213 n59 112.10 41 112.13–15 26 n17 114.15–16 44 114.19–22 28 n27 114.21–24 27 n19, 29, 216 n64 114.24–115.2 29 115.22–28 40 115.22–27 37 115.25 41 115.27–28 40 116.3–5 40 119.4–5.11 41 n56 119.11 41 The New Testament 2 Corinthians 1:17 122 James 5:12 122 Olympiodorus, Commentary on Plato’s Phaedo 2.14 243 n86 Origen, Against Celsus 6.71.5–7 (Borret) 241 n78, 257 n12 Origen, Commentary on the Gospel of John 21.10 203 n32

310

 

Origen, De Oratione 2.368 221 n22, 258 n17

Plutarch, Quaestiones convivales 740c 38 n50

Origen, De Principiis 3.1.4 275 3.1.5 213 n59

Plutarch, Quomodo adulescens poetas audire debeat 33d 202

[Heliodorus of Prusa], In Ethica Nicomachea paraphrasis 50.8–16 107 50.10–11 107 52.25–27 25 52.25–27 37

Plutarch, On Stoic self-contradictions Ch. 47 253 1037f 85, 97, 200 n22 1043b 203 1045b–c 229 1045c 230, 258 n16 1049f–1050d 223 n38 1049f–1050a 224–25 n39 1049f 236 n56 1050a 224–25 n39 1050b–c 182 n73 1050c 224–25 n39 1050c–d 221 n22, 258 n17 1052c 219 n11 1053b 219 n11 1053f 279 n71, 279 n73 1054a 218 n6, 220 n20 1055d–f 198 n18 1055e 27 n22, 236 n56, 238 n67, 240 n75, 256 n10 1055f–1056d 271–74, 287 1055f 274 1056a–b 271 1056b–c 271 1056b 240 n75, 256 n10 1056c–d 271 1056c 197 n14, 221 n22, 224–25 n39, 236 n56, 240 n75, 256 n10, 258 n17, 271, 273, 273 n57 1056d 204 n39, 224–25 n39, 226 n46, 246, 271 1057a–c 200 n22 1057a–b. 271 n49, 274 n58 1057b 265 n38

Philo of Alexandria, De Aeternitate Mundi 19, p. 505M 245 Philo of Alexandria, Quod omnis probus liber sit 97 202 SVF 3.359–64 203 n31 Philo, De sobrietate, 3.603 203 n31 Philodemus, Index Stoicorum Col. 53.7 222 n28 Philodemus, De Ira 32 135 n23 Philodemus, De Signis 36 159 n20 36.11–17 187 n90 36.11–17 190 n95 36.14 196 n8 Philodemus, De Pietate Col. 11.32–col. 12.1 225 n40, 225 n41 Col. 11.31–col. 12.2 236 n56 Plato, Republic 617E 38 Plotinus, Enneads 2.1.2 236.30–1 239 n72 3.1.2.17–22 240 n75, 256 n11 3.1.3 187 n90 3.1.7 221 n22, 258 n17 3.1.9–10 44 3.2.10 44 6.8.4–6 213 n59 6.8.7 44 Plutarch, Against the Stoics on Common Conceptions 1054a 221 n26 1075e 237 n59 1076e 197 n14, 221 n22, 224–25 n39, 236 n56, 258 n17 1085b 220 n17 1085c–d 218 n8

Plutarch, De sollertia animalium 964e 187 n91, 190 n95 [Plutarch], Epitome of the Opinions of the Philosophers 1.7 (Diels, Doxogr. graec. 306.1–6) 237 n58 1.12 (Diels, Doxogr. graec. 310.6–7) 241 n80, 257 n14 1.27 199 n20 1.27.3 159 n19 1.27.4 (Diels, Doxogr. graec. 322.11–12) 159 n20, 239 n73 4.12.1–3 (Diels, Doxogr. graec. 401–2) 265 n38

  [Plutarch], De Fato 570f–572f 26 n16 571b–d 34 571b 27 n18 571c–d 27 n19, 28 n27 571c 28 n25, 28 n26, 29 n28 571d–e 44 571d 37 574e–f 35 n41 574e 27 n21, 197, 239 n73, 259 n22, 262 n29 Porphyry, Letter to Marcella 31.34 8P = 239, Arrighetti 1973, 567 145 n70 Sextus Empiricus, Against the Mathematicians 5.46–48 159 n17 5.46 159 n17 7.40–41 247 7.241 265 n38 8.263 220 n17 9.75–76 218 n6 9.75 218 n5, 221 n26, 240 n76 9.207 219 n16, 255, 255 n4 9.211 219 n13, 219 n15, 254 n3, 255 n4 9.236–37 272, 272 n52 9.238 289 n95 9.239 255 n4, 289 n97, 289 n98 9.242 272, 272 n52, 289 n95 9.250–51 289 n96 9.332 218 n2 10.218 221 n23 10.52 221 n24 11.169 151 n102 Sextus Empiricus, Outlines of Pyrrhonism 2.81 257 n13 3.14 220 n21, 258 3.25 219 n16, 255 n4 3.218 237 n58 Seneca, On Benefits 4.7 240 n76 Seneca, Letters 52.3–4 144 n65, 147 n86 52.4. 150 n101 65.2 218 n5, 218 n6, 256 65.4 220 n21, 258 113.18 249 121.6–15 248 n91 Seneca, Natural Questions 2.45 241 n79, 256 n11 Seneca, On the Happy Life 15.7 211 n51

311

Simplicius, Commentary on Aristotle’s Categories 212–13 221 n22, 258 n17, 282 n80 306.14–15 220 n18 Simplicius, Commentary on the Encheiridion 20 41 Stobaeus, Eclogues 1.4–5 159 1.31.13–14 240 n76, 256 n11 1.78.4–6 239 n69 1.78.18–20 237 n57 1.79 197 1.79.1ff 197 n14 1.79.1–12 237–38, 238 n61, 238 n67 1.79.1–2 237 n58, 256, 257 n12 1.79.5–7 236 n56 1.79.5–12 256, 257 n12 1.79.5–10 240, 240 n75 1.79.15–16 238 n63 1.79.17–18 238 n62 1.89.2–5 24 n11 1.89.2–5 26 n17 1.106.8–9 222 n30 1.106.20–23 222 n32 1.132.27–133.5 240 n77 1.133.3–5 237 n57 1.138–39 255 1.138.5–16 219 n13, 254 n2, 258 n21 1.138.15–16 219 n15 1.138.15 223 n35 1.138.23–139.4 253 1.138.23–139.2 219 n13, 254, 254 n2 1.139.3–4 241 n81, 255, 256–57 1.139.7–8 219 n13 1.139.7–8 219 n15 1.165–66 222 n28 1.165.15–17 221 n24 1.165.17–18 221 n25 1.165.19–21 221 n27 1.166–67 222 n28 1.166.24–167.14 221 n22, 258 n17 1.166.24–26 221 n25 1.166.26–27 222 n29 1.167.9–14 222 n30 1.368 225 n42 2.73.1 221 n22, 258 n17 2.76.13–15 207 2.82.11–17 221 n22, 258 n17 2.86–8 200 n22 2.95.6–8 221 n22, 258 n17

312

 

Stobaeus, Eclogues (cont.) 2.101.15–20 211 n54 2.159.25–160.11 206 n43 2.163–73 196 n7 2.164 38 n49 2.173.21 41 4.44 209 n49 4.60 209 n49 Summa Alexandrinorum LII–LIII 112 n70 Syrianus, Commentary on Aristotle’s Metaphysics SVF 2.308 240 n76

Tacitus, Annals 6.22 38 n49, 38 n50, 238 n65 Tatian, Oratio ad Graecos 7 213 n58 Tertullian, Apology 21 237 n57, 238 n66 Tertullian, De Anima 21.6 213 n59 William of Moerbeke (attrib), Translation of Aristotle’s Nicomachean Ethics Bk 3, Ch. 7, p. 418.10 108 n52

General Index accountable for (hupeuthunos) 196 act ‘out of character’ 135 action (praxis) 23, 24, 36, 37, 40, 55, 56, 66, 67, 69, 79, 200 bad actions 160 goal-directed action 68 good or bad actions 69 human action 65 reflex actions 65 refraining from action 56, 79, 206 shameful-and-bad action 68 spur-of-the-moment acts 65 voluntary action 188 actions vs affections 286 active principle (archē poioun) 11, 218, 219, 220, 226, 228, 236, 237, 238, 241, 254, 266 actuality (energeia) 55 n12, 63 n37 adiuvans 283, 284 adolescence 138 adult human being 132 advances in age 182, 183 aether 219, 242, 244, 246, 249 affirmation 79 age 190 agent/agency ability to do otherwise 129 agent’s awareness 61, 62 agent’s disposition 18 autonomy 155, 156, 167, 169, 182, 187, 193 causal responsibility of 129 causality 229 decision of 129 goodness or badness 69 rationality of 151 strong-willed 71 vicious agent (phaulos) 53, 64, 65, 72 virtuous agent (spoudaios) 53, 64, 65, 70, 71 weak-willed agent 71 with not yet fully developed character 72, 81 whole person model of agency 155, 156, 176, 177, 178, 179, 183, 184 aiming at the target and striking it 207 aitia 255 aitia, reason, logos 256 aitia vs aition 256, 257, 258 Albert of Saxony 109

Albertus Magnus 108, 115 Alcinous 25, 32, 38 Aldine editio princeps of Aristotle’s works 114, 116, 117 Alexander of Aphrodisias 2, 11, 17, 18, 19, 21, 22, 28, 34, 35, 36, 37, 39, 41, 42, 43, 44, 46, 47, 48, 50, 107, 212, 213, 215 Mantissa 44, 45, 46 On Fate 41, 46, 212 Quaestiones 41 Al-Farabi 112 alternative courses of actions 214 alternative options 61, 62 Ammonius 28, 33, 37, 40, 41 Andronicus of Rhodes 120 anger 72 animals 134, 188, 191 behaviour 65 non-rational 218 animus 170 Anonymous Commentator on Books 2 to 5 of the Nichomachean Ethics 24, 107 Antipater of Tyre 207, 222 n28, 242 antiqua traductio 114 apatheia 210 Apollodorus 221 n25, 222 apotelesma 255 appetite 71, 72 Aquinas, Thomas 108, 109 n54, 115 Arabic 111 Arguments 150 Argyropoulos, John 114, 115 Aristotelian syllogistic 59 Aristotle 2, 3, 4, 5, 23, 25, 26, 27, 37, 39, 77–92 passim, 93, 94, 97, 98, 99, 101, 102, 103, 104, 152, 153, 193, 196, 197, 214 dialectic of 83, 89, 103 Eudemian Ethics 23, 33, 41, 52 n3, 64 De Interpretatione 9, 29, 30, 31, 32, 33 n36, 61, 62 Commentaries on 49 Metaphysics 124 Nicomachean Ethics 3, 4, 23, 33, 41, 45, 93–127 passim, 195 modern reception of 114–18

314

 

Aristotle (cont.) purpose of 3.1–5 51, 52 reception of 104–27 Arrian 204 Aspasius 24, 30 n30, 107 assent 8, 37, 38, 200, 201, 205, 206, 209, 214, 265, 268, 271, 282 assumption 143 atomism 5, 6, 7, 161, 187, 188 mechanistic 193 atomistic determinism 193 atoms 131, 132, 134, 138, 141, 162, 169, 177, 178, 179, 187, 189 atomic level, the 184, 188 atomic structures 149, 188 collision of 188 fire-like; air-like; breath-like 137 n33, 144, 147 types of atomic motions 178, 180, 189 collision 178, 180 necessary downward movement 178 natural downwards movement 192 weight 178, 180 weight of 188 attributes (sumbebēkota) 222 Augustine 11, 196 n9 autexousios 205, 208 authority 42 autonomy (of the agent) 15, 16, 21, 40, 48 autotelēs aitia 275 Balbus 244 Bardesanes 213 n59 because of us (par’hēmas) 132, 133, 134, 158, 164, 166, 167, 168, 191 beginning (archē) of action 37, 47 beginning of motion 172, 173, 174 behaviour changing one’s behaviour 148 being master of (kurios + genitive) 61 being qualified (poion) 279, 282 Bekker 116 edition of the Nicomachean Ethics 119 beliefs 6, 184, 192 changes in 185 firm beliefs 192 value beliefs 184 beliefs (dogmata) 208 beliefs (doxai) 133, 134, 135, 136, 145, 146, 160, 164, 166 adopted on rational grounds 134 changing one’s beliefs 146, 148 empty beliefs 148 false beliefs 145 firmness of beliefs 149

new beliefs 149, 150 prejudicial and irrational beliefs 151 true beliefs 145, 148 strengthening of 145 blame 130, 164 blameworthy 72, 131, 132 blending (krasis) 218 Boethius 37, 40 Bruni, Leornardo 114, 115 built-in time release 190 Burgundio of Pisa 108 Buridan, John 109 Burley, Walter 109 Burnet, J. 123 Byzantine Anonymous commentary on Books 2–5 of the Nicomachean Ethics 107 Byzantine Anonymous paraphrase of the Nicomachean Ethics 107 Byzantine Greek 100 Calcidius 25, 37, 39 Camerarius, Joachim 117 canon 148 of Epicurean philosophy 149 capability 199 capacities of opposites (antikeimena) 27, 30, 31 capacity (dunamis) 27 Carneades 180, 197 causal chain 74, 153, 239, 240 causal factors 161 external 153, 267 internal 153, 267, 281, 288, 289 nature of the object 286, 288 causal laws, empirical 234 causal network 17, 49, 201, 209 causal predetermination 198, 215 of changes 221 causal responsibility 6, 7, 21, 74, 128, 129, 130, 133, 135, 136, 140, 150, 158, 162, 163, 164, 166, 176, 201 causal simultaneous determination of states 221 causally undetermined decision 43, 81, 91 causation 6, 33, 156, 219, 220, 253–90 passim downwards 184 efficient causation 10, 55, 56, 58, 65, 69 n54, 228 mechanistic 241, 244 Stoic accounts of 219 causes 253–90 passim, 288 active causes 220, 234, 236, 258 antecedent causes 10, 11, 221, 259, 260, 262, 263, 264, 265, 267, 268, 270, 271, 272, 277, 281–82, 285, 290

  auxiliary and proximate antecedent cause 287 auxiliary antecedent cause 248 Aristotelian formal, material, and final causes 220, 258 bodies 254 Cause (Aitia) 9, 240, 241 cause and effect in different ontological categories 219 relative (pros ti) 219 causes of change 11, 290 causes of states 11, 283, 290 classification of causes 277 cohesive (continens) cause 268, 269, 272 cohesive (sunektikon) cause 220, 246, 279, 288 cohesive 280, 281, 283 cooperation of 273, 277, 279, 287 cooperating causes 260, 274, 290 corporeal causes 255 efficient causes (causae efficiente) 278 n66 external causes 281 helping cause (sunergon aition) 279, 283 identification of cohesive and self-sufficient cause 289 individual cause (aition) 241 necessitating causes 268, 269, 270, 290 ourselves as causes 149 perfect and principal causes 260, 262, 263, 264, 267, 268, 270, 274, 275, 280 perfect and principal antecedent cause 286, 287 preceding causes, see ‘cause, antecedent causes’ principal causes 248, 265 proximate and auxiliary causes 260, 262, 263, 264, 268, 269, 270, 276, 283, 285 procatarctic (prokatarktikos) causes 271, 272, 273, 273–74, 275, 285, 288 proximate causes 265, 266 relative (pros ti) causes 259, 289 self-sufficient cause (autotelēs aition) 10, 271, 271–72, 272, 273, 274, 275, 276, 278, 280, 282 self-sufficient antecedent cause 287 self-sufficient causes of change 285–89 taxonomies of causes 253 that because of which (di’ho) 254, 255, 258, 267, 288 Chaldaeans 39 n52 Chance/chance events 59, 132, 187, 188, 189, 221, 222, 223, 225, 226, 227, 228, 248 by chance (apo tuchēs) 158, 191 change (kinēsis) 220, 259 changing one’s nature 184

315

character 18, 22, 38, 155, 215 determination 47 development 178, 182–84, 185, 186, 187, 192 non-necessitated 189 formation, see ‘character, development’ moral 33, 271 self-modification of one’s character 140, 141, 142, 184 character (ēthē) 58 character disposition 184, 192 character dispositions (ēthikai hexeis) 58, 64, 68, 75 voluntary 75 characteristic movements 230 Chase, D. P. 119 children 138, 139 small 134 chocolate 200 Choeroboscus, Georgius 88 n19, 100 choice/choices 36, 39, 40, 58, 65, 66 future choices 204 human choice 235 of good and bad 50 reversal of choice 75 undetermined choice 207, 208 choice (prohairesis) 57, 59, 63, 78 undetermined 78 Christianity 5 Christian authors 11, 122 early Christian writers and texts 2, 39 n52, 49, 125 Christian philosophy 12, 38, 48, 50 Christian tradition 125 Chrysippus 10, 11, 22, 35, 47, 85, 194, 196, 198, 199, 200, 201, 202, 203, 204, 205, 206, 208, 210, 211, 213, 214, 216, 217, 218, 224–36 passim, 236–52 passim, 242, 253–90 passim Chrysippean compatibilism 209 Chrysippus’ Specified Causal Principle 235 On Nature 224–36 passim On Providence 242 theory of fate 271 Cicero 179–81, 185, 189, 197, 203 n36, 277 On Fate 179–81, 259–71 passim Stoic Paradoxes 203 n36 circumstantial external influences 192 Cleanthes 197, 202, 209 climate 288 coercion, absence of 193 coherence 246 cohesion (hexis) 218, 242 of objects 218

316

 

cohesive (sunektikos) 276, 278 command 84, 85 commentators/commentaries of the Nicomachean Ethics 11 Arabic commentaries 112–14 Ancient commentaries 107 Byzantine commentaries 107 Latin commentaries 116 medieval Latin texts 107, 108 nineteenth-century commentaries 120 common nature, see ‘Universal nature’ compatibilism 215, 260 compatibilist determinist 167 compatibility of causal determinism and that which depends on us 213 compulsion 15, 56, 162, 165, 167, 180 external and internal 175 of the mind 179 cone 266 conflagration 245, 246, 247 consciousness 248 continens 284 contingency 27, 199 contingent, the 28, 29, 30, 34 continuity (monē) 221 continuity/coherence (sunochē) 222 continuum theory 9, 218, 227, 234 cosmic perspective: see global perspective could have done otherwise 15, 78, 136, 207 counterfactual reasoning 234, 235 counterfactual situations 213 counter-natural changes 225, 226, 227 counter-natural motions 286 counter-natural states 225, 226 criterion 148 culpability 50, 158 cultural prejudice 139, 141 cylinder analogy 10, 11, 260, 266, 267, 274, 278, 281, 282, 283, 287 Damascius 12 death 139, 145 decision 178 decision maker 157, 158, 160 decision-making faculty 48, 154 decision-making faculty model 156, 176 deliberate choice 23, 24, 25, 28, 34, 37, 49 deliberation (bouleusis) 39, 58, 59, 61, 62, 63, 66, 81 Democritus 12, 180, 181, 187 depend on us 195, 198, 209 desires 56, 146 changing one’s desires 150 conflicting desires 71, 73

deliberated desires 57, 62 desires (orexeis) 58 desire (prothumia) 133 goal-directed desire 72 non-rational desires 56 rational desires 56 destiny (peprōmenē) 209, 238 determinism 212, 217–252 passim causal determinism 9, 10, 11, 16, 19, 25, 33, 43, 48, 79, 152, 153, 156, 168, 169, 200, 215, 219, 228–36 passim, 270, 285 causal fate-determinism 199, 201, 209, 214, 215, 216 incompatibility with free will 153 mechanical causal determinism 236 modern theories of 227, 232 universal causal determinism 17, 18, 49, 197, 233 logical determinism 33 universal determinism 226, 227 Determinists 215 development 245 of animals 247 developmental patterns 188 developmental ‘potentials’ (spermata) 190 dialecticians 27 Diocles of Carystus 255 Diogenes Laertius 106 n34 Diogenes of Oenoanda 181–82 Diogenianus 159, 196 n9 dispositions (diatheseis) 227 dispositions 22, 38, 47, 137, 138, 144 initial dispositions 160, 184 mental dispositions 134, 135, 141, 143, 155, 180, 183, 190, 192 changing one’s mental dispositions 143, 148, 149, 184 overall mental disposition 135, 148, 155, 176 practical rationality as a mental disposition 251 moral dispositions 53, 74 voluntariness of 75 natural dispositions 72 vicious disposition 53, 65, 74 virtuous disposition 53, 64, 65, 67, 74 divine action 249 divine awareness 248, 249, 250 divine intention (hormē) 248, 249 divine perception 249 divine providence 11, 48, 49 divine rationality 242 divine reason 250 divine soul 251 DNA 246, 247, 249

  do act / don’t act 84–86 doing what is one’s own (autopragia) 203 drunk 73, 76 education 139, 183 effects 223 immaterial, incorporeal 219, 220, 255 eidola 177 einai 158 ekpurōsis 245 Eleutheria (freedom), see ’Freedom (eleutheria)’ Eleutheros, the 203, 210 embryo 247 emotions 145, 146 changing one’s emotions 150 emotional tendencies 137 moral emotions 155 end (telos) 55, 58, 63, 136, 145, 205, 207 environment (ta periechonta) 130, 133, 138, 139, 141, 146, 150, 162, 177 environmental impact 147 changing one’s environment 148 eph’hēmin 8, 15, 19, 34, 40, 43, 77, 80, 81, 88, 89, 90, 157, 195, 196, 206 indeterminist concept of 43 not a technical term 197 that which depends on us (to eph’hēmin) 8, 17, 27, 28, 32, 35, 196, 197, 204, 208, 210, 211, 212, 213, 214 things that depend on us (ta eph’hēmin) 8, 201 things that do not depend on us (ta ouk eph’hēmin) 206 two-sided 206 n41 up to us (eph’hēmin) 81, 90, 157, 167, 168 Epictetus 2, 8, 37, 38, 39, 41, 194, 196, 204–11 passim, 212 Epicureans 5, 128–151 passim later Epicureans 157, 159 n17 Epicurean ethics 6, 185 forward-looking 151 Epicurus 5, 6, 21 n7, 128–151 passim, 152–192 passim, 196 n11 On Nature 25, 156 epistemic indistinguishability 231, 233 eternal (i.e. unchanging) things 59 eternity 248 of the universe 189 ethics 202, 203, 205, 208, 216 function of 150–51 Eustratius of Nicaea 100, 107 events (gignomena) 223 exousia, see ‘power (exousia)’ external and internal factors 270 external circumstances 41, 199, 201, 215, 246

317

external factors 70, 134 external force (or hindrances) 56, 198, 205, 206 absence of 92 external influences 180 external stimuli 143 facts 223 faculty of the will, see ‘will, faculty of ’ fatalism 21 n7, 212 fate (heimarmenē) 9, 10, 11, 17, 24, 26, 35, 38, 39, 132, 171, 197, 198, 200, 201, 209, 213, 217, 224, 236, 236–52 passim, 238, 249, 261, 271, 272, 274, 277, 285 blaming fate 251 hypothetical fate 38, 48 unfolding of through time 248 Felicianus, Johannes Bernardus 117 ‘for the most part’ 28, 29, 34, 181 force, by 165 forced (coactu) 171 motion 172 foreseeable consequences 74 free action 7 free will 1, 3, 4, 5, 11, 16, 19, 22, 47, 48, 63, 92, 193, 196, 211, 212 incompatibility with causal determinism 153 free-will problem 2, 7, 10, 16, 47, 48, 152, 153, 154, 156, 168, 194, 202 modern problem of free-will and determinism 195 Freedom 8, 15, 156, 195 ambiguity in the English language 212 freedom (eleutheria) 8, 15, 195, 202–4 passim, 210–11, 212, 213 freedom from determination by (external and) certain internal causal factors 14 freedom from determination by external causal factors 14, 21 freedom from force and compulsion 14, 21, 24, 49 freedom from predetermination 1 freedom of action 172 n49 freedom of choice 3, 4, 7, 77–92 passim, 91, 92, 94, 153, 168, 179, 186, 193 freedom of decision 13, 39, 41, 43, 46, 50, 142, 153, 154, 168, 179, 187, 193, 216 freedom of the sage 194, 195 freedom of the will 13, 153 freedom to do otherwise 13, 20, 27, 31, 40, 43, 44, 46, 48, 74, 167 metaphysical freedom 195, 214, 215 political freedom (liberty) 195 un-predeterminist freedom 14, 35, 36, 48, 49

318

 

future 205, 206, 243 future contingents 30, 33 future events 31, 197, 238, 239 future propositions 29, 31, 49 Galen 279 Garve, Christian 118 General Causal Principle 232, 233, 235 generation 245 Gerald of Odo 109 gignesthai 158 Gillies, John 118 global perspective 225, 226, 227, 240 Gnosticism 39 n52 god 219, 237, 241 god’s plan 207, 241 gods 159 good-and-noble action 68 Gorgias 2 Grant, A. 121 Grosseteste, Robert 108, 121 Growth 245 happiness (eudaimōnia) 136, 151, 205 Heliodorus of Prusa 25, 40, 107 Hellenistic philosophy 5–11, 158, 188 hereditary and environmental factors 6, 139, 161, 162, 191 hereditary development 179 hereditary internal influences 192 Hermachus 145 Hermannus Alemannus 112, 114 Hierocles 247 hormetikē phantasia 243 human agency 243, 244 human being 191 human nature 56 human self-determination 196 hylomorphism 12 Iamblichus 12, 41 Ibn Rushd 112, 113, 114, 116, 117, 122 Middle Commentary of the Nicomachean Ethics 112–14, 116, 117, 122 Ibn Sina 112 Ibn Zur‘ah (translator) 112 ignorance 57, 71, 72, 73 images 173 impossibility 198 impressions 37, 38, 201, 214, 265, 266, 269, 271, 274, 275, 282 caused in ordinary people by the sages 271 false impressions 271

impulsory impression, see ‘impulse-prompting impressions’ impulse-prompting impressions (phantasiai hormetikai) 37, 200, 201, 244 use of impressions 206, 209 impulse (hormēma) 133, 164, 165, 166, 169 impulse/intention (hormē) 200, 225, 243, 244 impulses 265 in equal parts 28, 29, 34, 35 in our power (in nostra potestate) 196 n9, 269, 271 in our power 75 inanimate objects 218 incompatibility of determinism and free will 153 incorporeal 220, 221, 222 void, time, place, sayable (lekta) 221 indeterminism 20, 74, 92, 157, 167, 168 indeterminism, causal 19, 25, 31, 79, 129, 153 indeterminist free choice 77, 80, 81 indeterminist freedom 1, 2, 13, 14, 30, 31, 35, 36, 46, 49, 154 indeterminist freedom of decision 155, 215 indeterminist freedom to do otherwise 45 indeterminist two-sided conception of what depends on us 22, 33 indifferents 206 indistinguishable 229, 230, 231, 232, 233, 235 individual nature 225, 226, 227, 228, 245 individuation 218 inexorable (aparabatos) 238, 247 information 228 n49, 246, 247, 249 initial constitution (hē ex archēs sustasis) 129, 130, 131, 133, 137, 138, 141, 145, 147, 150, 161, 171, 177, 180, 185, 190 initial make-up, see ‘initial constitution’ inner-worldly perspective 218, 225, 226, 240 inseparability of human soul from body, see ‘soul, inseparability from body’ intellect (nous) 63 n38 intention 8, 206, 214 intermediate in action, the 53, 64, 70, 71 interweaving (epiplokē) 239, 240 involuntariness 57, 71 n58 Ishāq 110 : b. Hunain
Joannes XI Beccus 100 n15 judgement (krisis) 62, 63 kata prohairesin 159 knowledge (epistēmē) 247 kurios 42, 205, 278 Lambin, Denys 115 language 138

  law 73 law-breaking 73, 74 present-day criminal law 57 Law of Excluded Middle 7 laws of nature 228, 233 Lazy Argument 251 Lefèvre d’Étaples, Jacques 114 libertarians 156, 214, 215 libertarian opponents 268 life cycle 247 like causes—like effects 233, 234 live in accordance with nature 205 logic 7 Lucretius 143, 144, 165, 166, 168–79 passim, 180, 181, 182, 183, 184, 185, 186, 189 De rerum natura 168–79 passim Maghribī (script) 112 Maghribī manuscript, the 107 Manuel II Palaeologus 100 n15 Marcus Aurelius 257 Master Argument 27 master 214 ‘master of ’ (kurios + genitive) 74 materialism 12 matter (hylē) 257 maturity 139 Maximus of Tyre 41 means-end relation 58 mechanical 9 mechanical elements 217 Megarics 27 mental development 137, 139, 140, 179, 183, 184, 185–91 pre-moral 139 mental processes empirically inaccessible, non-observable 266 metaphysical controversy 231 methodological challenges 5 Metrodorus 145 Michael of Ephesus 107 Middle Platonism 2, 8, 12, 25, 26, 27, 28, 30, 31, 32, 34, 35, 36, 38, 39, 40, 41, 44, 49, 216 theory of hypothetical fate 48, 209 mind adult human 145 atomic structure of 134, 137, 140, 144, 147, 148, 149 initial nature of human mind 143, 183 maturing of 138 nature of one’s mind 143 original nature of 141 restructuring of 150

319

mind (mens) 169, 170, 171, 175 atomic structure of 177, 178, 180, 186 development of 182 individual constitution of 173 motion of 171 qua conglomerate of atoms 176 relatively stable atomic structure 190 modalities 189, 277 modal logic 28 Stoic 198 modal theory 49 moral accountability 208, 261, 267 moral appraisal 15, 40, 128, 205 moral beings 142 innate potential for becoming 139 gradual process of becoming 142 pre-moral beings 142 moral decision 142 moral development 5, 6, 128, 135–50 passim, 161, 183 pre-moral stage 135 moral improvement, see ‘moral progress’ moral progress 6, 134, 142–150 causal factors involved in 146–48 never too late 144 moral responsibility 15, 21, 46, 55, 128–151 passim, 154, 155, 156 158, 160, 163, 164, 190, 191, 192, 193, 195, 196, 201, 208, 211, 215, 216, 217, 252, 282, 288 based on ability of agent to do otherwise 129, 142 based on agent’s causal responsibility 129 based on agents qua rational agents 142 compatibility with determinism 196 morally responsible agents 250 moral role model 145 Moses Maimonides 112 motion (kinēsis) 258 movements (kinēseis) 279 Mower Argument 27 Musonius 204 n38, 209 natural motions 286 natural philosophers 159 nature 146, 217, 246 agency of 244 as a rational agent 243 can do good but no bad 251 laws of 153 nature and nurture 6, 130 nature’s awareness, see ‘divine awareness’ nature’s intention, see ‘divine intention’ no cause external to 247 of the world 245

320

 

Nature (phusis) 9 Necessary, the, (anagkaion, to) 27 Necessity 24, 29, 130, 132, 137, 138, 158, 161, 162, 163, 166, 171, 175, 181, 182, 189, 198, 200, 201, 238, 261, 262 by necessity (kat’anagkēn) 158, 159, 165, 167 external 161, 177, 185 internal 176, 177, 182–84, 185 mechanical 168 necessitarianism 21 n7 necessitation 15, 156, 180, 198 externally necessitating factors 188 internal 177, 183, 191 internally necessitating factors 188 mechanical 161 universal 161, 163, 181 negation 79 negation particle 111 Nemesius 25, 29, 35, 36 n42, 38, 40, 41, 44, 196 n7 Neoplatonism 12, 44 New Testament 122 Newton, Isaac 189 Nicolaus of Damascus (Peripatetic) 112 no external obstacles to the development of the world 246 noble or shameful 55, 90, 91 non-necessity 167, 189, 198 non-necessary 188, 190 that which is non-necessary 199 non-predeterminism 18 non-volitional (invitus) 171 not taking care 73 nourishment 245 objects of choice (prohaireta) 62, 63 determined 62 objects of deliberation (bouleuta) 59, 60, 62 one-sided causative 157, 158, 159, 160 ontological indistinguishability 231 ontologically distinct 233 order in the universe 187 Oriental Arabic 112 Origen 11, 50 origin (archē) 55 original constitution (hē ex archēs sustasis), see ‘initial constitution’ Pachymeres, George 107 par’hēmas 196 not a technical term 197 passive principle 218, 220 past, present, and future 247 perception 178

Perion, Ioachim 115 Peripatetics 2, 8, 11, 18, 19, 23, 26, 35, 45, 48, 159, 214, 215, 216 personhood 22 Peters, F. H. 123 Philodemus 187 n90 philosophy function of 151 modern 187 n90 popular philosophy 106 physicalism 187 n90 plants 188, 191, 218 Plato 2, 12, 25, 26, 38, 39, 85 Demiurge 243 Myth of Er 38, 40 Republic 38 Timaeus 242 Platonism 11, 12, 38, 48, 50 pleasure 170, 190 pleasure and pain 136 Plotinus 11, 12 Plutarch 226, 229 On Stoic Self-Contradictions 271–74 passim [Plutarch] 25, 45 On Fate 45 pneuma 11, 218, 220, 228, 237, 239, 241, 243, 256, 257, 259, 279, 281, 282, 288 politics 203 popular culture 104, 105 Porphyry 12, 107, 112, 196 n7 Posidonius 209, 219 n13, 242 possibility 198 one-sided 27, 28 possible, the 32 that which is possible 198–99 two-sided 27, 28, 30, 199 power (exousia) 36, 40, 41, 42, 43, 50 power (potestas) 174 power, in my own (autoexousion) 43 practical reasoning 59, 64, 72, 85 practical (i.e., action-related) syllogism 59 praise and blame 53, 55, 68, 69, 74, 75, 132, 150, 163, 181, 196, 198, 202, 208 allocation of 151 justified 76 praiseworthy 70, 131, 132 prayers 239 preceding causes 31 preconceptions (Epicurean) 130, 131, 138, 145, 163, 164 predetermination 10, 31, 33, 39, 46, 49, 169, 170, 171, 209, 214, 238 divine predetermination 10, 251, 252 eternally predetermined 179, 244, 248

  predetermination, freedom from 1 theological predetermination 217 predicate 219, 222, 223 actualization of 223 incorporeal 254 predictability 236 principalis causa 277 principalis 277, 278 Principle of Bivalence 29, 232 principles of universal causality 232 printing press, invention of 114 probability 29 procatarctic (prokatarktikos) 273, 283 Proclus 12 proēgoumenon 259 prohairesis 23, 37, 38, 39, 41, 207 n46 Epictetus’ concept of 49 prohairetikē dunamis 41 propositions (axiōmata) 247 proposition, dialectical 83 proto-indeterminist 79 providence 236, 237 proximus 283 psychological development 5, 161 psychological, internal factors 71 punishment and reward 196 qualitative states (scheseis) 220, 221, 222, 223, 224, 225, 226, 227, 228, 258, 259 causes of 220 qualities (poiotēs) 279, 280 quantum leaps 189 quantum mechanics 156 Queen’s College, Oxford, the 116 Rackham, H. 124 random change 186 random motion 6, 170, 179, 186, 187 rationality 192 rational beings 218 rational organizing principle 227, 228 rationally desired 63 reason 65 reason (logos) 241 reason (logos spermatikos) 244 reason of the cosmos 219 reception 93 contemporary reception of the Nicomachean Ethics 125 reductionism 187 n90 regeneration (palingenesis) 245 regret 46, 155 relative (pros ti) 254, 255 Renaissance 114, 116

321

repetition 149 reproduction 188 reservation (hupexairesis) 205 Rieckher, J. 121 Ross, W. D. 120 ruling part of soul 218, 241, 242, 249 ruling principle 219 sage, see ‘wise person’ sage, free, see ‘wise person, free’ saying ‘yes’ and saying ‘no’ 77–83 scales and dice 229, 230, 231, 234 science and theology 250, 251 coexistence of 217 two aspects of the same theory 250 scientific explanation 10, 250, 251 sea battle 31, 33 n36 seeds 244, 245 self-addressed imperative 86 self-improvement 6 self-motion 188, 230, 231 self-origination 205 self-sufficient (autoteles) 288 seminal 244 Seneca 11, 144, 205, 241, 256 sense perception 189 Simplicius 12, 41 (sita) in nobis 196 n9 slave 202, 203 slavery (douleia) 203 soul 38, 48, 55, 64, 140, 171, 172, 188, 215 deliberative part 64 desiring part 58, 64 dispositional state 63 n37 incorporeal 12, 48, 49 not separable from body 48 rational part 58 ruling part in the agent’s 58 Spectator, The 118, 123 spontaneous 231, 232 spontaneous motion 172, 229, 230 stable patterns of behaviour 188 Stahr, Adolf 122 states (scheseis) 279, 280 Stewart, J. A. 121 Stoics 1, 7–11, 18, 19, 22, 27, 33, 35, 40, 44, 48, 50, 193, 194–216 passim compatibilist position 213 continuum theory 222, 242 cosmology 217, 224 determinism 211 early Stoics 8, 9, 11, 22, 153, 196, 197–204 passim, 206, 209, 210, 212, 217–252 passim

322

 

Stoics (cont.) ethics 194, 206 late ancient Stoics 8, 152, 194 212–16 modal logic 198 philosophy 194–216 passim physics 203, 217, 218–224, 227, 228 psychology 8, 203 theory of fate 46 theory of sympathy 234 structured world (cosmos) 218, 246 Summa Alexandrinorum 112–14 sunektikos (cohesive) 278 sunergon, ambiguous between ‘helping’ and ‘holding together’ 284 sunexis (cohesion) 278 swerve 5, 6, 7, 156, 168, 169, 170, 171, 172, 173, 174, 175, 176, 178, 179, 180, 181, 185, 191, 248 argument for the existence of 168 correlation interpretation 192 deviation of (elachiston) 189 function of 185–91 frequency of 189 minimalist interpretation 192 necessary but not sufficient for autonomous agency 187 syllogism 85 sympathy (sympatheia) 239, 242 Syriac 110, 112 n69

to mē 86, 87 to nai 87 to par’hēmas 156–61 passim toddler 132, 136 torture 56 trajectories 190 tranquillity (ataraxia) 8, 137, 144, 145, 147, 151, 192, 210 transference of the moral aspect from choice to action 65–9 translation 93–127 passim word-for-word 108, 110 nineteenth century 125 translations Arabic translations 109–12 passim, exegetical translations 115 Hebrew translations 112 Latin translations 114–16, 121 medieval Latin texts 107, 108 modern-language translations (19th and early 20th centuries) 118–125 vice-versa translations 87, 92 truth (alētheia) 247, 257, 257 n13 two-sided capacities 30, 41, 42 general two-sided capacity 167 two-sided freedom 7 two-sided notion of that which is in our power 61 two-sided potestative 157, 158, 159

Taylor, Thomas 119 teaching 148 teleological 9, 197, 226, 228, 241, 249 teleological determinism 50, 224–28 passim teleological predetermination 161 teleology 238, 244 tenor 218, 220 tensional movement (tonikē kinēsis) 279 terminology, discussion of 195 that which depends on us: see eph’hēmin that which happens through us 214 that [which happens] through ourselves (to di hēmōn autōn) 130, 164, 165, 166 that which is not (to mē on) 45 the cause from ourselves (hē ex hēmōn aitia) 130 the things in our power (ta eph’hēmin) 59 n26, 60 theology 248 theological explanation 10 things in the agent’s power 59 through us (di’ hēmōn) 18, 44 thunderstorms 147, 148 time, indefinite divisibility of 222 to act (prattein) 86, 87

undetermined 32, 33 universal nature 224, 226, 227 universal regularity 232 un-predetermined 15, 229 up to us: see eph’hēmin vice 41 Victorian times 122, 124, 125 virtue 54, 66, 136 virtue and vice 90 virtues and vices 23, 35, 36, 57, 68, 75, 227 in action 67 in our power 68, 69 virtuous does not have the ability to be vicious 216 volition 22, 168, 170, 171, 173, 179, 181, 183, 186 existence of 171 faculty of 173 formation of 168 free 7, 156 human volition 6 power of 175, 176 volition for something 191 volitional motion 172, 181

  voluntas (volitional act) 165, 166, 169 act of volition vs. power of volition 169 libera voluntas 171 voluntariness 54, 57, 66, 70, 72, 155 property of actions 60 voluntary consequences of action (indirect voluntariness) 73, 74, 75, 76 voluntary (hekousios) 56, 67, 71 wanting (boulēsis) 58, 64, 66, 72 weak-willedness (akrasia) 71 what depends on us 21, 25, 29, 46 Neoplatonist conception of 44 what depends on us (eph’hemin), one-sided causative 19, 20, 21, 44, what depends on us (eph’hemin), two-sided potestative 19, 20, 21, 35, 44, 46, 49 will 41 n57, 63 faculty of the 16, 47, 49, 196

323

freedom of the 49 will (boulēsis) 64 n43 William of Moerbeke 108, 121 Williams, Robert 123 wisdom 144 practical wisdom 58, 70 wise person 203, 204 free 211 wish (boulēsis) 64 n43 without master (adespoton) 156, 158, 166, 196 n8 world, best possible 244 world reason 257 world seed 246 Zeno 197, 204 n39, 217, 219 n13, 230 n52, 236, 238, 243, 244, 257 n12, 258, 258 n21 Republic, 202 Zeus 224, 227, 236 Zwinger, Theodor 115